P H I L O SO P H IC A L F O U ND A T I O NS O F T H E L A W O F T OR TS

PHILOSOPHICAL
FOUNDATIONS
OF THE LAW
OF TORTS
EDITED BY

JOHN OBERDIEK

1
 3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
# The several contributors 2014
The moral rights of the authors have been asserted
First Edition published in 2014
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Crown copyright material is reproduced under Class Licence
Number C01P0000148 with the permission of OPSI
and the Queen’s Printer for Scotland
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2013952799
ISBN 978–0–19–870138–5
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
 For Ray Solomon
colleague, mentor, and friend
 TABLE OF CONTENTS

Contributors xii
Table of Cases xiii

INTR O DU C TIO N: PHILOS O PHICA L FOUN DA TIONS OF

T H E LA W O F T O R TS 1
John Oberdiek

PART I: FOUNDATIONS OF TORT LAW

1. TO RT L AW AN D R E S P O N S IB I L I TY 17
John C.P. Goldberg and Benjamin C. Zipursky
I. Introduction 17
II. Responsibility Theories of Tort Law 19
III. Civil Recourse Theory as a Responsibility Theory of Tort Law 26
IV. Concluding Thoughts: The Importance of Recognizing
Responsibility-Based Accounts of Tort Law 36
2. TO RT S , RIGH T S, AND R I SK 38
Stephen Perry
I. Introduction 38
II. Duty of Care and Rights 40
III. Risk, Harm, and Rights 44
IV. Harm and Fundamental Moral Rights 49
V. The Role of Reasonable Foreseeability 60
3. CO M P E N S A T I O N AS A TO R T NO R M 65
Mark A. Geistfeld
I. Introduction 65
II. Injury Compensation and Liberal Egalitarianism 67
III. A Compensatory Tort Right and the Correlative Compensatory Duty 70
IV. Compensation and Corrective Justice 79
V. Conclusion 85
4. TORT A S A S UBS TIT UTE F OR R EV ENGE 86
Scott Hershovitz
I. Introduction 86
II. Corrective Justice 89
III. Corrective Justice Corrected 92
IV. The Message in the Money 96
V. Tort or Revenge? 98
VI. Corrective Justice and Criminal Law 99
VII. Conclusion 100
viii TABLE OF CONTENTS

5. S T RU C TU RE A N D J US TI F I CA T IO N I N C O N TR A C TU AL I ST
T O RT TH E OR Y 103
John Oberdiek
I. Introduction 103
II. Introducing Contractualist Tort Theory: Keating’s Social
Contract Theory 106
III. Rawlsian Distributive Justice and the Bilateral Structure of Tort Law 108
IV. A Primer on Justification in Scanlonian Contractualism 113
V. Justification to a Subject and “the Palsgraf Perspective” 114
VI. Conclusion 121
6. O N T H E “ PROPERTY ” A N D TH E “ TORT ” I N T R E S P A S S 122
Eric R. Claeys
I. Introduction 122
II. Property in a Lockean Morality of Labor 125
III. The Unconsented-Entry Paradigm 128
IV. Affirmative Defenses 130
V. Conforming the Prima Facie Tort to the Underlying
Substantive Right 132
VI. Harm-Based Exceptions to Rights-Based Torts 135
VII. Property Foundations and Tort Implementation 137
VIII. Making Accident Torts Complement Rights-Based Torts 141
IX. Conclusion 146
7 . T O R T L A W AN D P U B L I C F U N C T IO N S 148
Peter Cane
I. Introduction 148
II. Public Institutions and Public Functions 151
III. Demarcating the Province of Tort Law 153
IV. Theorizing the Tort Law of Relationships of Juridical Inequality 161
V. Conclusion 168

PART II: HARMS, WRONGS, RESPONSIBILITY, AND LIABILITY

8. W H A T M I G H T HA V E B E E N 171
Victor Tadros
I. Introduction 171
II. The Currency of Harm and Compensation 173
III. Incomparable Problems 177
IV. Meeting Comparative Complaints 178
V. Why Time Is Not of the Essence 181
VI. In Defense of Counterfactualism 185
VII. Conclusion 191
9. W H Y R E P AR A TI O N S? 193
Rahul Kumar
I. Introduction 193
II. Two Approaches to Reparative Obligations 195
 TABLE OF CONTENTS ix

III. Rehabilitating the Compensatory Model 201

IV. Are Living African-Americans Wronged by Chattel Slavery? 204
V. Is Saying “Sorry” Enough? 210
10 . R E P A I R IN G H A R MS A N D ANS W E RI N G F O R W RO N G S 212
R.A. Duff
I. Introduction 212
II. Punishing Wrongs and Repairing Harms 212
III. A Gap—and How Civil Recourse Can Fill It 218
IV. Questions about Civil Recourse: (i) The Accounting 222
V. Questions about Civil Recourse: (ii) The Remedy 225
VI. Civil Recourse, Cost-Allocation, and Criminal Law 229
11 . T OR T PRO C E S S E S A N D R E LA TI O N A L R E P AI R 231
Linda Radzik
I. Introduction 231
II. What Stands in Need of Correction? 233
III. Damaged Relations and the Moral Obligation to Repair Them 236
IV. Forms of Corrective Justice 239
V. Contributors to Corrective Justice 242
VI. From Morality to Law 245
VII. Conclusion 248
1 2 . T OR T LI A B IL I T Y A N D TA K I N G RE S P ON S I B I L I T Y 250
David Enoch
I. Introduction 250
II. New Zealand and Apologies 252
III. Taking Responsibility: The Intuitive Idea, and Some Examples 254
IV. The Taking and the Responsibility 258
V. Tort Liability and Taking Responsibility 266
VI. Is This Enough? 270
13. E XPLORING T H E R ELATIONSHIP B ET WE EN CONSENT ,
A S SU MPT I ON O F R IS K , A N D V IC T I M N E G LI G E N CE 272
Kenneth W. Simons
I. Introduction 272
II. Legal Background 275
III. The Phenomenology and Structure of Consent 277
IV. Comparing Consent (IT) with AR 281
V. Conclusion 290
14. STRICT LIABILITY WRONGS 292
Gregory C. Keating
I. Introduction 292
II. Harm-Based Strict Liability 296
III. Right-Based Strict Liability 297
x TABLE OF CONTENTS

15 . NO R MA TI V E T H E O R IE S OF P U N IT I VE D AM A GE S : TH E C AS E
O F DE TE RR E N CE 312
Anthony J. Sebok
I. Introduction 312
II. The Law of Punitive Damages 314
III. Theories of Punitive Damages 318
IV. The Incoherence of Deterrence Theory 321
V. Conclusion 328

PART III: DISTRIBUTIVE JUSTICE IN TORT LAW

16 . WH A T IS T OR T LA W F O R? PA RT 2 . TH E PLA C E OF
D I ST RI B U TI V E J U S TI C E 335
John Gardner
I. Distributive Justice in a Corrective Context 335
II. The Distribution of Correction 338
III. Distributive Justice between the Parties 346
IV. On “Risk-Distributive” Justice 350
1 7 . T OR T LA W A N D DI S T R I B U T I V E JU S T IC E 354
Hanoch Sheinman
I. Introduction 354
II. Background 356
III. The Aristotelian Distinction 359
IV. An Alternative Account 364
V. Reparative Justice in Tort Law 371
VI. The Priority of Corrective Justice 377
VII. Similar Views 379
VIII. The Mixed Tort Law 380
IX. Conclusion 383

PART IV: SKEPTICAL PERSPECTIVES

1 8 . F I N D I N G N O F A U L T WI T H NE G L I G E N C E 387
Heidi M. Hurd
I. Introduction 387
II. Finding Fault in the Action, Not the Actor 391
III. Finding Fault in Prior Risky Actions 394
IV. Finding Fault in the Failure to Exercise Capacities of Advertence 395
V. Finding Fault in the Possession of Defective Physical, Cognitive,
or Conational Attributes 398
VI. Finding Fault in Attributes of Poor Character 400
 TABLE OF CONTENTS xi

19 . C O N F US E D C UL P A B IL I TY , C O N TR I V E D C AU S AT I ON , AND
T H E C O LL A P SE O F T O RT TH E OR Y 406
Larry Alexander and Kimberly Kessler Ferzan
I. The Target Field 406
II. Is There a Duty in the House? 407
III. Causation 416
IV. Conclusion 425

Bibliography 426
Index 443
 C O N T R I B U T OR S

Larry Alexander is Warren Distinguished Professor of Law at the University of San Diego
School of Law.
Peter Cane is Distinguished Professor of Law at the Australian National University.
Eric R. Claeys is Professor of Law at George Mason University School of Law.
R.A. Duff is Professor of Law and Russell M. and Elizabeth M. Bennett Chair in
Excellence at the University of Minnesota Law School.
David Enoch is Professor of Philosophy and Jacob I. Berman Professor of Law at the
Hebrew University of Jerusalem.
Kimberly Kessler Ferzan is Distinguished Professor of Law at the Rutgers School of Law.
John Gardner is Professor of Jurisprudence at the University of Oxford.
Mark A. Geistfeld is Sheila Lubetsky Birnbaum Professor of Civil Litigation at New York
University School of Law.
John C.P. Goldberg is Eli Goldston Professor of Law at Harvard Law School.
Scott Hershovitz is Professor of Law and Professor of Philosophy at the University
of Michigan.
Heidi M. Hurd is David C. Baum Professor of Law and Professor of Philosophy
at the University of Illinois, Urbana-Champaign.
Gregory C. Keating is William T. Dalessi Professor of Law and Philosophy at the
University of Southern California School of Law.
Rahul Kumar is Associate Professor of Philosophy at Queen’s University.
John Oberdiek is Professor of Law at the Rutgers School of Law.
Stephen Perry is John J. O’Brien Professor of Law and Professor of Philosophy
at the University of Pennsylvania.
Linda Radzik is Professor of Philosophy at Texas A&M University.
Anthony J. Sebok is Professor of Law at the Benjamin N. Cardozo School of Law
at Yeshiva University.
Hanoch Sheinman is Professor of Law and Philosophy at Bar-Ilan University.
Kenneth W. Simons is Professor of Law and the Honorable Frank R. Kenison
Distinguished Scholar in Law at Boston University School of Law.
Victor Tadros is Professor of Law at the University of Warwick School of Law.
Benjamin C. Zipursky is Professor of Law and James H. Quinn ’49 Chair in Legal
Ethics at Fordham University School of Law.
 TABLE OF CASES

532 Madison Ave Gourmet Foods Inc v Finlandia Center Inc,

96 N.Y. 2d 280 (2001) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Addie & Sons, Ltd. v Dumbreck [1929] A.C. 358 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Alberts v Schultz, 975 P. 2d 1279 (N.M. 1999) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Alcorn v Mitchell, 65 Ill. 553 (1872) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86–102 passim
Anns v London Borough of Merton [1978] A.C. 728 (H.L.) . . . . . . . . . . . . . . . . . . . . 43, 44
Ashley v Chief Constable of Sussex Police [2008] 1 A.C. 962 . . . . . . . . . . . . . . . . . . 219, 228
B.M.W. of North America, Inc. v Gore, 517 U.S. 559 (1996) . . . . . . . . . . . . . . . . . . . . 316
Baker v Howard Cnty. Hunt, 188 A. 223 (Md. 1936) . . . . . . . . . . . . . . . . . . . . . . . . . 142
Baltimore & Ohio R.R. v Goodman, 275 U.S. 66 (1927) . . . . . . . . . . . . . . . . . . . . . . . . 392
Bamford v. Turnley [1862] 122 E.R. 27 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Berry v Sugar Notch Borough, 43 A. 240 (Pa. 1899) . . . . . . . . . . . . . . . . . . . . . . . 420, 422
Beswick v City of Philadelphia, 185 F. Supp. 2d 418 (E.D. Pa. 2001) . . . . . . . . . . . . . . . . 45
Bigelow v R.K.O. Radio Pictures, 327 U.S. 251 (1946) . . . . . . . . . . . . . . . . . . . . . . . . . 297
Blondell v Consolidated Gas Co., 43 A. 817 (Md. 1899) . . . . . . . . . . . . . . . . . . . . . . . . 136
Breunig v American Family Insurance Co., 173 N.W. 2d 619 (Wis. 1970) . . . . . . . . . . . 396
Brown v Dellinger, 355 S.W. 2d 742 (Tex. Civ. App. 1962) . . . . . . . . . . . . . . . . . . 133, 135
Bryant v Lefever, 4 C.P.D. 172 (1878-79) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
Buch v Amory Manufacturing Co., 44 A. 809 (N.H. 1898) . . . . . . . . . . . . . . . 143, 144, 145
Burch v American Family Mut. Ins. Co., 543 N.W. 2d 277 (Wis. 1996) . . . . . . . . . . . . . . 30
Chicago v Troy Laundry Machinery Co., 162 F. 678 (7th Cir. 1908) . . . . . . . . . . . . . . . 138
Cleveland Park, Inc. v Perry, 165 A. 2d 485, 488 (D.C. 1960) . . . . . . . . . . . . . . . . . 135, 297
Cooper v Hobart [2001] 3 S.C.R. 537 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43, 44, 63
Cooper v Wandsworth Board of Works, (1863) 43 E.R. 414 . . . . . . . . . . . . . . . . . . . . . 166
Cooper Indus. v Leatherman Tool Group, Inc., 532 U.S. 424 (2001) . . . . . . . . . . . . 318, 327
Crook v Sheehan Enterps. Inc., 740 S.W. 2d 333 (Mo. Ct. App. 1987) . . . . . . . . . . . . . . 133
Day v Woodworth, 54 U.S. 363 (1852) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Diana Shooting Club v Lamoreaux, 89 N.W. 880 (Wis. 1902) . . . . . . . . . . . . . . . . . . . 139
Donoghue v Stevenson [1932] A.C. 562 (H.L.) . . . . . . . . . . . . . . . . . 42, 51, 54, 62, 63, 342
Dougherty v Stepp, 18 N.C. 371 (1835) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Duda v Phatty McGees, Inc., 2008 S.D. 115 (2008) . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Exxon Shipping Co. v Baker, 554 U.S. 471 (2008) . . . . . . . . . . . . . . . . . . 312, 314, 316, 329
Fay v Parker, 53 N.H. 342 (1873) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Fibre Co. v Chicago, M. & St. P.R. Co., 232 U.S. 340 (1914) . . . . . . . . . . . . . . . . . . . . . 144
Flast v Cohen, 392 U.S. 83 (1968) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Fletcher v Rylands [1865] 159 E.R. 737 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Gasperini v Center for Humanities, 518 U.S. 415 (1996) . . . . . . . . . . . . . . . . . . . . . . . 327
Glidden v Szybiak, 63 A. 2d 233 (N.H. 1949) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Hay v Cohoes, 2 N.Y. 159 (1849) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Hinman v Pacific Air Transp. Co., 84 F. 2d 755 (9th Cir. 1936) . . . . . . . . . . . . . . . . . . 136
xiv TABLE OF CASES

Hotson v East Berkshire Area Health Authority [1987] A.C. 750

(H.L.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45, 46, 47, 48
Ira S. Bushey & Sons, Inc. v U.S., 398 F. 2d 167 (2d Cir. 1968) 306 . . . . . . . . . . . . . . . 307
Isle Royale Min. Co. v Hertin, 37 Mich. 332 (1877) . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Jacque v Steenberg Homes, Inc., 548 N.W. 2d 80 (Wis. Ct. App. 1996) . . . . . . . . . . . . . 140
Jacque v Steenberg Homes, Inc., 563 N.W. 2d 154 (Wis. 1997) . . . . . . . . . . . . 133, 138, 139
Kline v 1500 Massachusetts Ave. Apt. Corp., 439 F. 2d 477 (D.C. Cir. 1970) . . . . . . . . . 145
Koepnick v Sears Roebuck & Co., 762 P. 2d 609 (Ariz. Ct. App. 1988) . . . . . . . . . . . . . 135
Konradi v United States, 919 F. 2d 1207 (7th Cir. 1990) . . . . . . . . . . . . . . . . . . . . . . . 306
Ladd v County of San Mateo, 911 P.2d 496 (Cal. 1996) . . . . . . . . . . . . . . . . . . . . . . . . 301
Leffler v Sharp, 891 So. 2d 152 (Miss. 2004) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Leichtman v W.L.W. Jacor Communications, Inc., 634 N.E. 2d 697
(Ohio. Ct. App. 1994) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100, 101, 102
Longenecker v Zimmerman, 175 Kan. 719 (1954) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Losee v Buchanan, 51 N.Y. 476 (1873) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
MacPherson v Buick Motor Co., 217 N.Y. 382 (1916) . . . . . . . . . . . 39, 40, 42, 43, 54, 60, 62
Madison v Ducktown Sulphur, Copper & Iron Co., 83 S.W. 658
(Tenn. 1904) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Marshall v Ranne, 511 S.W. 2d 255 (Tex. 1974) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Mathias v Accord Economy Lodging, Inc., 347 F. 3d 672 (7th Cir. 2003) 87 . . . . . . . . . . 317
Maye v Yappen, 23 Cal. 306 (1863) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Meistrich v Casino Arena Attractions, Inc., 31 N.J. 44 (1959) . . . . . . . . . . . . . . . . . . . . 273
Mohr v Williams, 104 N.W. 12 (Minn. 1905) 298 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Oliver v Pullam, 24 F. 127 (C.C.D.N.C. 1885) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Owens-Ill., Inc. v Zenobia, 325 Md. 420 (1992) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Palmyra v Morton, 25 Mo. 593 (1857) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Palsgraf v Long Island Railroad Co., 248 N.Y. 339 (1928) . . . . . . . . . . . . 40, 42, 43, 44, 51,
53, 54, 60, 61, 62, 63, 115, 116, 118
Paxson v Sweet, 13 N.J.L. 196 (1832) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
People v Acosta, 284 Cal Rptr. 117 (Cal. Ct. App. 1991) . . . . . . . . . . . . . . . . . . . . . . . 425
People v Decina, 138 N.E. 2d 799 (N.Y. 1956) 394 . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
People v Emmert, 597 P. 2d 1025 (Colo. 1979) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Perkins v F.I.E. Corp., 762 F.2d 1250 (5th Cir. 1985) . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Ploof v Putnam, 71 A. 188 (Vt. 1908) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131, 132, 137
Plummer v Dill, 31 N.E. 128 (Mass. 1892) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Rowland v Christian, 443 P. 2d 561 (Cal. 1968) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Sargent v Ross, 308 A. 2d 528 (N.H. 1973) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Schloendorff v Society of New York Hospital, 105 N.E. 92 (N.Y. 1914) . . . . . . . . . . . . . 298
Scott v Shepherd [1773] 96 E.R. 525 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Southern Counties Ice Co. v R.K.O. Radio Pictures, 39 F. Supp.
157 (D. Cal. 1941) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Spano v Perini Corp., 250 N.E. 2d 31 (N.Y. 1969) 83 . . . . . . . . . . . . . . . . . . . . . . . . . 141
State v Shack, 277 A. 2d 369 (N.J. 1971) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
State Farm Mut. Auto. Ins. Co. v Campbell, 538 U.S. 408 (2003) . . . . . . . . . . . . . . . . . 316
Swetland v Curtiss Airport Corp., 41 F. 2d 929 (N.D. Ohio 1930) . . . . . . . . . . . . . . . . . 136
Taber v Maine, 45 F. 3d 598 (2d Cir. 1995) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
 TABLE OF CASES xv

Tarasoff v The Regents of the University of California, 17 Cal. 3d 425 (1976) . . . . . . . . . 42

The Amiable Nancy, 16 U.S. 546 (1818) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
The T.J. Hooper, 60 F. 2d 737 (2d Cir. 1932) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
U.S. v Carroll Towing Co., 159 F. 2d 169 (2d Cir. 1947) 389 . . . . . . . . . . . . . . . . . . . . 409
Vaughan v Menlove [1837] 132 E.R. 490 . . . . . . . . . . . . . . . . . . . . . . . . . . . 388, 398, 409
Vincent v Lake Erie Transp. Co., 124 N.W. 221 (Minn. 1910) . . . . . . . . . . . . . . . . 131, 214,
224, 296, 301, 302, 303, 304, 311, 373
Vosburg v Putney, 80 Wis. 523 (1891) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
White v University of Idaho, 797 P. 2d 108 (Idaho 1990) . . . . . . . . . . . . . . . . . . . . . . 298
Zaslow v Kroenert, 29 Cal. 2d 541 (1946) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
 Introduction: Philosophical
Foundations of the Law of Torts
John Oberdiek

Philosophy of tort law, ironically, owes its preoccupations and shape, in large part, to
an altogether different discipline: economics. H.L.A. Hart and Tony Honoré’s 1959
masterpiece, Causation in the Law, was a key influence in the development of
philosophy of tort law, but it was the surge of economic analysis of tort law in the
1960s and 1970s that stimulated the reaction that accounts for so much of what
philosophy of tort law is today. Law and economics represented the culmination of
various strands in the history of torts, providing a coherent and seemingly powerful
lens through which tort law could be viewed. That efficiency and the maximization of
aggregate wealth guided economic analysis, however, revealed the approach to be a
particular, and particularly crude, version of consequentialism. Consequentialism in
its various forms had long ruled moral and political philosophy, so it was only natural
that it would penetrate the normative domain of tort law. But in the 1970s the tide
began to turn. And it was in the wake of the general revival of non-consequentialism
that George Fletcher, Jules Coleman, and Ernest Weinrib staked out non-consequen-
tialist alternatives to the economic analysis of tort law.
Their early efforts were important as first-order contributions to theoretical debates
about tort law, of course, but they also helped to substantiate philosophy of tort law’s
standing as a distinct subfield within philosophy of law. David Owen provides a
succinct history of tort theory as a discipline in his introduction to Philosophical
Foundations of Tort Law, Oxford University Press’s 1995 predecessor to the present
volume, and identifies a Law and Philosophy symposium as a signal event in the
development of the field. What was so important about that two-issue symposium
published in 1982 and 1983, according to Owen, was that philosophers, including
Fletcher, Coleman, and Weinrib, presented a unified front against the economic
analysis of tort law, and that Coleman and Weinrib presented early sketches of their
quite different corrective justice theories of tort law alongside each other.
2 JOHN OBERDIEK

The work of all three, along with Honoré’s ongoing interventions, would prove to be
extraordinarily influential. Coleman extended his criticism of the economic account of
torts in a series of papers, including a searching 1988 Yale Law Journal review of
William Landes and Richard Posner’s The Economic Structure of Tort Law and Steven
Shavell’s Economic Analysis of Accident Law, both published in 1987. In “The Struc-
ture of Tort Law,” Coleman argued that economic analysis could not make sense of the
constitutive bilateral plaintiff-defendant structure of a tort suit and, thus, had to fail.
That criticism still stands and has influenced both critics and defenders of law and
economics. Coleman presented his expansive and mature views in 1992’s Risks and
Wrongs and 2001’s The Practice of Principle, each of which is a landmark of tort theory.
Weinrib joined Coleman in his assault on law and economics, but he went further.
Weinrib’s work reached its apotheosis in 1995’s The Idea of Private Law, where he
articulated and defended a formalist theory of private law as a whole, with tort law at
its center. Weinrib agreed with Coleman that economic theories could not make sense
of the structure of tort law (or private law generally), and that such theories therefore
failed. But it was the explanation that Weinrib gave of why economic theories could
not make sense of that structure that set his account apart: economic theories failed
because they were instrumental theories. They did not respect the modes of reasoning
internal to tort law, but sought to explain or justify the law of torts by reference to an
extrinsic goal. In this respect, Weinrib’s was not so much a criticism of economic
analysis, but of any analysis that did not appraise and accept tort law on its own terms.
Non-consequentialist theories that were instrumental, then, could also run afoul of
Weinrib’s formalism. Few have adopted Weinrib’s view in all of its particulars, but his
position remains influential—and his 2012 collection of essays, Corrective Justice, is
sure to reinvigorate debate about formalism—because it emphasizes the importance of
taking seriously the intelligibility and actual of practice of tort law.
Fletcher, for his part, by and large pivoted from tort theory to criminal law theory.
Still, his seminal 1972 Harvard Law Review article, “Fairness and Utility in Tort
Theory,” remains a pole star. While Fletcher’s Rawlsian approach to the law of torts
has not gained the kind of following that Rawls’s own theory of justice has—a high
standard indeed—there is a critical mass of tort theorists who find at least some
element in Fletcher’s account illuminating. It was Fletcher, after all, who first argued
that distributive justice plays an important role in tort law, that tort law could be
interpreted in contractualist terms, and that symmetrical and asymmetrical risk
impositions needed to be distinguished from one another. Each of these planks in
Fletcher’s platform have their defenders today.
While Hart, too, turned his attention to criminal law theory and, of course, to
general jurisprudence, Honoré continued to develop his views about tort law following
the initial publication of Causation in the Law and its 1985 update. Some of Honoré’s
most important work was collected in his 1999 book, Responsibility and Fault. In those
papers, Honoré renewed his exacting inquiry into causation, but he also expounded
upon the role of luck and agency in the assignment of legal responsibility and liability
and examined the relative places of corrective and distributive justice in tort law,
 INTRODUCTION : PHILOSOPHICAL FOUNDATIONS OF THE LAW OF TORTS 3

among other topics. Honoré’s contributions to tort theory have been very wide-
ranging, as well as fertile, spawning a great deal of insightful work by others.
If Fletcher, Coleman, Weinrib, and Honoré are philosophy of tort law’s pioneers,
their many intellectual heirs now join them in leading an established field. Every
contemporary tort theorist stands on the shoulders of these four in ways obvious and
subtle. Citing some clear examples, the lineage of Gregory Keating’s Rawlsian
approach to torts can be unmistakably traced to Fletcher, Stephen Perry’s conception
of corrective justice and outcome-responsibility owes much to the work of Coleman
and Honoré, respectively, and John Goldberg and Benjamin Zipursky’s civil recourse
theory shares much in common with Weinrib’s approach, even if not its high
formalism. Like any mature scholarly enterprise, philosophy of tort law builds on
itself, relying upon and reworking the insights of those who have gone before to forge
clearer insights and better theories.
The present volume is a testament to this rich heritage. Its contributors include
Keating, Perry, and Goldberg and Zipursky, as well as other leading figures who have
made important contributions to tort theory, like Peter Cane, Heidi Hurd, John
Gardner, Mark Geistfeld, Anthony Sebok, and Ken Simons. Younger philosophers
of tort law, like Scott Hershovitz, Hanoch Sheinman, and me, also contribute. This
group is evidence enough of tort theory’s vigor. While many of the aforementioned
people also work outside of tort theory, it is a distinguishing characteristic of this
volume that the balance of contributors work primarily outside of the field. Thus,
Larry Alexander and Kimberly Ferzan, Eric Claeys, Antony Duff, David Enoch, Rahul
Kumar, Linda Radzik, and Victor Tadros work principally in some other subfield of
legal theory or philosophy, whether criminal law theory, property law theory, or moral
theory. It was an important goal of this volume that it include scholars from cognate
disciplines. And it seems to me another sign of philosophy of tort law’s vitality as a
discipline that it can reach across its borders and attract so many excellent contribu-
tions on questions of common concern from theorists whose work primarily lies
elsewhere. Such inter-disciplinarity breathes new life into tort theory’s old questions,
offers new perspectives, and also helps keep the discipline honest.
This book is divided into four parts: Foundations of Tort Law; Harms, Wrongs,
Responsibility, and Liability; Distributive Justice in Tort Law; and Skeptical Perspec-
tives. Part I: Foundations of Tort Law comprises seven chapters that address questions
revolving around the nature and aims of tort law. It is fitting that Part I and indeed the
book as a whole should open with John Goldberg and Benjamin Zipursky’s “Tort Law
and Responsibility,” as those frequent co-authors joined the field after Owen’s 1995
volume and have made some of the most important contributions to philosophy of
tort law since that time. Arguing for the centrality of responsibility to tort law, their
chapter at once valorizes and challenges Stephen Perry’s Honoré-inspired conception
of outcome-responsibility. Goldberg and Zipursky maintain that, as powerful as
Perry’s account of tort law is, their now-famous account, which they have dubbed
civil recourse theory, is more powerful still. Goldberg and Zipursky present civil
recourse theory as corrective justice’s cousin—related, but nevertheless quite different.
4 JOHN OBERDIEK

It is worth noting that many who advocate corrective justice, including Ernest Weinrib,
have challenged that characterization, suggesting that civil recourse theory is closer
to corrective justice’s fraternal twin, or even its doppelgänger. In any case, Goldberg
and Zipursky do emphasize certain features of tort law that most corrective justice
theories do not. Civil recourse theory holds that the state institution of tort law enables
private individuals, who have suffered legally cognizable harm, to seek redress against
those who are purportedly responsible for their injury. The theory self-consciously
adopts many elements central to corrective justice accounts of tort law, to be sure, but
according to Goldberg and Zipursky, what drives civil recourse theory is not correc-
tion as such, but the opportunity for recourse. Civil recourse theory highlights the role
played by a plaintiff in initiating a tort suit—the institution of tort law is not some
government agency that seeks out unjustly harmed parties and automatically com-
pensates them—as well as the fact that not all recourse comes in the form of corrective
compensation, as corrective justice supposedly maintains. In this respect, Goldberg
and Zipursky argue that civil recourse theory provides an interpretation of tort law’s
doctrines that is superior to corrective justice’s in general, and in their chapter here, to
Perry’s in particular.
Among Goldberg and Zipursky’s most prominent contributions to tort theory has
been their defense of the duty of care element in the tort of negligence. It is here that
civil recourse theory runs most closely parallel to many corrective justice theories,
which also take the duty element seriously. On this shared view, the duty element
introduces a distinctive question of law that cannot be reduced to a question of breach
of duty or a question of proximate causation: namely, whether the defendant owed a
duty of care to the plaintiff in the first place. William Prosser famously contended that
the duty element is just a placeholder for the multifarious public policy considerations
that should inform a judge’s determination of whether to allow a negligence suit to go
forward. The growing numbers of philosophers of tort law who dissent from this
instrumental view argue, in reply, that the doctrinal platitude that the duty element is
keyed to a non-instrumental conception of reasonable foreseeability is in fact defens-
ible. In his contribution, “Torts, Rights, and Risk,” Stephen Perry joins that defense.
But he also shows just how complex the defense must be, as it cannot but raise hard
questions about reasonable foreseeability, the nature of risk, and the concept of rights.
While Perry is keen to offer a suitable interpretation of the positive law of negligence,
he also seeks to clarify what he calls the “moral character” of duty.
Focusing on the duty element from the other side, as it were, Perry explores the
content of the right that anyone has not to be harmed. When one does this, on his
view, one sees that the torts of negligence and strict liability, often considered to be in
conflict, share the same moral core: the two torts are different ways of accommodating
anyone’s qualified right not to be harmed. In the case of negligence, one has a right not
to be harmed through the careless conduct of another, which correlates with a duty to
exercise reasonable care towards others. In the case of strict liability, one has a right
not to be harmed through reasonably foreseeable and seriously risky conduct of a
certain kind, which correlates with a duty not to cause “reasonably foreseeable,
 INTRODUCTION : PHILOSOPHICAL FOUNDATIONS OF THE LAW OF TORTS 5

sufficiently proximate physical harm as a result of engaging in a certain type of

activity”. This leads Perry to expand on his illuminating prior work on the nature of
harm, and ultimately, to expand on his equally illuminating and highly influential
work on the nature of risk. The notions of acting carelessly towards another and of
reasonable foreseeability depend upon a particular understanding of risk, and no one has
contributed more than Perry to the understanding of risk in a legal and moral context.
After rehearsing his well-known argument for why risks are not themselves harms in some
fundamental sense, he argues at some length against those, like me, who have criticized
different aspects of his views on risks, harms, and rights, before turning to reasonable
foreseeability itself. Here, Perry fires back at Goldberg and Zipursky, essentially arguing
that, their efforts to highlight the importance of the duty element notwithstanding, they
fail to take that element seriously enough, for they underplay the importance of reasonable
foreseeability to duty.
If different forms of responsibility figure as the lodestar of Goldberg and Zipursky’s
and Perry’s contributions, Mark Geistfeld counters that compensation is tort law’s
fundamental concern in “Compensation as a Tort Norm.” Geistfeld has long been a
leading figure in tort theory, and is distinguished in part because, though trained as an
economist, he takes seriously philosophical approaches to the subject. Indeed, Geist-
feld has faced criticisms of the economic analysis of torts head-on, and has attempted
to broker a détente of sorts between economists and corrective justice theorists. In his
contribution here, Geistfeld continues the rapprochement.
There would seem to be nothing more obvious about tort suits than that plaintiffs
bring them in order to be compensated for an injury that they have suffered. Yet,
Geistfeld notes, few tort theorists have been moved by that explanatory fact to
conclude that a norm of compensation justifies the practice of tort law. Geistfeld
seeks to establish that commonsense conclusion. But the project faces an immediate
interpretive challenge posed by the fact that negligence, and not strict liability, is the
default tort used to compensate. If a norm of compensation truly justified tort law, one
might expect all injuries to warrant compensation. Generally speaking, however, only
those injuries caused carelessly warrant compensation. Drawing on Ronald Dworkin’s
celebrated equality-of-resources theory of distributive justice, Geistfeld maintains that
exercising reasonable care in conformity with the tort of negligence distributes risk so
as not to violate anyone’s compensatory right. According to Geistfeld, the compensa-
tory norm so understood also merges with the norm of deterrence—an animating
value within many economic accounts of tort law. And the result, surprisingly enough,
is a compensatory tort right that also satisfies the demands of corrective justice.
Upon consideration, it is not especially surprising that Geistfeld’s compensation-
based theory might implement corrective justice. Compensation seems to be exactly
what the corrective aspect of corrective justice is about. Indeed, it is the close
connection between corrective justice and compensation that lends credence to Gold-
berg and Zipursky’s criticism that corrective justice is blind to tort law’s other forms of
recourse. But even they recognize that the central case of recourse within the law of
torts is compensation. Might tort law also embrace more fulsome recourse? Scott
6 JOHN OBERDIEK

Hershovitz thinks so. He makes the case that tort law accords a prominent role to
revenge, arguing in “Tort as a Substitute for Revenge” that tort suits can take the place
of revenge. It does this in a number of ways. Tort suits can displace revenge by
placating a plaintiff with money damages—a payoff of sorts. Or they can displace
the hot revenge of violence with the cooler revenge of punishment, either in the
straightforward sense reflected by punitive damages, or in the subtler sense in which
exacting payment from another necessarily constitutes punishment. Like Geistfeld,
Hershovitz casts his argument within the frame provided by corrective justice: tort
serves as a substitute for revenge, according to Hershovitz, because both serve cor-
rective justice. And just as revenge can send a message, so too does corrective justice. It
sends a message about the plaintiff ’s standing and the defendant’s responsibility.
Hershovitz goes on to defend this novel expressive conception of corrective justice
against the traditional reparative Aristotelian version.
Responsibility, compensation, and revenge are candidates for tort law’s grounding
values and concepts. In my contribution, “Structure and Justification in Contractualist
Tort Theory,” I explore and defend the possibility that a contractualist conception of
justification underlies tort law. Where George Fletcher’s and Gregory Keating’s the-
ories, for all of their important differences, converge in deploying a contractualist
framework owing to John Rawls, I look to Thomas Scanlon’s distinctive version of
contractualism. It seems to me that the pride of place that Scanlonian contractualism
accords to a particular compelling conception of moral justification, which accords
normative primacy to claimants and their claims, imbues a theory of tort law founded
on it with essential virtues. Principally, I argue that such a theory can account for tort
law’s primary and not merely secondary obligations, unlike corrective justice or civil
recourse theories, as well as make sense of the definitive bilateral plaintiff-defendant
structure of tort suits, unlike either economic or Rawlsian contractualist theories.
Beyond these interpretive virtues, though, the theory I advance is unapologetically
moral—I offer it as a justificatory account to rival economic accounts, highlighting the
sense of wronging at its core. Though my contribution only introduces an overlooked
contractualist understanding of tort law, I hope that in my presentation it is plausible
enough that others might join me in exploring the theory’s promise.
I assume in my chapter that any plausible theory of tort law must accord priority to
primary over secondary obligations and make sense of the bilateral structure of a tort
suit. In “On the ‘Property’ and the ‘Tort’ in Trespass,” Eric Claeys pursues related
questions, examining the possibility that rights-protection is more fundamental to
normative justifications for tort, while corrective justice is more important for explain-
ing tort’s structure and organizing concepts. But he does so in an especially creative
way. Most who work principally in tort theory take negligence to be the signal tort, and
assume that what there is to understand about tort law can be best understood by
unpacking the tort of negligence. Claeys is first and foremost of scholar of property
law, however, and he brings to bear his special insight into property torts to illuminate
the terrain. Just as switching from a road map to a topographical map can make
familiar territory new, Claeys’s focus on the interplay between substantive property
 INTRODUCTION : PHILOSOPHICAL FOUNDATIONS OF THE LAW OF TORTS 7

rights in land, on the one hand, and various actions and defenses protecting those
rights, on the other, is at once alien and interesting. Ultimately, he defends the
possibility he sets out to examine: property torts do indeed effect corrective justice,
but they also depend upon an antecedent account of substantive property rights.
In the final chapter of Part I, Peter Cane ventures into territory that truly is
uncharted. If the signal tort is negligence, then the paradigm tort suit involves one
private party suing another. A great deal of theory has been spilled trying to delineate
and appreciate the nature and importance of the juridically equal relationship between a
private plaintiff and a private defendant. And yet, in “Tort Law and Public Functions,”
Cane points out that next to nothing has been written about juridically unequal
relationships, which are hardly peripheral to tort law. A juridically unequal relationship,
as Cane explains it, holds between a private plaintiff and a public agent acting on the
public’s behalf—basically, a citizen versus the state. If one desideratum of a tort theory is
that it be able to make sense of the bilateral structure of a tort suit, then according to
Cane, another is that it be able to accommodate juridically unequal relationships. It is
hard to argue with Cane on this point, which makes it all the more embarrassing that
tort theory has largely overlooked the importance and distinctiveness of such actions.
Wisely prescinding from debate between non-instrumental and instrumental theories,
Cane seeks to show how each theoretical family might make room for relationships of
juridical inequality—his goal is refreshingly non-partisan. Regarding non-instrumental
accounts, Cane maintains that our conception of justice will have to be developed, for
the justice of interactions between a citizen and the state will have a different cast than
justice between two private citizens. Instrumental accounts, also, will have to be
reconceived. For example, instead of focusing on the incentives that a given liability
regime creates to guide private parties towards efficient behavior, instrumental theories
should highlight how tort law can be used as a tool to promote accountability, and thus
legitimacy, in the exercise of state power. Cane’s suggestions on this score are sensible.
More important, though, is the new adequacy condition that he proposes: it is no longer
good enough for a theory of tort law to dwell only on the juridically equal relationship
between private parties. As the ever-growing administrative state increases the scope of
government power, and thus the opportunity for the state to exploit its citizens, Cane’s
proposal should be taken very seriously indeed.
Where Part I takes up foundational questions about tort law as such, the eight
chapters in Part II focus on fundamental questions within tort law. As the compound
title, Harms, Wrongs, Responsibility, and Liability attests, the questions explored here
revolve around the nature and significance of harm, wrongdoing and the responses it
warrants, and the conditions and limits of responsibility and liability. Victor Tadros
opens with “What Might Have Been,” a characteristically subtle inquiry into the nature
of harm. If tort law is going to compensate aggrieved parties for the harm they have
suffered, then it is crucial that the conception of harm assumed by tort law’s compen-
satory aim be sound. Tadros powerfully defends a counterfactual account of harm: an
event harms a person if that event renders the person worse off than she would have
been in some other possible world—hence the title of his chapter. But Tadros
8 JOHN OBERDIEK

introduces an important wrinkle. On his view, the possible world that serves as a
benchmark is not necessarily the world in which the event would not have occurred,
for other possible worlds may be relevant. Arguing against absolutist and temporal
accounts of harm, Tadros also fends off the charge that over-determination cases
undermine his counterfactual theory. He actually reverses the threat. On Tadros’s
view, only a counterfactual theory offers an adequate account of cases where more
than one cause would be sufficient to bring about some harm.
Tadros assumes that harm triggers compensation. In “Why Reparations?,” Rahul
Kumar expresses doubts about that assumption. (It is independently interesting that
both Tadros and Kumar, each of whom work primarily outside of tort theory, assume
a conception of tort law like the one that Geistfeld defends.) Kumar approaches the
issue from an interesting angle, by querying the normative basis of claims for historical
intergenerational injustice. He argues that it is not harm that triggers reparations, but
wrongdoing understood in a particular way. On his view, it is a serious distortion to
characterize claims for reparations as calls for compensation for harm done. A harm-
based view, which Kumar assimilates to tort law’s approach, recognizes the import-
ance of the fact that what was done matters morally, but it mistakenly emphasizes the
harm that was done. The chief alternative to a harm-based view, emphasizing recon-
ciliation, however, overlooks the importance of the past, according to Kumar. Justice is
inherently backward-looking, but reconciliation looks to the past only instrumentally
to determine what would make (essentially forward-looking) reconciliation most
effective. Still, Kumar notes that the reconciliation-based approach recognizes the
importance of people standing in a relation of civic trust to one another. So, Kumar
opts for a middle way, which takes on board the sound elements in each of the two
rival accounts: the basis of reparations claims lies in how some wrongdoer has related
to some wronged person or peoples. Kumar thus favors a normative basis of repar-
ations that appropriates a harm-based emphasis on responding to what was done in
the past, and a reconciliation-based emphasis on the importance of the relation in
which people stand to one another. He goes on to illustrate the merits and reach of his
theory by discussing the basis of African-American claims to reparations for slavery.
Antony Duff ’s contribution fruitfully intersects with Tadros’s and Kumar’s discus-
sions. Like Tadros and Kumar, Duff works primarily outside of tort theory, and like
Tadros, he works primarily in criminal law theory—he is, in fact, one of the world’s
leading criminal law theorists. In “Repairing Harms and Answering for Wrongs,” Duff
considers civil recourse theory as developed by Goldberg and Zipursky, or “Gold-
ursky” as he refers to them, and relates the conception of tort law that it underwrites to
criminal law. Duff applauds the ability of civil recourse theory to get beyond compen-
sation for harm, and to enable wronged persons to hold to account those who have
wronged them. This is not surprising, as Duff can surely see his own important views
about criminal law and the role of criminal trials finding an independent footing
within tort theory. But that is the rub for civil recourse theory. For Duff argues that a
civil recourse process must mirror the criminal process much more that Goldberg and
Zipursky allow. From such a perspective, tort law appears to be unduly permissive:
 INTRODUCTION : PHILOSOPHICAL FOUNDATIONS OF THE LAW OF TORTS 9

third-party insurance coverage, which is the tail that wags the tort law dog, as well as
pre-trial settlements that preempt the vast majority of civil trials, for example, would
seem to undermine the efficacy of authentic civil recourse. At the same time, tort law
seems unduly restrictive: criminal law’s wide variety of justifications and excuses
would seem to deserve a place within tort law that they presently lack. Duff ’s attention
to the details of civil recourse theory reveals a number of questions that Goldberg and
Zipursky and those who follow them must confront, and it also stands as proof that
tort theory is not for tort theorists alone.
Just as Duff shows how civil recourse theory relates to criminal law and exerts
pressure on the traditional doctrines and boundaries of tort law, Linda Radzik argues
that corrective justice resonates well beyond the confines of tort law in “Tort Processes
and Relational Repair.” According to Radzik, a reconciliation-based conception of
corrective justice whose home lies outside of tort law is far more dynamic than the
variety officially recognized by tort law. Once it is recognized that corrective justice is
dynamic in this way, playing roles in myriad social, political, and moral debates, the
version on display in tort law looks exceedingly pale and cramped by comparison.
Radzik goes so far as to argue that the contribution of tort processes to corrective
justice is actually quite limited. Worse still, if founded on corrective justice as so many
contend, then tort law is almost perverse. For on Radzik’s view, if one has to resort to
tort processes for corrective justice, then true repair—relational repair—is likely
impossible.
The fuller form of corrective justice represented by Radzik’s notion of relational
repair suggests that corrective justice, in at least certain of its incarnations, is a value
that we should not wish to live without, even if tort law fails to offer a variety of
corrective justice that is particularly attractive. But perhaps tort law, and specifically
tort liability, offers something worth wanting after all. David Enoch tentatively believes
that it does. In “Tort Liability and Taking Responsibility,” he maintains that tort liability
has moral merit, at least in part, because it enables people to take responsibility—that is,
to actively assume responsibility where one does not already bear it. Tort liability, on
this view, thus presupposes a (literally) powerful conception of human agency, one in
which it is possible to make it the case that one is responsible for another’s injury.
There is much to be said in favor of the capacity to take responsibility as Enoch
understands it, and that value should give one pause before one rejects tort law for
having no moral merit. On Enoch’s view, however, it remains an open question what
to do after that pause. A society without tort law will lack the particular way of taking
responsibility that tort law enables, to be sure, but it does not follow that such a society
will necessarily lack other adequate ways of taking responsibility. In this, Enoch’s
position resembles Coleman’s, who has voiced something like the following qualified
defense of corrective justice: the form of personal responsibility that tort law institu-
tionalizes is a value that ought not to be forsaken, such that if society abandons tort
law, it had better find some other way of valorizing personal responsibility.
Personal responsibility and the idea of taking responsibility also figures in “Explor-
ing the Relationship between Consent, Assumption of Risk, and Victim Negligence,”
10 JOHN OBERDIEK

by Ken Simons. Where Enoch and those who share his sense of the importance of
responsibility usually have in mind the responsibility of one who caused harm to
others, Simons focuses on the responsibility for harm that is born by the person who
suffered harm. Thus, he meticulously analyzes consent to harm, assumption of risk,
and comparative fault, as well as their interconnections. Simons argues that the
concept of consent differs in certain respects from assumption of risk, but that they
nevertheless share a common core. He also maintains that both are complete defenses:
they outright preclude and do not merely diminish damages, as under comparative
fault. At the same time, however, assumption of risk rarely gains traction as a defense,
while consent regularly does. This, Simons argues, is due not to any deep conceptual or
normative difference between the defenses, but to significant factual differences
between the most common scenarios in which each arises. Specifically, in standard
cases of consent but not assumption of risk, the plaintiff and defendant each benefit
from the harmful interaction, or at least the plaintiff justifiably relies on the defend-
ant’s apparent consent.
Simons’s contribution underscores the fact that those who suffer harm are not
necessarily true or innocent victims. For they can be responsible, innocently or not, for
their own injuries. As this much makes clear, there are limits to a defendant’s liability
that have nothing to do with the defendant’s own conduct. Of course, there are also
defendant-derived limits to liability: some kind of fault is usually a condition of
liability for harm. This is a consequence of the centrality of the tort of negligence in
the wider constellation of torts. Weinrib is probably the most famous contemporary
philosopher of tort law to argue that fault must be a condition of liability. In throwing
down this gauntlet, he stands against the defensibility of strict liability. Strict liability
appears to lack any justification if one approaches it, as Weinrib does, assuming that
wrongdoing requires fault. In his contribution here, “Strict Liability Wrongs,” Gregory
Keating forcefully dissents from such an assumption. Keating argues that strict liability
torts are genuine wrongs that violate rights. In making this case, he distinguishes
between two domains of strict liability wrongs: harm-based wrongs and sovereignty
wrongs. One commits a harm-based wrong that triggers liability if one harms another
but fails to redress it. For example, if one engages in an ultra-hazardous activity like
dynamiting a quarry that injures someone, even after taking great care not to cause
harm, then one is duty-bound to compensate the injured person and wrongs them if
one fails to do so. Keating maintains that one commits a sovereignty tort if one violates
the fundamental autonomy rights of another, as when one uses another’s property
without permission. Even if one uses the property for shelter from a vicious storm, and
thus uses the property justifiably, one nevertheless wrongs the property owner, for one
violates the property owner’s sovereignty. Keating, it is clear, defends a more expansive
conception of liability than many are willing to countenance. Keating has his reasons,
to be sure: limiting the liberty of would-be defendants is a lower price to pay than
sacrificing the safety of would-be plaintiffs, especially when insurance is readily
available to would-be defendants.
 INTRODUCTION : PHILOSOPHICAL FOUNDATIONS OF THE LAW OF TORTS 11

A regime of strict liability clearly expands the liability of those subject to it. But one’s
liability can be expanded along a different dimension as well. Just as fault-based
liability can be supplemented by no-fault strict liability, compensatory damages can
be supplemented by punitive damages. The final chapter in this section, by Anthony
Sebok, addresses this latter expansion. In “Normative Theories of Punitive Damages:
The Case of Deterrence,” Sebok sizes up the deterrence argument for punitive dam-
ages. Contrasting tort law’s punitive damages with the full-throated punishment that
public law recognizes, Sebok makes a pair of key observations. First, the technical
resources necessary to achieve deterrence are unavailable in common law punitive
damages, making it unlikely that punitive damages could achieve their putative goal.
Second, the common law of punitive damages eschews any appeal to political legit-
imacy to justify its choice of punishments. This, Sebok concludes, is a shortcoming.
The chapter concludes with the revisionist claim that political legitimacy, rather than
deterrence, should be foregrounded as the justification of punitive damages at com-
mon law.
Parts III and IV are shorter, including just two chapters apiece. Part III, Distributive
Justice in Tort Law, includes a pair of sequels. John Gardner’s “What is Tort Law For?
Part 2. The Place of Distributive Justice,” follows up on his already influential 2011
Law and Philosophy article, “What is Tort Law For? Part 1. The Place of Corrective
Justice,” while Hanoch Sheinman’s “Tort Law and Distributive Justice,” follows up on
his excellent Law and Philosophy article from 2003, “Tort Law and Corrective Justice.”
Questions about the role of distributive justice in tort law have long been a part of
debate in tort theory, even if debate surrounding corrective justice has garnered more
attention. Fletcher, Coleman, Weinrib, and Honoré, among others, have all weighed
in, in quite different ways, on the place of distributive justice in tort law. Here, Gardner
and Sheinman each offer an account of tort law’s special distributive role—the
distributive role that tort law cannot fail to play. Gardner identifies two ways that
tort law is concerned with distribution: it distributes rights to corrective justice between
classes of people, and it distributes the loss between the parties in a tort suit. Though he
recognizes ways in which tort law implicates distributive justice, however, Gardner also
recognizes that corrective justice remains supreme. For what these two ways have in
common is that they are both incidental to tort law’s primary corrective goal. Shein-
man disagrees in an interesting way with Gardner. Instead of seeing distributive justice
as subsidiary to corrective justice, Sheinman actually identifies tort law’s special
distributive role with its corrective role. On his view, tort law’s distributive justice
just is its corrective justice. In this way, Sheinman’s account rejects the prevailing
assumption that corrective and distributive justice must be distinguished, for he rejects
the widespread view that the two forms of justice are categorically distinct.
Part IV, Skeptical Perspectives, includes two chapters that raise troubling questions
about the tort of negligence and tort law as a whole, respectively. Heidi Hurd’s
contribution, “Finding No Fault with Negligence,” concludes that negligence liability
is, despite appearances, a species of strict liability. Hurd begins by arguing that
carelessness is not, in fact, blameworthy. She next equates fault with blameworthiness.
12 JOHN OBERDIEK

And on the force of this identification, Hurd reaches her startling conclusion: the tort
of negligence is actually a no-fault tort, entailing that negligence liability is strict.
Hurd contends that this presents the many corrective justice theorists who object to
imposing liability without fault with the following trilemma: (1) they must give up
their conviction that tort law should correct injustices; or (2) they must give up their
conviction that injustices occur only when blameworthy persons cause harms; or (3)
they must denounce negligence liability and work towards the adoption of a tort
system that maps civil liability onto conditions of genuine moral fault.
Hurd does not jettison all of tort law, she just dislodges its keystone. Frequent co-
authors Larry Alexander and Kimberly Ferzan take on the bigger job. They witheringly
argue that tort law is an ad hoc collection of doctrines that are either incoherent or
normatively indefensible or both. According to Alexander and Ferzan, tort law’s
mistakes begin with the conception of duty at its heart. There is no duty not to
cause harm, for harm is reciprocally caused—every plaintiff necessarily plays a causal
role in his or her own injury. Nor is there a duty not to cause harm through faulty
conduct. Sharing Hurd’s view, Alexander and Ferzan hold that there is no fault in not
adverting to some risk, and thus that negligence liability amounts to strict liability. The
only duty that Alexander and Ferzan recognize is a duty not to cause harm through
culpable or blameworthy conduct. And yet even this duty must be revised, for one can
be blameworthy without causing harm. This leads Alexander and Ferzan to conclude
that tort law errs in requiring the causation of harm. This in turn leads them to
catalogue the myriad supposed flaws in tort law’s understanding of both actual and
proximate causation. If tort law has any legitimate aims, Alexander and Ferzan
conclude that they would be better met by some other institution. For my part, I am
skeptical that Hurd’s and Alexander and Ferzan’s skepticism is warranted. Theirs are,
however, powerful challenges that demand answers that will exercise anyone who
seeks to defend tort law traditionally conceived.
This volume should enrich the study of torts and the broader issues that tort law
implicates for some time to come. It is my hope that it will come close to the
contribution made by Philosophical Foundations of Tort Law, which deserves a
prominent place in any update to Owen’s thumbnail history of tort theory. If it
achieves that influence, it will represent a remarkable turnaround. For this book was
literally born under a cloud. It grew out of a two-day conference in November 2012
sponsored by the Rutgers Institute for Law and Philosophy and held at the Rutgers
School of Law in Camden, New Jersey, which convened just days after Superstorm
Sandy had ravaged that state and the surrounding region. Amazingly, the storm forced
just one cancellation, though it threw a spanner into the travel plans of many others,
whether they were traveling from across the globe or down the New Jersey Turnpike.
Still, the show did go on. And as the chapters here suggest, it was a great success.
I am grateful to the presenters and other participants for the high quality of the
discussion at the conference. But as anyone who has organized a conference knows, it
is not merely the participants who deserve credit for a successful scholarly event.
Conference planner extraordinaire Carol Shaner handled a wide variety of crucial
 INTRODUCTION : PHILOSOPHICAL FOUNDATIONS OF THE LAW OF TORTS 13

details expertly and efficiently and was, as always, a pleasure to work with. I also
enjoyed broader institutional support, the importance of which cannot be overstated.
Dean Rayman Solomon deserves thanks from the community of scholars who con-
tributed to and will benefit from this volume. No member of any faculty could ask for a
better leader or a more ardent supporter of a law school’s scholarly mission. And I am
as lucky to count him as a friend as I am to call him my dean. I dedicate this volume to
him.
I owe thanks to others for their roles in bringing the resulting volume to publication.
I am grateful to Thomas Benton, Justin Corbalis, and especially Michael Edelman for
their editorial skill and overall dependability in assisting me with the final product.
Their work was tedious but essential. I also thank Alex Flach for initiating this volume
and entrusting it to me, Natasha Flemming and Clare Kennedy for shepherding it
through many phases of the publication process with so much patience, and Erin
Pearson for seeing it through to press. Last but certainly not least, I owe my wife, Patty,
and my three kids, Sophie, Lucas, and Greta, my infinite gratitude for their own
patience with this project.
 PART I

FOUNDATIONS OF
TORT LAW
 1
Tort Law and Responsibility
John C.P. Goldberg and Benjamin C. Zipursky*

I. Introduction
When a court enters a judgment for a pedestrian who has sued a driver for negligence,
it holds the driver responsible to the pedestrian. The basis of the driver’s being held
responsible, obviously, is the driver’s careless injuring of the pedestrian. The same is
true for a judgment entered against a manufacturer on a claim by a consumer who is
injured by a poorly designed product, for a private citizen defamed by a magazine, for
an investor defrauded by a swindler, and for a child molested by a caretaker. In all of
these cases, the plaintiff has suffered an injury because of the defendant’s wrongful
conduct, and she demands that the court hold the defendant responsible for that
injury. In ordinary parlance, the defendant must compensate the plaintiff for her
injury because it was the defendant’s fault.
Tort law is in the foregoing senses a law of responsibility. It allows for persons to
be held responsible (or accountable) for having wrongfully injured others. When
lawyers say that tortfeasors are “subject to liability,” they mean that, in light of what
the tortfeasor did and the injury suffered by the plaintiff, the tortfeasor is vulnerable to
being held responsible or accountable to the victim through the court system.
Much of the debate in contemporary tort theory has pitted corrective justice theory
against efficient deterrence theory. Among the problems with this framing is that it
deflects attention away from what should be the main contenders in this domain,
namely, responsibility-based theories of tort law. To be sure, there are many overlaps
between responsibility-based theories and corrective justice theories. But the respon-
sibility view is not best described as a version of corrective justice theory. This is for
both positive and negative reasons. As to the positive: responsibility and accountability
are the concepts at work on the face of the law; “justice” is not. As to the negative: it

* Thanks to John Oberdiek for arranging this volume and for organizing the conference at Rutgers School
of Law from which it grew. We received many helpful comments from our fellow conferees, and especially our
commentator Rahul Kumar, as well as from participants in the Notre Dame Law School Faculty Workshop.
Remaining errors are ours.
18 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

is frequently untrue in a tort judgment that any kind of correction really occurs, and
it is frequently the case that the judgment better conforms to a notion of who is
responsible than it does to the more aspirational idea that the defendant must pay in
order for justice to be achieved.
Responsibility-based views take many forms. Particularly influential versions have
been offered by scholars from the UK, including Honoré and Hart,1 as well as Strawson.2
(Contemporary corrective-justice theory, by contrast, has been developed most fully by
North American scholars including Epstein, Coleman, and Weinrib, who in turn have
relied on Lockean, Aristotelian, and Kantian political and moral theory.3) A wide range
of contemporary theorists probably are rightly treated as responsibility theorists, includ-
ing self-proclaimed members of the corrective justice camp such as Coleman, Arthur
Ripstein,4 and Stephen Perry,5 but also Peter Cane,6 William Lucy,7 Martin Stone,8 and
ourselves.
In what follows, we first outline Perry’s impressive effort to craft a responsibility-
based account of tort law. We invoke it to demonstrate the ability of responsibility
theories to capture basic features of, and important modern developments in, tort law.
Along the way, we contrast it briefly with views of tort that are problematic either
because they fail to give responsibility a central place, or because they draw too close a
connection between holding persons responsible and doing justice.
Having invoked Perry’s theory to establish the plausibility and value of viewing tort
law in terms of responsibility, we next argue that the account that we have developed
under the banner of “civil recourse theory” provides a better version of responsibility
theory. Tort law is best understood as law that defines duties not to injure others and
leaves those who have breached such duties vulnerable to their victims’ demands for
responsive action.
By way of conclusion, we offer some thoughts on why it is especially important
today to recognize the central place of responsibility in the law of torts.

1
See, e.g., H.L.A. Hart and Tony Honoré, Causation in the Law (Oxford: Clarendon Press, 2d ed. 1985),
130–307; Tony Honoré, “Responsibility and Luck,” 104 Law Q. Rev. 530 (1988).
2
Peter Strawson, “Freedom and Resentment,” 48 Proceedings of the British Academy 1 (1962). Obviously
Strawson was not a tort theorist. His philosophical discussion of responsibility nonetheless has been broadly
influential among moral and legal theorists interested in exploring legal responsibility.
3
Richard A. Epstein, “A Theory of Strict Liability,” 2 J. Legal Stud. 151 (1972); Jules L. Coleman, Risks and
Wrongs (New York: Oxford University Press, 1992), 197–385; Ernest J. Weinrib, The Idea of Private Law
(Cambridge, MA: Harvard University Press, 1995).
4
See, e.g., Arthur Ripstein, “As if it Never Happened,” 48 Wm. & Mary L. Rev. 1957 (2007).
5
Stephen R. Perry, “Responsibility for Outcomes, Risk, and the Law of Torts,” in Gerald Postema (ed.),
Philosophy and the Law of Torts (Cambridge: Cambridge University Press, 2001), 72–130 [hereinafter Perry,
Responsibility for Outcomes]; Stephen R. Perry, “The Moral Foundations of Tort Law,” 77 Iowa L. Rev. 449
(1992) [hereinafter Perry, Moral Foundations].
6
Peter Cane, Responsibility in Law and Morality (Oxford: Hart Publishing, 2002).
7
William Lucy, Philosophy of Private Law (Oxford: Oxford University Press, 2007).
8
Martin Stone, “The Significance of Doing and Suffering,” in Gerald Postema (ed.), Philosophy and the Law
of Torts (Cambridge: Cambridge University Press, 2001), 131–82.
 TORT LAW AND RESPONSIBILITY 19

II. Responsibility Theories of Tort Law

A. Perry on responsibility in tort law
Responsibility theories of tort make two basic claims. First, for any successful tort
claim in which a defendant is deemed liable to a plaintiff, the court is holding the
defendant responsible to the plaintiff. Second, the defendant’s responsibility rests on
the defendant’s being responsible for having injured the plaintiff.9 Obviously these
claims are quite general, which is why responsibility theories can take different forms,
depending on how they describe the terms on which the defendant is held responsible
to the plaintiff and the grounds on which the defendant is deemed eligible for bearing
that responsibility.

9
John Finnis has urged us to discuss how our usage of “responsibility” and its cognates relates to Hart’s
famous taxonomy, which identifies four aspects of the concept: (a) role-responsibility, (b) causal-responsibility,
(c) liability-responsibility, and (d) capacity-responsibility. H.L.A. Hart, Punishment and Responsibility
(Oxford: Oxford University Press, 2d ed. 2008), 210–30.
Hart had relatively little to say about iterations (a), (b), and (d). As he defined it, the term “role-
responsibility” refers to substantive duties that attend certain reasonably well-defined roles. To use his
example, a ship’s captain incurs a responsibility for the safety of his ship by virtue of assuming the position
of captain. Tort law’s duties of non-injury (and with them, the possibility of liability for injury) are often role-
dependent or relationship-dependent. For this reason, any plausible account of tort law will incorporate
notions of role-responsibility. However, these kinds of duties are not our primary focus here.
“Causal-responsibility” refers to uses of the term “responsibility” that provide normatively spare or agnostic
descriptions of cause-effect relationships, as in the sentence: “Wilt Chamberlain was responsible for the
National Basketball League changing its rules for free throws.” A great deal of confusion stems from conflating
causal-responsibility with the normatively richer notion of a person being responsible for having caused some
state of affairs. (Indeed, Hart in his work on causation with Honoré was arguably guilty of such conflations.) To
avoid these problems, we try to avoid using “responsibility” and its cognates as synonyms for “causation” and
its cognates, and we follow that practice here.
“Capacity-responsibility” refers to characteristics that render a person eligible for attributions of responsi-
bility, including faculties of reason, self-control, and the like. Any account that depicts tort law as law that holds
actors responsible for injuries caused to others presupposes some conception of capacity-responsibility. It is
not our present concern to specify that conception.
Hart gave a bit more attention to the idea of “liability-responsibility.” To the extent it is helpful to situate this
chapter and our work in relation to his, it is fair to say that our main concern is also with this form of
responsibility. This is hardly surprising. As he suggested by his choice of label, Hart argued that liability-
responsibility is closely connected to concepts of legal liability and moral blame. Determinations of liability-
responsibility, he argued, require attention to the elements or components of liability and blame, including the
mental or psychological attributes of an actor’s actions, the possible causal connection between the conduct of
the actor and the injury for which the actor is potentially liable or blameworthy, and the relationship between
the actor and whomever or whatever brought about the injury in question.
As we note below, Stephen Perry’s account of responsibility, on which we will focus in this chapter, builds
directly on the work of Tony Honoré, who, famously, was Hart’s co-author, and who developed an account of
liability-responsibility roughly along these Hartian lines. We, like Perry, suppose that in the vicinity of what
Hart called “liability-responsibility” there is a familiar usage of the term “responsibility”—one that alludes to or
connotes a connection between actor, action, and injury, and does so as part of an inquiry into the propriety of
deeming an actor liable or to blame. What those connections are (on our view) and how this view of these
connections relates to the ideas of Perry (and, to a lesser extent, Honoré and Hart) are sketched below.
20 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

In his superb article on “The Moral Foundations of Tort Law,”10 Stephen Perry
offered a powerful version of responsibility theory. Building on the work of Honoré,
Coleman, and Weinrib, Perry maintained that tort law holds a defendant responsible
to the plaintiff in the particular manner of enforcing the defendant’s moral duty to
repair the plaintiff ’s loss. That duty is ultimately grounded, in significant part, in the
defendant’s being “outcome-responsible” for that loss.
Roughly speaking, a person is outcome-responsible for a loss if the person’s
volitional action was a necessary condition for the loss’s having occurred, and if the
loss was avoidable, in that the person could reasonably have foreseen that his action
might cause the loss, and the person was capable of acting so as not to cause it.11
Critically, for any given loss, there can be more than one outcome-responsible actor,
including the victim herself.12
Perry calls this a “volitionist” conception because it centers on the idea that: (a)
volitional action creates an agency connection between a person and certain outcomes,
and (b) that agency-connection generates agent-relative reasons for action.13 More is
needed, however, to travel the full distance from the notion of outcome-responsibility
to a defendant’s being required, through a tort-based legal obligation, to heed his
moral duty to repair another’s loss. Outcome-responsibility generates moral reasons
for action in relation to an outcome. But it is a relatively thin conception of respon-
sibility. Again, one is outcome-responsible for a loss merely by virtue of having caused
it where it was avoidable. Fault is not required. (Indeed, as Perry notes, Honoré
originally invoked the concept of outcome-responsibility as part of an effort to provide
a moral justification for strict liability.14)
Correspondingly, outcome-responsibility often generates moral reasons for action
falling short of reparation. If I carefully back my car out of my driveway, but
nonetheless strike my neighbor’s trashcan, I am outcome-responsible for its being
knocked over and dented. My having been the one who collided with it gives me a
moral reason to do something with respect to that outcome. However, given that I was
driving carefully, it may be that I incur nothing more than a responsibility to pick up
the can and place it in an appropriate spot. Making a difference in the world gives rise
to reasons to take further actions, but the particular actions I have reason to take will
depend on additional considerations.
According to Perry, to move from outcome-responsibility to a moral duty of repair
requires the invocation of a distinct set of “distributive” considerations.15 Suppose a
pedestrian is crossing a residential street and a driver runs into him, knocking him
over and breaking his arm. Because collisions of this sort are, unfortunately, com-
monplace, the scenario was reasonably foreseeable to both the driver and the pedes-
trian, each of whose actions (we can suppose) were necessary for the loss to occur.

10
Perry, Moral Foundations (note 5). We recognize that Perry has developed his account of torts in
subsequent writings. See, e.g., Perry, Responsibility for Outcomes (note 5). We nonetheless focus on this
articulation as a particularly clear and powerful statement of a responsibility-based account of tort.
11 12
Perry, Moral Foundations (note 5) at 505. Perry, Moral Foundations (note 5) at 498.
13 14
Perry, Moral Foundations (note 5) at 507, 513. Perry, Moral Foundations (note 5) at 491.
15
Perry, Moral Foundations (note 5) at 509–10.
 TORT LAW AND RESPONSIBILITY 21

Thus the pedestrian’s loss is one for which both the pedestrian and the driver are
outcome-responsible. If the pedestrian were to sue the driver in negligence, she would
be seeking the state’s assistance in holding the driver responsible. The law of negli-
gence, however, will require the pedestrian to establish that the driver was at fault. It
does so, Perry maintains, as a way of answering the question of which of these two
outcome-responsible persons should, in fairness, bear the cost of the broken arm. The
plaintiff ’s proof that the defendant was driving carelessly (and that she was crossing
carefully) establishes that the defendant is not only outcome-responsible for the loss,
but also morally responsible, such that he now has reason to respond to the loss by
indemnifying the plaintiff.
In sum, according to Perry, when we assess whether an actor is morally responsible
for a loss, we are seeking to ascertain the answer to a question of “localized distributive
justice”—of who in fairness should bear a given loss—by reference to the relative
faultiness of the conduct of all outcome-responsible persons.16 When tort law deter-
mines legal liability by reference to this criterion, it is doing so because it is a scheme
for holding actors to the moral responsibility to repair that they incur by virtue of
being both outcome-responsible and at fault for a given loss.
It is worth noting one other aspect of Perry’s analysis. He argues that an appreci-
ation of the precise role that fault plays in the legal analysis of a claim such as the
pedestrian’s enables one to grasp why legal “fault,” though a moral concept, departs to
some degree from notions of culpability or blame that tend to attach to notions of
fault. Because fault is being invoked to resolve the distributive question of who should
bear a given loss and because the universe of candidates eligible to bear that loss is
determined by outcome-responsibility, it is appropriate, he argues, for negligence law
to focus on the nature of the defendant’s action, rather than on the defendant’s
blameworthiness for having acted in that manner. When we are looking for a reason
to distribute a loss among outcome-responsible persons, the failure to meet negligence
law’s “objective” standard is a good enough reason, even if the person who fails to live
up to the standard cannot be blamed for failing to do so.17
Perry’s particular responsibility theory of tort is powerful because of its ability
to order in an intuitive manner a complex set of considerations that bear on moral
responsibility, and to do so in a way that likewise makes sense of well-established
features of tort law, particularly negligence law. In assigning a critical role to
foreseeability, in isolating the importance of causation and injury, in permitting
the recognition of multiple responsible actors (including the possibility of victims
being partly responsible for their own injuries), and in creating room for the use of
a less-than-full-blooded version of fault, it captures a set of widely held moral
judgments and a parallel set of tort concepts and doctrines. In turn, it offers the
promise of explaining the contours of tort law by reference to moral judgments, and
thereby legitimizing key features of tort law. In this respect, the project is cheering to

16 17
Perry, Moral Foundations (note 5) at 513. Perry, Moral Foundations (note 5) at 509–10.
22 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

those of us who believe we should have reason to suppose a domain of law is

justifiable before we are willing to stand by it.

B. Responsibility in the real world: Products liability,

comparative fault, and affirmative duties
Although moral-philosophical theories of tort law are criticized for being unrealistic,
responsibility theories are often more grounded and more helpful than supposedly
down-to-earth instrumental theories in making sense of settled law and important
modern doctrinal developments. Here, we briefly discuss three such developments: the
emergence of strict products liability, the adoption of comparative fault, and the
recognition of new affirmative duties in negligence.
Strict products liability is typically defended on overtly instrumental grounds. Some
say it provides an insurance-like mechanism for spreading losses. Others say that it
incentivizes those in the best position to avoid product-related accidents to take
appropriate steps to prevent those accidents. Yet products liability law contains
fundamental limitations on liability that are difficult to justify by reference to these
instrumental considerations.
Courts continue to demand of a products liability plaintiff that she come prepared
to show a causal connection between the defendant’s product and her injury. More-
over, causation is not enough, for the product must also be “defective.”18 While there
has been much debate over how to define “defect,” it is clear that courts require a
products liability plaintiff to identify a problematic feature of the defendant’s product
in order to recover.
In this respect, products liability law resembles negligence law. One can reject
(as we do) the claim made by the reporters for the Third Restatement’s products
liability provisions that, for cases of design defect and failure to warn, proof of
“defect” amounts to proof of seller carelessness.19 Still, they were right in recognizing
that courts have rejected the notion that products-related liability should be divorced
entirely from notions of wrongdoing and responsibility. Even in “strict” products
liability, tort liability involves holding an actor responsible for having wrongfully
injured another.
Responsibility theory on the model of Perry’s helps us to see how this can be the
case. A seller can cogently be deemed morally responsible for its product having
caused an injury when the injury was an avoidable consequence of selling the product.
And the seller becomes accountable in tort law when the seller is not only outcome-
responsible, but when it acted in a wrongful (even if not particularly blameworthy)
manner by sending out into the world a product containing a defect. As far as modern
tort law is concerned, it is the seller’s responsibility to ensure that its products are safe

18
Restatement (Third) of Torts: Products Liability, } 2 (1998).
19
Restatement (Third) of Torts: Products Liability, } 2 (1998), comment a.
 TORT LAW AND RESPONSIBILITY 23

for ordinary use, and sellers are held responsible to injury victims when they fail to do
so and that failure culminates in injury.
The widespread replacement in the 1970s and 1980s of the all-or-nothing contribu-
tory negligence regime with comparative fault is hailed as a signature progressive
development in modern tort law. It is sometimes explained and defended on essen-
tially political grounds—for example, as an instance of judges being less determined to
protect businesses from liability. But the abandonment of contributory negligence is in
many respects the recognition in law of commonsense notions of responsibility. When
a person is knocked down as a result of a sidewalk collision, or one child accidentally
injures another while playing, we often and unproblematically think that each was
partly responsible for the bad outcome. An observer of such an incident might
instinctively judge that both the injurer and the victim should have been paying
more attention and conclude that each is properly deemed partially responsible for
the ensuing injury.
Perry’s notion of localized distributive justice among outcome-responsible persons
is consistent with these moral intuitions and their legal counterparts. We start with all
persons appropriately connected to a loss, and then ask who among them should bear
which portion of the loss. Whether liability and loss are apportioned on a pro rata basis
or in proportion to each actor’s relative fault need not be resolved at the level of tort
theory. Likewise, one can grant that apportionment is appropriate in many cases, yet
also maintain that there are some instances in which the victim is blameless, or
conversely, that the victim’s contribution to her injury is so significant relative to the
defendant’s that there is reason to deny her recovery altogether (as is done in
“modified” comparative fault systems and systems that still recognize implied assump-
tion of risk). It is enough to observe that comparative fault often tracks ordinary
notions of responsibility and fault, so much so that it is now difficult to grasp why
courts were once attracted to an across-the-board rule of contributory negligence.
Today we think it obvious that, on many occasions, each of those whose faulty conduct
was a necessary condition of a foreseeable injury is at least partially responsible for that
injury. Everyday notions of moral responsibility thus help explain why comparative
fault seems like an obvious improvement over contributory negligence.
A third modern doctrinal development of note concerns the increased willingness of
courts to allow plaintiffs to look past an immediate injurer (such as an assailant) to a
background actor whose carelessness is alleged to have set the stage for the injury (e.g.,
the owner of a parking garage or apartment building who fails to provide adequate
security).20 Fifty years ago, courts were, on the whole, more willing to accept defense
arguments that the immediate injurer’s wrong—often involving intentional and crim-
inal misconduct—functioned as a bar to the imposition of liability on the background
actor. (The immediate injurer’s actions were said to constitute a “superseding cause”

20
See John C.P. Goldberg and Benjamin C. Zipursky, “Intervening Wrongdoing in Tort: The Restatement
(Third)’s Unfortunate Embrace of Negligent Enabling,” 44 Wake Forest L. Rev. 1211 (2009) (discussing the
judicial recognition of, and limits on, remote-actor liability for injuries inflicted more immediately by another).
24 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

that blocked the attribution of responsibility to the background actor, even assuming
that the injuring would not have occurred but for that actor’s carelessness.) Today,
property owners are held liable, to varying degrees, for injuries that would not have
occurred had they not been careless in failing to prevent criminal attacks on their
properties. Likewise, the adult host of a party for high school students is subject to
liability if an underage guest gets drunk at the party, drives while intoxicated, and
injures another driver.
It is not hard to identify instrumental concerns that might seem to explain this
pattern. Often the immediate injurer is judgment-proof, whereas the background actor
can pay for the plaintiff ’s losses. Thus, one might suppose that this doctrinal trend is
another testament to the strength of the plaintiff ’s bar and the willingness of judges to
let plaintiffs search for deep pockets. But this account flounders on the fact that the
expansion of this form of liability has been quite circumscribed. Even before Federal
legislation blocked them, negligent marketing and public nuisance claims against gun
manufacturers were rejected by American courts. More generally, courts have rejected
product liability claims where the alleged defect is that an over-the-counter product can
too readily be put to criminal use. They have likewise overwhelmingly rejected “social
host” liability for cases in which the drunk driver is an adult rather than a minor guest.
One can fashion an instrumental explanation for these limits too, but the more
plausible account of the overall pattern is that courts are looking to distinguish
instances in which the background actor can plausibly be deemed responsible for the
victim’s injury. It is common ground in such cases that there is quite a lot to be said,
from the point of view of prudent conduct and foresight, in favor of requiring the
background actor to act so as to reduce the probability that a direct injurer (e.g., a
burglar in a crime-infested neighborhood, an illegal-gun-toting assailant, a drunk
driver) would seriously injure a person in the position of the plaintiff. Thus, for
example, one can argue that a welfare-conscious landlord in a dangerous neighbor-
hood should ensure that there are working locks on a building’s doors, a welfare-
conscious manufacturer of handguns should refuse to sell to downstream commercial
gun distributors with a record of illegal retail sales, and a welfare-conscious social host
should be vigilant about guests’ sobriety. The question in such cases is whether, after a
victim has been injured by the wrongful conduct of the immediate actor, the victim
should prevail in a negligence claim against the background actor on the ground that
his injuries were a foreseeable result of the actor’s failure to take steps like those
outlined above. Overwhelmingly, the courts seem to be moved by the following
question: Even granted that it would have been appropriate and praiseworthy for
the background actor to take the sort of precaution that he or she failed to take, does it
make sense to say that the background actor is therefore responsible for the tenant’s
being attacked by an intruder, responsible for the criminal assailant’s shooting of an
inner-city teen, or responsible for the drunken guest’s careless injuring of others on the
road? When it comes to commercial owners of property, courts routinely say that the
safety of tenants from intruders is in part the responsibility of a property owner.
Conversely, courts tended to deny that the safety of ordinary persons from the attacks
 TORT LAW AND RESPONSIBILITY 25

of criminals who have illegally purchased handguns is the responsibility of gun

manufacturers. Similarly, they deny that adults are required to treat their adult guests
as children who need to be minded, and hence refuse to deem hosts responsible to
persons injured by their adult guests’ drunk driving. In other words, they hold strong
views concerning whether a background actor is appropriately deemed morally
responsible for an injury inflicted wrongfully by a third party, and from there attempt
to determine when there might sensibly be legal responsibility and, with it, liability.

C. Corrective justice theory and responsibility theory

Most corrective justice theorists believe that a defendant who has wronged a plaintiff is
properly vulnerable to the plaintiff ’s claim against her. At least in the sense that the
defendant is deemed properly answerable to a plaintiff, the defendant is deemed legally
responsible for having injured the plaintiff. To the extent that corrective justice
theorists derive the claim that a defendant is properly vulnerable to the plaintiff ’s
claim from the defendant’s having breached a duty to the plaintiff, there is a substan-
tial isomorphism between responsibility theories and corrective justice theory. Where,
as has been the case in the work of Coleman, Ripstein, and Perry himself, the putative
corrective justice theorist actually describes his view in terms of the defendant’s
responsibility for the plaintiff ’s injury or loss, it is more than an isomorphism.
Nonetheless, there is at least one idea that is central to corrective justice theories that
need not be any part of responsibility theories. Corrective justice theorists deploy a
notion of rectification or correction that is said to be central to the normative structure
of tort law. In requiring that the defendant pay compensation to the plaintiff for
injuring the plaintiff, tort law is said to see to it that the wrongful injuring of the
plaintiff is corrected. The payment of compensation is in this sense the doing of
justice. The notion of correction here is teleological (the legal system aims to realize
some valued state of affairs—the rectification); it is dynamic (something happens over
time—an injustice is rectified); and it imports the notion of an equilibrium (a state of
affairs that once obtained is restored—the plaintiff is returned to the status quo ante).
There is no obvious reason why a responsibility theorist is required to reject the core
claims of corrective justice theories. Indeed, in Risks and Wrongs, Coleman seemingly
aims to be both a responsibility theorist, in the sense described above, and a corrective
justice theorist.21 (So too does Perry, though he seems less insistent that his moral
reconstruction of tort renders tort a scheme of corrective justice.) On the other hand,
there is no reason why a responsibility theorist is required to accept any of the above.
One could, for example, take the position that what is demanded of a tortfeasor, in
light of his having wrongfully injured the plaintiff, is to take responsibility by apolo-
gizing, and that tort law, by requiring that damages be paid, is requiring something
akin to an apology rather than requiring the defendant to correct the wrong or the loss.

21
Coleman, Risks and Wrongs (note 3) at 345–7.
26 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

Some corrective justice theorists, including perhaps Scott Hershovitz, might want to say
that, even on this understanding, tort law would be treating the payment of damages as
“making things right,” and would, in that sense, be seeing to it that corrective justice is
done.22 Perhaps. The point is that a theorist need not take this position; he might think
nothing will make things “right” between the parties (even in an extended sense of
“right”) but that tort law nonetheless requires the responsible party to own up to what he
has done to the plaintiff. Being a responsibility theorist permits being a corrective justice
theorist; it does not require it.
As is clear from our prior work, we think there are reasons to stick to responsibility
without a corrective justice gloss. Indeed, we are skeptical about all three aspects of the
concept of correction or rectification described above. Because we believe the state, in
tort law, empowers a plaintiff to seek redress, but does not itself have the power to
bring a victim’s tort claim, we do not conceive of the state itself as aiming to see to
it that compensation is paid by tortfeasors to victims. The normativity of liability-
imposition lies in the empowerment of plaintiffs to obtain redress if they choose.
Because we believe that myriad reasons lie behind a plaintiff ’s choice to seek redress,
and myriad circumstances lie behind a defendant having legally wronged the plaintiff,
and a range of variables typically characterize the relative positions of a defendant and
plaintiff, we are skeptical of the claim that justice is done whenever damages are paid
(though no doubt it is sometimes done). Because we reject the reduction of tort law’s
wrongs to interference with property or goods, and we do not understand how to
stretch the idea of an equilibrium to the notion of wrongs detached from goods, we
find the use of the notion of equilibrium either unpersuasive, mysterious or both.

III. Civil Recourse Theory as a Responsibility

Theory of Tort Law
A. Civil recourse theory
The theory we have developed over the past fifteen years to make sense of the structure
and substance of tort law belongs to the family of responsibility theories we have
sketched above. For a variety of reasons, we often use the phrase “civil recourse” to
denote our overall view, and others have found that label convenient too. We will
continue to do so, but we note here that the distinctiveness of the term “civil recourse”
was never intended to convey a rejection of the normative concepts that are pervasive

22
Scott Hershovitz, “Corrective Justice for Civil Recourse Theorists,” 39 Fla. St. L. Rev. 107 (2011), 118–25.
23
The title of our Torts casebook, co-authored with Tony Sebok and first published in 2004, makes this
connection explicit. See John C.P. Goldberg, Anthony J. Sebok and Benjamin C. Zipursky, Tort Law:
Responsibilities and Redress (New York: Wolters Kluwer, 3d ed. 2012). Likewise, in a recent book chapter,
we argued that one must tease apart the different notions of rights at work in tort precisely in order to grasp the
particular form of responsibility that tort law instantiates and implements. John C.P. Goldberg and Benjamin
 TORT LAW AND RESPONSIBILITY 27

in tort law and tort theory.23 Indeed, we maintain that our theory gives us a better
purchase on the ways in which concepts of duty and right are at work in tort law. Still,
we have not always succeeded in conveying the centrality of responsibility to our view.
We aim to remedy that deficiency here.
As we have noted elsewhere,24 civil recourse theory maintains that Anglo-American
tort law is best understood in terms of three interlocking features: (1) wrongs, (2) rights
of action, and (3) remedies. The theory maintains that these features hang together
to form a body of law that provides recourse through law to victims of a certain kind
of wrong.
Torts are wrongs. Each recognized tort stems from a norm of conduct that enjoins
us not to mistreat others in certain ways. Because these norms are legal norms, torts
are legal wrongs. Though the wrongs of tort tend to track the wrongs of ordinary
morality, an actor’s conduct being a moral wrong is neither necessary nor sufficient for
it to be a tort. For each tort, the norm enjoining conduct is a legally authoritative
directive or rule, even if a directive or rule only implicit in precedent.
Within the category of legal wrongs, torts are further distinguished because the
substance of tort law’s directives tends to be set by law rather than by agreement—a
familiar way of separating tort from contract. Torts are also distinctive as legal wrongs
in that they are injury-inclusive and relational wrongs. Absent an injury to someone,
there is no tort, and even where there is an injury connected to wrongful conduct,
there is still no tort unless the conduct was not merely wrongful in a generic sense but
wrongful as to the injury victim.
Civil recourse theory further identifies as critical to tort law a particular linkage
between the wrongs identified as torts and the idea of a right of action. The commis-
sion of a tort, we claim, confers on the tort victim a particular legal power, namely, the
power to demand and (if certain conditions are met) to obtain responsive action from
the tortfeasor. Liability is the Hohfeldian flipside of this legal power. The commission
of a tort leaves a tortfeasor vulnerable to a claim initiated by the victim and backed by
the power of the state. Because the vulnerability is to the victim, the wrongdoer’s fate is
to a substantial degree in the victim’s hands. The victim, not a government official,
decides whether to press her claim or not, and the victim in principle also decides
whether or not to accept a resolution of the claim short of judgment. If the claim is
successful, of course, the victim can enlist the state’s aid in her effort to enjoin ongoing
wrongful conduct, or to demand responsive action from the wrongdoer in recognition
of the wrong done to her.
It is hardly coincidental that courts and legislatures have seen fit to connect wrongs
and rights of action in the way that our tort law does. For the provision of tort law is

C. Zipursky, “Rights and Responsibility in the Law of Torts,” in Donal Nolan and Andrew Robertson (eds.),
Rights and Private Law (Oxford: Hart Publishing, 2012), 271–3.
24
The following outline of civil recourse theory borrows substantially from a description we have offered
elsewhere. See John C.P. Goldberg and Benjamin C. Zipursky, “Civil Recourse Theory Defended: A Reply to
Posner, Calabresi, Rustad, Chamallas and Robinette,” 88 Ind. L.J. (2013).
28 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

itself a political duty that the state owes its citizens. Following Locke and others, we
have suggested that this duty is rooted in the natural privilege of individuals to respond
to mistreatment by others. Insofar as individuals delegate such privileges to govern-
ments, and insofar as governments justifiably deny individuals the privileges of self-
help and self-assertion in the name of civil peace and justice, it becomes governments’
duty to provide alternatives. By granting to individuals who have been injuriously
wronged a legal power to exact a remedy from the wrongdoer through the courts, a
government complies with the principle of civil recourse—the principle that a person
who is wronged is entitled to an avenue of civil recourse against the wrongdoer.
The third level at which our theory operates is at the level of remedies. Civil recourse
theory asserts that the question of remedy in a tort case turns on the question of what a
person who has proven that she has been wronged is entitled to demand of the
wrongdoer. This, we insist, is a question apart from the question of what sort of
response the defendant is duty-bound to provide. It is about the victim’s right to
redress for the injurious wrong done to her.
To be sure, the idea of “making whole” figures centrally in modern tort practice, a
fact that has misled scholars of various stripes to suppose that tort law is all about
making whole. We argue instead that making whole is but one remedial rule, albeit one
that is in many instances a perfectly reasonable one to adopt, and that has, in fact,
become quite salient (although less than sometimes assumed) in tort law. In other
words, it is a rule that reflects a judgment regarding what constitutes reasonable
redress for the victim of a tortious wrong. Redress is a capacious concept that is
compatible with judicial provision of remedies ranging from injunctions to nominal
damages. This is why the civil recourse account can address more satisfactorily
than competing theories pressing contemporary questions about punitive and non-
economic damages.
Thus defined, civil recourse theory can rather plainly be seen to carry the hallmarks
of responsibility theory that we identified above. First, it emphasizes the significance
for tort of a notion of accountability. It starts with a political-theoretic picture
according to which, as a default matter, each person enjoys an immunity against
certain demands by others. By prevailing in a tort suit, a plaintiff surmounts this
default immunity and establishes that the defendant is properly subject to a legally
enforceable demand for redress. In legal terms, we say that the defendant is subject to
liability to the plaintiff because the defendant committed a tort upon the plaintiff, or
tortiously injured him or her. This is but a legally institutionalized version of the more
general idea that a person is properly subject to a demand from the victim for
compensation or conduct ameliorating her injury where the defendant wronged her.
Second, the ground for the defendant’s answerability to the plaintiff resides in the
defendant’s having wrongfully injured the plaintiff. Each tort is a wrongful injuring of
another. An instance of negligence, for example, is an instance in which an actor
injures another by failing to heed a duty owed to the other to take care not to cause
such an injury. The defendant is subject to liability because it was the defendant’s
 TORT LAW AND RESPONSIBILITY 29

wrong that brought about the plaintiff ’s injury, and the notion of a “wrong” is akin to
that of a blameworthy moral wrong while not being identical to it.
Third, understood as civil recourse law, tort law formalizes and institutionalizes
non-legal notions of wrongfulness, injuriousness, and redress. A tort defendant is
deemed liable to the plaintiff on the basis of having committed a legally defined
injurious wrong against the plaintiff. Tort law’s definitions of wrongdoing depart to
some degree from full-blooded moral wrongs. However, this scheme of responsibility
runs parallel to, and, in this sense still implements, notions of moral responsibility.
Hence, it is fair to explain why tort law imposes liability on certain people by saying
that it deems them responsible for having injured certain others and infers from this
responsibility a right in those others to demand compensatory damages.
In sum, civil recourse theory understands responsibility in tort law as accountability
or answerability for what one has done to another. It is, in part, because there are
certain acts upon others that count as wrongs upon them that it makes sense to regard
some persons as accountable for having injured another. To say they are “responsible”
is not necessarily to say yet what they should do. It is, however, to say that they are
fairly treated as vulnerable to a claim by the plaintiff.

B. Civil recourse theory contrasted with Perry’s

responsibility theory
Like Perry’s, then, ours is a responsibility theory of tort. There are, however, at least
two major differences between our view and his. Ours is not dependent, as his is, on
the claim that there is a general moral duty of repair. And ours is dependent on a
notion of wrongs, whereas his is dependent on a notion of responsibility for loss.
Unsurprisingly, we believe that these differences count in favor of civil recourse theory.

1. Rights of redress without duties of repair

A critical question for any responsibility theorist is why the defendant’s action in
bringing about an injury should generate in the victim a claim against the defendant.
Perry’s answer is this: because the at-fault defendant is outcome-responsible for the
plaintiff ’s harm and more fairly bears the plaintiff ’s loss in light of his fault, the
defendant owes a duty of repair to the plaintiff. The defendant’s outcome-responsi-
bility for bringing about the loss, combined with the defendant’s fault, generates a
moral duty of repair, which in turn generates a moral right in the plaintiff to claim that
compensation is owed. The legal duty of repair is an institutionalized version of the
moral duty of repair, stemming from the defendant’s responsibility for the loss, and
the legal right to be paid damages flows from the legal duty of repair.
We reject Perry’s account for reasons that, even though basic and far-reaching, are
concededly nuanced. First, it relies on an indefensible picture of the structure of tort
liability. There are many areas of law in which the state empowers an individual or
entity to prevail in a claim against another for payment because the defendant has
a legal duty to make that payment. Classic examples involve the federal government,
30 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

through the Internal Revenue Service, bringing an enforcement action against a

taxpayer who has not paid taxes owed, and a Creditor bringing a breach of contract
action for payment of a debt. Defendants in tort cases are not in this position. One of
us (Zipursky) made this argument more than a decade ago against a variety of tort
theorists, and it has never received a serious reply.25 There has, moreover, been
significant scholarship by others strengthening this critique.26
The second cluster of reasons for rejecting Perry’s view assumes arguendo that
there are moral duties of repair stemming from outcome-responsibility for a loss
combined with fault, but observes that the domain of tort liability is poorly matched in
substance to the domain of moral duties of repair. The mismatch occurs across several
dimensions.
Whereas the make-whole norm serves as a default rule of remedy in tort law, we
doubt that there is a comparable default rule in ordinary morality. Few, we suspect,
would sign on to the idea that one who carelessly knocks over a fellow pedestrian
incurs a moral duty to make the victim whole, at least if that entails paying tens of
thousands of dollars to cover lost wages, pain and suffering, and the like. Whatever
may be required from the careless injurer by way of repair, compensation of this
magnitude is more than ordinary morality seems to demand. Indeed, the contrast
between plausible conceptions of the moral duty to repair and the redress afforded by
tort law is precisely what renders the thin-skull rule such a jarring feature of tort law.
Similarly, whereas moral judgments of others’ conduct tend to take account of
certain kinds of excuses, there is no comparable leniency in tort law. Again, negligence
law is quite unforgiving of a person who is incapable of consistently meeting the
objective standard of ordinary care. As matter of formal doctrine, the law’s commit-
ment to objectivity is so stark as to entail a willingness to hold accountable even
persons who, at the time of acting, suffer from a serious mental illness or defect that
renders them incapable of appreciating the dangerousness of their actions.27 The
irrelevance of this sort of excusing condition to tort liability contrasts sharply
with ordinary moral intuitions that the duty of repair in such a case is at least
diminished. The obligatory quality of acts of repair, within ordinary morality, springs
in part from the full-fledged wrongfulness of the injurer’s conduct. The fact that tort
liability springs from acts that might be wrongful in only a thin sense is yet another
reason to think a moral duty of repair does not provide the key link between an
actor’s commission of a tort and the imposition of liability. If the payment of
damages were conceived as a sort of concrete mea culpa, we would expect tort doctrine
to be anchored in a different, and more robust conception of culpability than it
actually is.

25
Benjamin C. Zipursky, “Civil Recourse, Not Corrective Justice,” 91 Geo. L.J. 695 (2003).
26
Peter Jaffey, “Liabilities in Private Law,” 14 Leg. Theory 233 (2008); Nathan Oman, “Why There Is No
Duty to Pay Damages: Powers, Duties, and Private Law,” 39 Fla. St. L. Rev. 137 (2011); Stephen A. Smith,
“Duties, Liabilities, and Damages,” 125 Harv. L. Rev. 1727 (2012).
27
See, e.g., Burch v American Family Mut. Ins. Co., 198 Wis.2d 465, 543 N.W.2d 277 (1996).
 TORT LAW AND RESPONSIBILITY 31

Finally, tort notions of redress contrast sharply with otherwise comparable moral
notions in treating victim need, third-party need, and the injurer’s capacity to pay, as
essentially irrelevant. In ordinary morality, we would suppose, the magnitude and
gravity of a wrongful injurer’s duty of repair is sensitive to how badly it is needed by
the victim, to the needs of others to whom the wrongdoer may have duties, and to the
wrongdoer’s practical ability to pay. If a single parent who is struggling to provide a
decent life for her children carelessly injures an affluent victim, we might well count
the parent’s obligations to the children, and the disparity between the economic
situation of the parent and the victim, as bearing on the question of what the parent
owes the victim. One might, say, for example, that the parent owes it to the victim to
make amends in a manner that is consistent with her situation. Tort law, of course,
invites no such inquiry. In this respect, as in the two respects previously noted above,
tort law seems to demand of the injurer far more than any plausible rendering of the
injurer’s moral duty of repair.
The third cluster of reasons for questioning Perry’s effort to make sense of tort law
as a means of holding wrongful injurers to their moral duty of repair is more
affirmative than the other two. Simply put, there is no need to rely on a putative
moral duty of repair to explain why tort victims enjoy a legal right of action against
tortfeasors. In tort, the defendant’s responsibility for having brought about the injury
through tortious conduct generates a moral right in the plaintiff to demand compen-
sation of the defendant. This right stems from having been wronged, much the way a
moral right to use force to defend oneself against another person stems from being
threatened with imminent harm by that person. Just as one’s freedom from the
aggression of others is conditioned on one not presenting that person with threats,
so one’s freedom from the compensatory demands of others is conditioned on one’s
not having wronged others. The law of torts renders concrete, enforceable, and legal
the right of a victim to demand compensation for having been wronged. The legal
right to make an enforceable demand for compensation therefore arises from the
defendant’s responsibility for having injured the plaintiff. It does so in a way that
makes no reference to a moral duty of repair.
It is worth noting that there is nothing in this account that involves denying
that responsibility for an injury normally gives rise to a prima facie moral duty of
repair. We need not take a position on this issue. The point is that it is neither
necessary nor especially helpful in explaining tort law to say that it does. Our own
inclination is to suppose that under a range of circumstances, which is not nearly as
broad as that which is found in tort, injuring another through morally wrongful
conduct does give rise to some form of reparative duty, which again is typically
substantially less robust than the matching right to compensation would be in tort
law. More importantly, perhaps, we also think that in a range of circumstances where
others have a right to demand compensation because of one’s responsibility, it is a
display of virtue to offer to provide compensation, and in some, it is a display of vice to
fail to offer to provide compensation.
32 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

Part of the lure of the responsibility view, we believe, is that it offers an appealing
moral ideal of a person who takes responsibility for the damage she has inflicted on
others. One might spell out this view by imagining one who is willing to admit when
she is responsible for having brought about another person’s injury and willing to
compensate the victim for the harm inflicted. This sort of scenario is often what one
has in mind when describing someone as “taking responsibility” for what one has
done. Even in this view, one typically imagines a victim asserting that an injurer has
caused damage and the responsible person responding by providing compensation if
she believes the victim’s claim is true. The notion of “responsible” at work in the claim
“you are responsible for my injury!” connotes “you are fairly subject to a claim by me
to account for this!” The “responsible” of the statement “Pat is a responsible person;
she will pay” connotes that Pat has the attribute of responding appropriately to sound
demands based on what she has done; she will pay.
Our view is thus that “responsible,” like its moral cognates “accountable,” and
“answerable,” like its legal cognate “liable,” ends up denoting, in tort law, what
might be called a normative vulnerability. If a person has wrongfully injured another,
that person is, in the standard case, properly vulnerable to a demand for a certain kind
of response.

2. Wrongs, not outcome-responsibility plus fault

Perry treats foreseeability, loss-causation, and fault as a set of conditions, the sequen-
tial satisfaction of which generates a moral duty of repair: first foreseeability, then
causation, and then fault. At the last stage, fault is used as a distributive criterion—as
providing a reason why the defendant’s moral responsibility takes the particular form
of a duty to repair.
On our view, Perry’s effort to utilize fault as a distributional tiebreaker is untenable.
His conception of fault goes simply to the assessment of the defendant’s conduct as
measured against some standard; fault is an attribute of an act or set of acts or
behaviors of the defendant, standing on its own, that marks the defendant as an
appropriate loss-bearer. Wrongfulness in tort law, by contrast, is not an attribute of
acts per se. It is instead an attribute of acts qua injurious interactions. A trespass to
land, a careless running down of a pedestrian, a defrauding of a retiree out of his life
savings: these are all wrongs in the sense that tort law use the term “wrongs.” They
involve the violation of norms directing people not to mistreat others in certain ways—
norms enjoining certain kinds of wrongs against others. (Other torts involve the
violation of legal norms directing people to treat one another appropriately in certain
ways, that is, norms demanding the protection of others against certain kinds of
injury.) It is in the nature of these wrongs that if they have occurred, someone has
been injured. Conversely, it is in the nature of the injuries associated with these wrongs
that the injuries are, so to speak, at the victim-end of the wrong.
In a meritorious tort claim, the plaintiff holds the defendant responsible for having
legally wronged him or her. The responsibility for the legal injury and the responsi-
bility for having legally wronged her are one and the same. We understand these legal
 TORT LAW AND RESPONSIBILITY 33

wrongs to have the same general form as a certain subset of moral wrongs: they are
relational, injury-inclusive wrongs. One is responsible for having committed a legal
wrong upon another just as one is responsible for having committed a moral wrong
upon another. “You hit me!,” “You lied to me!,” “You damaged my car,” “You stole my
boyfriend,” “You ruined my party,” “You killed my dog,” “You scared the living
daylights out of me!”—these are all very natural protestations; complaints of a sort
that are often followed by a demand for responsive conduct.
It is true that in both ordinary life and in law, an injured person will often begin a
discussion of responsibility by identifying some loss or harm suffered, and then
attribute responsibility for that loss or harm to another person. (“My car is destroyed!
You are responsible for this!”) Often the assertion both in and outside of law is
that, because of a causal link between the conduct and the harm, and because of
certain attributes of the conduct, the person engaging in that conduct is responsible
for the loss. In this sense, it is natural, understandable, and unobjectionable to
speak of responsibility for outcomes. However, there is not necessarily any inconsist-
ency between speaking of responsibility for outcomes, in this sense, and speaking of
responsibility for wrongs.
Consider an example in which the plaintiff leaves her car parked on the street, and
the defendant carelessly smashes into it as he backs his car out of his driveway. The
plaintiff, upon discovering the damage, might exclaim: “Who is responsible for this?!?”
In identifying “the defendant” as the answer to this question, one arguably invokes a
notion of outcome-responsibility. But typically one is also inquiring whether a wrong
has been done by focusing on the injury-end of the wrong and then seeking to identify
the author of the wrong. In other words, one is asking whether the injury is part of, or
appropriately connected to, misconduct by someone. Note also that a person can be
connected to an injury not by causing it, but by failing to protect the victim against
it. An ambassador is dead. Who is responsible? Terrorists may be responsible
for his death because his dying is part of an act of killing him for which the terrorists
are responsible. But the President may also be responsible for allowing the ambassador
to be killed by terrorists, and the ambassador’s death may be part of the President’s
wrongful failure to protect the ambassador against being killed. In this scenario,
the President is perhaps not outcome-responsible, in Perry’s sense, but he is still
responsible.
There are many reasons why it is important to our account of tort law that it be
wrongs-based rather than loss-based. Many torts, such as harmless trespasses, involve
no loss. A loss-based view also fails to explain the diversity of remedies in tort law,
including punitive damages. It also fails to explain tort law’s “substantive standing”
requirements, the distinction between predicate injuries and parasitic damages, and
various agency-related doctrines that are core to tort law. Or so we have argued
previously.28 Here we want to address two narrower and somewhat more defensively

28
John C.P. Goldberg and Benjamin C. Zipursky, “Torts as Wrongs,” 88 Tex. L. Rev. 917 (2010).
34 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

framed questions: first, insofar as tort law is about responsibility, isn’t it more naturally
understood as being about holding people responsible for losses? Second, isn’t it an
advantage of Perry’s approach that it makes room for strict liability in a way that a
wrongs-based account cannot? In other words, isn’t Perry’s blend of foreseeability,
causation, and fault—lying, as it does, between Honoré’s broad defense of strict
liability and a wrongs-based approach—much closer to the interpretive sweet spot?
We will answer these questions in reverse order.
As noted from the outset of this chapter, one of the advantages of responsibility
theories of tort law is their capacity to make sense of tort law’s not requiring full-
blooded fault or culpability as a condition of liability. From an interpretive point of
view, this is just where a tort theory ought to be; for it is quite clear that this is how
lawyers and courts have long understood the terms on which liability is imposed.
Perry’s account quite interestingly accommodates this aspect of tort law by depicting
the ultimate judgment of who is responsible by reference to a comparative inquiry
about fault for purposes of resolving the question of localized distributive justice. Fault
is normally the basis for allocating a loss to the defendant, he says. But, in the absence
of subjective fault, objective fault provides a sufficient reason for the law to shift a loss
from an innocent victim to the at-fault actor.
Our approach finds a middle ground between strict liability and full-fledged moral
culpability by a different route, one that we believe is more powerful and systematic.
To say, as we do, that torts are wrongs, is to say that they are violations of relational
directives of conduct that prohibit (or require) certain ways of treating others. Again,
the wrongs of tort and the directives in connection with which they are defined are
legal. Only where the directive can be said to have the status of belonging to the legal
system does the wrong in question count as a legal wrong. Members of the legal
community understand the law to include such directives of conduct and understand
them to enjoin, prohibit, and render not-to-be-done certain ways of interacting with
others. The addressees of these directives owe relational duties to certain others to
refrain from injuring them in certain ways. These duties correspond to rights in
potential victims, in the sense of sets of interests that are not to be interfered with.
Given that we are dealing here with legal directives, not moral directives, potential
victims’ rights are legal rights, just as defendants’ duties are legal duties.
Obviously the content of relational legal directives can vary among legal systems.
But there is nothing that prevents them from being defined in such a way—and in our
system they are in fact often defined in such a way—that they can be violated even
when a person who is subject to them acts in a diligent manner. This is plainly true of
the torts of trespass to land and battery, as well as the torts of libel and slander, at least
as defined at common law. Likewise, once one understands the senses in which the
standard of care in negligence law is objective, or what it means (and does not mean)
to deem a product “defective,” one sees that lack of defendant diligence is not a
condition of liability in the law of negligence or products liability. And yet a person
who injures the plaintiff in the requisite way by engaging in action that falls below
what the directive requires has still committed what can cogently be described as a
 TORT LAW AND RESPONSIBILITY 35

legal wrong. It makes perfect sense to conceive of the norm “Do not invade or occupy
another person’s land without permission” as designating such conduct as wrongful.
And it makes perfect sense to imagine members of a community internalizing the
norm that this is conduct one is duty-bound to refrain from engaging in, and that this
form of interference is one that others have a right to be free from. Nonetheless, a
person might act diligently and reasonably and yet, through a mistake, end up
violating the directive. To do so is to commit the wrong of trespass to land.
For a number of reasons, tort law tends to include legal directives that closely
resemble customary norms of conduct—socially accepted moral directives. Nonethe-
less, the norms of tort are sharper in some places and duller in others, and more
structured in certain respects, too. The result is that there is a reasonably good match
between legal liability in tort and moral responsibility, but the standards of liability are
also over-inclusive, relative to moral culpability (for example, the imposition of battery
liability for an intentional touching in which the person doing the touching genuinely
meant no harm or offense) and under-inclusive, relative to moral culpability (for
example, the absence of liability for blatant breaches of uncontroversial affirmative
moral duties).
In sum, our wrongs-based framework is capable of capturing at least as well as
Perry’s how and why the wrongs of tort law are both related to moral wrongs yet
defined in ways that depart from the dictates of morality, including by allowing
liability without strong forms of culpability. Now, focusing on cases in which wrong-
doing is accompanied by loss, let us consider whether the framework captures
the terms on which tort law imposes liability for losses as well as (or better than)
Perry’s view.
It states the obvious to say that, in a broad swath of tort cases involving losses, a
principal motivation for the plaintiff is to recover an award of compensatory damages
for the harm she has suffered. From the perspective of the plaintiff and her lawyer, the
thought that is first and foremost in their minds may well be that the defendant is
responsible for causing this harm, and thus should pay. Is this not clearly outcome-
responsibility in action, i.e., an effort to hold an actor who has caused harm respon-
sible for having caused the harm, as opposed to an effort to hold an actor responsible
for a wrong?
Our answer is “no.” To assert, in a case like this, that the defendant should be held
liable for the harm because he is responsible for it, is to maintain that the harm was
brought about by the defendant’s breach of a duty to not to harm her through careless
conduct. Take the case of a simple accident between strangers. The plaintiff in such a
case demonstrates that the defendant is responsible for the harm by showing that her
being harmed was the realization of a risk that the defendant carelessly took with
respect to her physical wellbeing by behaving as he did. In other words, the allegation
that the defendant is responsible for the plaintiff ’s harm links the harm suffered by the
plaintiff to the defendant’s breach of a duty to be careful not to harm her. The
plaintiff ’s case for outcome responsibility, in a successful negligence claim, is built
upon her case that the defendant not only failed to avoid causing her loss, but also
36 JOHN C . P . GOLDBERG AND BENJAMIN C . ZIPURSKY

wronged her. Making the case for outcome responsibility actually requires drawing
upon a notion of wrong. It is because the defendant is answerable for having wronged
the plaintiff that he is responsible for the harm suffered.

IV. Concluding Thoughts: The Importance

of Recognizing Responsibility-Based
Accounts of Tort Law
In our conclusion we want to return to the primary aims of this chapter: recognition of
the similarity of responsibility theories of tort law to one another, of their superiority
to a variety of other theoretical approaches to tort law, and of their strength at both an
interpretive and a normative level.
Initially, it is worth reiterating the essentials of a responsibility-based conception of
tort law. The basic idea is that tort law is built upon widely accepted moral principles
according to which one person is sometimes responsible for another’s injury because
she brought it about through action that is wrongful. In a tort claim, the injured person
is empowered to hold the injurer to account for having injured her. The simplest way
to understand tort liability is that it is a concrete, institutionalized, and practical form
of moral responsibility for having wrongfully injured someone. Infighting among
philosophical theorists of tort law notwithstanding, there is actually very substantial
agreement on these core ideas. And, critically, these ideas provide an entirely different
perspective on tort law than efficient deterrence theory or compensation-deterrence
theory, the latter of which is probably still today the dominant account of tort law
among judges and jurists.
An irony of responsibility theory is that, like some other parts of jurisprudential
theory and philosophical work generally, it is a surprisingly short step from the
obvious to the rarified. Few normative ideas are more basic than the thought: “This
is your responsibility!” It is a huge advantage of responsibility theories over other
approaches that they engage the language of the participants in the legal system, and
engage the quotidian discourse of lawyers and judges about what they are doing when
they structure and resolve a battle over tort liability: they are determining who shall be
held responsible.
Although this chapter has not principally aimed at explaining or justifying particu-
lar areas of tort doctrine, much of our prior work has done so, as has the work of
Perry and numerous other responsibility theorists. From Hart and Honoré, to Perry,
to work by ourselves and John Gardner, it is clear that the responsibility-rooted
analysis of causation at the center of Honoré’s work has a real capacity to illuminate
causation doctrine in the law of torts. On foreseeability, duty, fault, and strict liability,
responsibility theories have a great deal to offer. Civil recourse theory, as a form of
responsibility theory, has addressed all of these, as well as a range of issues pertaining
to defenses, liability limitations, remedies, and beyond.
 TORT LAW AND RESPONSIBILITY 37

Responsibility theories, including Perry’s and our own, tend to adopt an interpretive
stance that straddles the sharp divide that is sometimes posited between description
and prescription. In our case, as in Perry’s and Dworkin’s, the methodological
straddling is by design, not the result of ambivalence. One especially important role
of the common law legal theorist involves articulation of the law’s rules and principles
in a manner that displays them as normatively justifiable but that is also consistent
with what might be termed their “normative defeasibility.” Tort law, on our view,
hangs together as a largely coherent scheme for holding wrongful injurers responsible
to their victims, and in doing so, it sits well with the values that law and morality tend
to regard as important, such as liberty of action, security against injury, attention to the
interests of fellow-citizens, and the like. However, it may be that, at least in certain
applications, tort law ought to give way to a scheme, ordered on different principles,
that better permits the realization of these values or other important values. This is
the sense in which the case for tort law, understood as a law of responsibility, is
defeasibly normative. Responsibility theories of tort law—far more than efficient
deterrence, instrumentalist, teleological, or purely deontological accounts—sit well
with this kind of interpretivism.
Finally, responsibility theories usefully depict tort on terms that cut across another
familiar, yet unhelpfully stark, divide: the divide between the public and the private.
Tort law, according to these theories, is fundamentally private because, even if a
defendant is being held responsible through the courts and therefore through the
state, it lies within a private person’s discretion whether to exercise the power to have
the defendant held responsible to her. Yet the empowerment of individuals to bring a
claim—to have one’s injurer held responsible to one—in turn expresses and realizes
important political or public values. For example, tort instantiates a notion equality by

29
See Jason Solomon, “Civil Recourse as Social Equality,” 39 Fla. St. L. Rev. 243 (2011). Although we share
Solomon’s view that tort law instantiates a notion of equality, we are less inclined to suppose that any such
notion is rightly characterized as providing the normative foundation for tort law. John C.P. Goldberg and
Benjamin C. Zipursky, “Civil Recourse Revisited,” 39 Fla. St. L. Rev. 243 (2011), 356–8.
imposing duties upon each not to mistreat others and by conferring upon each rights
not to be mistreated, regardless of status. Among the most important of the many
ways our political system treats people equally is by empowering each to hold others
accountable on the same terms.29
 2
Torts, Rights, and Risk
Stephen Perry*

I. Introduction
It is commonplace to say that theories of tort law divide into two main camps.
Theories in the first camp, which are identified by a wide range of labels including
instrumentalist, economic, functionalist, pragmatist, welfarist, utilitarian, and conse-
quentialist, represent the dominant strand of theorizing about torts in the United
States. Such theories hold that the point or purpose of tort law as a whole is to achieve
certain kinds of moral goals, such as the maximization of welfare or the promotion of
economic efficiency, which can broadly be characterized as collective or aggregative in
character. Although for various reasons I do not think the term is entirely apt, I will,
following widespread usage, refer to this entire category of theories by the label
“instrumentalist.” Theories in the second camp again go by a wide range of labels,
which variously make reference to such notions as corrective justice, rights, duty,
responsibility, fairness, reciprocity, non-instrumentalism, and deontology. Many
scholars in this camp have identified their theoretical views by using the term
“corrective justice.” There is, however, so much disagreement about both the nature
and the concept of corrective justice that I prefer for present purposes to refer to this
entire category of theories as “rights-based.” As I hope the exposition to follow will
make clear, this term captures the essence of most theories in the second camp.
Rights-based and instrumentalist theories of tort, as these have actually been put
forward and defended in the literature, tend to take very different approaches to
both the explanation and the justification of the institution of tort law. Rights-based
theories can be said, as a general matter, to claim to take seriously what Jules
Coleman has called the “structure” of tort law1 and Ernest Weinrib has called its
“self-understanding.”2 These notions of structure and self-understanding refer, at least

* I am grateful to participants at the Conference on the Philosophical Foundations of the Law of Torts, held
at Rutgers Camden Law School, and to participants at the North American Workshop on Private Law Theory,
held at McGill Law School, for their helpful comments on an earlier draft.
1
Jules L. Coleman, Risks and Wrongs (New York: Cambridge University Press, 1992), 374–5.
2
Ernest J. Weinrib, The Idea of Private Law (Cambridge, MA: Harvard University Press, 1995), 50.
 TORTS , RIGHTS , AND RISK 39

as an initial matter, to the doctrinal articulation of the tort of negligence as consisting

of the distinct elements of duty of care, standard of care, injury, cause in fact, and
proximate cause; further structural elements come into play at the stage of remedy. On
a rights-based view, these doctrinal elements have relatively stable meanings which are
drawn from commonsense concepts that tend to be partially morally in character and
partially nonmoral. (An example that tends to be almost wholly moral is the duty of
care; an example that tends to be almost wholly nonmoral is cause-in-fact.) Instru-
mentalist theories, by contrast, which generally see tort law as an instrument for the
pursuit of collective goals such as economic efficiency and loss-spreading, tend to view
tort doctrine, and in particular negligence doctrine, in a much more plastic and
malleable manner, and hence as being subject to perhaps quite radical reform in the
name of better serving these goals. Thus, for example, Guido Calabresi has offered a
notoriously revisionist, probabilistic account of the doctrinal requirement of cause in
fact.3 Defenders of rights-based theories of tort law have frequently criticized instru-
mentalist theories as simply not possessing either the conceptual or the normative
resources that would be adequate to explain not just the requirement of cause-in-fact
but also other basic doctrinal elements such as the duty of care and proximate cause.
In this chapter I wish to explore some questions concerning the relationship
between torts, rights, and the concept of risk, taking as my starting point the doctrinal
element of a duty of care in negligence law. The idea of a general duty of care has
almost disappeared from the prevailing American academic understanding of negli-
gence law. This historical development has been comprehensively and perceptively
discussed by John C.P. Goldberg and Benjamin C. Zipursky.4 As they persuasively
argue, the reasons for the development have to do with the sway of legal realism within
the American legal academy and an associated and growing acceptance, over the
course of the twentieth century, of an instrumentalist understanding of tort law that
has been strongly influenced by the work of William Prosser but that can ultimately be
traced back to the writings on the common law of Oliver Wendell Holmes. This has
given rise to a predominant academic understanding of the nature of tort law in the
United States which in certain respects is strikingly different from the understanding
that still prevails in England and the Commonwealth. I try in what follows to elucidate
the element of duty of care in negligence law by examining its underpinnings in moral
theory, with a view to reinforcing the traditional understanding of the duty element
and, in the American context, to perhaps contributing to its rehabilitation. As will
emerge, I do not think that the concept of a duty of care is as clear-cut or as
uncomplicated as rights theorists sometimes seem to assume. This is because the
relationship between the duty of care, the concept of rights, and the concept of risk
is quite complicated. My hope is that clarification of the moral character of the duty
element and its relationship to risk and rights will lend at least indirect support for a

3
Guido Calabresi, “Concerning Cause and the Law of Torts,” 43 U. Chi. L. Rev. 69 (1975).
4
John C.P. Goldberg and Benjamin C. Zipursky, “The Moral of MacPherson,” 146 U. Pa. L. Rev. 1733
(1998).
40 STEPHEN PERRY

general understanding of negligence law, and ultimately of tort law as a whole, which is
predominantly rights-based.

II. Duty of Care and Rights

On any understanding of tort law, instrumentalist or rights-based, the tort of negli-
gence subjects persons to a mandatory norm of conduct. In doctrinal terms this is of
course the standard of reasonable care, which requires that one act so as not to subject
others to unreasonable or excessive risks. A mandatory legal norm is, by its very nature,
obligatory, which is simply to say that one has a duty to comply with it. So negligence
law cannot do without the concept of duty in at least this limited sense, and this point is
of course recognized by instrumentalist theories. Thus the American Law Institute’s
Restatement of the Law of Torts, Third: Liability for Physical and Emotional Harm,5
which incorporates and indeed effectively codifies a predominantly instrumentalist
understanding of torts, does not ignore the concept of duty entirely. Section 7(a) says
that “[a]n actor ordinarily has a duty to exercise reasonable care when the actor’s
conduct creates the risk of physical harm.” Although } 7 is entitled “Duty,” and } 7(a)
states conditions under which a person has a duty, this is clearly just a formulation of
the familiar standard of reasonable care. It states a mandatory legal norm of conduct,
and of course there is always a duty to comply with a mandatory norm. There is very
little in the Restatement that corresponds with the element of duty of care as it figures in
English and Commonwealth law or, indeed, in Justice Cardozo’s classic opinions in
MacPherson v Buick Motor Co6 and Palsgraf v Long Island Railroad.7
What is the difference between the } 7 “duty” and the rights-based, traditional
notion of a duty of care? Almost without exception, rights theorists have characterized
tort law as being “relational” in character, a feature which has also been variously
referred to as “interactional,” “transactional,” “bipolar,” “bilateral,” and “correlative.”8
What rights theorists have in mind in employing these terms is quite straightforward,
and involves two separate but closely related points. The first is that tort law in general,
and negligence law in particular, regulates certain kinds of interactions or transactions
between persons, who in the paradigmatic case are two in number. The second point,

5
Restatement (Third) of Torts: Liability for Physical and Emotional Harm.
6
MacPherson v Buick Motor Co., 217 N.Y. 382, 385, 11 N.E. 1050 (1916).
7
Palsgraf v Long Island Railroad, 248 N.Y. 339, 162 N.E. 99 (1928).
8
See, e.g., Ernest J. Weinrib, “Understanding Tort Law,” (1989) 23 Val. U. L. Rev. 485 (1989), 494
(“bipolar”); Coleman, Risks and Wrongs (note 1) at 311–18 (“relational”); Peter Benson, “The Basis of
Corrective Justice and its Relationship to Distributive Justice,” 77 Iowa L. Rev. 515 (1992), 533 (“relational,”
“intelligible as interaction”) and 569 (“corresponding” duty and right); Stephen Perry, “The Moral Founda-
tions of Tort Law,” 77 Iowa L. Rev. 449 (1992), 507 (“correlative rights and duties”); Weinrib, The Idea of
Private Law (note 2) at 114–44 (“correlativity”); Goldberg and Zipursky, “The Moral of MacPherson” (note 4)
at 1744 (“relational”); Jules L. Coleman, The Practice of Principle: In Defence of a Pragmatist Approach to Legal
Theory (New York: Oxford University Press, 2001), 13–24 (“bilateral”); Arthur Ripstein, “Tort Law in a Liberal
State,” 1(2) Journal of Tort Law 2 (“bilateral”) and 14–15 (“transactional”).
 TORTS , RIGHTS , AND RISK 41

which is the more important of the two because it ultimately underlies the first,
concerns the normative character of the duty of care. While negligence law on any
view involves a duty to comply with a mandatory norm, the traditional understanding
of the duty of care, and the understanding advocated by rights theorists, is that of a
duty which is owed to someone. Duties to comply with mandatory norms do not
necessarily have this feature. Laws making it an offence to litter or jay-walk, for
example, create mandatory legal norms which impose duties not to litter or jay-
walk. But such duties are not ordinarily thought to be owed to anyone, or, at the
very least, they are not thought to be owed to other citizens. To say, however, that a
duty is owed to another person is to say that the other person has a correlative right.
Duties can exist without rights in some contexts, but in a rights-based theory of tort
law, duties cannot exist without rights and rights cannot exist without duties. That is
one of the reasons that I prefer to label the second category of theories of tort law that
I identified in the Introduction by using the term “rights-based” rather than one of the
other terms in current use, such as “corrective justice.”
Instrumentalist theories of tort law take a variety of forms, but one of the most
influential variants sees tort law in general, and negligence law in particular, as a
means for promoting economic efficiency. The general idea is that actors, both
potential defendants and potential plaintiffs, should be given incentives to conform
their behavior to the standard of reasonable care, which is then typically defined in
cost-benefit terms along the lines of Judge Learned Hand’s famous formula.9 The core
of the efficiency-based understanding of tort law is thus the standard of reasonable
care, understood as a mandatory norm which all persons have a duty to comply with.
The purpose of the norm is to eliminate, through deterrence, inefficient levels of risk-
taking, and the associated duty is simply a duty to comply with the norm, in exactly the
same way that one has a duty to comply with legal norms which prohibit jay-walking
or littering. The duty is not one that is owed to anyone else. On the pure efficiency-
based approach, then, there is neither need nor room for the full-blooded notion of
a duty of care, understood as a duty which is owed to persons whose interests an
actor’s activities have foreseeably put at risk, and who can therefore be said to have a
correlative right that the duty-holder look out for their interests. What matters, for
the efficiency-based view, is simply that persons conform their conduct to the standard
of reasonable care. It is thus often said that, on this view, tort law is best viewed simply
as a variant of a general regulatory approach, where regulation happens to take a
private form.
In light of the influence and pervasiveness of the economic interpretation of tort law
in the United States, it is not in the least surprising that the element of duty of care in
negligence law has for many years been de-emphasized by instrumentalist theorists
and that now, with the advent of the new Restatement, has disappeared almost entirely

9
United States v Carrol Towing Co., 159 F. 2d. 169 (2d Cir. 1947), 173. According to Hand J, an actor is
negligent if, for that actor, B < PL, where B is the burden or cost of precautions, L is the magnitude of the loss
were it to occur, and P is the probability of its occurrence on any given occasion.
42 STEPHEN PERRY

from the most influential formulation of American tort doctrine. To the extent that the
formal idea of a duty owed to another has survived in some vestigial form in American
instrumentalist accounts of negligence law, it is often understood, either explicitly or
implicitly, in terms of Holmes’s notion of “a duty of all the world to all the world.”10
Superficially this phrase might seem to suggest a relational view, but, as Holmes
himself made clear—and as Andrews J did as well in his borrowing of the idea in his
dissent in Palsgraf—“the duty of all the world to all the world” does not involve
anything like correlative rights, but simply states the general idea that everyone should
conform their conduct to the general standard of reasonable care. For instrumentalists
who are more pragmatic or more pluralist than advocates of, say, the pure efficiency-
based approach, the existence of a duty is sometimes understood not in this Holmesian
sense but rather as always depending directly—either on a case-by-case basis or at least
on a category-of-case by category-of-case basis—on the balancing of a wide range of
policy considerations. Typically these include, among many other factors, deterrence,
loss-spreading, and the concern for avoiding overly extensive liability in certain
contexts, such as in cases of pure economic loss. On the “general balancing” approach,
the various relevant policy considerations are balanced by the court to determine
whether or not there is a “duty of care.” This differs from the Third Restatement’s
approach, which begins with a “general rule” of reasonable care—meaning a universal
rule like Holmes’s “duty of all the world to all the world”—which is then subject to
restriction by various policy considerations (so-called “no-duty rulings”). However,
neither approach includes the idea that the relevant duty is relational and correlative of
a right on the part of the plaintiff.11 This is in strong contrast to the understanding of
duty of care found in the classic cases on the subject. Thus, in Donoghue v Stevenson
Lord Atkin spoke of the duty of care as “a general conception of relations,”12 and in

10
“The Theory of Torts,” 7 Am. L. Rev. 652 (1873), 662. This article has no attributed author but, according to
the Third Restatement, Holmes is widely credited as having written it. The Restatement cites this article in support
of what it calls the “the general rule” of } 7(a) that “[a]n actor ordinarily has a duty to exercise reasonable care when
the actor’s conduct creates a risk of physical harm.” See Reporter’s Note to comment a, } 7. As has already been
noted in the text, no notion of a correlative right is mentioned in the Restatement’s discussion of } 7.
11
On a related point, the “duty to all the world” approach, which is adopted by the Third Restatement,
there is clearly no room for the restrictive requirement of reasonable foreseeability that figures so prominently
in the classic cases of Donoghue v Stevenson [1932] AC 562 (HL) 580, MacPherson, and Palsgraf. As for the
“general balancing” approach, some American courts include reasonable foreseeability among the factors to be
balanced, and some do not. For an example of the “general balancing” approach that apparently does not
include reasonable foreseeability among the relevant policy concerns, see 532 Madison Ave Gourmet Foods Inc
v Finlandia Center Inc, 96 N.Y. 2d. 280, 288–9, 750 N.E. 2d. 1097 (2001). For a version of the general balancing
approach that not only includes foreseeability of harm to the plaintiff as one of the considerations to be
balanced, but states that foreseeability is the most important such consideration, see, e.g., Tarasoff v The
Regents of the University of California, 17 Cal. 3d 425, 434, 551 P. 2d 334 (1976). The Third Restatement
explicitly rejects the California approach: “So long as the actor’s conduct created a risk of harm—the predicate
for a duty under this Section—foreseeability has no role under this Section and Restatement in a determination
that a duty exists vel non” (} 7, Reporters’ Note, comment j). For further discussion of } 7, see Stephen Perry,
“The Role of Duty of Care in a Rights-Based Theory of Negligence Law,” in Andrew Robertson and Tang Hang
Wu (eds.), The Goals of Private Law (Oxford: Hart Publishing, 2009), 84–91.
12
Donoghue v Stevenson [1932] AC 562 (HL) 580.
 TORTS , RIGHTS , AND RISK 43

Palsgraf Cardozo CJ was even more explicit that what is at stake is a right on the part of
the plaintiff: “What the plaintiff must show is a ‘wrong’ to herself, i.e., the violation of
her own right, and not merely a wrong to someone else, nor conduct ‘wrongful’
because unsocial, but not a ‘wrong’ to anyone.”13 For all intents and purposes, modern
American tort law has abandoned this traditional understanding of the duty of care.
Modern Commonwealth case law on the duty of care stems from Anns v London
Borough of Merton,14 which laid out a two-stage test for determining the existence of a
duty of care. According to the first stage, which builds on Lord Atkin’s “neighbor
principle,” one asks whether:
there is a sufficient relationship of proximity or neighbourhood such that, in the reason-
able contemplation of the [alleged wrongdoer], carelessness on his part may be likely to
cause damage to the [person who has suffered damage]—in which case a prima facie duty
of care arises.

At the second stage, one asks whether there are any considerations which ought to
negative, limit, or reduce the scope of a prima facie duty that arises at the first stage. This
two-stage test differs from both of the American approaches I have distinguished—i.e.,
the “duty to all the world” and the “general balancing” approaches—in that it expressly
emphasizes the relationship between the parties, and it does so in a way that preserves
the central role of reasonable foreseeability in defining that relationship. As McLachlin
CJ interprets the first stage of the Anns test in the Supreme Court of Canada case of
Cooper v Hobart, “reasonable foreseeability must be supplemented by proximity”
where proximity is:
clearly intended to connote that the circumstances of the relationship inhering between
the plaintiff and the defendant are of such a nature that the defendant may be said to be
under an obligation to be mindful of the plaintiff ’s legitimate interests.15

As this passage makes quite clear, McLachlin accepts a relational conception of the
duty of care.
McLachlin goes on to state that the factors relevant to the determination of
proximity include “expectations, representations, reliance and the property or other
interests involved.”16 It is worth emphasizing that these factors are among the most

13
Palsgraf, 248 N.Y. at 343–4. The misinterpretation of Cardozo CJ’s opinions in MacPherson and Palsgraf
in the service of various instrumentalist causes has been superbly documented by Goldberg and Zipursky (note 4).
For a particularly egregious instrumentalist misconstruing of Palsgraf, see Calabresi, “Concerning Cause and the
Law of Torts” (note 3) at 91–102. For an excellent defense of the view that Cardozo CJ in fact held a systematically
rights-oriented understanding of tort law which can be discerned across the full range of his opinions, see Peter
Benson, “The Problem with Pure Economic Loss,” 60 S.C.L. Rev. 823 (2009).
14
Anns v London Borough of Merton [1978] AC 728 (HL) 751–2.
15
Cooper v Hobart [2001] 3 SCR 537, 551 and 552.
16
Cooper v Hobart [2001] 3 SCR at 552. It should be noted that the term “proximity,” as it is used by
Commonwealth courts in the context of duty of care, means something quite different from the American term
“proximate cause.” The Commonwealth usage that corresponds to the American term “proximate cause” is
“remoteness.”
44 STEPHEN PERRY

important considerations that bear on the question of whether one person has, as
a moral matter, a right as against another person that the latter be mindful of the
former’s legitimate interests.17 It is only at the second stage of the Anns test that policy
considerations of the sort that figure directly in the one-stage American general
balancing test—in McLachlin’s words, considerations that “are not concerned with
the relationship between the parties, but with the effect of recognizing a duty of care on
other legal obligations, the legal system, and society more generally”18—come into
play. This two-stage, structured approach to the existence of a duty of care is very
naturally interpreted as giving priority to considerations that bear on the existence of a
duty on the defendant’s part which is correlative of a right on the part of the plaintiff,
but which also allows policy considerations in the broader sense to come into play in a
secondary or ancillary manner.19

III. Risk, Harm, and Rights

I said in the preceding section that rights theorists see the element of duty of care as a
feature of negligence law that captures its true relational character. That this is so is not
an entirely straightforward matter, however, at least insofar as one of the conditions of
adequacy of a rights-based theory is that it take seriously what I earlier referred to,
following the helpful terminology of Coleman, as the “structure” of tort law. An
important potential difficulty is that, if the concept of tort law’s structure is to be
anything more than a pure doctrinal invention, each structural component—in the
case of negligence law, each of the five main elements of doctrine—must correspond to
a meaningful feature of morality. But it is not entirely clear that the duty of care, for
example, really does correspond to, say, an independent moral duty. The content of
the duty of care would appear to be, at the level of doctrine, a duty to be mindful of the
interests of persons placed at reasonably foreseeable risk by one’s actions, where those
other persons have a correlative right that one be so mindful. As Cardozo memorably
put the point in Palsgraf, “[t]he risk reasonably to be perceived defines the duty to be
obeyed, and risk imports relation; it is risk to another or to others within the range of
apprehension.”20 Putting the duty of care together with the standard of care, the result
is a duty not to subject persons whom one’s actions place at reasonably foreseeable risk
to unreasonable or undue levels of such risk. (For purposes of brevity, I shall

17
See Stephen R. Perry, “Protected Interests and Undertakings in the Law of Negligence,” 42 U. Toronto
L. Rev. 247 (1992).
18
Cooper v Hobart [2001] 3 SCR 537, 551 and 554.
19
It should be pointed out that, in subsequent cases, the House of Lords has to some extent distanced itself
from the Anns two-stage test, especially as regards the second, policy-oriented stage. The relevant cases are
insightfully discussed by Robert Stevens in Robert Stevens, “Law of Torts,” in Louis Jacques Blom-Cooper,
Gavin Drewry and Bruce Dickinson (eds.), The Judicial House of Lords 1876–2009 (Oxford: Oxford University
Press, 2009).
20
Palsgraf, 248 N.Y. at 344.
 TORTS , RIGHTS , AND RISK 45

sometimes refer to this duty as the duty of care, or the duty not to subject others to
unreasonable risk.) As has often been noted, it is possible, certainly as a doctrinal
matter, to breach one’s duty of care—which essentially means, as was just noted, to
subject another person to an unreasonable level of risk—without causing that person
physical harm, and therefore without committing the tort of negligence. At least two
considerations might lead us to think that the duty of care might not correspond to an
independent moral duty. The first is that it is very plausible to think that our most
fundamental moral duties are, quite simply, duties not to harm others in various ways,
rather than duties not to subject others to the risk of harm. The second consideration is
that even if our most fundamental duties are duties not to harm, it is not clear that
there is room within the appropriate moral framework for even an ancillary or derived
duty to not subject others to risks of harm. The overall concern, in other words, is that
there is simply no such thing as a moral duty not to subject others to (unreasonable or
undue) risks,21 and hence there is nothing which can serve as the moral counterpart of
the doctrinal duty of care. I will address these two considerations in Section IV below.
First, however, it will be helpful to say something about one particular claim that has
sometimes been made in support of the conclusion that there is a moral duty not to
subject others to risk. The basic idea is that risk itself is a form of harm, and we are
duty-bound not to subject others to (unreasonable) levels of risk because to do so is to
harm them. This inquiry will be useful because, among other things, it will tell us a
great deal about the morality of risking. A number of prominent commentators have
argued for the view that risk is at least sometimes harm in itself,22 and in the United
States several courts have recognized such claims in the medical malpractice context.23
In order to bring the concept of risk into sharper moral focus, I will develop an
example based on the English medical malpractice case of Hotson v East Berkshire
Area Health Authority.24 Hotson was not a duty of care case, but rather a case
concerning the requirement in negligence law that the plaintiff have suffered a legally
cognisable injury. The plaintiff ’s claim in Hotson was that he had, as a result of the
defendant’s negligence, lost a chance of avoiding an adverse physical outcome, and
that this lost chance could properly be characterized as harm in its own right, of a kind

21
The view that there is no such thing as a right to be free from (unreasonable) risk—which is just the
correlate of the duty not to subject others to (unreasonable) risk—is argued for in Heidi M. Hurd, “The
Deontology of Negligence,” 76 B.U. L. Rev. 249 (1996); Heidi M. Hurd and Michael S. Moore, “Negligence in
the Air,” 3(2) Theoretical Inquiries in Law, Art 3 [2002], 17–21.
22
See, e.g., Joseph H. King, “Causation, Valuation, and Chance in Personal Injury Torts Involving
Preexisting Conditions and Future Consequences,” 90 Yale L.J. 1353 (1981); Jane Stapleton, “The Gist of
Negligence, Part 2,” 104 Law Q. Rev. 389 (1988).
23
See, e.g., Alberts v Schultz, 126 N.M. 807, 975 P. 2d 1279 (1999); Beswick v City of Philadelphia, 185 F.
Supp. 2d 418 (E.D. Pa. 2001). The Third Restatement seems cautiously to endorse these cases: Restatement
(Third) of Torts, } 26, comment n. For criticism of this approach, see Stephen R. Perry, “Risk, Harm, and
Responsibility,” in David G. Owen (ed.), Philosophical Foundations of Tort Law (Oxford: Clarendon Press,
1995), 330–9.
24
Hotson v East Berkshire Area Health Authority [1987] AC 750 (HL). I have discussed Hotson in earlier
work. See Perry, “Risk, Harm, and Responsibility” (note 23) at 330–9; Perry, “The Role of Duty of Care in a
Rights-Based Theory of Negligence Law” (note 11) at 92–6.
46 STEPHEN PERRY

that was capable of sustaining a negligence action. To say, however, that a person has
suffered a reduced or lost chance of avoiding an adverse physical outcome is simply to
say that the person has been subjected to an increased risk of suffering that same
outcome.
The facts in Hotson were that the plaintiff had suffered a fracture of his left femoral
epiphysis, which the defendant health authority negligently failed to diagnose. The
trial judge found that at the time the plaintiff arrived at the defendant’s hospital there
was a 75 per cent chance that even with proper treatment a serious condition called
avascular necrosis would have occurred, but that the delay in treatment which resulted
from the defendant’s negligent diagnosis made the development of this condition a
certainty. The plaintiff argued, and the trial judge accepted, that the defendant had
caused the plaintiff to lose a 25 per cent chance of avoiding avascular necrosis, and
that this lost chance should itself be treated as damage compensable in tort. In
the House of Lords Lord Mackay put the plaintiff ’s argument this way:
[Counsel] who appeared for the plaintiff . . . said that in relation to the facts of this case as
found by the trial judge what was meant by a chance was that if 100 people had suffered
the same injury 75 of them would have developed avascular necrosis . . . and 25 would not.
This, he said, was an asset possessed by the plaintiff when he arrived at the authority’s
hospital . . . It was this asset which [counsel] submits the plaintiff lost in consequence of the
negligent failure of the authority to diagnose his injury properly.25

The plaintiff ’s argument, as summarized by Lord Mackay, is a crystal-clear statement

of the idea that a risk or probability should be understood by reference to the notion of
relative frequency. The 25 per cent chance of avoiding avascular necrosis which the
plaintiff claimed to have lost is expressed as a relative frequency of the occurrence of
the disease within a so-called reference class, which in this case Lord Mackay describes
as a group of 100 persons who are said to have suffered the “same” injury. This chance
is, moreover, said by the plaintiff to be an asset which the plaintiff possessed at the time
he arrived at the defendant’s hospital. If this claim can be sustained, and if it can be
further shown that in an appropriate sense the defendant caused the plaintiff to lose
this asset, then it is at least arguable that the defendant set back an interest of the
plaintiff ’s in a way that might be said to constitute harm.
It is, however, notorious within the philosophy of probability that for any given
situation involving probability or risk in the relative frequency sense, there is no
determinate and unique way to specify a reference class which can be regarded as
the canonical or correct one. This gives rise to problems for the claim that risk is a
harm in itself, which we can bring out by further reflection on the facts of Hotson. As
mentioned, the reference class averred to by Lord Mackay is that of “100 people
[who] . . . suffered the same injury.” But no two people ever suffer exactly the same
injury, and we can therefore imagine a variety of different reference classes that might
be formulated in a case like Hotson. In the plaintiff ’s case, the immediate cause of the

25
Hotson [1987] AC 750 (HL) 783.
 TORTS , RIGHTS , AND RISK 47

injury was a reduction in the supply of blood vessels reaching the epiphysis. It was not
known how many blood vessels had originally been ruptured, but suppose the number
was known to be relatively low. We could then formulate a different reference class
consisting of 100 persons who had suffered a fracture of the left femoral epiphysis and
who had also ruptured such-and-such (relatively low) proportion of vessels supplying
blood to the epiphysis. We could then expect to find, as an empirical matter, that the
relative frequency of persons who will develop avascular necrosis even with proper
treatment is, within this reference class, something less than 75 per cent. It might be,
say, 40 per cent. Alternatively, suppose we knew that the proportion of ruptured blood
vessels was relatively high. Then we could in parallel fashion formulate yet another
reference class in which the relative frequency of persons who will develop avascular
necrosis can be expected to be higher than 75 per cent. And we can further imagine
coming to know more and more about the causes of avascular necrosis so as to be able
to formulate, at least in principle, a reference class consisting of persons who fractured
their left femoral epiphysis and who also possessed certain other characteristics, call
them X, such that the relative frequency of developing avascular necrosis even with
proper treatment was 100 per cent. Similarly, we can imagine that increased empirical
knowledge would allow us to formulate a reference class consisting of persons who
fractured their left femoral epiphysis and who also possessed certain other character-
istics, call them Y, such that the relative frequency of developing avascular necrosis
even with proper treatment is, within this reference class, 0 per cent. And of course, if
we knew enough about the plaintiff ’s particular injury so as to be able to say either that
he possessed the X characteristics or that he possessed the Y characteristics, then we
would be in a position to say that his chance of developing avascular necrosis even
with proper treatment was either 100 per cent or 0 per cent.
As, I hope the example begins to make clear, both the existence and the magnitude
of a so-called lost chance of avoiding physical harm of some kind—or, what comes to
the same thing, a risk of suffering that kind of physical harm—is a function of our
current state of knowledge about both the general type of injury and the specific
plaintiff ’s particular condition. Perhaps it would be more accurate to say that the
existence and magnitude of the risk is a function of our current state of ignorance
about these matters. But whichever way we put the point, it is a very odd form of harm
the very existence of which depends in this way on the state of our knowledge. Either
way, there is a clear sense in which we can say that the existence and magnitude of risk
are epistemic phenomena, because in all practical cases statements of risk are relative
to a (greater or lesser) degree of uncertainty. Notice, however, that, in the example
based on Hotson, the accumulation of greater empirical knowledge which permits us
to define new reference classes does not justify the conclusion that statements of
relative frequency that were made in a state of greater ignorance were erroneous.
Thus a relative frequency of 75 per cent might be accurate for the admittedly rather
vaguely identified reference class that figures in the quote from Lord MacKay, and
the 40 per cent figure might be accurate for the different reference class that I described
in my variant hypothetical. Both statements of relative frequency are (or at least
48 STEPHEN PERRY

might be) accurate, for the reason that a statement of relative frequency is always made
relative to the specification of a reference class, and, on the variant of the facts in
Hotson that I described, the plaintiff belongs to both of these reference classes. There is
thus a sense in which relative frequencies, and thus statements of risk, are not
epistemic; they are facts about the world.26
Strictly speaking, a probability, understood as a relative frequency, is a property of
specified reference classes and not a property of the individuals who happen to fall into
those classes. Any given individual will fall into an indefinitely large number of reference
classes, where the relative frequency of harm will be, for each one, different. For this
reason, counsel for the plaintiff in Hotson was wrong to argue that the plaintiff ’s
supposed 25 per cent chance of avoiding avascular necrosis if proper treatment had
been administered was an asset that the plaintiff possessed at the time that he entered the
hospital. Even if we assume that the 25 per cent figure was accurate for some reference
class that could be identified with some reasonable degree of specificity, the plaintiff also
belonged, as a matter of objective fact, to indefinitely many other reference classes—
many of which we simply will never know about—for which the relative frequency of
avoiding the physical harm was different. In the hypothetical I developed, the plaintiff
simultaneously belongs to reference classes for which the relative frequency of avoiding
the harm if proper treatment had been administered is 25, 60, and 100 per cent. But
surely the plaintiff did not simultaneously possess assets of a 25 per cent chance, a 60 per
cent chance, and a 100 per cent chance of avoiding the harm if appropriate treatment
had been rendered. This is just another way of saying that a decreased chance of avoiding
a physical harm—or, equivalently, an increased risk of suffering that harm—cannot
plausibly be viewed as itself being a form of harm.27
The House of Lords rejected the plaintiff ’s argument in Hotson that risk is a form of
harm, and I have argued that it was correct to do so. Of course Hotson was not a duty

26
The fact that a risk of suffering harm (or, equivalently, a reduced probability or chance of avoiding harm)
is always, practically speaking, a function of some specified level of knowledge has led some theorists to say that
risk is by its very nature a purely epistemic phenomenon which has no existence in the physical world. From
the perspective of an omniscient being, the risk is always either 0 per cent or 100 per cent, and therefore these
are the only risks that can, according to these theorists, actually be said to exist. See, e.g., Hurd, “The
Deontology of Negligence” (note 21) at 263–4; Hurd and Moore, “Negligence in the Air” (note 21) at 19.
However, these claims are not exactly right, because statements of risk have both an epistemic and a non-
epistemic dimension. For discussion, see Stephen R. Perry, “Responsibility for Outcomes, Risk, and the Law of
Torts,” in Gerald Postema (ed.), Philosophy and the Law of Torts (New York: Cambridge University Press,
2001), 98–9; Stephen Perry, “Risk, Harm, Interests and Rights,” in Tim Lewens (ed.), Risk: Philosophical
Perspectives (New York: Routledge, 2007), 193–8.
27
If the argument that the plaintiff advanced in Hotson were correct, it is very difficult to see how to avoid
the conclusion that every risk that one person imposes on another is an instance of harm. Some rights theorists
have, however, argued for the proposition that every risk imposition is not a harm but a rights violation, and
that our most fundamental moral rights are precisely rights not to be subject to any risk of physical harm
whatsoever, whether known or unknown. See David McCarthy, “Rights, Explanation, and Risks,” 107 Ethics
205 (1997); David McCarthy, “Liability and Risk” 25 Philosophy and Public Affairs 238 (1996). Ultimately,
however, this proposal goes awry because of the impossibility, in a given risk situation, of delineating a
determinate and unique reference class. For detailed criticism of McCarthy’s theory, see Perry, “Risk, Harm,
Interests, and Rights” (note 26) at 203–5.
 TORTS , RIGHTS , AND RISK 49

of care case, and it might seem that we have strayed rather far from our original
questions about the true moral character of the defendant’s duty and the plaintiff ’s
correlative right. But we have not strayed so far as all that, because in the course of this
digression we have learned something about the nature of risk, and the morality of
risking is the central concern of the doctrinal requirement of duty of care. I have
argued elsewhere that because of the epistemic dimension of a statement to the effect
that an agent A imposed a risk of physical harm on another person B, such a statement
must, for purposes of both morality and the law, always be made relative to an
epistemic perspective.28 An epistemic perspective involves a specification, either impli-
cit or explicit, of the knowledge of the relevant facts that A either possessed (actual
epistemic perspective) or ought to have possessed (constructed epistemic perspective)
at the time of acting. A constructed epistemic perspective is one that is attributed to
A for moral and legal purposes. The specification of a constructed epistemic perspec-
tive will involve, usually as an implicit rather than an explicit matter, the idea that a
certain description of the kind of action engaged in by A—and, hence, that a certain
characterization of a background reference class of types of actions or types of effects,
together with an associated relative frequency of a specified kind of harm within that
reference class—is the appropriate or correct description for purposes of morally and
legally evaluating A’s actual act of risk imposition. The elements in negligence law of
both the duty of care and the standard of care involve, among other things, a
constructed epistemic perspective—what the defendant ought to have known about
the riskiness of her actions—to which she is held regardless of her actual state of
knowledge about the risk she was imposing. There is, however, another implication of
the epistemic dimension of risk. If, because of this dimension, risk cannot be regarded
as harm, one reason to think that the right not to be unreasonably risked is a
fundamental moral right—and that the correlative duty not to subject others to
unreasonable risk is a fundamental moral duty—is eliminated.

IV. Harm and Fundamental Moral Rights

How can we further assess the claim that among our most fundamental moral rights
are rights not to be subject to risk? The most straightforward path forward is to make a
positive case for the obvious alternative view, and indeed the most intuitively appeal-
ing view, that our most fundamental moral rights are, quite simply, rights not to be
physically harmed. This issue in tort theory has an analogue in criminal law theory.
The distinction between duties not to cause harm and duties not to attempt to cause
harm is the crux of the fundamental distinction between so-called objectivist and
subjectivist theories of the criminal law. Framing the issue in terms of duties rather
than rights, John Gardner characterizes the basic question in moral theory as a

28
See Perry, “Risk, Harm, Interests, and Rights” (note 26) at 192–3, 197–8; see also Perry, “Responsibility
for Outcomes, Risk, and the Law of Torts” (note 26) at 97–101.
50 STEPHEN PERRY

distinction between a duty to succeed, meaning in this context a duty to succeed in not
causing harm, and a duty to try, meaning a duty to try to avoid causing harm.29
Adopting language that is more amenable to the issue as it arises in tort law, Arthur
Ripstein and Benjamin Zipursky draw a roughly similar distinction between what they
call duties of “non-injury” and duties of “non-injuriousness.”30 It is, as I say, intuitively
appealing to suppose that our most fundamental rights are rights not to be harmed, as
opposed to rights that others try not to harm us or rights that they not subject us to
certain risks of harm, and this is certainly the view that seems to be embedded in
ordinary language and in the language of the criminal law. Criminal law follows
ordinary language in defining most offences in terms of causally complex acts like
killing, wounding and stealing, and the subjectivist has his work cut out for him in
showing that what the criminal law is really up to is prohibiting acts of trying to kill or
wound, or acts that risk killing or wounding.31 Despite the clear intuitive appeal of the
view that our most fundamental moral duties are duties not to cause harm, and that
our most fundamental moral rights are correlative rights not to be caused harm, this is
nonetheless a very difficult issue in moral philosophy, which I cannot fully address
here.32 But we must venture at least some way into this difficult philosophical terrain
in order to arrive at a proper understanding of both the morality of risking and the role
of the duty element in negligence law.
On one view, which has been defended by the moral philosopher Judith Jarvis
Thomson33 and, in a slightly different form, by the libertarian legal theorist Richard
Epstein in his early, rights-based work on torts,34 our most fundamental moral right is
a right not to be harmed tout court, which would translate legally into a standard of
absolute liability. I have argued elsewhere against this view on the grounds that, inter
alia, any such right would be unavoidably indeterminate in content and morally
extremely unattractive.35 I will not repeat those arguments here. The point to which

29
John Gardner, “Obligations and Outcomes in the Law of Torts,” in Peter Cane and John Gardner (eds.),
Relating to Responsibility: Essays in Honor of Tony Honoré on His 80th Birthday (Oxford: Hart Publishing,
2001), 120–5. The language of a duty to try to avoid causing harm seems more at home in the criminal law than
in the law of torts. Gardner nonetheless argues that the duty of care in negligence law involves a duty to try, in
his relatively technical sense of that term. For persuasive criticism of this view, see Ori Herstein, “Responsibility
in Negligence: Why the Duty of Care is not a Duty ‘To Try’,” 23 Canadian Journal of Law and Jurisprudence
403 (2010).
30
See Arthur Ripstein and Benjamin Zipursky, “Corrective Justice in an Age of Mass Torts,” in Gerald
J. Postema (ed.), Philosophy and the Law of Torts (New York: Cambridge University Press, 2001), 219.
31
For a distinguished recent defense of the subjectivist view in criminal theory, see Stephen J. Morse,
“Reason, Results, and Criminal Responsibility,” 2004 U. Ill. L. Rev. 363 (2004).
32
For philosophical arguments in favor of the intuitively appealing view, see the valuable discussions in
Heidi M. Hurd, “What in the World is Wrong?,” 5 Journal of Contemporary Legal Issues 157 (1994), and in
John Gardner, “Obligations and Outcomes in the Law of Torts” (note 29) at 134–43.
33
J.J.S. Thomson, The Realm of Rights (Cambridge, MA: Harvard University Press, 1990), 227–48.
34
See, e.g., Richard Epstein, “A Theory of Strict Liability,” 2 Journal of Legal Studies 151 (1973).
35
See Stephen R. Perry, “The Impossibility of General Strict Liability,” 1 Canadian Journal of Law &
Jurisprudence (1998); Stephen R. Perry, “Libertarianism, Entitlement, and Responsibility,” 26 Philosophy &
Public Affairs 351 (1997). Epstein has recently responded to these criticisms in Richard Epstein, “Toward
a General Theory of Tort Law: Strict Liability in Context,” 3:(1) Journal of Tort Law (Online) (2010). For
 TORTS , RIGHTS , AND RISK 51

I wish to draw attention for present purposes is that a right not to be harmed need not
be a right not to be harmed full stop, but can instead be—to adopt the useful
terminology of Ripstein and Zipursky36—qualified in various ways. Thus, the most
fundamental moral right underlying negligence law is, very roughly, a right not to be
harmed as a result of someone else acting negligently towards one. The correlative
moral duty is a duty not to harm others as a result of acting negligently towards them.
We can begin to fill out this qualified moral right as a right not to be harmed as a
proximate result of having been subjected to an unreasonable degree of risk. This is still
only a first approximation of the underlying moral right because the idea of acting
negligently “towards” another, or “subjecting” another to (unreasonable) risk needs to
be further unpacked, so as to take account of the distinctive role played in negligence law
by the notion of reasonable foreseeability.37 This notion is an essential aspect of the duty
of care which was famously addressed by Cardozo CJ in Palsgraf and by Lord Atkin in
Donoghue v Stevenson. I will take up this aspect of what it means to act negligently
towards another in the following section. In the meantime, I will simply note that the
manner in which the fundamental moral right underlying negligence law is qualified
entails, among other things, that an actor does not violate her correlative moral duty
unless her conduct falls below the standard of reasonable care. The standard of reason-
able care is a mandatory legal norm that applies to behavior and behavior alone, in the
sense that it is stated independently of whether or not the undesired outcome—physical
harm, in this case—occurs or not. One can comply with the standard of care and still
cause harm, and one can fall below the standard of care without causing harm. Because
violation of a behavioral norm like this is a necessary condition for the tort of negligence
to be committed, negligence is properly characterized as a fault-based tort. But the
manner in which a moral right not to be harmed can be qualified does not necessarily
require the breach of a mandatory norm in this sense, and for that reason there can be,
certainly as a conceptual matter, true torts of strict liability.
In American law, a vigorous version of strict liability can be found in the doctrine of
abnormally dangerous activities. According to this doctrine, the qualification of the
relevant right not to be harmed simply involves, in the first instance, the fact that the
defendant engaged in an abnormally dangerous activity (the paradigmatic example
of which is blasting). An abnormally dangerous activity is one that (1) “creates a
foreseeable and highly significant risk of physical harm even when reasonable care is
exercised by all actors” and (2) “is not of common usage.”38 Notice that, though this

further criticism of the Epsteinian view, see John Oberdiek, “Specifying Rights Out of Necessity,” 28 Oxford
Journal of Legal Studies 127 (2008).
36
Ripstein and Zipursky, “Corrective Justice in an Age of Mass Torts” (note 30) at 219–20.
37
This first approximation also needs to be further filled out because more analysis is needed to understand
when a risk is “unreasonable” (element of standard of care), and when the result of a tortious act is “proximate”
(element of proximate cause). I have made a start on discussing these elements in Perry, “Responsibility for
Outcomes, Risk, and the Law of Torts” (note 26) but say nothing about them here. In the present chapter my
main doctrinal concern is the duty of care.
38
Restatement (Third) of Torts, } 20.
52 STEPHEN PERRY

definition mentions the standard of reasonable care, the standard plays only a negative
role: The idea is that one has a duty not to cause physical harm as a result of engaging
in a certain kind of activity even if one took reasonable care. Unlike the tort of
negligence, strict liability does not involve a duty of care. This is because, in the case
of strict liability, the defendant has no duty, not even a derivative one, not to create
unreasonable risks, but instead has a duty not to cause, to reasonably foreseeable
persons or classes of persons, sufficiently proximate physical harm as a result of
engaging in a certain type of activity. Whether or not one has complied with the
standard of reasonable care is neither here nor there. A number of prominent rights
theorists, including Ernest Weinrib and Alan Brudner, have argued that, doctrinal
appearances to the contrary notwithstanding, most so-called torts of strict liability
really involve a fault standard, and that any other understanding would be morally
incoherent or otherwise morally problematic.39 Although I cannot take up the issue in
detail here, I believe that they are mistaken in this regard.40 It is important to
emphasize that arguments for the view advocated by Weinrib and Brudner must be
made on moral grounds, and once it is accepted that our fundamental moral rights
are rights not to harmed in certain qualified ways, it is easy to see that the tort of
negligence and true torts of strict liability have much more in common than is often
thought to be the case.41 Put crudely, the difference between them simply concerns the
manner in which fundamental moral rights not to be harmed are qualified. Contrary
to what is sometimes suggested, it is not the case even in American tort law, which is
dominated by instrumentalist thinking, that torts of strict liability can only be
defended on instrumentalist grounds.
To summarize: the most important torts that protect against physical harm involve
a duty not to cause harm, but not a duty not to cause harm full stop. A duty not to
cause harm full stop would entail absolute liability, which is unacceptable both
conceptually and morally. The duty not to cause harm must be qualified in one or

39
Weinrib, The Idea of Private Law (note 2) at 1284–1303; Alan Brudner, The Unity of the Common Law:
Studies in Hegelian Jurisprudence (Berkeley: University of California Press, 1995), 190.
40
Let me very briefly address one of Weinrib’s arguments. Weinrib maintains that “[t]o ascribe liability to
an action, regardless of culpability, for whatever harmful effects it has simply because they are its effects, is to
hold the agent liable for being active.” Since, Weinrib continues, strict liability judges action by its effects in just
this way, it “treats the defendant’s agency as an incoherent normative phenomenon”: Weinrib, The Idea of
Private Law (note 2) at 181.
It is a mistake, however, to think that strict liability “hold[s] the agent liable for being active.” The
correct normative characterization of strict liability is that it holds persons to an appropriately qualified
duty not to cause harm of some specified type. Liability for violation of such a duty is thus liability for
causing harm as a result of engaging in certain activity, where so engaging is the appropriation qualifi-
cation on the duty; it is not liability for simply “being active,” which would be absolute, not strict, liability.
If one wishes to speak in terms of “ascribing liability to an action,” the appropriate characterization of the
action will be a causally complex description that includes the effect in question (for example, an act of
blasting, or an act of killing).
41
This point is made particularly well by John Gardner. See Gardner, “Obligations and Outcomes in the
Law of Torts” (note 29) at 123–5.
 TORTS , RIGHTS , AND RISK 53

the other or both of two possible ways. The first way, which is typified by most forms
of strict liability, is to restrict the duty to persons engaging in a particular activity or
type of activity, such as abnormally dangerous activities, where the risks associated
with the relevant activities are both foreseeable and uncommonly serious. The second
way, which is typified by negligence law, is to embed a secondary duty, namely, a
mandatory standard of care, into the duty not to cause harm. This talk of “embedding”
a standard of care into a duty not to cause harm is illuminating, because it helps clarify
the structure of the relevant complex duty, but it can also be misleading, because it
suggests that the complex duty—what I shall call, with apologies to Lewis Carroll, a
“portmanteau” duty and its correlative “portmanteau” right—is literally constituted of
simpler elements, namely, among other things, a duty not to harm others and a norm
specifying that one not subject others to unreasonable risks. This is potentially
misleading because neither the duty not to cause harm nor the standard of reasonable
care can morally exist on its own so as to constitute the building blocks, as it were, of
the complex duty. The complex duty—i.e., as characterized so far, the duty not to
harm others as a result of acting negligently towards them—is, despite being complex,
morally fundamental. The duty not to harm cannot stand on its own because, among
other things, it would entail absolute liability and as such would be radically indeter-
minate. If there is any sense in which the duty of reasonable care can stand on its own,
it is because it is a secondary or ancillary duty which is morally derived from the
complex, portmanteau duty.
It has been suggested that Cardozo, in adopting in Palsgraf a relational and hence
rights-based understanding of the duty of care, committed himself to the position that
the fundamental moral right underlying negligence law is a right not to be subjected to
certain risks, rather than, as I have been suggesting, a right not to be caused harm as a
result of another person acting negligently towards one. Thus Heidi Hurd and Michael
Moore summarize what they take to be Cardozo’s chain of inferences in the following
fashion:
[N]egligence implies (in the sense of presupposes) wrongdoing; wrongdoing implies a
rights-violation; a rights-violation occurs only when the rights-holder is unreasonably
risked, not when the rights-holder is caused harm. This latter is true because a victim’s
rights (with respect to non-intentional interference) are rights against being placed at risk
of harm.42

This is, in my view, an uncharitable and forced interpretation of Cardozo’s reasoning

in Palsgraf.43 Although he may not have stated in so many words that the fundamental

42
Hurd and Moore, “Negligence in the Air” (note 21) at 16.
43
I believe that Hurd and Moore’s interpretation of Cardozo goes astray for the following reason. While
they accept, correctly, that our most fundamental moral rights are rights not to be caused harms of various
kinds, they apparently do not accept that these rights can and indeed must be subject to inherent qualifications
of the sort I have discussed in the text. Thus on the view I have outlined, the right not to be harmed that
underlies negligence law is partly constituted by the qualification that the right is only violated if the actor acted
negligently (i.e. unreasonably) towards the person harmed. Hurd and Moore’s view is, rather, if I understand
54 STEPHEN PERRY

underlying moral right at stake in negligence law is a qualified right not to be harmed,
rather than a right not to be subject to certain risks, the former understanding of rights
clearly informs his thinking throughout his opinion, as for example when he states that
“bodily security is protected, not against all forms of interference or aggression, but
only against some.”44 The point here is that it is bodily security itself that is protected
against interference, not a supposed interest in not being subject to risk of bodily
injury. Cardozo made the point explicitly and very clearly in his earlier decision in
MacPherson: “A poison falsely labelled is likely to injure anyone who gets it. Because
the danger is to be foreseen, there is a duty to avoid the injury.”45 Lord Atkin made the
point just as clearly in Donoghue v Stevenson, when he famously wrote that “[t]he rule
that you are to love your neighbour becomes in law, you must not injure your
neighbour.”46
It is true that Cardozo’s doctrinal focus in Palgraf was on the duty of care, and it is of
course also true that an actor can breach her duty of care even without causing
physical injury. How that fact about doctrine can be reconciled with the view that
the most fundamental moral right underlying negligence law is a qualified right that
others not cause one physical harm by reason of acting negligently towards one, rather
than a right that others simply not subject one to an unreasonable degree of risk of
physical harm, is the issue to which we turn next.
It will be helpful to begin by saying something more about the concept of harm.
Harm, as I have argued elsewhere,47 is a relatively specific moral concept which
requires that a person have suffered serious interference with one or more interests
that are particularly important to human wellbeing, and which for that reason are
appropriately designated as fundamental. Interests that can plausibly be thought to
fall into this category include life, health, dignity, the physical integrity of the body,
autonomy and freedom of movement, the interest in not experiencing severe pain, the
interest in not experiencing severe mental or emotional distress, and certain kinds of
property interest. Not surprisingly, these interests overlap to some extent. It is also not
a coincidence that this set of interests, or some very similar set, turns out to be
of primary concern for both deontological approaches to morality and to theories of

them correctly, that the fundamental right underlying negligence law is (like all fundamental rights) a right not
to be harmed full stop. Violations of this fundamental right do not, on their view, have moral or legal
implications unless they are accompanied by some form of culpability, which in the case of the criminal law
is usually one of the standard forms of mens rea and in the case of negligence involves a breach of the standard
of reasonable care. The fundamental, essentially free-standing right they posit is violated whenever one
individual causes harm to another individual, but this rights-violation does not have moral or legal implica-
tions unless an appropriate form of culpability is present. While I cannot discuss the issue in detail here,
I believe that Hurd and Moore go wrong in thinking that fundamental rights and their correlative duties
cannot be qualified along the lines I have been discussing. Their error is, in effect, to adopt a conception of
rights which is too closely oriented to the distinction in criminal law between actus reus and mens rea.
44 45
Palsgraf, 248 N.Y. at 345. MacPherson, 217 N.Y. at 385 (emphasis added).
46
Donoghue v Stevenson [1932] AC 562 (HL) 580 (emphasis added).
47
See Stephen Perry, “Harm, History, and Counterfactuals,” 40 San Diego L. Rev. 1283 (2003); Perry, “Risk,
Harm, Interests, and Rights” (note 26).
 TORTS , RIGHTS , AND RISK 55

rights that are at least partly nonconsequentialist in character. Nor is it a coinci-

dence that these same interests also figure prominently among the interests
that are protected by both criminal law and torts, which are the areas of law that
are most plausibly thought to have a deontological or nonconsequentialist norma-
tive structure.
I have suggested that there are certain aspects of human wellbeing which can be
regarded as fundamental or core moral interests, and that setbacks to these interests, or
at least sufficiently serious setbacks, constitute harm. In Section III, I argued that risk,
by itself, is not harm, which should now be taken to mean that it does not constitute a
setback to a fundamental or core interest of the kind I have just been discussing. Every
harm is of course a setback to an interest, but I have not suggested, nor is it plausible to
think, that all setbacks to interests are harms. Suppose I am right in thinking that the
imposition of a risk does not, by itself, constitute a setback to any fundamental
interests.48 Is it not the case, it might be asked, that there must still be some sense in
which it is against my interests to live, say, with a rusting barrel of toxic waste under
my house? I have argued elsewhere that the answer to this question is yes, because the
very existence of a set of core interests necessarily gives rise to secondary interests of
various kinds, and at least some of those interests will be second-order interests,
meaning interests that I have in my interests.49 To see that this is so, consider a type
of case other than risking. Suppose you attempt to physically injure me but fail. To
avoid unnecessary complications, assume that at no time was I aware of your attempt
and that your action was not of a kind that interferes with my dignity. Your failed
attempt thus does not set back any of my fundamental interests in life, bodily integrity,
autonomy, and so on, and therefore cannot plausibly be thought to have caused me
any harm. Even so, I wish to claim, it sets back a second-order interest I have that
persons not even attempt to set back those first-order, fundamental interests. As it is
with attempts, so it is with risk: I have a second-order interest that others not subject
me to risk, meaning a second-order interest that they not engage in actions that are
unreasonably risky. And just as failed attempts do not necessarily cause me any harm,
neither do actions that subject me to risks but that do not, in the event, lead to any
injury.50 But even though (unreasonably) risky actions do not in and of themselves
harm me, it is at least arguable that the second-order interest that I have in not being
subject to (unreasonable) risk is, on an interest theory of rights,51 protected by a right.

48
I have also argued elsewhere that a setback to welfare, and, in particular, a setback to a preferentialist or
partially preferentialist conception of welfare, does not constitute harm in the sense that I have characterized in
the text. It follows that the fact that I would prefer not to live with a certain risk does not show that imposing
that risk on me is, eo ipso, harm. See Perry, “Risk, Harm, Interests, and Rights” (note 26).
49
Perry, “Harm, History, and Counterfactuals” (note 47) at 1307–9.
50
Heidi M. Hurd suggests that it follows that there cannot be a right not to be subject to a risk. Hurd, “The
Deontology of Negligence” (note 21) at 267. I criticize her view in Perry, “Responsibility for Outcomes, Risk,
and the Law of Torts” (note 26) at 75–81.
51
See, e.g., Joseph Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986), 166 (“ ‘X has a right’ if
and only if X can have a right and, other things being equal, an aspect of X’s well-being (his interest) is a
sufficient reason for holding some other person(s) to be under a duty.”).
56 STEPHEN PERRY

The interest I have in not being subject to unreasonable risk may not by itself be strong
enough to protect with a right, but there is instrumental value in protecting such an
interest because unreasonably risky actions have a higher probability of leading to
harm than do actions which are not unreasonably risky. Notice that this reliance on
there being a higher probability of harm, i.e., a higher degree of risk, does not
undermine the proposition that risk is not harm. The epistemic character of risk
means that assessments of risk carry information, and we should use whatever
modicum of information that we possess or ought to possess so as to prevent
unwanted physical harm; it simply does not follow that the use of such information
about risk to specify second-order rights which indirectly protect against physical
harm entails that risk is any sense a harm in itself.
Barbara Fried has written that nothing “operationally” turns on the distinction
I have drawn between first- and second-order interests and rights, but that is not so.52
There is a saying in philosophy that second-order is second-best,53 and a second-order
right that protects physical interests indirectly and instrumentally is simply not as
important, morally speaking, as the first-order, portmanteau right that others not
cause one physical harm as a result of subjecting one to unreasonable risk. In terms of
legal doctrine, the second-order, derived right that one not be subjected to unreason-
able risk is correlative of the duty of care. Breaching one’s duty of care is not the same
as committing the tort of negligence, since the latter requires, inter alia, causation of
harm; for that reason one can breach one’s duty of care without giving rise to liability
of any kind. At least part of the explanation for this is that the second-order right that
one has not to be subject to unreasonable risks is, morally speaking, simply less
important than, because it is derivative of, the first-order right not to be harmed as a
result of another’s unreasonably risky action. For example, the latter right receives
legal protection, in the form of a remedy if the right is violated, in a way that the
former does not. A similar response to the one I have made to Fried can also be offered
to John Oberdiek’s objection that my argument about second-order interests and
rights “cheapens” rights because there can in principle be “ever-higher orders” of
rights, and “[t]he resulting proliferation would force us to take rights less seriously in
practice than we should intuitively take them.”54 It is true that my account entails that
there can be higher and higher order interests, but, just as second-order is second-best,
so third-order is third-best, and so on. The moral, legal, and practical significance of
higher order interests runs out sufficiently quickly that it becomes quite implausible to
think that there are higher order rights above the second level or so. The view does not
lead to the runaway proliferation of rights that Oberdiek suggests.
Suppose, however, that I am wrong to think that there is a second-order interest in
not being subject to unreasonable risks of physical harm, and that for that reason there

52
Barbara Fried, “The Limits of a Nonconsequentialist Approach to Torts,” 18 Legal Theory 231 (2012),
247–8.
53
Thomas Hurka brought this saying to my attention.
54
John Oberdiek, “Towards a Right Against Risking,” 12 Law and Philosophy 367 (2009), 381–2.
 TORTS , RIGHTS , AND RISK 57

can be no second-order right not to be subject to unreasonable risks. In my view the

idea that we have interests in our interests is a natural and intuitive one, but on a
different view perhaps it might appear to be merely a technical maneuver which has
been adopted to facilitate the desired conclusion, namely, that there can be pure moral
rights not to be risked which are instrumentally valuable because they promote the end
of avoiding physical harm. However, it seems to me that one can also justify an
instrumentally valuable right not to be risked not only by appealing to a second-
order interest, but also by appealing directly to the interests one has in not being
physically harmed in one’s person or property. After all, the interest theory of rights, as
characterized by Joseph Raz, holds that a person has a right if and only if an aspect of
the person’s well-being—his interest—“is a sufficient reason for holding some other
person(s) to be under a duty.”55 There is no obvious reason to think that, on such a
theory of rights, the interest ultimately protected by a right must itself actually be set
back by each and every action that violates the right. Assuming that my fundamental
interests in life, dignity, bodily integrity, autonomy, and so on are sufficiently import-
ant to justify, say, a right that other persons not intentionally harm me in my person or
property, there is no reason in principle why those same interests cannot be suffi-
ciently important to justify a right that others not attempt to injure me, or a right that
they not subject me to unreasonable risks. The truth of this statement is not affected by
the fact that many attempts and most unreasonably risky actions do not, in fact, cause
any injury at all.
Suppose, however, that I am wrong even about this, and that there is simply no such
thing as a moral right not to be subject to pure risk (i.e., risk that does not lead to
harm). This would have minimal impact on our understanding of the law of negli-
gence, because, I have argued, the right underlying the tort of negligence is a unified,
complex, portmanteau right of a certain kind, namely, a right not to be caused
reasonably foreseeable physical harm as a proximate result of another person’s
engaging in unreasonably risky conduct towards one. If there were a moral right not
to be subject to pure risk, it wold be derivable from this more complex but morally
more fundamental right, and not the other way around. The resulting lack of corres-
pondence between the doctrinal duty of care—which, as previously noted, is framed as
a duty not to risk rather than a duty not to harm—and the (by hypothesis nonexistent)
moral duty not to subject others to risk would not pose a serious problem for a rights-
based theory. It would simply become necessary explicitly to recognize that the basic
elements of the tort of negligence—duty of care, standard of care, injury, cause in fact
and proximate cause—amount, in essence, to an artificial separation of different
aspects of what is, morally speaking, a single complex duty (and correlative moral
right). It is true that the fact that this separation is an artificial, doctrinal one would
place the standard claim of rights theorists to take seriously the structure of tort law in
a somewhat curious light, since the duty of care in law would not then directly

55
Raz, The Morality of Freedom (note 51) at 166.
58 STEPHEN PERRY

correspond to a distinct duty in morality. But it would correspond to one aspect of the
actual moral duty, and that seems to me to be sufficient to sustain a plausible, rights-
based interpretation of negligence law. From the practical perspective of the litigation
process, there is obviously an enormous amount to be said in favor of artificially
separating different aspects of the single right into distinct doctrinal requirements that
the plaintiff must meet one by one. That is a form of instrumentalism, but in my view
it is instrumentalism of a kind that should be completely acceptable to a rights theorist.
The question of whether or not there is ever a moral right not to be subject to pure
risk is a complex and controversial one, and I have not even begun to exhaust the
subject. For present purposes, however, I will limit further discussion to a number of
observations about Oberdiek’s own very interesting theory of when there is such a
right. Oberdiek begins by arguing that “[w]hether [pure] risk impositions themselves
are capable of being impermissible or wrong is . . . the central question of the moral
significance of risking.”56 It is far from clear that this is true, since it ignores the
possibility, defended in this chapter, that the central moral significance of risk is its
role in the fundamental complex right described in the preceding paragraph. As we
have seen, a right not to be subject to risk as such may or may not be derivable from
this more fundamental right, but, either way, risk obtains all the moral significance
that it will ever need from the fact that it figures as an element in this right. Oberdiek is
thus also mistaken in arguing that “[u]nless imposing risk can be impermissible,
unjustified risking—which is to say negligence—is literally impossible.”57 The tort of
negligence is made “possible” by its direct correspondence to the complex right.
Negligence in the sense of breaching one’s duty of care is, as we have seen, made
possible either as a moral right derived from the more complex right, or as a
doctrinally (and therefore artificially) recognized aspect of the complex right.
But let us set aside the complex right for the moment. Oberdiek argues powerfully
that there is an understanding of risking which does not locate “[its] moral signifi-
cance . . . in its perceived potential for harm.”58 (This latter kind of approach, a version
of which I have defended, is dismissed as a commonsense view.) Oberdiek maintains
that “[p]ure risk impositions are morally significant precisely because of their negative
but nonmaterial effect on autonomy, whose integrity is of momentous concern.”59
A pure risk imposition is one that does not lead to physical harm. Autonomy is
understood in Raz’s sense of having a range of acceptable options which is large
enough to permit one to be the author of one’s own life.60 To take one example, if
I pose risks of physical injury to you by laying mines in a field where you habitually
walk, I have interfered with your autonomy by removing a certain number of safe
options from your range of otherwise available options, and this is true whether or not

56
John Oberdiek, “The Moral Significance of Risking,” 18:(3) Legal Theory 339 (2012), 344.
57
Oberdiek, “The Moral Significance of Risking” (note 56), 339.
58
Oberdiek, “The Moral Significance of Risking” (note 56) at 343.
59
Oberdiek, “The Moral Significance of Risking” (note 56) at 354.
60
Raz, The Morality of Freedom (note 51) at 369–99.
 TORTS , RIGHTS , AND RISK 59

you ever step on a mine, and even if you are not aware of the mines. Imposing pure
risks is, in Oberiek’s metaphor, like laying traps for a person—a metaphor that is
particularly apt in the case of the mine example—and if you set enough traps, you can
“utterly annihilate”61 the other person’s autonomy even if the person is never phys-
ically injured. Elsewhere, Oberdiek has argued that because of the importance of
autonomy, in the sense of a range of acceptable options in life, morality recognizes a
right against (certain instances of) pure risking.62
I agree with Oberdiek that risking can have a “nonmaterial effect” on autonomy (i.e.,
an effect that does not flow from physical injury). But I disagree that the central “moral
significance” of pure risking consists in its effect on autonomy. I take it that Oberdiek
would agree that the central moral significance of non-pure risking—i.e., risking that
does lead to physical harm—is precisely the fact that it has led to physical harm. But
one cannot tell in advance which instances of risking will lead to physical harm and
which will not. The duty of care and standard of care in negligence law—whether or
not they together correspond to a distinct moral right against risking—do not distin-
guish between pure and impure risking for the very good reason that one cannot
distinguish between them in advance. It is in fact very difficult to see how one could
differentiate between a right not to be risked because of the potential for physical harm
and a right not to be risked because of the potential for nonmaterial effects on
autonomy, since both cases are concerned in the first instance precisely with the
potential for physical harm; after all, there is only an effect on nonmaterial autonomy
because the potential for physical harm reduces one’s safe options. I have outlined
above a number of ways in which a moral right not to be physically risked could be
instrumentally justified by reference to the complex right not to be negligently
physically harmed, and it might make sense to say that such a risk also protects, as
an ancillary matter, a broad interest in autonomy. (By a “broad” interest in autonomy,
I mean one that can be interfered with by means other than physical injury.) But a
stand-alone right that protects only the broad interest in autonomy is implausible. For
one thing, it is less important, morally speaking, than a right not to be risked because
of the potential for physical harm. Relatedly, a pure right not to be risked appears to
receive virtually no legal protection. One of the few places it does receive protection in
the law of torts is from the tort of false imprisonment (and, arguably, from the tort of
assault), but false imprisonment is an intentional tort which is justified in ways quite
different from the tort of negligence. It might be argued that a right against pure
risking receives protection from the law of negligence because it corresponds to the
duty of care, but I have already argued that the duty of care (and its correlative right) is
better understood in other ways; in any event, breach of the duty of care does not give
rise to liability on its own. In his article, Oberdiek refers at a number of points to the
fact that drunk driving is a wrong. However, it should be noted that one can be guilty
of criminal drunk driving even if no individual was put at risk by one’s activity.

61
Oberdiek, “The Moral Significance of Risking” (note 56) at 352.
62
Oberdiek, “Towards a Right Against Risking” (note 54).
60 STEPHEN PERRY

In addition, criminal drunk driving that results in no physical injury does not give rise
to a right in criminal law on any individual’s part. Criminal laws prohibiting drunk
driving are best regarded as measures that protect general public safety, as opposed to
measures that only protect individual rights.

V. The Role of Reasonable Foreseeability

I said in the preceding section that the fundamental moral right underlying negligence
law is, roughly, a right not to be caused reasonably foreseeable physical harm as a
proximate result of another person’s engaging in unreasonably risky conduct towards
one. I also said that this rough approximation of the underlying moral right needs to
be further unpacked so as to take account of the distinctive role played by the notion of
reasonable foreseeability. Foreseeability of not just the type of harm but also of a
person or class of persons placed at risk by the defendant’s conduct was central to both
Lord Atkin’s and Cardozo’s view of the relational nature of the duty of care. Thus the
proposition that the defendant engaged in unreasonably risky conduct “towards” the
plaintiff must be further unpacked so as to include the idea that the defendant ought to
have foreseen that he or she was acting so as to create an unreasonable risk of
proximate physical harm for a reasonably foreseeable class of persons that included
the plaintiff. It is, of course, true that an actor can only take account of the potentially
harmful effects of her actions if she has the capacity to foresee that her actions might,
in fact, cause harm to another person. But instrumentalists often argue that foresee-
ability is appropriately dealt with in the calculus of risk which is, they claim, built into
the standard of care. The Third Restatement endorses this view, and explicitly rejects
the idea that foreseeability has a role to play at the doctrinal level of duty.63 What is
perhaps surprising is that some rights theorists essentially agree with the instrumen-
talists on this point.
Even such prominent defenders of a rights-based, relational understanding of tort
law as Goldberg and Zipursky appear to downplay the role of reasonable foreseeability
in determining a defendant’s duty of care, and they interpret Cardozo’s landmark
opinions in MacPherson and Palsgraf as doing likewise. Goldberg and Zipursky argue,
correctly, that as a conceptual matter, even an understanding of the duty of care that is
relational, in the sense of being rights-based, could be universal, in the sense that the
duty is owed to every other individual in the world. In the famous language of
Andrews in his dissenting opinion in Palsgraf, the duty, while relational, would

63
See note 11. Recall that the Third Restatement implicitly rejects the idea of a relational duty, and does not
even employ the traditional terminology of a “duty of care.” Duty is envisaged, rather, simply as a general duty
to conform one’s conduct to the standard of reasonable care. See note 10. Thus the Restatement’s view about
the relationship between duty and foreseeability is, strictly speaking, the view that foreseeability should not be
taken into account in so-called “no-duty rulings”; it should only be taken into account in determinations by the
trier of fact that the standard of reasonable care has been breached.
 TORTS , RIGHTS , AND RISK 61

nonetheless be “owed to the world at large.”64 Goldberg and Zipursky argue that,
doctrinally and morally, the duty of care is not, generally speaking, universal in this
sense, and they interpret Cardozo as recognizing as much in Palsgraf. But they
nonetheless suggest that, in the core cases of causing physical injury or property
damage to others, duties of care are “universal . . . in the relational but unrestricted
sense”: “These are duties of care each person owes to each other person.”65 There is a
tension here, however, at least insofar as they are attributing this view to Cardozo,
because Cardozo clearly regarded Palsgraf as just such a core case. Goldberg and
Zipursky suggest that, in Palsgraf, “[c]learly the railroad did owe a duty of care to its
customer and there was no need for any discussion of reasonable foreseeability in
order to establish this conclusion.”66 While it is certainly true that, as a general
doctrinal matter, the special relationship of common carrier and passenger gives rise
to a duty of care, the fact remains that this consideration played absolutely no role in
Cardozo’s reasoning in the case. He clearly would have reached exactly the same
conclusion as he did even if Mrs Palsgraf had been a stranger and a bystander to the
defendant, for example if she had happened to be injured as a result of flying debris
that hit her while standing outside the station property. Cardozo CJ held against Mrs
Palsgraf because it was not reasonably foreseeable to the defendant that she was, when
considered either as an individual or as a member of a class of persons, at risk of being
physically harmed by its negligent action, and on this basis he concluded that the
defendant did not owe her a duty of care.67

64
Goldberg and Zipursky, “The Moral of MacPherson” (note 4) at 1821–2, correctly recognize that
Andrews J himself was not defending a relational or rights-based understanding of tort law, but was rather
drawing on the instrumentalist views that even in 1928 were already strongly influential in the United States.
Hurd and Moore, “Negligence in the Air” (note 21), mistakenly interpret Andrews J as defending the view that
the duty of care is both relational and universal.
65
Goldberg and Zipursky, “The Moral of MacPherson” (note 4) at 1833. It is an interesting question why
Cardozo CJ did not take account of the fact that Mrs Palsgraf was the railroad’s customer. Possibly he wanted
to take the opportunity to emphasize that, in all types of negligence cases, reasonable foreseeability is a
necessary determinant of the duty of care.
66
Goldberg and Zipursky, “The Moral of MacPherson” (note 4) at 1820. It should be noted, however, that in
another article Zipursky argues vigorously for the view that foreseeability has a role to play in duty determin-
ations, but adds the caveat that such determinations take place at a “category level” rather than at the level of
individual cases. See Benjamin C. Zipursky, “Forseeeability in Breach, Duty, and Proximate Cause,” 44 Wake
Forest L. Rev. 1247 (2009), 1264.
67
Other rights theorists go much further than Goldberg and Zipursky in rejecting a role for reasonable
foreseeability in the determination of the duty of care. See, e.g., Dilan Esper and Gregory Keating, “Putting
‘Duty’ in Its Place: A Reply to Professors Goldberg and Zipursky,” 41 Loy. L.A. L. Rev. 1225 (2008), 1243–55.
Esper and Keating defend what they call a “weakly relational” view of duty, which they contrast with the
“strongly relational” view that they attribute both to Goldberg and Zipursky and to Cardozo CJ in Palsgraf.
According to Esper and Keating, a strongly relational view mistakenly views duty as “personal,” in the sense of
involving an obligation that “runs from this named defendant to this named plaintiff”; in their view,
“particularizing” duty in this way has the effect, among other things, of eroding the line between judge and
jury. Esper and Keating, “Putting ‘Duty’ in Its Place” at 1250–1. They reject outright Cardozo CJ’s view of duty
of care, arguing that “duty is owed to everyone . . . and it is ordinarily triggered simply by acting in a way
that poses a ‘reasonably foreseeable’ risk of harm to anyone at all,” Esper and Keating, “Putting ‘Duty’ in Its
Place” at 1255. Esper and Keating’s defence of their view is complex, and cannot be dealt with properly here.
62 STEPHEN PERRY

Goldberg and Zipursky argue that the significance of the fact that harm to Mrs
Palsgraf was not foreseeable was not that she was not owed a duty of care, but rather
that the defendant simply had not, as a matter of law, breached the duty it owed her.68
The “subtle” duty issue raised by the case is, they claim, “why it should matter that the
railroad did not breach its duty to Mrs Palsgraf, given that the railroad was already
presumed to have breached the duty of care it owed to the other passenger.”69 The
answer, they maintain, is simply that the notion of duty should be viewed relationally
rather than non-relationally. Cardozo took the first view of the matter, while Andrews
took the second. This is correct, so far as it goes, but it is simply irrelevant to the
further question of whether or not reasonable foreseeability has a substantive role
to play in determining the existence of a duty of care in the relational sense. In my
view Cardozo is most naturally interpreted as holding, in both MacPherson and
Palsgraf, that the answer to this question is “Yes.” Cardozo’s general view was that it
is a necessary condition for the existence of a duty of care in negligence law that it be
reasonably foreseeable to the defendant that the defendant’s contemplated action place
the plaintiff, either individually or as a member of a class of persons, at risk of suffering
harm. In ordinary misfeasance cases involving physical harm between strangers,
reasonable foreseeability will generally be not just a necessary, but also a sufficient
condition for the existence of a duty of care. In cases involving some special feature—
for example, an affirmative rather than a negative duty, or the fact that the harm was of
a nonstandard kind, such as economic loss or emotional distress—reasonable foresee-
ability of harm to the plaintiff will still be a necessary condition for the existence of a
duty of care, but it will not generally be a sufficient condition. For example, there is,
traditionally, no general right in negligence law not to be caused pure economic loss.
The fact that it was reasonably foreseeable to the defendant that the plaintiff was at risk
of suffering such loss is thus not sufficient to establish a duty of care; some further
factor, such as the fact that the defendant gave the plaintiff an undertaking of an
appropriate kind, is required.70
In English and Commonwealth law, it has been clear since Donoghue v Stevenson
that reasonable foreseeability is always a necessary condition for the existence of a duty
of care, and subsequent case law has emphatically confirmed that answer. In fact the
clear tendency in modern English and Commonwealth law has been to add, under
the head of “proximity,” further restrictions, generally of a rights-based character, to

The only point I would make for present purposes is that it is a mistake to think that the difference between a
duty of care that is relational and universal and one that is relational but non-universal is that in the latter case,
but not the former, the duty is “personal.” Rights and duties in the relational sense are always personal, but it is
sufficient to establish this that the plaintiff belonged to a class of persons who were placed at risk by the
defendant’s conduct. Neither Cardozo CJ nor Goldberg and Zipursky can be plausibly interpreted as holding
the view that duty of care must run from a named defendant to a named plaintiff.
68
Goldberg and Zipursky, “The Moral of MacPherson” (note 4) at 1819–20 and 1821–4.
69
Goldberg and Zipursky, “The Moral of MacPherson” (note 4) at 1820.
70
See, e.g., Benson (note 13); Perry, “Protected Interests and Undertakings in the Law of Negligence” (note
17) at 262–302; Stevens, “Law of Torts” (note 19).
 TORTS , RIGHTS , AND RISK 63

the class of persons to whom a duty of care is owed, both in particular cases and
in categories of cases.71 But allegiance to the basic requirement that reasonable
foreseeability is always a necessary condition of the existence of a duty of care has
remained steadfast. It would be exceedingly odd if an appropriately rights-based
interpretation of American law differed so starkly on such a fundamental point from
English and Commonwealth law, which, as noted in Section II, has generally con-
tinued to lend itself much more straightforwardly to interpretation in rights-based
terms.
Doctrine in American tort law is complicated by the fact that civil cases in the
United States are still generally tried to juries, and almost certainly one reason that
many commentators support a shift of the role of reasonable foreseeability from the
duty of care to the standard of care or to the element of proximate cause is a common
belief, often shared by proponents of quite different theoretical standpoints, that
the foreseeability question should be decided by juries and not by judges.72 This belief
may well be correct, but it does not settle the answer to the following question: as a
theoretical matter, is reasonable foreseeability an appropriate consideration to be
taken into account in determining, to borrow the language of McLachlin CJ in Cooper
v Hobart, whether “the circumstances of the relationship inhering between the plaintiff
and the defendant are of such a nature that the defendant may be said to be under an
obligation to be mindful of the plaintiff ’s legitimate interests?”73 After all, even if the
determination of reasonable foreseeability should generally fall to juries, juries can be
charged conditionally. Why might one think that reasonable foreseeability should not
only play a role, but play the central role, in determining the existence of a duty of care
in negligence law?
In Donoghue v Stevenson Lord Atkin famously wrote:
I owe a duty of care to persons who are so closely and directly affected by my act that
I ought reasonably to have them in contemplation as being so affected when I am directing
my mind to the acts or omissions which are called in question.74

71
See, e.g., Cooper v Hobart (note 15) at 552.
72
This is the view explicitly adopted in the Restatement (Third) of Torts } 7, comment j. The very
complicated set of questions concerning the proper role of judge and jury in negligence cases, including in
particular the proper delineation of the duty element, is illuminatingly debated in the following two articles:
Dilan Esper and Gregory Keating, “Abusing ‘Duty’,” 79 S. Cal. L. Rev. 265 (2006); and John C.P. Goldberg and
Benjamin Zipursky, “Shielding Duty,” 79 S. Cal. L. Rev. 329 (2006). The debate is not easily summarized, and
both sets of authors make a number of insightful points. I think that Esper and Keating are correct in observing
that at least sometimes “[w]hen all the risks of physical injury created by some conduct are so remote that they
are not reasonably foreseeable, no tort duty of care is owed to anyone”: Esper and Keating, “Abusing ‘Duty’ ” at
283–4. Elsewhere, however, Esper and Keating appear to reject this view. See Esper and Keating, “Putting ‘Duty’
in Its Place: A Reply to Professors Goldberg and Zipursky” (note 67). In the latter article, they defend the view
that Cardozo CJ was mistaken to make reasonable foreseeability a central determinant of the duty of care, and
part of the reason for this is that doing so “erodes the line between judge and jury,” Esper and Keating, “Putting
‘Duty’ in Its Place” at 1250. In their view reasonable foreseeability is most appropriately taken into account at the
stage of proximate cause, and most casebook editors are therefore correct to treat Palsgraf as a proximate cause
case and not a duty of care case. Esper and Keating, “Putting ‘Duty’ in Its Place” at 1253–5, 1261–4.
73 74
Cooper v Hobart [2001] 3 SCR 537, 552. Donoghue v Stevenson [1932] AC 562 (HL) 580.
64 STEPHEN PERRY

He presented this proposition neither as self-evident nor as following from received

legal doctrine, but rather as an answer to the abstract question: “Who, then, in law is
my neighbour?” The idea that we are neighbors not just in a temporal or spatial sense,
but in a moral sense as well, is a powerful metaphor for the relational conception of
rights and duties. As Lord Atkin recognized, the morality of tort law is not the morality
of love but rather the morality of rights, and he was therefore correct to emphasize that
the lawyer’s question “receives a restricted reply.” As moral neighbors, we do not owe
one another a universal duty not to cause one another harm, and that is so even if
harm is reasonably foreseeable to someone other than the person actually harmed.
I only owe you a duty of care if you are among the persons foreseeably placed at risk by
my action. This answer is the correct one because unless I can foresee the possibility
that I might injure you, there is no way that I can take steps to avoid injuring you; to
impose a duty on me even in the absence of a capacity to avoid injuring you would not
be a proper expression of our general relationship as moral neighbors.75 It was
Holmes, in a brilliant passage in The Common Law, who gave most eloquent expres-
sion to this thought:
The requirement of an act is the requirement that the defendant have made a choice. But
the only possible purpose of introducing this moral element is to make the power of
avoiding the evil complained of a condition of liability. There is no such power where the
evil cannot be foreseen.76

There is a particular irony to the fact that this powerful formulation of the moral
centrality of foreseeability to a relational conception of tort law should have been
offered by the most influential architect of the prevailing view in the American legal
academy that tort law should be understood in instrumentalist terms.

75
I have tried to show at greater length elsewhere that possession of a capacity to avoid a certain type of
outcome is a necessary moral condition of the existence of a relational duty to avoid bringing about an instance
of that outcome. See Perry, “Responsibility for Outcomes, Risk, and the Law of Torts” (note 26) at 91–108.
76
Oliver Wendell Holmes, Mark DeWolfe Howe (ed.), The Common Law (Cambridge, MA: Belknap Press,
1963), 77.
 3
Compensation as a Tort Norm
Mark A. Geistfeld*

Honeste vivere (to live honorably), alterum non laedere (not to injure others), and
suum cuique tribuere (to render to every man his due), were the three general
precepts to which Justinian reduced the whole doctrine of the law . . . And what of
alterum non laedere? “Thou shalt do no hurt to thy neighbor.” Our law of torts,
with all its irregularities, has for its main purpose nothing else than the develop-
ment of this precept.1

I. Introduction
In this passage from one of the first treatises on tort law, Sir Frederick Pollock invoked
a compensatory rationale to explain the development of tort law. As Percy Winfield
subsequently explained, Pollock “consistently adopted the . . . view” that tort liability is
“based on the principle that . . . all injuries done to another person are torts, unless
there is some justification recognized by law.”2 This principle, according to Winfield,
enables the common law to recognize new torts. If, to use Winfield’s phrasing, “the law
of tort is based upon a general principle that all harm to another person is presump-
tively unlawful,” then one who causes injury to another commits a tort and is subject
to liability for compensatory damages, unless there is some justification for the denial

* I’m grateful to Ronald Dworkin for his helpful input on an earlier iteration of this project, and for the
helpful comments on this iteration that I received from John Goldberg, Henry Smith, and participants in the
Private Law Workshop at Harvard Law School. Financial support was provided by the Filomen D’Agostino
and Max E. Greenberg Research Fund of the New York University School of Law.
1
Frederick Pollock and James Avery Webb, A Treatise on the Law of Torts (St. Louis: F.H. Thomas Law
Book Co., 3d ed. 1894), vol. 1, 12–13. The term “injury” was employed by early legal scholars such as
Blackstone to refer to “a completed wrong that has been committed by one person against another.” John
C.P. Goldberg, “Two Conceptions of Tort Damages: Fair v. Full Compensation,” 55 DePaul L. Rev. 435 (2006),
437. By equating injury with hurt or harm, however, Pollock referred to the other meaning of injury
increasingly used by legal scholars and courts throughout the nineteenth century and commonly used today,
namely, “a loss or setback that a person has suffered.” Goldberg, “Two Conceptions of Tort Damages.”
2
Percy H. Winfield, The Province of the Law of Tort (New York: MacMillan, 1931), 32–3.
66 MARK A . GEISTFELD

of liability.3 When derived from a compensatory norm, a cause of action for compen-
satory damages can be viable even if the nominate tort had not previously been
recognized by courts. The evolutionary growth of tort law can be animated by a
norm of injury compensation.
By the middle of the twentieth century, scholars had reached a consensus that “tort
law ought primarily to be a means for compensating injured people” rather than “an
instrument for admonishing currently undesirable civil conduct.”4 Since then, scholars
have coupled the function of compensation with that of deterrence, yielding “the
baseline proposition . . . , repeated at the outset of countless law review articles pub-
lished in the last fifty years” that “the function of tort law is to compensate and deter.”5
Despite this history, scholars now roundly reject the proposition that tort law
implements a compensatory norm, relying on a reason that would seem to foreclose
further inquiry about the matter: “Measures of compensatory liability sometimes
exceed, sometimes fall short of, and sometimes bear no relation to what is required
to make the claimant whole.”6 Tort law relies on a default rule of negligence liability
that primarily values the duty to exercise reasonable care and accordingly limits the
availability of compensatory damages, fundamental attributes that appear to be wholly
inconsistent with a compensatory norm.
In contrast to the prevailing skepticism about the matter, in my view tort law
implements a norm of compensation. As I have argued at length elsewhere, a compen-
satory tort right that is justified by the value of individual autonomy or equal freedom
can persuasively explain the important tort doctrines governing physical harm, includ-
ing those that limit liability.7 Having concluded that tort law can be plausibly described
by a compensatory tort right and its correlative compensatory duty, I will now try to
show that compensation is a defensible norm of justice for answering “questions about
who is to get how much of what and why (i.e., on what grounds).”8
The argument proceeds in three parts. Part I argues that a compensatory duty can
be justified by the principle of liberal egalitarianism, illustrating the claim with the
conception of equality articulated by Ronald Dworkin. Part II then specifies the
substantive content of a compensatory tort right and explains why the correlative
compensatory duty is either largely or fully satisfied by the exercise of reasonable care,
eliminating any obligation to pay compensatory damages in a wide range of cases. This
compensatory norm justifiably limits liability in the manner suggested by Pollock’s
general conception of tort law. It also unifies the functions of compensation and

3
Winfield, The Province of the Law of Tort, at 36.
4
G. Edward White, Tort Law in America: An Intellectual History (New York: Oxford University Press,
1980), 146–7.
5
John C.P. Goldberg, “Twentieth-Century Tort Theory,” 91 Geo. L.J. 513 (2003), 525.
6
Emily Sherwin, “Compensation and Revenge,” 40 San Diego L. Rev. 1387 (2003), 1388.
7
See generally Mark A. Geistfeld, Tort Law: The Essentials (New York: Walters Kluwer, 2008) (hereinafter
Tort Law).
8
John C. Gardner, “What Is Tort Law For? Part I. The Place of Corrective Justice,” 30 Law & Phil. 1 (2011),
8 (defining norms of justice in these terms).
 COMPENSATION AS A TORT NORM 67

deterrence, yielding a coherent conception of the compensation-deterrence reasoning

that is now so prevalent within the practice of tort law.9 Part III completes the
argument by showing how such a compensatory tort right implements the principle
of corrective justice in a conceptually interesting and determinate manner.

II. Injury Compensation and Liberal

Egalitarianism
Liberal egalitarianism justifies distributive schemes that strive to give everyone the
same, just starting point so that each can pursue his or her own conception of the good
life. Different pursuits typically generate different levels of individual wealth or
welfare, and so “[t]he essence of this moral conception is equality of treatment rather
than impartial concern for well-being.”10 To use Ronald Dworkin’s terminology, the
distributive principle should be “endowment-insensitive” and “ambition-sensitive.”11
One’s position in life should reflect ambitions and choices rather than the arbitrary
circumstances of endowment beyond one’s control.
To identify the types of wealth redistributions that can be justified by liberal egalitar-
ianism, Dworkin constructs a hypothetical auction for determining the initial distribution
of resources that would satisfy the principle of equality, an outcome he calls “equality of
resources.” To conduct such an auction, the political system must have previously
specified various legal entitlements, including those constitutive of tort law. These entitle-
ments are grounded on the principle that individuals should incur the costs foreseeably
caused by their autonomous choices, thereby justifying a compensatory tort duty.

A. Equality of resources
As Dworkin stipulates, equality of resources is a general theory of distributional
equality that treats individuals “as equals when it distributes or transfers so that no
further transfer would leave their shares of the total resources more equal.”12 So
defined, the theory is “very abstract” because there are “different theories about what
would count as equality of resources.”13 Dworkin expends a considerable amount of
his own resources to develop a “suitable conception,” but the basic idea is that once
everyone has the same, just starting point, each can pursue his or her conception of the
good life.14 Wealth differences attributable to one’s ambition and autonomous choices

9
For more sustained argument on this point, see Mark A. Geistfeld, “The Coherence of Compensation-
Deterrence Theory in Tort Law,” 61 DePaul L. Rev. 383 (2012).
10
Thomas Nagel, Equality and Partiality (New York: Oxford University Press, 1991), 106.
11
Ronald Dworkin, “What is Equality? Part 2: Equality of Resources,” 10 Phil. & Pub. Aff. 283 (1981), 311.
12
Ronald Dworkin, Sovereign Virtue: The Theory and Practice of Equality (Cambridge, MA: Harvard
University Press, 2000), 12.
13 14
Dworkin, Sovereign Virtue, 12. Dworkin, Sovereign Virtue, at 65–119.
68 MARK A . GEISTFELD

are just, whereas differences stemming from one’s (unchosen) endowments, including
disease and disability, are unjust. Consequently, “equality of resources requires that
people pay the true cost of the lives they lead.”15
To determine what counts as a “cost” for distributive purposes, Dworkin constructs
a hypothetical auction in which participants have equal resources, defined as things
external to the individual bidder. Like any other competitive auction, Dworkin’s
hypothetical auction yields prices that reflect opportunity costs or “fix the value of any
transferable resource one person has as the value others forego by his having it.”16 The
resulting distribution would satisfy an “envy test” because each participant would prefer
his or her own bundle over one purchased by anyone else (otherwise the individual
would have purchased such an alternative bundle). The distribution is equal in this
fundamental respect, making opportunity costs (the auction prices) the normatively
appropriate measure for evaluating distributional equality.
The opportunity cost or price obtained from any auction depends on how the
underlying entitlement for the resource has been specified. Dworkin’s hypothetical
auction accordingly requires a “background or baseline liberty/constraint system” that
defines the particular liberties or entitlements associated with the resources to be
auctioned.17 This baseline must be justified by the same principle that justifies equality
of resources, namely, “in the more abstract egalitarian principle, which requires a
community to treat each of its members with equal concern.”18 The abstract egalitar-
ian principle measures equal shares in terms of opportunity costs, so the baseline itself
must be constructed by reference to “what we might call the true opportunity costs of a
set of resources.”19
Dworkin then develops this concept by reference to a principle of abstraction:
This principle recognizes that the true opportunity cost of any transferable resource is the
price others would pay for it in an auction whose resources were offered in as abstract a
form as possible, that is, in the form that permits the greatest flexibility in fine-tuning bids
to plans and preferences.20

When described at this level of generality, the scheme itself is highly abstract.
Nevertheless, it has specific implications for the substantive content of tort law.

B. Opportunity costs and tort compensation

The baseline of entitlements required by the hypothetical auction includes those
specified by the tort system. As Dworkin explains, “any competent baseline liberty/
constraint system would include a principle of security: this would mandate con-
straints on liberty necessary to provide people with enough physical security and
enough control over their own property to allow them to make and carry out plans and

15 16
Dworkin, Sovereign Virtue, at 76. Dworkin, Sovereign Virtue, at 149.
17 18
Dworkin, Sovereign Virtue, at 143. Dworkin, Sovereign Virtue, at 147.
19 20
Dworkin, Sovereign Virtue, at 149. Dworkin, Sovereign Virtue, at 151.
 COMPENSATION AS A TORT NORM 69

projects.”21 Tort rules governing accidental harms, therefore, must be formulated by

reference to the general consideration applicable to all aspects of the baseline—they
must capture the “true opportunity costs of a set of resources.”22
In Sovereign Virtue, Dworkin describes tort law as a system for constraining liberty
that “would correct for externality.”23 To ensure that individuals internalize the “true”
opportunity cost of their risky behavior, tort law could adopt a rule of strict liability.
One who engaged in risky behavior would be obligated to pay compensatory damages
to those who were foreseeably harmed by this autonomous choice, thereby correcting
for externality as required by Dworkin’s formulation of liberal egalitarianism.
This reasoning finds further expression in Dworkin’s earlier discussion of torts in
Law’s Empire, in which the appeal of strict liability is made evident by his conception
of just distribution:
The theory of private responsibility we are testing explains why relative cost figures in
these moral decisions. According to that theory we must act as if the concrete rights we
cannot both exercise had not yet been distributed between us, and we must distribute these
ourselves as best we can, in the way equality of resources commends . . . If compromise is
not possible in the circumstances . . . we must each act so as to minimize the inequality of the
distribution we achieve, and that means so that the loser pays less.24

This principle of comparative harm can be satisfied by a rule of strict liability.

A dutyholder subject to strict liability would choose to create a foreseeable risk of
harm whenever the net private benefit from engaging in the activity exceeds the
compensatory obligation that would be owed in the event of an accident causing
injury to the rightholder. Under these conditions, the dutyholder benefits from the
risky interaction, and an award of fully compensatory damages in the event of an
accident ensures that the interaction does not make the “loser” or injured rightholder
worse off. As compared to the outcome in which the rightholder receives no compen-
sation, strict liability minimizes the inequality of distribution between the interacting
parties by minimizing the loss suffered by the rightholder as “loser” in that interaction.
Alternatively, if the dutyholder rationally decides not to create the risk, then the net
private benefit that she would have derived from the risky activity must be less than the
compensatory obligation. Now the dutyholder is the “loser,” but her opportunity cost
(the lost benefit of the risky activity) is necessarily less than the opportunity cost that
would otherwise be created by the conduct in question (measured by the total
compensatory obligation owed to the rightholder). Once again, the rule of strict
liability minimizes the loss or opportunity cost that must necessarily be incurred by
at least one of the parties, thus satisfying the principle of comparative harm as
formulated by Dworkin.

21 22
Dworkin, Sovereign Virtue, at 148–9. Dworkin, Sovereign Virtue, at 149.
23
Dworkin, Sovereign Virtue, at 157.
24
Ronald Dworkin, Law’s Empire (Cambridge, MA: Harvard University Press, 1986), 303 (emphasis
added).
70 MARK A . GEISTFELD

The appeal of no-fault tort compensation is not limited to Dworkin’s formulation of

liberal egalitarianism. According to Will Kymlicka, liberal egalitarianism can be
generally characterized in terms of an abstract principle of the type developed by
Dworkin: “Treating people with equal concern requires that people pay for the costs of
their own choices.”25 This abstract principle provides a morally coherent role for a
compensatory tort obligation for reasons that are fully illustrated by Dworkin’s
conception of equality of resources.
To be sure, tort law does not ordinarily entitle accident victims to compensatory
damages. Compensation, however, is not wholly defined by the compensatory dam-
ages remedy. Within the context of a nonconsensual interaction or forced exchange, a
compensatory payment comprises the resources required to satisfy a compensatory
obligation. One’s compensatory obligation, in turn, is defined by the correlative
compensatory entitlement held by the other party. The attributes of a compensatory
tort right accordingly determine the compensatory properties of tort law, and so until
that right has been fully specified, it is an open question whether tort law can
implement a compensatory norm without granting an entitlement to compensatory
damages in all cases.

III. A Compensatory Tort Right and the

Correlative Compensatory Duty
Any evaluation of a compensatory tort norm must begin with a more complete
statement of a compensatory tort right and its implications for tort liability. Under
at least one formulation, a compensatory tort right can justify the default rule of
negligence liability. In a wide range of cases, the negligence rule distributes risk in a
manner that fully satisfies the demands of a compensatory rightholder, yielding
outcomes in which the dutyholder makes the full compensatory payment by exercising
reasonable care rather than by paying compensatory damages in the event of acciden-
tal harm. In the remaining cases, the compensatory right justifies a rule of strict
liability that must be supplemented by negligence liability to ensure that risk is
distributed in an adequately compensatory manner. The compensatory properties of
risk distribution can justify the default rule of negligence liability.

A. The substantive content of a compensatory tort right

According to the Restatement (Second) of Torts, an individual interest that “is
protected against any form of invasion . . . becomes the subject matter of a ‘right’.”26

25
Will Kymlicka, Contemporary Political Philosophy: An Introduction (New York: Oxford University Press,
1990), 75.
26
Restatement (Second) of Torts, } 1 (1965), cmt. b.
 COMPENSATION AS A TORT NORM 71

The specification of such a right necessarily prioritizes the protected interest of the
rightholder over the conflicting interest of the dutyholder, making it possible for
the tort rule to burden the subordinate interest of the dutyholder in order to protect
the prioritized interest of the rightholder. A rule that protects the individual interest in
physical security, for example, gives the security interest of the rightholder some sort
of legal priority over the conflicting or invading liberty interest of the dutyholder. To
do so, the tort rule must first distinguish these interests in a manner that justifies a
priority for the security interest. The nature of the priority then defines the substantive
content of the tort right and correlative duty. Rights-based tort rules, therefore, can be
characterized in terms of an underlying priority that gives one set of interests legal
protection over another set of conflicting or invading interests of the dutyholder.
A prioritized set of interests is valued more highly by tort law than the subordinate
set of conflicting interests. How the interests are normatively valued for this purpose
depends on the underlying principle of equality. Pursuant to liberal egalitarianism, tort
law can give different values to the individual interests in physical security and liberty
based on their relative importance for the equal exercise of the general liberal egali-
tarian right to autonomy or self-determination.
Because individuals must first be adequately secure in order to fully exercise
autonomy, tort law can prioritize the individual interest in physical security.27 The
exercise of liberty is also essential for living a meaningful life, so the requirement of
equal treatment prevents the rightholder’s security interest from having an absolute
priority that fully negates the value of the dutyholder’s conflicting liberty interest.
Under liberal egalitarianism, a priority of the security interest must account for the
value of liberty, explaining why “[m]ost of the rights of property, as well as of
person . . . are not absolute but relative.”28
Based on a relative priority of the security interest, tort rules can be formulated “to
give compensation, indemnity or restitution for harms”—the first purpose of liability
according to the Restatement (Second) of Torts.29 If a dutyholder’s exercise of liberty
foreseeably causes physical harm to a rightholder, a compensatory obligation burdens
the dutyholder’s subordinate liberty interest to compensate harms it caused to the
prioritized security interest of the rightholder; neither legal fault nor an unreasonable
liberty interest is required to justify the compensatory obligation. This duty permits
individuals to engage in risky behavior by relying on compensation to protect the
rightholder’s security interest, the type of outcome required by a right to security that
is relative to a right of liberty.
To be justifiable, a compensatory norm must address any normative problems
created by the rightholder’s lack of consent and the poor manner in which

27
See Richard Wright, Justice and Reasonable Care in Negligence Law, 47 Am. J. of Jurisprudence 143
(2002), 170–94 (explaining why leading justice theorists reject the utilitarian approach of weighing all interests
equally and instead maintain that rights-based tort rules prioritize the individual interest in physical security
over the conflicting liberty and economic interests of others). See also text accompanying note 21.
28
Losee v Buchanan, 51 N.Y. 476, 485 (1873). 29
Restatement (Second) of Torts, } 901(a) (1965).
72 MARK A . GEISTFELD

compensatory damages might otherwise protect the rightholder’s autonomy. Most

obviously, a tort duty limited to the payment of monetary compensation for a
nonconsensual harm can be deeply corrosive of the rightholder’s autonomy (consider
rape). To ensure that a dutyholder avoids behavior that disvalues the rightholder’s
autonomy, a compensatory tort norm can prohibit behavior of this type, justifying
extracompensatory damages that punish the dutyholder for having engaged in such
reprehensible behavior.30 A compensatory tort norm can define the types of behavior
for which a compensatory obligation adequately protects the rightholder’s autonomy.
In most cases, however, risky behavior entails no disrespect for the autonomy of
others; the risk is an unwanted byproduct of the activity. To establish liability in these
cases, a compensatory norm does not require culpability or personal fault. For cases of
accidental harm in which the interacting parties are blameless, “it is a fait accompli
that some innocent party will be burdened . . . Therefore, it cannot be a moral require-
ment that no party lose out as a consequence of his own blameless conduct. All that
remains open for decision is how the loss is to be apportioned.”31 By prioritizing
the rightholder’s interest in physical security, the compensatory norm places the loss
on the dutyholder as risky actor. The dutyholder’s exercise of liberty establishes the
requisite form of responsibility for the foreseeable outcomes of the autonomous
choice.32 The occurrence of foreseeable injury, not any moral shortcoming in the
behavior itself, can then trigger the obligation to pay compensatory damages.
This form of outcome responsibility is embodied in the common law maxim sic
utere tuo ut alienum non lædas, which for present purposes loosely translates into the
principle to use your own so as not to injure another.33 The maxim locates the
compensatory duty in the injury-causing conduct rather than the unreasonableness
of the injurer’s behavior, and so it has frequently been invoked by courts and
commentators to justify rules of strict liability.34
Such a compensatory norm can be used not only to justify rules of strict liability, but
also to explain why the tort system relies on a default rule of negligence liability to
govern cases of accidental physical harm. The reason involves the manner in which the
compensatory properties of a tort rule depend on how it distributes risk.

30
See generally Mark A. Geistfeld, “Punitive Damages, Retribution, and Due Process,” 81 S. Cal. L. Rev. 263
(2008) (discussing the role of punitive damages within a compensatory tort system and showing that this role
persuasively explains the relevant tort rules).
31
Loren E. Lomasky, “Compensation and the Bounds of Rights,” 33 Nomos 13 (John W. Chapman (ed.),
1991), 34 (discussing cases of necessity).
32
For more extended discussion of this conception of individual responsibility, see Stephen R. Perry,
“Responsibility for Outcomes, Risk, and the Law of Torts,” in Gerald Postema (ed.), Philosophy and the Law of
Torts (New York: Cambridge University Press, 2001), 92–3.
33
The maxim means “[u]se your own property in such a manner as not to injure that of another.” Black’s
Law Dictionary (St. Paul, MN: West Publishing Co, 5th ed. 1979), 1238. As applied to risky behavior not
involving the use of property, the maxim yields a principle that “under the common law a man acts at his
peril.” Oliver Wendell Holmes, The Common Law (Boston, MA: Little, Brown & Co., 1881), 82 (stating that
“some of the greatest common law authorities” held this view).
34
See, e.g., Perkins v F.I.E. Corp., 762 F.2d 1250, 1254–56 (5th Cir. 1985) (noting that the sic utere maxim is
the basis for the rule of strict liability governing ultrahazardous activities under Louisiana law).
 COMPENSATION AS A TORT NORM 73

B. Compensation as risk distribution

In a compensatory tort system, the appropriate formulation of liability rules critically
depends on context. Different types of risky interactions create different types of
compensatory problems. The different compensatory problems have different solu-
tions, most of which do not include an entitlement to compensatory damages in all
cases. Tort rules can instead distribute risk in a manner that fully satisfies the demands
of a compensatory rightholder.
A compensatory tort right prioritizes the rightholder’s interest in physical security
over conflicting liberty interests of the dutyholder for reasons of autonomy. If that
priority applies to an interaction between the two parties, it justifies the rightholder’s
entitlement to compensatory damages in the event of injury. Such an interpersonal
conflict of interests, however, does not exist in two important classes of nonconsensual
harms. For cases in which the rightholder and dutyholder are engaged in reciprocally
risky interactions or are otherwise in a direct or indirect contractual relationship, the
tort rule governs an intrapersonal conflict of the rightholder’s security and liberty
interests. In these cases, the rightholder does not prioritize the security interest and
instead has compensatory demands that are fully satisfied by a negligence rule
requiring the dutyholder to exercise the cost-minimizing amount of reasonable care.
First, consider tort rules governing reciprocal risks. For example, as two automo-
biles go past one another on the road, each driver simultaneously imposes a risk of
physical harm on the other. For perfectly reciprocal risks, the interacting individuals
are identical in all relevant respects, including the degree of risk that each imposes on
the other, the severity of injury threatened by the risk, and the liberty interests
advanced by the risky behavior. Very few risky interactions will actually satisfy these
conditions. But due to the requirement of equal treatment, tort law evaluates risky
behavior under an objective standard that, in this instance, asks whether the activity is
common in the community.35 Automobile driving is such an activity, so as an
objective matter, tort rules governing automobile accidents apply to reciprocally
situated parties, even for cases in which the victim was walking or riding a bicycle.
Reciprocity eliminates any relevant differences between the interacting parties. For
example, each automobile driver has the identical right against the other, each owes an
identical duty to the other, and each expects to derive a benefit, on balance, by
participating in the activity of driving.
In these circumstances, neither party prioritizes the security interest over the liberty
interest. Each interacting individual instead prefers a cost-minimizing duty of reasonable

35
Compare Geistfeld, Tort Law (note 7) at 93–5 (explaining why the autonomous choices made by a
rightholder, such as the decision not to drive automobiles, would violate the principle of equal treatment if
these choices were to determine unilaterally whether the dutyholder would be subject to negligence or strict
liability, thereby justifying a rule that evaluates reciprocity in the objective terms of whether the activity is
common in the community); with Restatement (Third) of Torts: Liability for Physical and Emotional Harms, }
20 (2010), cmt. j (“Whenever an activity is engaged in by a large fraction of the community, the absence of
strict liability can be explained by considerations of reciprocity.”).
74 MARK A . GEISTFELD

care that requires a safety precaution only if the benefit of risk reduction (fully
accruing to the individual as reciprocal rightholder) exceeds the burden or cost of
the precaution (also fully borne by the individual as reciprocal dutyholder).36 By
minimizing accident costs, the negligence rule maximizes the net benefit that each
driver expects to gain by participating in the activity.
A tort rule that rejected each individual’s preference for a cost-minimizing negli-
gence rule by instead prioritizing the security interest under a rule of strict liability
would be unreasonable or contrary to the autonomy interests of both parties to the
risky interaction. For this class of cases, the reasonable demands of the compensatory
rightholder—those conforming to the underlying value of equal autonomy—are
fully satisfied by a negligence rule that requires the dutyholder to exercise the cost-
minimizing amount of care.
In these cases, the dutyholder fully satisfies the compensatory obligation by exercis-
ing the amount of reasonable care required by the compensatory tort right. Doing so
does not necessarily eliminate risk, creating the possibility that the interaction might
accidentally injure the rightholder. In that event, however, the compensatory tort right
does not entitle the victim to an award of compensatory damages—the dutyholder’s
exercise of reasonable care has already fully satisfied the rightholder’s compensatory
demands. A compensatory tort obligation does not entail the payment of compensa-
tory damages in all cases of accidental harm.
The same outcome occurs for cases in which the rightholder and dutyholder
are in a direct or indirect contractual relationship, as in product cases involving
consumers and manufacturers.37 The consumer rightholder purchases the product
on the expectation that doing so, on balance, will be advantageous. By selling the
product, the manufacturer creates a risk of physical injury to which the consumer is
exposed. A tort rule that makes the manufacturer liable for these injuries will affect
product costs, price, aggregate demand, and net profits. The distributive impact of tort
liability, however, must be defined in relation to the normatively justified tort rule.38

36
For more rigorous demonstration, see Mark Geistfeld, “Placing a Price on Pain and Suffering: A Method
for Helping Juries Determine Tort Damages for Nonmonetary Injuries,” 83 Cal. L. Rev. 773 (1995), 851–2.
37
Unlike the consumer–manufacturer relationship discussed in text, in other types of contractual relation-
ships, the rightholder sells something to the dutyholder. The most important example is the employment
relationship (the sale of labor), in which the employee must be compensated for facing work-related risks
either by an increase of wages or receipt of compensation for work-related injuries. The employer minimizes
this total compensatory obligation by adopting cost-effective safety measures and compensating employees for
the residual risks. Employees currently receive both forms of compensation, albeit outside of the tort system
(workplace injuries are governed by workers’ compensation schemes that provide guaranteed compensation
for work-related injuries). Workplace injuries accordingly provide further support for the conclusion that the
law regulates accidental harms in a compensatory manner, with the different compensatory legal rule in these
cases (one of strict liability) stemming from the different form of contractual relationship (the rightholder as
seller rather than buyer).
38
To identify the distributive effects of liability, one must first specify the appropriate baseline for analysis.
This baseline cannot be derived by economic analysis because cost-benefit analysis depends on prices, which
ultimately depend on the initial allocation of legal entitlements or property rights. See Lewis A. Kornhauser,
“Wealth Maximization,” in Peter Newman (ed.), The New Palgrave Dictionary of Economics and the Law
 COMPENSATION AS A TORT NORM 75

At this baseline, the consumer pays for the full cost of tort liability, as the equilibrium
product price must cover all of the seller’s costs, including its liability costs. Consumer
interests are the only ones that factor into the distributive analysis required by the
normatively justified tort rule, explaining why products liability law recognizes that “it
is not a factor . . . that the imposition of liability would have a negative effect on
corporate earnings or would reduce employment in a given industry.”39 For risks
not threatening injury to bystanders, product cases only implicate an intrapersonal
conflict of consumer interests: those involving physical security, liberty (regarding
product use), and money (product price and other financial costs of product use).40
In comparing her own security and liberty interests, the consumer gives no special
priority to either one. The consumer prefers to pay for product safety only if the
benefit of risk reduction (accruing to the consumer) exceeds the cost of the safety
investment (also borne by the consumer via the associated price increase or decrease of
product functionality). Consumers reasonably expect product-safety decisions to be
governed by a cost-benefit calculus because that decisional rule maximizes consumer
welfare. A product that does not satisfy reasonable consumer expectations is defective
and subjects the seller to liability under the widely adopted rule of strict products
liability.41 This rule does not entitle consumers to compensatory damages in all cases.
Due to the relatively high cost of tort compensation as compared to other forms of
insurance, consumers do not reasonably expect to receive tort compensation for
injuries caused by nondefective products.42 The reasonable compensatory demands
of consumer rightholders are fully satisfied by cost-minimizing tort rules that limit
liability to the physical harms caused by defective products.43
As in cases of objective reciprocity, the dutyholder in product cases fully satisfies
the compensatory obligation by making the cost-minimizing investments in safety
required by the compensatory tort right. Doing so does not necessarily eliminate risk,
but the dutyholder (having fully satisfied the compensatory tort right) is not obligated
to pay compensatory damages for injuries caused by the residual (or reasonable) risks
inherent in most nondefective products. The demands of the compensatory right-
holder, once again, are fully satisfied by a negligence rule that does not require the
dutyholder to pay compensatory damages in all cases.

(New York: Stockton Press, 1998), vol. 3, 679. The initial entitlements define the normatively justified tort rule,
making it the baseline for distributive analysis.
39
Restatement (Third) of Torts: Products Liability, } 2 (1998), cmt. f.
40
For risks threatening injury to bystanders, the analysis involves the interpersonal mediation of security
and liberty interest characteristic of more general forms of tort liability. See Mark A. Geistfeld, Principles of
Products Liability (New York: Foundation Press, 2d ed. 2011), 309–20.
41
See generally Geistfeld, Principles of Products Liability.
42
See Geistfeld, Principles of Products Liability at 61–7.
43
See Geistfeld, Principles of Products Liability at 256–66 (explaining why consumers do not reasonably
expect to receive compensatory damages for pure economic loss and standalone emotional harms caused by
defective products).
76 MARK A . GEISTFELD

C. Risk distribution as non-ideal compensation

In a wide range of cases, the negligence rule can attain the ideal compensatory
outcome by distributing risk to maximize the net benefit that a rightholder expects
to derive from the risky interaction, so the rightholder is not made worse off, ex ante,
than she would otherwise be in a world without the risk (and the associated benefit to
be gained from the risky activity). The only remaining cases involve rightholders who
are not in a contractual relationship with a dutyholder who creates an objectively
defined nonreciprocal risk of physical harm. In these cases, the negligence rule can still
distribute risk in the manner reasonably required by the compensatory tort right, but
the compensation is not ideal, even when supplemented by a rule of strict liability.
These cases involve activities that are not common in the community and create
risks above the ordinary level of background risk. A paradigmatic example involves the
use of dynamite for construction purposes, although objectively defined nonreciprocal
risks are also created in myriad other ways, including instances in which the duty-
holder’s lack of intelligence or skill creates dangers above the background level
(defined by ordinary intelligence and skill).
For this class of cases, the tort rule must mediate an interpersonal conflict between
the dutyholder’s interest in liberty and the rightholder’s interest in physical security.
A compensatory tort rule resolves these conflicts by prioritizing the rightholder’s
security interest, justifying a right to compensatory damages for these injuries—the
same outcome that is attained by the rule of strict liability for abnormally dangerous
activities and the pockets of strict liability within the objectively defined negligence
standard of reasonable care.44
The compensation afforded by these forms of strict liability, however, does not fully
satisfy the compensatory obligation. In the event of a fatal accident, the dutyholder is
not obligated to pay for the decedent’s loss of life’s pleasures because the damages
award cannot compensate a dead person, a problem that substantially reduces and
potentially eliminates the compensatory damage award for wrongful death.45 The most
severe type of physical harm cannot be fully redressed by a rule of strict liability.
To solve the compensatory problem inherent in a rule of strict liability, the right-
holder reasonably prefers to supplement this rule with a behavioral obligation of
reasonable care that directly protects against the threat of premature death. Such a
safety obligation must be derived from the compensatory duty, which can be defined
by the total burden that a dutyholder would incur under ideal conditions in which the
rightholder is always fully compensated. Because the dutyholder does not bear this

44
See Geistfeld, Tort Law (note 7) at 92–7.
45
See Romo v Ford Motor Co., 6 Cal. Rptr. 3d 793, 811 (Ct. App. 2003) (ruling on a punitive damages award
in a wrongful death case involving an award of zero compensatory damages); Edward A. Adams, Venue
Crucial to Tort Awards: Study: City Verdicts Depend on Counties, N.Y.L.J., Apr. 4, 1994, at 1, 5 (reporting
results of empirical study finding, among other things, that the average tort award in New York City between
1984 and 1993 was three times higher for brain damage than for wrongful death, which was only twice as much
as the average damage award for a herniated disc).
 COMPENSATION AS A TORT NORM 77

entire compensatory burden under a rule of strict liability, tort law can eliminate the
compensatory shortfall by shifting that component of the compensatory obligation
from the compensatory damages remedy into the duty to exercise reasonable care.46
Such a negligence rule requires the dutyholder to satisfy the compensatory obliga-
tion, in part, by incurring these expenses through the exercise of reasonable care.
These safety expenditures, when added to the cost-minimizing precautions that the
dutyholder would otherwise take under ideal compensatory conditions, further reduce
risk or the likelihood that the rightholder will suffer injury. The supplemental rule of
strict liability then fulfills the compensatory obligation with respect to the remaining,
residual risks that are not eliminated by the exercise of reasonable care. These
abnormally dangerous or nonreciprocal risks are subject to strict liability, but the
default rule of negligence liability continues to distribute these risks in the manner
reasonably required by the compensatory tort right.47
Nonetheless, the risk distribution in these cases is not ideal for the rightholder,
unlike the distribution that occurs in cases involving reciprocal risks or contractual
relationships. As we have found, risk distribution can be fully compensatory for
rightholders who (1) incur the burdens of the compensatory duty (as reciprocally
situated dutyholder or consumer) and (2) participate in the risky activity (such as by
driving or using a product) engaged in by the dutyholder (another driver or a product
manufacturer). For nonreciprocal risky interactions that occur outside of contractual
relationships, neither condition applies. The rightholder does not bear the full burden
of the compensatory duty or otherwise derive a sufficient benefit from the risky activity
engaged in by the dutyholder, so it is not possible for tort law to distribute risk in a
manner that would fully compensate the rightholder (consider premature death).
This compensatory problem, however, does not justify a ban of the risky behavior.
The compensatory right is based on a relative priority of the security interest, not an
absolute priority that negates, or gives no value to, conflicting liberty interests.48 By
exercising reasonable care and paying compensatory damages for the harms foresee-
ably caused by the residual nonreciprocal risks, the dutyholder fully satisfies the
compensatory obligation. This exercise of liberty has normative value that is not
negated simply because social conditions make it infeasible to attain the ideal com-
pensatory outcome. The reasonable compensatory demands of the rightholder—those
that give equal concern to the autonomy of the dutyholder—do not justify a ban of the
dutyholder’s exercise of liberty. These interactions can leave the rightholder worse off
than she would otherwise be, but tort law still distributes risk in the manner that fully
satisfies the reasonable demands of the compensatory rightholder.

46
For more rigorous discussion of the argument in this paragraph, see Mark A. Geistfeld, “Reconciling
Cost-Benefit Analysis with the Principle that Safety Matters More than Money,” 76 N.Y.U. L. Rev. 114 (2001).
47
This reasoning explains why a strictly liable dutyholder who reprehensibly rejects the duty to exercise
reasonable care is subject to punitive damages. Cf. Owens-Ill., Inc. v Zenobia, 325 Md. 420, 601 A.2d 633, 653
(1992) (adopting majority rule requiring proof of “actual malice” to justify punitive damages under strict
products liability).
48
See Part II.A.
78 MARK A . GEISTFELD

D. Breaches of the compensatory duty

Breach of the primary duty to exercise reasonable care creates a second-order duty to
pay compensatory damages for the physical harms proximately caused by the breach.
Though intrinsically related, these two duties are not substantively equivalent. Due to
the inherent limitations of the compensatory damages remedy, the second-order duty
to pay compensatory damages does not fully substitute for the first-order duty to
exercise reasonable care.
The most severe physical harm governed by tort law is wrongful death, and yet
monetary damages cannot compensate a dead rightholder for the premature loss of
life. Compensatory damages also do not make the plaintiff rightholder “whole” in
cases of bodily harm, nor does this remedy strive to do so.49 Premature death and
bodily injury are paradigmatic examples of an irreparable injury, although this com-
mon law category also encompasses damage to real or tangible property.50 The entire
category of physical harms—bodily injury or damage to real or tangible property—is
comprised of irreparable injuries that ordinarily cannot be fully compensated by the
damages remedy.
For irreparable injuries, breaches of the primary compensatory obligation to exer-
cise reasonable care will usually not be remedied in a fully compensatory manner.
The exercise of reasonable care is the only way for a dutyholder to fully satisfy the
reasonable compensatory demands of the rightholder. The superior compensatory
attributes of risk distribution accordingly explain why negligence is a “behavioral” rule
defined by a first-order safety obligation that is substantively more important than the
second-order remedial duty to pay compensatory damages for breach.51
To protect the integrity of the first-order duty, tort law prohibits the dutyholder
from consciously rejecting or recklessly ignoring the primary duty to exercise reason-
able care. A defendant who engaged in this prohibited conduct and breached the
first-order duty is subject to punitive damages, regardless of whether the defendant
was always willing and able to pay compensatory damages.52 The extracompensatory
award of punitive damages is required to vindicate the compensatory tort right due to
the inherent inadequacy of the compensatory damages remedy.

49
See Restatement (Second) of Torts, } 903 (1965), cmt. a (stating that a damage award for the loss of life’s
pleasures is not supposed to “restore the injured person to his previous position” but should instead only “give
to the injured person some pecuniary return for what he has suffered or is likely to suffer”).
50
Mark A. Geistfeld, “The Principle of Misalignment: Duty, Damages, and the Nature of Tort Liability,”
121 Yale L.J. 142 (2011), 164.
51
See Mark A. Geistfeld, “Tort Law and the Inherent Limitations of Monetary Exchange: Property Rules,
Liability Rules, and the Negligence Rule,” 4 J. Tort Law, No. 1, Art. 4 (2011), at <http://www.bepress.com/jtl/
vol4/iss1/art4> accessed October 21, 2013.
52
Geistfeld, “The Principle of Misalignment” (note 50) at 165–9 (identifying the types of behavior
prohibited by the negligence rule and providing citations to cases holding that a defendant who engaged in
such behavior cannot avoid liability for punitive damages even if fully willing and able to pay compensatory
damages).
 COMPENSATION AS A TORT NORM 79

But even in these cases, a dutyholder breaches the primary duty only if the unrea-
sonable conduct proximately causes the rightholder to suffer compensable harm. The
failure to exercise reasonable care, no matter how reprehensible, creates no further
compensatory obligation in the absence of injury. There is simply nothing left to
compensate. In cases of injury, by contrast, a breach of the first-order compensatory
duty to exercise reasonable care creates a compensatory shortfall that triggers the
second-order duty to pay compensatory damages. Tort liability is based on the
occurrence of injury for obvious compensatory reasons in accord with “ordinary
moral evaluation” that careless behavior causing injury is “deemed worse” than
careless behavior that does not ripen into harm.53
By focusing on the consequences of breach in cases of irreparable injury, it becomes
apparent why a compensatory negligence rule is primarily concerned about the
prevention of injury through the exercise of reasonable care. According to a leading
nineteenth-century treatise, in cases of irreparable injury “judges have been brought to
see and to acknowledge . . . that a remedy which prevents a threatened wrong is in its
essential nature better than a remedy which permits the wrong to be done, and then
attempts to pay for it.”54 In seeking to prevent irreparable injuries, the common law
has also long recognized the principle that the tort obligation cannot impose undue
hardship on the dutyholder.55 When derived from a compensatory duty, a primary
obligation to reduce the risk of irreparable harm through the exercise of reasonable
care does not impose undue hardship on the dutyholder. Compliance with this duty
distributes risk in the manner reasonably demanded by the holder of the compensa-
tory tort right, making it possible for tort law to compensate rightholders for physical
harms that cannot be fully repaired by the damages remedy.

IV. Compensation and Corrective Justice

To be a form of corrective justice, tort liability must repair the inequality created by a
dutyholder’s violation of a correlative tort right. Not only does the compensatory tort
right satisfy the requirements of corrective justice, such a tort right is arguably
required in order to make corrective justice a conceptually interesting and adequately
determinate interpretation of tort law.

53
John C.P. Goldberg and Benjamin Zipursky, “Torts as Wrongs,” 88 Tex. L. Rev. 917 (2010), 942 (arguing
in favor of interpretations of tort law that can incorporate this “framework of moral thought that people deploy
regularly in their daily lives”).
54
John Norton Pomeroy, A Treatise on Equity Jurisprudence as Administered in the United States of
America (San Francisco, CA: A.L. Bancroft & Co., 1883), 389; see also Douglas Laycock, “The Death of the
Irreparable Injury Rule,” 103 Harv. L. Rev. 687 (1990), 699 (“Judges act on these premises, whether or not they
consciously acknowledge all that Pomeroy imputed to them.”).
55
Cf. Laycock, “The Death of the Irreparable Injury Rule” (note 54) at 732–9 (discussing the rule that
monetary damages provide the remedy for harms that would otherwise be irreparable when equitable relief
would interfere with countervailing rights or impose undue hardship on the dutyholder).
80 MARK A . GEISTFELD

A. Compensation as a form of corrective justice

The importance of corrective justice within tort law has been extensively analyzed by
Jules Coleman, who explains the concept in these terms:
Corrective justice claims that when someone has wronged another to whom he owes a
duty of care, he thereby incurs a duty of repair. This means that corrective justice is an
account of the second-order duty of repair. Someone does not incur a second-order duty of
repair unless he has failed to discharge some first-order duty. However, the relevant first-
order duties are not themselves duties of corrective justice. Thus, while corrective justice
presupposes some account of what the relevant first-order duties are, it does not pretend to
provide an account of them.56

By exclusively addressing the second-order duty of repair, this formulation of cor-

rective justice lacks explanatory power. When the duty of repair is triggered only by the
plaintiff ’s exercise of a power that subjects the defendant to liability, the judgment
itself creates the compensatory obligation.57 Any judgment requiring the payment of
compensatory damages for a rights violation in a suit between private litigants is a form
of corrective justice. This formulation cannot identify anything distinctive about the
practice of corrective justice within tort law, nor can it fully specify the behavioral
requirements of the first-order tort duty to exercise reasonable care.
Due to this lacuna, Barbara Fried has concluded that “cost/benefit analysis is
currently the only game in town for determining appropriate standards of conduct
for socially useful acts that pose some risk of harm to others (a category that describes
almost all noncriminal conduct).”58 The apparent indeterminancy of the first-order
behavioral obligation under corrective justice also supports Jody Kraus’s conclusion that
“economic theories appear to have the edge on deontic theories because their explanations
of judicial decisions systematically yield more determinate results, at least in principle.”59
These problems exist whenever the domain of corrective justice is limited to the
second-order remedial duty that is generated by a judgment in the lawsuit. When
compensation is a first-order duty, however, the corrective-justice interpretation of
tort law is adequately determinate and conceptually interesting.60
A first-order compensatory duty and correlative right are abstract and become
concrete only in the context of a particular interaction between a dutyholder and

56
Jules L. Coleman, The Practice of Principle (New York: Oxford University Press, 2001), 32.
57
Benjamin Zipursky, “Civil Recourse, Not Corrective Justice,” 91 Geo. L.J. 695 (2003).
58
Barbara H. Fried, “The Limits of a Nonconsequentialist Approach to Torts,” 18 Legal Theory 231
(2012), 231.
59
Jody S. Kraus, “Transparency and Determinacy in Common Law Adjudication: A Philosophical Defense
of Explanatory Economic Analysis,” 93 Va. L. Rev. 287 (2007), 304.
60
Cf. Fried, “The Limits of a Nonconsequentialist Approach to Torts” (note 58) at 244, 250 (recognizing
that contradictions or paradoxes inherent in deontological accounts of tort law do not exist for a compensatory
account); Zipursky, “Civil Recourse, Not Corrective Justice” (note 57) at 710–12 (arguing that corrective justice
provides a conceptually uninteresting description of tort law if “the recognition of a right of action in tort” is
not “isomorphic with the recognition of a duty of repair”).
 COMPENSATION AS A TORT NORM 81

rightholder. The concrete form of the compensatory obligation ordinarily reduces to

the duty to exercise reasonable care.61 The associated requirements of reasonable
care—the conduct required of the dutyholder—can be specified with the same amount
of determinacy, in principle, as that attained by economic formulations of the duty.62
In addition to being adequately determinate, a compensatory tort duty makes the
principle of corrective justice a conceptually interesting rationale for tort law. The
compensatory duty is primary and not merely a secondary remedial obligation,
thereby explaining why corrective justice has distinctive importance for tort law.
Consider the reasons why compensatory tort rules satisfy the requirements of
corrective justice. According to John Gardner, “[s]ome transactions need not be
wrongful in order to call for correction. They are wrongful only if they go uncor-
rected.”63 A transaction of this type is embodied in the first-order compensatory duty
to exercise reasonable care, which deems risky behavior to be wrongful if the duty-
holder failed to correct for the risky behavior by not satisfying the compensatory
obligation through the exercise of reasonable care. This type of wrongdoing is also
entailed by the supplemental rules of strict liability for objectively nonreciprocal risks,
which involve behavior that is reasonable (for satisfying the first-order compensatory
duty of care) and not inherently wrongful (such as blasting for construction purposes).
These forms of behavior only become wrongful if the dutyholder failed to satisfy the
remaining component of the compensatory obligation by not correcting for the
occurrence of injury through the payment of compensatory damages. By satisfying
both the first-order compensatory duty to exercise reasonable care and any supple-
mental rules of strict liability, a dutyholder fully satisfies the rightholder’s compensa-
tory entitlement and the corresponding demands of corrective justice.
So, too, a breach of the first-order compensatory duty to exercise reasonable care
constitutes a corrective injustice. The breach of a first-order duty is wrongful as a
matter of corrective justice, according to Gardner, if “[t]he reasons not to do whatever
one did, the thing that now calls for correction, suffice to make that action wrongful
even if it is corrected.”64 This requirement is satisfied by a breach of the first-order
compensatory duty to exercise reasonable care that proximately caused the rightholder
to suffer physical harm—an irreparable injury that cannot be fully repaired by the
compensatory damages remedy.65 In these cases, the dutyholder’s payment of com-
pensatory damages “still leave[s] too great a rational remainder behind, too much in
the way of unsatisfied or imperfectly satisfied reasons, for the wrongdoing to have been
averted by the act of correction [via the payment of compensatory damages] alone.”66
The inherent inadequacy of the compensatory damages remedy fully explains why a

61 62
See Part II.B–C. See Geistfeld, Tort Law (note 7) at 191–204.
63
See Gardner, “What Is Tort Law For? Part I. The Place of Corrective Justice” (note 8) at 34.
64
Gardner, “What Is Tort Law For? Part I. The Place of Corrective Justice” (note 8) at 34.
65
See Part II.D.
66
Gardner, “What Is Tort Law For? Part I. The Place of Corrective Justice” (note 8) at 34.
82 MARK A . GEISTFELD

breach of the first-order compensatory duty to exercise reasonable care is a wrong that
can be redressed by corrective justice.
Having breached the first-order duty, a defendant incurs the second-order remedial
obligation to pay compensatory damages for the wrongful injuries proximately caused
by the breach. This form of corrective justice, however, is substantively different from
other remedial forms of corrective justice. A defendant’s payment of tort damages does
not simply redress a prior wrong; it redresses a prior corrective injustice (the failure to
satisfy the primary compensatory obligation through the exercise of reasonable care).
Under Gardner’s continuity thesis, “[i]f all else is equal, the reasons that were
capable of justifying a primary obligation are also capable of justifying a secondary
one.”67 A primary or first-order compensatory duty to exercise reasonable care
provides the strongest possible justification for a secondary remedial obligation to pay
compensatory damages for breaches of the primary duty—the payment of compensatory
damages supplies the “next best conformity” with the first-order compensatory duty.
This rationale then extends to the payment of compensatory damages under the
supplemental rules of strict liability. Negligence and strict liability are constitutive
elements of a single compensatory tort entitlement.68 For cases in which the exercise of
reasonable care does not fully satisfy the compensatory obligation, the injury-causing
conduct is wrongful unless accompanied by the payment of compensatory damages
under the rule of strict liability. This payment, however, does not fully substitute for
the primary obligation to exercise reasonable care. Both obligations must be satisfied
to prevent the conduct from being wrongful, providing the requisite continuity
between the duty of care and the compensatory damages award afforded by the rule
of strict liability.
Punitive damages can also be justified by the compensatory norm, making this
liability a form of corrective justice for reasons missed by others.69 By unifying the
first-order behavioral duty with the full set of second-order remedial duties, the com-
pensatory tort norm yields a unified body of tort rules that fully instantiate the principle
of corrective justice.

B. Can a compensatory tort right be just?

Although a compensatory tort right justifies the default rule of negligence liability, it
also justifies supplemental rules of strict liability for activities that are not common

67
Gardner, “What Is Tort Law For? Part I. The Place of Corrective Justice” (note 8) at 333.
68
If the compensatory entitlement cannot be adequately protected by a rule of strict liability, the resultant
compensatory shortfall can be eliminated by redirecting that compensatory obligation into the standard of
reasonable care, illustrating the inherent relation between the remedy of strict liability and the underlying
obligation to exercise reasonable care. See Part II. C. For more rigorous argument showing that negligence and
strict liability can be constitutive elements of a single entitlement, see Geistfeld, “Tort Law and the Inherent
Limitations of Monetary Exchange: Property Rules, Liability Rules, and the Negligence Rule” (note 51).
69
Compare notes 30 and 52 (explaining how punitive damages protect the integrity of a compensatory
right) with Ernest J. Weinrib, “Civil Recourse and Corrective Justice,” 39 Fl. St. L. Rev. 273 (2011), 290
(“Punitive damages are . . . inconsistent with corrective justice for reasons both of structure and content.”).
 COMPENSATION AS A TORT NORM 83

in the community and create risks above the ordinary level of background risk.70 This
attribute of a compensatory right would seem to be problematic, however, for “[s]trict
liability is widely thought to be unjust because there is liability without fault.”71 Unless
it would be just to impose a compensatory obligation not limited by fault, a compen-
satory tort system cannot implement the principle of corrective justice.
The injustice created by a rule of strict liability has been fully identified by Ernest
Weinrib:
The inequality in strict liability emerges from the principle that the defendant is to be liable
for any penetration of the plaintiff ’s space. What is decisive for the parties’ relationship is
the demarcation of the domain within which the law grants the plaintiff immunity from the
effects of the actions of others; the activity of the defendant is then restricted to whatever falls
outside this sphere. Thus the interests of the plaintiff unilaterally determine the contours of
what is supposed to be a bilateral relationship of equals.72

Although strict liability would be unjust if it were to grant the plaintiff rightholder
an “immunity” from being accidentally harmed by a defendant dutyholder, tort law
does not formulate rules of strict liability in this way. For example, the rule of strict
liability for the abnormally dangerous activity of blasting recognizes that the activity is
reasonable and does not prohibit the conduct with an “immunity” from injury granted
to the rightholder; it only requires the blaster as dutyholder to pay compensatory
damages to an injured rightholder.73 The compensatory duty, therefore, is not unjust
for giving the defendant “a duty to do something that is beyond him,” which is the
criticism most commonly leveled against rules of strict liability.74 The protected
interests of the plaintiff also do not “unilaterally determine the contours of what is
supposed to be a bilateral relationship of equals” as Weinrib claimed. Strictly liable
actors are free to impose these nonconsensual, reasonable risks on others, subject only
to the duty that they compensate the ensuing foreseeable harms. Properly understood,
strict liability is “liability rule” that does not impose any behavioral obligations on the
dutyholder beyond the requirement to pay compensatory tort damages.75

70
See Part II.C.
71
Peter Jaffey, “Duties and Liabilities in Private Law,” 12 Legal Theory 137 (2006), 153.
72
Ernest J. Weinrib, The Idea of Private Law (Cambridge, MA: Harvard University Press, 1995), 177 (italics
added).
73
E.g., Spano v Perini Corp., 250 N.E.2d 31, 34 (N.Y. 1969) (explaining that the plaintiff ’s claim of strict
liability does not seek to “exclude the defendant from blasting” but instead “merely seek[s] compensation for
the damage”). If these activities were presumptively unreasonable and prohibited, the mere choice to engage in
them would subject the dutyholder to punitive damages. See note 52.
74
Jaffey, “Duties and Liabilities in Private Law” (note 71) at 153 (identifying the impossibility of engaging in
risky behavior without ever harming another as the reason why strict liability “is widely thought to be unjust”).
75
A “liability rule” exclusively relies on the compensatory damages remedy to protect the rightholder’s
interests, unlike a “property rule” that immunizes these interests from harm absent the rightholder’s consent
and accordingly employs injunctive relief as a remedy. See Guido Calabresi and A. Douglas Melamed,
“Property Rules, Liability Rules, and Inalienability: One View of the Cathedral,” 85 Harv. L. Rev. 1089
(1972). For reasons discussed in the text, the rule of strict liability criticized by Weinrib is a property rule
rather than a liability rule.
84 MARK A . GEISTFELD

An absolute right to physical security would create the injustice identified by

Weinrib—the security of such a rightholder would have absolute dominion over the
conflicting liberty interest of a dutyholder—but a compensatory tort right is defined by
a relative, default priority of the rightholder’s interest in physical security over the
dutyholder’s interest in liberty.76 The relative priority recognizes that both security
and liberty are required for the exercise of autonomy within the liberal egalitarian
community, thereby giving rightholders and dutyholders an equal opportunity for
self-determination. There is nothing inherently unjust about a compensatory rule of
strict liability
Indeed, if the compensatory duty is overly onerous for the liberty interest, then tort
law limits the duty accordingly. For example, the rule of strict liability for abnormally
dangerous activities does not apply to socially valuable activities.77 When social value
is categorically defined by reference to the autonomy interests of all parties who would
be governed by the duty, then strict liability can be justifiably limited if it would cause a
loss of social value (or limitation of autonomy for the relevant category of liberty
interests) that exceeds the gain in social value (promotion of autonomy by categoric-
ally protecting the security interests of rightholders), yielding the rule of negligence
liability for this category of risky interactions. The same principle then applies to
negligence liability and explains both the partial limitations of duty and the full
immunities from tort liability.78 The way in which tort liability can unduly curtail
the exercise of liberty only justifies the varied limitations of the tort duty rather than
the wholesale rejection of strict liability.

C. The relation between corrective and distributive justice

As a form of corrective justice, a compensatory tort system resolves a tort dispute
without any reliance on the principle of distributive justice. Such a compensatory tort
system or its functional equivalent (sufficiently demanding and comprehensive regu-
lation plus social insurance), however, is essential for implementation of a liberal
egalitarian scheme of distributive justice. By implementing corrective justice, a com-
pensatory tort system establishes the normatively appropriate baseline of wealth
and resources against which the complementary distributive scheme operates.79
A compensatory tort system, therefore, is a form of corrective justice that is distinct
from a scheme of distributive justice such as equality of resources. The two forms of
justice are instead complementary or morally coherent in that each one finds justifi-
cation in the same underlying abstract principle of equality.

76
See Part II.A. 77
Restatement (Second) of Torts, } 520(f) (1965), cmt. k.
78
See Geistfeld, Tort Law (note 7) at 91–7.
79
For more rigorous demonstration, see Mark A. Geistfeld, “Efficiency, Fairness, and the Economic
Analysis of Tort Law,” in Mark D. White (ed.), Theoretical Foundations of Law and Economics (New York:
Cambridge University Press, 2009), 234–52.
 COMPENSATION AS A TORT NORM 85

V. Conclusion
Tort law entitles a rightholder to an award of compensatory damages under quite
limited conditions, a fundamental feature of liability that would seem to foreclose a
compensatory conception of tort law. A compensatory tort right, however, does not
necessarily entail an entitlement to compensatory damages in all cases. The exercise of
reasonable care by the dutyholder can fully satisfy the compensatory obligation. Such a
compensatory duty finds justification in the principle of liberal egalitarianism that
makes an individual responsible for the foreseeable consequences of her autonomous
choices. Compensation is a defensible norm of justice that can persuasively explain
tort doctrine, despite the limited availability of the compensatory damages remedy.
 4
Tort as a Substitute for Revenge
Scott Hershovitz*

I. Introduction
In 1872, the Supreme Court of Illinois decided a case called Alcorn v Mitchell.1 It was
not the first litigation between the parties. Some years earlier, Alcorn had sued
Mitchell for trespass.2 That suit did not go well, and at the close of the trial, just
after the court adjourned, Alcorn spit in Mitchell’s face.3 Mitchell then turned the
tables and sued Alcorn for battery. He won a judgment for $1,000, which was a lot of
money back then—depending on how you think about the change in value of money
over time, the present day equivalent would range from just about $20,000 to a bit
more than $250,000.4 Alcorn appealed. The question put to the Illinois Supreme Court
was whether he had “been made to pay too dearly for [his] indulgence.”5
The court did not think so. Toward the end of its opinion, it noted that Alcorn was a
“man of wealth.” 6 But it placed more emphasis on what he did than on who he was. “It
is customary,” the court observed, “to instruct juries that they may give vindictive
damages where there are circumstances of malice, willfulness, wantonness, outrage

* Professor of Law and Professor of Philosophy, University of Michigan, Ann Arbor. Thanks to Andrew
Gold, John Goldberg, Don Herzog, Bill Miller, John Oberdiek, and the participants in the Legal Theory
Workshop at the American University Washington College of Law for helpful comments and conversations.
1
Alcorn v Mitchell, 63 Ill. 553 (1872).
2
The only published opinion in the litigation between Alcorn and Mitchell is the 1872 Illinois Supreme Court
opinion from the second suit. It refers to the parties as plaintiff and defendant, without saying who was who.
However, the docket in that case (available in the Illinois State Archives) makes clear that Alcorn was the plaintiff
in the first suit and Mitchell the plaintiff in the second. See Pleas and Proceedings in the Circuit Court, Alcorn v
Mitchell, 65 Ill. 553 (1872) (Supreme Court Docket No. 88, Agenda No. 15) (herinafter “Pleas and Proceedings”).
3
The opinion in the later case describes Alcorn as an “exasperated suitor.” Alcorn, 63 Ill., at 554.
4
Simply updating for inflation suggests that the equivalent award today would be about $19,400. But we
might get a better read on the magnitude of the award by gauging its relationship to per-capita gross domestic
product (GDP), which was $195.64 in 1872. Since per-capita GDP in 2012 was $49,927.74, a similarly sized
award today would be roughly $255,000. I obtained both measures from calculators at <http://www.
measuringworth.com> accessed October 23, 2013.
5
Alcorn, 63 Ill., at 554. Alcorn also challenged the jury instructions, but the court did not address them.
6
Alcorn, 63 Ill., at 554.
 TORT AS A SUBSTITUTE FOR REVENGE 87

and indignity attending the wrong complained of. The act in question was wholly
made up of such qualities. It was one of pure malignity, done for the mere purpose of
insult and indignity. ”7 Alcorn, the court said, gratified “his malignant feelings in [a]
despicable mode.”8 And worse than that, he did so in “the temple of justice . . . in the
very presence of its ministers.”9
So it seems fair to say that the court did not think much of Alcorn’s spit. But why
damages, and why such a substantial sum? There is no hint that Mitchell suffered an
injury beyond the indignity, no hint that he suffered an injury that money might
repair. Yet, the court still saw a purpose to the payment: “The act in question was one
of the greatest indignity, highly provocative of retaliation by force, and the law, as far
as it may, should afford substantial protection against such outrages, in the way of
liberal damages, that the public tranquility may be preserved by saving the necessity
of resort to personal violence as the only means of redress.”10 In other words, Alcorn
had to pay so that Mitchell would not have to strike back. The tort suit was to
substitute for the revenge the court worried Mitchell might take absent a civil means
of redress.
The idea that tort suits are a substitute for revenge is still with us today.11 But it is not
clear how the substitution is supposed to work. One possibility is that tort suits are a
substitute for revenge simply because some people who would seek revenge are willing
to forego it if they can obtain money damages in a tort suit instead. On this picture, tort
damages are a payoff to prevent private violence. This is a rather thin sense of
substitution, as it does not presuppose any connection between tort and revenge beyond
the preferences of potential plaintiffs. In this sense, a scheme of taxpayer-funded trips to
Tahiti might substitute for revenge, if victims would forego violence in favor of it.
There is a hint of the payoff picture in Alcorn, when the court says that liberal
damages are warranted so that the “the public tranquility may be preserved.”12 But
I suspect that the court meant to say more by way of justifying the award. After all, the
payoff picture is consistent with a dim view of Mitchell and his pursuit of damages.
A better man, the court might have thought, would let the slight slide and not hold the
court hostage to the threat of retaliation. But there is no hint in the opinion that the
court regrets the need to purchase public tranquility at the cost of liberal damages. Quite
the opposite. The court seems pleased to approve the judgment, explaining that “[s]o
long as damages are allowable in any civil case by way of punishment or for the sake of
example, the present, of all cases, would seem to be a most fit one for the award.”13
The talk of punishment here suggests another way we might think of the substitu-
tion. Today, we would call the damages that Alcorn was ordered to pay “punitive.” But
though the court clearly saw the damages as a form of punishment, it labeled them

7 8
Alcorn, 63 Ill., at 554. Alcorn, 63 Ill., at 554.
9 10
Alcorn, 63 Ill., at 554. Alcorn, 63 Ill., at 554.
11
See Mathias v Accord Economy Lodging, Inc., 347 F.3d 672, 677 (7th Cir. Ill. 2003) (Posner, J.) (“[A]n an
age-old purpose of the law of torts is to provide a substitute for violent retaliation against wrongful injury.”).
12 13
Alcorn, 63 Ill., at 554. Alcorn, 63 Ill., at 553.
88 SCOTT HERSHOVITZ

“vindictive,” not “punitive.” Formally, there is no difference. In most jurisdictions,

“punitive” and “vindictive” are used interchangeably to denote extra-compensatory
damages awarded for willful or wanton misconduct. But “vindictive” is a revealing
label, as it suggests that an award of such damages is responsive to the attitude
associated with revenge. And it is a short step from thinking of the damages as
vindictive to regarding them as a kind of revenge inflicted through a tort suit. On
this picture, the substitution is not so much tort for revenge, as it is a bloodless revenge
for a bloody one.
This thought can be developed in different ways. One version carves off punitive
damages from the rest of tort, holding that revenge comes into play only when
plaintiffs are permitted to seek extra-compensatory damages.14 Another takes the
label “vindictive” to make manifest as to punitive damages what is true of all tort
damages. After all, punitive damages are not the only damages that are painful to pay.
This has led some to suggest that a tort suit provides a plaintiff an opportunity to take
revenge by inflicting harm on the person who harmed her, even when only compen-
satory damages are on offer. And that suggestion is sometimes buttressed by the
observation that compensatory damages are not nearly so tied to the extent of the
plaintiff ’s injury as the name might suggest. By operation of several different doc-
trines, they often run beyond or fall short of the plaintiff ’s losses, making it hard, some
say, to take seriously tort’s talk about making plaintiffs whole.15
I do not doubt that tort can serve as an outlet for vindictive motives, and I think that
true even when punitive damages are not in play. But it is a mistake to regard tort
damages as a kind of revenge. Tort works rather differently, and not just because it is
civil rather than violent. We will see some of the differences later on. For now, I want
to continue to catalog ways in which tort might substitute for revenge. We started with
the thought that tort might be a way of paying off plaintiffs, so that they refrain from
revenge. Then, we considered the possibility that a tort suit (especially, but perhaps not
only, a successful one) just is a kind of revenge. The last possibility I want to put on the
table is that tort suits are a substitute for revenge because they can be used for the same
purposes.
This is obviously true, in at least one respect. Both tort and revenge can be used to
deter. You would think twice if the person you spit on might spit back. And you would
also think twice if you might be forced to pay $20,000 for the privilege.16 When the
Alcorn court said that it was appropriate to award damages “for the sake of example,”
deterrence is surely what it had in mind. Of course, there is nothing distinctive about
this relationship between tort and revenge. Regulation is a tool for deterrence, and
criminal prohibition is too. As means of deterrence, each has its costs and benefits.
But though a pairwise comparison between tort and revenge might make tort seem

14
See Anthony J. Sebok, “Punitive Damages: From Myth to Theory,” 92 Iowa. L. Rev. 957 (2007).
15
See Emily Sherwin, “Compensation and Revenge,” 40 San Diego L. Rev. 1387 (2003).
16
The standard idiom here is a bit misleading: damages do not purchase a privilege, not even in retrospect,
as paying damages does not legitimize the conduct that led to them.
 TORT AS A SUBSTITUTE FOR REVENGE 89

attractive, it is far less certain that tort holds up well against the others, at least for
much of the conduct that can underwrite a tort suit. It seems doubtful, for example,
that the best way to deter medical malpractice, or risky driving for that matter, is
through a system of private tort suits. So if we have deterrence in mind, we might end
up thinking that tort is a substitute for revenge, but far from the best available.
Deterrence, however, may not be the only purpose to which both tort and revenge
could be put. Another possibility is that they can both be used to do corrective justice.
This will not seem an odd thought as to tort, as philosophers have long argued that tort
suits enforce moral duties of repair.17 But it will strike many as an odd thought as to
revenge, which we are often told has little to do with justice.18 Of course, that is a
distinctly modern attitude. Francis Bacon pronounced revenge “a kind of wild just-
ice,”19 even as he warned against it. And for those who followed the law of the talion,
getting justice meant getting even, often through violent retaliation.
I think that Bacon and adherents of the talion had it right: revenge is a tool for doing
corrective justice. And tort is a substitute for revenge because it is too. Of course, the
thought that tort is a tool for doing corrective justice is not novel. But it is often
misunderstood, and reflecting on revenge can help us see why. My aim in this chapter
is to think through what Mitchell might have gotten from revenge and what tort
offered him in its place. But before we do that, it will help to think a bit about
corrective justice.

II. Corrective Justice

Contemporary thinking about corrective justice is dominated by a picture that we owe
to Aristotle. Aristotle imagined two parties in an initial position of equality, repre-
sented by a line divided into equal segments. When one party disturbs that equality,
lengthening her own segment and shortening the other’s, corrective justice demands
that we restore the equality, taking from the party with the lengthened segment just
what is necessary to restore the shortened one.20
On this way of thinking, corrective justice is all about allocation. Here is how John
Gardner explains it:

17
The classic statements of the corrective justice take on tort law are: Jules Coleman, Risks and Wrongs
(Cambridge: Cambridge University Press, 1992), and Ernest J. Weinrib, The Idea of Private Law (Cambridge,
MA: Harvard University Press, 1995).
18
See Gabriel Seltzer Mendlow, “Is Tort Law a Form of Institutionalized Revenge,” 39 Fla. St. U. L. Rev. 129
(2011) (“Most would agree that institutions of revenge have no place in a just society . . . ”); See also Robert
Nozick, Philosophical Explanations (Cambridge, MA: Belknap Press, 1981), 366–8 (distinguishing revenge
from retribution and suggesting that the latter has a connection to justice that the former does not).
19
Francis Bacon, “Of Revenge,” in Brian Vickers (ed.), Francis Bacon: The Major Works (Oxford: Oxford
University Press, 2002), 347.
20
Aristotle, Nicomachean Ethics, bk. V, Roger Crisp (ed. & trans.) (Cambridge: Cambridge University
Press, 2000), 88.
90 SCOTT HERSHOVITZ

Something has shifted between . . . two parties. The question of corrective justice is not the
question of whether and to what extent and in what form and on what ground it should
now be allocated among them full stop, but the question of whether and to what extent and
in what form and on what ground it should now be allocated back from one party to the
other, reversing a transaction that took place between them.21

The Aristotelian picture fits some forms of wrongdoing well. If Alcorn had stolen
Mitchell’s horse, corrective justice might demand that he give it back, reversing the
wrongful transaction.
But Alcorn did not steal from Mitchell. He spit on him. To be sure, something
shifted between them. But it hardly seems that the question of corrective justice in this
case could be about the allocation of Alcorn’s spit. After all, the problem here is not
that Mitchell possesses something to which he has no right; it is that he was subjected
to treatment that he had a right to be free from. And there is no way to reverse that
transaction. This is not because we cannot force the spit back on Alcorn. Perhaps we
could. But even if we did, that would not change the fact that Mitchell was spit upon.
Of course, we could say much the same when Alcorn steals Mitchell’s horse. Mitchell
had a right to his horse, which Alcorn infringed when he stole it. Corrective justice
might well demand that Alcorn return the horse, but there is no changing the fact that
Mitchell stole it. And we might wonder whether corrective justice requires that we do
anything about that. After all, returning the horse is just what we would expect Alcorn to
do had he found it, so it hardly seems an adequate response to the fact that he stole it.
Now one possibility is that corrective justice demands only that Alcorn return the
horse, leaving other moral principles to determine whether a further response to his
wrongdoing is warranted. This is Gardner’s view. He says that norms of justice are
“norms for tackling allocative moral questions, questions about who is to get how
much of what.”22 Within that domain, norms of corrective justice address questions
about allocating something back from one person to another, once it has shifted as the
result of wrongdoing.23 But that is all corrective justice does. If there is more to be done
in response to wrongdoing, it is not because corrective justice calls for it.24
I find this conception of corrective justice much too cramped. It has the consequence
that many of the most serious wrongs do not call for corrective justice at all, as they do
not involve transfers that can be reversed through reallocation. If Alcorn had raped
Mitchell, for example, there would be no corrective justice to be done, as there would be
no horse to hand over. So too with the spit. And that is a rather startling conclusion.
Part of the problem here stems from the fact that wrongdoing does not always
generate a gain for the wrongdoer that matches the victim’s loss. Had Alcorn stolen
Mitchell’s horse, he would have gained the very thing that Mitchell lost. But Alcorn

21
John Gardner, “What is Tort Law For? Part I. The Place of Corrective Justice,” 30 Law & Philosophy 1
(2011), 9–10 (hereinafter “What is Tort Law For? Part I”).
22
Gardner, “What is Tort Law For? Part I” (note 21) at 6.
23
Gardner, “What is Tort Law For? Part I” (note 21) at 9–10.
24
Gardner, “What is Tort Law For? Part I” (note 21) at 47.
 TORT AS A SUBSTITUTE FOR REVENGE 91

may not have gained anything from spitting on Mitchell, and if he did gain something
(satisfaction, perhaps), it is neither the thing that Mitchell lost, nor something that he
could hand over. So the Aristotelian imagery is off: Mitchell’s segment of the line may
have been shortened, but Alcorn’s does not seem to have been lengthened. And even if
it was, we cannot rejigger the lines to get back where we started.
This sort of worry leads some to tweak the Aristotelian picture, so that corrective
justice requires restoring Mitchell’s segment of the line, even if that will leave Alcorn
with a shorter segment than he started with. The idea here is that the loss that Alcorn
inflicted on Mitchell is properly his to deal with, so he must restore Mitchell to his
rightful position and handle the consequences on his own.25 And there are cases in
which it looks like this might work. If Alcorn destroyed Mitchell’s horse, for example,
then requiring that he pay Mitchell its market value would seem to make the loss
Alcorn’s to deal with. Mitchell would be able to acquire more or less what he lost and,
Alcorn would be out money, not Mitchell.
But the problem here should be apparent, as it is symmetrical to the one we just
faced. Many kinds of wrongdoing do not involve losses that can be shifted to the
wrongdoer. It is not obvious, for example, that Mitchell lost anything when Alcorn spit
on him, but if he did (some measure of dignity, perhaps), it is not the sort of loss he can
transfer to Alcorn. This would be even more stark had Alcorn physically injured
Mitchell. If he had put his eye out, for example, we could transfer some of the
consequential losses (medical expenses, lost income, and the like) and make them
Alcorn’s problem, rather than Mitchell’s. But Mitchell would surely miss his sight, and
he could not pass that loss on to Alcorn.
Of course, there is a touch of this problem in the horse case too. As I said, Mitchell
might use his damages to acquire more or less what he lost. But he cannot recover just
what he lost—that horse has left the barn—and the difference might well matter. So it
turns out that, even when we tweak the Aristotelian imagery, it can serve as a picture of
corrective justice only for a narrow set of cases: those in which the victim has lost a
commodity, and just that. In all other cases, the victim must suffer at least some of her
loss herself.
The Aristotelian picture holds out a false promise.26 It makes it seem like we can get
back where we started; we just have to rejigger the line. But we can never get back
where we started. When one person wrongs another, the wrong is part of our history,
indelibly, and the challenge is to figure out how to go on, not how to go back. Of
course, everyone who writes about corrective justice appreciates this, at least to some

25
See, e.g., Arthur Ripstein, Equality, Responsibility, and the Law (Cambridge: Cambridge University Press,
2001), 58 (“But insofar as they enable a plaintiff to adapt to his or her situation, money damages are an
appropriate way of transferring the loss so that it becomes the injurer’s problem to decide how to deal with
what is properly his or her loss.”).
26
I develop this argument further in Scott Hershovitz, “Corrective Justice for Civil Recourse Theorists,” 39
Fla. St. L. Rev. 107 (2011), 100–17. John Goldberg and Ben Zipursky sound similar themes when they argue
that loss-focused conceptions of corrective justice do not make sense of tort doctrine. See John C.P. Goldberg
and Benjamin C. Zipursky, “Torts and Wrongs,” 88 Texas L. Rev. 917 (2010), 954–7.
92 SCOTT HERSHOVITZ

degree. So they douse their claims in qualifiers. They say that corrective justice consists
in “undoing, to the extent possible, the very injustice that has been committed,”27 or
“restoring things, at least in some measure, to where they would have been had one not
occasioned their loss.”28 And maybe we should do these things when we can. But these
qualifiers are a dodge. We can never restore things to just the way they were, and all
too often, we cannot restore them at all. So if that is what corrective justice demands,
we are bound to fall short in all cases, and we will not be able to do anything that
counts as corrective justice in many.
That would be depressing, if it were true. It would be nice to think that we could do
corrective justice for Mitchell, whether Alcorn steals his horse, spits on him, or
sexually assaults him. And it would be nice to think that our ability to do so does
not depend on the happenstance of Mitchell having incurred losses that we could shift
to Alcorn through a damage award. Happily, I think that we can offer Mitchell
corrective justice in all these cases. But to see how, we need to set the Aristotelian
picture aside and make a fresh start.

III. Corrective Justice Corrected

Let us turn our attention back to the actual facts of Alcorn v Mitchell. Alcorn had sued
Mitchell for trespass. The charge seems to have been that Mitchell took down Alcorn’s
fence without permission.29 Mitchell testified on his own behalf, apparently in a
manner that Alcorn found damaging. When the court adjourned, Alcorn approached
Mitchell and spit in his face.
But it was not just any spit, at least not according to Mitchell. Here is how his
complaint described the event:
[T]he plaintiff . . . avers that because he the plaintiff did and would testify the truth the
whole truth and nothing but the truth and because the same proved to be matter against . . .
Alcorn in his said suit . . . Alcorn then and there did willfully, maliciously and Devilishly
spit upon the plaintiff, and then and there for the cause and reason aforesaid squirted into
the face and eyes of the plaintiff a large quantity of filthy matter out of and from the mouth
of the defendant, consisting of saliva, mingled with divers filthy noisome and disgusting
drugs, whiskey and other kinds of nastiness then and there prepared by the defendant for
the express purpose of being by him squirted in the plaintiff ’s face as aforesaid.30

The court’s worry that Mitchell might strike back was not fanciful. His complaint
continued:

27
Ernest J. Weinrib, Corrective Justice (Oxford: Oxford University Press, 2012), 36 (emphasis added).
28
Gardner, “What is Tort Law For? Part I” (note 21) at 37 (emphasis added).
29
Testimony of J.B. Gillaspie, Records from Marion County Proceedings, 3, Alcorn v Mitchell, 65 Ill. 553
(1872) (Supreme Court Docket No. 88, Agenda No. 15) (hereinafter “Testimony of J.B. Gillaspie”).
30
Pleas and Proceedings, at 4–5.
 TORT AS A SUBSTITUTE FOR REVENGE 93

[T]he people who were present demand[ed] that the plaintiff ought to have then and there
killed the defendant, which temptation the plaintiff resisted out of respect to the laws of
God and the State, and to the great displeasure and disgust of the numerous. To wit, One
hundred bystanders who were indignant that so gross an insult should pass unrevenged.31

It is hard to know how seriously to take all that, but the court took it seriously enough
to uphold significant damages, on the ground that the act in question was highly
provocative of retaliation by force.
But our question is not whether Mitchell would have actually taken revenge had he not
been able to file a tort suit. It is what he would have gotten from revenge had he taken it.
For guidance, we can look again to the complaint, which recounts the fallout from having
let the insult pass: “[I]n consequence the plaintiff hath been greatly injured in his feelings
in his circumstances, his credit and his reputation.”32 That sounds a bit like boilerplate,
but, on these facts, it is easy to believe, and it is worth asking why.
The answer, I think, is that Alcorn’s spit sent a message. Here is how Jeffrie Murphy
explains the messages implicit in moral injury:
One reason we so deeply resent moral injuries done to us is not simply that they hurt us in
some tangible or sensible way; it is because such injuries are also messages—symbolic
communications. They are ways a wrongdoer has of saying to us, “I count but you do not,”
“I can use you for my purposes,” or “I am here up high and you are there down below.”33

That last message, especially, would have been easy for Alcorn’s audience to take away.
As the court’s opinion noted, Alcorn was a wealthy man; estimates of his net worth at
trial ranged from $50,000 to $200,000 (or somewhere between $1 and $4 million,
updated for inflation).34 Mitchell, on the other hand, was said to be worth just
$2,000.35 So people may well have been primed to hear the message that Alcorn was
up high and Mitchell down low. Indeed, it is easy to imagine that the very point of
Alcorn’s spit was to reassert that message in the wake of the court crediting Mitchell’s
testimony rather than his own.
Messages of this sort matter. As Murphy says: “Most of us tend to care what others
(at least some others, some significant group whose good opinion we value) think
about us—how much they think we matter. Our self-respect is social in at least this
sense, and it is simply part of the human condition that we are weak and vulnerable in
these ways.”36

31 32
Pleas and Proceedings, at 5. Pleas and Proceedings, 5.
33
Jeffrie G. Murphy, “Forgiveness and resentment,” in Jeffrie G. Murphy and Jean Hampton, Forgiveness
and Mercy (Cambridge: Cambridge University Press, 1990), 25.
34
The estimates of Alcorn’s net worth appear in transcripts of testimony contained in Records from
Marion County Proceedings, 2, Alcorn v Mitchell, 65 Ill. 553 (1872) (Supreme Court Docket No. 88, Agenda
No. 15). Once again, I have taken rough inflation figures from <http://www.measuringworth.com>. If you
prefer to compare the figures to per-capita GDP, then present equivalent of Alcorn’s net worth would range
from $12 million to $50 million.
35
Argument and Brief of Appellee, Alcorn v Mitchell, 65 Ill. 553 (1872) (Supreme Court Docket No. 88,
Agenda No. 15).
36
Murphy, “Forgiveness and resentment” (note 33) at 25.
94 SCOTT HERSHOVITZ

That vulnerability makes messages of the sort Alcorn’s spit sent threatening, and the
threat lingers if we do not respond to them. Here’s how Pamela Hieronymi puts it:
I suggest that a past wrong against you, standing in your history without apology,
atonement, retribution, punishment, restitution, condemnation, or anything else that
might recognize it as a wrong, makes a claim. It says, in effect, that you can be treated
this way, and that such treatment is acceptable.37

To be clear, the threat that Hieronymi means to highlight is not that you might be
subjected to the same treatment again, though that may be a worry if people come to
think it is acceptable to treat you that way. Rather, the threat is to your social standing,
which is diminished if people think it was okay to treat you that way, even if it is not
likely to happen again.38
The simplest way to recognize a wrong is to resent it, and many philosophers have
thought resentment vital to self-respect. Heironymi argues that resentment is war-
ranted when you hold three judgments:
(1) The act in question was wrong; it was a serious offense, worthy of moral attention. (2)
The wrongdoer is a legitimate member of the moral community who can be expected not
to do such things. As such, she is someone who can be held responsible and she is worth
being upset by. (3) You, as the wronged, ought not be wronged. This sort of treatment
stands as an offense to your person.39

If Mitchell had been able to deny any of these propositions, he would have had no need
for resentment. If, for example, he thought that Alcorn was not worth being upset by,
he could simply dismiss him. But Mitchell probably thought that Alcorn mattered, and
given the crowd’s reaction, he surely knew that other people did. In those circum-
stances, if Mitchell could not even muster so much as resentment in protest of Alcorn’s
spit, then he would seem to concede that he may be spit upon, at least by the likes of
Alcorn.
But though resentment would have been a way for Mitchell to resist the thought that
he may be spit upon, it would not have been, by itself, a terribly effective way of
countering the threat to his social standing. Hence, Hieronymi’s list: apology, atone-
ment, retribution, punishment, restitution, and condemnation. To different degrees,
these are all public performances that would mark Alcorn’s behavior as wrong, and
thus stand as a counterweight to the message that his spit sent. Revenge does not make
Hieronymi’s list, but she later adds that “taking revenge might be one way to ‘correct’
the historical significance of the event, marking it as a wrong.”40
I think that is right, and I do not see any reason for the scare quotes. Had Mitchell
done as the crowd demanded and killed Alcorn, he would have sent the message that

37
Pamela Hieronymi, “Articulating an Uncompromising Forgiveness,” 62(3) Philosophy and Phenomeno-
logical Research 529 (2001), 546.
38
Hieronymi, “Articulating an Uncompromising Forgiveness” (note 37) at 548, n. 31.
39
Hieronymi, “Articulating an Uncompromising Forgiveness” (note 37) at 530.
40
Hieronymi, “Articulating an Uncompromising Forgiveness” (note 37) at 548.
 TORT AS A SUBSTITUTE FOR REVENGE 95

he may not be spit upon, that he regards it is an offense to his person, which he takes
seriously and will not tolerate. Of course, killing Alcorn seems much too much, as it
suggests that the wrong was much graver than it in fact was. There were surely
responses short of death that would have countered the message in Alcorn’s spit.
Indeed, there’s a certain genius in the law of the talion, with its famous formula an eye
for an eye. When a victim treats a wrongdoer just the way the wrongdoer treated her,
she communicates that she was wronged and will not stand for it. But she also claims a
kind of equality—an entitlement to treat others as they treat her.
That equality is central to the way that revenge works, or at least it was for adherents
of the talion. For them, getting justice was about getting even. But the evenness
involved was not about allocation, as the Aristotelian conception of corrective justice
would have it. Rather, it was an evenness in a certain kind of social standing, expressed
in a formula that inverts the golden rule: you may do unto others as they do unto you.
Much modern commentary about revenge misses this. The mistake is on display in
Robert Nozick’s suggestion that “[r]etribution sets an internal limit to the amount of
punishment, according to the seriousness of the wrong, whereas revenge internally
need set no limit to what is inflicted.”41 There’s a logic to the talion, and the logic is
largely about limits; getting even is not going postal.42
To be sure, most revenge cultures left some play in the joints. The rules were not so
rigid as to require that the victim do to the wrongdoer just what the wrongdoer did to
her. But it better be possible to sell the response as a way of getting even, or it would
not be seen as a justified act of revenge.43 And that leads me to doubt that Mitchell
would have killed Alcorn over the spit, or that he would have been treated well had he
done so. It also leaves me a little suspicious of his claim that the crowd demanded it.
But it might have: Alcorn’s spit presumably posed a threat to many in the crowd (if it
was okay for Alcorn to spit on Mitchell, it might be okay for him to spit on them too),
and in the heat of the moment crowds often take things too far.
Whatever the crowd demanded, we are now in a position to appreciate at least one
thing that Mitchell would have gotten from taking revenge. He would have countered
Alcorn’s message about his social standing. And if his revenge was properly calibrated,
he would have asserted, in at least one fundamental sense, that he was Alcorn’s equal.
That strikes me as a kind of justice, and indeed corrective justice. But what corrective
justice so construed corrects is the historical significance of the wrong—the threat that
it poses—and not an aberration in the prevailing allocation of goods.

41
Nozick, Philosophical Explanations (note 18) at 367.
42
William Ian Miller, Eye for an Eye (New York: Cambridge University Press, 2006), 24 (“[W]e think of
revenge as going postal and blasting away, but revenge cultures did not think of it that way. For them, revenge
was not just an ethic but an aesthetic, the aesthetic of proportion and balance.”).
43
I explore getting even in more detail in Hershovitz, “Corrective Justice for Civil Recourse Theorists” (note
26) at 117–25. As I emphasize there, getting even has a performative dimension, and success in the
performance depends in part on prevailing assessments of value.
96 SCOTT HERSHOVITZ

IV. The Message in the Money

What did Mitchell get from his tort suit? Money, of course. But the suggestion I want
to make is that the money was paid in service of a message that is much the same as the
message Mitchell would have sent by taking revenge. In finding Alcorn liable for
battery, the jury marked his conduct as wrong, rebutting any suggestion that it was
permissible for him to spit on Mitchell. And this was not just a pro forma prelude to
the damage award. Quite the contrary. The message seems to have been just what
Mitchell was after when he filed suit. In settlement negotiations, Mitchell told Alcorn
that “all he wanted” was for “Alcorn to acknowledge that he’d done wrong.”44 Indeed,
he disclaimed any interest in Alcorn’s money, explaining that he filed suit only after his
attorney, among others, told him that he “ought not to stand it to be spit on.”45 So it
seems apt to say that the lawsuit was Mitchell’s way of standing up for himself.
Of course, the jury did not just find Alcorn liable; it imposed $2,000 in vindictive
damages (the trial court reduced the award to $1,000 for reasons that are not clear
from the record). Why damages, and why such a substantial sum? For the same reason
that Alcorn spit on Mitchell. He could have hurled insults at him, but his spit showed a
contempt that words could hardly communicate. (As the old saying goes, actions
speak louder than words.) The jury was in much the same position. It could have
simply said that Alcorn had wronged Mitchell, entering a nominal damage award, if
that was formally required. By attaching substantial damages to its verdict, the jury
made its message unmistakable: Alcorn’s conduct was wrong, and seriously so. That
message still comes through loud and clear today. My students are routinely shocked
at how seriously the jury took Alcorn’s spit, as none of them can imagine being
ordered to pay $20,000, let alone $250,000, for a similar transgression today.
This is not the story that corrective justice theorists commonly tell about punitive
damages. They tend to argue that punitive damages are surreptitiously compensatory,
so that they fit into the Aristotelian picture, or that they are out of place, since a
plaintiff does not have a claim to the money (or anything equivalent) prior to the
wrong.46 But punitive damages can clearly contribute to corrective justice if one takes
on board the suggestion that corrective justice is about correcting the message that
wrongdoing sends, not about altering allocations in an effort to restore the status quo

44
Testimony of J.B. Gillaspie, 3. Alcorn did eventually apologize. According to Gillaspie, he acknowledged
that “it was a little thing to spit on any man.” He also said it was “no credit to him” that he had done so, that it
was “wrong” and that he was “sorry.” Nevertheless, the case did not settle. Some witnesses suggested that
attorneys’ fees held up the settlement; Alcorn offered to pay $25 toward attorneys’ fees, but Mitchell’s attorneys
likely stood to do much better than that with a contingency fee. However, there are reasons to doubt that the
attorney stood in the way. One witness reported that Mitchell’s lawyer told him that he “could do as he pleased
about settling,” but that “he would not let any man spit in his face for ten thousand dollars.” Testimony of
A.J. Douherty, Records from Marion County Proceedings, 3, Alcorn v Mitchell, 65 Ill. 553 (1872) (Supreme
Court Docket No. 88, Agenda No. 15). I think it is possible that Mitchell used his attorney as a convenient
excuse. Apologies are easy to fake, and by the time Mitchell filed suit, he may well have come to prefer public
vindication to an apology offered only as an inducement to settle.
45
Testimony of J.B. Gillaspie. 3.
46
See Benjamin C. Zipursky, “A Theory of Punitive Damages,” 84 Texas L. Rev. 105 (2005), 136.
 TORT AS A SUBSTITUTE FOR REVENGE 97

ante. Punitive damages help mark the gravity of a wrong. And because juries are
permitted to consider the defendant’s wealth when imposing them, they can ensure
that the victim’s dignity will not be seen as something available for purchase (at least
not without a price that is painful).
But what should we think, then, about compensatory damages, which often do seem
like an effort to restore the status quo ante, or at least aspects of it? They serve a
communicative purpose too. Once again, here is Hieronymi:
[A]ny wrongdoing leaves in its wake some amount of damage or cost, be it physical,
financial, emotional, relational, or social. This is damage which the offender usually cannot
repair (“you can’t take it back,” as children learn), and which the offended will, in any case,
incur. The persistence of the damage threatens any attempt to leave the past in the past,
insofar as the damage testifies to the deed.47

Hieronymi is refreshingly upfront about the problem that corrective justice theorists tend
to gloss over: a victim almost always suffers damage that cannot be repaired. Hieronymi
argues that when a victim forgives a wrongdoer she agrees to bear that damage without
further protest.48 And that strikes me as right. But her observation can also help us
explain why we might require a wrongdoer to undo the damage, or mitigate its costs
when he can. So long as the damage persists, it “testifies to the deed,” and leaves the threat
to the victim’s social standing in place: it just might be okay to treat her that way, because
she was treated that way, and she has been left to absorb the consequences, even though
they might have been repaired. When we require repair, we affirm the message that the
conduct in question was wrong and the results the wrongdoer’s responsibility.
The same observation can also help explain why we invite juries to award compen-
sation for injuries that cannot be repaired, or even sensibly valued in monetary terms.
Pain and suffering damages, for example, do not fix pain and suffering. And there is
typically no price at which people would agree to endure serious pain and suffering, so
the damages cannot be set to simulate a market rate. But pain and suffering damages
still serve a purpose: they mark the pain and suffering as the responsibility of the
defendant. To do that effectively, their magnitude must reflect the weight of the injury
that the plaintiff must bear. Thus, serious suffering merits a serious sum, even though
the serious sum cannot be taken seriously as a price or means of repair.
The bottom line is that tort law does not do corrective justice only, or even
primarily, by enforcing moral duties of repair. It does, to be sure, enforce duties of
repair. But not all tort damages are reparative, and they are no less a part of corrective
justice for that. Moreover, the damages that are reparative do corrective justice in the
same way that the non-reparative damages do. They do justice through the message
that they send about the victim’s standing and the wrongdoer’s responsibility.49

47
Hieronymi, “Articulating an Uncompromising Forgiveness” (note 37) at 550.
48
Hieronymi, “Articulating an Uncompromising Forgiveness” (note 37) at 551.
49
Stephen Smith gestures toward a similar view of the relationship between tort damages and corrective
justice near the end of his “Why Courts Make Orders (And What This Tells Us About Damages),” 64 Current
Legal Problems 51, 84-86 (2011). However, he stops short of embracing a communicative picture of either
98 SCOTT HERSHOVITZ

V. Tort or Revenge?
The lesson so far is that, in at least one important respect, tort offered Mitchell the
same thing that revenge did. It offered him a way of countering the message that
Alcorn’s spit sent, a way of correcting the historical significance of Alcorn’s wrong.
Tort law did not just pay off Mitchell to purchase public tranquility; it saved him the
necessity of resort to private violence by offering him a civil means to do what he
might have done through force. This is the chief reason tort and revenge are substi-
tutes: they can both be used to do corrective justice.
Which is the better way to do it? Tort has one obvious advantage. It is civil rather
than violent, and to the extent we can, it seems better to correct the historical
significance of wrongs without inflicting further injuries. This is true for many reasons.
Justice is just one virtue among others. We should also aim to treat people humanely.
Sometimes that might require that we accept less justice, or forego it altogether. But it
might just require that we do justice differently, accepting civil means of redress when
they provide a suitable substitute for revenge.
Tort has another important advantage too: when a tort suit is successful, represen-
tatives of the community hold the wrongdoer responsible. That public vindication
may abate the threat that wrongdoing poses to its victim more effectively than any
action the victim might take on her own. As Murphy says, “our self-respect is social,”50
and because of that, victims have special reason to prize the communal affirmation
reflected in a tort judgment.
But we should be careful not to overestimate the difference between tort and revenge
on this point. Revenge is rarely just a private performance. Recall that the crowd
demanded that Mitchell take revenge. That provided some communal affirmation
even before he filed his tort suit. And that sort of thing was not uncommon. Represen-
tatives of the community often played a role in revenge, sometimes by taking sides in the
dispute, and sometimes by judging what it would take to get victim and wrongdoer even.
Indeed, I have argued elsewhere that we can view tort as the institutionalization of the
oddman, a character who was often available in revenge cultures to weigh in on what
would get the parties even when they could not reach an agreement on their own.51 So
when it comes to public vindication, the difference may be in degree, not kind.
But elsewhere, revenge has clear advantages over tort. A victim can take revenge on
her own, and if she needs help, she can get it from family or friends. She does not need
to invoke complicated legal proceedings or hire a lawyer to help her navigate them.
And there may be special value in vindicating one’s worth oneself (you would rather be

corrective justice or tort. Smith comes at the problem from a different angle, and I am not sure that I agree with
everything he says along the way. But I do think he is gesturing in the right direction, and I hope here to have
explained why.
50
Murphy, “Forgiveness and resentment”(note 33) at 25.
51
Hershovitz, “Corrective Justice for Civil Recourse Theorists” (note 26) at 124–5. For more on oddmen,
see Miller, Eye for an Eye (note 42) at 9–11.
 TORT AS A SUBSTITUTE FOR REVENGE 99

the kid who can handle things on her own than the one who has to appeal to the
teacher for help). But I think the most important advantage revenge has is that it can
send a wider array of messages than tort. Tort has limited tools—money damages, and
occasionally injunctions. Sometimes, that will be enough. But in cases that involve
serious wrongdoing, those remedies will not be able to send a very satisfying message.
Ordering a rapist to pay medical expenses, for example, or even punitive damages,
does not really get at the gravity of the wrong. Here you can see the beginnings of an
argument for the fearsome consequences that criminal law imposes. I will return to
those shortly. The point for now is that in the most extreme cases, revenge can deliver
more complete corrective justice than tort law can. At the limit, there is always death,
and ever more gruesome ways of dispensing it. And that makes the range of messages
available almost limitless.52
There is more to say, on behalf of both tort and revenge, but I will stop for now.
There is no chance that we will replace tort with a regime of revenge, and I do not
think that we should. But we should remember that revenge has virtues that tort
cannot match.

VI. Corrective Justice and Criminal Law

Though the relationship between tort and criminal law is too large a topic to take on
here, I want to say a few words about how criminal law fits into the picture. A standard
way of carving legal institutions assigns tort the aim of corrective justice and criminal
law the aim of retributive justice. This is much too neat. As I just suggested, when the
tools available to tort law are not up to the task, it is natural to look to the criminal law
to correct the historical significance of wrongdoing. Sometimes, this may be because
the tort remedies are in themselves unsatisfying. Other times, it may be because the
wrongdoer is judgment proof, and hence immune to the main tool tort uses to send its
messages. In both cases, criminal punishment might do what tort cannot: affirm the
victim’s standing by responding to the wrongdoing in a way that conveys its gravity.
Of course, this is not all criminal law does, and when it does this, it does it indirectly.
In a tort suit, a plaintiff files a complaint against a person she believes wronged her,
and the point of the process is to decide whether that plaintiff was wronged and what
should be done about it if she was. The person who would be a plaintiff in a tort suit is
typically a bit player in criminal proceedings. If she has a role at all, she is there to offer
evidence, not make a claim of her own. Victim impact statements and criminal
restitution orders blur this line a bit. But in the main, criminal proceedings are
about the defendant, and not the claims that a victim might have on the defendant.
There may be a distinctive sort of justice pursued through criminal law. Certainly,
many people think that criminal law deals in retributive justice, which they take to be

52
Almost, but not quite. History has given us more than a few monsters against whom no revenge would be
satisfying.
100 SCOTT HERSHOVITZ

concerned with the question how a wrongdoer deserves to be treated, quite apart from
the question what the wrongdoer might owe his victim. I am not sure that there is a
distinct kind of justice here, but if there is, I am quite certain that criminal proceedings
can do corrective justice too—and not just when restitution orders are tacked on.
Outside the courtroom, we find it natural to think that criminal punishment delivers
justice for the victims of a crime. That thought is hard to square with standard notions of
retributive justice, which is not normally taken to be personal in that way.53 But it is easy
to understand on a communicative picture of corrective justice. Punishing a defendant
for a wrong that had a victim helps abate the threat to the victim’s social standing by
affirming that she was mistreated and by marking the gravity of that mistreatment. By
the same token, declining to prosecute crimes that have victims risks leaving the threat
in place, and perhaps reinforcing it, especially where there are no other public proceed-
ings that mark the conduct as wrong. There are good reasons not to give victims the
same control over criminal proceedings that they have over civil proceedings, as there is
more at stake in a criminal trial than corrective justice.54 But victims are stakeholders in
the criminal process because corrective justice is among the things at stake.

VII. Conclusion
I want to close in a courtroom several hundred miles from the one in which Alcorn
spit on Mitchell, with a suit that was filed more than a hundred years later. The
plaintiff was an anti-smoking advocate named Ahron Leichtman. He was a guest on
Bill Cunningham’s radio show, invited to discuss the dangers of secondary smoke.
According to Leichtman’s complaint, another host at the station, Andy Furman,
repeatedly blew cigar smoke in Leichtman’s face during the interview. Cunningham
egged him on. Leichtman sued Cunningham, Furman, and the radio station for
battery. The trial court dismissed the claim, but the Ohio Court of Appeals revived it.55
The court’s opinion is short, and the lion’s share of the discussion addresses the
elements of battery. The court concludes that the particulate matter in smoke can

53
Nozick, Philosophical Explanations (note 18) at 367 (distinguishing revenge from retribution in part on
the ground that revenge is personal, whereas retribution is not).
54
For a guide to some of those reasons, see John Gardner, “Crime: In Proportion and in Perspective,” in
Offences and Defences: Selected Essays in the Philosophy of Criminal Law (Oxford: Oxford University Press,
2007), 213–38. I agree with much of what Gardner says, but I think he has too narrow a view of the relationship
between criminal law and revenge. He argues that one function of the criminal law is to displace revenge by
satisfying the retaliatory desires of victims, so far as consistent with its other aims. I do not disagree with that,
but Gardner seems to have something like the payoff picture in mind, as he shows little respect for those
retaliatory desires. He describes them as often excusable, but rarely justifiable. To the contrary, I think revenge
is often warranted, or at least it would be absent institutions like tort and criminal law that remove the warrant
by providing alternative means of doing justice. And providing alternative means of doing justice strikes me as
the more important function of the criminal law, at least insofar as revenge is concerned.
55
Leichtman v WLW Jacor Communications, Inc., 92 Ohio App. 3d 232, 634 N.E.2d 697, 698 (Ohio. Ct. of
App. 1994).
 TORT AS A SUBSTITUTE FOR REVENGE 101

make contact, and that blowing it in someone’s face is offensive. Thus, it holds that
Leichtman stated a valid battery claim and sends the case back to the trial court for
further proceedings. But not without comment. These passages appear near the end of
the opinion:
Arguably, trivial cases are responsible for an avalanche of lawsuits in the courts. They delay
cases that are important to individuals and corporations and that involve important social
issues. The result is justice denied to litigants and their counsel who must wait for their day
in court. . . .
This case emphasizes the need for some form of alternative dispute resolution operating
totally outside the court system as a means to provide an attentive ear to the parties and a
resolution of disputes in a nominal case. Some need a forum in which they can express
corrosive contempt for another without dragging their antagonist through the expense
inherent in a lawsuit.56

The contrast with the attitude of the Alcorn court could hardly be more stark. This too
was an act “done for the mere purpose of insult and indignity.”57 But the court does
not say that the law should afford liberal damages, or make an example out of Furman.
Instead, it calls the case “trivial” and “nominal,” and it suggests that Leichtman, not
Furman, needs a better outlet for his “corrosive contempt.”
There are, of course, reasons to think Furman’s wrong less significant than Alcorn’s.
I would rather have smoke blown in my face than suffer Alcorn’s spit, and I suspect
I am not alone in that. We also take insults less seriously nowadays. And, of course, the
place where this one happened matters too. A radio studio is not a temple of justice.
Finally, it is tempting to say that Leichtman knew what he was getting into when he
went on the show. These were well-known shock jocks, and Leichtman had mixed it
up with them before.58 We must tread carefully here, however. We are talking about a
battery, and “he was asking for it” is just one letter off from “she was asking for it.” But
even if you think (mistakenly) that going to a place where you know you might be
mistreated mitigates the mistreatment, at best all these differences add up to reason to
limit Leichtman’s damages, not reason to bar his claim from court.
Unless, that is, you do not think Furman’s treatment of Leichtman mattered. And
that is the unmistakable message of the court’s parting comment. There are important
social issues to resolve (some of them presented by corporations!), and the affront to
Leichtman’s dignity is not one of them. Of course, there is the half-hearted suggestion
that Leichtman should have some other forum—totally outside the court system—
in which to express his corrosive contempt for Furman and friends. But kicking
Leichtman’s claim to kiddie court and casting him as the person whose behavior is

56 57
Leichtman, 634 N.E.2d, at 700. Alcorn, 63 Ill., at 554.
58
The suit arose from Leichtman’s second appearance on the radio station. During an earlier interview,
Furman smoked a cigar. Afterward, Leichtman filed a complaint with the city health department, which cited
the station for violating an ordinance that banned smoking in private workplaces. “Cigar smoke can be used as
weapon, Ohio court rules,” The Tuscaloosa News (January 30, 1994), 2A.
102 SCOTT HERSHOVITZ

corrosive already downgrades his dignity. Sure, he was battered, but why bother
with that?
The court did not want to do corrective justice for Leichtman. It sent his case back
grudgingly, since it thought the Ohio Constitution required it. But in doing so, it
undermined the message that a verdict for Leichtman would send. I have some
sympathy for the court’s attitude. Leichtman was mistreated, but his suit may have
been motivated more by a desire to gain publicity than by the slight.59 (When the suit
settled, Leichtman got an hour of airtime to say as he pleased.)60 So Leichtman may
not have been a compelling candidate for corrective justice. But the court’s hostility
was not reserved for him; it took his case as representative of an avalanche of trivial
suits. And there is a lesson to learn from its screed. Tort is structured to do corrective
justice. But its capacity to do corrective justice is not just a function of its structure.
Justice is not just about who has what, and courts do not do justice just because they
move money around in response to wrongdoing. They do justice, when they do,
through the messages that they send about our relations to one another. So if we
care about corrective justice, we have to make sure that courts have the tools to send
the right messages, and that they care enough to use them.
This is all the more important in a society in which revenge is not a serious option.
Mitchell might well have killed Alcorn, or seriously injured him, but Leichtman was
not going to hurt Furman. We do not do that anymore, or at least most of us do not.
We are too civilized, or too timid. Or a bit of both. And it is not clear that it would
work anymore anyway. We are so steeped in anti-revenge discourse that we may not
hear the messages that revenge aims to send. But if revenge is not an option, then tort
better be, at least for any wrongdoing that we think worth taking seriously. Otherwise,
victims will not have revenge or an adequate substitute, and they will be left without
corrective justice.

59
For what it is worth, Leichtman did claim that he was injured. He said that the morning after the
interview, he had a corneal abrasion, which prevented him from wearing his contact lenses. See Junda Woo,
“Blowing Smoke Around Others May Be Battery,” The Wall Street Journal (April 11, 1994), B1.
60
George W. Jarecke and Nancy Plante, Seeking Civility: Common Courtesy and the Common Law (Boston:
Northeastern University Press, 2003), 20.
 5
Structure and Justification in
Contractualist Tort Theory
John Oberdiek*

I. Introduction
Much as contractualist theories of social justice and morality have been developed as
systematic alternatives to utilitarian and more broadly consequentialist theories of
those domains,1 contractualist theories of tort law have been advanced as alternatives
to consequentialist, and specifically economic, theories of tort law. Consequentialist
theories of any kind revolve around the aggregate goodness of impersonal states of
affairs, so that what anyone ought to do in any situation is to maximize whatever is
held to be good. Economic theories of tort law simply adopt consequentialism’s
maximizing theory of right acts and, in one prominent version,2 specify the maximand
as social wealth. On this view, one ought not act, and one is concomitantly liable for
any harm one proximately causes, just in case the expected overall economic benefit of
one’s conduct does not exceed its expected overall economic harm.3
The lodestar of contractualist theories, in contrast, is a form of mutual justification
modeled by hypothetical agreement. John Rawls’s account of social justice, “Justice as

* An earlier version of this chapter was presented as a paper at Yale Law School and I am grateful to the
participants at that session for an excellent discussion, including Leora Katz, Scott Shapiro, Sam Shpall, John
Witt, Gideon Yaffe, and especially Greg Keating, who also provided me with extensive written comments.
Thanks also to Rahul Kumar for his helpful comments.
1
See John Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971), 3 (“My guiding
aim is to work out a theory of justice that is a viable alternative to [intuitionism and utilitarianism] which have
long dominated our philosophical tradition”); T.M. Scanlon, “Contractualism and Utilitarianism,” in Amartya
Sen and Bernard Williams (eds.), Utilitarianism and Beyond (New York: Cambridge University Press, 1982),
128 (“I have put forward contractualism as an alternative to utilitiarianism . . . ”).
2
See, e.g., Richard Posner, “The Concept of Corrective Justice in Recent Theories of Tort Law,” 10 Journal
of Legal Studies 187 (1981), 206 (“[W]ealth maximization is the ultimate objective of the just state”).
3
Richard Posner, “A Theory of Negligence,” 1 Journal of Legal Studies 29 (1972), 33 (“Perhaps, then, the
dominant function of the fault system is to generate rules of liability that if followed will bring about, at least
approximately, the efficient—the cost-justified—level of accidents and safety”).
104 JOHN OBERDIEK

Fairness” is, of course, the paradigmatic contractualist theory, and it has inspired
theories of tort law in the same cast: thus, George Fletcher defends the “paradigm of
reciprocity” in “Fairness and Utility in Tort Theory,” according to which “the interests
of the individual require us to grant compensation whenever the disproportionate
distribution of risk injures someone subject to more than his fair share of risk,” and in
“A Social Contract Conception of the Tort Law of Accidents,” Gregory Keating
maintains that “[r]easonable care . . . is the level of care that fairly reconciles the
conflicting liberties of injurers and victims.”4 On this picture, debate in tort theory is
continuous with debate in moral and political philosophy—it is one more proxy war in
the ongoing global conflict between consequentialism and its critics, with both sides
fighting over which substantive moral standard justifies or should govern the law of
torts. Tort law, so conceived, just happens to be a suitable battlefield for this wider
justificatory war.
Waging this war is important. For it is important to determine morally defensible
rules and principles of conduct and liability. It is no exaggeration to say that if tort
law’s rules and principles are morally unfounded or otherwise mistaken, tort law’s
justifiability is cast into serious doubt. So, it is crucial to get tort law’s moral founda-
tions right. But if we are interested in developing and justifying a theory of tort law, it
also matters whether the moral foundations on offer are in fact those of tort law.
A justification of some practice only morally sustains the law of torts if the under-
written practice is, in certain fundamental ways, tort law. That is, a theory justifying
tort law is a theory justifying tort law if, but only if, the theory takes as its object salient
features of tort law’s substance and structure.
It has become commonplace to suppose that, whatever their justificatory merits,
economic theories fail to satisfy this interpretive adequacy condition. Whatever the
economic approach is approaching, the argument goes, it is not torts. In particular,
economic theories are supposed to have trouble making sense of the “correlative” or
“bilateral” plaintiff–defendant structure of tort suits.5 The apparent failure of eco-
nomic theories to adequately account for this structural feature of tort law, in turn,
supposedly counsels in favor of corrective justice theories of tort law, which in Jules
Coleman’s representative words hold that “individuals who are responsible for the
wrongful losses of others have a duty to repair the losses.”6 Corrective justice theories
claim to be able to explicate in a principled way why tort law treats particular plaintiffs
and defendants as a tandem, bound together: the defendant has allegedly wrongfully
harmed the plaintiff and so, if found liable, the defendant must right that wrong by
paying damages commensurate with the plaintiff ’s harm. Corrective justice theories
claim victory, then, not (or at least not merely) because theirs is the more compelling

4
George Fletcher, “Fairness and Utility in Tort Theory,” 85 Harv. L. Rev. 537 (1972), 550–1; Gregory
C. Keating, “A Social Contract Conception of the Tort Law of Accidents,” in Gerald J. Postema (ed.),
Philosophy and the Law of Torts (New York: Cambridge University Press, 2001), 43.
5
See Jules Coleman, “The Structure of Tort Law,” 97 Yale L.J. 1233 (1988); Ernest Weinrib, The Idea of
Private Law (Cambridge, MA: Harvard University Press, 1995).
6
Jules Coleman, The Practice of Principle (New York: Oxford University Press, 2001), 15.
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 105

justification as such, but because what they justify is recognizable as tort law while
what economic theories justify is not.
Yet there are two independent problems with this common narrative. First, the
claim that economic theories cannot provide an account of the bilateral structure of
tort suits is arguably question-begging. Economic theories have an explanation of that
structure.7 It is of course true that economic accounts explain tort law’s bilateralism
differently than corrective justice theories do; namely, as a merely contingent, if
entrenched, feature of tort law (more on this below). But that is simply due to the
unabashed instrumentalism of the economic approach’s consequentialism.8 The struc-
tural critique of economic theories of tort law is intended to be autonomous, however,
prescinding from debate about the economic approach’s inherently instrumental
consequentialism.
Second, whether or not economic theories can explain tort law’s bilateralism, it is
just a mistake to conceive of corrective justice theories of tort law as comprehensive
alternatives to economic theories. Any economic theory provides a substantive
account of tort law’s primary obligations, but no corrective justice theory even
attempts to offer such an account. As Coleman again puts it, “while corrective justice
presupposes some account of what the relevant first-order duties are, it does not
pretend to provide an account of them.”9 Instead, corrective justice theories only
explain the basis of tort law’s secondary, remedial obligations that are incurred by the
violation of wholly unspecified antecedent primary obligations—remedies are trig-
gered by wrongful harms or their prospect. A viable alternative to an economic
approach to torts will make sense of tort law’s bilateralism, to be sure, but it will
also support a comprehensive account of tort law’s primary obligations (and indeed
accord those primary obligations conceptual priority).10 Corrective justice theories do
not do this and are therefore too incomplete to rival economic theories.
This returns the discussion to contractualism, for I maintain that a contractualist
theory of tort law can meet these desiderata. There is a contractualist theory that can
provide a comprehensive account and justification of the primary and secondary
obligations of tort law and also make sense of the bilateral structure of tort suits.
What’s more, a contractualist account does not bifurcate these inquiries, but treats
them as interconnected: tort law’s primary obligations are a function of tort law’s

7
See, e.g., Mark Geistfeld, “Economics, Moral Philosophy, and the Positive Analysis of Tort Law,” in
Gerald J. Postema (ed.), Philosophy and the Law of Torts (New York: Cambridge University Press, 2001),
250, 252–9.
8
John Gardner makes this point, though not exactly in these terms, in “Backward and Forward with Tort
Law,” in Joseph Keim Campbell, Michael O’Rourke, and David Shier (eds.), Law and Social Justice (Cambridge,
MA: MIT Press, 2005), 255, 274–8. For the view that economic accounts cannot provide the right kind of
explanation of tort law’s bilateral structure, see, for example, Benjamin Zipursky, “Pragmatic Conceptualism,”
6 Legal Theory 457 (2000), 459.
9
Coleman, The Practice of Principle see (note 6) at 32.
10
See Hanoch Sheinman, “Tort Law and Corrective Justice,” 22 Law and Philosophy 21 (2003).
106 JOHN OBERDIEK

bilateral structure. It is my aim, here, to explore certain fundamentals of a contrac-

tualist account of tort law.
As the introductory paragraphs make clear, I am hardly the first to defend a
contractualist theory of tort law. In Section II, I provide a brief overview of Keating’s
account of accident law, which I take to be the leading extant contractualist tort theory,
in order to introduce basic elements of a contractualist approach to torts. In
Section III, I examine whether Keating’s social contract theory is susceptible to the
charge, typically pressed against economic accounts, that it cannot motivate tort law’s
bilateralism. Keating’s theory can appear susceptible to this criticism because the
theory revolves around distributive justice, and it is difficult to see how distributive
justice can accommodate the bilateral plaintiff-defendant relationship central to tort
law. In this, Keating’s account appears to share the economic approach’s indifference
to tort law’s status as private law. I argue that at least one version of this criticism is
misplaced, but that questions remain and that switching from the Rawlsian social
contract framework that Keating favors to the moral contractualism developed by
T.M. Scanlon permits one to retain what is attractive in Keating’s account while
providing a more perspicacious account of tort law’s bilateral structure. For Scanlon’s
is essentially a theory of interpersonal justification. More specifically, I argue that the
form of justification built into the bricks of Scanlonian contractualism illuminates the
notion of relational duty that is at the heart of tort law’s bilateralism and grounds tort
law’s status as private law. In Section IV, I introduce that form of justification and
highlight its role within Scanlonian contractualism, and in Section V, I catalog its
virtues as a basis for a theory of tort law. It seems to me that much of what unites the
dominant band of broadly non-consequentialist/non-economic tort theories, whether
under the banner of “corrective justice” or “civil recourse”—I call what unites these
views “the Palsgraf perspective,” borrowing Benjamin Zipursky’s phrase—can be best
explained and justified when founded on Scanlonian contractualism. My goal, then, is
at once modest and ambitious: it is modest in that I simply aim to introduce a different
contractualism than is current within tort theory, but it is ambitious in that I hope to
present that brand of contractualism as an appealing and novel comprehensive
foundation for the law of torts.

II. Introducing Contractualist Tort Theory:

Keating’s Social Contract Theory
Keating develops an account of the law of accidents owing to Rawls and, through him,
to Kant. Its touchstones are freedom and fairness.11 On Keating’s view, the law of

11
Keating, “A Social Contract Conception of the Tort Law of Accidents” (note 4) at 22. My exposition
of Keating’s view centers on “A Social Contract Conception of the Law of Accidents.” See also Gregory
C. Keating, “Reasonableness and Rationality in Negligence Theory,” 48 Stan. L. Rev. 311 (1996); Gregory
C. Keating, “Rawlsian Fairness and Regime Choice in the Law of Accidents,” 72 Fordham L. Rev. 1857
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 107

accidents governs risk impositions, which are “the by-product of beneficial activities”12
characteristic of modern life but which “pit the liberty of injurers against the security
of victims.”13 Faced with these “competing freedoms,”14 which are both “precondi-
tions of effective rational agency,”15 it is the point of accident law to “fix terms that are
fair.”16
Following Kant’s and Rawls’s lead, Keating locates a normative conception of the
person as free and equal at the center of this theory.17 And following Rawls’s later
political turn, Keating interprets the normative conception of the person as having a
political cast: persons are conceived of as free and equal democratic citizens with
irreducibly diverse aims and values. For Keating, this underscores that “the problem of
accidental harm is a problem of social choice, a matter of reconciling the competing
claims of liberty and security for a plurality of persons.”18 In addition to being free and
equal, persons so conceived are also understood to be both rational and reasonable,
and, thus, we must ascribe to them prudence as well as a sense of justice that accords
due regard to the interests of others, respectively.19 Relying on this distinction, one can
recognize that, while it might be rational to take certain risks, it would be unreasonable
to impose those same risks on others. Fair terms of interaction, and specifically of risk
imposition, then, will be those that such persons “might reasonably accept so long as
others do so as well.”20 More specifically, “[i]t is reasonable to expose other people to
risks of serious injury and even death when it is fair to do so; and it is fair to do so,
when they, too, stand to gain from the imposition of those risks.”21 Thus, on Keating’s
view, reasonable and so fair risk impositions, which reconcile liberty and security
across persons, will be those that are objectively mutually advantageous: it is fair to
subject others to mortal risk “when doing so is in the long run expected advantage of
the person imperiled.”22
Keating contrasts his Rawlsian tort theory with Fletcher’s, which favors a regime of
negligence liability for cases of symmetrical risk imposition. If people are subjecting
one another to roughly the same risks and taking care in doing so, then on Fletcher’s
view, any harms that eventuate ought to be borne by those who suffer them—the

(2004). Keating’s most recent work addresses certain of the concerns I raise in what follows, though not
explicitly from within a Rawlsian social contract framework. See Gregory C. Keating, “The Priority of
Respect Over Repair,” 18 Legal Theory 293 (2012). A fuller discussion of that work falls outside the
purview of the present discussion.
12
Keating, “A Social Contract Conception of the Tort Law of Accidents” (note 4) at 27.
13
Keating, (note 4) at 23.
14
Keating, (note 4) at 23.
15
Keating, (note 4) at 28.
16
Keating, (note 4) at 23.
17
Keating, (note 4) at 25.
18
Keating, (note 4) at 28.
19
Keating, (note 4) at 27.
20
Keating, (note 4) at 26.
21
Keating, (note 4) at 29.
22
Keating, (note 4) at 30.
108 JOHN OBERDIEK

injurer owes no compensation. For, given symmetrical risk imposition, no one is

taking advantage of anyone else and so the interaction between persons is fair. Only
if one fails to take care against a background of symmetrical risk imposition or
imposes asymmetrical risks outright, should one be liable for the harm that one causes,
on Fletcher’s view, because in such a case risks are no longer reciprocally imposed.
With due care, however, harms consequent to reciprocal, symmetrical risk impositions
lie where they fall.
Keating believes that this is a mistake because reciprocity of risk does not corres-
pond closely enough to reciprocity of harm, and it is the harm that can eventuate from
risk impositions that matters, not the risk impositions themselves. This is because risk
as such seldom “impairs the ability to pursue a conception of the good over the course
of a complete life; it is harm—physical injury and death—that wreaks havoc with
people’s lives.”23 More attention needs to be paid to the ex post problem of harm and
less to the ex ante problem of risk, according to Keating, because it is the physical
integrity of the person that grounds the interests that a given liability regime must
honor.24 In this respect, security, understood as freedom from physical harm, is a more
important freedom than liberty, understood as the freedom to engage in risky conduct.
This leads Keating, ultimately, to favor a more expansive regime of strict liability than
Fletcher accepts or that tort law now recognizes. Though a strict liability standard
would, of course, hem in the liberty of injurers far more than would a negligence
standard—those involved in risky activities would have to be exceptionally careful in
order to avoid liability—Keating contends that a negligence standard would thwart the
security of victims far more. Diminished liberty of action, in short, is not as bad as
serious injury or death. And thus, in the pairwise comparison of burdens that guides
contractualist evaluation, strict liability is actually less burdensome and so a fairer
liability regime for mortal risk cases.

III. Rawlsian Distributive Justice and the

Bilateral Structure of Tort Law
Though Fletcher and Keating disagree fundamentally about liability regimes, they
appeal to the same normative framework to reach their disparate conclusions. The
animating ideal of both Fletcher’s and Keating’s theories is distributive justice, as befits
their shared foundation in Rawlsian political philosophy. Where Fletcher argues that
one should not be subject to more than one’s fair share of risk, Keating maintains that
one should not have to endure more than one’s fair share of harm. Importantly, in
both cases, what matters is the distribution of risk or harm, respectively, across all of
society’s members. An individual’s share of risk or harm is, in that sense, a function of

23
Keating, “A Social Contract Conception of the Tort Law of Accidents” (note 4) at 35.
24
Keating, (note 4) at 34.
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 109

a wider social distribution. Keating underscores this in describing the problem of

accidents as a “problem of social choice.”25 Framing the problem this way is natural,
given the provenance of the approach in Rawlsian political philosophy, for it was
Rawls’s aim to identify the principles of justice that should govern institutions
constituting the so-called basic structure of a well-ordered society.26 Rawls’s own
focus was explicitly social. Fletcher and Keating repurpose that framework, and
focus on the institution of tort law and its attendant liability regimes to solve a
particular social problem as well.
But accepting certain fundamentals of Rawls’s framework might seem to pose a
challenge to any theory of tort law. Focusing on Keating’s theory, the challenge is that,
as a problem of social choice, the distribution of harm entailed by a given liability
regime must be justified to society at large, not most fundamentally to the parties to a
tort suit. This assumes, but does not motivate, tort law’s bilateral structure—indeed, it
leaves the bilateral structure of tort law mysterious. Gerald Postema puts the challenge
this way:
Distributive justice takes the perspective of the community as a whole. But, then, members
of this community other than injurers and victims might also have stakes in the norms
adopted for regulating risky conduct, since they or the community at large stand to benefit
from and to bear some of the costs of conduct (or reductions in the level of certain
activities) in accord with the rules of liability adopted . . . Nothing in the theoretical frame
that Keating proposes requires that we pay special attention to specific injurers and
victims.27

Postema’s criticism is not implausible. Distributive justice, as Rawls understands it,

concerns society as a whole. While transactions between particular individuals might
affect the distributions that matter from the perspective of distributive justice, indi-
vidual conduct itself is not a matter of Rawlsian distributive justice. On Rawls’s view,
norms of distributive justice do not govern individual conduct directly, whether
prospectively or retrospectively, but instead govern the institutions that constitute
the basic structure of society. This aspect of Rawls’s position has attracted well-known
criticism from all sides of the political spectrum, which has in turn elicited equally
well-known replies on Rawls’s behalf.28 The merit of Rawls’s position on this score is

25
Keating, “A Social Contract Conception of the Tort Law of Accidents” (note 4) at 28.
26
See John Rawls, Political Liberalism, Expanded Edition (New York: Columbia University Press, 2005),
Lecture VII.
27
See Gerald J. Postema, “Introduction: Search for an Explanatory Theory of Torts,” in Gerald J. Postema
(ed.), Philosophy and the Law of Torts (New York: Cambridge University Press, 2001), 7; see also Stephen
R. Perry, “The Moral Foundations of Tort Law,” 77 Iowa L. Rev. 449 (1992), 471.
28
The most prominent criticism of this aspect of Rawls’s view from the left is G.A. Cohen, “Where the
Action Is: On the Site of Distributive Justice,” 26 Philosophy & Public Affairs 3 (1997), and more generally,
G.A. Cohen, Rescuing Justice and Equality (Cambridge, MA: Harvard University Press, 2008). For a defense of
Rawls, see Thomas Pogge, “On the Site of Distributive Justice: Reflections on Cohen and Murphy,” 29
Philosophy & Public Affairs 137 (2000). The most famous criticism of Rawls from the right is Robert Nozick,
Anarchy, State, and Utopia (New York: Basic Books, 1974). Thomas Nagel criticizes Nozick and defends a
broadly Rawlsian position in “Libertarianism Without Foundations,” 85 Yale L.J. 136 (1975).
110 JOHN OBERDIEK

irrelevant in the present context. More important is what these exchanges underscore:
a seemingly inconvenient fact for Keating, namely, that Rawlsian distributive justice
takes as its object society as a whole. This raises the concern that, from the theoretical
perspective that Keating adopts, the relationship between a specific plaintiff and a
specific defendant at the center of any tort suit is essentially invisible. Given its prima
facie plausibility and significance, it is worth developing Postema’s criticism to deter-
mine whether it is ultimately sound.
Economic theories of tort law are considered susceptible to structural criticism
because they simply assume the bilateral relationship between plaintiff and defendant
in their explanations and justifications of tort law, without motivating that structural
feature. It might appear that Keating’s theory is open to a particularly sharp version of
the same objection. His theory seems to be on all fours with economic accounts in just
assuming the bilateral structure of a tort suit. Where unabashedly instrumental
economic accounts have at least the resources to mount a response to the structural
criticism, whether ultimately successful or not, however, the basis of Keating’s theory
in Rawlsian distributive justice would seem to prevent it from being able to respond to
this redirected structural critique at all from within its own terms.
If an economic account attempts to make sense of the bilateral structure of tort law by
maintaining that it is an historical artifact that is suboptimal but would be too costly to
alter, that response at least makes sense within the theory’s terms.29 Efficiency is the
master value of any economic account, and it is efficiency that explains both why the
bilateral structure of tort law would not be replicated by social planners starting from
scratch as well as why, because we are not starting from scratch, it is most efficient
simply to preserve the existing fundamentals of tort law, including its admittedly
inefficient bilateral structure. In other words, economic accounts can conclude that it
would be inefficient, all-things-considered, to correct certain local inefficiencies, thus
leaving in place those local inefficiencies on grounds of efficiency. There is no principled
reason why tort law has a bilateral structure according to this view; but again, economic
theories are not in the business of justifying legal institutions that way.
The charge is that a tort theory founded on Rawlsian distributive justice, like
Keating’s, cannot similarly cope with the structural critique. According to this criti-
cism, Keating cannot make moves analogous to the economist and argue that,
although the focus on an individual plaintiff and defendant has no principled basis
in Rawlsian distributive justice, a tort theory founded on Rawlsian distributive justice
can nevertheless justify the focus on an individual plaintiff and defendant on grounds
of Rawlsian distributive justice. It makes no sense within a Rawlsian framework even
to speak of local distributive injustices that call for redress because there can be no
local distributive justice or injustice according to Rawls.30 This is precisely the feature
of Rawls’s conception of distributive justice that, for different reasons, draws Robert

29
See Geistfeld, “Economics, Moral Philosophy, and the Positive Analysis of Tort Law” (note 7).
30
See Stephen R. Perry, “On the Relationship Between Distributive and Corrective Justice,” in Jeremy
Horder (ed.), Oxford Essays in Jurisprudence, 4th ed. (Oxford: Oxford University Press, 2000), 237.
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 111

Nozick’s fire from the right and G.A. Cohen’s fire from the left. Thus, the criticism
concludes, Keating’s theory would appear to be especially vulnerable to the charge,
more commonly leveled against economic accounts of tort law, that it neither motiv-
ates nor makes sense of the fundamental bilateral structure of tort law. For local
distributive justice has no place within Keating’s Rawlsian framework.31
It is in presenting the problems of tort law as social problems that call for a solution
at the level of social choice that Keating’s theory invites trouble. Focusing on social
choice appears to align Keating’s theory with the economic theories that it rivals in a
crucial respect: if tort suits are nothing more than opportunities to make good public
policy, then it would not be surprising if both Keating’s theory and economic theories
failed to honor bilateralism. Any broad public policy-based approach, whether keyed
to Rawlsian distributive justice or efficiency, overlooks the essential fact that tort suits
are fundamentally private matters whose resolution (ideally) depends not upon the
multifarious considerations that must inform matters of social choice and public
policy, but upon those considerations that matter, first and foremost, to the parties
to the lawsuit.
This limit on the considerations that may be brought to bear in standing on one’s
tort-based rights is, in part, what makes tort law a branch of private law. Of course
public law values have probably always had some influence on and otherwise pene-
trated tort law, as well as private law’s other domains. But one need not endorse a strict
demarcation of private law from public law to defend the position that private law
employs a distinctive form of justification, which is distinct from the form of justifi-
cation appropriate to public law. Private law, after all, concerns the discrete, interper-
sonal dealings of specific (natural and corporate) persons, while public law concerns
overarching policy problems. Where private law is the realm of individual claims,
public law is the realm of regulation. Where private law asks, in a paradigm case, what,
if anything, the Long Island Railroad owes to Mrs. Palsgraf as a matter of right for her
injuries, public law aims to address the cost of accidents in general.
Yet Keating’s Rawlsian tort theory is clearly animated by an interpersonal concern
with fairness, not a public policy-based impersonal concern with bringing about some
favored state of affairs. The Rawlsian framework that Keating adopts calls upon
individuals to relate to each other as fellow participants in a cooperative scheme that
is fair. Each is to relate to all others as individuals in a way that respects their standing
as persons who are free and equal. The basic structure’s demand that one regulate
one’s conduct in a way that is fair is a duty that one owes to each of one’s fellow
participants in the fair cooperative scheme, not to the scheme itself. Violations of the
norms of the scheme thus constitute wrongs to one’s fellow participants, not an offense
to some impersonal goal. Keating cannot therefore be charged with simply replacing
the impersonal goal of efficiency with the impersonal goal of distributive justice. His is
not an impersonal goal-based tort theory.

31
I should emphasize that I argue that local distributive justice does not make sense within a Rawlsian
framework; I remain agnostic about local distributive justice as such.
112 JOHN OBERDIEK

Still, the concern remains that Keating’s Rawlsian approach is incompatible with
tort law’s bilateral structure, even if it is thoroughly interpersonal. The remaining
problem, and another way of posing Postema’s challenge, is that a failure to operate
within the terms of the fair cooperative scheme turns out to be a case not merely of
wronging the one with whom one has interacted in a way that the basic structure
forbids. Rather, in doing so, one wrongs everyone to whom the obligation to comply
with those norms is owed. It would seem that the remedial duties one thereby incurs
are thus owed not simply to the most obvious victim or victims of the wrongdoing, but
to everyone. For the relevant duty is to ensure that the benefits and burdens attending
the basic structure are fairly distributed. Now, so understood, the obligation to
cooperate on terms that are fair to all is owed not to everyone impersonally, but
interpersonally: violating those terms is not interpersonally justifiable. But it remains
to be seen whether the interpersonality at the heart of Keating’s theory is sufficiently
personal, as it were. This is because, on Keating’s account, if one is victimized, one’s
complaint must be based not on one’s own mistreatment as such but on the
mistreatment of all the participants in the (otherwise) fair cooperative scheme.
One who has been victimized can appeal to the fact that another has failed to comply
with the fair terms of cooperation governing everyone’s mutual relationships to
show that one has been wronged, but not to show that one has been wronged qua
someone standing in a unique relationship to whomever has wronged one. There is,
to be sure, a kind of bilateralism here. It is, however, derivative, for there is
bilateralism only because there is omnilateralism. This is hardly a conclusive criti-
cism. But, it seems to me, this fact counsels against embracing Keating’s theory in all
of its particulars.
Still, Keating’s approach has its attractions. I would submit that Rawlsian social
contract theory owes what appeal it has as a basis for tort law to its comprehensive
prescriptivity—distinguishing it from exclusively remedial corrective justice theories
of torts—and its systematic non-consequentialism—distinguishing it from (prescrip-
tive) economic theories of torts. These are, indeed, signal virtues and they should not
be foregone. Nor need they be. There is a different version of contractualism than the
one that Rawls develops and Keating adopts that, nevertheless, shares its comprehen-
sive prescriptivity and systematic non-consequentialism, but which better harmonizes
with tort law’s bilateralism.
Scanlon’s version of contractualism, while widely considered to be the most
promising systematic alternative to consequentialist moral theory, has been largely
overlooked as a foundation for tort law. One feature of Scanlonian contractualism in
particular makes it especially promising as a foundation for torts, and that is what
I highlight in what follows: the conception of justification at the theory’s core. It is
that conception of justification that makes sense of the bilateral structure of tort law
in a principled way and illuminates how the structure and content of tort law are
mutually reinforcing. Scanlonian contractualism thus holds the promise of provid-
ing a foundation for tort law that is at once interpretively plausible and morally
compelling.
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 113

IV. A Primer on Justification in Scanlonian

Contractualism
According to Scanlon’s conception of contractualism, “an act is wrong if its perform-
ance under the circumstances would be disallowed by any set of principles for the
general regulation of behavior that no one could reasonably reject as a basis for
informed, unforced general agreement.”32 While this formula’s concision belies its
complexity along both normative and metaethical dimensions, here I wish simply to
elucidate its contractual form. For the contractualist form of the formula highlights the
central role that a particular model of justification plays in the theory.
The contractualist formula accords pride of place to hypothetical agreement, as that
which no one can reasonably reject forms the basis of a reasonable agreement.33 There
are two reasons why the agreement sought is hypothetical, not actual: impracticality
and under-normativity. If actual unanimous consensus about principles of conduct
were required, but there were some who simply refused to participate in any consen-
sus-building, then moral justification would become a hostage to fortune. Requiring
actual agreement would therefore be impractical. And yet, even if everyone were to
participate, but some people were sufficiently out of touch with their interests, then the
content of whatever actual consensus was reached would lack moral force—it is the true
interests of each person and not what each person (merely) believes his or her interests
to be that matter. Actual agreement would therefore also be insufficiently normative.
Hypothetical agreement lies at the core of the contractualist formula, then, because it
does justice to both the practical and the probative aspirations of moral reasoning.
More fundamentally, however, hypothetical agreement occupies the central place it
does in Scanlonian contractualism because norms that are objects of hypothetical
agreement are valid.34 This in turn is because hypothetical agreement expresses the
value of justifiability: “Thinking about right and wrong is, at the most basic level,
thinking about what could be justified to others on grounds that they, if appropriately
motivated, could not reasonably reject.”35 In this gloss on the contractualist formula,
Scanlon omits reference to agreement and instead casts the principle’s lodestar in
terms of justifiability. As consequentialism also helps itself to the concept of justifiability,

32
T.M. Scanlon, What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998), 153.
33
Scanlon maintains that “what is fundamental to morality is the desire for reasonable agreement.”
T.M. Scanlon, “Contractualism and Utilitarianism,” in Amartya Sen and Bernard Williams (eds.), Utilitarian-
ism and Beyond (Cambridge: Cambridge University Press, 1982), 115 n. 10. On the equivalence of non-
rejectability and unanimous acceptability in Scanlon’s contractualism, see Brian Barry, Justice as Impartiality
(New York: Oxford University Press, 1996), 70.
34
I leave this intentionally ambiguous because I wish to remain agnostic about the metaethics of
contractualism.
35
Scanlon, What We Owe to Each Other (note 32) at 5. Scanlon takes justifiability to be basic in two ways: it
provides both the normative basis of the morality of right and wrong, and a general characterization of its
content. See Scanlon at 189.
114 JOHN OBERDIEK

however, it cannot be justifiability simpliciter that distinguishes contractualism, but

some more specific conception of it.36
The best way to make sense of contractualism’s appeal to justifiability, I believe, is to
understand it as appealing to a particular form of justification that we can follow
Thomas Nagel in calling “justification to a subject.”37 Justification to a subject, which
I explore below, animates Scanlonian contractualism. When one is able to justify one’s
actions to others according to justification to a subject, one stands in what Scanlon
calls “a relation of mutual recognition,” the value of which “underlies our reasons to do
what morality requires.”38 As Scanlon puts it, contractualism “locates the source of the
reason-giving force of judgments of right and wrong in the importance of standing in a
certain relation to others”—a relation of mutual recognition.39 On this reading, the
idea of justification to a subject simply is the idea of a relation of mutual recognition,
and the contractualist formula employs the evocative language of hypothetical agree-
ment in order to express this abstract master value.

V. Justification to a Subject and “the

Palsgraf Perspective”
Nagel contrasts justification “to a subject as a subject” (or simply “to a subject” for
short) with justification “to the world at large.”40 Justification to the world at large is
fundamentally impersonal. When one goes about assessing whether some conduct
is permissible according to this conception of justification, one considers whether it is
acceptable not to anyone in particular, but whether it promotes some valuable state of
affairs—it must be acceptable to “the world at large.” Justification between individuals
under this form of justification is indirect, as it is mediated by some valuable imper-
sonal state of affairs to which justifications between people must refer. Individuals

36
For example, writing from a consequentialist perspective, Shelly Kagan contends that the Kantian
argument from respect “tries to claim that a certain kind of act’s being disrespectful is the basis of its being
unjustified. But I believe that the reverse is closer to the truth . . . ” See Shelly Kagan, The Limits of Morality
(New York: Oxford University Press, 1989), 176.
37
Thomas Nagel, “War and Massacre,” in Mortal Questions (New York: Cambridge University Press,
1979), 66–8. Nagel coined this term before Scanlon first developed his brand of contractualism. My claim is
only that the concept appropriately captures the form of justification at contractualism’s core. Still, it is worth
noting that Nagel has signed on to certain fundamentals of contractualism. See Thomas Nagel, Equality and
Partiality (New York: Oxford University Press, 1991).
38
Scanlon, What We Owe to Each Other (note 32) at 162.
39
Scanlon (note 32) at 177–8. See also Christine Korsgaard, who maintains, “[t]he subject matter of
morality is . . . how we should relate to one another.” Christine M. Korsgaard, “The Reasons We Can Share:
An Attack on the Distinction Between Agent-Relative and Agent-Neutral Values,” in Creating the Kingdom of
Ends (New York: Cambridge University Press, 1996), 275; and Rahul Kumar, who maintains that morality
“provides a framework in which individuals can relate to one another and form valuable relationships with one
another, on a basis of mutual respect for one another’s status as a person.” Rahul Kumar, Consensualism in
Principle: On the Foundations of Non-Consequentialist Moral Reasoning (New York: Routledge, 2001), 81.
40
Nagel, “War and Massacre” (note 37).
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 115

have claims on this view only in an attenuated and denuded sense, for to the extent
people have claims, those claims are entirely derivative. This is because individuals are
not true claimants according to justification to the world at large—individuals them-
selves are not the sources of claims. They are instead bestowed with whatever “claims”
they have if and only if doing so can be justified to the world at large.
In contrast, under justification to a subject, persons are not derivatively morally
relevant but are instead the sources of claims. Anyone’s conduct must, therefore, be
justified directly to other individuals. As Nagel puts it, justification to a subject requires
that what we do to others “must be aimed at him as a subject, with the intention that he
receive it as a subject. It should manifest an attitude to him rather than just to the
situation . . . ”41 Under justification to a subject, concern is shown for persons because
they are themselves owed justification. Comparing the two conceptions, we can say
that under justification to a subject, behavior is justified to the world at large in virtue
of being (directly) justified to every individual as an individual, whereas under
justification to the world at large, behavior is (indirectly) justified to individuals in
virtue of being justified to the world at large.
Even with this much, one can see how Scanlonian contractualism, with justification
to a subject built into its bricks, grounds a promising, and also familiar, approach to
tort law. The approach to tort law that Scanlonian contractualism grounds is familiar,
especially in the work of Ernest Weinrib and the duo of John Goldberg and Benjamin
Zipursky, and we can follow Zipursky in calling it “the Palsgraf perspective.”42 That
perspective is the conjunction of two linked theses: tort law is private law, wherein
litigants press or defend their tort-based rights within a bilateral structure, and its
right-based duties are relational, being owed to particular parties. The perspective of
course owes its name to and finds clear application in Palsgraf itself, where these two
commitments are on display in Judge Cardozo’s majority opinion that Mrs. Palsgraf
ought not to recover against the Long Island Railroad (LIRR) for her injuries.43 While
Cardozo concedes that the LIRR injured Mrs. Palsgraf, the LIRR owed Mrs. Palsgraf a
duty of care, and the LIRR was in breach of its duty of care as to someone, he determines
that Mrs. Palsgraf nevertheless must lose because the LIRR’s injury-causing breach was
not of the duty owed to Mrs. Palsgraf.44 The harm that Mrs. Palsgraf suffered was not
reasonably foreseeable to the admittedly careless LIRR. Rather, the LIRR was careless

41
Nagel, “War and Massacre” (note 37) at 66. The emphasis on intention and on mental states generally is
an inessential feature of Nagel’s view, and in any case, it is an inessential feature of the view that I endorse.
42
Benjamin C. Zipursky, “Palsgraf, Punitive Damages, and Preemption,” 125 Harv. L. Rev. 1757 (2012),
1792. Zipursky has developed this position, although not under this label, in numerous places, often with John
Goldberg. See, e.g., John C.P. Goldberg and Benjamin C. Zipursky, “The Moral of MacPherson,” 146 U. Pa.
L. Rev. 1733 (1998); Benjamin C. Zipursky, “Rights, Wrongs, and Recourse in the Law of Torts,” 51 Vand.
L. Rev. 1 (1998); and John C.P. Goldberg and Benjamin C. Zipursky, “Torts as Wrongs,” 88 Tex. L. Rev. 917
(2010). Ernest Weinrib defends the tenets constituting the Palsgraf perspective in Weinrib, The Idea of Private
Law (note 5). The discussion that follows draws from text and develops ideas explored in John Oberdiek,
“Method and Morality in the New Private Law of Torts,” 125 Harv. L. Rev. F. 189 (2012).
43
Palsgraf v Long Island Railroad Co., 248 N.Y. 339, 162 N.E. 99 (1928).
44
Zipursky, “Palsgraf, Punitive Damages, and Preemption” (note 42) at 1763–9.
116 JOHN OBERDIEK

only as to those whose possible injuries were a reasonably foreseeable consequence of

its conduct.
According to the Palsgraf perspective, one cannot stand in the shoes of another
when bringing a negligence claim. As Cardozo famously maintains, “[t]he victim does
not sue derivatively . . . to vindicate an interest invaded in the person of another.”45
According to Cardozo and the Palsgraf perspective, “[w]hat the plaintiff must show is
a ‘wrong’ to herself, i.e., the violation of her own right, and not merely a wrong to
someone else . . . ”46 Duty is thus relational on this view—duties are always owed to
someone in particular. Duty is relational, moreover, because of tort law’s status as
private law. Cardozo is clear that individuals lack plenary authority to enforce tort-
based rights generally, as a legislature might; people can stand only on their own
rights.47 This is what makes tort law private law, according to the Palsgraf perspective,
and it stands in stark contrast to what Zipursky calls “the private attorney-general
model” of torts, where individuals are empowered to pursue sound public policy as a
legislature does.48
These two ways of conceiving of tort law reflect the two quite different patterns of
relations between people realized by the distinct conceptions of justification canvassed
above. This is not, I think, mere coincidence. Any theory of tort law that has
prescriptive aspirations will be founded on some conception of justification and will
thus commit, at the deepest of levels, to a particular understanding of how conduct is
judged to be tortious. Indeed, much of what distinguishes different prescriptive tort
theories—as well as normative ethical theories—revolves around the distinct forms of
justification on which the theories are based. I maintain that the Palsgraf perspective is
a manifestation of contractualist justification to a subject and that that form of
justification has significant virtues that make it attractive as a foundation for a theory
of tort law.
One clear virtue is that the Palsgraf perspective, so founded, has the resources to
account for primary obligations and not merely secondary remedial ones. For justification
to a subject has as much normative traction prospectively as it has retrospectively—it
provides the resources to define primary, and not merely secondary, obligations. This
dramatically enhances the power of the Palsgraf perspective. Recall that corrective
justice theories of tort cannot be comprehensive alternatives to economic theories
because they do not define or generate, but presuppose, the primary obligations to
which its secondary remedial obligations answer. On this front, at least, Goldberg and
Zipursky’s civil recourse theory is no different than corrective justice theories: all focus
exclusively on the past, as they revolve around the assignment of responsibility for
redressing tortious harms. Scanlonian contractualism, on the other hand, earns its

45
Palsgraf, 162 N.E. at 101, quoted in Zipursky, “Palsgraf, Punitive Damages, and Preemption” (note 42)
at 1768.
46
Palsgraf, 162 N.E. at 100.
47
Zipursky calls this the “substantive standing” requirement. Zipursky, “Palsgraf, Punitive Damages, and
Preemption” (note 42) at 1769.
48
Zipursky, “Palsgraf, Punitive Damages, and Preemption” (note 42) at 1758.
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 117

stripes as a moral theory—and thus as a foundation for a comprehensive prescriptive

tort theory—by delineating primary obligations and permissions that are prospectively
applicable.49 My aim here is not to articulate and defend the content of contrac-
tualism’s primary obligations, but simply to draw attention to the uncontroversial
fact that contractualism does indeed provide an account of primary wrongs. When
harnessed to justification to a subject, the Palsgraf perspective need not rest content
with redressing antecedently defined wrongs, despite how its principal architects
understand the perspective. Given a contractualist grounding, the Palsgraf perspec-
tive is more robust, enjoying greater explanatory and prescriptive power than
it possesses without such a foundation, and promises to offer a comprehensive
tort theory.
Further virtues of this approach are brought into relief by considering the pattern of
relations between people instantiated by an economic account of torts. The content of
anyone’s tort-based rights and duties on such a theory is supposed to answer ultim-
ately to the impersonal goal of efficiency. The ideal of efficiency mediates the norma-
tive relationship between individuals. The justificatory relationship that holds between
persons on such a view is therefore indirect. Everyone’s conduct is judged by its
promotion of an impersonally good state of affairs, and that aim regulates how
individuals should treat one another. It is efficiency that determines the grounds of
liability through the choice of an overarching liability regime. Individuals’ tort-based
rights and duties are, therefore, a function of an overarching liability regime selected
for how well it advances an impersonal goal. So, some impersonal valuable state of
affairs drives the selection of liability regimes, which in turn shape the contours of
anyone’s rights and duties within those liability regimes.
In the direct justificatory relationship underwritten by justification to a subject,
however, individuals’ tort-based rights and duties are a function of what can be
justified between individuals, full stop. A liability regime emerges, but it does so as a
consequence of this bilateral, relational form of justification. Likewise, a valuable state
of affairs obtains—something like global relations of mutual recognition—but, again,
what is doing the work in fixing the content of individual rights and duties here is not
the emergent super value but the direct justificatory relationship that gives rise to it.
What we have, then, is the inverse of the theoretical ordering displayed by an
economic account: individuals justify their conduct to each other directly within a
bilateral relationship, which gives rise to a particular liability regime, which in turn
expresses some encompassing value.
Founded on contractualist justification to a subject, tort law does not set out to
achieve efficiency or any other impersonal goal. Still, the morality embedded within
the Palsgraf perspective has a “social function,” to use Warren Quinn’s words, which
“is to define our proper powers and immunities with respect to each other, to specify
the mutual authority and respect that are the basic terms of voluntary human

49
This is not to say that it does not also make room for retrospective moral evaluation. See T.M. Scanlon,
Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, MA: Harvard University Press, 2010).
118 JOHN OBERDIEK

association.”50 This aim makes sense of Cardozo’s concern with relational duty in
Palsgraf. If a normative domain like tort law is founded upon justification to a subject
and is therefore fundamentally about defining appropriate interpersonal powers and
immunities, and a (non-relational) duty of care with the scope imagined in Judge
Andrews’s Palsgraf dissent would run roughshod over such considerations, then the
kind of “prevision” that Mrs. Palsgraf expected of the LIRR’s employees would most
certainly be “extravagant.”51
Mrs. Palsgraf could not have reasonably expected the LIRR’s employees to foresee
her injury; she could not have justified that expectation directly to them. That finding
for Mrs. Palsgraf might have been efficient or might have advanced distributive justice
is inapposite, then, because it could not have been justified to those who would have
been held liable in the service of those impersonal ends. How people are to relate to
one another—what “our proper powers and immunities with respect to each other”
are—varies depending on whether the duties to which they are subject are relational or
non-relational. If one is liable to others for harms that one could not have reasonably
foreseen causing them, as Mrs. Palsgraf alleges and Andrews favors, then to avoid
liability one would have to be so vigilant of others’ well-being that, as Cardozo puts it,
“life will have to be made over.”52 Cardozo recognizes that unless one’s tort-based
duties are sensitive to what anyone can reasonably expect of another, those duties will
be unjustifiable to those on whom they bear—perverse even—because they will be
incompatible with a valuable, recognizably human life that is not simply given over to
the protection of others. Scanlonian contractualism, better than any other theory,
makes sense of this concern.
The Palsgraf perspective’s commitment to justification to a subject yields another
virtue as well, which Cardozo was clearly alive to, concerning its characterization of
wrongness. Cardozo famously contends, “[w]hat the plaintiff must show is ‘a wrong’ to
herself, i.e., a violation of her own right, and not merely a wrong to some one else, nor
conduct ‘wrongful’ because unsocial, but not ‘wrong’ to any one.”53 What makes the
relational conception of duty described here so attractive is that it captures the
distinctive moral significance of the wrongful treatment of people: there are victims
of tortious behavior who have a unique claim to better treatment. Only tort theories
founded upon justification to a subject can recognize this.
There is no party who is wronged by conduct that cannot be justified to the world at
large. Normative theories founded upon justification to the world at large can of
course determine that certain actions are or would be wrong. But they do not tie
that wrongness to anyone who is or would be wronged because, under such theories,
individuals are not owed any non-derivative justification. Justification to a subject,
on the other hand, zeroes in on the parties who are or would be wronged by certain

50
Warren Quinn, “Actions, Intentions, and Consequences: The Doctrine of Doing and Allowing,” in
Morality and Action (New York: Cambridge University Press, 1993), 173–4; see also Scanlon, What We Owe to
Each Other (note 32) at 162.
51 52 53
Palsgraf, 162 N.E. at 100. Palsgraf, 162 N.E. at 100. Palsgraf, 162 N.E. at 100.
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 119

conduct, for it is people as sources of claims who are owed justification. Consequently,
to the extent that some action is wrong, it is wrong because it wrongs someone in
particular (or many, taken individually). Goldberg and Zipursky express this in terms
strikingly similar to Nagel’s, maintaining that the duties recognized by tort law “are
owed by certain persons to others: duties that, when breached, constitute wrongs in
those others, as opposed to wrongs to the world.”54
In this way, justification to a subject isolates a distinctive kind of wrongness: what is
wrong is so because it wrongs.55 The reason one must not do wrong is that those who
would be its victims have a claim that one not do so—the wrongful conduct cannot be
directly justified to those who would be victimized by the wrong. Likewise, retrospect-
ively, one who has been victimized stands in a special relation to the person who has
wronged him or her, giving rise to a claim of redress by the wronged person against the
wronging person—a central insight of corrective justice and civil recourse theories that
contractualist tort theory so understood seamlessly accommodates.
Failures to respect one’s obligations under justification to the world at large will be
“‘wrongful’ because unsocial,” but they will not wrong anyone. Yet plaintiffs have
claims. They ought to be treated morally because of their claims, not because it will
enhance efficiency or even distributive justice—those are the wrong kinds of reason.
Plaintiffs do not go to court to correct an inefficient state of affairs or general
unfairness. The relational conception of duty embedded within justification to a
subject is the only one that takes seriously the claims that arise within interpersonal
relations. Only theories founded upon justification to a subject can recognize authentic
claimants and claims, thus capturing the character of wrongdoing. The Palsgraf
perspective, based upon contractualist justification to a subject, makes sense of the
distinctiveness of moral wrongdoing.
Viewing a theory of tort law through the lens provided by its foundational form of
justification also highlights the kinds of consideration that are relevant to justification:
justification to the world at large accepts, in Scanlon’s parlance, only “impersonal
reasons,” while justification to a subject accepts only “personal reasons,” as coins of the
realm.56 This further substantiates the correspondence between the Palsgraf perspec-
tive and Scanlonian contractualism and constitutes a further virtue of the approach.
For the kinds of reason bearing on justification to a subject are precisely those that tort
law’s bilateral structure admits as exclusively relevant. What this shows is that only
tort theories incorporating justification to a subject can make sense of tort law’s status
as private law.

54
Goldberg and Zipursky, “Torts as Wrongs” (note 42) at 918. Emphasis in the original. There remains a
question about just how relational this relationality must be, underscored by my unease with Keating’s
derivative bilateralism, which I do not pursue here.
55
For illuminating discussion, see Rahul Kumar, “Defending the Moral Moderate: Contractualism and
Common Sense,” 28 Philosophy and Public Affairs 275 (1999), 280–1. The first contemporary statement along
these lines is found in G.E.M. Anscombe, “Who is Wronged?,” 5 Oxford Review 16 (1967).
56
See Scanlon, What We Owe to Each Other (note 32) at 219.
120 JOHN OBERDIEK

The considerations relevant to justification to the world at large are impersonal

in that they derive from no one—impersonal reasons make no reference to any-
one’s point of view. Instead, they refer only to the net effect on whatever state of
affairs is being pursued, e.g., efficiency or distributive justice. In this respect, any
individual’s approval or disapproval of some liability regime or of specific conduct
within a liability regime is not the approval or disapproval of anyone in particular,
except in a trivial sense. Under justification to the world at large, there is no one
on whose behalf one assesses the justifiability of conduct or of some broader
liability regime—the assessment must be based, fundamentally, on impersonal
grounds. This should have a familiar ring: the ubiquitous “public policy consider-
ations” invoked in torts cases to protect or advance some larger value—a concern
with “floodgates” is a standard example—are impersonal reasons. Such reasons are
necessarily extrinsic to the bilateral relationship that defines a tort suit. Tort
theories founded on justification to the world at large, thus, fail to respect tort
law’s status as private law.
In contrast, the only reasons that are relevant under justification to a subject are
personal reasons, namely, those that derive from individual points of view. This
stricture respects the status of persons as the sole sources of claims and evinces a
commitment to normative individualism that is not shared by theories founded on
justification to the world at large. The personal reasons restriction embedded within
contractualist justification to a subject is, thus, a moral analogue of sorts to the legal
requirement of standing, which “focuses on the party seeking to get his complaint
[heard] and not on the issues he wishes to have adjudicated.”57 For one’s claim to meet
a requirement of standing, it is not enough that there be some blight on the world to
which someone objects; the claim must be the claimant’s in the relevant sense. The
considerations that any party to a tort suit may invoke to press a claim must, therefore,
arise from his or her point of view—valid reasons must be intrinsic to the bilateral
relationship. It is this that makes tort law fundamentally private law. Only theories
founded on justification to a subject respect that status in their very design.
Furthermore, explaining the personal reasons requirement in this way shows how it
is linked to the idea that something is wrong because it wrongs someone in particular:
a tortious wrong will wrong someone in particular and its unjustifiability will be
explained by reference to the wronged individual’s personal reasons. These two aspects
of justification to a subject—that wrongs are wrongs to someone and that the reasons
bearing on justification must be personal reasons—are therefore corollaries of each
other. This, in turn, infuses the personal reasons restriction with the theoretical virtue
and explanatory power of justification to a subject’s form of wrongness—that some-
thing is wrong because it wrongs. And in doing so, it imbues the Palsgraf perspective
with greater theoretical and normative depth.

57
Flast v Cohen, 392 U.S. 83, 99, 88 S.Ct. 1942, 1952 (1968) (emphasis added). See also Zipursky’s
discussion of “substantive standing” central to the Palsgraf perspective at note 38.
 STRUCTURE AND JUSTIFICATION IN CONTRACTUALIST TORT THEORY 121

VI. Conclusion
Contractualist theories of tort law are not monolithic, and, in some cases, they share
only a family resemblance. The Scanlonian version of contractualism that I have
introduced here as a basis for tort law does not revolve around distributive justice
like the well-known Rawlsian contractualist tort theories expounded by Fletcher and
Keating, even as it shares other commitments. Instead, it centers on a particular
conception of justification. This form of justification, justification to a subject, is
independently attractive and will shore up the moral merits of any tort theory that
is based upon it. I have tried to show that a theory of tort law based upon it also makes
good interpretive sense, unifying and deepening our understanding of various elem-
ents in the so-called Palsgraf perspective and, indeed, expanding that perspective’s
scope. Many are left cold by economic theories of torts, either on interpretive or more
thoroughly moral grounds. If there is to be a true rival to the economic approach to
torts, it seems to me that the contractualist theory that I have begun to sketch here is
the best hope.
 6
On the “Property” and the
“Tort” in Trespass
Eric R. Claeys*

I. Introduction
Over the last two decades, philosophical scholars have vied to identify tort law’s most
important priority. According to the better-known view, tort embodies corrective
justice, i.e., the field’s main goal is to rectify wrongs. According to a less prominent
view, even if tort seems to focus on wrongs, it is more logical and fundamental for the
field to protect rights from any possible wrong.1 Recently, however, disputants have
suggested that this debate is framed badly. Wrong-correction and rights-protection
supply answers to different theoretical questions. If the question is, “What supplies the
most basic foundation for having tort law?,” the answer is “Rights-protection.” If the
question is, “What seems best to explain tort’s structure in practice?,” the answer is
“Wrong-correction.”2

* I thank Lloyd Cohen, Pat Kelley, Bruce Kobayashi, Adam Mossoff, and Elina Treyger for helpful
comments, and Peter Cane, John Gardner, Amir Nezar, and Adam Scales for especially penetrating comments
and criticisms.
1
On corrective justice, see Jules Coleman, The Practice of Principle: In Defence of a Pragmatist Approach to
Legal Theory (Oxford: Oxford University Press, 2001), 31–7; Ernest J. Weinrib, The Idea of Private Law
(Cambridge, MA: Harvard University Press, 1995), 56–83. On protection, see Gregory C. Keating, “The Priority
of Respect Over Repair,” 18 Legal Theory 293 (2012); John Gardner, “What Is Tort Law For? Part I—The Place of
Corrective Justice,” 30 L. & Phil. 1 (2011); Stephen R. Perry, “Comment on Coleman: Corrective Justice,” 67 Ind.
L.J. 381 (1992), 388–9. The latter position is associated as well with distributive justice—in the sense that tort
distributes benefits by way of securing individual rights, and burdens in the form of obligations to respect those
rights. See Peter Cane, “Distributive Justice and Tort Law,” 2001 New Zealand L. Rev. 401.
A possible third option is civil recourse theory. See John C.P. Goldberg and Benjamin Zipursky, The Oxford
Introductions to Law: Torts (New York: Oxford University Press, 2010). In this chapter, I assume that civil
recourse is subsumed within corrective justice. See Scott Hershovitz, “Corrective Justice for Civil Recourse
Theorists,” 39 Fla. St. U. L. Rev. 107 (2011).
2
See Jules Coleman, “Epilogue to Risks and Wrongs,” 24–6 (unpublished manuscript) available at <http://
ssrn.com/abstract=1679554> accessed February 14, 2013. This chapter qualifies and supersedes observations
I have made in previous scholarship about the relationship between property norms and corrective justice, see
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 123

I believe that this reconciliation is accurate and insightful (although I also believe
that rights-protection gets priority because the question it answers is more fundamen-
tal).3 That reconciliation points to an important field for follow-on scholarship:
to clarify how tort manages to protect substantive moral rights within a corrective
structure. This inquiry may seem unambitious, for it does not stake out a distinctive
grand unifying theory of tort. Yet recall the parable in which six blind men argued
whether elephants are walls, tree trunks, snakes, fans, snakes, or ropes. Each grand
unifying theory overgeneralized from a valid insight about one body part all elephants
have to an absurd conclusion about elephants as wholes. In tort, durable conceptual
inquiry should make tort’s protective and corrective functions intelligible, and also
explain how each complements the other to make tort operate as a whole.
Two torts are better suited than all others to make concrete the interplay in
tort between rights-protection and wrong-correction: battery and trespass to land.
Although many tort scholars assume that negligence is the paradigmatic tort, trespass
and battery focus on wrongs simpler and more paradigmatic than negligence.4 In this
chapter, I focus on trespass. Lawyers and scholars appreciate that “property” and “tort”
are largely autonomous, yet partly complementary. They intuit that property justifies
the normative interests protected by property rights (viz., the exclusive control over
and use of land). They also intuit that tort supplies important legal infrastructure for
enforcing those rights (viz., trespass to land).5 By contrast, there is no field of “liberty
law” or “bodily-safety law” autonomous from tort. The lack of such a field makes it
difficult for tort scholars to appreciate the interplay between battery’s person-protect-
ing and wrong-correcting functions.
In this chapter, then, I explain how trespass to land and related doctrines implement
normative goals associated with the law and policy of property. I hope to make the
interplay between rights-protection and wrong-correction concrete by way of case
study. I also hope that this explanation provides partial confirmation that tort is
simultaneously protective in its ultimate goals and corrective in its operating structure.
I use the term “explanation” intentionally. I assume that trespass and related torts are
meant to effectuate property-related policies; I focus on descriptive and conceptual
issues about how those torts do effectuate these policies within a corrective structure.

Eric R. Claeys, “Jefferson Meets Coase: Land-Use Torts, Law and Economics, and Natural Property Rights,” 85
Notre Dame L. Rev. 1379 (2010). I make these qualifications and revisions primarily in response to Coleman,
“Epilogue to Risks and Wrongs” (this note); and Gardner, “What Is Tort Law For? Part I—The Place of
Corrective Justice” (note 1).
3
For my prior efforts to reconcile rights-protection and wrong-correction, see Eric R. Claeys, “Private Law
Theory and Corrective Justice in Trade Secrecy,” 4 Journal of Tort Law no. 2, article 2 (2011), 13–27. I believe
that rights-protection is more fundamental than wrongs-correction because the norm that justifies having and
enforcing a legal rule is more fundamental normatively, and seems more fundamental in legal reasoning, than
a second norm that constrains the structure by which that first norm is implemented in law. See John Finnis,
Natural Law & Natural Rights, 2d ed. (Oxford: Oxford University Press, 2011), 9–19.
4
See Coleman, “Epilogue to Risks and Wrongs” (note 2) at 22.
5
See Eric R. Claeys, “Exclusion and Private Law Theory: A Comment on Property As the Law of Things,”
125 Harv. L. Rev. Forum 133 (2012), 147–50.
124 ERIC R . CLAEYS

I also hope that this case study sheds light on how tort law protects substantive rights
more generally. Conceptually, the ultimate priority of the land-related property torts is
to protect land-related property rights. Because tort law has a corrective structure,
however, it discharges this priority indirectly and backhandedly. Tort doctrines will
seem indirect to property scholars. The normative interests in exclusive control over
and use of land are defined not by any one tort doctrine but by the coordination of all the
relevant ones in the right land-use disputes. The same doctrines will seem backhanded
to tort scholars. The fact that tort simultaneously discharges two functions complicates
how it discharges each. When trespass and other doctrines protect substantive property
rights, they do so not by declaring the rights but by remedying wrongs to those rights.
These wrong-focused actions manage to stay protective because the wrongs are parasitic
on logically-prior accounts of the underlying rights.
Here, some readers may wonder: why does tort have such a convoluted structure?
Readers may find at least one of three possible answers satisfying. First, who cares?
There is scholarly value in “begin[ning] not at the top, but in the middle, by asking what
principles if any, are embodied in the legal practices we are presently engaged in.”6
Second, because “[t]he backbone of tort is a set of moral—natural law—principles
identifying as wrongful all choices precisely to harm or to deceive.”7 That assertion may
be somewhat overdrawn. The moral “backbone” for any particular scheme of tort law is
not automatically “natural” simply because legal decision makers who apply it believe
that it is. Nevertheless, it confirms natural law that decision makers assume that
particular judgments they make must fit intelligibly onto to some backbone. So too
that decision makers take such backbones for granted while settling particular disputes.
My last possible answer is a pragmatic conceptualist one:8 “It’s the worst system
available, except for all the others.” A wrong-based approach trades on people’s
selfishness. Citizens may reliably be expected to assert their claimed rights to their
fullest—and to accuse anyone who stands in the way of committing a wrong. A wrong-
based approach also accommodates value pluralism. A system of law is less likely to get
bogged down if public officials target wrongs than if they dispute what values best
justify the wronged rights—utility, fairness, equality, autonomy, flourishing, or some
other value. In addition, civil judicial proceedings are the likeliest venues for clarifying
the scope of rights—and they are almost certain to be more effective if they are focused
on the wrongs. Criminal law can and sometimes does declare property rights.9 But
substantively, criminal law focuses more often on punishing clear aggression than it

6
Coleman, “The Practice of Principle: In Defence of a Pragmatist Approach to Legal Theory” (note 1) at 5.
7
John Finnis, “Natural Law: The Classical Tradition,” in Jules Coleman and Scott Shapiro (eds.), The
Oxford Handbook of Jurisprudence and Philosophy of Law (Oxford: Oxford University Press, 2002), 1, 45.
8
Cf. Benjamin C. Zipursky, “Pragmatic Conceptualism,” 6 Legal Theory 457 (2000).
9
See, e.g., People v Emmert 597 P 2d 1025 (Colo S Ct, 1979) (holding that recreational boaters are not
entitled to boat on rivers not navigable under traditional tests of navigability without the consent of the owners
of the beds beneath the rivers); State v Shack 277 A 2d 369, 370 (NJ S Ct, 1971) (holding that a landowner’s
possessory interest in control does not entitle him to exclude government caseworkers or publicly funded legal
advocates from meeting migrant farmworkers residing on his land).
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 125

does on determining the precise limits on rights allegedly aggressed-on. And institu-
tionally, criminal lawsuits are initiated and resolved by public prosecutors, not owners.
The latter have strong incentives to clarify the scope of their property rights in
litigation; the former hardly ever do. Property rights could also be declared in in
rem actions—as claims to title are in quiet-title proceedings. But it is far easier, less
tedious, and less threatening to non-parties’ due process interests for a court to
pronounce that a defendant is more culpable than the plaintiff than it is for the
court to enumerate all of plaintiff ’s rights in relation to his res and everyone else in
the world.10
Yet even if the tort system generally avoids making sweeping pronouncements
about property rights, it does operate as if such pronouncements exist. I hope to
demonstrate as much by showing how five salient doctrines implement a single
property-related political morality. I deem “salient” here: one paradigm “property”
tort (trespass to land); one “tort” defense to trespass (necessity); another “property”
defense to trespass (adverse possession); the tort that gets pride of place among torts
(negligence); and one orphan action usually overlooked in both property and tort
(trespass to chattels).

II. Property in a Lockean Morality of Labor

Throughout most of this chapter, I will interpret the tort doctrines I have selected
assuming that a labor-based morality justifies the property rights being secured by
trespass and the other doctrines selected. Readers need not find labor-based morality
normatively persuasive to follow my conceptual use of it here. Although such a
morality is not in vogue today, many, if not most, of the seminal English and American
trespass, nuisance, and other related cases relied on some such morality.11 Conceptual
interplays are easier to follow when the cases rely on a single political morality than
when—as in current tort law—different cases and scholarship assume different foun-
dations. In addition, in Anglo-American law, judges have relied on labor-based
morality to generate fairly hard-edged practical prescriptions about property rights.
These hard-edged prescriptions may be, and have been, applied consistently across a
wide range of property torts. The simpler and more consistent the substantive
prescriptions, the easier it is to track conceptual interplays between substantive right
and tort structure.

10
See Richard A. Epstein, “Possession As the Root of Title,” 13 Ga. L. Rev. 1221 (1979), 1222–4.
11
See Thomas M. Cooley, A Treatise on the Law of Torts or the Wrongs Which Arise Independent of
Contract (Littleton, CO: Fred B. Rothman & Co., 1993) (Chicago, IL: Callaghan, 1880), 1–6, 8–19; James Kent,
Commentaries on American Law (New York: Da Capo Press, 1971) (New York: O. Halsted, 1827), 2:255–76; 2
William Blackstone, Commentaries on the Laws of England *1–*15 (A.W. Brian Simpson ed., 1979) (1766);
Claeys, “Jefferson Meets Coase: Land-Use Torts, Law and Economics, and Natural Property Rights,” (note 2) at
1394–430.
126 ERIC R . CLAEYS

Let me supply an extremely compressed restatement of a common-denominator,

labor-based justification for property rights.12 In this justification, legal property rights
are justified by their tendencies to secure to members of the political community their
natural rights to labor. Here, “natural rights” must be understood against the backdrop of
an egoistic morality. In principle, this morality is grounded in human flourishing. In
practice, however, this morality encourages legal decision makers to focus on the lowest
and most urgent aspects of flourishing. Government actors are competent enough to
determine whether particular actions contribute to their own or others’ “Subsistence and
Comfort,”13 but they are usually incompetent to settle citizens’ disputes how to rate
different comforts.14 By focusing on a natural right to labor, this morality declares that
tort law and other conventional rules are judged by how well they secure nonconventional
moral interests to those interests’ claimants. By describing those interests in terms of the
rights, the morality encourages claimants to assert their own just interests spiritedly.
In this context, “labor” refers to planned activity, intended and reasonably likely to
generate for the laborer subsistence or improvement.15 (Henceforth, I will refer to such
goals as “self-preservation and self-improvement” when context requires distinction,
and “prosperity” or “benefit” when it does not.) In relation to an external resource, the
right to labor on the resource means the right to “use” the resource “beneficially” or
“productively.” Property rights are justified because they give proprietors the priority
and security they need to labor to further their plans for using resources beneficially.
Labor theory imposes four separate correlative constraints on the nonconventional
right to claim property for labor. To enjoy a labor-based claim over a resource, a
laborer must actually “make use of” it “to any advantage of life before it spoils.”16 (I call
this requirement the responsibility not to waste.) The laborer must also mark the
resources he appropriates so others know not to appropriate them, by “put[ting] a
distinction between them and common.”17 (I call this requirement “claim-marking.”)
A laborer must defer to others’ interests in acquiring and consuming resources “as will
keep [them] from extream [sic] want, where [they have] no means to subsist other-
wise.”18 (This requirement, which I call here the “necessity”19 proviso, confirms how

12
The following restatement compresses and adapts to relevant context Claeys, “Jefferson Meets Coase:
Land-Use Torts, Law and Economics, and Natural Property Rights” (note 2) at 1398–404; Eric R. Claeys,
“Productive Use in Acquisition, Accession, and Labor Theory,” in James E. Penner and Henry E. Smith (eds.),
Philosophical Foundations of Property Law (Oxford: Oxford University Press, 2013), 13, 17–32.
13
John Locke, “Two Treatises of Government,” in Peter Laslett (ed.), Two Treatises of Government
(Cambridge: Cambridge University Press, 1988) (London: Awnsham & Churchill, 1689), v.I, } 92, p.209.
14
On Locke’s aversion to judging differences in different people’s taste, see John Locke, An Essay
Concerning Human Understanding, Peter Nidditch (ed.), (Oxford: Clarendon Press, 1979) (London:
Awnsham & Churchill, 1700), v.II, ch.xxi, } 55, p.269–70.
15
See Blackstone, “Commentaries on the Laws of England”(note 11) at 2:*2–*3 (finding a “foundation in nature
or in natural law” for the right to use property for “such things as [man’s] immediate necessities required”).
16
Locke, “Two Treatises of Government” (note 13) at v.II, } 31, p.290.
17
Locke, “Two Treatises of Government” (note 13) at v.II, } 28, p.288.
18
Locke, “Two Treatises of Government” (note 13) at v.I, } 42, p.170.
19
See Hugo Grotius, De Jure Belli Ac Pacis Libri Tres, Francis Kelsey (trans.) (Oxford: Clarendon Press,
1925), bk.II, ch.2, }}6–9, p.193–5. Locke referred to this liberty not as a right of necessity but rather as a right
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 127

labor theory prioritizes self-preservation over self-improvement in its many forms.)

Lastly, because every citizen’s right to labor is equal to every other’s, the “sufficiency
limitation” requires each citizen to leave others enough and as good opportunities to
appropriate and use resources for their own prosperities.20
These non-conventional foundations justify any conventional system that seems
practically likely to enlarge citizens’ opportunities to labor beyond the opportunities
they would have in an unorganized community. As a result, the morality any com-
munity enforces relies only in part on non-conventional labor rights; it also relies in
part on somewhat-contingent community opinions how labor will best be facilitated.
Here, Locke and seminal Anglo-American jurists largely agreed on a practical and
rough empirical judgment: the best way to encourage concurrent labor in relation to
land is to endow proprietors of land with broad rights of exclusive control, possession,
and managerial discretion over the future uses of their lots. (I will refer to this collection
of rights here as “control rights.”) Such rights free proprietors to make long-term
investments into, and far-reaching transformations of, their lots. Land can generate on
the order of 100 times more life benefits than it could when used temporarily by
individuals foraging off of it while exercising usufructs.21 Such control rights do not
violate the sufficiency proviso—at least, not as long as the political community institutes
a reliable currency and open markets for labor and commerce. Under such conditions,
people who do not possess land themselves may acquire the goods they could have
appropriated from land by some combination of work and exchange.
Although labor-based morality does supply an indirect-consequentialist justifica-
tion for control rights, in principle the justification may become too attenuated.
The justification breaks down if a non-proprietor has a genuine claim based on the
necessity proviso. It also breaks down when if a particular proprietor violates
the nonwaste responsibility—e.g., if he lets “the Grass of his Inclosure [rot] on the
Ground, or the Fruit of his planting [perish] without gathering.”22 If legal control
rights assign proprietors broad control, but do so in a manner that does not clearly
delineate what those proprietors own, the legal rights may undermine property’s
claim-marking function. Additionally, positive law control rights may prevent non-
proprietors from exercising their just sufficiency interests.
Separately, positive law may justifiably qualify positive law control, when doing so
seems practically likely to enlarge the likely productive uses most proprietors will make
of their lots. Assume that a city requires all homeowners to lay sidewalks according to

(a “Title”) of charity the needy could claim against the propertied. Locke, “Two Treatises of Government”
(note 13) at v.I, } 42, p.170.
20
Locke, “Two Treatises of Government” (note 13) at v.II, } 27, p.288. See also Locke, “Two Treatises of
Government” (note 13) at v.II, } 4, p.269; Jeremy Waldron, The Right to Private Property (Oxford: Oxford
University Press, 1988), 209–18.
21
Locke, “Two Treatises of Government” (note 13) at v.II, } 40, p.296. Hence Locke’s contrast between the
king of a native American tribe and a day laborer in England. Locke, “Two Treatises of Government” (note 13)
at v.II, } 41, p.296–7.
22
Locke, “Two Treatises of Government” (note 13) at v.II, } 38, p.295.
128 ERIC R . CLAEYS

specifications it sets. The city invades those owners’ uses of the ground beneath the
sidewalks, but if its plans are well executed, it expands their liberties to travel more
than it diminishes their control.23
Last, although principles like “control rights,” “waste,” “sufficiency,” “necessity,” and
“clear marking” all have some content, the legal norms that implement them remain
somewhat contingent and open-texture. Sometimes, labor-based morality may require
legal decision makers to make a slightly arbitrary choice among several plausible
property rules—as a broad natural right to bodily safety requires legislators to choose
one of many possible speed limits. Often, practical judgment often requires legal
decision makers to determine how to reconcile or trade off between different related
goals. At least as often, labor-based morality requires legal decision makers to use
indirect-consequentialist reasoning. And such reasoning often requires decision
makers to make predictions that are implicitly empirical but practically impossible to
verify with precision. For these and other reasons, labor-based morality does not supply
fully determinate answers or prescriptions in property, tort, or any other field. That
said, such a morality does frame and narrow the questions those fields ought to answer.

III. The Unconsented-Entry Paradigm

When justified by labor-based morality, tort hones in the proper content of such rights
by a series of approximations. I will classify such approximations as consisting
of five separate strategies. The first strategy is quite general: to identify a paradigm
for the control rights proprietors deserve to enjoy over land. In trespass and
other related land-use tort doctrines, that paradigm is supplied by the uncon-
sented-entry test. Blackstone explained how this test implements labor-based
moral interests. Once:
the right of meum and tuum, or property, in lands [is] established, it follows as a necessary
consequence, that this right must be exclusive; that is, that the owner may retain to himself
the sole use and occupation of his soil: every entry therefore thereon without the owner’s
leave, and especially if contrary to his express order, is a trespass or transgression.24

The individual natural right to “use” land (i.e., to use it to benefit human life) supplies
a moral foundation for property in land. Given land’s use potential, however, land-
owners25 deserve substantive interests not only in their use of land but also in their
“occupation” of it. That “occupation” means control rights—as Blackstone confirms
by assuming that “occupation” must be “exclusive” and “sole.” To be sure, when he
announces that boundary crossings best specify this right of sole and exclusive

23
See Palmyra v Morton 25 Mo 593 (Mo S Ct, 1857); Paxson v Sweet, 13 NJL 196 (NJ S Ct, 1832).
24
Blackstone, “Commentaries on the Laws of England” (note 11) at 3: *209.
25
Blackstone refers to “owners” here, and I will as well in the remainder of this chapter for ease of
exposition. Strictly speaking, however, many of the land-use torts protect not only absolute owners but also
tenants and other proprietors with shared, finite, or defeasible interests in present possession.
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 129

occupation, Blackstone makes a practical implementation choice. This choice may be

contested on several different grounds. A boundary rule may not be the best indirect-
consequentialist means to secure owner control; any prediction that it is such a means
is implicitly empirical but hard to verify. Additionally, any such indirect-consequen-
tialist prediction is likely to have exceptions and seem arbitrary. Assume, however, that
the members of the relevant political community have determined that a boundary-
crossing rule is good enough for government work. At that point, the community has
settled on a “property” prescription, which then generates several important concep-
tual implications in tort.
To begin with, tort is structured assuming that there exist substantive moral rights
logically prior to tort law, and that these rights “seem to merge entitlements to do, have,
omit or be something with claims against others to act or refrain from acting in certain
ways.”26 Blackstone confirms in his account of trespass; his justification for the tort is
layered on a prior account of a substantive property right in exclusive control. Black-
stone justifies owners’ having an entitlement (a Hohfeldian liberty27) to do things—
occupy the land, to direct its use without outside interference, and to enjoy benefits
resulting from such use. He also justifies owners’ having claims (a Hohfeldian claim-
right) against others—to be free from interference with use-choices made in the
legitimate exercise of the liberty. Yet the autonomy declared by this latter claim-
right runs no farther than appropriate to secure the normative interests justifying
the autonomy—for Blackstone, the normative interest in “use,” or productive labor.
Next, Blackstone’s account of trespass provides a paradigmatic example of what a
“tort” is in relation to such rights. Trespass makes central the relation between torts
and autonomy. “Torts” consist of civil wrongs to rights as just explained, where the
wrongs do not arise out of a violation of a contract.28 Because a mere unconsented
entry is a trespass, the wrong in trespass lies neither in the landowner’s economic
losses nor in his property damage, but in the interference with his rightful autonomy
over his land. If a paradigmatic tort is one that “most vividly captures and illuminates
what it is that makes a tort a tort,”29 the centrality of autonomy to trespass makes vivid
and illuminates how torts focus on autonomy generally. Alternately, tort may be
described as being layered on and implementing three-way accountability relation-
ships, between two parties and a range of act-situations implicating their interests.30
Trespass layers rights and obligations on the social relationships implicit in the saying,
“Good fences make good neighbors.”

26
Joel Feinberg, “The Nature and Value of Rights,” 4 J. Value Inquiry 243 (1970), 256.
27
See Wesley Newcomb Hohfeld, “Some Fundamental Legal Conceptions as Applied in Judicial Reason-
ing,” 23 Yale L.J. 16 (1913), 30–44. I use “liberty” where Hohfeld uses “privilege,” in part because “liberty”
seems closer to non-specialized lay usages about moral rights, and in part because “privilege” is a term of art
referring to some defense-side claims in tort.
28
See Goldberg and Zipursky, “The Oxford Introductions to Law: Torts” (note 1) at 1–5, 6–7; Scott
Hershovitz, “Two Models of Tort (and Takings),” 92 Va. L. Rev. 1147 (2006).
29
Coleman, “Epilogue to Risks and Wrongs” (note 2) at 22. See also Coleman, “The Practice of Principle: In
Defence of a Pragmatist Approach to Legal Theory” (note 1) at 32.
30
Coleman, “Epilogue to Risks and Wrongs” (note 2) at 12.
130 ERIC R . CLAEYS

Some readers may doubt the justice in such relationships. After all, Blackstone’s
unconsented-entry paradigm seems to accentuate the “relation . . . between an owner
and a thing,” and to eliminate the relation “between the owner and other individuals in
reference to things.”31 True, some passages of Blackstone—especially his description of
property as “sole and despotic dominion”—give color to this perception.32 Both
conceptually and normatively, however, any one owner’s property in land is part of
a package deal. Normatively, if labor-based morality is reasonably persuasive, it is just
to force individuals to refrain from asserting their sufficiency-based claims on one
person’s land if they are given ample opportunities to labor for their own prosperities
on other tangible resources. Conceptually, property and tort scholars err if they
complain that trespass creates an “analytical mismatch” between a landowner’s legal
rights and his and others’ normative interests in the use of land.33 Conveyancing laws
make it possible to acquire land. Lease and contract law make it possible to rent land
within non-owners’ means. Rules establishing a currency, commercial regulations, and
employment regulations make it possible to generate prosperity by work and savings.
Welfare entitlements supply a backstop. Normatively and empirically, one may ques-
tion whether—in English or American economies—these and other similar legal
entitlements have given land non-owners sufficient opportunities to labor. Conceptu-
ally, however, hard-edged trespassory rights do not need exceptions expressly
embodying sufficiency claims if these other fields of law justly secure those claims.34
Lastly, Blackstone’s unconsented-entry paradigm defines the rights trespass presumes
backhandedly—in reference to wrongs by non-owners. It would be tedious to list all of
the legitimate uses owners might make with their Hohfeldian liberty to determine the
uses of their lots. It would be even more tedious to enumerate all the potential duty-
holders obligated not to interfere with those uses—let alone all the activities by non-
owners that might jeopardize those uses. It is far more economical to declare that owners
of land hold an in rem claim-right to be free from unconsented entries. It is even more
economical for tort to declare that right by focusing on the correlative in rem duty, that
any non-owner commits a wrong if she enters an owner’s close without his consent.

IV. Affirmative Defenses

Now, the unconsented-entry paradigm does not supply a direct rule for legal decision-
making. The test institutes a rebuttable normative presumption, and the other four
strategies by which tort implements property rights all respond to that presumption.

31
Morris Cohen, “Property and Sovereignty,” 13 Cornell L. Q. 8 (1924), 12.
32
Blackstone, “Commentaries on the Laws of England” (note 11) at 2:*2.
33
Avihay Dorfman, “The Normativity of the Private Ownership Form,” 75 Modern L. Rev. 981 (2012), 982.
34
Or, perhaps, as long as legal decision makers in the English and American legal systems have believed
that trespassory legal rights have been part of a package adequately accommodating all citizens’ substantive
sufficiency interests, and their grounds for so believing have been reasonable enough to pass a laugh test. This
is what I mean when I assume in text that tort implementation choices are “adequate” or “reasonable” enough
for conceptual purposes.
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 131

The second strategy is to use affirmative defenses to override the unconsented-entry

test, when positive-law control seems likely to give owners too much control and non-
owners too little access to land for their respective needs. Assume that, in a recurring
land-use dispute, one of the four responsibilities or provisos associated with labor-
based property defeats the justification for control rights. Assume also that this dispute
occurs relatively infrequently, but just often enough to deserve its own doctrine. In
such conditions, it is reasonable to leave the prima facie case for trespass broad and
encompassing, and then use affirmative defenses to winnow out the (relatively few)
false positives caught in that case. If one takes tort’s rights-protecting function as
paramount, landowners’ substantive property rights are not specified by the prima
facie case for trespass in isolation. Rather, they are specified instead by the “interplay
among [prima facie] rules” in and “positive defenses”35 to trespass.
Let me illustrate with two representative defenses to trespass. The necessity privilege
basically embodies and recognizes in law the moral necessity proviso attached to labor
claims.36 Of course, the owner’s control rights and the necessitous entrant’s liberty
of access need to be harmonized. Necessity doctrine does so, and reasonably so. The
privilege gives priority to the entry only if the entrant faces a bona fide threat to his self-
preservation, and the privilege lasts only as long as the threat. The privilege also requires
the entrant to use the owner’s property with reasonable care, and to hold the owner
harmless for any damage caused by his temporary commandeering of the property.37
Adverse possession supplies a complete defense to trespass when a defendant
adversely occupies a plaintiff ’s land (i.e., with intent to appropriate it) exclusively,
notoriously, and continuously for the applicable limitations period.38 Many elements
of adverse possession determine subsidiary legal details relatively remote from the
doctrine’s core justifications. On labor-based foundations, however, those justifications
are to enforce the two responsibilities labor rights attach to ownership: to mark claims of
ownership, and not to waste owned resources. If a title owner “neglects to assert his
rights in the manner provided by law,” he confuses the occupier and his neighbors, he
confirms he is not using his land beneficially in any sense, and he must then “accept the
result of his own folly and negligence.”39 Such a negligent title owner then legitimizes
the adverse possessor’s having a claim, marked by long and notorious occupancy, to
own land “improved by his labor, and enriched by the sweat of his brow.”40

35
Jules L. Coleman, Risks and Wrongs (Cambridge: Cambridge University Press, 1992), 216. See also
Richard A. Epstein, “Defenses and Subsequent Pleas in a System of Strict Liability,” 3 J. Leg. Stud. 165 (1974).
36
See, e.g., Ploof v Putnam 71 A 188, 189 (Vt S Ct, 1908). For a more elaborate explanation why necessity is
not only a “tort” but also a “property” doctrine, see Dennis Klimchuk, “Property and Necessity,” in Penner and
Smith (ed.), Philosophical Foundations of Property Law 47 (manuscript on file with author).
37
See Vincent v Lake Erie Transp. Co. 124 NW 221, 221–2 (Minn S Ct, 1910).
38
See Thomas W. Merrill and Henry E. Smith, Property, Principles and Policies, 2d ed. (New York:
Foundation Press, 2012), 199 n. 3.
39
Oliver v Pullam 24 F 127, 131 (Cir Ct of NC, 1885).
40
Ransom H. Tyler, A Treatise on the Remedy by Ejectment and the Law of Adverse Enjoyment in the United
States: Embracing in Full the Statutory Policy of the Several States, in Respect to the Action for the Recovery of
Real Property (Albany, NY: William Gould & Son, 1876), 854. See Eric R. Claeys, “Productive Use in Trespass,
132 ERIC R . CLAEYS

When necessity and adverse possession specify property rights, they also do so
backhandedly. Adverse possession converts a wrong into a non-wrong. When com-
pleted, an adverse possession ceases to be a prima facie trespass and becomes a
legitimate occupancy and use of land. Necessity doctrine illustrates even more power-
fully. In Ploof v Putnam, Ploof, his family, and his boat all were injured when he tried
to tie the boat to Putnam’s dock during a storm and Putnam’s servant pushed the boat
away from the dock. Logically, the most important question was whether the storm
threatened Ploof and his family enough that they deserved a temporary power to
commandeer the dock. Doctrinally, however, the parties backed into that question.
Each accused the other of trespass against his own rights. Ploof sued Putnam for
trespass to chattels, i.e. interfering with his rightful control over his boat. When
Putnam pleaded defense of property,41 he implicitly accused Ploof of trespassing on
his dock. Necessity came into the case because it was Ploof ’s ground for demurring to
Putnam’s defense. Consistent with the rights-protective account, the parties’ dispute
was resolved by inquiring whether Ploof ’s substantive right of self-preservation
legitimately took priority over Putnam’s substantive right to control the dock for his
future uses. Consistent with the wrong-correction account, the court learned what it
needed to know to decide that question as each litigant proved his accusation that the
other had committed a wrong.

V. Conforming the Prima Facie Tort to the

Underlying Substantive Right
The third strategy by which tort implements substantive property rights is by config-
uring the prima facie case for a tort to conform to the paradigm suggested by the
underlying right. In trespass, an unconsented entry is the prima facie test.
One can confirm the same point by studying doctrinal elements not required in
trespass. The tort most familiar to practitioners—negligence—has a harm element;
trespass to land does not.42 Trespass’s rights-based character protects autonomy as
means to secure underlying interests in laboring. In the absence of extenuating
circumstances, even the mere entry onto land unduly threatens the secure use justi-
fying owner control. That is why it is “an elementary principle, that every unauthor-
ized, and therefore unlawful entry, into the close of another, is a trespass.”43
Treatise writers also like to classify torts as strict, intentional, or fault-based. In
practice, most trespasses are deliberate or the product of (careless) turn-a-blind-eye
mistakes. Yet trespass litigation can settle which of two parties with designs on a lot

Adverse Possession, and Labor Theory,” (unpublished) (<http://ssrn.com/abstract=1759551> accessed

October 24, 2013).
41 42
See Ploof, 71 A (note 36) at 188–9. Subject to exceptions recounted in the next part.
43
Dougherty v Stepp 18 NC 371, 372 (NC S Ct, 1835). See also Restatement (Second) of Torts } 158 (1965).
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 133

has better possession or title. Even if a defendant trespasses innocently and carefully,
his mere entry may threaten significant damage to the land or activities on it. It is far
easier for an owner (or his insurer, creditors, or neighbors) to monitor whether
entrants are on his land than whether they are on it carelessly or deliberately. If “[t]he
chief characteristic of ownership is th[e] right to complete dominion,” such that “[t]he
line of a man’s private domain, like the boundary line between nations, is not to be
crossed without permission,”44 then even innocent trespasses are culpable. The need for
“dominion” in property creates a strong presumption that liability be strict in tort.
Because property imperatives drive trespass liability to be strict, commonsensical
“fault”—intentional, knowing, or careless scienter—plays only a secondary role in
trespass. Commonsensical fault aggravates the wrong in a trespass. Such fault may
expose an already-liable defendant to punitive damages. It may disentitle that defend-
ant from pleading undue hardship if the plaintiff prays for an injunction.45 But the
trespass is wrongful without such aggravation; “[g]ood faith [should] not excuse [a
defendant] . . . , the law requiring him at his peril to ascertain what his rights are, and
not to invade the possession, actual or constructive, of another.”46
By the same token, although trespass’s strict character settles liability for accidents
caused in the course of trespasses, trespass is not an “accident” tort. All torts declare
and protect rights. Trespass to land provides landowners with property rights of
exclusive control. Trespass holds defendants “civilly liable for the consequences
which directly flowed from their unauthorized” trespasses,47 no matter how unfore-
seeable those consequences are,48 because those consequences are parts of the wrongs
to the property rights. So trespass determines liability for accidents, but only inciden-
tally, in the course of securing rights.
Trespass’s strictness makes it seem wrong or strange to many scholars. Readers who
find trespass’s strictness normatively unpersuasive should keep in mind that nothing
about trespass’s conceptual structure requires it to be a strict tort. A political commu-
nity could institute control rights narrower than the rights explained in Parts I and II,
and use explicit fault-based scienter requirements to limit owners’ rights under
positive law. Other scholars may assume that strict torts are inconsistent with cor-
rective justice. Not so. Pure strict liability is inconsistent with corrective justice,49 but
strict liability in trespass to land is neither pure strict liability nor inconsistent with
corrective justice. Corrective justice requires rectification of wrongs to owners’ legit-
imate substantive rights. Understood in proper context, corrective justice says nothing

44
Cooley, “A Treatise on the Law of Torts or the Wrongs Which Arise Independent of Contract” (note 11)
at 302.
45
See Jacque v Steenberg Homes, Inc. 563 NW 2d 154 (Wis S Ct, 1997) (upholding punitive damages); Isle
Royale Min. Co. v Hertin 37 Mich 332, 336–7 (Mich S Ct, 1877) (undue hardship).
46
Isle Royale Min. Co. 37 Mich (note 45) at 335.
47
Brown v Dellinger 355 SW 2d 742, 747 (Tex Civil App, 1962).
48
See Crook v Sheehan Enterps. Inc., 740 SW 2d 333 (Mo App Ct, 1987).
49
See Weinrib, The Idea of Private Law (note 1) at 171–83; Stephen R. Perry, “The Impossibility of General
Strict Liability,” 1 Can. J.L. & Juris. 147 (1988).
134 ERIC R . CLAEYS

about the content of those substantive rights, it assumes that some logically prior
political morality has specified those rights, and it layers corrective duties over logically
prior substantive duties correlative to those rights. As long as the unconsented-entry
paradigm fairly represents the situations in which a non-owner unjustly interferes with
an owner’s rightful exclusive control, the prima facie case for trespass embodies a
moral relation of fault even though that case lacks an express requirement of fault.
My use of “fault” here may prompt another objection: if “fault” is pliable enough to
cover a careful, innocent, and mistaken trespass, my use of the term must be incoher-
ent or idiosyncratic. This objection may hold trespass to expectations informed by
negligence, which has an express “fault” requirement. If so, that reaction just goes to
show that too many tort scholars confer on negligence an undeserved “imperialism,”
attributable to the fact that most reported torts cases sound in negligence.50 The
objection may also hold the “fault” in trespass to expectations about fault that stress
subjective state of mind—like criminal law.51 But this objection makes as little sense as
it would to deny that a square peg is a peg because it does not fit in a round hole.
Conceptually, “fault” operates as a general concept throughout many fields of law and
social interaction, as a specific concept in criminal law, and then again as a specific
concept in tort. The two specific concepts may and should differ from one another, to
accommodate salient differences between their fields of application. In criminal law,
“fault” focuses on subjective state of mind because that field focuses on rights-
invasions with strong public repercussions. Such repercussions arise most often
when rights-aggressors act with especially deliberate or aggressive intentions. (For
that reason, trespass’s rules about punitive damages and the undue-hardship defense
reinforce in civil law priorities more central to criminal law.) To a far greater degree
than criminal law, tort specifies the content of substantive rights. As not only land torts
but also speeding laws confirm, quite often, rights cannot be enjoyed unless they are
clearly defined and marked. To get that clarity, sometimes both property and civil
traffic law must prohibit conduct that is voluntary, subjectively well-meaning, but
objectively in violation of bright-line markers. In both fields, it is correspondingly
reasonable to classify innocent violations of objective duties as “faulty.”
Whether one agrees with this account normatively, trespass doctrine supplies
powerful confirmation that judges follow it intuitively and conceptually. If one wanted
to make trespass track criminal law-like conceptions of fault, the easiest way to do so
would be to graft a subjective intent requirement onto the prima facie case for trespass.
The Restatement of Torts does in fact classify trespass as an intentional tort.52 But take
a case in which a trespassing defendant appropriates resources from the plaintiff ’s
land carefully and mistakenly believing the land is his own,53 he causes property

50
Peter Cane, Tort Law and Economic Interests (New York: Clarendon Press, 2d ed. 1996), 11. For further
elaboration, see Part VII, infra.
51
See Larry Alexander and Kimberly Ferzan, “Confused Culpability, Contrived Causation, and the Collapse
of Tort,” Chapter 19 in this volume.
52
See Restatement (Second) of Torts } 158(a) (1979).
53
See Maye v Yappen 23 Cal 306 (Cal S Ct, 1863).
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 135

damage specifically intending not to do so,54 or he trespasses mistakenly and the

trespass enables a third party or an extraordinary natural force to damage the
property.55 In criminal law, the defendants in these hypotheticals are at best not
culpable or at worst only minimally culpable. In the civil law, however, all defendants
are culpable. In each, the plaintiff deserved secure control over his land, the defendant
upset that control, and the plaintiff ’s losses flow directly from the defendant’s invasion
of his control rights. No surprise, then, that in civil trespass, judges “gut the traditional
concept of intentional harm,”56 so that “intent” becomes indistinguishable from
the voluntariness a person displays in the course of acting.57 Even when trespass is
revised to fit opinions about fault from commonsense usages or criminal law, judges
recast the revisions—to keep trespass in accord with controlling judgments about
property.

VI. Harm-Based Exceptions to Rights-Based Torts

Similarly, tort scholars sometimes “wonder why trespass to land does not . . . have a
harm requirement,” especially seeing as trespass to chattels does.58 A trespass to
chattel may be defined as a direct interference with a plaintiff ’s possession of a chattel,
in a manner that causes the plaintiff harm.59 Cognizable “harms” include proof that
the chattel was damaged, that the owner was dispossessed of it, or that he lost its use
for a significant duration.60
Yet there is no compelling reason why trespass to land must be symmetrical to
trespass to chattels—not for pure symmetry’s sake, not to embody corrective justice,
nor to make trespass to land more like criminal law or negligence. As the same
scholars acknowledge, “each tort involves the violation of a norm that specifies how
one must treat others in light of certain important interests.”61 These scholars do not
consider why distinctions between different forms of trespass might reasonably
accommodate salient differences between the likely uses and management of land
and chattels. Proprietary norms encouraging the productive uses of resources may
justify limits on rights of exclusive control; trespass may implement those norms by
instituting harm-based limitations to trespass for particular situations where they seem
appropriate. Compare a paradigmatic trespass to land against what the Restatement

54
See Brown, 355 SW 2d (note 47) at 747.
55
See, e.g., Southern Counties Ice Co. v RKO Radio Pictures 39 F Supp 157, 159 (US Dist Ct SD Cal, 1941).
56
Richard A. Epstein, Torts (New York: Aspen Publishers, 1999), } 1.3.1, p.8.
57
See, e.g., Cleveland Park, Inc. v Perry 165 A 2d 485, 488 (DC Ct App, 1960) (“intent to do the physical act
which released the harmful force”); Restatment (Second) of Torts } 163, cmt. b (1979).
58
Goldberg and Zipursky, “The Oxford Introductions to Law: Torts” (note 1) at 248.
59
See Prosser and Keeton on the Law of Torts, 5th ed. (St. Paul, MN: West Publishing Co., 1984), } 14,
p.85–7.
60
See Restatement (Second) of Torts } 218 (1979); Koepnick v Sears Roebuck & Co. 762 P 2d 609, 617–19
(Ariz Ct App, 1988).
61
Goldberg and Zipursky, “The Oxford Introductions to Law: Torts” (note 1) at 2.
136 ERIC R . CLAEYS

(Second) of Torts portrays as a paradigmatic non-trespass to a chattel—a dispute in

which a child pulls on the ears of a dog that does not belong to her.62 The dog, like
most chattels, is movable; the land is not.63 Land cannot be brought into common
spaces like sidewalks or highways. Dogs can—and then create obstacles or risks of
accidents for others. Land cannot be moved easily out of the way of others’ dangerous
conduct; dogs can.64
To accommodate the differences between ownership of the land and the dog, tort
deploys a fourth strategy to approximate substantive property rights: qualify prima
facie rights-based torts, to secure directly the substantive interests that autonomy
usually secures indirectly. There are exceptions in both trespass to land and trespass
to chattels, but the exceptions confirm the rules. There are rights-based trespass to
chattels—for chattels that are immovable. If a company sells gas regulators for utility
customers to install on gas pipes, tort and remedy doctrine will presume that the
unconsented attachment of the regulator to the pipe creates the “harm” requisite for a
property tort. Since the gas pipes are immovable, if there is any possibility that the
regulators may cause explosions it is practically likely that only “sole control” gives
the utility company an adequate “right of protecting itself from loss caused by
interference.”65
And in trespass to land, sometimes doctrine qualifies the prima facie case for
recurring land-use disputes. In early American law, hunters had easements to traverse
private land as long as the land was not occupied, they did not remain on the land any
longer than necessary to hunt, and they caused no property damage.66 During the
nineteenth century, courts and legislators instituted limitations on suits involving
cattle trespasses. The limitations excused accidental and harmless trespasses by cattle—
but not deliberate or property-damage-causing trespasses.67 In the twentieth century,
courts limited prima facie trespass causes of action for over-flights. Landowners could
no longer sue for trespass at altitudes above the floors set for air flight by regulation,
and they could sue beneath those floors only when over-flights caused actual damage
to their lots.68 Each of these rules diminishes owners’ control over their lots. In the

62
Restatement (Second) of Torts } 218, cmt. e, illus. 2 (1979). The hypothetical comes from the facts of
Glidden v Szybiak 63 A 2d 233 (NH S Ct, 1949).
63
It is reasonable to design trespass to chattels law assuming it applies only to moveable chattels.
Immoveable chattels are usually deemed to be legal accessories to real estate by operation of the ratione soli
and fixture rules. See Cooley, “A Treatise on the Law of Torts or the Wrongs Which Arise Independent of
Contract” (note 11) at 426–32; Claeys, “Productive Use in Acquisition, Accession, and Labor Theory” (note 12)
at 39–43.
64
See Richard A. Epstein, “Intel v Hamidi: The Role of Self-Help in Cyberspace?,” 1 J.L. Econ. & Pol’y 147
(2005), 150.
65
Blondell v Consolidated Gas Co. 43 A 817, 818–19 (Md Ct App, 1899).
66
McConico v Singleton 9 SCL 244 (S Car Const Ct, 1818).
67
Lazarus v Phelps 152 US 81, 86 (1894). See also Restatement (Second) of Torts } 504 (1979).
68
See Hinman v Pacific Air Transp. Co. 84 F 2d 755 (US Ct of Apps (9th Cir), 1936); Swetland v Curtiss
Airport Corp. 41 F 2d 929 (US Dist Ct for ND of Ohio, 1930); Epstein, “Intel v. Hamidi: The Role of Self-Help in
Cyberspace? ” (note 64) at 154–5; Eric R. Claeys, “On the Use and Abuse of Overflight Columns,” 2 Brigham-
Kanner Property Rights Conference Journal 61 (2013), 76–88.
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 137

right socioeconomic conditions, however, owners receive reciprocating advantages in

the beneficial uses of others’ land—for sustenance (hunters’ rights), for pasturage
(cattle), or for travel and commerce (by air). Still, in each of these qualifications, the
unconsented-entry paradigm helps focus the tradeoffs and limit the scope of any
exception. It is reasonable to force owners to waive their claim-rights to blockade
innocuous entries if they are simultaneously endowed with reciprocal liberties to enter
others’ lots to hunt for sustenance, to graze livestock, or to fly. These implied
reciprocal bargains break down when trespasses generate actual property damage or
(as in cattle cases) when non-owners trespass with scienters as aggressive or deliberate
as those targeted in criminal law.
As do property-based affirmative defenses, harm-based intentional property torts
implement property prescriptions backhandedly. Harm-based trespass to land doc-
trine confirms when non-owners have proprietary rights—usufructs—in relation to
others’ lots of land. Like the defenses in Ploof, however, trespass doctrine declares these
usufructs backhandedly, by declaring that a non-owner commits no trespass when he
enters another’s land in the legitimate exercise of his usufruct. Harm-based trespass to
chattels is even more intricate. Black-letter sources declare that owners hold an interest
in the “inviolability” of their chattels.69 Such declarations make harmless contacts
wrongful, and they entitle chattel owners to exercise legal self-help. At the same time,
because chattels doctrine requires proof of harm, harmless contacts are not wrongful
in the sense that they trigger liability in trespass. That qualification subtly authorizes
non-owners to use their own self-help powers to move others’ chattels aside.

VII. Property Foundations and

Tort Implementation
Some readers may wonder whether, when tort law structures prima facie torts, it
always implements property prescriptions as automatically as I have suggested by my
examples in the last two parts. In easy cases, yes; in hard cases, no.
The formalism exhibited in the last two parts helps keep simple cases that could
readily be portrayed as complicated. Assume that a defendant discharges water
underneath a land-owning plaintiff ’s property. The land suffers no damage until
substantially later, when the plaintiff releases the water by accident while driving
support piles into his soil. The defendant’s water emission constitutes one of many
run-of-the-mill threats to the owner’s secure control over and future use of his land.
Since the plaintiff installed support piles after the defendant emitted water, however, it
would not be difficult to argue that the plaintiff should be reciprocally and proximately
responsible for his own losses.70

69
Restatement (Second) of Torts } 218, cmt. e (1979).
70
See R.H. Coase, “The Problem of Social Cost,” 4 J.L. & Econ. 1 (1960), 13.
138 ERIC R . CLAEYS

Tort law precludes this argument because it is structured to protect and secure moral
rights. In tort, “causation” means moral causation. In a trespass dispute, causation
focuses only on losses attributable to one party’s violation of the other’s rights of
exclusive control over the trespassed-on property. In the pile-driving example, then,
the “trespass was the proximate cause, legally, the sole cause, for the innocent act of the
owner in driving piles cannot be used as a shield by defendant any more than could the
innocent acts of the bystanders in the squib case.”71
Yet every novel question in a property-tort dispute provides another opportunity to
specify the scope of substantive property rights along a different margin. In such
marginal cases, property norms create presumptions or preferences for certain ranges
of results without automatically requiring any particular result in tort. In such cases,
while judges try to secure corrective justice, they simultaneously specify the substan-
tive rights that generate corrective obligations.
Let me illustrate using a recent case on trespass, Jacque v Steenberg Homes, Inc.72
Steenberg Homes needed to deliver a mobile home to meet a delivery date in a
purchase contract. Thanks to a blizzard, the most direct route to the delivery address
was blocked by snow, in drifts of seven feet at some spots. Steenberg Homes wanted to
circumvent the drifts by towing the mobile home across a field on the Jacques’ lot and
behind their house. Although company staff tried to bargain with the Jacques, they
refused and said “it was not a question of money; [they] just did not want Steenberg to
cross their land.” The Jacques had lost property to adverse possession and prescription
claimants in the past, and they probably refused Steenberg Homes’s request in the
(legally mistaken) belief that a crossing license would have given the company grounds
for asserting adverse possession. A Steenberg Homes assistant manager ordered employ-
ees to cross the field anyway, telling employees, “I don’t give a — what [Mr Jacque] said,
just get the home in there any way you can.” Steenberg Homes employees towed the
mobile home across the Jacques’ field, committing no significant damage to the field in
the process. After employees called and told the assistant manager that the crossing
was complete, he “reacted by giggling and laughing.”
The jury awarded the Jacques nominal damages of $1, found no actual damage to
their land, and awarded punitive damages of $100,000. Although the punitive damage
award was set aside by the trial court, the Wisconsin Supreme Court reinstated it.73 To
support punitive damages, the Supreme Court needed first to establish that the Jacques
had a rights-based cause of action for Steenberg Homes’s harmless trespass. In other
words, the Supreme Court needed to specify and clarify the scope of the substantive

71
Chicago v Troy Laundry Machinery Co. 162 F 678, 679 (US Ct of Apps (7th Cir), 1908) (referring to Scott
v Shepherd [1773] 96 ER 525, KB).
72
563 NW 2d 154 (Wis S Ct, 1997). The following discussion accords with, but adds considerable
specification to, my previous treatments of Jacque, in Claeys, “Jefferson Meets Coase: Land-Use Torts, Law
and Economics, and Natural Property Rights” (note 2) at 1407–9; Eric R. Claeys, “Property 101: Is Property a
Thing or a Bundle?,” 32 Seattle U. L. Rev. 617 (2009), 640–1.
73
Jacque, 563 NW 2d (note 72) at 157.
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 139

interest in control informing the Jacques’ complaint. The court concluded that the
Jacques’ interest in control was implicated by the intentional character of Steenberg
Homes’s trespass. Such a trespass, the court asserted, “causes actual harm to the
individual [owner], regardless of whether that harm can be measured in mere dollars.”74
In what follows, I will illustrate how a labor-based approach would have focused the
Wisconsin Supreme Court’s reasoning about the scope of the Jacques’ possessory
interest in control. Even though the court reasoned primarily in a utilitarian frame-
work, it relied on labor- and rights-based precedents about trespass75 and focused on
several of the policies a labor-based approach would make salient. I think that the
Wisconsin Supreme Court came to the right conclusion, and I also think that labor-
based foundations create a strong presumption for that conclusion. Here, however,
I am more interested in showing how those (property-based) foundations both
constrain and leave play in tort reasoning.
A labor-based approach would not automatically require that the Jacques be entitled
to a rights-based cause of action. If landowners are not entitled to causes of action for
harmless trespasses in hunters’ access, cattle-trespass, or over-flight disputes, the
Jacques were not entitled to exclude Steenberg Homes just by asserting “it was not a
question of money.” In labor-based terms, a responsible decision maker would need at
least to consider whether, on facts like those presented in Jacque, the autonomy
granted by the unconsented entry test might not secure the normative interests that
autonomy usually furthers.
Labor-based norms would institute some starting presumption in favor of the
Jacques’ enjoying autonomy. Although the unconsented entry paradigm operates as
a metaphor, not as a declaration of policy, it does embody several property-related
indirect-consequentialist presumptions. As in the tunnel example just discussed, the
rights-based approach freed the Jacques from needing to insure against accidental
property damage not caused by their own activities. Separately, the rights-based
approach conserves and reinforces all the advantages that clear boundaries provide
in a wide range of land-related transactions.
That said, the presumption for the rights-based approach is rebuttable, and the
norms that structure legal property rights help identify the legal reasons that count as
acceptable grounds for rebuttal. In labor-based terms, owner control deserves to be
suspended if the non-owner is commandeering property to deal with a necessity. In
Jacque, it was not Steenberg Homes’s fault that the regular delivery route was blocked
by snow. In addition, owner control should be overridden if the owner’s use claims
seem far weaker than those of the entrant. Hunters’ access defenses embody this
limitation, by protecting the hunter when the owner is absent and the land suffers
no damage. In Jacque, a harm-based approach could have recognized that the Jacques’

74
Jacque, 563 NW 2d (note 72) at 161.
75
See Jacque, 563 NW 2d (note 72) at 160 (affirming “[e]very person[’s] constitutional right to the exclusive
enjoyment of his own property for any purpose which does not invade the rights of another”) (quoting Diana
Shooting Club v Lamoreaux 89 NW 880, 886 (Wis S Ct, 1902)).
140 ERIC R . CLAEYS

stated grounds for refusing access were unfounded, and that Steenberg Homes was
trying to complete a useful contract.
Nevertheless, if tort law construed necessity claims in too freewheeling a fashion, or
if it granted harm-based exceptions too often, it might undermine some of the
indirect-consequentialist benefits property supplies. Ordinarily, labor-based property
norms strongly discourage legal decision makers from balancing the intended land
uses of one particular owner against those of a particular non-owner. Necessity lets
non-owners plead their rights of self-preservation against owners’ claims of control
and exclusion—but trespass doctrine makes it impossible for trespassers to argue that
their planned uses are likelier to improve their lives than the owner’s current use is
improving his. In Jacque, the Jacques were actually occupying their lots (like land-
occupying owners immune from hunters’ claims of access). They were also were
“using” the field behind their home in some productive sense—even if only to enhance
their aesthetic enjoyment of their home. Meanwhile, Steenberg Homes did not face a
life-threatening or property-destroying necessity; it could have waited to deliver the
home. If Steenberg Homes’s claimed use was not absolutely compelling, and if the
Jacques’ reasons for denying the license could possibly have been grounded in
legitimate claims of occupancy and beneficial use, perhaps it was better to err on the
side of the Jacques—and, prospectively, all occupants in possession.
Last, Steenberg Homes precipitated the confrontation. The Jacques had made clear
before the day of the confrontation that they did not want to license a crossing, and the
company precipitated a confrontation by moving the home anyway.76 The company
was also vicariously responsible for the contempt its assistant manager displayed
toward the Jacques. It was thus not unreasonable for the Wisconsin Supreme Court
to conclude that Steenberg Homes’s trespass was wrongful even if it did not cause
actual damage. Like an intentional cattle trespass, the company’s crossing was delib-
erate, even motivated by pleasure at the Jacques’ distress. No surprise, then, that (in
utilitarian jargon) the court professed to find decisive society’s “interest in punishing
and deterring intentional trespassers[, to] preserv[e] the integrity of the legal system”
and to discourage private vengeance by landowners.77
To be clear, the Wisconsin Supreme Court could have used necessity precedents or
harm-based trespass precedents without undermining a labor-based approach. But
holdings like these would have strained such an approach. In a common law system,
precedents gradually establish specific factual distinctions—like the distinction
between rights-based trespass for occupants and harm-based trespass for absentee
owners—determining when a new case should be governed by the presumptive rule or
the relevant exceptions. In a close case, these are the means by which a substantive
morality supplies a foundation for legal reasoning without totally determining that
reasoning. If readers disagree with the Wisconsin Supreme Court’s conclusion, those

76
See Jacque v Steenberg Homes, Inc. 548 NW 2d 80, 81 (Wis Ct of Apps, 1996), rev’d, 563 NW 2d 154 (Wis
S Ct, 1997).
77
Jacque, 563 NW 2d (note 76) at 160.
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 141

disagreements are almost certainly normative, not conceptual. In Jacque, relevant

background norms of property justified not only rights of control but also rights of
necessity and usufructuary access. From the perspective of tort law, even if background
property norms did not settle the parties’ dispute dispositively, they still focused the
dispute by identifying the salient issues. From the perspective of property law, Jacque’s
holding in tort was resolving a dispute arguably covered by several competing property
policies and paradigms.

VIII. Making Accident Torts Complement

Rights-Based Torts
A. The relation between accident torts and trespass
Thus far, we have focused primarily on how moral opinions about rightful control
over property inform trespass torts—the torts that focus most on such control. Yet
many other act-situations may threaten owners’ interests in controlling or using
property incidentally. Which takes us to the last main strategy by which tort imple-
ments and specifies property rights: to make sure that more general torts specify and
secure substantive rights consistently with the policies declared and enforced in
specific rights-based torts.
In land-use disputes, this strategy is implemented not only in one subject-specific tort
(nuisance) but also in the two backstop accident torts, negligence and strict liability.
All three fields qualify the unconsented-entry paradigm to achieve the same goal as
harm-based trespass doctrine. As one nuisance case explained, such qualifications are
“as much for the advantage of one owner as of another; for the very [harm to property]
the one complains of, as the result of the ordinary use of his neighbor’s land, he himself
will create in the ordinary use of his own.”78 To demonstrate this suggestion fully, one
would need to study nuisance,79 land-based strict liability,80 and negligence compre-
hensively. For reasons of space and focus, I will focus primarily on negligence, which
probably seems to most readers the doctrine most inconsistent with my claims thus far.

B. Harm and property in negligence

Negligence seems to confound my hypothesis that property torts are protective.
Trespass is a strict and rights-based tort, but a landowner who suffers property damage

78
Bamford v Turnley [1862] 122 ER 27, 33, Exch. (opinion of Bramwell, L.).
79
See Claeys, “Jefferson Meets Coase: Land-Use Torts, Law and Economics, and Natural Property Right-
s”(note 2) at 1419–23; Richard A. Epstein, “Nuisance Law: Corrective Justice and Its Utilitarian Constraints,” 8
J. Leg. Stud. 49 (1978), 74–94.
80
See Spano v Perini Corp. 250 NE 2d 31 (NY Ct of Apps, 1969); Fletcher v Rylands [1865] 159 ER 737,
Exch. Ct., rev’d, L.R. 1 Ex. 265 [1866], aff ’d sub nom Rylands v Fletcher [1868] L.R. H.L. 330; Hay v Cohoes, 2
NY 159, 161 (NY Ct of Apps, 1849).
142 ERIC R . CLAEYS

not caused by a trespass must prove actual harm81 and the defendant’s negligence.82
Trespass strongly presumes that the victims of ongoing encroachments deserve protec-
tion by injunctions.83 By contrast when landowners are exposed to negligently caused
risks that have not yet generated actual harm, tort and remedies law both dispel the
“suggestion that [a] plaintiff [may] enjoin future acts of negligence by defendants.”84 Is
negligence really protecting property rights if it protects owners’ control with rights
more qualified and remedies weaker than in trespass?
Negligence reasonably adapts control rights to the act-situations it covers. In land-
use disputes, negligence serves as the last possible backstop. Trespass regulates risks of
accidents created by entries by persons or large objects. Nuisance regulates risks
created by non-trespassory but still regularly-invasive pollution. Traditional land-
based strict liability regulates risks arising from trespassory explosions, flooding, and
other severe accidents, when these risks are greater than the risks incidentally created
by beneficial land uses common in the locale. If both parties’ activities are generally
legitimate and beneficial, and if neither activity imposes on others unusual risks of
accident for the neighborhood, all local owners’ use interests are enlarged if they are all
required to sacrifice the power to get prophylactic injunctive protection against any
possible property damage. As one leading nuisance case explained:
[I]n a case of conflicting rights, where neither party can enjoy his own without in some
measure restricting the liberty of the other in the use of property, the law must make the
best arrangement it can between the contending parties, with a view to preserving to each
one the largest measure of liberty possible under the circumstances.85

In this and other similar cases, courts rely on property-based norms to specify and
limit the remedies available to plaintiffs with prima facie nuisance claims.86 To
implement parallel substantive property prescriptions in negligence (and also strict
liability),87 tort institutes a harm requirement, and the law of remedies institutes a
general presumption against prophylactic injunctive relief.
Some readers may find it incoherent or damning for a theory of moral rights to limit
positive-law rights to accommodate “broad considerations of public welfare” or “courts’
institutional capacities and constraints.”88 As the Introduction suggested, and my
interpretation of trespass thus far has confirmed, tort’s basic structure is an accommo-
dation to courts’ institutional capacities; tort vindicates general rights most effectively by
focusing on correcting specific wrongs. As for the relation between moral rights and

81
In negligence, nuisance, or strict liability.
82
If neither nuisance nor land-related theories of strict liability apply.
83
See, e.g., Baker v Howard Cnty. Hunt 188 A 223 (Md Ct of Apps, 1936).
84
Geller v Brownstone Condominium Ass’n 402 NE2d 807, 810 (Ill App, 1980).
85
Madison v Ducktown Sulphur, Copper & Iron Co. 83 SW 658, 667 (Tenn S Ct, 1904).
86
See Eric R. Claeys, “Exclusion and Exclusivity in Gridlock,” 53 Ariz. L. Rev. 9 (2011), 36–43; Jules
L. Coleman and Jody Kraus, “Rethinking the Legal Theory of Rights,” 95 Yale L.J. 1335 (1985), 1340–52.
87
On strict liability, see Gregory C. Keating, “Strict Liability Wrongs,” Chapter 14 in this volume.
88
Dan Priel, “That Can’t Be Rights,” 2 Jurisprudence 227 (2011), 235–6 (book review).
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 143

social welfare: respectable contemporary political theorists find it “irrational” and

“crazy” to suggest that a moral theory could justify a claim that we have certain rights
without considering the consequences the existence of such rights would entail.89 Now,
labor-based morality reconciles labor-based rights to a relatively focused understanding
of public welfare—to all citizens’ enjoying “the largest measure of liberty possible in the
circumstances.” That focus may make labor-based morality unattractive—but it does
not make it incoherent. And the same may be said of many other theories of rights.
Separately, some readers may believe that negligence gives negligent tortfeasors an
entitlement to pollute.90 This view makes a conceptual mistake, for it “completely
misrepresents the normative guidance” of negligence.91 In a land-use dispute, even if
the defendant’s primary activity was otherwise morally legitimate, a judgment of
negligence declares that he commits a wrong by having managed carelessly a risk of
accident that caused a neighbor actual damage. Negligence refrains from awarding
prophylactic protection or stronger damages to facilitate the free exercise of property
rights. Since the defendant is adjudged “negligent,” however, he is condemned. The
limits on remedies do not create a liberty to inflict the rights-invasion triggering the
condemnation.

C. Duty and property in negligence

As Parts IV and V showed, trespass’s prima facie elements are parasitic on logically
prior substantive property rights. The same parasitic relation applies in negligence as
well. To be sure, negligence doctrine makes the parasitic relationship harder to see.
Negligence encourages legal decision makers to consider individualized factors too
numerous to consider in detail here. Before decision makers apply negligence prin-
ciples to all of these individualized considerations, however, they must put the
considerations in context within an accountability relationship. In negligence doctrine,
the duty element supplies the focal point where decision makers determine which
accountability relationship to apply. Other disputes will implicate different norms
about bodily safety, accurate reputations, and so forth; property owners’ rights and
duties will parasitize on controlling property norms.
To illustrate, I will focus here on the case law regulating landowners’ duties of care,
primarily from the period when such duties were determined within a labor-based
morality. The relationship between property right and negligence duty is illustrated by
Buch v Amory Manufacturing Co., well-known for its warning that the Good Samaritan
principle does not apply to legal negligence. In Buch, an eight-year-old boy trespassed

89
John Rawls, A Theory of Justice (Cambridge, MA: Belknap Press of Harvard University Press, 1971), 30.
See also Claeys, “Productive Use in Acquisition, Accession, and Labor Theory” (note 12) at 28.
90
See Guido Calabresi and A. Douglas Melamed, “Property Rules, Liability Rules, and Inalienability: One
View of the Cathedral,” 85 Harv. L. Rev. 1089 (1972), 1108–9.
91
J.E. Penner, The Idea of Property in Law (Oxford: Oxford University Press, 1997), 66; Claeys, “Exclusion
and Exclusivity in Gridlock” (note 86) at 37, 40–3; Coleman and Kraus, “Rethinking the Legal Theory of
Rights,” 1358, 1364–5.
144 ERIC R . CLAEYS

onto the premises of a mill where his older brother worked. He crushed his hand in a
machine his brother was trying to teach him to operate.92 Assume, as the New
Hampshire Supreme Court did, that labor- and control-based property norms should
determine the mill’s accountability to Buch. That assumption requires normative
justification; one would need to determine whether it unduly constrains an owner’s
rightful control over land to expect him to anticipate and prevent accidents to trespass-
ers. Assuming that owner control does require immunity from responsibility for
trespassers, negligence concepts closely track property norms. The property norms
entitled Amory Manufacturing to a Hohfeldian liberty generally to direct “the conduct
of [its] business and management” within its own premises. While Buch was trespassing,
it followed, he held a correlative no-right in relation to the possibility that the mill might
ignore his safety. The easiest way to embody those analytical relations in negligence
doctrine is to declare that the mill owes Buch no duty of care. As long as Amory
Manufacturing and its agents did “nothing, let [Buch] entirely alone, in no manner
interfere[d] with him, he [could] have no cause of action against them for any injury that
he may [have] receive[d].”93
The same reasoning shapes defenses to negligence. If a railroad negligently causes
damage to a farm adjacent to its tracks, black-letter law holds that the farm owes no
duty of care to minimize the damage. As a matter of property policy, as long as the
owner of the farm does “not interfere with nor embarrass the operation of the
railroad,” his choice to use the land as a farm remains “of itself a proper use.” If
contributory negligence doctrine imposes a duty on the farmer to take reasonable care
against spark fires, it gives rise to the “anomaly” that “property may be subject to the
servitude of the wrongful use by another of his property.” To implement property
norms, tort doctrine—here, the duty element in contributory negligence—holds that
the landowner is not required “so to use his own property that it may not be injured by
the wrongs of another.”94
Like all of the other examples considered thus far, these negligence doctrines secure
property rights backhandedly. In both Buch and train sparks cases, courts declared
property rights by denying that owners owed tort duties. In both situations, however,
the duty-denials accorded with and effectuated background substantive judgments
about owners’ legitimate rights to determine the uses of their lots.
As trespass law can vary between rights- and harm-based models, so too can
negligence vary the duty it imposes on landowners. In labor-based terms, tort
doctrine may enhance an owner’s likely beneficial uses of his lot if it pre-commits
him to take reasonable care to protect guests accomplishing some goal common to
owner and guest. It is also reasonable to scale the duty of care in proportion to the

92 93
44 A 809, 809–10 (NH S Ct, 1898). 44 A 810.
94
LeRoy Fibre Co. v Chicago, M. & St. P.R. Co. 232 US 340, 349 (1914). Similarly, a defendant prima facie
liable may not argue that the plaintiff assumed a risk of accident precipitated by the defendant’s trespassory
conduct. See Marshall v Ranne 511 SW 2d 255, 260 (Tex S Ct, 1974); Claeys, “Jefferson Meets Coase: Land-Use
Torts, Law and Economics, and Natural Property Rights,” (note 2) at 1393–4, 1417.
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 145

commonality of the goal. These are the judgments implemented (roughly) by the
licensee and invitee categories.95
To be sure, the traditional premises liability categories of trespass, licensee, and
invitee have been repudiated in many situations, especially as it applies to apartments
in urban areas.96 Yet my conceptual insights do not stand or fall with the labor-based
approach. I have focused here on Buch and traditional premises liability because both
implement the same property-based prescriptions about owner control. Assume that
judges believed that owner control should be narrower and non-owner rights of access
and use should be broader. Such judges would embrace a different morality of
property, and they could use negligence to implement that morality.
For example, although Rawls’s two principles of justice almost certainly do not require
any single approach toward property rights, one could interpret a system applying
traditional principles of trespass and modern premises liability as the product of a
Rawlsian compromise. Trespass might reasonably be viewed as implementing Rawls’s
first principle. Trespass gives all owners, no matter how well or poorly advantaged,
wide autonomy and privacy consistent with others’ enjoying the same. In most
premises-liability disputes, however, the defendant owner (a landlord) is usually
more advantaged than the plaintiff (a tenant, or a guest of a tenant). So modern
premises liability might reasonably be viewed as implementing Rawls’s second prin-
ciple, in a repeat act-situation in which the least-advantaged is likely to be the plaintiff
and the better-advantaged the defendant.97
Such a reconciliation of trespass and premises liability is normative, not conceptual—
but it could be implemented in tort consistent with my conceptual account. In
property terms, perhaps tenants’ leasehold rights and guests’ rights of access cannot
be enjoyed meaningfully without landlords’ owing stronger responsibilities than
required at traditional common law to keep apartment buildings reasonably safe.
The duty element in negligence provides the analytical point where tort implements
narrower proprietary conceptions about owner rights and stronger conceptions of
owner proprietary responsibilities. In doing so, the doctrine backhandedly declares
that tenants deserve stronger proprietary rights to use their leaseholds. The doctrine
also declares that tenant guests deserve stronger interests (probably grounded not in
property but in personal safety and locomotion) in visiting apartments free from
dangerous conditions. It is harder to reconcile the property rights an apartment
owner holds under traditional trespass and modern premises liability than it is the
rights of owners in trespass and under Buch. But the difficulty stems in part from the
fact that Rawls’s principles of justice do not apply very determinately to premises

95
See Leffler v Sharp 891 So 2d 152, 157 (Miss S Ct, 2004); Plummer v Dill 31 NE 128, 129 (Mass S Jud Ct,
1892); Addie & Sons, Ltd. v Dumbreck [1929] A.C. 358, 364–5, HL (opinion of Hailsham, L.).
96
See, e.g., Kline v 1500 Massachusetts Ave. Apt. Corp. 439 F 2d 477 (DC Cir Ct, 1970); Rowland v Christian
443 P 2d 561 (Cal S Ct, 1968).
97
Rawls, A Theory of Justice (note 88) at 60–1, 153, 302–3. I interpret Sargent v Ross 308 A 2d 528, 533 (NH
S Ct, 1973), as using such a Rawlsian approach.
146 ERIC R . CLAEYS

liability, and in part from the fact that traditional trespass and modern premises
liability stand in slight tension.

IX. Conclusion
If I were asked whether tort is a law of rights-protection or wrong-correction, I would
probably answer that it is a little bit of both, though slightly more the former.98 Although
that answer sounds weak, I hope my argument in this chapter has proven it right—and
clarified in which senses tort sounds both in rights-protection and wrongs-correction.
On one hand, trespass to land and other related doctrines focus ultimately on securing
to owners their due rights of control. On the other hand, those doctrines are structured
not to declare control rights but to supply rectification for wrongs to those rights. Back
on the first hand, however, the same doctrines do protect control rights—by specifying
the relevant wrongs to prohibit conduct inconsistent with the underlying rights. So then
again on the second hand, positive-law tort prohibitions do declare control rights—by
using trespass, the duty element in negligence, harm requirements, and defenses to
secure to owners the package of claim-rights, duties, liberties, and no-rights they deserve
consistent with the controlling morality of property.
In recent philosophical tort scholarship, this interplay has been noticed, but it has
not received the attention it deserves. I hope that my study here of land-use torts
confirms and clarifies in what precise respects the law of torts protects rights and
corrects wrongs. In particular, I hope that this study has clarified how tort law declares
and implements moral prescriptions about rights. Tort’s corrective functions are
obvious, because torts prohibit and rectify actions declared to be wrongs. Its protective
functions, by contrast, need to be reverse-engineered from many related torts and
defenses. But recall the parable of the car driver who loses his keys on the dark side of a
parking lot but looks for them where the light is. Tort scholars must not study tort’s
normative structure in that spirit.
Finally, although my intended contributions are descriptive and conceptual, I hope
I have at least suggested why it might be reasonable for tort to protect rights by
correcting wrongs. Wrong-correction is certainly not a logical way to protect rights.
Yet wrong-correction almost certainly helps finesse deep debates about whether labor
theory, Rawlsian principles, or other encompassing theories of morality supply the
most satisfying account of rights. Perhaps self-interest gives most citizens the proper
incentives to assert their rights—to complain only when others wrongly interfere with
their free exercise of their rights. Perhaps courts are better at resolving resource
disputes with two parties at a time than they are at predicting systematically who
among a wide range of claimants deserves to use a resource in different situations. And

98
See Claeys, “Jefferson Meets Coase: Land-Use Torts, Law and Economics, and Natural Property Rights,”
(note 2) at 1394–97.
 ON THE “PROPERTY ” AND THE “TORT” IN TRESPASS 147

perhaps most people reason better about particular disputes than they do about the
fundamental opinions they assume and apply when they judge those disputes. And if
one or more of these four assumptions is reasonable, then what seems at first blush a
convoluted process is actually a very elegant accommodation to some basic limitations
on human coordination and reasoning.
 7
Tort Law and Public Functions
Peter Cane*

I. Introduction
In mainstream Anglophone tort theory, both instrumentalist and non-instrumentalist,1
tort law (in its central or focal sense, at least) is explained in terms of interactions
between parties who are “juridically equal” in the sense that they enjoy, by virtue of
their status as persons, equal freedom to pursue their own projects and purposes.2
Relatively little attention is paid to the question of how tort law applies to what I shall
call “relationships of juridical inequality” between citizens (or “private agents”) and

* Australian National University College of Law. I am very grateful to participants in the Oxford Public Law
Discussion Group for stimulating conversation. Special thanks to Roderick Bagshaw, Liz Fisher and Martin
Matthews. François du Bois, Carol Harlow, Nick McBride and Sandy Steel were kind enough to read drafts of
the chapter and each made typically perceptive comments.
1
These terms are not ideal, but they are adequate for present purposes because nothing in the argument
turns on the choice of language to describe the two strands of theorizing.
2
So far as non-instrumentalism is concerned, it may be worthwhile distinguishing theory rooted in the civil
law tradition from that rooted in the common law tradition. Although I think the statement in the text is
essentially true of both, it may be true for different reasons. By “theory in the civil law tradition” I am referring
primarily to thinking inspired by Kant. Theorists of this stripe include Allan Beever, Arthur Ripstein, and
Ernest Weinrib. The distinction between private law and public law is more sharply drawn in civil law than in
the common law. As John Henry Merryman puts it in his classic study, “[i]n private legal relations [including
tort law] the parties are equals and the state is referee. In public legal relations the state was a party, and as
representative of the public interest (and successor to the prince) it was a party superior to the private
individual.” John Henry Merryman, The Civil Law Tradition: An Introduction to the Legal Systems of Europe
and Latin America, 3d ed. (Stanford, CA: Stanford University Press, 2007), 94. In the common law tradition by
contrast (encapsulated most famously by A.V. Dicey in his concept of the rule of law), tort law (i.e., private law)
applies to public officials as such precisely in order to subject the exercise of public power to “the ordinary law”
administered by “the ordinary courts” (rather than a regime of public liability) and in that way to establish a
sort of equality between private citizens and “citizens in uniform” as potential tortfeasors. See John Gardner,
“Criminals in Uniform,” in R.A. Duff et al. (eds.), The Constitution of Criminal Law (Oxford: Oxford
University Press, 2012). In both traditions, a concept of equality underpins tort law. The contrast between
the two traditions can be helpfully understood in terms of the English tendency to think of the public realm as
populated by “real people” rather than in terms of the continental idea of a metaphysical “state.” For an
illuminating general discussion see Janet McLean, Searching for the State in British Legal Thought: Competing
Conceptions of the Public Sphere (Cambridge: Cambridge University Press, 2012), esp. Chapter 7.
 TORT LAW AND PUBLIC FUNCTIONS 149

the state or, more precisely (as I shall explain in Section II), agents performing public
functions (“public agents”).3 The tort law of the theorists is private law, not public law.
Its dramatis personae are juridical equals, and tort rules and principles are understood
as expressing and giving effect to that equality.4 By contrast, because the state exists to
promote the public interest (the collective interests of its members) rather than its own
interests and to that end claims a monopoly of legitimate coercion (or authority) over
its citizens (the individual members of the state) that potentially constrains their
freedom,5 the dramatis personae of public law—the state and its citizens—are jurid-
ically unequal. This juridical inequality has two aspects. On the one hand, unlike
private agents, public agents are obliged (and, therefore, empowered) to pursue and
promote the public interest even at significant expense to the freedom of private agents
to pursue their own projects and purposes. On the other hand, public agents are under
an obligation not to pursue their own projects and purposes or those of any particular
private agent or group.
As a result of its focus on relationships of juridical equality, mainstream tort theory
is constructed as a form or application of moral theory, concerned with interpersonal
rights, obligations and relationships, not as a form or application of political theory,
concerned with the powers and duties of government and the relationship between
government and citizen.6 Whereas the central issue of moral theory is “what we owe to
each other,”7 the central preoccupation of political theory is how to justify the state’s
claim to a monopoly of legitimate coercion. The basic moral theory of legal instru-
mentalism is utilitarianism, while Kantianism is the main inspiration for non-
instrumentalists.
This is not to say that the state plays no part in mainstream tort theory. On the one
side, for instrumentalists, the purposes that tort law serves—compensation, deterrence
and so on—are public purposes. Public institutions (the legislature and the courts)
determine the content of, and play a role in enforcing, the rights conferred and
the obligations imposed to promote those purposes. On the other side, according
to one prominent non-instrumentalist account, courts are needed to “actualize

3
Because my concern in this chapter is with relationships between private agents and public agents, I leave
two issues aside. The first is how relationships between public agents are best understood in terms of juridical
equality and juridical inequality. The other is whether relationships between private agents are always best
understood in terms of juridical equality. The assumption that the law treats all humans (or, at least, those of
full age and capacity) as equals is a reasonable starting point, but it may not be universally valid. For the
purposes of this chapter, I also ignore the distinction between the state (or the Crown or the government) as a
metaphysical entity and officials and agencies of the state. As a result, I say nothing about the distinction
between direct and vicarious liability.
4
For the claim that the morality underlying tort law is similarly equality-based see Stephen Darwall and
Julian Darwall, “Civil Recourse as Mutual Accountability,” 39 Fla. St. L. Rev. 17 (2012).
5
Nicholas Barber, The Constitutional State (Oxford: Oxford University Press, 2010), Chapter 2.
6
The distinction being drawn here between the moral and the political should not be confused with the
Kantian distinction between law (external) and morality (or “ethics”) (internal). See Ernest Weinrib, The Idea
of Private Law (Cambridge, MA: Harvard University Press, 1995), 99. In terms of this latter distinction, the
political and the legal realms are both “external,” as is the moral realm in the sense I am using the term here.
7
Thomas M. Scanlon, What We Owe to Each Other (Cambridge, MA: Belknap Press, 1998).
150 PETER CANE

and guarantee”8 private rights and to make them “public” and “systematic.”9 Accord-
ing to another, tort law is best explained in terms of an obligation resting on the state
“to provide a body of law that defines wrongs and empowers victims of wrongs to
respond to those who have wronged them.”10 In such accounts, public institutions play
a theoretically central role in making and enforcing tort law, but they do not figure as
bearers of rights conferred and obligations imposed by tort law.
And yet, of course, in our (common law) legal world, state institutions are not only
involved in making and enforcing tort law: they are also subject to it. Public agents can
commit, and can be victims of, the very same torts that private agents can commit and
suffer. Public agents can make tort claims against private agents and private agents
can make tort claims against public agents.11 Moreover, in some systems, there are
“wrongs,” conceptualized as “torts,” that only public agents can commit: for instance,
“constitutional torts” in US law, and the tort of misfeasance in public office in the law
of England and other common law jurisdictions.12
The purpose of this chapter is to explore the theoretical implications of the fact that
although the central or focal application of tort law, being private law, is to relation-
ships of juridical equality, it also applies in a secondary or non-focal way to relation-
ships of juridical inequality. My basic argument will be that in order to explain the
operation of tort law between private agents on the one side and public agents on the
other—with respect to relationships of juridical inequality—we need a theoretical
framework different from that used to explain tort law as it applies to relationships
of juridical equality. I will suggest how such a theory might be constructed in non-
instrumental terms of the “justice” of the relationship between citizens and the state,
and in instrumental terms of the contribution tort law can make to the legitimation of

8
Ernest Weinrib, “Private Law and Public Right,” 61 U.T.L.J 191 (2011), 195.
9
Weinrib, “Private Law and Public Right” (note 8) at 196–7. See also Arthur Ripstein, Force and Freedom:
Kant’s Legal and Political Philosophy (Cambridge, MA: Harvard University Press, 2009), esp. Chapter 6. The
basic argument is that because living in society requires individuals to accept limitations on their freedom to
pursue their own projects and purposes for the sake of the equivalent freedom of others, public institutions
(particularly courts) are needed to provide individuals with the assurance that if they observe publicly
announced, clear and specific limits on their freedom, others will (or will be forced to) observe equivalent
limits on their freedom.
10
John Goldberg and Benjamin Zipursky, “Rights and Responsibilities in the Law of Torts,” in Donal Nolan
and Andrew Robertson (eds.), Rights and Private Law (Oxford: Hart Publishing, 2012), 268.
11
Historically, in English law, the “Crown” was absolutely immune from tort liability—i.e., tort law did not
apply to the Crown. Public officials and agencies did not enjoy such immunity. Public officials and agencies still
benefit from various limited immunities—more in some jurisdictions than in others. However, it does not
follow from the fact that tort law applies to state agencies and officials that it applies to them in the same way as
it applies to private agents.
12
In both doctrinal and theoretical discussions of the application of tort law to state agencies and officials,
such entities typically figure only as bearers of obligations and not also as beneficiaries of rights. The fact that
they are both is a corollary of the fact that tort law is private law, concerned (in its central application) with
relationships between juridical equals who necessarily have mutual (as well as correlative) rights and obliga-
tions. This might justify the conclusion that “torts” that can be committed only by state institutions and
officials—i.e., torts that impose obligations on them but give them no rights—are “torts” in only a non-central
sense.
 TORT LAW AND PUBLIC FUNCTIONS 151

the state’s coercive power. In a common way of thinking, states exist to secure for their
citizens as a whole benefits that could not be secured at all or, at least, not as well
without government. In this sense, states pursue public projects and purposes, and a
public interest, which are distinguishable from and may conflict with the personal
projects, purposes and interests of particular private agents. To be justified within this
framework of thought, the state’s claimed monopoly of legitimate coercion must be
used (only) in the public interest and to promote public purposes, but consistently
with preserving for private agents the largest possible measure of freedom to pursue
their own projects and purposes. Tort law can contribute to the legitimation of state
power in two ways: positively by securing to the state entitlements that it needs to
enable it to pursue the public interest, and negatively by constraining abuse of its
coercive power to the detriment of citizens.
Section II of the chapter explores in more detail the nature (or, perhaps, the source)
of relationships of inequality. Section III considers a possible response to the claim that
in order to explain the application of tort law to relationships of juridical inequality we
need a theoretical framework different from that which explains its application to
relationships of juridical equality, namely that while we need a theory of tort-like
liability tailored to relationships of juridical inequality, it will not be a theory of tort
law because juridical equality is intrinsic to the concept of tort law. This section
concludes (1) that mainstream tort theorists in fact accept that tort law, as traditionally
understood, applies to relationships of juridical inequality but also (2) that they
provide no satisfactory theoretical explanation of this feature of tort law. Section IV
explores how non-instrumentalist and instrumentalist explanations of the application
of tort law to relationships of juridical inequality might be developed. Section V is a
very brief conclusion.

II. Public Institutions and Public Functions

A common way of framing the topic of this chapter is in terms of the tort liability of
“public authorities” or “the state” or “the government.” However, in the past thirty
years or so, in response to changes in public administration, judges and legal scholars
have developed a much improved understanding and account of the role of non-
governmental entities in performing “public functions.” A public function may be
defined as one that is to be performed on behalf and in the interests of the public
(in the sense of society as a whole) rather than on the functionary’s own behalf (self-
interestedly), or (in a partisan way) on behalf and in the interests of any particular
individual or group within society. Whether or not any particular function is public in
this sense is a normative question that need not be addressed here. This new under-
standing of the nature of “governance” has obvious implications for thinking about the
juridical relationship between the citizen and the state. When non-governmental
entities perform public functions vis-à-vis citizens they exercise public (i.e., the state’s)
power and, for that reason, are in a relationship of juridical inequality with those
152 PETER CANE

citizens. The source of inequality is not that the exercise of power by the non-
governmental entity performing the public function is underwritten by the state’s
claimed monopoly of legitimate force, because exercises by citizens of private legal
powers (such as the power to make contracts) are also underwritten in this way.
Rather, the source of inequality is the fact that public power is being exercised over
citizens.
Another corollary of the distinction between institutions and functions is that state
entities may exercise powers that are conceptualized as being private, such as the
power to make contracts. Common law systems do not distinguish categorically
between contracting power exercised by citizens and contracting power exercised by
state entities. The power to contract is private in the sense that citizens are free to
exercise the power in pursuit of their own projects and purposes: it is not a power that
must be exercised in the public interest. In this sense, just as non-governmental entities
may perform public functions, so governmental entities may exercise private powers.
Does it follow from the premise that the relationship between two non-governmen-
tal entities will be juridically unequal when one is performing a public function, that
the relationship between a private agent and a state entity will be juridically unequal
only when the state entity is performing a public function and not when it is exercising
a private power? There are two different approaches to answering this question. In one
view, everything that state entities do is public because in everything they do, their
obligation is to serve the public interest: state entities have no interests of their own
that they may legitimately pursue and promote, whether by exercising public power or
private power. According to this approach, in all their legal interactions the state and
its citizens are juridically unequal. In a different view, the state has—as it were—a
“private life.” In this line of thought, for instance, a public official driving a publicly
owned car from A to B on public business would typically be engaged in a private, not
a public, activity. In driving the car, the official would not be required to pursue public
purposes and promote public interest, and would not be permitted to claim that they
were doing so, but would be deemed to be a private agent. In such cases, public officials
would interact on equal terms with private agents. For present purposes it is not
necessary to pursue this difference of opinion further because it is independent of the
issue of whether tort law can apply between state entities and private agents interact-
ing as juridical unequals—i.e., in cases where the state entity is performing a public
function and the private agent is pursuing their own personal projects and purposes.
The upshot of the “functional turn” is that the issue addressed in this chapter should
be framed not in terms of the application of tort law to “public authorities” or “the
state” or “the government,” but rather in terms of its application to the performance of
public functions. This is because—as the new understanding makes clear—what
generates juridical inequality is not the fact that the state interacts with its citizens
but rather that entities (whether governmental or non-governmental) performing
public functions (public agents) interact with private entities pursuing their own
projects and purposes (private agents). Nevertheless, regardless of whether the issue
is framed in terms of agents or in terms of functions, projects and purposes, we need
 TORT LAW AND PUBLIC FUNCTIONS 153

an explanation of tort law as it applies to relationships of juridical inequality different

from the explanation of tort law as it applies to relationships of juridical equality.

III. Demarcating the Province of Tort Law

A. Non-instrumentalism
An obvious response to this claim might be to say that tort law just is a body of law that
applies to relationships of equality between agents pursuing their own projects and
purposes; and that whatever the law that applies to relationships of inequality might
be, it is not tort law. Even if we need a theory of “liability law” as it applies to relationships
of inequality, it will not be a theory of tort law. Perhaps surprisingly, however, this is not
the response implicit in the accounts of prominent non-instrumentalist tort theorists.
For instance, Ernest Weinrib gives the example of negligence claims against “public
authorities or public champions [who] cause injury while attending to emergencies.”13
He accepts that in assessing the reasonableness of the defendant’s conduct, the fact
that D was “charged with responsibility of acting for the public good” may be taken
into account, but suggests that “the social role of the defendant moves the litigation
closer to the judicial review of administrative action where a court might be properly
deferential about substituting its assessment for that of officers with specialized
expertise.”14 Weinrib does not say, however, that the defendant’s special position
moves the claim outside private law into public law. Even more pertinent is Weinrib’s
discussion of the application to public authorities of the principle that tort law does
not impose duties to confer benefits or prevent harm. For Weinrib (as for many
corrective justice theorists), this fundamental principle follows from the juridical
equality of the dramatis personae of tort law as agents entitled to pursue their own
projects and purposes. However, Weinrib accepts that a statutory provision that
imposed on a public authority a duty to confer benefits on individual citizens or to
protect individual citizens from harm would be actionable in tort. This, he says,
“reflects the difference between the juridical standing of private parties, whose freedom
would be infringed by the coercion of a benefit, and that of public authorities, which
exist for the public good.”15
For Robert Stevens, a tort is an infringement of a right that one person has against
another and a breach of a duty correlative to that right.16 In his theory, infringement of
a right recognized by tort law is a tort regardless of whether the infringer is a private
agent or a public agent. Moreover, the fact that the alleged infringer is a state entity,
or that the alleged infringement arose out of the performance of a public function,

13
Weinrib, The Idea of Private Law (note 6) at 151, n. 12.
14
Weinrib, The Idea of Private Law (note 6) at 151, n. 12.
15
Weinrib, The Idea of Private Law (note 6) at 153, n. 16.
16
Robert Stevens, Torts and Rights (Oxford: Oxford University Press, 2007).
154 PETER CANE

is irrelevant to determining whether the agent’s conduct infringed the right.17 In

Stevens’s scheme, for instance, if, in exercising a public power (as opposed to per-
forming a public duty) a public agent infringes a right recognized by tort law, the
action will be tortious regardless of whether the power was exercised legally or illegally
(intra vires or ultra vires) according to principles of (public) administrative law. On
the other hand, however, statutory and constitutional provisions may impose on
public agents duties (for instance, to confer benefits or prevent harm) that private
agents do not have and that are correlative to rights, infringement of which can be
tortious.18 Again, public agents may enjoy statutory protection from tort liability that
private agents do not.19 Even more significantly, Stevens accepts that “deliberate
infliction of loss, absent violation of a right” by a public agent may be actionable in
tort as misfeasance in public office.20
Like Stevens, Allan Beever finds the essence of tort law (or, at least, of negligence
law) in the infringement of rights. However, unlike Stevens, he sees no problem in
importing administrative law concepts (and with them, considerations of public
interest) into tort law. If a statute gives “a public authority discretion to act within a
certain sphere,” he says, “and the public authority was acting within that sphere, then
it cannot be liable . . . This matter is appropriately settled in accordance with the rules
of administrative law.”21 Like Stevens, Beever thinks that breaches of statutory and
constitutional provisions that confer rights on individual citizens and correlative
duties on the state are actionable as torts even if the right is to a benefit or to protection
from harm. On the other hand, absent such a provision, he seems to think, positive
legal obligations of the state towards its citizens are in some sense inconsistent with
“the structure of tort law.”22
In their different ways, Weinrib, Stevens, and Beever are all non-instrumentalists in
the sense that they think that legal categories—such as private law, tort law and
negligence law—are best understood in terms of their juridical features or components
rather than their functions, purposes or goals. Each attributes to individual rights and
some concept of equality a central role in explaining the juridical features of the legal
categories they respectively choose to analyze;23 and their respective accounts all rest

17
Stevens, Torts and Rights (note 16) at 229–30.
18
Stevens, Torts and Rights (note 16) at 219–20.
19
Stevens, Torts and Rights (note 16) at 225–8.
20
Stevens, Torts and Rights (note 16) at 242. For a very brief account of this tort in English law, see Peter
Cane, Administrative Law, 5th ed. (Oxford: Oxford University Press, 2011), 218–19. However, some theorists
are more troubled than Stevens about the status of this cause of action as a tort. See, e.g., John Murphy,
“Misfeasance in a Public Office: A Tort Law Misfit?,” 32 O.J.L.S. 51 (2012); Erika Chamberlain, “Misfeasance in
a Public Office: A Justifiable Anomaly within the Rights-Based Approach?,” in Donal Nolan and Andrew
Robertson (eds.), Rights and Private Law (Oxford: Hart Publishing, 2012). See also note 12.
21
Allan Beever, Rediscovering the Law of Negligence (Oxford: Hart Publishing, 2007), 335.
22
Beever, Rediscovering the Law of Negligence, (note 21) at 340.
23
The role of equality is particularly clear and prominent in the thinking of Weinrib and other corrective
justice theorists such as Arthur Ripstein, Equality, Responsibility and the Law (Cambridge: Cambridge
University Press, 1999) and Richard Wright (see, e.g., “Substantive Corrective Justice,” 77 Iowa L.R. 625
(1992)). For Weinrib, equality lies at the heart of corrective justice, which in turn explains tort law. See, e.g.,
 TORT LAW AND PUBLIC FUNCTIONS 155

on a public/private distinction of some sort. It is striking, therefore, that in all these

accounts, tort law applies, in certain respects and to a certain extent, to relationships of
juridical inequality between citizens and the state.
Initially, at least, this is particularly surprising in the case of Weinrib, who insists on
quarantining private law (including tort law), which (he says) expresses corrective
justice, from “politics,” which (he claims) expresses distributive justice.24 In Weinrib’s
account, just as oil and water cannot be inter-mixed, so distributive justice and
corrective justice cannot join in a single “juridical ensemble.” They differ in various
ways. Corrective justice pertains to correlative, bilateral relationships between indi-
vidual, juridically equal doers and sufferers of harm. In Ofer Grosskopf ’s terminology,
for Weinrib corrective justice is concerned with vertical equality.25 Distributive justice,
on the other hand, is concerned with horizontal equality—for instance, between
individual sufferers or individual doers of harm. This explains Weinrib’s claim that
explaining tort law in terms of compensation or deterrence treats it as an expression of
distributive justice, not corrective justice: the compensation goal expresses the value of
horizontal equality between sufferers of harm as a group, and the deterrence goal
expresses the value of horizontal equality between doers of harm as a group. Neither
goal joins individual doers and sufferers in a vertical, correlative relationship. By
contrast, “corrective justice holds the parties to the equality inherent in their imme-
diate interaction.”26
Weinrib also claims that whereas doing distributive justice requires a political
choice between many possible schemes of distribution, doing corrective justice
requires no such choice between schemes of correction but only “specification” of
“the meaning of corrective justice with respect to the transaction in question.”27
Because doing distributive justice requires a political choice that affects “the interests
of all members of the community,”28 it is properly the function of politically account-
able state entities “that have the capacity and authority to evaluate the full range of
possible distributions”29 and not of judges and courts. Conversely, because doing
corrective justice requires no such choice, it is properly the function of courts and
judges. Constitutional separation of powers and ideas of the comparative competence
of public institutions are integral to Weinrib’s account of private law.

Weinrib, The Idea of Private Law (note 6) at 57–8. Beever says, for instance, “to achieve justice . . . the law of
negligence requires a standard for judging the behaviour of the defendant that treats the claimant and the
defendant as equals.” Beever, Rediscovering the Law of Negligence (note 21) at 81. For Stevens, the rights that
tort law protects are grounded in “morality” and in determining what they are, “the starting point is the
negative formulation of the golden rule. We should not do unto others what we would not want done unto
ourselves.” Stevens, Torts and Rights (note 16) at 332.
24
Weinrib, The Idea of Private Law (note 6) at 206–14.
25
Ofer Grosskopf, “Horizontal Equality and the Law of Torts,” in Daniel Friedmann and Daphne Barak-
Erez (eds.), Human Rights in Private Law (Oxford: Hart Publishing, 2001), 357.
26
Weinrib, The Idea of Private Law (note 6) at 212.
27
Weinrib, The Idea of Private Law (note 6) at 212.
28
Weinrib, The Idea of Private Law (note 6) at 211.
29
Weinrib, The Idea of Private Law (note 6) at 211.
156 PETER CANE

Although Weinrib himself does not link his discussion of the relationship between
corrective and distributive justice to his brief comments about tort claims against
public agents, others have done so. In an admirably subtle discussion of the tort
liability of public authorities for failure to protect citizens from harm, Hanna Wilberg
argues (albeit somewhat tentatively) that although the common law (i.e., judges) may
properly impose on public agents obligations of protection that private agents do not
have (primarily on the basis that public agents may have powers that private agents
lack), judges should stop short of imposing on them duties “to the world at large,”
owed to individuals “simply as members of the public.”30 Instead, as in the case of
private agents pursuing their own projects and purposes, the common law should
recognize (or create) obligations of protection owed by public agents only when there
is some special relationship between the public agent and one particular individual
who needs protection or a “definable and foreseeable group”31 of such individuals.
Unlike Weinrib, Wilberg does not think that distributive justice and corrective
justice are incompatible or categorically distinct, or that in doing corrective justice
judges do not (have to) choose between different “schemes” of correction. However,
for much the same reasons as Weinrib gives for limiting courts to doing corrective
justice, Wilberg argues that recognition of duties to protect owed “to the world
at large” (as opposed to individuals with which the agent was in some special
relationship) would involve judges in making (impermissible) “wholesale” distribu-
tive decisions rather than (permissible) “retail” distributive decisions.32 If wholesale
duties of protection are to be created, Wilberg argues, this should be done by the
legislature, not the courts.33
Wilberg’s approach explains the inclusion in the province of tort law of “public”
obligations of protection more demanding than those imposed on private agents, but
at the expense of Weinrib’s insistence on the rigid separation of corrective and
distributive justice and only by introducing a difficult distinction between retail and
wholesale distributive decisions.
François du Bois also deploys the distinction between corrective and distributive
justice to argue that obligations to pay damages for breaches of “human rights,” as
such, are not part of tort law but properly belong to “human rights law.”34 Human

30
Hanna Wilberg, “In Defence of the Omissions Rule in Public Authority Negligence Claims,” 19 Torts L. J.
159 (2011), esp. 181ff. I interpret these phrases as referring to duties owed to each and every member of the
public, rather than to duties owed to the public as a whole but to no individual member of it. Wilberg’s
argument that duties to the world at large would be inconsistent with the bilateral structure of tort law perhaps
conflates the two meanings.
31
Wilberg, “In Defence of the Omissions Rule in Public Authority Negligence Claims” (note 30) at 182.
32
Wilberg, “In Defence of the Omissions Rule in Public Authority Negligence Claims” (note 30) at 187.
33
Another argument that Wilberg uses, but Weinrib does not, is that an award of damages is an
inappropriate remedy for breach of duties owed to the world at large because it “directly allocates resources
by ordering payment out of public moneys.” Wilberg, “In Defence of the Omissions Rule in Public Authority
Negligence Claims” (note 30) at 188. For consideration and rejection of this argument, see Peter Cane,
“Damages in Public Law,” 9 Otago L. Rev. 489 (1999).
34
François du Bois, “Human Rights and the Tort Liability of Public Authorities,” 127 Law Q. Rev. 589 (2011).
 TORT LAW AND PUBLIC FUNCTIONS 157

rights are rights of citizens against the state. They give effect to the “special normative
relationship between states and their citizens”35 by justifying the imposition on public
agents of (negative) obligations not to harm, and (positive) obligations to protect from
harm and to confer benefits, which are more onerous and extensive than those
imposed on private agents. The imposition of such obligations, du Bois argues, is
“premised on the notion that membership of a political community entitles individ-
uals to certain services by the state;”36 and it rests on a conception of society “as a joint
enterprise engaged in by its members for their common good.”37 It follows, he argues,
that decisions about who should be entitled to what services, and what the conse-
quences should be of failure to provide the services “concern the management of
society’s resources . . . [and] therefore raise questions of distributive justice.”38
However, for du Bois, this is not a conclusive argument against including the state’s
human rights obligations within the province of tort law because, like Wilberg and
unlike Weinrib, du Bois does not think that distributive justice and corrective justice
are incompatible or that it is inappropriate for courts to make distributive decisions
(or, in other words, to take account of the public interest) in deciding tort claims. The
real problem (for du Bois) with treating human rights obligations (especially positive
obligations) as part of tort law is that they express the relationship of juridical
inequality between citizens and the state, whereas tort law is concerned with inter-
actions between juridical equals.39 On the other hand, he apparently sees no problem
in courts taking account of human rights standards in adjudicating tort claims between
citizens as claimants and public agents as defendants, whether by “reassessing public
interest considerations” relevant to duty of care or standard of liability, or by imposing
a stricter standard of liability.40 The crucial distinction for du Bois, it seems, is between
acknowledging the juridical inequality between citizen and state by taking account of
human rights standards in adjudicating tort claims (permissible) and treating breaches
of human rights standards, per se, as torts (impermissible).
Central to the arguments of both Wilberg and du Bois is the asserted link between
recognition of positive obligations and distributive justice. But to my mind, neither
satisfactorily explains the link. In Weinrib’s theory, at least, corrective justice (as an
account of tort law) has both a structural component and a substantive component.
The structural component is correlativity between the doer and the sufferer of harm,
between the former’s duty and the latter’s right. The substantive component is the
juridical equality of the two parties. As we have seen, Weinrib accepts (as do Wilberg

35
du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34) at 595.
36
du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34) at 595.
37
du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34) at 596.
38
du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34) at 596.
39
du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34) at 602.
40
du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34) at 590–1. See also the
discussion of the horizontal application of human rights in François du Bois, “Social Purposes, Fundamental
Rights and the Judicial Development of Private Law,” in Nolan and Robertson (eds.), Rights and Private Law
(Oxford: Hart Publishing, 2012).
158 PETER CANE

and du Bois) that the juridical inequality of citizen and state is, to a certain extent,
(legitimately) recognized in tort law, for instance in the imposition on public agents of
positive duties that private agents do not have. As a result, to the extent that Wilberg
and du Bois consider such recognition to be inconsistent with corrective justice, they
explain the inconsistency in terms of the structural component of corrective justice.41
However, imposition on a public agent of a positive duty to a private agent is no more
inconsistent with the correlative structure of corrective justice than the imposition of a
positive duty on one private agent to another, or of a negative duty on a public agent to
a private agent. Moreover, this is true even where the duty is owed to all private agents
simply by virtue of their membership of “the public.”
What would be inconsistent with the correlative structure of corrective justice is the
enforcement, through tort law, of a positive (or negative) duty that was owed to the
public as a whole or to some social group but not to any individual member of society
or the group. Suppose, for example, that a court, in response to a tort claim made by an
individual or group of individuals against a public agent for breach of a positive duty
owed to the public at large or to some social group, ordered the public agent to
establish a fund to be distributed amongst the public or group without tying the size of
the fund or the rules of distribution to tort rules and principles governing the
assessment of damages. Such an order certainly would be inconsistent with the
correlative structure of corrective justice and tort law. Moreover, giving effect to
such an order would require decisions about how the fund was to be distributed
amongst individuals, and such decisions would, in a significant sense, be “external” to
the rationale for creating the fund, which would make no reference to individuals.
This, of course, is not the sort of case that Wilberg and du Bois have in mind. Their
objection is to the recognition (or creation) by tort law of certain types of obligations
owed to individuals, and their objection to such obligations must be that they are
inconsistent with the substantive component of corrective justice not (despite what
they say) with its structural component. However, since they (with Weinrib) are
prepared to accept that in certain cases and to a certain extent tort law does regulate
relationships of juridical inequality, the correct conclusion is that corrective justice
understood substantively in terms of relationships of equality does not provide a
complete understanding of tort law. What is missing is an account of the role of tort
law in regulating relationships of inequality.

B. Instrumentalism
For instrumentalists—especially those, such as economic analysts, who attribute to law
one function rather than various functions—distinctions between traditional legal
categories are of empirical rather than theoretical significance. Legal categories, such

41
See Wilberg, “In Defence of the Omissions Rule in Public Authority Negligence Claims” (note 30) at
184–8; du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34) at 603, n. 106 and
accompanying text.
 TORT LAW AND PUBLIC FUNCTIONS 159

as contract law, tort law and criminal law, or private law and public law, represent
different tools or techniques for promoting particular ends, and the relevant question
about the various tools is which promotes those ends most effectively and efficiently.
The main features of the tort law tool are that (1) it creates or recognizes rights and
imposes or recognizes duties, (2) that operate bilaterally, (3) that may be enforceable
by and against strangers (i.e., parties who had no “personal” relationship prior to the
relevant interaction), (4) that need not be bought or paid for, and (5) the main remedy
for interference with which is monetary damages, typically to compensate for harm
inflicted on the claimant but possibly also to capture benefits acquired by the tortfea-
sor, to punish the tortfeasor, or to “vindicate” the claimant’s right. In Calabresi’s
words, “what characterizes tort law is the liability rule . . . tort law and the liability rule
are the middle way between contract law and criminal law/regulation.”42
There has been some debate amongst instrumentalists about the utility and efficacy
of tort law as a tool for regulating relationships of juridical inequality. David Cohen
poses the question of whether “we should transpose institutions and solutions which
provide the framework for private ordering to administratively generated losses.”43
Cohen answers this question negatively for three main reasons. First, he argues, the
starting point of tort law is that agents interact as juridical equals; and it adopts this
starting point even when the parties are juridically unequal—i.e., when the alleged
tortfeasor is a public agent. Second, he says, tort liability is “far less likely”44 to deter
undesirable behavior when imposed on public agents than when it is imposed on
private agents, because the state “can shift costs to the general public through taxation
policy.”45 Third, Cohen argues, “Leviathan is not a monolith, and adjustment policies
[i.e., liability rules] must be sensitive to the enormous range of institutional structures
and programs that characterize the modern . . . state.”46 Tort law’s starting point is that
all agents are equal and equally free to choose projects and purposes. Tort law is
generally insensitive to individual differences of need, skill, capacity, personality and
so on; and to the particular and widely diverse projects and purposes that various
people choose to pursue. For these reasons, Cohen concludes, the “homogeneity” of
tort law should be replaced by a set of “program-specific adjustment policies.”47 He
concedes that the rule-of-law symbolism of applying private law to public agents has
some value. However, he argues that it is outweighed by the huge administrative cost
of the tort system as a means of delivering compensation, its inefficacy as a deterrent
in the case of public agents, and the possibility of promoting rule-of-law values by
other means.
Cohen’s complex arguments can be boiled down to two main propositions. First,
tort law under-deters public agents (as compared with private agents). This claim is

42
Guido Calabresi, “Toward A Unified Theory of Torts,” 1 Journal of Tort Law (2007), Issue 3, p. 1.
43
S. David Cohen, “Suing the State,” 40 U.T.L.J. 630 (1990), 632–3.
44
Cohen, “Suing the State” (note 43) at 647. 45
Cohen, “Suing the State” (note 43).
46
Cohen, “Suing the State” (note 43) at 648. 47
Cohen, “Suing the State” (note 43) at 660.
160 PETER CANE

frequently challenged both by scholars who think that tort law performs a useful
deterrent function in this context48 and, more commonly perhaps, by those who think
that it over-deters public agents and chills promotion of the public interest.49 Second,
Cohen argues that although (as is desirable) tort law compensates only selectively for
“administratively caused losses,” the criteria it uses to decide which losses deserve
compensation are unsatisfactory, partly because they are unclear50 and partly because
they deny compensation in some situations where it should be paid. This claim is
attacked by William Bishop.51 He accepts that the tort rules regulating relationships
of juridical inequality are unclear, but not that they are any less clear than many
other legal rules. He argues against expansion of the damages liability of public
agents to private agents on the basis that this would discourage the latter from
desirable self-protection and also on the ground (somewhat inconsistently) that
there are other legal techniques—notably “validity review”—available to “remedy”52
government illegality.
Standing back from the details of this debate, two points become clearer. First,
Cohen frames his contribution in terms of a choice between two different regimes of
liability rules: tort law or a system of program-sensitive adjustment policies. By
contrast, Bishop frames his contribution in terms of a choice between (or, perhaps
more accurately, a combination of) a liability regime (tort law) and a judicial (“valid-
ity”) review regime. Second, Cohen’s starting point is that the juridical inequality of
private agents and public agents is relevant to the instrumental role of tort law. By
contrast, Bishop denies its relevance: “In the economic stylization tort law is a system
designed to give incentives to both parties to take optimal care. The theory works just
as well in the public as in the private arena.”53 However, Bishop does not deny the
juridical inequality of private agents and public agents or that this is relevant to the
design and evaluation of legal liability regimes. Indeed, he accepts that the issue of why
we have a two-track public law regime of liability incentives and validity review “is an
interesting theoretical question, and why we assign certain tasks to one part of the
system and others to the other is an acute practical question.”54
I conclude, therefore, that for both non-instrumentalists and instrumentalists the
application of tort law (in the former case) and liability law (in the latter) to relation-
ships of juridical inequality raises a theoretical issue separate from that raised by their
application to relationships between juridical equals. To be sure, there is a strong

48
See, e.g., William Bishop, “The Rational Strength of the Private Law Model,” 40 U.T.L.J. 663 (1990).
49
See, e.g., Carol Harlow, State Liability: Tort Law and Beyond (Oxford: Oxford University Press,
2004), 25ff.
50
This argument is made at length in S. David Cohen and J.C. Smith, “Entitlement and the Body Politic:
Rethinking Negligence in Public Law” 64 Can Bar R. 1 (1986). See also Bruce Feldthusen, “Failure to Confer
Discretionary Public Benefits: The Case for Complete Negligence Immunity,” 5 Tort L. Rev. 17 (1997).
51
Bishop, “The Rational Strength of the Private Law Model” (note 48).
52
Bishop, “The Rational Strength of the Private Law Model” (note 48) at 664.
53
Bishop, “The Rational Strength of the Private Law Model” (note 48) at 663.
54
Bishop, “The Rational Strength of the Private Law Model” (note 48) at 664.
 TORT LAW AND PUBLIC FUNCTIONS 161

tradition of opposition to the “public/private distinction” in instrumentalist public law

thinking. However, such opposition seems to rest not on a denial that the relationship
between citizen and state is one of juridical inequality but rather on a particular view
about what the law should do about it.

IV. Theorizing the Tort Law of Relationships

of Juridical Inequality
A. Non-instrumentalist theory
In equality-based non-instrumentalist tort theory, tort law has a structural element
and a substantive element. Structurally, tort law arranges the social world into one-on-
one, bilateral relationships between juridically equal agents. Substantively, tort law
reflects the equality of those agents and gives it juridical expression (in terms of rights
and correlative duties) by striking a balance between the freedom of each to pursue
their own projects and purposes. The role of courts, in adjudicating tort disputes, is to
specify publicly and authoritatively the content of tort law understood as an expression
of the juridical equality of the two disputing agents. In doing so, courts may have to
take account of constitutional or statutory provisions that, for instance, create rights
or impose duties; but they (must) do so consistently with tort law’s nature as an
expression of juridical equality in bilateral relationships.55 In non-instrumentalist
theory of this stripe, the concepts of autonomy and equality are used to explain
various features of tort law and to explain others away.56
As it applies to relationships of juridical inequality, tort law has the same bilateral,
correlative structure it has when it applies to relationships of juridical equality. By
contrast, because the substance of tort law, as understood by mainstream non-instru-
mentalist theory, reflects and gives juridical expression to the relationship between the
two parties, it should (according to the theory) reflect the difference between relation-
ships of juridical equality and relationships of juridical inequality by striking a
different balance (in terms of rights and correlative duties) between the parties to a
relationship of juridical inequality than it strikes between juridical equals.
An analogy may be drawn here with Weinrib and Weinrib’s account of the
“horizontal” impact of human rights on tort law.57 The Weinribs distinguish between
the role of human rights in the juridical regulation of relationships of inequality (or, in
their terms, relationships between citizens and the government) and their role in the
regulation of relationships of equality (or, in their terms, in private law). In the former

55
Laura Weinrib and Ernest Weinrib, “Constitutional Values and Private Law in Canada,” in Friedmann
and Barak-Erez (eds.), Human Rights in Private Law (Oxford: Hart Publishing, 2001). See also du Bois, “Social
Purposes, Fundamental Rights and the Judicial Development of Private Law” (note 40).
56
See, e.g., Weinrib and Weinrib, “Constitutional Values and Private Law in Canada” (note 55) at 69–70.
57
Weinrib and Weinrib, “Constitutional Values and Private Law in Canada” (note 55) at 58.
162 PETER CANE

case, human rights apply directly, whereas in the latter case they apply only indirectly.
The distinction between the two types of case is reflected in a contrast between two
different concepts of proportionality. In cases of direct application, proportionality as
a test of compliance with human rights by the government expresses a balance
between the protected right and the public interest. In cases of indirect application
“a version of proportionality geared to private law is available.”58 Here, the balance to
be struck is between the interests of two juridical equals; but the fact that the interest of
one is a human right (such as freedom of expression) shifts the balance in that party’s
favor. Such cases, the Weinribs say, involve “a balancing process in which a central
aspect of one normative principle [such as freedom of expression] is granted priority
over a comparatively more marginal aspect of another [such as protection of reputa-
tion].”59 In their view, this “private law” concept of proportionality “expresses the
form of transactional equality relevant to the presence of constitutional values.”60
However, because the presence of the constitutional value gives the interest of one
party greater weight than it would have in its absence, we might just as plausibly say
that the constitutional value introduces an element of inequality into the relationship
between the parties. So understood, this account provides a model for explaining non-
instrumentally how tort law might reflect and give juridical expression to the rela-
tionship of inequality between private and public agents.
In the most abstract terms, tort law does this by adopting the misleadingly called
“equality principle.” The equality principle is founded on an understanding of tort law
as centrally concerned with relationships of juridical equality. However, its purpose
and effect (as reflected in the qualification that public agents should be treated in the
same way as private agents “as far as possible”) is to adjust the juridical balance that
tort law strikes between juridical equals to reflect the juridical inequality of private and
public agents. In this way, the equality principle resembles the Weinribs’ private law
version of proportionality. So understood, the equality principle can be contrasted
with what we might call the “immunity principle”—that is, the idea that because tort
law assumes juridical equality, it should not be applied at all to relationships of
juridical inequality.61 The function of both the qualified equality principle and the
private law version of proportionality is to inject into tort law values that are incon-
sistent with its strict construal in terms of juridical equality, and in that way—in the
Weinribs’ graphic but somewhat obscure words—“to soften the hard edges”62 of the
rights that private parties have against one another.

58
Weinrib and Weinrib, “Constitutional Values and Private Law in Canada” (note 55) at 57.
59
Weinrib and Weinrib, “Constitutional Values and Private Law in Canada” (note 55) at 58.
60
Weinrib and Weinrib, “Constitutional Values and Private Law in Canada” (note 55) at 58.
61
In relation to the “constitutional values” discussed by the Weinribs, the equivalent of the immunity
principle would be a principle that the common law should not be developed even indirectly by absorption of
such values.
62
Weinrib and Weinrib, “Constitutional Values and Private Law in Canada” (note 55) at 58.
 TORT LAW AND PUBLIC FUNCTIONS 163

There is another feature of mainstream non-instrumentalist tort theory that should

be noted in this context. Mainstream theorists picture the process of making a tort
claim as an equilateral triangle with two private agents at either end of the base and a
public agent at the apex, mediating between them. By contrast, in cases between
juridically unequal parties, public agents occupy two corners of the triangle, and
only one is occupied by a private agent. This situation creates two contrasting but
related dangers, and to avoid both the court must take due account of the fact that the
parties are in a relationship of juridical inequality. One danger is that the court will
unduly favor the public party to the relationship and give too little weight to the
interests of the private party. The other danger, by contrast, is that the court will
unduly favor the private party and in that way “usurp” the public agent’s responsibility
to decide how to protect and promote the public interest consistently with preserving
for private agents the greatest possible measure of freedom to pursue their own
projects and purposes. Protection against both dangers is found in the constitutional
value of separation of powers; but it is no easy solution, and the various stratagems that
courts have adopted in its name (such as the concept of justiciability, and the policy/
operational distinction) have attracted much scholarly criticism. Nevertheless, the
dangers exist and there is no feasible (or perhaps even imaginable) institutional
arrangement in which they would not inhere.
There is one final issue to be addressed. I have suggested a non-instrumentalist
account of the application of tort law to relationships of inequality by analogy with the
Weinribs’ analysis of the horizontal, indirect application of human and constitutional
rights. However, there are “public torts” that are better analogized to the vertical, direct
application of human rights: most obviously “constitutional torts” in the US, the
“statutory tort” of breach of Convention rights under s 7 of the Human Rights Act
1998 (UK) (“HRA”), and the common law tort of misfeasance in public office. Such
torts may protect rights that are also protected by “traditional” tort law—that is, tort
law understood as an expression of juridical equality. Nevertheless, such torts reflect
and give juridical expression to the relationship of inequality between citizen and state
directly, not indirectly. This distinction between direct and indirect expressions of
juridical inequality is consistent with the fact that public torts impose duties but confer
no rights on public agents vis-à-vis private agents, whereas under traditional tort law
public agents can have rights as well as duties vis-à-vis private agents. It also explains
why “just satisfaction” rather than compensation is the basic remedy under s 8 of the
HRA.63 Some versions of non-instrumentalism, such as “rights” theory, may be able to
accommodate (some) public torts, but it is hard to see how juridical-equality theories
could.64

63
Cane, “Administrative Law” (note 20) at 222–3.
64
du Bois, “Human Rights and the Tort Liability of Public Authorities” (note 34). However, the Weinribs
think that they can explain how new “traditional torts” could be created indirectly. See Weinrib and Weinrib,
“Constitutional Values and Private Law in Canada” (note 55) at 64–5.
164 PETER CANE

B. Instrumentalist theory
The goals that instrumentalists most commonly attribute to tort law are compensa-
tion, deterrence, loss distribution and horizontal equality (between men and women,
for instance, or between races). All of these goals are, in Bishop’s words, “abstract and
neutral as between the private or public identity of the two central actors in the [tort]
drama: the injurer and the victim.”65 Bishop made this statement with reference to
“the concepts in which [tort law] is formulated;”66 and it is true, by and large,67 not
only of tort law but of other legal tools that can be used to promote these various goals,
such as contract law and criminal law—in between which, it will be recalled, Calabresi
placed tort law in the legal economy of the “mixed society.”68
There is also a body of law that is not neutral as between public and private agents,
namely administrative law; and associated with administrative law is a goal particu-
larly appropriate to relationships of juridical inequality, namely accountability. The
concept of accountability may provide an attractive theoretical framework for thinking
instrumentally about relationships of juridical inequality because it addresses the
fundamental political question of how public power can be justified and legitimized.
Central to liberal democracy is the idea that the exercise of public power should, to a
significant and meaningful extent, be under the control of the people—individual
citizens, groups of citizens and the citizen body as a whole. Administrative law
provides a tool—or perhaps a toolbox—that the people can use to exercise a degree
of control over the exercise of public power. Tort law, as it applies to relationships of
juridical inequality, can be understood as providing another tool or set of tools for
holding public agents accountable for the exercise of public power.
As an accountability tool, tort law differs in significant ways from administrative
law. First, it is available for use only by individual private agents. By contrast, adminis-
trative law may also be available for use by unincorporated groups or even by the
public at large (as represented by an individual or group, or a public official such as an
attorney-general or an equality commission). Second, administrative law remedies are
different from those provided by tort law: not monetary awards but rather the undoing
of what has been done, possibly coupled with an order to do it again (to get it right);
and orders to do X or not to do X. Indeed, the characteristic administrative law
remedy—undoing coupled with an order to do again—is unknown to tort law; and
traditionally, at least, the characteristic tort remedy—monetary compensation—was
unknown to administrative law. Third, administrative law imposes on public agents
“public” duties—of procedural fairness, for instance—that are unknown to tort

65
Bishop, “The Rational Strength of the Private Law Model” (note 48) at 663.
66
Bishop, “The Rational Strength of the Private Law Model” (note 48) at 663.
67
But not completely: witness, e.g., misfeasance in public office and criminal liability for abuse of
public power.
68
Calabresi, “Toward a Unified Theory of Torts” (note 42). See also Guido Calabresi, “Torts—The Law of
the Mixed Society,” 56 Texas L. Rev. 519 (1978).
 TORT LAW AND PUBLIC FUNCTIONS 165

law; and it creates “public” wrongs (illegalities)—such as error of law—that are

unknown to tort law.
Such differences are grist to the instrumentalist’s mill because they suggest an
understanding of tort law in terms of the distinctive contribution it can make to the
project of holding public power to account, of making public agents answerable to the
people and in that way, giving citizens a significant degree of control over the exercise
of public power. That contribution can be described in Calabresian terms as a
“liability” regime, and contrasted with administrative law’s regime of (what we
might call) “civil (i.e., ‘non-criminal’) regulation.” Whereas a regime of civil regulation
tells public agents what they must do, must not do or must do again, a liability regime
redresses harm caused by breaches of duties and interferences with rights. Despite the
fact that the two regimes contribute to accountability in quite different ways, the
traditional common law rule was that monetary compensation was not available to
redress harms caused by public law wrongs and breaches of public law duties even in
cases where the wrong consisted of interference with a right or the duty was correlative
to a right.
In this respect, public “civil” wrongs are analogous to criminal wrongs. Tradition-
ally, criminal wrongs do not, as such, attract compensation, but only in cases where the
criminal wrong is also a private wrong. Elsewhere, I have explained this in terms of a
distinction between a criminal model of responsibility and a civil model.69 The
criminal model, much more than the civil model, focuses on the agent. Victims are
marginalized in criminal law and the criminal process. The classic instrumentalist
explanation and justification of a regime of criminal law is protection not of individual
members of the public but of the public and the public interest. Although harm to an
individual victim is an element of the actus reus of many crimes, there are victimless
crimes, and attempting a crime can itself be a crime. The classic instrumentalist
explanation and justification for criminal punishment is general and specific deter-
rence. To the extent that pursuit of this goal is qualified in the name of justice, it is
justice to the criminal, not to victims.
By contrast, the civil model of responsibility focuses on victims as much as on
agents. Tort law’s bilateral structure and its substantive role of expressing and giving
effect to the relationship between the two parties necessarily qualify its pursuit of the
public interest. There are no victimless torts, and attempting a tort is not itself a tort.
The primary function of tort remedies is compensation for harm. Non-compensatory
monetary tort remedies are exceptional and more-or-less controversial, and
non-monetary tort remedies are exceptional and normally linked to the inadequacy
of monetary remedies. The weakness of austerely deterrence-focused (Posnerian) instru-
mentalist accounts of tort law is that they do not recognize the independent value of the
civil model of responsibility and compensation as a goal. On the other hand, the
weakness of severely formal, justice-based accounts of tort law is that they fail to

69
Peter Cane, Responsibility in Law and Morality (Oxford: Hart Publishing, 2002).
166 PETER CANE

recognize that tort law serves the public interest as well as the interests of individuals
and to that end may pursue goals “external” to the relationship between the parties.
Administrative law, of course, is neither private law nor criminal law. It is not
criminal law because public law remedies are not punishments. Rather they express
and give effect to the principle that underpins the legal accountability of public agents
to private agents both individually and collectively—namely that everything that
public agents do in their capacity as such must be positively authorized by law (as
opposed, negatively, to being not prohibited). Administrative law remedies provide
individuals, groups and the public with tools for enforcing the “principle of legality”
and, in that way, they satisfy a necessary—but not, of course a sufficient—condition for
the legitimacy of public power.
Administrative law is not private law because it is not a liability regime. Neverthe-
less, there are two quite different ways of thinking about administrative law, which
I shall respectively call the public model and the private model. These two models
differ in respect of the way that administrative law relates to private law. In the private
model, administrative law provides additional protection for interests (“rights”) also
protected by tort law (for instance) by providing non-compensatory remedies for
wrongs not recognized as such by private law (public wrongs)—but does no more. So,
for instance, administrative law protects against procedurally unfair interferences with
private property.70 In the public model, by contrast, administrative law goes further in
two directions. First, it protects interests of the public and of social groups by
providing non-compensatory remedies for public wrongs. Second, by providing such
remedies it also protects individual interests that tort law does not recognize as rights
(that Reich famously called “new property”).71
This latter process of recognizing (“public law”) rights that private agents have
against public agents but not against other private agents (constitutional rights and
human rights, for instance)72 is the mechanism by which the liability-based account-
ability of public agents has been extended in the past fifty years or so. It has also
provided a springboard for suggestions that liability should be used even more widely
as an accountability mechanism.73 However, there is disagreement about whether (or,
perhaps, the extent to which) what we might call “public law liability” for interference
with public law rights, the rationale of which is to hold public agents accountable in
the name of legality (and, perhaps, other norms of “good administration”), should be
assimilated to private law (tort) liability, the prime rationale of which is to compensate
for harm caused by interferences with rights (and breaches of duties).74 So, for
instance, some would argue that monetary rights against public agents should primarily

70
See, e.g., Cooper v Wandsworth Board of Works (1863) 14 C.B.N.S. 180, 143 ER 414.
71
Charles Reich, “The New Property,” 73 Yale L.J. 733 (1964).
72
And which, conversely, public agents do not have against private agents.
73
See, e.g., Tom Cornford, Towards a Public Law of Tort (Aldershot, Hants: Ashgate, 2008).
74
And under which, therefore, public agents may not only owe duties to private agents and be obliged to
respect the rights of private agents but may also have rights against private agents and be owed duties by
private agents.
 TORT LAW AND PUBLIC FUNCTIONS 167

be a matter of bureaucratic accountability (via ombudsmen and administrative—

statutory or non-statutory—compensation schemes) rather than legal accountability
(to courts and other judicial bodies).75
There is also much debate about whether the measure of liability should be the same
when a liability regime is used as a tool of public accountability as when it is used as a
vehicle to allocate interpersonal responsibility. On the one hand, there are those who
argue that the compensatory measure is less likely to deter public agents than to deter
private agents, and that punitive damages should be more readily available against
them.76 On the other hand, some assert to the contrary that public agents are likely to
be over-deterred by standard compensatory liability, and that the prime aim of
monetary awards in a public law liability regime should be neither to compensate
nor to punish but to provide what the European Convention on Human Rights calls
“just satisfaction.” Furthermore, there are debates about whether a public law liability
regime should use damages to “vindicate” rights, independently of either compensa-
tion for harm or punishment for wrongdoing. Such debates can be understood as
being concerned with the respective instrumental merits of a private law liability
regime and a public law liability regime as tools of accountability for the exercise of
public power.
Recall Bishop’s comment that the existence of a liability regime and a civil-regula-
tory regime side-by-side raises an interesting theoretical issue. In his discussion of the
relationship of tort law to contract law on the one hand and criminal law on the other,
Calabresi associates contract law with individualism, criminal law with communitar-
ianism and tort law with a blend of individualist and communitarian values.77 By
analogy, the issue that Bishop raises in relation to accountability for the exercise of
public power might be addressed by associating a private law liability regime with
individual autonomy, a civil regulatory regime with the common good and a public
law liability regime with a blend of the values of individual autonomy and the common
good. Calabresi argues that the respective provinces of tort law, contract law and
criminal law, understood as tools for regulating the pursuit by private agents of their
own projects and purposes, will vary from society to society according to the relative
value each places on the individual and the community respectively. By analogy, it
might be argued that the respective provinces of regimes of private law liability, civil
regulation and public law liability, understood as tools of accountability for the
exercise of public power, will vary from society to society according to the relative
value each places on autonomy and the common good respectively.
The main strengths of an accountability-based theory of the civil liability of public
agents are that it takes seriously the distinction between relationships of juridical

75
For a robust argument along these lines see Harlow, State Liability: Tort Law and Beyond (note 49), esp.
Chapter 3.
76
See, e.g., Carol Harlow, “A Punitive Role for Tort Law?,” in Linda Pearson, Carol Harlow and Michael
Taggart (eds.), Administrative Law in a Changing State: Essays in Honour of Mark Aronson (Oxford: Hart
Publishing, 2008).
77
Calabresi, “Toward a Unified Theory of Torts” (note 42).
168 PETER CANE

equality and relationships of juridical inequality; and that it can easily accommodate
both public law and private law liability regimes. Its main weakness is that it cannot
easily explain the fact that under the private law liability regime of tort law, public
agents can have rights as well as obligations. This would certainly be a problem for a
monistic instrumentalist account in which the law’s only (or, at least, its central and
overriding) goal was accountability, but perhaps not for pluralistic instrumentalist
accounts in which accountability was only one of various goals of liability law.

V. Conclusion
In this chapter, I have argued that mainstream theory lacks an account of the
application of tort law to relationships of juridical inequality between private agents
and public agents. I have suggested ways in which non-instrumentalist theory and
instrumentalist theory respectively might be developed to take account of relationships
of inequality. In the case of non-instrumentalist theory, the key lies in a distinction
between the “justice” of interactions between private agents and the “justice” of
interactions between public and private agents. In the case of instrumentalist theory,
the development involves a shift from thinking about a liability regime as a tool for
providing private agents with incentives to pursue their own projects and purposes in
certain ways and not others, to imagining it as a tool for holding public agents
accountable for the exercise of public power. The accounts I have suggested may or
may not be considered plausible. However, because the common law lacks a categor-
ical, institutionally entrenched distinction between public law and private law, some
such accounts are needed.
 PART II

HARMS, WRONGS,
RESPONSIBILITY,
AND LIABILITY
 8
What Might Have Been
Victor Tadros*

I. Introduction
If one person has wrongfully harmed another, the one often has a duty fully to
compensate the other. My question is: what is full compensation? The answer to this
question is closely related to the answer to another: what is harm? The reason why
these answers are closely related is that the magnitude of full compensation depends
on the magnitude of harm that the person wronged has suffered. Does it solely depend
on the magnitude of the harm suffered? Perhaps not—perhaps compensation must be
paid for harmless wrongdoing. But even if not, it surely depends heavily on it.
For this reason, philosophers of tort ought to attend to the following:
1) What is the currency of harm?
2) What is the measure of harm?
Harm is a subset of the bad things that might happen to a person. In Section I, I will
briefly address that subset. I canvas three views about the currency of harm and
compensation:
The Preference View. Whether E harms or compensates V depends on whether the state
that E causes V to be in is preferred by her.
The Alienation View. Whether E harms or compensates V depends on the relationship(s)
between the state that a person is and her will.
The Wellbeing View. Whether E harms or compensates V depends on the effect of E on V’s
wellbeing, where a person’s wellbeing is not reducible to preference satisfaction.

* Parts of this chapter were presented at the Universities of Manchester and Warwick, the Ethics Group of
the Rutgers Philosophy Department, the political theory seminar at University College London, and the
Philosophical Foundations of Tort conference at the Rutgers Institute for Law and Philosophy. I am grateful
to the participants of these events for all the excellent discussion that was had. I am especially grateful to Kim
Ferzan for her excellent comments, and to Kimberley Brownlee, Matthew Clayton, Christoph Hoerl, Robert
Jubb, Jeff McMahan, John Oberdiek, Bill O’Brian, Jon Quong, Adam Slavny, and Zofia Stemplowska.
172 VICTOR TADROS

The goodness or badness of a state that a person is in is a function of its quality and
duration.1 I will briefly defend the Wellbeing View against its rivals.
My main focus is the measure of harm. Harm comes in degrees. What is the form of
the scale on which these measures are marked? Let wellbeing be the currency of harm.
How do we determine how much E harms V? Is it simply the noncomparative level of
wellbeing that E causes V to be in? Or is the magnitude of harm a comparative matter?
I canvas three views of the measure of harm:
The Noncomparative View. The magnitude of harm that some event, E, causes the victim,
V, depends on the noncomparative badness of the state of affairs that E causes V to be in.
The Temporal View. The magnitude of harm that E causes V depends on the difference
between the state that V was in prior to E and the state that E causes V to be in.
The Counterfactual View. The magnitude of harm that E causes V depends on the
difference between the state of affairs that E causes V to be in and some other state of
affairs that V could have been in.

I defend a version of the Counterfactual View over its rivals. After showing significant
problems with the other two views canvassed, I mount a novel defense of the
Counterfactual View. The Counterfactual View, I argue, is the only view capable of
handling the problem of overdetermination adequately. Let me briefly explain. Whilst
counterfactual accounts of harm are intuitive in standard cases, they appear to be
threatened by the problem of overdetermination. The problem is as follows. A simple
counterfactual view holds:
The Simple Counterfactual View. E harms X only if X is worse off than he would have been
had it not been for E.

The problem of overdetermination demonstrates the falsity of this view.

For example:
Finger. D chops off one of V’s fingers, preempting that finger and another being chopped
off by X.

E harms V even though V would have been harmed to an even greater degree were it
not for E. Hence, E does not render V worse off than he would have been had it not
been for E. In some cases, where E preempts E2, E harms V even though E also
prevents V from suffering a worse harm.
Many people think that the overdetermination problem threatens counterfactual
views.2 On the contrary, I argue. Overdetermination cases provide powerful support

1
See Matthew Hanser, “The Metaphysics of Harm,” 77 Philosophy and Phenomenological Research 432
(2008), for a defense of an interesting event-based view of harm against a state-based view. I do not consider that
here. For criticisms, see Judith Jarvis Thomson, “More on the Metaphysics of Harm,” 82 Philosophy and
Phenomenological Research 436 (2011). For Hanser’s response, see Matthew Hanser, “Still More on the
Metaphysics of Harm,” 82 Philosophy and Phenomenological Research 459 (2011). Further important criticisms
of Hanser’s view are developed in Adam Slavny “Harm” (unpublished manuscript, on file with the author).
2
See, e.g., Judith Jarvis Thomson, The Realm of Rights (Cambridge, MA: Harvard University Press, 1990),
261–2; Seana Shiffrin, “Harm and its Moral Significance,” 18 Legal Theory 357 (2012), 367–8.
 WHAT MIGHT HAVE BEEN 173

for a more complex counterfactual view—one that allows more than one contrast to be
drawn between the actual world and other worlds in which the relevant event did not
occur:
The Complex Counterfactual View. E harms X only if X is worse off than he would have
been in a relevant possible world where E did not occur. Comparison with more than one
possible world may be warranted in a single case, yielding different verdicts about harm
and benefit.

To find out why, read on.

II. The Currency of Harm and Compensation

E harms V only if V is bad for E. However, some things are bad for V but are not
harmful. Harm is a subclass of things that are bad for a person. Accounts of harm can
be criticized for breadth, narrowness, or both.
For example, Joel Feinberg’s view that a person is harmed if her interests are set
back3 is too broad. Some of my interests can be set back without me being harmed.4
For example, I have an interest in not being disrespected, but disrespect is not harmful.
Similarly, I have an interest in others not attempting to harm me but attempts to harm
are not harmful. Were this not so, attempts to harm others would necessarily succeed.5
Of course, disrespecting people and attempting to harm people can cause harm—for
example, they can cause psychological harm. But a theory of harm is concerned with
what is intrinsically harmful, not with what causes harm.
Andrew Simester and Andreas von Hirsch, in contrast, defend a view that is too
narrow. They claim that E harms V only if V’s opportunity to engage in valued
activities and relationships, and to pursue self-chosen goals, is impaired.6 This
implausibly implies that suffering pain is not intrinsically harmful. Furthermore,
neither nonhuman animals nor tiny infants can be harmed on this view.7

A. Rejecting the Preference View

I will offer reasons to support the Wellbeing View over two rivals. The first is the
Preference View. Here is Robert Nozick’s account of compensation:

3
Joel Feinberg, Harm to Others (Oxford: Oxford University Press, 1986), 33.
4
It is also too narrow in excluding pain, nausea, and so on. See Feinberg, Harm to Others (note 3) at 46.
5
I suspect that it follows that the idea that risking harm to V harms V is similarly false. For a defense of the
view that risking harm harms, see Claire Finkelstein, “Is Risk a Harm?,” 151 U. Pa. L. Rev. 963 (2003). For
doubts, see Stephen Perry, “Risk, Harm, and Responsibility,” in David G. Owen (ed.), Philosophical Founda-
tions of Tort Law (Oxford: Oxford University Press, 1995). Of course, attempting harm and risking harm can
be harmful in an incidental way, for example, by frightening people or making them take precautions.
6
See A.P. Simester and Andreas von Hirsch, Crimes, Harms, and Wrongs: On the Principles of Criminal-
isation (Oxford: Hart Publishing, 2011).
7
See Shiffrin, “Harm and its Moral Significance” (note 2) at 370–1.
174 VICTOR TADROS

Something fully compensates a person for a loss if and only if it makes him no worse off
than he otherwise would have been; it compensates person X for person Y’s action A if X is
no worse off receiving it, Y having done A, than X would have been without receiving it if
Y had not done A. (In the terminology of economists, something compensates X for Y’s act
if receiving it leaves X on at least as high an indifference curve as he would have been on,
without it, had Y not so acted.)8

Like me, Nozick defends a version of the Counterfactual View about the measure of
harm. My interest, though, is in another feature of his view. In the passage in brackets,
Nozick implies an idea familiar from economics—that X’s preferences are the currency
of harm and compensation. Whether a person is indifferent between two states of
affairs depends only on her preferences about these states of affairs.
The Preference View is both too broad and too narrow. One problem is that a person
may have preferences about being harmed or benefited. She may prefer to be harmed
or prefer not to be benefited. She may also prefer to avert a smaller harm rather than
receiving a greater benefit or prefer a greater harm to a smaller benefit. Let me
illustrate:
Expiator. Y wrongly chops X’s finger off. Because X believes that he deserves to suffer for
his sins, he prefers to have lost a finger to not having lost it.

The Preference View yields the verdict that Y does not harm X. But even X agrees that
he has been harmed—he just thinks he deserves it. The preference-based view of
compensation implies that X is fully compensated if he is provided with nothing.
X may believe that he doesn’t deserve to be compensated. That does not imply that he
has been, though.
Here is the converse case:
Manicurist. Y damages X’s fingernail and the fingernail cannot be fully repaired. It will
always look odd. X is vain. He would prefer to have beautiful nails to receiving £1,000,000.

The Preference View yields the verdict that X has not been fully compensated if Y gives
him £1,000,000. This seems false. This is not simply because Y ought not to be required
to pay that much compensation to Y. If Y acknowledges, as he may, that his life goes
better with the £1,000,000, he will recognize that he has been fully compensated. He
may prefer to have nice nails to full compensation.
Mistakes about wellbeing pose further problems; consider:
Amnesia. Y crashes into X, causing X to suffer partial amnesia for one year. X is
immediately provided with a very small sum of money in compensation, and is happy
with that sum.

X has been harmed to a much greater degree than she realizes. She wrongly believes
that she has been fully compensated for the harm that she suffers. The Preference View
struggles to explain this verdict.

8
Robert Nozick, Anarchy, State, and Utopia (Oxford: Blackwell, 1974), 57.
 WHAT MIGHT HAVE BEEN 175

Defenders of the Preference View may respond that the time after she has suffered is
not the appropriate time to assess her preferences. However, it is difficult to see how to
set the appropriate time without referring to the objectively correct assessment of
wellbeing. To see this, consider two other times at which we might assess her
preferences: t1 (just before the accident) and t3 (when she has had time to think
about it). The challenge for the Preference View is to pick a time that yields plausible
results without referring to objective criteria.
A person may make mistakes that render her preferences an unreliable guide to how
much she has been harmed at any of these times. For example, before she has been
harmed she may not realize how important a good memory is. She may be the kind of
person who is willing to enter Nozick’s famous Experience Machine,9 which makes a
person happy but makes her forget all of her actual commitments. Now suppose that at
t3 she realizes that she was wrong about this. The tiny amount of compensation
required to satisfy her pre-accident preferences is surely insufficient.
Perhaps her considered preferences provide a better measure. However, the mere
fact that she has considered things for longer or more carefully at t2 than at t1 is not
decisive. To see this, suppose that at t3, after reflecting carefully on things, X resents
Y so badly that she can never accept that she will be fully compensated however well
Y makes her life go. The Preference View implausibly implies that this belief is akin to
a self-fulfilling prophecy. Furthermore, the reason we typically have to respond to a
person’s considered over unconsidered preferences is that by considering things, a
person’s preferences about her life are more likely to align with what is good for her.
That is, after all, what a person is typically trying to do when she considers her self-
regarding preferences.
Perhaps the Preference View can be revised to meet these concerns. For example, we
might consider a person’s ideal preferences—the preferences that she would have were
she fully rational with full information. This version of the preference-based view is an
improvement. However, as preferences change over time, it would still face the
problem of identifying the appropriate moment at which to idealize a person’s
preferences. Overall, I doubt that preferences are very important in determining
whether a person has been harmed or compensated.

B. Rejecting the Alienation View

Seana Shiffrin tentatively defends the Alienation View.10 This is a complex view. On
Shiffrin’s own account, there are various ways in which conflicts between a person’s
state and her will may constitute harm.
Shiffrin’s account is confused in one way that needs addressing before assessing the
Alienation View. Shiffrin presents the Alienation View as a noncomparative account of

9
Nozick, Anarchy, State, and Utopia (note 8) at 42–5.
10
For her outline of this view, see Shiffrin, “Harm and its Moral Significance” (note 2) at 385–97.
176 VICTOR TADROS

harm. But her defense of the Alienation View is independent of her defense of
noncomparativism. The Alienation View is compatible with all three accounts of the
measure of harm described above. The Alienation View is better understood as a view
about the currency of harm.
To see this, notice that counterfactualists and temporalists may endorse the Alien-
ation View. A counterfactualist might claim that E harms V if E causes V to be in a
state where she is more severely alienated from her will than she would have been in
some other possible world. A temporalist might claim that E harms V if E causes V to
be more seriously alienated from the will than she was. I leave assessment of Shiffrin’s
defense of noncomparativism (or rather, her attack on comparativism) to Section II.
Much of Shiffrin’s discussion of alienation of the will is otherwise attractive.
Whether some state of affairs is harmful to a person, and the degree to which it is
harmful, depends at least in part on the relationship between the state that the person
is in and the person’s will. This idea is well captured by the Wellbeing View as well. As
Joseph Raz11 and T.M. Scanlon12 emphasize, successful pursuit of valuable projects is
plausibly an important component of wellbeing. The will is thus naturally important
on the Wellbeing View. The question is whether alienation of the will is either
necessary or sufficient for harm. I have some doubts about both ideas.
Alienation of the will does not seem necessary for harm. Causing pain to newborn
babies or nonhuman animals harms them. That is not in virtue of the relationship
between their will and the pain caused. Furthermore, a person can will herself to be
harmed, for example in the Expiation case, above. Alienation of the will is also
insufficient for harm. Wrongful interference with autonomy is not always, or even
typically, harmful. A person may be wronged when others interfere with her autonomy
even when she is not harmed—even when she is benefited.
We should not morally overburden the concept of harm.13 One reason why Shiffrin
is driven to her account of harm is that she wishes to capture the central role that harm
has in morality. It is true that harm is central to morality, but not every wrong, nor
every value, relates to harm. Shiffrin, I think, tends to give harm a more prominent role
in morality than it has.
The reason why wrongfully alienating a person from her will is not always harmful
is that autonomy is often to be respected even when a person chooses what is not good
for her. Consider Shiffrin’s claim that fraud is necessarily harmful in that it is
“inconsistent with respect for one’s status as an independent autonomous agent.”14
I agree that fraud is typically wrong in virtue of being disrespectful. I doubt that this
shows that it is harmful. Fraud can be wrong even if the victim of the fraud is benefited.

11
Joseph Raz, The Morality of Freedom (Oxford: Oxford University Press, 1986), Chapter 12.
12
T.M. Scanlon, What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998),
Chapter 3.
13
I offer more support for the independent significance of autonomy and harm in Victor Tadros, “Harm,
Sovereignty, and Prohibition,” 17 Legal Theory 35 (2011) and Victor Tadros, “Consent to Harm,” 64 Current
Legal Problems 23 (2011).
14
Shiffrin, “Harm and its Moral Significance” (note 2) at 385.
 WHAT MIGHT HAVE BEEN 177

For example, fraudulently inducing a person to end a bad relationship is wrong even if
the person is benefited. This idea seems difficult to capture on Shiffrin’s view.
Wellbeing, I think, provides the currency of harm.15 Preferences and the will gain
their importance because of their relationship with wellbeing. Wellbeing is probably
best understood as multidimensional. Preferences play at most a minor role in well-
being. The will plays a major role. But wellbeing is not reducible to the relationship
between a person’s state and her will. And the moral importance of a person’s will is
not reducible to its role in securing her wellbeing. Whilst I lack a complete account of
wellbeing, this should not hamper my discussion of the measure of harm. We can
avoid problems about the nature of wellbeing by examining cases where wellbeing is
set back on any plausible account of the currency of harm.

III. Incomparable Problems

Let us evaluate noncomparativism. If wellbeing is the currency of harm and noncom-
parativism is true, to be harmed is to be caused to have a low level of wellbeing.
Noncomparativists need not be anti-relativists about harm. They may believe that
whether having some property puts a person in a harmed state is relative to other
properties of that agent. As Thomson notes, having the mental capacity of a two-year-
old may be a harmed state for an adult, but not for a two-year-old.16 Noncompar-
ativists deny that being harmed is relative to some other state that the person was or
could have been in.
Shiffrin has given the most sustained defense of noncomparativism to date.17 She
defends it mainly by attacking comparativism.18 Here, I outline some familiar and
some new problems for noncomparativism. I then meet Shiffrin’s complaints about
comparativism.
One problem is that noncomparativism cannot explain why making a well-off
person slightly worse off harms her.19 Consider:

15
There may be another dimension to the currency of harm that I lack the space to explore here, which has
to do with the strength of the psychological connections that would obtain between the person as she is at the
moment when she is harmed and her potential future selves. I am unsure whether this is best understood as
part of the currency or the measure of harm. I leave this complex problem aside, but for an excellent discussion
in the context of the badness of death, see Jeff McMahan, The Ethics of Killing: Problems at the Margins of Life
(Oxford: Oxford University Press, 2002), Chapter 2.
16
Thomson, “More on the Metaphysics of Harm” (note 1) at 438.
17
See Seana Shiffrin, “Wrongful Life, Procreative Responsibility, and the Significance of Harm,” 5 Legal
Theory 117 (1999) (as originally stated), and Shiffrin, “Harm and its Moral Significance” (note 2) (as stated
more recently and in a more sustained way).
18
As I suggest above, her positive proposal is not best understood as an alternative to comparative accounts
of harm.
19
See also Hanser, “The Metaphysics of Harm” (note 1).
178 VICTOR TADROS

Novelist. You are writing a brilliant novel. Out of jealousy, I mildly poison you. This
slightly diminishes your talents. The novel you write, though brilliant, is worse than it
would have been.

Surely I have harmed you even though you are noncomparatively well off.
Shiffrin thinks that it is a threat to counterfactual accounts of harm that small losses
to the very well off are harmful. Billionaires are not harmed by losing one thousand
dollars, she claims, even “assuming that the billionaire has a stake in his stockpile, as
many billionaires do.”20 But this seems intuitive only because the loss of one thousand
dollars has no impact on the billionaire’s wellbeing. At most, this case demonstrates
that losing something that one has a stake in is insufficient for harm. That is better
understood as a claim about the currency of harm than about the measure of harm.
Here is an even worse problem for noncomparativists. We have powerful reasons
to benefit those who are very badly off, even if those benefits are very small, and
hence do not lift the person out of a noncomparatively bad state. For example,
providing food to the starving is required even if the person remains sickly after
receiving the food. It is difficult for the noncomparativist to explain the idea that this
person is benefited as she is caused to be in a bad state—her state is bad but better than
it was and would have been. But “better” is a comparative judgment.21
The basic problem with noncomparativism is this: events cause harm. Whether
some event causes harm depends not simply on the state in which the event leaves the
person affected by the event. It depends on whether the event improves or erodes the
person’s quality of life. Shiffrin thinks that it is unclear why comparisons ought to be
central to assessments of harm.22 The answer is that without comparisons, we have not
described harm but something else. We distinguish between whether a person is well
off or badly off, which is a noncomparative matter, and whether a person is harmed or
benefited, which is a comparative matter. As we need both comparative and non-
comparative ideas in descriptive and normative work, we are better retaining a
comparative account of harm, and using other words such as “badly off ” in making
noncomparative assessments.

IV. Meeting Comparative Complaints

Here, I address the four main challenges that Shiffrin mounts to comparativism. First,
comparativism implies that it is possible that A is harmed and B is benefited by some
event, and yet A is left better off than B.23 Shiffrin thinks this is counterintuitive.
Shiffrin is wrong. If a very poor person steals my wallet, she harms me and benefits

20
Shiffrin, “Harm and its Moral Significance” (note 2) at 371.
21
See also Thomson, “More on the Metaphysics of Harm” (note 1) at 441.
22
Shiffrin, “Harm and its Moral Significance” (note 2) at 369.
23
Shiffrin, “Harm and its Moral Significance” (note 2) at 372–3.
 WHAT MIGHT HAVE BEEN 179

herself. That is so even though she remains worse off than me, and even if the stealing
is permissible.
Or consider:
Transplant. A has two arms and two legs. B has no limbs. If I flip a switch an arm will be
transplanted from A to B. No other harms or benefits will be caused.

Shiffrin implies that it is a weakness of comparative accounts of harm that they imply
that flipping the switch harms A and benefits B. This verdict seems intuitive though.
Perhaps Shiffrin would explain this verdict by appealing to the legitimate expect-
ations and the investments that A and B have in their long-term projects. But whilst it
is true that legitimate expectations and investments make a difference to how harmful
some event is, that does not fully explain Transplant. Whether A is harmed does not
depend on his legitimate expectations and projects. Suppose that A and B are newborn
babies or monkeys. They have no projects or expectations. The transplant still harms
A and benefits B.
Second, Shiffrin claims that comparative accounts of harm cannot adequately
capture the idea that a person who has severe disabilities at birth is harmed if those
disabilities, or their cause, alter the personal identity of the person. Non-identity cases
are complex, and they warrant more attention than I give them.
Here are three possible responses. First, we do not need the idea of harm accurately
to characterize non-identity cases where the person is very badly off. A person who is
born with very severe disabilities is born in a very bad state. It adds nothing to this to
say that the person is harmed. That the person is not harmed does not imply that the
person causing her to come into existence owes her no duties. She may be owed duties
simply in virtue of the fact that she is badly off. If we can make adequate evaluative and
normative judgments about non-identity cases without referring to harm, what is
objectionable about the view that these people are not harmed?
A second response is that a comparative account of harm can be supplemented by
noncomparativism for this case. For reasons given above, a noncomparative account
would seem only adequate for this special case.24 Otherwise, it is difficult to avoid the
unwelcome implication that B is harmed in Transplant.
A third response is that there are relevant counterfactual claims that can be made to
yield the verdict that the person born in a very bad state is harmed in virtue of the fact
that she is worse off than she would have been in some relevant possible world. It is
true that the Simple Counterfactual View does not warrant the verdict that the person
is harmed in all non-identity cases. More complex counterfactual may do so.25 Overall,
comparativists have plenty of resources to respond to this complaint.

24
This is more or less the conclusion reached in Jeff McMahan, “Causing People to Exist and Saving
People’s Lives,” Journal of Ethics (forthcoming). See also Thomson, The Realm of Rights (note 2) at 262, n. 7.
25
For a brief discussion of this possibility, see Thomson, “More on the Metaphysics of Harm”(note 1) at
450–3. Thomson plausibly thinks that only some of these cases, cases where the child’s disabilities are genetic,
are non-harming cases. If she is right, the first response is sometimes the right one.
180 VICTOR TADROS

Shiffrin’s third complaint is that comparativism cannot explain the intuitive idea
that causing pain is harmful. As I suggest above, I agree with her that short sharp
shocks are intrinsically harmful. But here again Shiffrin confuses an issue about the
currency of harm with an issue about the measure of harm. If E causes V to suffer pain,
comparativists may claim that E harms V in virtue of the fact that E causes V to be in
pain rather than to be free from pain.
Furthermore, comparative accounts of harm seem better placed than noncompara-
tive accounts to handle the case of pain. If E slightly relieves V’s pain E benefits V, even
if V remains in great pain. It is not clear how noncomparativism can explain this
verdict. Overall, the problems with noncomparativism extend to the problem of pain,
and pain provides no problem for comparativism.
Shiffrin’s fourth complaint is that comparativism, especially the Simple Counterfac-
tual View, cannot adequately explain the fact the moral asymmetry between harming
and failing to benefit. Shiffrin writes that “on these accounts, enduring harm and
not being benefited, for example, suffering an opportunity cost, are identical.”26 This
complaint is unclear to me. Perhaps Shiffrin worries that counterfactualists cannot
explain the moral difference between harming a person and failing to provide her with
a benefit in virtue of the fact that they can offer no reason to pick out any particular
possible state of affairs with which to compare the person’s current state. If this is her
complaint, she might be clearer about why she thinks this.
I doubt the force of this complaint. Consider:
Leg 1: X cuts V’s leg off, leaving V with one. He could easily have cut two legs off.
Leg 2. V has no legs. Y could easily provide V with two prosthetic legs but provides him
only with one.

Perhaps Shiffrin worries that comparativists must treat these cases as morally identi-
cal. In both Leg 1 and Leg 2, we compare the one-legged V with the two-legged V.
But comparativists need not treat these cases alike. For example, the standard
comparison that counterfactualists make is between V’s state and the state that
V would have been in had E not occurred. X harms V even though he could easily
have harmed him more. Y benefits V even though he could have benefited him more.
This helps to explain why X’s wrong is more serious than Y’s. Perhaps Shiffrin thinks
that the decision to set the baseline that this view relies on is arbitrary—why not
compare V’s circumstances with the circumstances that he would have been in had
X cut off both legs? It is true that counterfactualists need to answer that question, and
it is difficult to answer. But until we are shown that it cannot be answered, we lack a
decisive reason to reject the Counterfactual View.
Furthermore, as I noted earlier, we should not morally overburden harm. Here is
another difference between Leg 1 and Leg 2. In Leg 1 X interferes with V in a way that
V would not consent to. That is not true in Leg 2. This fact seems at least as important

26
Shiffrin, “Harm and its Moral Significance” (note 2) at 372.
 WHAT MIGHT HAVE BEEN 181

as the difference between the harm and benefit done. But its significance does not
depend on harm.
In general, as I have already suggested, the wrongness of interfering with a person’s
autonomy does not depend powerfully on the magnitude of harm caused. For
example, the wrongness of killing a person does not depend in a very significant
way on the magnitude of harm that death causes the person. It is typically much more
harmful to die when one is twenty than it is to die when one is seventy. Yet the
wrongness of killing a seventy-year-old is almost as bad as the wrongness of killing a
twenty-year-old.27 The fact that this verdict is not yielded by an account of harm does
not count against that account of harm. Harm simply has a more limited role in
morality than Shiffrin thinks. And finally, given that D is left with a single leg in both
Leg 1 and Leg 2, it seems that it is noncomparativism rather than comparativism that
fails adequately to capture the moral difference between harming and benefiting.
Shiffrin’s attacks on comparativism do not seem decisive. Her noncompara-
tivism faces insurmountable difficulties. I conclude that the best account of harm is
comparative.

V. Why Time Is Not of the Essence

If comparativism is right, “E harms X” implies that E causes X to be worse off than she
would be in some other state. Two comparativist possibilities present themselves—
temporalism and counterfactualism. Temporalism, which I focus on here, is the
view that E harms V (if and)28 only if E causes V to be worse off after E than she
was before E.
Temporalism seemingly has the virtue that it plausibly explains the verdict that
D harms V in Finger. In determining whether E harms V, we determine whether
E caused V to be in a state that is worse than the state that he was in prior to E. This
yields the appropriate verdict in Finger—V is worse off than he was in virtue of D’s
chopping. Furthermore, at first blush it seems to have plausible implications for the
magnitude of harm done in this case. It asks: how much worse off is V without a finger
than he was? Finally, the Temporal View appears to avoid preemption problems. V is
worse off than he was, but not worse off than he would have been but for the finger
chopping. Hence, temporalism seems preferable to counterfactualism.
Nevertheless, the Temporal View faces significant problems. One problem arises
from the fact that the extent to which E harms V depends on the duration of its effects.
It is difficult to find an appropriate comparison to make about duration without

27
See Jeff McMahan, The Ethics of Killing: Problems at the Margins of Life (note 24) 16, at 189–203.
28
This part of temporalism is bracketed for the reason that Thomson, in The Realm of Rights (note 2),
endorses temporalism but rejects the bracketed element. In The Realm of Rights she offers no defense of
temporalism other than a swift rejection of counterfactualism. In “More on the Metaphysics of Harm” she
indicates that she now prefers counterfactualism to temporalism.
182 VICTOR TADROS

appealing to counterfactuals. For example, suppose D negligently causes my mobility

to be restricted for two months. The duration of my immobility is obviously relevant to
how much harm I suffer. How harmful is this? Temporalism invites a comparison
between the immobility that I now have for the duration of two months and the
mobility that I had. To make a comparison, though, we need some duration of the
mobility that I had to compare my two months of immobility with.
The answer cannot be the moment prior to being harmed, for wellbeing without
duration is inconceivable. The most natural answer for temporalists compares the two
months after the harm-causing event with the two months prior to the harm-causing
event. But this answer has unappealing implications. Suppose that I was in a coma for
the two months prior to being harmed. I was not at all mobile for the two months prior
to being harmed. Temporalism implausibly seems to imply that I have not been
harmed. Counterfactual comparison is required to yield the right result: what matters
is how mobile I would have been over the two months following the harm-causing
event.
Relatedly, temporalism seems unable to capture one important way to harm people:
by diminishing their good prospects.29 Similarly, it struggles to explain how amelior-
ating a person’s bad prospects benefits him or her. Consider:
Recovery. A is recovering from a serious illness. His recovery is assisted by the use of a
drug. B tampers with the drug. The drug is less effective in assisting A’s recovery than it
would have been had B not tampered with it.

Intuitively, B harms A. But as A is better off after taking the tampered drug than he was
before taking it, temporalism yields the opposite verdict.
Now consider:
Decline. A has a degenerative disease. He is declining quickly. B gives him a drug to
ameliorate his decline.

Intuitively, B benefits A. But, as A is worse off than he was before being given the drug,
temporalism yields the opposite verdict.

A. Nature and options

The recovery problem has not led all scholars to abandon the Temporal View. Stephen
Perry notices the problem but thinks it nonfatal. He tentatively offers two responses.
Neither seems ideal to me. He considers the problem of a person whose recovery from
injury is interfered with by another person, and agrees that the person is harmed. His
first tentative suggestion is this person is harmed is in virtue of the fact that a natural
process of healing is interfered with.30 This suggestion is insufficiently broad to meet

29
See also Hanser “The Metaphysics of Harm” (note 1).
30
Stephen Perry, “Harm, History, and Counterfactuals,” 40 San Diego L. Rev. 1283 (2003), 1297–8.
 WHAT MIGHT HAVE BEEN 183

the concern. In Recovery, B does not interfere with a natural process, but rather with a
drug-assisted process, yet B harms A.
Perry’s second response is as follows: if a person is denied a benefit that she would
have otherwise received, she is harmed in virtue of the fact that her options are
restricted. To evaluate this response, we need to characterize it more carefully. It is,
of course, true that B denied A the option of recovering. But it is not clear how
referring to A’s options makes a difference here. After all, it is the recovery that A cares
about, not the option of recovering. If B harms A, this is not in virtue of denying A an
option, it is in virtue of interfering with A’s recovery.
I think that Perry means to rely the idea that a person’s options are valuable to a
person in a “selection-independent” way. That is, the value of a person having an
option is not reducible to the benefits to him if he selects the option. A person may be
better off by having an option, even if he does not select it. I think that Perry means
this because it would allow him to point to a pre-event advantage that is removed in
cases such as Recovery. If the option of recovering is valuable to A independently of
whether he actually recovers, B harms A by removing that option.
I agree that options can have selection-independent value.31 Perhaps the most
important reason is that options improve autonomy independently of their selection.
But this explanation why B harms A in Recovery is unappealing. To see this, notice that
A is harmed even if A is unconscious, a newborn baby, or a nonhuman animal in
Recovery. But options lack selection-independent value to unconscious people, new-
born babies, and nonhuman animals.
Moreover, the magnitude of harm that A suffers in Recovery is not determined by
the value of the option of recovering independently of whether A selects that option.
We could consider the selection-independent value of the option by imagining that
A would not have selected the option. The harm that B does to A is more plausibly
calculated, though, by comparing A’s state with his circumstances where he selects the
option. What is most important in recovery is not that A had the option of recovering.
It is that A would have recovered.

B. Prospects
David Velleman develops the Temporal View in a way that has more general applica-
tion, and thus might be preferred.32 Velleman thinks that how well off a person is
depends on his prospects. In Recovery, B harms A, Velleman might suggest, by
diminishing A’s prospects. Similarly, B benefits A in Decline by improving A’s

31
For further discussion, see Victor Tadros, Criminal Responsibility (Oxford: Oxford University Press,
2005), 200–7.
32
J. David Velleman, “Persons in Prospect,” 36 Philosophy and Public Affairs 221 (2008), 242–3. See also
Thomson, “More on the Metaphysics of Harm” (note 1) at 445 (suggesting that the Temporal View might be
rescued from this objection by referring to a person’s chances). I suspect, though I will not demonstrate it here,
that my objections to Velleman extend to Thomson.
184 VICTOR TADROS

prospects. As the unconscious, newborn babies and nonhuman animals have pros-
pects, the challenge outlined to Perry’s proposal can be evaded.
Here is a problem with Velleman’s proposal: when it is adjusted for prospects, the
Temporal View is not obviously distinct from the Counterfactual View, for a person’s
prospects are understood counterfactually. It follows from this that the apparent
advantage that temporalists claim over counterfactualists—that temporalism can
respond more adequately to preemption cases—is lost. To see this, recall Finger.
Had D not chopped off V’s finger, X would have chopped two. If X’s chopping is
part of V’s prospects, Finger is like Decline. As D has improved V’s prospects, does
temporalism not imply implausibly that D has not harmed V? But, intuitively,
D harms V in Finger, whereas B does not harm A in Decline. There is no obvious
way distinguish Finger from Decline without relying on counterfactual comparison.
To illuminate this problem even more brightly consider:
Recovery 2. A is recovering from a serious illness using a drug. B tampers with the drug.
The drug is less effective in assisting A’s recovery than it would have been had B not
tampered with it. Had B not tampered with the drug, C would have tampered with the
drug more severely, rendering it even less effective.

B harms A. If C’s potential action is part of A’s prospects, though, this verdict seems
unavailable to the temporalist. The judgment that B harms A relies on us excluding
what C would have done from the picture. We must compare A’s state with his state
had neither B nor C acted. The temporalist can deliver the verdict that B harms A only
by including the effects of the unaffected drug as part of A’s prospects but excluding
C’s potential act. Prospect-adjusted temporalism thus provides no improvement over
counterfactualism in explaining the case. Once prospects are built into the Temporal
View, it seems a confusing restatement of the Counterfactual View, inheriting all of its
strengths and weaknesses.

C. Regret
Velleman offers a further argument in support of the Temporal View. He claims that
judgments about harm and benefit are judgments about which we sensibly feel
regretful or gratified. The Counterfactual View of harm and benefit, Velleman thinks,
cannot make sense of these emotions. It suggests that we compare the person as he is
and the person as he might have been. But as the people that he might have been are
not objects of self-concern, they do not appropriately figure in judgments of regret or
gratification.33
This argument is mysterious. For one thing, nonhuman animals can be harmed
even though they are incapable of feeling regret or gratification. This suggests that
these emotions are less central to the idea of harming than Velleman suggests. More

33
Velleman, “Persons in Prospect” (note 32) at 242.
 WHAT MIGHT HAVE BEEN 185

importantly, the idea of regret depends on us imagining how things might have been.
For example, suppose I go to a restaurant with my partner. She chooses steak, and
I choose fish. Our dishes come. The fish looks rubbish and the steak delicious. I regret
choosing the fish.
How is regret explained? Here is how. I compare “me with steak” to “me with fish”
and find that I would have been better off being “me with steak.” Velleman implies that
this natural way of seeing the decision is inappropriate. “Me with steak” cannot be an
object of self-concern, and, hence, can play no role in feeling regret. But I do not need
to be concerned with the person that I would have been to experience regret. I can be
concerned with the person that I am, but wish that my circumstances had been
different. And that involves imagining what it would have been like to have steak.
Seeing how Velleman’s own account of regret is misplaced bolsters this simple idea
further. The temporal account does as though regret involves comparing how I am
with how I was. But when I regret ordering fish I don’t compare “me with fish” with
“me prior to ordering, with the option of fish or steak.” We can see this in virtue of the
fact that prior to ordering, although I had options, I may also have been grouchy and
hungry. Insofar as we can compare my circumstances with my circumstances prior to
ordering, my choosing of fish improved my circumstances, though not by as much as
they would have been improved had I ordered steak.
Perhaps it might be argued that I was better off prior to ordering because I have the
option of choosing the steak and that option is preferable. But the value of that option
depends on how well off I would be were I to receive the steak. The value of an option
depends on the value of its content. And the value of its content depends on how
things would be were I to select it. Regret is filled with counterfactual imagining.34 This
reinforces the idea that prospect-adjusted temporalism is simply a confusing restate-
ment of counterfactualism.

VI. In Defense of Counterfactualism

I offer a further argument in support of the Counterfactual View. The argument
depends on the following idea. In cases of overdetermination, various judgments
about harm and benefit are relevant to normative enquiry. These judgments draw
on various different comparisons. The range of comparisons involves imagining
various different ways in which the world might have been. Hence, various counter-
factual comparisons are involved in assessing these cases. This provides a compelling
reason to believe a version of the Complex Counterfactual View. This argument also
has a nice irony about it—preemption cases are normally thought to be a weakness in

34
This is also the standard view in the psychology of regret. See, e.g., Marcel Zeelenberg and Eric van Dijk,
“On the Comparative Nature of Regret,” in D.R. Mandel and D.J. Hilton (eds.), The Psychology of Counter-
factual Thinking (London: Routledge, 2005).
186 VICTOR TADROS

counterfactual accounts of harming. On the contrary, I argue, they reveal their

strength.

A. Harm, liability, and preemption

Recall:
Finger. D chops off one of V’s fingers, preempting that finger and another being chopped
off by X.

Here are three verdicts about Finger that are relevant to normative enquiry. First,
D harms V by chopping off his finger. Second, D benefits V by preempting V being
harmed by X. Third, D benefits V overall in virtue of the fact that the harm that he
causes is less serious than the harm that he preempts.
Not all counterfactualists think these verdicts are warranted. Derek Parfit accepts
that it is intuitive that D harms V. He claims that we should revise the concept of harm
and adopt the Simple Counterfactual View.35 This view yields the verdict that D does
not harm V in Finger in virtue of the fact that V is better off than he would have been
had D not acted. As his view makes it more difficult clearly to ask the range of morally
salient questions that we wish to ask about Finger, we should reject Parfit’s view.
For example, we wish to ask:
a) Does D owe V compensation for the chopping even though the chopping
preempted an even worse chopping? And
b) Does D’s obligation to pay compensation to V depend on whether D was aware
of the chopping that he preempted? And
c) Does D’s obligation to pay compensation to V depend on the reasons why
D chopped?
Our questions can naturally be stated in a more general way:
a) Does one person owe compensation to another for harming him even though he
preempts a greater harm?
b) Where the harm caused is overdetermined, does the obligation of the person
causing the harm to compensate the victim depend on whether she knew that
the harm was overdetermined?
c) Where the harm caused is overdetermined, does the obligation of the person
causing the harm to compensate the victim depend on the reasons for which she
acted?
At least the last of these questions is difficult to answer. Our account of harm should
make them easy to ask.

35
See Derek Parfit, Reasons and Persons (Oxford: Oxford University Press, 1984), 69–70.
 WHAT MIGHT HAVE BEEN 187

The answer to a) is surely: at least sometimes. This is because the answer to b) is

clearly yes. If D is unaware that he has preempted X’s more harmful action, D owes
V compensation. X’s potential action cannot vitiate D’s liability. If X does nothing and
has no influence over D’s action, X owes V nothing. We do not owe compensation for
the harm we might have caused, only for the harm that we have caused. If D owes no
compensation, V is not entitled to compensation. That is hard to believe.
Here is a brief defense of the idea that D is liable to compensate V. Consider a world
in which D is not liable to compensate V in virtue of the fact that he preempted X’s
action. X’s plan to harm V vitiates D’s liability to V. In such a world, how ought we to
respond to the harm that V has suffered? We could let the harm remain with V. If so,
V remains very badly off as a result of the combination of D’s action and X’s plan. We
have strong reasons not to let V remain in this state.
Alternatively, other citizens could compensate V. They will then bear the costs for
the harm caused by the combination of D’s action and X’s plan. There would seem
little more reason to require these citizens to compensate V than to provide funds for
other citizens that are badly off. There is more to say about this possibility, but I believe
that it should be rejected.
There are strong reasons, in this world, to hold X liable to compensate V. Here is
why. If X is not liable to compensate V, X will have rendered V much worse off than he
would have been had X not formed a plan to harm V; for in that case, V would have
been able to secure compensation from D for the harm that he has suffered. Hence, in
this world, X would have rendered V worse off than he would have been had he not
formed his plan.
The fact that X would have harmed V supports X’s liability to compensate V. It may
not support it decisively—perhaps fault is also required. A good reason to make fault a
condition of compensatory liability is that this allows people to avoid becoming liable
to pay compensation.36 They can do so by refraining from acting in a faulty way.
Suppose, then, that it is wrong for X to plan to chop V’s fingers off. X’s objection to
being held liable is weak in virtue of the fact that he could have avoided being held
liable simply by refraining from forming a plan to chop V’s fingers off. The two most
important components of liability: causation and avoidability are both present.
To summarize, in the world that we are considering, we have three reasons to hold
X liable to pay compensation: 1) were we not to do this, X would have rendered
V worse off than he would have been had X not formed his plan; 2) even if liability is
imposed, it could easily have been avoided by X; 3) if X does not compensate V,
innocent people will have to bear the cost.
Now return to the actual world of Finger. From our discussion, we can see that
someone, either D or X, ought to be liable to compensate V. If D is not liable to
compensate V, X is. Hence, either D or X is required to compensate V. Who should we

36
For more on avoidability, see Scanlon, What We Owe to Each Other (note 12) at Chapter 6. I develop
Scanlon’s account of the significance of choice in Victor Tadros, The Ends of Harm (Oxford: Oxford University
Press, 2011), Chapters 3 & 8.
188 VICTOR TADROS

hold liable, then, D or X? The main reason in favor of holding D liable is that D has
actually affected V. It is only if we refrain from holding D and X liable for the harm
done to V that X will have rendered V worse off than he would have been had he not
formed his plan. If D is liable to pay compensation to V, X has not affected V. The idea
that we are especially responsible for the outcomes that we actually cause is, of course,
somewhat controversial. It is also highly intuitive. I will not attempt to defend it here.
I only claim that if this plausible idea is true, it follows that D rather than X is liable to
compensate V.
This verdict that D is liable to compensate V relies on the idea that D has harmed
V and hence rejection of the Simple Counterfactual View. We thus have two reasons to
reject Parfit’s revisionist account of harm. First, it makes the relevant questions
difficult to ask. Second, in Finger it is highly plausible that if D is unaware of X’s
potential action, D is liable to compensate V for the harm that he has caused. Parfit’s
revision renders this description unavailable.

B. Multiple comparisons
Some may think that the analysis of Finger above provides a reason to favor non-
comparativism or temporalism. It doesn’t. It provides a reason to reject the Simple
Counterfactual View, but only in favor of the Complex Counterfactual View. Recall:
Recovery 2. A is recovering from a serious illness using a drug. B tampers with the drug.
The drug is less effective in assisting A’s recovery than it would have been had B not
tampered with it. Had B not tampered with the drug, C would have tampered with the
drug more severely, rendering it even less effective.

The intuitive verdict that B harms A relies on a comparison between the recovery that
A actually has with the recovery that he would have had if neither B nor C had
tampered with his drugs. This comparison also provides an intuitive account of the
magnitude of harm that A suffers. The amount that A has been harmed is the
difference between his actual state and the state that he would have been had no one
tampered with his drug.
This judgment is only one of a range of morally significant judgments to make about
the case. Importantly, as well as harming A, B has benefited A. It is only in virtue of B’s
act that C did not harm A. There are two different judgments to make here. First, we
can consider the magnitude of the threat that B’s act averts. This is the difference
between the state that A would have been in had B not acted and the state that he
would have been in had C but not B acted. Second, we can consider how much better
off A is than he would have been had B not acted. This is the difference between the
state that A would have been had B not acted and A’s actual state.
Here is a more complete account. The numbers at the end of the descriptions are
illustrative of the magnitude of the difference in wellbeing between the first state and
the second:
 WHAT MIGHT HAVE BEEN 189

1) The harm B causes. The difference in A’s wellbeing between the world where
B tampers with the drug and the world where no one tampers with the drug. (5)
2) The magnitude of C’s threat. The difference in A’s wellbeing between the world
where C severely tampers with the drug and no one tampers with the drug. (12)
3) The overall benefit that B provides. The difference in A’s wellbeing between the
world where C severely tampers with the drug and the world where B tampers
with the drug. (7)
We use all three judgments in this simple and natural characterization of B’s act:
B harms A (judgment 1), but he also benefits A by preempting a greater harm
(judgment 2), benefiting A overall (judgment 3). Thus, B has harmed A by five units
of wellbeing, he has preempted a harm to A of twelve units of wellbeing, and hence he
has benefited A overall by seven units of wellbeing.
It is a strength of the Counterfactual View that it yields a range of verdicts in
preemption cases. These verdicts describe preemption cases more fully than the
Absolute View or the Temporal View (insofar as the latter is distinct from the
Counterfactual View). Neither the Absolute View nor the Temporal View pick out
the idea that B both harms A but also benefits him.
This is because only single verdicts are available on the alternative views. On the
Absolute View, A is harmed only if he remains badly off. If he does remain badly off,
though, it cannot pick out the idea that B benefits A. On the Temporal View, A is
harmed if he is worse off than he was prior to B’s act. But without counterfactual
comparison, there is no way to represent the idea that B is also rendered better off by
A. Hence, its ability to handle preemption cases such as Recovery 2 is a comparative
strength of counterfactualism.
Some might doubt that we need both the verdict that B harms A and the verdict that
B benefits A. To see why both verdicts are necessary, consider whether B is required to
compensate A. Earlier, I supported the conventional view that the person who causes
harm preempting a greater wrongful harm, owes full compensation for the harm that
he causes if he was unaware of the benefit that he provides. Similarly, in Recovery 2, if
B was completely unaware that he was preempting C’s action B owes A compensation
for five units of harm.
Now suppose that B acts only in order to benefit B by preempting C’s action. There
is no way for B to benefit A other than by tampering with the drug. Furthermore,
B cannot seek A’s consent. In that case, intuitively, B does not owe A compensation.
It would be surprising if doing this noble deed for the sake of A rendered B liable
to compensate A. If anything, A owes B a debt of gratitude. If anyone owes
A compensation in that case, it is C. C may owe A compensation in virtue of wrongly
having created the conditions in which B is justified in harming A. I conclude that
B does not owe compensation to A only if he was aware that he was preempting C’s
act. Only if he was aware that he was preempting C’s act can he point to the benefit
that he has provided to vitiate A’s claim to compensation.
190 VICTOR TADROS

I do not claim that B’s awareness of C’s potential action is sufficient for him to avoid
liability to pay compensation to A. B’s liability may also depend whether B was
motivated to assist A. If B was aware of C’s action, but hated A so much that he
wanted to be the one who harmed A, it is plausible that B owes A compensation. On
this view, awareness of a reason that would justify one’s action is insufficient to avert a
person’s liability to pay compensation for causing harm. To avert liability, one’s harm-
causing action must not only be justifiable, it must be justified.37 Another way to put
this is that in assessing B’s liability it is relevant to compare A’s current state with the
state that he would have been in had C harmed A only if B was aware of C’s potential
action, and perhaps also only if he was motivated to avert it.
So far I have established that there is good reason to refer both to the harm that
B causes and the benefit that B provides overall. Depending on further facts in the case,
these judgments will determine whether B is liable to pay compensation, deserves
gratitude, or ought to be compensated for any losses that he suffered in benefiting
A overall.
Do we also need to use our second judgment—the magnitude of the threat that
C posed to A? There are a number of reasons why we may wish to determine not only
the magnitude of harm that B averted overall, but also the magnitude of the threat that
C posed. Determining what would have happened had B not intervened seems
important for a number of reasons. First, there is an important sense in which B has
not only benefited A, but has also benefited C. He has benefited C by preventing
C from harming A. This benefits C not least because C would have been liable to
compensate A for a loss of twelve units of wellbeing had B not intervened. Though it is
even more unlikely than the cases that I have described, we could imagine a variation
of the case where the tampering that C would have done had A not intervened would
have been negligent rather than intentional. C would then be grateful to B for
intervening.
Second, referring to the harm that C would have caused had B not intervened is
important in guiding C’s conduct in the future. It has attractively been argued that
appealing to the role that causation has in practical reasoning can vindicate a coun-
terfactual analysis of causation.38 We develop causal concepts, it has been suggested, in
order to learn how to intervene to bring about certain results. It is for this reason that
we distinguish between causation and correlation. The fact that one event, E, correlates
with another, E2, helps us in predicting whether E2 will occur—it will occur whenever
E occurs. It does little to help us bring E2 about. The fact that changes in a barometer
reading correlates with a storm does not allow us to control the weather. If we discover

37
For a parallel view in the context of criminal law, see John Gardner, “Justifications and Reasons” in
Offences and Defences: Selected Essays in the Philosophy of Criminal Law (Oxford: Oxford University Press,
2007); see also Tadros, Criminal Responsibility (note 32) at Chapter 10. The problem of overdetermination also
helps to establish that intentions are relevant to permissibility. For further discussion, see Tadros, The Ends of
Harm (note 37) at Chapter 7.
38
E.g., James Woodward, Making Things Happen (Oxford: Oxford University Press, 2003).
 WHAT MIGHT HAVE BEEN 191

that E is a cause of E2, in contrast, we learn how to bring E2 about. That depends on it
at least typically being the case that bringing E about will bring E2 about.
Whatever the right account of causation may be, the idea that our account of
harming has the practical function of securing our level of wellbeing by guiding us
to manipulate the world is attractive. C has good reason to learn how not to cause
harm. He can find that out by investigating how much harm he would have caused
had B not intervened. C should make that judgment given that he cannot normally rely
on B to intervene.
Finally, if C plans to act out of bad motives, he may be liable to be harmed even
though he does not pose a threat. C is an attempter. It is plausible that the magnitude
of harm that may be imposed on attempters through punishment is determined in part
by the harm that they attempted to cause. In this case, C attempted to cause twelve
units of harm. This is important in determining his magnitude of criminal liability.
The second verdict is normally unimportant in determining whether, and how much,
compensation is owed. It is important for a range of other reasons.

VII. Conclusion
Perhaps some may doubt the importance of the defense that I have provided for the
Counterfactual View. They may believe that my cases are so unusual that our concep-
tion of harm ought not to be driven by them. These cases are, of course, unusual.
Unusual cases help us to clarify the implications of different theories. But overdeter-
mination itself is extremely common. It is very often true that the harm that one
person causes to another is overdetermined, and it is thus very important that our
theory of harm is adequate to handle these cases.
Furthermore, the problem of overdetermination has provided a key reason why
counterfactual views of both harm and compensation have been rejected. Were it not
for odd overdetermination cases, almost everyone would be a counterfactualist. If this
apparent weakness of counterfactual views turns out to be a significant strength, a
great deal of support has been provided to counterfactualism more generally.
This is a welcome and significant result. Counterfactualism is highly attractive in
cases where harm is not overdetermined. In these cases, the Simple Counterfactual
View is ideal in determining the magnitude of harm suffered and the magnitude of
compensation owed. That test is inadequate in more complex cases. But as we have
seen, the Complex Counterfactual View performs excellently in these cases. Our theory
of compensation and harm is thus nicely unified.
Of course, large problems remain. As there are many counterfactuals that we could
refer to, how are we to pick out those that are salient for purposes of compensation?39

39
Context will clearly be important in picking out the relevant counterfactuals. For discussion in the
context of causation, see, e.g., Christopher Hitchcock, “Counterfactual Availability and Causal Judgment,” and
Peter Menzies, “The Role of Counterfactual Dependence in Causal Judgements,” both in Christoph Hoerl,
192 VICTOR TADROS

The Counterfactual View itself does not provide us with a solution to this complex
problem. We need principles within that view to develop solutions. But even if the
Counterfactual View does not provide us with all of the answers, it does help us to get
clear about the best way to ask the questions.

Teresa McCormack, and Sarah Beck (eds.), Understanding Counterfactuals, Understanding Causation: Issues
in Philosophy and Psychology (Oxford: Oxford University Press, 2011).
 9
Why Reparations?
Rahul Kumar

I. Introduction
Is it possible that a past action, say one that took place a few hundred years ago, can be
said to have wronged not only individuals alive at the time of the wrongdoing, but also
individuals alive now? Certain familiar ways of conceiving of the normative basis of
the claims of the living to reparations for past wrongdoing appear to presuppose that
it is.1 Recent philosophical discussions of the normative foundations of reparations
claims, however, suggest that it is in fact surprisingly difficult to articulate a plausible
account of why it is that living individuals may well have a legitimate claim to
reparations for a past wrong doing, such as the African-American claims to repar-
ations for chattel slavery.
At the heart of the problem, it seems to me, is the tacit assumption in many
discussions that we ought to understand reparations claims as claims to some kind of
compensation for harm done, analogous to a corrective justice claim of the kind whose
natural home is in tort law. Thinking about the normative grounds of certain kinds of
reparations claims in such terms may be appropriate in, for instance, some cases of
claims to reparations for the confiscation of property. But it seriously distorts the nature
of the claim when what is of concern is not property loss or damage, but what was done
to persons, and in particular, how the wrongdoer has related to the wronged.
That intergenerational reparations claims ought not to be interpreted as a kind of
tort claim is not a new suggestion; the point has been offered as a reason for the
wholesale rejection of the idea of reparations as a demand for compensation. But this is
to overlook an important insight captured, imperfectly, by the idea of reparations
claims as claims to compensation: the grounds of reparations claims have to do with
what is owed in virtue of what was done, rather than the positive consequences of taking
such claims seriously. Where the characterization of reparations-as-compensation goes
wrong is not in the way such claims are grounded in past events, but in the assumption

1
For purposes of discussion, “past wrongdoing” concerns wrongdoing that took place at least far enough in
the past that it can reasonably be thought of as having consequences for who now exists.
194 RAHUL KUMAR

that what is relevant to the assessment of whether a living individual has been wronged,
in a way that would justify her claim to reparations, is whether and to what extent she
has been harmed as a result of what was done in the past.
A plausible philosophical account of the normative foundations of reparations
claims ought, I believe, to satisfy three desiderata. It should (i) offer an illuminating
explanation for how it is that a past wrongdoing can be said to wrong both living
individuals who arguably would not now exist but for that very wrongdoing and
individuals alive at the time of the wrongdoing; (ii) offer an account of the wrongdoing
that can make sense of the kinds of symbolic and material measures associated with
reparations claims; and (iii) characterize the wrongdoer against whom it is appropriate
to press claims to reparations, given the character of the wrong for which reparations
are owed. A characterization of reparations claims, not as demands for compensation,
but as a matter of the wronged person holding the wrongdoer accountable for his or
her (or its) wrongdoing satisfies all three desiderata. Further, it is one that, I believe,
does justice to the insights of the compensation model while avoiding its pitfalls.
To be justified in holding another accountable for what she has done, it need only be
true that the person held accountable have wronged another by having either ignored
or flouted certain entitlements having to do with how others may relate to her, which
are constitutive of what respect for her standing as a person requires. It need not be
true that the wronged person has been made worse off than she otherwise would have
been (or suffered some kind of set back to her interests). Nor need it be true (though it
is compatible with the possibility) that the wronged person is entitled to demand
compensation for the harm (or loss) imposed upon her. An account meeting these
desiderata would be able to preserve what is right about, while not being vulnerable to,
certain powerful objections that can be pressed against a general characterization of
reparations claims as claims to compensation.
The case for why this is so falls into five parts. In Section II, I argue that the two most
prominent patterns of justification for the claims of the living to reparations for past
wrongdoing—what I call reparations as compensation and reparations as restoration—
are subject to serious objections that cast serious doubt on their plausibility. Section III
introduces, somewhat schematically, a better conception of what it is for one person to
have wronged another. The particular appeal of this approach for understanding
reparations claims lies, I suggest, in the fact that it identifies a person’s claim to have
been wronged in considerations having to do with the character of the wrong, rather
than its consequences. In Section IV, the approach to wronging sketched in Part II is
illustrated by attention to the case of the claims of living African-Americans to
reparations for chattel slavery, claims that are thought by many to be undermined
by the argument that the claimants would not now exist were it not for chattel slavery.
I argue that this skepticism is unfounded on the grounds that the approach to
wronging sketched in section three focuses attention on an important respect in
which chattel slavery can be said to have wronged both those enslaved and generations
of African-Americans who came after them, including the present generation, even if
it is true that many of them would not now exist were it not for chattel slavery.
 WHY REPARATIONS ? 195

II. Two Approaches to Reparative Obligations

Discussions of the validity of reparations claims often appeal to the thought that the
disadvantages, both material and psychological, that presently burden members of
certain identifiable groups can only be properly understood by attention to the past
injustices or wrongs visited upon past members of that group in virtue of taken to
belong to that group. One cannot, for instance, properly understand the sense of
cultural alienation and inferiority that is widely explored in post-colonial literature
without attention to the ways in which colonialism seriously wronged the indigenous
populations. To understand the present, one has to understand the past from which it
has emerged.
On its own, this is a relatively innocuous idea about the possibility of there being any
valid reparative claims for past wrongdoing to which not even the skeptic ought to
object. For even if it is the case that, for instance, many of the disadvantages currently
faced by African-Americans are causally connected to chattel slavery, that doesn’t
show that anyone is in any way entitled to anything because of chattel slavery. The
causal story might amount to nothing more than salutary history, useful for dispelling
certain serious misunderstandings concerning the roots of the disadvantages currently
faced by many African-Americans, but nothing more. After all, does anyone really
want to claim that, had history been different, such that no sense could be made of
chattel slavery being a distinct causal factor in bringing about the injustices that
presently disadvantage African-Americans, the state would have no responsibility to
eradicate the sources of those unjust disadvantages? If the answer is “no,” then it
is hard to resist the conclusion that the reason the history is relevant is because it is
important to uncover the roots of morally repugnant attitudes and patterns of thinking
that still retain their social currency, and to crush any propensity to a sanctimonious
sense of moral superiority. That is, what the history brings to light are considerations
that are important for advancing the cause of bringing about a more just world in the
future. What it does not bring to light are valid grounds for claims of the descendants
of the victims of past injustice for reparations for those injustices.
This argument is of course too quick. All it shows is that reparative justice claims,
grounded in facts about history, and the requirements of distributive justice, whose
force is drawn from the injustice of the current state of affairs, sometimes converge on
what is required.2 It doesn’t show that there couldn’t be valid reparative claims that
require the state take responsibility for improving the quality of life of, and oppor-
tunities available to, a badly off segment of its population because of its past role in
bringing it about that they are badly off, a responsibility grounded in considerations
distinct from those that ground its duty to ensure that its citizens do not suffer

2
The distinction between reparative justice and present state justice does not claim that present state
theories cannot be attentive to facts about the past, or how a present state of affairs came about. It only requires
that reparative justice have as its subject matter the “putting right” of past wrongdoing or injustice, rather than
the identification of present injustices whose identification may require attention to facts about past choices or
transactions.
196 RAHUL KUMAR

arbitrary disadvantages due to structural facts about the organization of basic social
institutions.
One way of developing this line of thought is nicely captured by what I will call the
compensatory model. Say I destroy your car’s engine (absent mindedly, I pour wind-
shield wiper fluid into the oil tank, inattentive to the label marked “oil only”). Because
of what I did, I owe it to you to make things right, by compensating you for the damage
done to your car. This responsibility for making amends is mine and no one else’s.
Further, the reason I have to make amends just is the fact that I wrongfully destroyed
your car engine. It may also be true that if I do not make amends, it will destroy our
friendship, or that you cannot afford to fix it yourself, so if I don’t pay for the repairs,
you will be left without transport. But these are additional reasons to make amends;
the duty to compensate stands independently of them.
Though something along these lines is often taken as a starting point for the
investigation of the normative grounds of reparations claims, there are objections of
varying seriousness to its value as a way of characterizing the normative grounds of
such claims. Let me mention just two. First, the compensatory model requires that
compensation be made to the wronged in order to restore them to the state they would
have been in had they not been wronged. Determining what is owed is a difficult
matter when the original victims of the wrongdoing are still alive, but at least sense can
be made of the question of what the victim’s life would now be like had the wrong-
doing not occurred. It is not clear, however, that sense can always be made of the
counter-factual when it is asked in the context of assessing the reparative claims of
the descendants of the original victims of the wrongdoing. Consider, for example, the
question of whether or not certain living African-Americans are owed reparations for
chattel slavery (granting, for the sake of argument, that chattel slavery is a significant
contributory cause to the serious disadvantages now faced by many African-
Americans). The compensatory model invites us to ask “how well off would they
have been had it not been for chattel slavery?,” as it identifies as a necessary condition
of having a claim to redress that one have been made worse off by the relevant event
for which compensation is thought to be owed. The problem here is not just that any
attempts at calculation are more likely than not to yield an indeterminate result.
Rather, it is that the most plausible answer to the question “how well off would they
be” is that they wouldn’t have been better off because they would not have been at all.
If it were not for chattel slavery, it is implausible to claim that those who met as a result
of slavery and had children together would have met and had they never been
enslaved. Because, in this kind of case, the fact of wrongdoing proves to be an
identity-fixing fact, the idea that the descendants of the slaves would be better off
had it not been for chattel slavery proves to be literally nonsensical.3

3
A good recent discussion of this issue, which nicely exposes some of the deep difficulties involved in
grappling with it, is George Sher, “Transgenerational Compensation,” 33:2 Philosophy and Public Affairs 118
(2005); see also Bernard Boxill, “A Lockean Argument for Black Reparations,” 7 Journal of Ethics 63 (2003).
 WHY REPARATIONS ? 197

This problem, known as the non-identity problem, is arguably the most serious
conceptual, rather than epistemological, obstacle to putting the compensatory model
to work in service of vindicating as valid many of the claims for reparations pressed by
the descendants of the colonized and the enslaved. Not all, however, accept that the
non-identity problem undermines the value of the compensatory model for under-
standing reparations for historical injustice. The mistake, it has been argued, is in
thinking that the reparative demands of the living seek redress for the wrongdoing
initially perpetrated against their ancestors. The real basis of reparative claims has to
do with the failure to appropriately compensate the original victims of the wrong-
doing. Reparations for chattel slavery might be owed, for instance, to the descendants
of the slaves not because of the wrong of enslavement, but because the freed slaves
were not properly compensated for their labor under chattel slavery after its dis-
mantlement by, for example, the payment, at a fair rate, of back wages for their work
on the plantations. Had the emancipated slaves been appropriately compensated for
their labor, never mind what they were owed by way of compensation for the harm
done to them, the argument goes, the current generation of African-Americans would
be much better off than they are now.
Arguably, this way of framing the demand for reparations using the resources of the
compensatory model is not so obviously vulnerable to the non-identity problem. The
defense, however, comes at a heavy price: that of trivializing the moral force of the kind
of reparative claim in question. It ends up characterizing the wrongdoing for which
reparations are owed as akin to that of someone having failed to compensate my great-
grandfather for accidentally killing the ox that pulled his plow, a failure, let us say, that
is causally related to my life not having gone as well as it would have done had he been
compensated. This is not an illuminating characterization of the grievance that drives,
for instance, the reparative demands made on behalf of African-Americans. It leaves
out of the picture the way in which chattel slavery involved the systematic, state
sanctioned degradation of a class of human beings because of the pigmentation of
their skin.
The problem here—that of leaving out of the grounds for reparative claims the
morally objectionable character of what was done, focusing instead on what
happened as a result of what was done—is not simply an artifact of the example
I have offered. Rather, I take it to be a general problem with using the compen-
satory model to understand reparative claims. One reason for thinking that this is
so emerges clearly in the debate over the value of unjust enrichment suits as a
vehicle for seeking reparations for chattel slavery. The charge of trivialization of the
wrong of slavery is one that has been forcefully pressed by critics of such suits. In
defense of them, it has been argued that the critics misinterpret the symbolic
significance of unjust enrichment claims in the context of slavery. Slaves did not
receive a wage for their labor because in the moral economy of chattel slavery they
had the status of farm animals, not persons. To press for damages due to unjust
198 RAHUL KUMAR

enrichment is really a way to seek acknowledgment of this degradation of a class of

human beings.4
This strategy escapes the charge of trivializing the nature of the wrongdoing for
which redress is sought, but does so by abandoning the thought, central to the
compensatory model, that compensation is the essentially correct form that redress
for wrongdoing ought to take, and downgrading it to an instrument for getting at what
is really sought, namely acknowledgment of the nature of the wrongdoing. Compen-
sation for past wrongdoing, at least on the grounds the compensatory model identifies
as salient, seems to be just the wrong kind of vehicle for redress of past wrongdoing
like slavery, as it appears to be only able to address the consequences of the wrong-
doing, but not what was morally objectionable about the wrongdoing.
These two problems—one distinctively philosophical (non-identity), the other more
intuitive (trivialization)—are two of the considerations that have led many to conclude
that the compensatory model is just not the right way to think about what justice
requires as a response to past mass violence or systematic abuse and degradation. As
one critic of the compensatory model has noted, “No market exists for the value of
living an ordinary life, without nightmares or survivor guilt. Valuing the losses from
torture and murder strains the imagination.”5 We should, the thought goes, abandon
thinking about the considerations that underwrite reparative demands as having to do
with compensation, and move to an understanding of reparations as restoration. The
intuitive idea driving this model is often explained using the analogy of a personal
relationship. If you wrong your lover by making overtures to another, the right way to
address the betrayal is not to offer compensation, but to admit guilt and seek
forgiveness. The legacy of past injustice is that of fractured moral relationships
between citizens, a lack of civic trust, feelings of alienation and a lack of self-respect
amongst members of the historically wronged group, as well as economic hardship
and a lack of adequate opportunities. Taking demands for reparations seriously,
through acknowledgment of, and taking responsibility for, the past wrongdoing and
the subsequent decades of refusal to acknowledge it and its consequences, is necessary,
though obviously not sufficient, for the restoration of relations of mutual respect and
trust amongst all those who are entitled to standing in the civic community. The
purpose of reparations for past injustice, according to this model, then, is not to
compensate but to communicate atonement for past wrong, and thus heal the rela-
tionship between the wrongdoers and the wronged.6
As an approach to understanding the basis of reparative claims, this model has
several key strengths. First, unlike the compensatory model, it requires attention not

4
A line of argument clearly articulated in Dennis Klimchuk, “Unjust Enrichment and Reparations for
Slavery,” 84 B.U. L. Rev. 1257 (2004), 1257–75.
5
This line of argument is advanced is Martha Minow’s much discussed Between Vengeance and Forgiveness
(Boston: Beacon Press, 1998).
6
As Janna Thompson puts it in her influential book, Taking Responsibility for the Past, “the point of
reparation is to bring about reconciliation.” Janna Thompson, Responsibility for the Past: Reparation and
Historical Injustice (Cambridge: Polity Press, 2002), 83.
 WHY REPARATIONS ? 199

just to what happened as a result of the wrongdoing, but attention to the character of
the relevant wrongdoing. In order to acknowledge one’s responsibility for a wrong-
doing and atone for it one must first, after all, be clear about the nature of the
wrongdoing for which one must atone if relations of civic trust and mutual respect
are to be restored.
Second, it requires getting clear on who are the wrongdoers and who are the
wronged. The compensatory model also requires identifying the persons or groups
to whom compensation is owed, but it does not stress the importance of the identity of
the wrongdoer as the one who provides the compensation. On the compensatory
model, what matters is that the wronged (or those who are entitled to press the claims
of the wronged) be compensated.7 Whether the compensation is provided by the
wrongdoer or through some third party is arguably not strictly relevant.
Third, the restorative approach helpfully illumines actual reparations cases, such as
the rejection by the vast majority of approximately 500 survivors of the estimated
200,000 “comfort women” exploited as sexual slaves during WWII, of an offer by the
Prime Minister of Japan of a letter of apology and monetary compensation. Two
reasons the offer was rejected are particularly relevant. The first is that the funding for
the compensation came from private sources, and the letter of apology was not a letter
of apology offered on the part of the Government of Japan. The second was that it was
felt by many that acceptance of compensation would be too much like prostitutes
presenting a bill for services rendered.8
It is of course not always easy to always identify who are the wrongdoers and who
are the wronged in intergenerational cases. Some argue that in cases such as chattel
slavery and colonialism, the state should be viewed as the wrongdoer, bearing corpor-
ate responsibility for the past injustice; citizens of the state are not strictly culpable for
the past wrongdoing, but must share part of the cost of complying with state’s
reparative duties. And the wronged, it has been suggested (partly in order to avoid
entanglement with non-identity issues), should not be characterized in terms of
individuals harmed at the time of the wrongdoing, but in terms of a people who
have been harmed by the wrongdoing.
These are interesting proposals; in particular, it seems to me there is something right
about the idea of the state as the relevant wrongdoer in many of these cases. They are,
however, independent of the proposal at the heart of the restorative model: that when
we ask what the reasons are that might ground legitimate reparative claims for past
wrongdoing, the relevant considerations have to do with restoring relations of mutual
respect and civic trust. But while mutual respect and trust are important elements of
any just political order, this proposal fixes reparations as owed for the wrong kind of
reason: not because of what has been done in the past, but in order to make things
better in the future. As Leif Wenar puts it:

7
Further, it is not clear that on the compensatory model, there is anything objectionable about transferring
one’s claim to compensation to another.
8
See Claudia Card, The Atrocity Paradigm (New York: Oxford University Press, 2002), 136.
200 RAHUL KUMAR

past injustice can hinder future justice when it leaves an undergrowth of disrespect and
distrust. Reparations can remove the hindrances to justice by clearing the ground so that
trusting relations can take root. Those who share attributes with previous aggressors can
signal with an official apology that they will not forget that the past injustice occurred, and
that they will work to transcend the attitudes of superiority they may have inherited.
Similarly, they can offer to fund the cause of reparations in order to show that they are
willing to sacrifice other ends for the sake of establishing sound moral relationships in the
future. Those who share attributes with previous victims can accept offers of reparations in
the spirit intended and with grace, likewise signaling that they are willing to extend trust to
those making a sincere effort to create a scheme of just cooperation.9

The kinds of considerations that the restorative model identifies as relevant to the
justification of reparative claims turn out to be considerations having to do with what
is necessary to prepare the civic soil so that a socially and distributively just society may
flourish. Past wrongdoing ceases to be the ground, on this account, of the entitlements
to reparations.
Maybe this isn’t a problem. If we take a pragmatic approach to reparative claims, we
can allow that the considerations that ground reparative claims are conceptually
indistinct from explicitly forward looking claims of social justice, but still allow that
pressing reparations claims can, in certain contexts, be valuable as an effective vehicle
for getting nations and governments to take steps towards improving relations with
other nations or their own citizens, by reminding them of their role in the history that
has resulted in those relationships becoming morally compromised in various
respects. Reparative claims can have this salutary effect even if it is true that a state’s
causal responsibility for bringing about the current state of affairs does not justify the
conceptual claim that it is responsible for rectifying injustice in virtue of its responsi-
bility for bringing about the current unjust state of affairs.
The appeal of this line of thought is clearest if we think of it as a justification for
reparative claims offered to those against whom the reparative claims are pressed.
What tells against its plausibility is the problem of thinking of it as a justification
offered to those who take themselves to have a claim to reparations for past wrong-
doing. From their point of view, their claims rest on considerations having to do
primarily with the wrongdoer’s past conduct. Appropriately responding to these
reparative claims may help to heal a fractured civic polity, and the goal of rebuilding
civic trust and solidarity is no doubt of great importance in thinking about the form
reparations for a particular wrongdoing in a particular case ought to take. But from the
point of view of those who take themselves to be victims past wrongdoing, these kinds
of considerations in no way figure in the grounds of their claim to reparations.
That an account of the basis of reparative claims cannot make sense of even the
general form of the rationale that those who press such claims take them to have is a

9
Leif Wenar, “Reparations for the Future,” 37:3 Journal of Social Philosophy 396 (2006); see also Christo-
pher Kutz, “Justice in Reparations: The Cost of Memory and the Value of Talk,” 32:3 Philosophy and Public
Affairs 277 (2004) (discussing the normative significance of reparations).
 WHY REPARATIONS ? 201

serious reason to doubt its plausibility. The restorative model, which advocates taking
seriously demands for reparations as justified on instrumental grounds, for the better
realization in the future of the requirements of social justice, is certainly prey to this
objection. Though not intended by its advocates, the restorative model’s justification
for the legitimacy of reparative claims has the flavor, from the point of view of those
who take themselves to be entitled to reparations, of what Bernard Williams famously
called “Government House utilitarianism.”
What a plausible account of the basis of claims for reparations for past wrongdoing
needs to be able to make sense of is how it is that living individuals can be wronged by
what was done in the past. The compensatory model tries to do justice to this thought,
but ends up being derailed both by conceptual difficulties in making sense of there
being a status quo ante to which compensation aims to restore certain individuals, and
intuitive doubts about whether it is even appropriate to think about the kind of redress
past wrongdoing calls for in term of compensation. The restorative model avoids these
difficulties, but it finds the grounds for the validity of reparative claims in consider-
ations that fail to make sense of reparations as having to do with redress for past
wrongdoing. Though those who favor the restorative model take themselves to be
making a case for the validity of certain claims for reparations, it is hard to see why
what they are arguing for, as a conceptual matter, is reparations, rather than a more
sophisticated understanding of what is required to bring about a more just society, in
which there is both social solidarity and civic trust.

III. Rehabilitating the Compensatory Model

Both the compensatory and the restorative models contain important elements of
truth. What the compensatory model gets right is that reparations are owed because of
past wrongdoing; the reason for complying with such obligations is not that doing so
promises to improve the future health of the political community or international
order. The restorative model, on the other hand, correctly draws attention to the extent
to which thinking about reparative demands as requiring restoration of the wronged
group to the status quo ante (or as close as one can get) through a program of
compensation constitutes, in many cases, a serious misunderstanding of the normative
significance of demands for reparations for past wrongdoing.
The elements of an account of the normative basis of reparative claims that does at
least some justice to what both these models get right emerges, I believe, if we
distinguish two questions that the compensatory approach seeks to answer. First, in
virtue of what does a person have a claim to have been wronged? Second, what
considerations do we need to take into account in thinking about what the wrongdoer
owes the wronged in virtue of her wrongdoing? The compensatory model’s answer to
the first is that a person has been wronged insofar as she has been harmed (either
counterfactually or made worse off from the relevant ex ante state), offering in
202 RAHUL KUMAR

response to the second that what redress for the wrongdoing requires is compensation
for the harm done that aims to make it as if the wrongdoing had never happened.10
Some of the objections to the model can be addressed if different answers to the
above questions are supplied. In particular, some of the problems that plague the
compensatory model can be avoided if the idea that there is a tight connection between
wronging and harming is rejected, and treat the question of whether an individual
(or group of individuals) has been harmed as something to be taken into account
in determining what the wrongdoer is required to do in virtue of having wronged
another, but not a relevant consideration for fixing whether another has in fact been
wronged.
The case for the claim that a person is not wronged simply in virtue of harm having
befallen her can be motivated with a simple thought experiment: say that I am harmed
in a bicycle accident (my brakes fail and I end up crashing into a hedge). I am injured,
but not dangerously so. But then I discover that my “accident” was not an accident at
all, but the result of sabotage by a foe. Whether the harmful event was accidental or the
result of meddling makes no difference to the harm I have sustained as a result of it.
Because of it, I am worse off than I was before, and my being worse off than I would be
had the accident not occurred. That is, the realization that my accident is really an
“accident” has no bearing on what has happened to me. But it does change the
normative significance of the event, as what I realize is that it’s not just that harm
has befallen me, but that my enemy has wronged me. This difference in normative
significance is most plausibly located not in facts about what has happened to me as a
result of my foe’s machinations, but in considerations having to do with the character
of her conduct. One might, for instance, characterize what is objectionable about
the character of her conduct in this case as having to do with her treating, in her
deliberations, how my life goes as something that it is permissible for her to direct
irrespective of my wishes.
This simple thought experiment does not, however, show that having been harmed
is not a necessary condition for having been wronged; at most, it shows that what it is
to have been wronged cannot be completely accounted for by appeal to considerations
of harm. Whether or not a person can be wronged even if she has not been harmed
turns on whether or not I would cease to have a grievance if my foe failed to succeed in
securing the result at which she was aiming. It seems to me that the fact that her plan
came to nothing as a result of her incompetence in executing it, or some other bit of
bad luck, makes no difference to my having a legitimate grievance against her (though
if she fails due to incompetence rather than some more external kind of bad luck,
I may choose not to press that grievance). If it is her plan that is morally objectionable,
the success of it ought to make no difference to whether or not I have been wronged by
her conduct, though it may make a difference with respect to what I am entitled to
demand of her in light of her having wronged me.

10
By “counterfactually” I have in mind the view that is common in tort “but for” reasoning, that a person
has been harmed if she would now be better off had it not been for another’s relevant action(s).
 WHY REPARATIONS ? 203

Returning to the two questions distinguished above, I believe that what it is for a
person to have been wronged by another is best explicated, roughly, as a person has a
claim to have been wronged by another insofar as the wrongdoer, either intentionally
or negligently, conducts herself in a way that does not conform with a way she would
be justified in conducting herself were she to take the relevant interests of the wronged
into account in her understanding of how it is permissible for her to conduct herself, in
the way to which the wronged is entitled as a matter of respect for her value as a
human being.11 Respect for the value of a person’s life, on this view, requires appro-
priate recognition of that individual as capable of living, and the importance to her
being able to live, a rationally self-governed life, meaningful life. This understanding of
wronging and how the value of a human life is appropriately respected are tied
together, because whether or not the value of one’s life is appropriately recognized
in another’s conduct could only matter in relations between creatures capable of
rational regulation of their deliberations and consciousness of a failure of another to
appropriately recognize one’s value in her practical deliberations. That is, what one
person’s conduct says about the value of another as a person only matters if both are
capable of meaning something by their actions.
That a person has been wronged, however, only shows that something is owed, not
what the wronged is entitled to in virtue of the wrongdoing. What is owed depends on
both the character of the wrongdoing and its relevant consequences for the wronged. Not
all wrongdoing involves a setback of a person’s interests. Some ways of wronging another
are largely symbolic, best characterized in terms along the lines of an assault on, or denial
of, one’s value as a human being. Central to redressing such wrongs are acknowledgment
and apology (where the requirements of apology often go well beyond just sincere
avowal). In other cases, though, the emphasis will be on providing material restitution,
or that steps be taken to improve the prospects and well-being of the wronged.
What is important here is that determining what a wrongdoer owes the person (or
persons) she has wronged is itself a matter of substantive moral argument that is
sensitive to the character of the wrongdoing and its consequences. There will be some
cases of wronging in which what the wrongdoer owes the wronged precisely is
compensation, and in which the reasons that fix it that compensation is what is
owed also require that what is offered as compensation be what will return the
wrongdoer to the status quo ante. But if this is so, it will be true because of features
of the particular case. It is not a general feature of a person having been wronged that
what she is entitled to is best understood as “compensation” (irrespective of whether or

11
Joseph Raz helpfully distinguishes “respect” for value from “engagement” with value. Out of respect for
the value of Michaelangelo’s Pieta, I have reason not to spray paint it, or attack it with an axe, or think of it as
something any old hack could have produced. Doing these things would count as failures to appropriately
recognize its value. Further, the reasons not to do or think these things, and to generally appropriately
acknowledge its value in thought and action, apply to me whether or not I take any further interest in it,
such as admiring it, reading about, or discussing it with others. It is this understanding of respect for value that
I employ here. See Joseph Raz, Value, Respect and Attachment (Cambridge: Cambridge University Press,
2001), 161–9.
204 RAHUL KUMAR

not compensation is understood as restoring an individual to the state she was, or

would have been, in prior to the wrongdoing).
This way of thinking about what it is for a person to have been wronged by another
preserves what is importantly right about the compensatory model, while avoiding its
weaknesses. Two respects in which this is so are particularly salient to the present
discussion. First, it avoids characterizing past wrongdoing as a matter of trying to
compensate for the past wrongdoing. Reparative claims are grounded in past wrong-
doing in virtue of which the wronged have legitimate claims against the wrongdoer,
but there is no reason to characterize what the wronged are entitled to as compensa-
tion for that which is not compensable, or some kind of compensation that is
independent of any serious symbolic acknowledgment of responsibility for having
wronged another (or others).
Second, it is not obviously vulnerable to the threat of the non-identity problem. The
non-identity problem, recall, draws attention to the fact that in order for a person to
have been harmed as a result of a particular event, there must be a possible world in
which that person both exists and in which the event did not occur that stands in a
branching relation to the world before the harmful event. Where there is no such
possible world, it is literally nonsensical to claim that a person has been harmed by a
past event, as there is no possible world in which she would have been better off. This
is only a threat to the possibility of legitimate intergenerational claims for reparations,
however, if it is the case that either (a) it must be the case that a person has to have
been harmed in order to have been wronged or (b) fixing what the wrongdoer owes the
wronged is a matter of determining compensation. An analysis of what it is for a person
to have been wronged of the kind proposed here, however, that denies that harm is either
sufficient or necessary for a person to have been wronged, is not so obviously vulnerable
to non-identity issues. And as a person having been wronged does not necessarily call for
compensation, non-identity concerns need not render the question “what are some
individuals owed as a result of past wrongdoing” a necessarily empty question.
Nothing that has thus far been said, however, shows that there is anything to be said
in favor of it being plausible to think of living persons being owed anything as a result
of wrongdoing that took place, in some cases, several generations ago. I do think the
legitimacy of such claims is defensible. My defense, however, will take an indirect
form, through an examination of how the approach to wronging sketched here can be
used to illumine the basis of the claims of living African-Americans to reparations for
chattel slavery.

IV. Are Living African-Americans Wronged

by Chattel Slavery?
There is no such thing as the wrong of chattel slavery. The question of whether or not
living African-Americans have a valid claim to reparations for chattel slavery, as I will
interpret it here, is whether there is an important respect in which chattel slavery can be
 WHY REPARATIONS ? 205

said to wrong both living African-Americans and to have wronged those enslaved
under the legal regime of chattel slavery. Put somewhat tendentiously, the question can
be characterized as that of whether there is a significant respect in which chattel slavery
simultaneously wronged those alive at the time and future generations of blacks in
America? Grounds for thinking there is emerge, I believe, if, taking our cue from the
discussion of wronging from the previous section, we focus on the significance of the
installation by the state of chattel slavery as a legal institution.
Slavery in the American colonies predated the installation of chattel slavery, initially
taking the form of indentured servitude.12 Chattel slavery went well beyond this,
transforming, through the enactment of a complex statutory framework by colonial
legislatures—later accommodated in various ways in the framing of the US
Constitution—the legal status of black American’s into that of property, subject not
to the laws that govern citizens, or even the treatment of persons who fall within the
legal ambit of the state’s authority, but to an elaborate codified body of slave law.13
Presupposed by these legislative changes was a particularly degrading understanding
of the value of human being with black skin. It was not as if they were understood to be
unlucky, in the way that a conquered people who lose a war might end up being
enslaved by the victors. Rather, their assignment of the status of being slaves—a form
of private property—was taken to be in some sense deserved, or fitting, given their
status as inferior beings, especially with respect to their capacity for rational self-
governance.
Acceptance of such a degraded understanding of the value of those identified as
racially “black” was a failure to appropriately respect the value of all those identified as
black at all times and all places, insofar as understanding of the value of blacks
presupposed by chattel slavery expressed an understanding of the value of a racial
type. Any token of the type would, therefore, be justified in taking the understanding
of the value of blacks that underwrote chattel slavery to be expressive of an attitude
towards her to which she could justifiably take offense. This is compatible with the
symbolism of the state’s taking blacks to be inferior having a significance for
those subject to the state’s authority that it does not have for those who are not, one
associated with the legitimacy that an attitude being the attitude of a legitimate
authority gives it. It is, for instance, one thing to be picked on by one’s classmates,
and quite another to have this go on with one’s teacher looking on and doing nothing.
The content of the teacher’s attitude may be the same as the students, but her position
as an authority figure confers upon it a legitimacy it would not otherwise have, as well
as lending an air of legitimacy to the attitudes of others.

12
The next three paragraphs are from Rahul Kumar and David Silver, “The Legacy of Injustice,” in Lukas
Meyer (ed.), Justice in Time (Baden-Baden: Nomos, 2004).
13
“Accommodate” here does not mean “endorse.” Nothing I say here requires the view that the Consti-
tution as drafted with an eye to the interests of the slaveholding states. It is enough for my purposes that the
Constitution can be thought of as accepting the understanding of blacks initially legitimated by colonial
legislatures in the slaveholding states.
206 RAHUL KUMAR

The attitude towards blacks expressed by the installation of chattel slavery certainly
wronged them, as taking them to be inferior represented a significant failure by the
state to appropriately recognize their value. But even if we accept the asymmetry noted
above, it is not, intuitively, the right kind of wrongdoing to appeal to in an account of
the possible grounds of claims of living African-Americans to reparations to chattel
slavery. They are claims whose grounds appeal to the way chattel slavery wronged
blacks in America that is distinct from the way it can be said to have wronged other
blacks. Looking to the symbolic significance of chattel slavery as an authoritative
endorsement of the inferiority of blacks does not identify a distinctive wrongdoing
for which those not subject to the state’s authority could not in principle claim
reparations.
Though this consideration is not decisive, I want to turn to another way of charac-
terizing the relevant wrongdoing that is arguably more illuminating, one suggested by
certain of the strengths of the symbolic interpretation. This proposal characterizes the
relevant wrongdoing as the introduction into the pubic reason of the state the
inferiority of blacks as a legitimate consideration relevant for guiding the state’s
understanding of what it owes blacks subject to its authority.
“Public reason” has a particular technical meaning in Rawls’s conception of demo-
cratic legitimacy. But there is a more informal way of understanding the idea that is the
relevant for present purposes, that of “public reason” as having to do with the kinds of
considerations that it is legitimate for the state to be guided by in how it regulates itself
with respect to the interests of those subject to its authority. The particular attitudes or
reasons that lead to the decision by the relevant legislative bodies to adopt a particular
piece of legislation, approve a treaty, adopt a policy, etc. are not relevant for fixing the
content of public reason, as the term is being understood here. What is relevant are
those considerations that are relevant for making best sense of the justification of how
the state exercises the particular normative powers constitutive of it as an at least
widely perceived to be legitimate authority. State action, broadly construed, then,
stands in a unique relation to pubic reason. It is the considerations that are relevant
for justifying how the state regulates itself with respect to the interests of those who fall
under the purview of its authority that determine the content of public reason.14
The link between how the state exercises its normative powers and public reason
allows for the isolation of one respect in which the transition from slavery as an
unregulated practice to chattel slavery wronged blacks subject to the state’s authority:
the legal establishment of chattel slavery introduced into public reason as legitimate a
certain understanding of the kind of consideration that the state owes those racially
black subject to its authority.
Doing so wronged blacks subject to the state’s authority in a respect in which it did
not wrong those not subject to it. For those subject to the state’s authority have an
interest in, and are owed by the state, recognition as entitled to the same privileges and

14
Public reason, on this view, then, is relevantly analogous to the stated justification for a judicial decision
fixing precedent.
 WHY REPARATIONS ? 207

protections that are partially constitutive of having the status as a full citizen as
Caucasians subject to the state’s authority. The legitimacy of understanding blacks
to be inferior provided the right kind of justification for this moral entitlement to be
denied African-Americans by the state in law, public policy, and in questions of the
importance of enforcing legally mandated rights. In this is a way in which chattel
slavery wronged both blacks in America at the time of the introduction of chattel
slavery and future generations of blacks, but did not, and does not, wrong blacks not
subject to the state’s authority.
It will help to clarify what is being suggested here concerning the character of the
wrong to recall the skeptical challenge to intergenerational reparations claims, to the
effect that an ongoing history of injustice and wrongdoing is not sufficient to justify a
claim to reparations. All that shows is that individuals in the past were wronged, and
that living individuals continue to be wronged. There is, therefore, good reason now to
end present injustices against those who are being wronged by particular extant
practices, and, perhaps, reason to acknowledge past failures to prevent the unjust
treatment of those no longer alive. The first duty is one that is owed to living African-
Americans, on the plausible assumption that they are in fact currently victims of
unjust treatment by the state. The second is one owed to victims of past injustice no
longer alive, and the best that can be done to retrospectively comply with the duty to
treat them justly and prevent them from being treated unjustly. There is, however,
nothing that living African-Americans are owed in virtue of how the state regulated
itself with respect to past generations of African-Americans.
The proposal on offer, while not denying the relevance of on-going unjust treatment
of African-Americans, does not rest on the idea that reparations are owed to living
African-Americans in virtue of the attitude of the state towards them, or how they are
treated by the official organs of the state. Rather, the thought is that the legitimation in
public reason of the inferiority of blacks in itself wronged all African-Americans, as
doing so provided the justificatory basis for the state (through its various official
organs) not recognizing (either actively or through indifference) the moral entitlement
of African-Americans to all the rights and protections that are partially constitutive of
the status of full citizen. This wronged African-Americans alive at the time and future
generations of African-Americans, quite apart from any consequences that this had for
African-Americans.
Distinguish between (i) the legitimation of it being justified for the state (through its
various organs) not to recognize the moral entitlement of African-Americans to all the
rights and protections of citizenship and (ii) the many ways African-Americans were
(and arguably are) wronged by the state in its treatment of and attitude towards them.
Even if, let us imagine, the state had not viewed African-Americans as second-class
citizens, or treated them unfairly, for many generations—without the state ever having
even acknowledged the way in which (i) wronged African-Americans—living African-
Americans would still have a claim to have been wronged in virtue of (i).
The skeptic’s case rests on the observation that the history of African-Americans in
the United States post-chattel slavery is a history of many different wrongs being
208 RAHUL KUMAR

visited upon them, often horrific, but not identical to those visited upon them as slaves.
The present proposal accepts the skeptic’s observation because it is not inconsistent
with the further claim that there is a relevant thread that links these diverse wrongs
with chattel slavery. The observation alone does not provide warrant for the skeptic’s
normative view. The real disagreement with the skeptical proposal is over whether this
history of seemingly diverse wrongs is necessary for living African-Americans to have
a valid claim to have been wronged in some respect by chattel slavery. The proposal on
offer here denies that it is, though it does not deny the relevance of this history for
assessing what is owed by way of redress to the wronged (a matter taken up in the next
section).
The proposed approach to characterizing the basis of the claims of living African-
Americans to reparations for chattel slavery, it is worth noting, is an example of an
intergenerational wronging, but not in the way that reparations claims are often
conceived of as being intergenerational in the philosophical literature on reparations.
It is (in principle) an intergenerational wrongdoing because what was legitimated in
public reason was not a degrading understanding of the value of particular individuals
with black skin, but the humanity of a type of individual. A “type” of individual is best
understood, for present purposes, as a way of referring to a normatively significant set
of characteristics that are instantiated by particular (token) individuals, or in this case,
a set of characteristics imbued with a particular normative significance by the Ameri-
can state. The legislative installation of chattel slavery, therefore, wronged not just
tokens of the type alive at the time of chattel slavery, but all those who entered or were
born in the United States afterwards. This includes those who would not have existed
were it not for chattel slavery, as it is both necessary and sufficient for a token of the
relevant type to have a valid claim to have been wronged that he or she in fact is a
token of the relevant type (being disadvantaged is not a relevant criterion).
It is not intergenerational, in another sense, however, as the claims of living African-
Americans to reparations for past wrongdoing is conceptually independent of whether
blacks in the generation, or generations, before them were wronged in this same way.15
That is, the claim of living African-Americans to be wronged is not one they have
inherited, nor is it one properly characterized as having been transmitted through the
generations. Rather, the wrongdoing is one that simultaneously wronged blacks whose
standing in the state was fixed by the legislative apparatus of chattel slavery, and all
future generations of blacks morally entitled to recognition by the state as full citizens.
The intergenerational character of the wrongdoing does not, however, show that it
makes sense to speak of a now existing wrongdoer whom living African-Americans are
entitled to hold accountable. What sense, one might wonder, can be made of the
persistence of a wrongdoer from the time of chattel slavery to the present? If the
wrongdoer were a particular person, this could be an insurmountable challenge. But in
this case, the wrongdoer is institutional, the wrongdoing having to do with how certain

15
Cf. Sher, “Transgenerational Compensation” (note 3).
 WHY REPARATIONS ? 209

normative powers exclusive to the state were exercised. To identify a persisting

wrongdoer who the wronged are entitled to hold accountable for the wrongdoing,
then, what is required is not the continuity of particular agents, but sufficient institu-
tional continuity (of the kind appealed to in legal contexts to make sense of continuity
of treaty obligations), of sufficient complexity to support the attribution of normative
powers to it, to justify holding the existing state authority accountable for the rectifi-
cation of the past wrongdoing because of its past wrongdoing.
The hard question is to what extent the particular way chattel slavery wronged
African-Americans identified here can be thought to have remained unrectified over
time.16 The issue is not one of the persistence of racism but, first, what would
constitute the state having done enough to expunge from public reason the under-
standing of blacks as being of inferior status? Second, has this been done by the state?
These are both difficult questions, for which there are no straightforward answers.
Guidance as to what might count as having “done enough” might be found in reflec-
tion on the kinds of legislative and policy initiatives during the period of the first
Reconstruction immediately following the Civil War, and later initiatives during the
second Reconstruction. Whether enough was done will crucially depend on how
legislative and judicial history, for a start, after the Civil War is interpreted. One
familiar suggestion is that though the right steps have been taken at various points over
the past 150 years, enough has never been done to firmly establish recognition by the
various organs of the state of the entitlement of African-Americans to the protections
and privileges that, as a matter of justice, the state owes its citizens. An understanding
of blacks as not entitled to what other citizens are entitled to, the thought goes,
survived Reconstruction and went on to vividly manifest itself in Jim Crow laws,
and continues to manifest itself today, particularly in the very visible way in which
racial social stratification continues to be supported by the state by its failure to
respond, and the official indifference to, certain disadvantages faced by African-
Americans that are strongly correlated with race.
The view on offer here cannot, on its own, settle the question of whether this
position is a plausible one. As an account of how to understand what it would be for
chattel slavery, in one important respect, to be a standing wrong—one that wronged
African-Americans in the past, and in the same way wrongs African-Americans
today—it offers an approach to framing the relevant questions that need to be asked
as part of a more general inquiry, attentive to the relevant aspects of the historical
record, into whether living African-Americans have a legitimate claim to be wronged,
in this particular respect, by chattel slavery. Philosophical inquiry alone cannot settle
the matter.

16
There is considerable historical debate concerning the handling of the slavery issue in the framing of the
United States Constitution. Nothing here requires that it be read as endorsing slavery; it is enough that the
understanding of blacks already salient in public reason justified not treating the securing of the equal status of
blacks as an urgent matter (which allowed various “pragmatic” compromises).
210 RAHUL KUMAR

V. Is Saying “Sorry” Enough?

How plausible an account of the normative basis of a certain kind of reparations claim
is thought to be will turn on both the plausibility of the characterization of the
wrongdoing for which reparations are owed and the content of what is thought to
be owed by way of reparations. The discussion thus far has focused on the first issue,
not the second. Assume, then, that the legitimacy of African-American claims to
reparations for chattel slavery can be vindicated. What might be owed by way of
reparations for the wrongdoing? One reasonable interpretation of what the proposal
on offer justifies by way of what is owed are largely symbolic measures: what the state is
obligated to do is to acknowledge the validity of the claims of living African-Ameri-
cans to have been wronged by what was done in the past and its responsibility for
rectifying the wrongdoing, though, for example, measures such as issuing a public,
official, apology, constructing museums and memorials which aim to depict a truthful
account of the injustices of the past, and ensuring that the history and treatment
of African-Americans is taught as part of the official history in the public school
system. It needs to do these thinks, and no doubt more, by way of acknowledging its
responsibility for the past wrongdoing and its rectification.17
That such measures are required in order to rectify the past wrongdoing is intui-
tively plausible. Consideration of why such measures are required by way of making
reparations, in light of the proposed character of the wrongdoing, suggests, however,
that the symbolic measures required may be more extensive than commonly recog-
nized. Recall that the wrong in question is not that of directly doing what harms living
African-Americans, nor is it expressing a disrespectful attitude towards African-
Americans, but of legitimating the inferiority of African-Americans as legitimate
consideration for determining how it appropriate for the state to regulate itself with
respect to African-Americans and their interests. What rectification of this wrong-
doing requires is that an understanding of African-Americans as not entitled to the
same privileges and protection of their interests as Caucasian citizens be expunged
from public reason.
Official, public recognition of the wrongdoing as the state’s wrongdoing—through
an official apology, is an essential part of what is required by a process of expunging
the corrupt understanding of the status of African-Americans in the polity from public
reason. Such an apology is important for the official recognition of the legitimacy of
both African-American claims to have been wronged by the state’s past conduct, and
the state’s responsibility to rectify its wrongdoing. Beyond that, what is required is the
vigilant implementation by the state of the kinds of measures to protect African-
Americans interests (in the realms of, for instance, health, education, and protection

17
What counts as appropriate rectification here has to be some kind of normative standard, as it may be
that whatever is done, the victims may never, and may never be able to, recognize that enough has been done to
repudiate the pass wrongdoing.
 WHY REPARATIONS ? 211

from racial discrimination) to which they are entitled, as a matter of social justice, as
citizens. That is, the way for the state to expunge the corrupt understanding of the
status of African-Americans introduced into public reason at the time of chattel
slavery is to ensure that African-Americans are recognized in the operations of the
various organs of the state as having the status of equal citizens.
On this view, then, it is a mistake to characterize what living African-Americans are
owed by way of reparations as compensation, whose aim is to “make up for” past
wrongdoing. Much of what they are owed in fact overlaps with the requirements of
distributive justice, understood in terms of what the state owes its citizens independ-
ently of any requirement to redress past wrongdoing. What distinguishes compliance
with such requirements in the name of distributive justice from compliance in the
name of redress for past wrongdoing has to do with the rationale offered for compli-
ance with the relevant requirements. That the rationale has to do with redress for past
wrongdoing may make no difference to what ought to be done (though it may make a
difference in the assessment of the urgency of certain measures). But it is important
for the meaning of what is done. That is, what makes measures undertaken in the
name of redressing past wrongdoing that have significant material implications for
the wronged reparative measures is symbolic, and it is this symbolic dimension that
distinguishes them from nonreparative social justice measures to which they are
extensionally equivalent.18
I have argued that both the idea of reparations for historical injustice as claims to
compensation—taking them to have the form, that is, of a tort claim—and reparations
as aimed at healing fractured relationships are deeply flawed approaches to under-
standing the normative basis of reparations claims. What I have sketched, with respect
to the case of reparations for chattel slavery, is an alternative model that I believe
captures the central insights of both approaches. This new approach requires ampli-
fication and discussion beyond what I can do here, but I hope enough has been said to
make its promise clear.

18
For this reason, concerns (expressed by Wenar and Kutz) that redressing reparations claims are in
tension with the requirements of distributive justice strike me as essentially misplaced.
 10
Repairing Harms and
Answering for Wrongs
R.A. Duff *

I. Introduction
This chapter is not about the proper interpretation of existing tort law. My interest is
normative: what kinds of law, serving what ends, giving institutional form to what
values, should we have? My ultimate aim is to determine the proper place of the
criminal law within a legal system—in particular the proper relationship, and the
proper division of labor, between criminal law and tort law. As a prelude to that
enterprise, this chapter considers two different conceptions of tort law: one portrays it
as a way of (re)distributing the costs of harm attributable to human agency, the other
as providing “civil recourse” for people who have been wronged. The first model, as we
will see in Section II, makes possible a clear distinction between the aims of criminal
law and of tort law, and explains some of their differentiating features; but it leaves a
familiar gap, which (we will see in Section III) the civil recourse model promises to fill.
However, as Sections IV—V will show, the civil recourse model can fill that gap only
by making tort law much more like criminal law: this will bring into sharper focus
questions about the relationship between criminal law and tort law.

II. Punishing Wrongs and Repairing Harms

Suppose that I recklessly cause damage to my neighbor’s property: I build a garden fire,
ignoring the obvious risk that it might damage her property; the risk is actualized; and
her garage is destroyed. Two kinds of legal process could then unfold.

* Thanks are due to the participants in the Rutgers Tort Theory Conference, in seminars at the Law Schools
of the University of Western Ontario, Macquarie University, and the University of Minnesota, and in the
Gerald Gordon Seminar at the University of Glasgow, at which earlier versions of this chapter were presented,
for their helpful comments; particular thanks are due to John Oberdiek, for his editorial guidance and patience.
 REPAIRING HARMS AND ANSWERING FOR WRONGS 213

First, I might face a criminal charge of arson. If I go to trial, and the prosecution
proves that my fire destroyed her garage and that I was reckless as to the risk of such
damage, I will be convicted.1 My sentence will depend on the seriousness of the crime:
the seriousness of the risk that I created, the damage that I caused, and my culpability
in creating it. My sentence, let us suppose, is a fine of $3,000, which I pay not to her,
but to the state.2
Second, she could sue me for damages in a civil court. If we go to court and she
proves that my fire caused the damage to her garage and that my conduct in lighting it
was faulty (typically, that I was negligent, failing to exercise care that I should have
exercised), the court will find in her favor. So long as I do not succeed in a counter-
claim, for instance of contributory negligence on her part, the court will award her the
full costs of making good the harm I caused: the cost of rebuilding the garage and of
replacing its contents. Let us suppose that the court fixes the amount of damages at
$3,000.
The two legal processes are sparked off by the same course of conduct and the same
salient effect of that conduct (my lighting the bonfire, the destruction of her garage).
They have, from my point of view, the same salient material outcome: a court orders
me to pay $3,000. If we are to understand the relationship between them, and the aims
that each can properly serve, we must take account of some of the obvious differences
between them. Five are worth noting: I describe them in over-simplified terms, since
my aim is to identify ideal types of each process.
First, a civil case is controlled by the plaintiff. My neighbor decides whether to sue,
whether to pursue the case, or to settle it if a settlement is offered, whether to enforce
any judgment in her favor. The material and psychological burdens of bringing the
case also fall on her. By contrast, a criminal case is controlled by criminal justice
officials: the police decide whether to investigate the matter; a prosecutor decides
whether and with what to charge me and negotiates any bargain that will avert a trial;
the sentence is enforced by officials. As the injured party, she might have de facto
influence on the process: the police might become aware of the crime only if she
reports it; they might not pursue it if we assure them that we are sorting it out
informally; the prosecutor might not charge me if she asks her not to. But she has
no formal standing to demand that I be, or not be, prosecuted or to influence my
sentence.3
Second, both cases aim to determine responsibility for the destruction of the garage,
in order to determine liability to a legal consequence; in both, the question of whether
the destruction was caused by my bonfire is central to that determination of respon-
sibility. But the determination of criminal liability depends on culpability in ways that

1
Criminal Damage Act 1971, s. 1; see also Model Penal Code } 220.3.
2
I leave compensation orders (see Powers of Criminal Courts (Sentencing) Act 2000, ss. 130–3) aside here.
3
This contrast is radically over-simplified: apart from the possibility of private prosecutions, some legal
systems allow the complainant a formal role in the initiation of criminal proceedings. As noted above, however,
my aim here is to identify ideal types of each process.
214 R . A . DUFF

the determination of civil liability does not. My civil liability might be established by
proving simply that my conduct fell below an objective standard of reasonable care; in
the criminal case, more is needed—in English law, that I was at least reckless as to the
risk of damage. A range of defenses are also available in the criminal case that are
unavailable in the civil case. I can avoid civil liability by showing that she consented to
my action or that I was defending myself against her unlawful attack: but it would not
avail me in the civil court, as it would in the criminal court, to show that I acted under
duress, even if the court agreed that it rendered my action excusable, or that I was so
mentally disordered as not to be criminally responsible for my actions.4 Indeed,
I might still be liable to pay for the damage to her shed even if the court found that
my action was justified by necessity.5 I could also mitigate my civil liability, but not my
criminal liability, by showing that her negligence contributed to the damage.
Third, the liability imposed by a civil court that finds for the plaintiff is, typically, a
liability to pay damages; in simple cases, the amount of damages is determined by the
cost of repairing the harm for which the defendant is held responsible. If it costs $3,000
to rebuild my neighbor’s garage and to restore, as far as is possible, the status quo ante,
that is what the court will award, regardless of my means and of the extent of my
culpability: it will matter neither whether I am rich or poor nor whether my conduct
was only modestly negligent or more seriously reckless.6 The liability imposed by a
criminal court, by contrast, is a liability to be punished, and in systems that take
culpability and proportionality seriously the severity of that punishment is determined
not by the cost of repairing whatever harm was caused, but by the seriousness of the
offense—orthodoxly understood as a function of the culpability of the agent and the
seriousness of the harm caused. If the punishment is a fine, it is also appropriate to
proportion its cash value to the offender’s means, since the fine’s penal severity
depends not on its absolute amount, but on what proportion of the defendant’s
resources it takes.
Fourth, in both cases a full trial might be avoided by pre-trial negotiation: a civil case
might be settled before it comes to trial; a criminal case might be (indeed in Britain and
the US usually is) decided by a guilty plea, which itself often results from pre-trial
bargaining, or be disposed of by a police caution or a “prosecutor fine,”7 or be
informally diverted from the criminal process. However, settlements in civil cases
often explicitly exclude any admission of liability, whereas a criminal plea bargain still
requires the defendant formally to plead guilty in court, and cautions similarly require
an admission of guilt.8

4
Nor would “unfitness to plead” bar a civil trial. See John Goudkamp, Tort Law Defences (Oxford: Hart
Publishing, 2013), Chapter 19, 472–3. I leave aside here “public policy” defenses in tort law.
5
See Vincent v Lake Erie Transportation Co., 109 Minn. 456, 124 N.W. 221 (1910). But see text at notes
26–27 of this chapter.
6
But see text at notes 36–8 (concerning punitive damages for willful or malicious conduct).
7
See Peter Duff, “The Prosecutor Fine,” 14 Oxford Journal of Legal Studies 565 (1994) (on prosecutor fines).
8
That is not true, however, of prosecutor fines. See Duff (note 7) at 574.
 REPAIRING HARMS AND ANSWERING FOR WRONGS 215

Fifth, we can insure ourselves against civil liability but not against criminal liability.9
If I am insured, my insurers might seek to settle without going to trial, in which case
I could take the case to trial only by giving up my insurance coverage; if my neighbor
was insured, her insurers might do the same. But whether the case is settled pre-trial or
by a verdict for my neighbor, the money will be paid by my insurers, not by me. The
affair will not cost me nothing: apart from the burdens of filling out an insurance form
and, perhaps, having to appear in court, I might find that my insurance premiums
have increased, or find it harder to obtain insurance, or have to pay the first $N of
the award; but I will not have to pay the $3,000 that the court awards her. But if
I am convicted in the criminal court, I must undertake or undergo the punishment:
I cannot expect my insurers to pay the fine (though the law cannot prevent someone
else paying it).
These contrasts suggest a familiar distinction of function between these two kinds of
legal provisions. The criminal law is concerned with wrongs, albeit, perhaps, only with
harmful wrongs. The criminal process aims to call alleged wrongdoers to account, to
censure, and to punish them if they are proved guilty. By contrast, tort law is
concerned with harms, albeit, perhaps, only with wrongful harms, and where their
costs should fall. A civil process aims to allocate the costs of harm, and to ensure as far
as possible that they are allocated away from those who innocently suffered them and
onto those who are properly held responsible for causing them.
To spell this out in a little more detail, the substantive criminal law’s definitions of
offenses purport to define wrongs that merit the polity’s formal condemnation. This
does not imply a form of positive legal moralism, according to which it is the
wrongfulness of any type of conduct that generates our positive reason for criminal-
izing it; all it requires is the negative legal moralist thesis that conduct can be
legitimately criminalized only if it is wrongful.10 In its procedural mode, the criminal
law then provides for a response to the alleged commission of such wrongs—one
which makes their wrongfulness salient: the criminal trial calls the defendant to answer
to a charge of criminal wrongdoing; a criminal conviction, which is warranted by
proof that he committed the offence charged, unless he can offer an exculpatory
defense for that commission, condemns his conduct as wrongful; the punishment
that normally follows conviction similarly communicates the censure that the crime
warrants.11 This need not imply a robust retributivism, according to which the
justifying purpose of criminal law is to impose deserved suffering on wrongdoers—it
leaves open the possibility that its justifying aim is, for instance, the prevention of
future harms of the kind that criminalized conduct might cause. It is, rather, to say that
what is distinctive of criminal law, as a particular mode of legal regulation, is its focus
on wrongdoing: if its goal is to prevent harm, it pursues that goal by defining a set of

9
In which context it is significant that insurance policies will not provide cover for punitive damages.
10
See also R.A. Duff, “Towards a Modest Legal Moralism,” 8 Criminal Law and Philosophy (forthcoming
2014).
11
See also R.A. Duff, Answering for Crime (Oxford: Hart Publishing, 2007).
216 R . A . DUFF

wrongs and by calling those who commit them to such condemnatory account. Those
who accept the Harm Principle will insist that only harmful wrongs may be crimin-
alized: but if what they offer is an account of criminal law, their focus must be on the
wrongfulness of such harmful conduct; the criminal law’s concern is with harmful
wrongs.
By contrast, tort law is focused on harms: typically, harms that have occurred and
can (to at least some extent) be repaired. The law enables one who suffers harm at
another’s hands to shift the cost of that harm onto the other person, by proving that he
was legally responsible for it. The law might include wrongdoing and fault as condi-
tions of responsibility, requiring the plaintiff to prove not only causation, but also a
breach of duty and at least negligence. Tort law would then be concerned not with
harms as such, but with wrongful harms: nonetheless, its focus is on the harm and
what is needed to repair it. In criminal law, the object of liability is the wrong: we are
liable for the wrong we committed (even if it should be a condition of criminal liability
that the wrong is harmful). In tort law as portrayed here, by contrast, the object of
liability is the harm: we are held liable for the harm we caused, even if it is a condition
of being thus held liable that our harm-causing conduct was wrongful.12 Sometimes
the harm can be prevented, for instance by an injunction, or reversed, as when
misappropriated goods are returned. But the focus is on remedying harm, and on
allocating its costs if it cannot be prevented. Further questions then arise about the
principles that should determine such cost-allocations: should they be grounded, for
instance, in considerations of economic efficiency, or of justice? Such questions do not
concern us here, however: for they presuppose a cost-allocation model of tort law.
This simplified, idealized distinction between “criminal” law and “tort” law neatly
explains the various differentia noted above. First, if the question is whether the cost of
harm is to be shifted from the person who suffers it, we can leave the process under her
control. The polity should enable those who suffer harm at others’ hands to seek
compensation, by providing a legal process to adjudicate their claims, and enforcing
judgments in their favor: but it is for the person who suffered harm to decide whether
to pursue the matter; it would be odd to insist that she must pursue it, or must be
compensated, if she does not wish to receive such compensation. If, however, the
question is whether someone is to be called to account, and condemned and punished,
for committing a wrong that merits public condemnation, the victim’s wishes should
not be decisive. A crime violates the values embodied in the criminal law: the response
to it should be public not merely (as is true of torts) in the sense that it involves an
authoritative judgment by a public court, but in the sense that it is decided and carried
through by the polity’s officials.13 I seek to recoup the costs of repairing the harm that

12
See Duff, Answering for Crime (note 11) at 82–9 (on objects and conditions of liability).
13
This is what it means to say that crimes are “public” wrongs: they are wrongs that rightly concern not
only their direct victim, but the whole polity. See S.E. Marshall and R.A. Duff, “Criminalization and Sharing
Wrongs,” 11 Canadian Journal of Law and Jurisprudence 7 (1998).
 REPAIRING HARMS AND ANSWERING FOR WRONGS 217

I suffered from the person who was responsible for it; we call to punitive account the
person who violated our shared public values.
Second, although there is plenty of room for argument about the proper grounds of
liability in both the civil and the criminal context, we can see why a civil process might
plausibly involve a less demanding standard of fault.14 If harm has been caused, the
cost of that harm must fall on someone. We could leave it to fall on the person harmed,
relying on others’ charity to help those ill-equipped to bear that cost. Or we could
create an insurance system to cover the most serious kinds of harm. Or we could, as a
matter of justice or of economic efficiency, provide a mechanism through which the
costs can be shifted from those who suffered the harm to those responsible for causing
it, at least when the causers were not “innocent” in the way that the sufferer was. If it is
my fault that my neighbor’s garage was destroyed, it is fair that I bear the cost of its
repair, but “fault” need not imply here the censurable culpability that it implies in the
criminal context: the fact that my conduct fell below a suitable standard of care could
suffice to warrant shifting the cost from her to me. When what is at stake is criminal
conviction and punishment, however, we are not allocating a burden that must fall
somewhere; given the burden’s condemnatory meaning, it should be imposed only on
those who culpably deserve it. That is why defenses that ward off criminal liability
might not ward off civil liability, unless they show that the person who suffered the
harm was responsible for it (as when the defendant can argue consent or self-defense).
Third, since the point of the civil process is to repair, or provide compensation for,
harm that was caused, damages are payable only if harm was caused, to the amount
necessary to repair it: a reckless bonfire-setting that fortuitously caused no harm will
not attract civil liability; and whilst it might be a matter of luck that Jones caused
damage costing $10,000, whereas Smith’s equally reckless conduct caused damage
costing only $1,000, Jones cannot complain that it is unfair that he must pay ten times
more than Smith. But since a criminal process focuses on condemning and punishing
wrongs, a reckless risk-creator who luckily causes no harm could still face criminal
liability for creating a risk;15 whilst it is controversial whether the fact or the extent of
criminal liability should depend on the fact or extent of actual harm, Jones should not
be punished ten times as harshly as Smith; and fines should be proportioned to the
offender’s means.
Fourth, we can see not only why pre-trial settlements should be encouraged in civil
cases (so long as the weaker party is protected from undue pressure), but also why
there need be no admission of liability. The point of assigning civil liability is to
allocate costs; if they can be allocated, in a way that satisfies both parties, without any
formal determination of liability, there is no reason to insist on such a formal
determination. The person who suffered harm might not receive as much as he
would have secured had the case gone to trial; but it is rational for him to settle
on the basis of a calculation of the chances of winning the case and the costs of

14
And, for similar reasons, a less demanding standard of proof.
15
See, e.g., Model Penal Code } 220.1(2)(b) (“Reckless Burning”).
218 R . A . DUFF

pursuing it. In criminal cases, by contrast, plea-bargaining is controversial, and

something seems amiss if a criminal defendant is allowed to pay a penalty without
admitting liability. For the point of the criminal process is to determine whether the
defendant is guilty of the wrong of which he is accused, and to impose an appropriate
punishment if he is proved guilty: the determination of liability is thus central to the
enterprise—a goal that plea-bargaining undercuts insofar as it distorts or conceals the
truth about the defendant’s guilt.
Fifth and finally, we can see why it can be appropriate to allow, or require, people to
insure themselves against civil liability but not against criminal liability. In a civil case,
the aim is to secure compensation for the person harmed: the defendant who caused
the harm may be ordered to pay damages; but what matters is not so much that she
bear the cost as that she is responsible for ensuring that the victim is compensated. She
can discharge her responsibility either by paying the damages, or by ensuring that they
will be paid by insuring herself; indeed, people engaging in activities involving
significant risks of harms may be required to insure themselves to ensure that compen-
sation can be paid. By contrast, the aim of a criminal case is to ensure that culpable
wrongdoers are held to account, and punitively censured, for their wrongs: that aim is
achieved only if the wrongdoer stands trial and suffers the punitive censure of
conviction. The conceptual point here is that I am not punished if I have an insurance
policy that pays my fine; the normative point is that the aim of the process is only
achieved if the offender bears the punitive burden.
So far, it might seem, so good: we have two kinds of legal process, criminal and civil,
each serving a distinct purpose to which it is suitably adapted. But we must now note a
significant gap in this structure—a gap that “civil recourse” offers to fill.

III. A Gap—and How Civil Recourse Can Fill It

If I have been wronged, I might seek a remedy: some response from the wrongdoer and
others that recognizes the wrong and its implications. Outside the law, we have
familiar ways of responding to wrongs. I expect my friends to respond with sympa-
thetic indignation: not just with sympathy for any harm I have suffered, but with a
recognition of the wrong done to me. I might challenge the person who wronged me,
and hope that my friends or colleagues, others involved in the form of life in which the
wrong was done, will support me. If I challenge the wrongdoer, I look for a particular
kind of response. Perhaps he will explain that my challenge is misdirected—he didn’t
do what I think he did; or he might offer a justification or excuse; or he might admit
the wrong. In that last case, I will look for more than mere admission, more than a
simple offer to pay for whatever harm was caused: the wrong requires apology, and
offers of repair or compensation will be suitable only if they serve to give substantive
force to the apology that they accompany.
Many of the wrongs that we commit and suffer should not concern the law: they are
properly dealt with in the extra-legal contexts in which they occur, by those directly
 REPAIRING HARMS AND ANSWERING FOR WRONGS 219

involved. But some wrongs should concern the law—wrongs that involve the violation
of a right, or the impairment of an interest, that the law should protect. What should
the victim of such a wrong expect from the law? Here is where we find the gap that civil
recourse is to fill.
If the victim looks to the criminal law, she finds that she loses the standing that she
has in the extralegal contexts in which we call those who wrong us to account. When
I am prosecuted for criminal damage, the case is brought not by my neighbor, but by
the polity; I am called to answer not to her, but to the polity; she has no formal role in
the case; if I am convicted and required to pay a penal debt (literally through a fine, or
metaphorically through “community payback” or imprisonment), that debt is owed
and paid not to her, but to the polity. Christie famously argued that the criminal law
“steals” the “conflicts” it treats as crimes from those to whom they belong: the
“victims,” the “offenders,” and those close to them.16 There is much with which to
quarrel in his portrayal of crimes as conflicts, but the criticism that our criminal
process deprives victims of their proper standing resonates with the complaints often
made from different normative starting points by victims’ groups—that they are
allowed no formal role in the criminal process. When I respond extralegally to a
wrong, I respond not as a patient to whom something has been done and something is
now owed, but as an agent who seeks what is due to him; but when the case is taken
over, as a criminal case, by the criminal law’s officials, I become a mere patient.17
If the victim then looks to the civil law, she will find herself as an agent again: it is for
her to bring a case, to decide whether and how to pursue it. However, the kind of case
that tort law (as portrayed in the previous section) allows her to bring might not be the
kind of case that she wants (or should want) to bring. What she wants is not (simply)
to shift the cost of repairing some harm she suffered onto the person who was
responsible for it, or to obtain compensation for that harm, but to call that person
to account for the wrong that he did her. More precisely, since it is no part of the law’s
function simply to give people what they want, this is something the victim could
reasonably seek, and something that a law that takes wrongdoing seriously should
enable her to obtain. Two of many possible examples will illustrate this point.18
In countries in which libel is not criminal, my only legal recourse if I am libeled is a
tort suit: but what can a cost-allocative tort law offer me? If my case is to fit the cost-
allocation paradigm, I must show that the lies published about me caused me harm to
which a monetary value can be attached, or for which a monetary payment could
compensate me. Now there might sometimes be a financially quantifiable loss (per-
haps the libel lost me business), but this is not usually what concerns a libel plaintiff:

16
Nils Christie, “Conflicts as Property,” 17 British Journal of Criminology 1 (1977).
17
To which it might be replied, (see note 3), that some European systems give the victim a more active
formal role in the criminal process; and that if there is a gap, of the kind identified in this section, in our
existing legal provisions, it should be filled not by resorting to a “civil recourse” mode of tort law, but by so
reforming the criminal process that it can give victims their due. That possibility, however, must be a topic for
another time.
18
See also Ashley v Chief Constable of Sussex Police [2008] 1 AC 962 for another good example.
220 R . A . DUFF

I do not sue because my good name is a financially valuable asset. As for monetary
compensation, we must wonder whether a monetary payment could, as such, be an
appropriate compensation. When it is said that “all right-minded persons agree that a
good name is precious beyond rubies, but Una would have preferred less honor and
more rubies,”19 we are to think Una misguided: honor and rubies are not thus fungible.
I want a correction as public as the original falsehood: but mere corrections belong
with reasonable mistakes rather than wrongful libels. For the libel, an appropriate
mode of reparation is an apology: but apologies differ from monetary compensation in
that while we can separate money, as compensation, from its source, we cannot
separate apology from its source. If $10,000 would compensate me for the harm that
I suffered, what matters is that I receive that amount: who should pay it is a further
question. But if I am owed apology, it must come from the person who wronged me:
insurers cannot apologize on his behalf. A New York firm offers corporations help
with apologies:
Sorry should not be the hardest word, not least because acknowledging poor service can
actually enhance customer or staff loyalty. We can provide an appropriately structured
corporate apology programme.20

Now it is possible (whether or not desirable) for someone else to script my apology: but
I cannot pay someone else to apologize to you for what I did; apology must come from
the wrongdoer, in the first person.
Second, consider those tragic cases in which a child dies as a result of medical
negligence, and the parents sue the hospital or doctor. There may be costs associated
with the death that the defendant can be required to pay, but the court can also award
“damages for bereavement,” set at £3,500 in English law.21 The parents might com-
plain that it is outrageous to value their child’s life at £3,500; but we must be clear
about the meaning of such a complaint. If it is expressed as the complaint that the
child’s life is valued at “only £3,500,” that might suggest that a larger sum—£35,000, or
£350,000, or £3,500,000—would be at least closer to adequacy. But that is not (we
hope) what such parents mean. Sometimes, even if money cannot fully repair the
harm, and no payment could leave a plaintiff as content as he would have been had the
harm not occurred, we can see a payment as incompletely adequate, or as an appro-
priate consolation. It is incompletely adequate when the compensation is incomplete
(I would prefer not to have suffered the harm), but falls short of what would constitute
complete compensation; it is consolatory if it enables me to say “Well, I lost X, but at
least I got £N.” There are interesting questions about when “but at least I got £N” is an

19
Sinclair Lewis, The Job (1917), 14; Proverbs 22.1 (“A good name is rather to be chosen than great riches.”).
20
Michael C. Fina, quoted by Jeremy Waldron, “Where Money and Markets Don’t Belong,” in New York
Review of Books, August 16, 2012 (LIX.13), p. 65. The firm’s current website (<http://www.mcfrecognition.
com/service-recognition/products/corporate-apology> accessed November 5, 2013) offers a “Sincerely Sorry”
service, which “provides customer-facing operatives with the agency to deflect potential business loss scenarios
by sending thoughtful gifts and/or apology cards when the occasion arises.”
21
Administration of Justice Act 1982, s. 1.
 REPAIRING HARMS AND ANSWERING FOR WRONGS 221

appropriate response, and when it marks a radical mis-valuing of X, but it is not an

appropriate reaction for the grieving parents whose child died through medical
negligence. It is inappropriate because it would seriously mis-portray the loss they
have suffered, and does not address the wrong that was done—to their child and to
them; and although they must sue in their own names, they might say that they are
suing for the child, to whom no monetary payment is possible.
What the parents properly want is an accounting, and an apologetic recognition not
merely of the harm caused, but of the wrong done: an explanation, from the people
concerned, of how their child died, of what went wrong; and if what went wrong
included negligence, an admission of that. But such an accounting, such apologetic
recognition of wrongdoing, cannot fit within a system designed to allocate the costs of
harm. This is not to say that monetary damages could not figure in cases of libel or of
wrongful death: only that we cannot understand such damages as allocating the costs
of harm. As we will see, damages in such cases make more sense if we focus on their
relation not to the harm suffered, but to the wrong done; not merely on what the
victim receives, but on what the wrongdoer must pay. Whilst “I lost my son/was
seriously libeled, but at least I got £N” is an inappropriate response, “I lost my son, but
at least the doctor whose negligence killed him had to pay £N,” or “I was libeled, but at
least the libeler had to pay £N,” is not so inappropriate. This will require us, however,
to understand monetary damages in different terms.
This is the gap that is left by the two kinds of legal process sketched in the previous
section: they have no place for the wronged person who seeks recourse, in her own
name, for the wrong she has suffered. That seems a serious gap; but it can be filled by
“civil recourse.” Civil recourse, as portrayed by Goldberg and Zipursky (hereafter,
Goldursky),22 is a process through which one who is wronged can bring a suit against
the wrongdoer. The primary purpose of the suit is not to secure damages or compen-
sation, as it is on the cost-allocation model of tort law: it is to hold the wrongdoer to
account for the wrong and to secure from the court a verdict that he is liable; only after
such a finding of liability does the question of a remedy for that wrong arise. Gold-
ursky offer “civil recourse” as a theory of our actual tort law, but my concern here is
with the theory’s normative attractions as portraying a distinctive legal process, one
that enables us to realize some of the values of responsible agency and citizenship.
Those attractions should be obvious. Civil recourse offers the wronged individual a
remedy for the wrong she suffered; it enables her to hold the person who wronged her
to account through a public legal process. Unlike the criminal process, she is in charge:
she decides whether to sue, whether to settle, and whether to enforce any award in her
favor. Unlike a civil process that deals only with allocating the costs of harm, however,

22
See, e.g., Benjamin C. Zipursky, “Rights, Wrongs, and Recourse in the Law of Torts,” 51 Vand. L. Rev. 1
(1998); Benjamin C. Zipursky, “Civil Recourse, Not Corrective Justice,” 91 Geo. L.J. 695 (2003); Benjamin
C. Zipursky, “A Theory of Punitive Damages,” 84 Tex. L. Rev. 105 (2005); Benjamin C. Zipursky, “Substantive
Standing, Civil Recourse, and Corrective Justice,” 39 Fla. St. U. L. Rev. 299 (2011); John C.P.Goldberg and
Benjamin C. Zipursky, “Torts as Wrongs,” 88 Tex. L. Rev. 917 (2010); John C.P. Goldberg and Benjamin
C. Zipursky, Torts (New York: Oxford University Press, 2010).
222 R . A . DUFF

the process focuses on the wrong she suffered: the harm (if any) caused by that wrong
will also figure in the process; but that is because it bears on the seriousness and
character of the wrong. It is also true that if the plaintiff wins her case, the most
common remedy will be an award of monetary damages, usually a sum sufficient to
“make her whole”—to repair the harm she suffered.23 However, apart from the facts
that other remedies are available, and that (as we will see) the significance of such
damages is different from that of a mere reallocation of costs, on this model the
remedy is not the initial focus of the process: the first question is not whether or how
much the defendant must pay the plaintiff; it is whether he is liable for the wrong that
the plaintiff alleges he did to her. If he is liable, the question of remedy arises: but the
finding of liability, the verdict that the defendant is responsible for the wrong that the
plaintiff alleged, is the prior point of the process; and its significance does not lie
simply in its role as ground for the award of a remedy. What matters initially is that the
wronged plaintiff can hold the person who wronged her to account; the remedy comes
later.
(It might be thought that I have created the appearance of a gap that civil recourse
could then fill only by ignoring corrective justice theories of tort law. I leave aside here
the debates between theorists of corrective justice and of civil recourse, and the
question of whether civil recourse is independent of corrective justice—though civil
recourse surely gives particular procedural form to one possible conception of cor-
rective justice; my concern is with aspects of the civil recourse model which do not
depend on its relationship to different conceptions of corrective justice.)

IV. Questions about Civil Recourse:

(i) The Accounting
There is much to be said for civil recourse: it satisfies a need unmet by the criminal and
cost-allocative processes sketched above; it gives wronged victims the power to seek
appropriate recourse. But more must be known about the model before we can see
what role it should play in a legal system, and just what recourse it offers such victims.
I focus here on two questions: one about the accounting it seeks, the other about the
meanings of the remedies to which it can lead. Civil recourse enables someone who has
been wronged to call the wrongdoer to account, but we must ask more carefully about
who can be called and what kind of account he can give. We should note first, however,
that the civil recourse model makes pre-trial settlements which involve no admission
of liability more problematic than does a cost-allocation model. If the aim is to call the
person who wronged me to account, and the remedy (damages, for instance) comes
into play only once liability has been determined, the determination of liability must
be the first step; a wrongdoer is not called to account if she secures a settlement that

23
But courts can award “punitive damages” for malicious or willful wrongs; see further text at notes 36–8.
 REPAIRING HARMS AND ANSWERING FOR WRONGS 223

involves no admission of liability. Plea bargains are already controversial, but at least
what is bargained is a plea—an admission of liability; a liability-free settlement misses
the point of civil recourse. Goldursky are aware of this problem, but point out that the
plaintiff is not compelled to settle, since she could insist on taking the case to trial; and
“settlements can themselves empower victims and impose responsibility.”24 But this is
unpersuasive: as we saw in Section III, the power that victims need is not the power to
secure a payment but the power to hold to account, which requires an accounting—a
determination or admission of liability; a liability-free settlement, especially when
negotiated by insurers, is not a process that holds the wrongdoer accountable, respon-
sible, for his wrongdoing. The plaintiff might settle because she thinks it more
important to secure the money on offer, as partial compensation for the harm; but
she then abandons civil recourse, in favor of partial cost-allocation.
As to who can be called to account, the obvious answer is that it is the (alleged)
wrongdoer, but there are some complications. Suppose, first, that the wrongdoer has
died since he committed the wrong. If it was just a matter of allocating the costs of the
harm that he caused, the plaintiff could sue his estate: by negligently causing the harm,
he incurred a duty to pay damages; while he cannot be sued in person for that debt, it
can be a charge on his estate. But if it is a matter of calling a wrongdoer to account,
death must bar the suit: a dead person cannot be held to account; nor can his heirs,
who inherit his debts, be called to account for his wrong. In this respect, civil recourse
resembles the criminal law: it requires a living defendant who can be called to account.
Suppose, second, that the wrongdoer has, since his wrongdoing, become so dis-
ordered that he cannot understand, or play any active role in, the proceedings. His
disorder is no bar to a civil suit that seeks to reallocate the cost of harm: if the plaintiff
can prove that he was liable for the harm, the court can award her appropriate
damages, to be taken from the defendant’s property. But if the process is to call the
defendant to account for his wrongdoing, his disorder must bar the case from
proceeding: we cannot call someone to account if he is incapable of understanding
or responding to the call. Here, again, civil recourse must resemble the criminal law:
the defendant must be fit for trial—able to answer to the charge of wrongdoing that he
faces.25
As to the kind of account that a defendant must be allowed to give, we must
consider some of the defenses that are available at a criminal trial, but unavailable in
a cost-allocative civil trial.
Suppose first that the defendant wants to offer a justification for his conduct: not
one based on the plaintiff ’s own conduct, which aims to shift responsibility for the
harm onto the plaintiff (consent, self-defense, or assumption of risk), but one that,
whilst not denying full responsibility for the harm, argues that the defendant acted
justifiably in causing it. If the case is simply about where the cost of the harm should
fall, it can be argued that it should fall on the defendant who caused the harm, however

24
Zipursky and Goldberg, Torts (note 22) at 67.
25
See Duff, Answering for Crime (note 11) at 179–81.
224 R . A . DUFF

justifiable his conduct;26 in which case, the defendant’s justification cannot be heard in
court, since it is irrelevant to the case. If, however, the case is a matter of civil recourse,
of calling the defendant to account for his alleged wrong, his justificatory explanation
is relevant and should be heard. Perhaps the defendant should still bear the cost of the
harm: we might say that even if his conduct were justified, the plaintiff was wronged,
and the cost should fall on him—although that raises questions about how justification
bears on wrongdoing.27 But if I am called to account for an alleged wrong to you, my
justificatory explanation, which shows my conduct to have been entirely reasonable, is
relevant to the accounting: it makes a difference to the character of that alleged wrong
and to my relationship, as its agent, to you as its patient. Once we separate the calling
to account from the remedy, as Goldursky do; once we insist that the case is not about
allocating the cost of harm, but about the wrong that the defendant allegedly did to the
plaintiff: we must take the accounting seriously and allow the defendant’s justificatory
account to be heard. (The remedy for the wrong should also surely be sensitive to
whether it was a justified wrong; I return to the question of remedies in Section V.)
A similar point applies to non-justificatory defenses: to what would count in
criminal law as excuses or exemptions. If you call me to account for a wrong, my
response might be to admit the wrong, but to explain that I was acting under a threat
that, while not enough to justify my action, should spare me from blame; or that my
capacity for rational action was so impaired by mental disorder that I cannot be held
culpably responsible for my conduct.28 Such explanations might not save me from
having to pay for the harm that I caused: but they should make a difference to our
understanding of the wrong that I did you and of my relationship as the agent of that
wrong to you as its victim. If you sue me as a matter of civil recourse, in order to hold
me to account for that wrong, you and the court should be willing to listen to such an
exculpatory account.
In these respects too, civil recourse should resemble criminal law: it should make
room for kinds of defense—justifications, excuses, and exemptions—that are relevant
in a criminal trial. They might not make a difference to a decision about who should
bear the cost of any harm that was caused: but that is not the primary focus of civil
recourse; if the process is to hold the defendant to account for a wrong he allegedly did
to the plaintiff, it must find room for accountings that affect our understanding of that
wrong—for defenses that would preclude criminal conviction.

26
Vincent, 124 N.W. at 221. See Goldberg and Zipursky, Torts (note 22) at 238–41. Goldberg and Zipursky
reject the orthodox analysis, according to which necessity grants an “incomplete privilege” to act as the
defendant did—incomplete because he must still compensate for the harm he causes.
27
We might say that the plaintiff ’s rights were infringed but not violated. But see John Oberdiek, “Lost in
Moral Space: On the Infringing/Violating Distinction and its Place in the Theory of Rights,” 23 Law and
Philosophy 325 (2004). Goldursky insist that in Vincent the defendant did commit the wrong of trespass, “for
which compensation was owed”: “reasonable trespasses . . . are trespasses nonetheless” Goldberg and Zipursky,
Torts (note 22) at 240. But a civil recourse process should still be sensitive to a wrong’s justified character.
28
Excuses admit responsibility for the wrong but deny culpability; exemptions deny responsibility. See John
Gardner, “The Gist of Excuses,” 1 Buff. L. Rev. 575 (1998), 1; Duff, Answering for Crime (note 11) at 284–91.
 REPAIRING HARMS AND ANSWERING FOR WRONGS 225

This leads, however, to a second set of questions about civil recourse. A successful
defense in criminal law saves the defendant from conviction: she is acquitted, and faces
no further legal consequences at the hands of the criminal court. But a defendant in a
civil recourse case might be required to pay for the harm that she caused even if she
can show that her conduct was justified or excusable, and the court might order the
same remedy absent any exculpatory explanation.29 If, however, the focus of the case is
on accountability for a wrong, it is not clear why the remedy should be so insensitive to
factors that make such a difference to the character of that wrong.

V. Questions about Civil Recourse: (ii) The Remedy

Suppose that I damage your newly planted hedge in a way that would make me liable
should you sue. You are away on holiday, and I am able to have the hedge replaced so
efficiently that, if I keep quiet, you will never know what happened. If what matters is
that the cost of the damage falls on me rather than on you, I need not tell you: if the
hedge is repaired, you have suffered no loss for which compensation is still due. So
should we conclude that I need not tell you about it?
Even if the damage was caused by nonculpable accident, we might think that
I should tell you—a thought that becomes stronger as the harmful conduct becomes
more culpable. If I caused harm negligently, “I’m so sorry, I carelessly damaged your
hedge, but I had it repaired” might suffice; something more is needed if I was seriously
reckless; and if I intentionally damaged your hedge out of malice (but then repented
my action), something yet more forceful is needed.30 If we keep our dealings out of the
law’s reach, perhaps the only material compensation required is still that I get the
hedge repaired: but I also owe you an apologetic explanation—an accounting; that is
how I show my recognition not just of the harm you suffered, but also of the wrong
that I did you. We can also see how, especially if the wrong was serious, I might think
that more is needed than a verbal apology and the repair of your hedge: to express my
apology more forcefully, to make amends for my wrong, I might offer to trim the
hedge for you or buy you a suitable “peace offering.”
What then of civil recourse: what kind of remedy might a plaintiff expect for the
wrong he suffered? Sometimes a monetary award sufficient to cover the costs of the
harm he suffered (if that harm can be costed), an award of the same value as he would
receive under a cost-allocation model, will suffice, especially if the defendant’s conduct
was only modestly negligent—when, even if it caused serious damage, its wrongfulness
was minor. Even here, however, if the award is to remedy a wrong, its meaning is not
the same as that of a purely cost-allocative award: for it should mark a suitable
recognition not just of the harm caused, but also of the wrong done. A remedy for

29
The default remedy in civil recourse. See Goldberg and Zipursky, “Torts as Wrongs” (note 22) at 962.
30
If I was justified in damaging it, I still owe you repair and explanation: my apologetic explanation will
express not my wish that I had not acted as I did, but my regret that I had to do it.
226 R . A . DUFF

the wrong you suffered, one that addresses the wrong, should have an apologetic
dimension: if it is a monetary award, it must be understood to constitute an apologetic
payment.
Insurance is, therefore, problematic in civil recourse. If what matters is that the
harmed person receives compensation, then that compensation can be provided by
anyone; it need not come from the person who caused the harm. But if what matters is
that the wronged person receives an apology, to be expressed in a monetary payment,
it must come from the wrongdoer: my insurers cannot apologize in my name. Gold-
ursky suggest that even when an insurer pays out in a tort case, the defendant still faces
some responsibility:31 but telephoning my insurers or filling out the forms hardly
constitutes the “responsive conduct” to which the person I wronged is entitled.32 So
why should a plaintiff who is seeking recourse for a wrong accept damages paid by an
insurer as an adequate remedy; why should the law, which does not allow criminal
offenders to shift the penal cost of their wrongdoing onto insurers, allow a tortious
wrongdoer to shift the remedial cost of his wrongdoing onto an insurer? Allowing
insurance for tort liability seems, instead, to mark a recognition that it is more
important to ensure that the harmed person receives compensation than it is to ensure
that the person who wrongfully caused the harm pays for it—that a suitable (re)
allocation of costs is more important than holding the wrongdoer accountable.
If our concern is with remedies for wrongs, we should ensure that the plaintiff
receives not merely appropriate compensation, but that she receives it from the
defendant. The significance of this point becomes clearer as we move away from
cases in which the harm can be costed, to cases in which such costing seems
inappropriate—as when I sue the doctor whose negligence killed my child. To see
how an award of damages could be anything other than a distortion in such cases, we
must focus not on the fact that the plaintiff receives a sum of money, but on the fact
that the defendant pays that sum to her: he is made to pay for what he has done.
Goldursky emphasize that monetary damages are not always concerned with the
reallocation of costs. A plaintiff might have incurred debts that damages can then pay,
or suffered losses that damages can rectify. But she might also have “experienced
setbacks that cannot be rectified as such but the impact of which money can help
ameliorate . . . A lost limb, a damaged reputation, being rendered paraplegic in a car
accident—all of these support payments that compensate for a kind of harm rather
than make good on a debt or loss.”33 However, money as such cannot “help amelior-
ate” such wrongful injuries. Even if it can ameliorate the loss of a limb or paraplegia, it
cannot ameliorate a damaged reputation or the death of one’s child; and to describe
the harms simply as lost limbs or paraplegia (harms that could have resulted from
natural causes) omits the fact that what requires remedy is the wrong, not just the
harm. We can, however, make sense of monetary damages in such cases by focusing

31
Goldberg and Zipursky, Torts (note 22) at 67.
32
Zipursky, “Substantive Standing” (note 22) at 336.
33
Goldberg and Zipursky, “Torts as Wrongs” (note 22) at 961. I discuss other kinds of remedy below.
 REPAIRING HARMS AND ANSWERING FOR WRONGS 227

not on the plaintiff ’s receipt of money, but on the defendant’s payment of money to
the plaintiff. What can ameliorate is not $10,000 or $100,000 itself, but that the person
who wronged me pays it to me.
We might still wonder how such payment can be appropriate in response to such
wrongs: is it not more important that the defendant apologize for what he has done?
Certainly an apology is required, and a court-ordered apology is one remedy that the
law can provide. It might be argued that only sincere apology can have remedial
value—and that court-ordered apology is unlikely to be sincere. But that would be to
underestimate the importance of ritual: by ordering an apology, the court expresses its
recognition of the wrong done; by undertaking this ritual of apology, the defendant
formally accepts that judgment on his wrongdoing.34 However, mere words are often
not enough to make emphatically clear to those concerned the seriousness of the
wrong to be addressed. By requiring not just apology, but also a substantial payment,
the court can communicate to the defendant (and to others) the seriousness of the
wrong in a language that is hard to ignore;35 by making the payment, the defendant
communicates his own recognition of that wrong.
If what the plaintiff properly seeks in civil recourse is “responsive conduct” from the
person who wronged her; if the remedy that the court orders is to constitute not
merely compensation for harm, but also a remedy for the wrong: what is required is
something apologetic—something whose imposition on the defendant makes clear
that he should apologize, and whose undertaking by the defendant could express that
apology. This demand becomes more stringent as the wrong becomes more serious. If
the defendant’s fault amounted to nothing more than simple negligence, which had
unluckily harmful effects, a finding of liability and an order to pay what is necessary to
repair or provide compensation for the harm might suffice: but as we move up the
scale of fault, the focus will shift more towards apology and what is needed to show a
proper recognition of the wrong.
The damages awarded in civil recourse are, therefore, always in a sense punitive,
even if they do not formally count as “punitive damages”: damages, as remedies for
wrongs, have the same significance as criminal punishment, understood as a commu-
nicative enterprise. The burdensome imposition in which criminal punishment con-
sists is an attempt to communicate to the defendant a recognition of his wrong, and is
the prescribed way in which he can formally communicate his own repentant recog-
nition of that wrong as a kind of moral reparation.36 Damages as remedies for wrongs
should also consist in something designedly burdensome, authoritatively imposed on
a wrongdoer as a proportionate response to his wrong: they should be punitive.
This, then, renders “punitive damages” unmysterious.37 If all civil recourse damages are

34
See also Christopher J. Bennett, The Apology Ritual (Cambridge: Cambridge University Press, 2008). But
see Nick Smith, “Against Court-Ordered Apologies,” 16 New Crim. L.R. 1 (2013), for objections.
35
See M.J. Radin, “Compensation and Commensurability,” 43 Duke L.J. 56 (1993).
36
See R.A. Duff, Punishment, Communication and Community (New York: Oxford University Press, 2001).
37
Zipursky, “A Theory of Punitive Damages” (note 22); Goldberg and Zipursky, Torts (note 22) at 352–9.
228 R . A . DUFF

punitive, what distinguishes “punitive” from ordinary damages is that they involve
heightened penal severity, given the more serious nature of the wrong. Ordinarily, the
fault involved is only negligence; ordinary damages are calculated on that assumption.
But if the wrong was committed maliciously, that aggravating factor justifies increased
damages. We need not talk of allowing the plaintiff to be “vindictive”:38 we can just
note that the always punitive remedy for a civil wrong should increase in severity if the
wrong is more serious. I do not suggest that this rationalizes the amounts actually
awarded as punitive damages: but such an aggravating factor focuses attention on how
much the defendant should pay rather than on how much the plaintiff should receive
(which does invite substantial awards when the defendant is a wealthy corporation).
Whilst civil recourse fills a gap left between criminal law and cost-allocative tort law,
it can do so only if it is like criminal law in its focus on calling wrongdoers to account
and in imposing punitive burdens on wrongdoers who are found liable. If the court is
to provide a remedy for the wrong done by the defendant to the plaintiff, its award
must be intended to burden the defendant: by imposing that burden, the court
communicates its judgment on the wrong; by undertaking that burden, the defendant
makes formal reparation for that wrong.
If this is right, we might ask why civil courts should not have access to the range of
punitive measures available to criminal courts. As Goldursky emphasize,39 they can
offer remedies other than monetary damages, such as declarations of the parties’ rights
or injunctions to desist from wrongful conduct; but why should other punitive
measures not be available? There are obvious reasons why imprisonment and proba-
tion should not be available as civil remedies: they are not directed, as civil remedies
must be, towards the plaintiff; and if we imagine plaintiff-directed modes of such
punishments, we can see that they would give individuals a power over others that they
should not have. But why should civil courts not be able to order a defendant to
undertake a private form of “community payback”—to work unpaid for the plaintiff
for a specified number of hours?40 The answer might be that this would threaten the
formal distance that monetary awards preserve between plaintiff and defendant. The
law of civil recourse, like the criminal law, focuses on our mutual dealings as citizens,
and in a liberal polity citizenship is an important but limited relationship; we should
mistrust remedies that require citizens to engage more intimately with each other.
Monetary awards might be seen as putting a price on that which is beyond price; but
they maintain the formal distance that liberal law requires.
If this account of civil recourse portrays a normatively attractive legal process, it also
raises a set of questions about the place of such a process in the law—questions that
I can only raise here.

38
See Goldberg and Zipursky, Torts (note 22) at 354 (drawing on the classical description of punitive
damages as “vindictive” damages). We could instead say that the damages are “vindicative.” See Ashley [2008]
1 AC (note 18).
39
Zipursky, “Torts as Wrongs” (note 22) at 955, 962–3.
40
See Andrew Ashworth, Sentencing and Criminal Justice (5th ed.) (Oxford: Oxford University Press,
2010), 341.
 REPAIRING HARMS AND ANSWERING FOR WRONGS 229

VI. Civil Recourse, Cost-Allocation,

and Criminal Law
One question concerns civil recourse and tort law. Goldursky offer it as a complete
account of tort law, but although my concern has not been with the interpretive or
normative plausibility of such a claim, the previous two sections should suggest that
tort law neither is, nor should be, concerned only with civil recourse. Sometimes, what
a plaintiff properly wants (and what the law should offer) is indeed a civil recourse
against the person who wronged her: a process through which she can call that person
to public account. But sometimes what properly matters is, instead, simply the
appropriate (re-)allocation of the costs of harm—something that can be achieved
(by settlements without admissions of liability, through the use of insurance) without
a formal calling to account. If the law should serve both these functions, we must ask
whether they can be served by the same process: can tort law function as a process
through which the costs of (wrongfully caused) harms can be suitably allocated, and
through which those who have been wronged can seek civil recourse?
Another question, or set of questions, concerns the relation of civil recourse to
criminal law. I have argued that civil recourse is much more like the criminal process
than its proponents seem to allow: in both cases, someone who is alleged to have
committed a wrong is called to account for it in a court of law; in both cases, if he is
held liable for that wrong, he is liable, also, to suffer the imposition of a legal
consequence (a civil remedy, a criminal punishment) that is essentially punitive in
its meaning. So do we need both kinds of process? If so, how should we decide which
kinds of wrong are apt for which process?
Many criminal wrongs, of course, lack identifiable victims who could pursue, or on
whose behalf others could pursue, civil recourse.41 But if we focus on legally cognizable
wrongs that do have identifiable victims, we must ask what form that legal cognizance
should take. Should it be to empower the victims of such wrongs to seek civil recourse
if they wish? The answer to that question should, presumably, be, in principle, “Yes.” If
I have suffered a wrong of which the law ought to take notice, it ought to enable me to
call the wrongdoer to legal account. But what then of the criminal law? One argument,
reflecting strands in abolitionist thought, might be that when wrongs do have identi-
fiable victims, some version of civil recourse is the appropriate response (so long as
victims can receive adequate help in bringing their cases), since that leaves ownership
of the wrong where it belongs.42 If that is implausible, if at least some such wrongs
should (also) be criminal wrongs, we must ask why, and which. What could give the
polity reason to take the wrong over from the victim (which is what criminal law does),

41
Unless we should say that in such cases the polity is the (only) victim, and thus acts as the plaintiff in a
criminal version of civil recourse.
42
See Christie, “Conflicts as Property” (note 16). Though many abolitionists, including Christie, would also
object to a focus on wrongs, and to the abstract formality of the legal process whether it be criminal or civil.
230 R . A . DUFF

and to render its perpetrator liable not (just) to the kind of remedy that civil recourse
can provide, but to the full range of criminal punishments? Should we focus, in
thinking about reasons to criminalize certain kinds of wrong, on punishment (what
makes a wrong criminalizable is that it merits or requires punishment); or on whether
the wrong should be seen as “our” collective wrong rather than the victim’s wrong—
but what could ground such judgments about the ownership of the wrong?43 Then,
once we have decided which victimizing wrongs to criminalize, should the law also still
offer a separate civil recourse process to the victims of such wrongs; or should we find a
way of giving them a more formal active role in the criminal process—not just as the
passive recipients of compensation orders, but as active participants in the process of
calling the wrongdoer to public account?
These are questions for another time: the most I hope to have achieved in this
chapter is to have shown that civil recourse offers a distinctive, normatively attractive,
legal response to some kinds of victimizing wrongs—one that fills a normative gap left
by cost-allocative models of tort law and orthodox models of criminal law; that it can
do so only because it resembles a criminal law response in some crucial ways—as a
process that calls alleged wrongdoers to public account, and renders them liable to
remedies that are themselves punitive; and that by asking what role civil recourse as
thus understood should play in a legal system, we can also highlight some central
questions about the proper role and aims of the criminal law.

43
See Marshall and Duff, “Criminalization and Sharing Wrongs” (note 13).
 11
Tort Processes and Relational Repair
Linda Radzik

I. Introduction
The last twenty-five years or so of thought about tort law have been remarkably
productive and dynamic, as the dominance of the law and economics model has
been challenged by theories that reintroduce the language of corrective justice. Over
this same time period, theorizing about corrective justice has sprung up in response to
a wide range of social, political, and moral issues. I have in mind work on restorative
theories in criminal justice;1 on postwar justice;2 on truth commissions, political
reconciliation and transitional justice in contexts of regime change;3 on official
apologies as responses to international crises, institutionalized abuse and historical
injustices;4 and on interpersonal apology, the making of amends, and forgiveness as
responses to everyday moral transgressions.5 In these writings, theorists explore what

1
E.g., Nils Christie, “Conflicts as Property,” Br. J. of Criminology. 17.1 (1977), 1–15; Howard Zehr,
Changing Lenses: A New Focus for Crime and Justice (Scottsdale, PA: Herald Press, 1990); R.A. Duff, Trials
and Punishments (New York: Cambridge University Press, 1991); John Braithwaite, “Repentance Rituals and
Restorative Justice,” J. of Political Philosophy 8.1 (2000).
2
Brian Orend, “Jus Post Bellum,” 31 J. of Social Philosophy 1 (2000); Gary J. Bass, “Jus Post Bellum,” 32
Philosophy & Public Affairs 4 (2004); Larry May, After War Ends: A Philosophical Perspective (New York:
Cambridge University Press, 2012).
3
Martha Minow, Between Vengeance and Forgiveness: Facing History after Genocide and Mass Violence
(Boston: Beacon Press, 1998); Robert I. Rotberg and Dennis Thompson (eds.), Truth v. Justice: The Morality of
Truth Commissions (Princeton, NJ: Princeton University Press, 2000); Colleen Murphy, A Moral Theory of
Political Reconciliation (New York: Cambridge University Press, 2010).
4
Jean Harvey, “The Emerging Practice of Institutional Apologies,” 9 Intl. J. of Applied Philosophy 2 (1995);
Aaron Lazare, On Apology (New York: Oxford University Press, 2005); Margaret Urban Walker, “Truth
Telling as Reparations,” 41 Metaphilosophy 4 (2010).
5
Jeffrie G. Murphy and Jean Hampton, Forgiveness and Mercy (New York: Cambridge University Press,
1988); Charles L. Griswold, Forgiveness: A Philosophical Exploration (New York: Cambridge University Press,
2007); Margaret Urban Walker, Moral Repair: Reconstructing Moral Relations after Wrongdoing (New York:
Cambridge University Press, 2007); Nick Smith, I Was Wrong: The Meanings of Apologies (New York:
Cambridge University Press, 2008).
232 LINDA RADZIK

people owe to one another in the aftermath of various kinds of wrongdoing, and
thereby subtly challenge our ideas about what should be included under the heading of
corrective justice. One new strand of thought, which appears in many of these debates,
suggests that corrective justice requires a form of reconciliation or relational repair.
My goal in this chapter is to introduce this idea and explore what it might offer to the
philosophy of tort law.
A reconciliation approach to corrective justice is quite different from the notions of
corrective justice that tend to appear in tort literature. I suspect that at least some of
the differences can be traced to the different ways that philosophers of law and
reconciliation theorists frame the questions they ask. In tort literature, corrective
justice is often presented as the counterpart to criminal justice. Criminal law and
tort law deal with different kinds of societal problems (or different aspects of those
problems) and offer different kinds of solutions. These two realms of law are well
established; each has its distinctive institutions and histories. The task for the phil-
osopher of law, then, is to offer an account of tort law that makes sense of its particular
domain and character. In contrast, a reconciliation approach to corrective justice takes
its topic to be simply: what does justice require in the aftermath of wrongdoing? Recon-
ciliation theorists tend to be moral and political theorists who do not feel beholden to
account for any particular institutions or else deal with contexts in which institutions
have failed to function properly (e.g., writers on transitional justice and criminal
justice reformers). From this point of view, both crimes and torts call out for
corrective justice, as do: the ordinary wrongs of everyday life, systematic institutional
abuse, histories of oppression, and civil war. In other words, where philosophers of
law tend to start with a body of law and set of institutions and ask what form of
justice they serve, a reconciliation theorist will start with an ideal of justice and then
ask whether and how tort law makes a contribution to it. I follow this latter approach
in this chapter.
I present the reconciliation view of corrective justice by contrasting it with a simple,
Aristotelian conception of corrective justice as compensation. I argue that the recon-
ciliation conception is preferable to the Aristotelian model for at least four reasons: it is
rooted in a more complete understanding of what is wrong with wrongdoing and
harm-causing; it recognizes that a wider variety of responses to wrongs and harms are
related to justice; it acknowledges that the work of corrective justice typically requires
the efforts of a number of different parties; and it can capture and explain the intuition
that some resolutions are better than others. In Sections II–V, I develop the reconcili-
ation theory by dealing with cases of ordinary moral wrongdoing. I do this for two
reasons. First, we all have experience with these sorts of situations. Second, if an ideal
of corrective justice—a full righting of wrongs—is possible anywhere, it will be here. In
Section VI, I turn from the theory of corrective justice to torts. While there are aspects
of tort law practices that seem incompatible with the ideal of justice that the rest of the
chapter presents, I suggest that we may be able to understand those practices as
approximating justice in non-ideal contexts.
 TORT PROCESSES AND RELATIONAL REPAIR 233

II. What Stands in Need of Correction?

Let us bring to mind a few cases of everyday moral wrongdoing. Think about cases of
betrayal, such as a friend who reveals a secret or a lover who cheats on her partner.
Think too of everyday cases of cruelty (such as schoolyard bullying) or unfairness
(such as a case where a boss docks an employee’s pay for something that was not his
fault). Think of a neighbor who borrows one’s property and carelessly damages it, or a
friend whose reckless horseplay blackens his friend’s eye. What does corrective justice
require in cases like these? What do the wrongdoers morally owe to their victims?
What can the victims legitimately demand of the wrongdoers?
How we conceive of the righting of moral wrongs depends on how we understand
the wrongs themselves. The Aristotelian conception of corrective justice suggests that
what is wrong with wrongdoing is that it causes harm.6 Wrongdoing takes something
away from the victim, or it damages something that belongs to the victim. It is the
damaged property or the lost pay that must be rectified. The way to correct wrong-
doing, then, is to restore the original good or else compensate the victim for its loss. On
this model, it is natural to conceive of wrongdoing as creating a kind of debt.
Correcting the wrong then becomes a matter of repaying the debt.
But as other commentators have noticed, the Aristotelian account seems to confuse
harms with wrongs.7 Not all harm-causing is wrongful; it might be blameless, justified,
or excused. Nor do all wrongs create harms. For example, even when an attempt to
harm fails, the attempt itself is wrongful. The wrong itself calls out for some form of
response. Furthermore, when wrongs do also cause harm, we should not conclude that
the only thing that is wrong with wrongdoing—the only thing standing in need of
correction—is the harm. If you are responsible but blameless for causing damage to
my property, you may well owe me compensation. If you damage my property
negligently, recklessly, or maliciously, you continue to owe me compensation; but
you owe me more than that. Your culpability for the harm, and not just the harm itself,
requires a response.
At this point, some theorists will respond that what wronging (as opposed to
harming) requires is punishment—that the form of justice required is punitive justice.8
If the moral wrong is also a legal wrong, the punishment will be a legal one; if it is not a
legal wrong, punishment might take the form of social sanctioning or self-imposed
guilt. According to this model, the suffering of punishment corrects the wrongfulness
of wrongful harming, whereas compensation corrects the harmfulness. Strong versions

6
Aristotle, Nicomachean Ethics, trans. T. Irwin, 2d ed. (Indianapolis, IN: Hackett Publishing, 1999), Book V.
7
See, e.g., Seth R.M. Lazar, “Corrective Justice and the Possibility of Rectification,” 11 Ethical Theory and
Moral Practice 4 (2008), 355.
8
See, e.g., Scott Hershovitz, “Corrective Justice for Civil Recourse Theorists,” 39 Fla. St. U. L. Rev. 1 (2011),
118.
234 LINDA RADZIK

of this view, which are usually known as retributivist theories, insist that only punish-
ment can correct wrongfulness.9
For reasons that I cannot adequately explore here, I reject this strong view.10 While
punishment can sometimes play a role in correcting the wrongfulness of wrongdoing,
so can other measures (such as apology). Furthermore, even when punishment is
corrective, it takes some doing to say why this is so. Retributivists rarely give us any
such explanation. Indeed, they frequently insist that it is simply obvious that wrong-
doing can only be corrected by the suffering of the wrongdoer. I suspect that what
underlies this confidence is often a kind of magical thinking: only blood can answer
blood; only suffering can undo suffering. But in what sense could punishment “undo”
wrongdoing? Why doesn’t punishment simply add new suffering to old?
Furthermore, as a response to wrongful harming, retribution, like compensation,
seems to have too narrow a view of what stands in need of correction. Consider the
case of schoolyard bullying. Suppose school authorities punish the bully by suspending
her from school. Suppose further that the bully is forced to pay her victim compen-
sation for having stolen his lunch money. Has the wrong been corrected? Most likely
not. For one thing, the victim continues to have reason to worry that the bully will
mistreat him in the future. The fact that others have inflicted penalties on her and
coerced compensation from her gives him no reason to believe that she has changed
her hostile attitude toward him. The victim continues to have reason to fear and
mistrust her. While the victim has received compensation for the material loss he
suffered, he may well continue to suffer the less material forms of harm that are typical
of bullied children: fear, anger, humiliation, lowered status in his social group, social
isolation and lowered self-esteem. Not only are these harms uncompensated, they are
harms that resist the model of compensation, involving as they do values that are not
fungible with goods that can be transferred from wrongdoer to victim. (The fact that,
in tort settings, people will accept financial compensation for harms such as emotional
distress and loss of reputation should not blind us to the inadequacy of financial
compensation for restoring such losses but instead signal to us that there might be
something else going on in the tort context.)
A reconciliation approach to corrective justice highlights the importance of these
sorts of psychological and interpersonal consequences of wrongful harming. The
bullied child did not merely lose his lunch money. He did not merely experience a
discrete number of unpleasant experiences that stay safely in the past. Instead, his
relationships with his abuser, his school community, and himself were damaged in
ways that threaten his future. Correcting the wrong that was done to him will require
the correction of these ongoing problems.

9
See, e.g., Michael S. Moore, “The Moral Worth of Retribution,” in Ferdinand Schoeman (ed.), Respon-
sibility, Character, and the Moral Emotions: New Essays in Moral Psychology (New York: Cambridge University
Press, 1987).
10
See generally, Linda Radzik, Making Amends: Atonement in Morality, Law, and Politics (New York:
Oxford University Press, 2009), Chapter 2.
 TORT PROCESSES AND RELATIONAL REPAIR 235

Compare a case in which you suffer a black eye as the result of an accident
(something that is no one’s fault) with cases in which your eye is blackened through
someone else’s negligence, recklessness, or malice. Your face is equally bruised in all
four scenarios. But the emotional and interpersonal significance of that injury differs.
In the latter three cases, the black eye is prima facie evidence of another person’s
carelessness, disregard, or hostility toward your interests. It is reasonable for you to feel
insulted by this person’s action and to see him as posing a possible threat to you in the
future. The black eye gives you a reason to reconsider how you relate to him—what
attitudes you should take toward him, whether you should trust him, and whether
you must take precautions against him. These meanings, reasons and attitudes, which
I will refer to as damaged relations, are part of the consequences of the wrongful,
harmful act.
A reconciliation theory of corrective justice insists a proper response to wrongdoing
or harming requires the correction of the damage that the wrong or harm does to the
relationships of the parties involved. This relational damage must be repaired for two
basic reasons. First, insofar as damaged relations are allowed to continue, the wrong
continues. Consider again the case of the bullied child. His money is returned to him.
The bully is punished. But these steps do not guarantee his future safety. They
probably do not remove the social stigma of having been bullied, or restore his self-
esteem. He continues to suffer from the bully’s misdeeds. Second, leaving damaged
relations unrepaired is dangerous. Resentment, fear, humiliation, and distrust threaten
to lead to new wrongs and harms. For this point, the example of bullying is sadly apt.
Not infrequently, bullied children become bullies themselves. In extreme cases, some
turn to self-harming and even suicide.
With its emphasis on relational damage, reconciliation theories of corrective justice
respond to aspects of what is wrong with wrongdoing and harm-causing that are
missed by Aristotelian and retributivist theories. Reconciliation theories are also able
to incorporate those things that Aristotelian and retributivist theories properly notice.
The Aristotelians emphasize the importance of compensating harms, or, at least, those
harms that are compensable through some kind of transfer to the victim. Reconcili-
ation theorists can agree that a principle of compensation is a part of corrective justice;
compensable harms intrinsically call out for compensation. To this, the reconciliation
theorists will add that leaving harms inexcusably uncompensated also damages rela-
tionships. Uncompensated harms are prima facie evidence of a continuing disrespect
for the victim.11 Retributive theorists insist that the wrong must be addressed as well as
the harm. Reconciliation theorists agree. But rather than simply insisting that wrongs
can only be repaired through the punishment of the wrongdoer, the reconciliation
theorist attends more closely to the significance of wrongdoing for the parties
involved—what it means to them, how it has affected them, and what they need in

11
Where this is not true—where uncompensated harms pose no threat to relationships, say, because the
debt is small or the victim has forgiven—the moral importance of compensation is diminished.
236 LINDA RADZIK

order to normalize relations. Punishment might play a role in repairing wrongdoing,

but other remedies may be just as, if not more, important.

III. Damaged Relations and the Moral

Obligation to Repair Them
So far, I have left reconciliation theory rather vague. What is a relationship? What sorts
of relationships ought we to have? What does it mean to repair a relationship? Which
relationships are we obliged to repair? Versions of reconciliation theory will differ in
their answers to these questions. In this section, I will fill out my preferred version of
reconciliation theory.12
A relationship with another person consists of patterns of interaction, the attitudes
one tends to take toward that person, and the expectations that one has of the other.
Different kinds of relationships (friend, employee, fellow citizen) are marked by
distinct sets of interactions, attitudes, and expectations. Different kinds of relation-
ships can also be distinguished by the norms to which they should adhere. The duties
we have to our friends differ from the duties we have to our employees. But according
to the broadly Kantian moral theory to which I subscribe, in order for any particular
relationship to be morally acceptable, it must be marked by certain forms of respect
and regard. Each party to the relationship must view the other as an equally valuable
moral person and (other things being equal) be prepared to show him at least a modest
degree of goodwill. When a relationship is morally healthy, each party also believes
that the other person sees her as an equally valuable moral person in return, and
reasonably accepts that the other person can be trusted to treat her with respect and
goodwill.
Wrongdoing affects this relationship in a couple of different ways, which we can
label objective and subjective. First, to morally wrong someone is to fail to treat her
with the respect and concern that she deserves (for a Kantian, this is definitional). The
appropriate terms for the relationship have been objectively violated (that is, they have
been violated whether or not the victim either realizes or cares that they have been).
Second, to morally wrong someone is to give her reason to suspect that one cannot be
trusted to live on terms of mutual respect and regard with her. It is to give her grounds
to subjectively experience both resentment and suspicion. We do not always come
to resent or distrust the people who wrong us; but we frequently do, and the reason for
this is quite straightforward. Indeed, the fact that wrongdoing does not always
undermine trust and goodwill is harder to explain. Even the most resilient relationship
will be undermined by repeated wrongdoing, so it seems fair to say that each wrongful
act poses at least some (perhaps very minor) threat to the relationship. In
sum, objective wrongdoing harms or threatens the subjective moral relationship of

12
For an extended discussion, see Radzik, Making Amends (note 10), Chapters 4 & 5.
 TORT PROCESSES AND RELATIONAL REPAIR 237

mutual respect, goodwill, and reasonable trust between the victim and the wrong-
doer.13 This consequence of wrongdoing is of chief concern to the reconciliation
theorist because of the threat it poses to future wellbeing and peaceful, morally
appropriate relations.
It is important to notice that the subjective moral relationship between the victim
and the wrongdoer is not the only one that might be threatened by wrongdoing. As in
the bullying case, being victimized can sometimes damage the relationship between
the victim and the community. The other children in the class may come to see the
victim as a low sort of person, who is unworthy of equal regard. Wrongdoing can also
damage the relationship between the wrongdoer and the community, as when the
classmates come to regard the bully with suspicion because of her cruel behavior.
Wrongdoing can also result in damage to both the victim’s and the wrongdoer’s
relationships to themselves (if the reader will allow me to speak of a relationship to
oneself). The victim may come to see himself as of lower value than other people. The
bully may come to feel shame over her actions and believe herself to be unworthy of
other people’s regard and trust.
Even mere harm-causing (as opposed to wrongdoing) can threaten subjective moral
relationships. If you harm me in a way for which you deserve no moral blame, but do
not explain to me what happened, then I may have reason to resent and distrust you. If
you leave me burdened by the consequences of your actions, then, even though you did
not wrong me, I may have reason to suspect that you do not regard me with the respect
and concern I deserve. Though the initial harm was not wrongful, your ongoing failure
to compensate me for that harm or to assure me of your respect and goodwill may
become wrongful.
According to reconciliation theory, in order to correct one’s wrongful or harmful
action, one must repair the damage that action caused. In this way, there is an affinity
between reconciliation theory and the Aristotelian notion of compensation. But here,
one’s duty of corrective justice requires one to repair the moral relationships that were
harmed (or shore up those that were threatened) by one’s deed, and not merely to pay
one’s debts. How much one will have to do to repair relationships—whether one must
respond only to the victim or also to third parties, and how much soul-searching and
personal reformation one will have to undertake—will depend on what the effects of
one’s wrongdoing were. To refuse to repair these relationships is to allow one’s deed to
continue doing harm and endangering future relations.
Importantly, reconciling or repairing one’s moral relationship with another person
does not amount to restoring one’s relationship to whatever it was before the wrongful
or harmful action occurred. After all, that relationship might itself have been prob-
lematic; it might not have conformed to moral standards. Instead, corrective justice
requires one to work to establish the sort of moral relationship that should pertain.

13
Subjective relational damage is a consequence rather than a criterion of wrongdoing. The view is not that
an action is wrong only insofar as the parties experience an undermining of mutual respect, regard, or trust.
238 LINDA RADZIK

Critics frequently object to interpreting corrective justice in terms of repairing

relationships because there are so many cases of wronging and harming where there
is no relationship between the victim and the transgressor, or where the victim simply
does not value that relationship. Indeed, tort cases, which are the focus of this volume,
frequently involve people who are strangers to one another except for the conflict
under consideration. Once the case is resolved, they typically never interact again.
However, I am not moved by this objection. First, I think that there is always a
moral relationship of the sort relevant to corrective justice between each and every
person. When people have little or no contact with one another, it is easy to overlook
the relationship; but it is there. The relationship becomes noticeable when the norms
that properly regulate it are breached. Once the breach has occurred, we express
attitudes, emotions, and claims that seem to presume that the relationship was there
all along: “You owed me better treatment than this! I had the right to expect it from
you!” Second, I think people direct a lot of thought and emotional energy to persons
whom they believe wronged or harmed them. The fact that the other person involved
in the car wreck was a stranger and that all one knows about him is that he drove
recklessly does not prevent one from feeling anger and resentment or make it easier to
venture out on the roads again (which requires trust in other strangers). Third, our
conception of corrective justice should not endorse a victim’s lack of concern for her
moral relationship with the person who wronged or harmed her. If a victim does not
want an apology because she feels no resentment, that is probably fine.14 There is no
need to repair an aspect of a relationship that was not impaired. But if a victim does
not care about improving her relationship with the transgressor because she feels
nothing but contempt toward him, then she denies him the recognition of his moral
personhood that is his due. Contempt is occasionally excusable, where wrongdoing
was extreme, but it is always morally worrying.
One might object that, though this defense of the ubiquity of moral relationships is
true of individuals, it is not true of collectives, such as corporations or institutions.15
Can one really stand in a relationship of mutual respect, regard, and trust with BP,
Penn State University, or the Bureau of Indian Affairs? Groups do pose genuinely
difficult puzzles to philosophical discussions of agency and responsibility. Our account
of corrective justice may need to be adjusted in some ways to deal with such cases.
However, for the moment, I would just like to point out that individuals do in fact hold
many of the same morally-laden attitudes towards collectives as they do towards
individuals. They expect a certain level of treatment from those institutions. They
feel angry, resentful, betrayed, distrustful, insulted or threatened when those expect-
ations are violated. (If you doubt this, simply search the internet for discussions of

14
I say “probably” because there are cases in which a victim’s failure to feel resentment is rooted in her
inability to see herself as deserving decent treatment. These cases are especially worrying if the victim’s lack of
self-respect is traceable to the wrongdoer’s own misdeeds. I address these sorts of complications in Radzik,
Making Amends (note 10), Chapter 5.
15
Thanks to Scott Hershovitz for this objection.
 TORT PROCESSES AND RELATIONAL REPAIR 239

recent scandals involving BP, Penn State, or the Bureau of Indian Affairs.) Similarly,
when the institutions with which we are affiliated violate other people’s legitimate
expectations, we sometimes do feel shame, regret, and a sense of the obligation to
make amends. There may well be something puzzling about such attitudes and
emotions, but I would argue that the contemporary world would be a far worse
place, morally speaking, if people were not susceptible to them. The motivations to
avoid other people’s indignation and our own feelings of guilt and shame are among
the chief incentives to moral behavior.
In this section, I have been concerned to describe what a reconciliation conception
of justice looks like as a response to moral wrongdoing, and my examples have focused
on everyday sorts of misdeeds. I would like to pause to point out how a conception of
corrective justice that focuses on repairing relationships might also be applied to other
types of wrongdoing as well. For example, philosopher Colleen Murphy applies such a
conception of corrective justice to transitional political contexts.16 She argues that, in
the aftermath of repressive regimes or civil conflict, justice requires the reconciliation
of distinctively political relationships, in which the parties respect one another as
equally valuable persons and reasonably trust one another with regard to the domain
of interaction that defines them as fellow citizens. This political relationship is thinner
in some ways than the moral relationship I have described (in that the domain of
interaction is more narrowly defined and the degree of goodwill required may be less),
and in other ways thicker (in that fellow citizens owe things to each other that they do
not owe to all moral persons, such as access to the vote). Different kinds of wrong-
doing (moral or political) damage different kinds of relationship (moral or political).
Yet one thing that becomes clear in reading Murphy’s discussion of democratic
transitions in places like South Africa or Argentina is just how intertwined these
relationships are. Being excluded from access to political influence, or living where the
rule of law has ceased to function, denigrates one as a morally valuable person.
Emotional reactions to political injustice, such as resentment, anger, fear of one’s
neighbor and mistrust of officials, must be assuaged before the political relationship
can function properly. Simply passing new laws and designing new institutions cannot
achieve just transitions. Subjective, interpersonal relationships among citizens must be
healed as well.

IV. Forms of Corrective Justice

So far, I have argued that a reconciliation theory of corrective justice provides a better,
more comprehensive understanding of what is wrong with both wrongdoing and harm-
causing than does the Aristotelian conception. The next advantage that I would like
to claim for a reconciliation theory is that it can recognize and explain why a wider

16
Murphy, A Moral Theory of Political Reconciliation (note 3).
240 LINDA RADZIK

variety of responses to wrongdoing are related to corrective justice. Compensation for

wrongfully taken or damaged property is both intrinsically appropriate and a step
toward repairing relationships. But other measures may be necessary too, such as
apologies, explanations of past wrongdoing, and assurances of better behavior, as well
as non-tangible, interpersonal forms of reparation, such as efforts to repair a victim’s
reputation in a community through subtle gestures of respect. These sorts of actions
are all matters of corrective justice in the sense that they are things that the wrongdoer
owes to the victim—and things that the victim can demand from the wrongdoer—in
virtue of the wrongdoer’s responsibility for the misdeed. In addition to compensation
for more material harms, the victim can, as a matter of justice, demand the correction
of the relational harms he suffered because of the wrongdoer’s action.
Consider apology more closely. In a paradigmatic case of an interpersonal apology,
the wrongdoer must correctly identify the wrongful action, honestly explain how she
came to perform it, acknowledge her culpability, condemn the action, communicate
remorse, and express her commitment not to repeat the offensive behavior in the
future.17 Reflecting on the character and consequences of wrongdoing helps clarify
how apologies are a form of corrective justice. To wrong someone is to treat him in a
way that is inconsistent with his value as a human being. The wrongful action
(whether one intends it to or not) sends the message that he is of low value.18 For
this reason, wrongs typically leave the victim feeling insulted and, often, threatened.
To apologize is to acknowledge responsibility for the action while at the same time
disavowing any continuing insult or threat. When the offender properly apologizes for
wrongdoing, she acknowledges that the victim deserves to have been treated better. In
showing her remorse, she lets the victim know that she does not take her mistake
lightly but is instead emotionally moved by it. Receiving an explanation of the wrong
allows the victim to better gauge whether he has reason to fear future mistreatment,
from either this wrongdoer or other members of the community.
Since apology helps to redress some of the emotional and relational harms that
wrongdoing may cause, we might be tempted to classify it as a kind of compensation.
But compensation is, at best, a metaphor here. It is not as if something of value that
was lost (e.g., the victim’s peace of mind) is replaced with something of equal value
(e.g., the wrongdoer’s remorse). Apologies repair harms not by repaying debts but by
disarming threats to relationships.
While apologies from wrongdoers to victims work to repair their relationship,
apologies performed in front of an audience can contribute to the broader project of
reconciliation. By publicly declaring who was in the wrong, and expressing regret
and respect, the wrongdoer can repair the victim’s image in the eyes of the community.
The public apology also provides the community with evidence of the wrongdoer’s

17
For more on non-paradigmatic apologies, such as apologies by political leaders for historic or collective
injustice, see Harvey, “The Emerging Practice of Institutional Apologies” (note 4); Radzik, Making Amends
(note 10), Chapter 7.
18
Cf. Murphy and Hampton, Forgiveness and Mercy (note 5), Chapters 1–2.
 TORT PROCESSES AND RELATIONAL REPAIR 241

moral improvement, and so can address the damage to the relationship between the
wrongdoer and the community. Public apologies are particularly important in cases
where persons other than the direct victim have reason to feel insulted or threatened
by the misdeed. For example, an employer who publically apologizes to the woman he
has sexually harassed can, thereby, make amends with his other female employees.
This analysis helps explain why apologies are so highly valued in our culture. It
helps explain, for example, why victims who receive heartfelt apologies sometimes
forswear any interest in material compensation or retribution, while compensation
and retribution rarely erase a victim’s desire for an apology.19 A well-formed apology
directly addresses the relational damage—the sense of having been insulted and
demeaned, the resentment, fear, and distrust—that wrongdoing leaves in its wake.
Still, even if one agrees that apologies and similar measures may be morally required
in the aftermath of wrongdoing, this does not establish that they are owed as matters of
justice. Not everything that we ought to do morally is something that can be demanded
from us. Apology might be thought to fall into this category. We often say, “She owes
him an apology” or “I demand . . . ” or “I deserve an apology.” That sort of language is
usually indicative of a claim of justice. But there are certain aspects of apology that
might seem to place it outside of the scope of justice. For example, apologies ought to
be sincere. They ought to be accompanied by emotions such as regret and remorse. But
how can sincere, emotional reactions be demands of justice?20
As Bonnie Raitt sings, “I can’t make you love me if you don’t/You can’t make your
heart feel something it won’t.”21 In these lyrics, we find an argument for placing
emotions and, thereby, apologies outside the scope of justice. Demanding an emo-
tional response from another person is pointless (“I can’t make you love me if you
don’t”), because people do not have sufficient control over their emotions (“You can’t
make your heart feel something it won’t”). But while this view makes for good country
music, and while it may be true of emotions like romantic love, it does not provide a
convincing account of remorse. At least when we are rational, our emotions about the
past are closely linked to our beliefs about the facts (who did what to whom) and our
moral commitments (what is right, what is wrong, who is worthy of decent treatment,
and whether moral norms are authoritative for us). If those moral commitments are
not sufficiently in our control, then morality as a whole is a fiction. If those moral
commitments are in our control, then our emotions of remorse are at least indirectly in
our control as well. If remorse is sufficiently in our control, then other people can
sensibly demand it from us.
A second argument for placing apologies outside of the scope of justice might
concede that sincerity and remorse are in the control of the wrongdoer, but point
out that these reactions are outside of the control of other people. The victim might

19
Several such examples are recorded in Lazare, On Apology (note 4).
20
On this question, see also Lazar, “Corrective Justice and the Possibility of Rectification” (note 7), 366.
21
“I Can’t Make You Love Me,” lyrics by Mike Reid and Mark Shamblin (Capitol Records, 1991).
242 LINDA RADZIK

say, “I demand an apology,” but neither she nor anyone else can back up that demand
with compulsive force. No one can make the wrongdoer reform her attitudes.
I have argued that apologies are matters of justice because they are both owed to
victims and because victims and other people can legitimately demand that they are
made. Victims can claim apologies as their due. The counter-argument suggests that a
response to wrongdoing cannot be owed as a matter of justice unless it can be, not just
demanded or claimed as one’s due, but compelled or guaranteed by force. But I see no
reason to add this provision. While it might well be impossible (psychologically) to
force another person to sincerely regret her past action, it is both possible and common
for people to pressure one another to do so. We rebuke those who fail to apologize to us
or who seem insincere. We give them dirty looks and cold shoulders. We berate them
to other people, and encourage those people to turn a cold shoulder as well. These are
awfully aggressive ways to treat wrongdoers. We should only be permitted to treat
people in these ways if justice is on our side. Rejecting the idea that apologies can be
owed as a matter of justice would commit one to rejecting these common moral
practices.
Apologies, explanations, promises of better behavior, and gestures of respect can all
be properly demanded of wrongdoers. Without these sorts of interactions, victims may
find it very difficult to put the past behind them and re-establish morally appropriate
relationships with their former abusers, their communities and themselves. Insofar as
a theory of corrective justice cannot explain the value and (at least in certain cases) the
necessity of such measures, it is incomplete and unsatisfactory.

V. Contributors to Corrective Justice

The next advantage I see in a reconciliation theory of corrective justice is that it
supports a better answer to the question of who owes duties of corrective justice. The
Aristotelian account of corrective justice is often criticized on these grounds.22 Cor-
rective justice cannot merely be a matter of compensation because compensation can
be made just as well by third parties as by wrongdoers. My friend Lois carelessly
damages my computer. A mutual friend of ours, Rich, kindly pays for the repairs. Rich
addresses the material aspect of the harm, but the wrong lingers. A reconciliation
theory of corrective justice has no trouble explaining why it seems so important that
Lois herself apologize and explain her behavior. Her negligence suggested a lack of
respect and concern for me. It made me doubt whether I can trust her in the future.
Until these relational harms are addressed, the wrong is not fully corrected.
But, another advantage to the reconciliation theory of corrective justice is that it
recognizes that third parties often can contribute to corrective justice. Rich’s money
cannot fully right Lois’s wrong, but it does help repair some of the harm. Another way

22
Cf. Jules Coleman, Risks and Wrongs (New York: Cambridge University Press, 1992), 309.
 TORT PROCESSES AND RELATIONAL REPAIR 243

in which third parties can contribute to corrective justice is through the social pressure
that they place on wrongdoers to make amends. Were Rich to make it clear to Lois that
he expects her to apologize to me, and that their own good relations are dependent
on it, Rich further helps to repair the wrong.
Third party contributions to corrective justice are possible; in some cases, they are
also necessary. Consider again the case of schoolyard bullying. Studies indicate that
around eighty-five per cent of school bullying is performed in front of an audience.23
A large part of the harm caused by bullying is the harm done to the relationships
between the victim and witnesses to his humiliation. Bystanding children often view
bullying victims with contempt, or fear contact with the victims lest they should attract
the bully’s aggressive attention to themselves. The wrong cannot be fully corrected
unless the bully herself does something to make amends. But this sort of wrong is also
unlikely to be fully corrected unless the bystanders make some effort as well. They
must not let themselves be contaminated by the bully’s contempt.
Given the sort of wrong that bullying is, we might ascribe even stronger duties of
corrective justice to the bystanders. I am inclined to argue that classmates have positive
obligations to come to one another’s defense. Where directly standing up to a bully
might be unreasonably dangerous for a bystander, the bystander at least has an
obligation to try to ameliorate the situation by communicating concern and caring
to the victim. Studies suggest that such simple responses can significantly reduce the
psychological damage to victims.24 Whether we classify such support by the classmates
as supererogatory acts of kindness or duties of corrective justice will depend on how
we understand the classmate relation. It looks more like justice than charity if you
think classmates have special moral obligations to protect and care for one another.
Teachers and school administrators clearly have such special obligations, and so they
clearly have duties of corrective justice in these cases. Since the harms done to the
relationship between the victim and the social group are so central to the phenomenon
of bullying, we might say that bystanders who remain passive become complicit in the
wrong itself. Unless they do something to help the victim, they allow themselves to
become tools through which the bully torments her victim. By failing to do something
to correct the wrong, these children lose their status as third parties and become
wrongdoers. The general points that I want to make are, first, that third parties can
make a contribution to corrective justice, and second, that sometimes they are obliged
to do so.
On my view, even victims have certain duties of corrective justice. They are certainly
obliged not to socially punish wrongdoers in ways that are disproportional to the
wrongs and harms. Though the point is controversial, I would also argue that, other

23
N. Pearce, D. Cross, H. Monks, S. Waters, and S. Falconer, “Current Evidence of Best Practice in Whole-
School Bullying Intervention and its Potential to Inform Cyberbullying Interventions,” 21 Australian J. of
Guidance & Counselling 1 (2011), 9.
24
P. Flaspohler, J. Elfstrom, K. Vanderzee, H. Sink, and Z. Birchmeier, “Stand by Me: The Effects of Peer
and Teacher Support in Mitigating the Impact of Bullying on Quality of Life,” 46 Psychology in the Schools 7
(2009).
244 LINDA RADZIK

things being equal, victims have a duty not to prevent the wrongdoer from making
amends. Victims do not have to accept wrongdoers’ apologies or reparations, but
blocking them from making the offers would be wrong. It would unfairly impair the
wrongdoer’s ability to improve her relationships, not only with the victim, but also
with herself and her community. The victim’s duty not to prevent wrongdoers from
attempting amends would be trumped, of course, if further contact with the wrong-
doer posed an unreasonable risk of harm to the victim. Another duty one might
ascribe to the victim is a duty to restore a moral relationship with a wrongdoer who has
offered sincere and proportional amends. The victim must at least recognize the other
as once again a member in good standing in the moral community, even if he chooses
to end the friendship or other special relationship that existed before the wrong.
In considering these ways in which victims might have duties of corrective justice in
the aftermath of wrongdoing, we find one more role for third parties to wrongdoing.
Third parties often play a role in moderating victims’ responses to wrongdoing. They
do so by expressing disapproval of disproportional demands for amends or unreason-
able resistance to reconciliation.
According to reconciliation theory, the achievement of corrective justice frequently
requires the efforts of the wrongdoer, members of the affected community and the
victim. What stands in need of correction in the aftermath of wrongdoing is primarily
the relational damage—the damage to relationships of mutual respect, goodwill, and
reasonable trust—caused among these three parties. It makes sense that the repair of
those relationships will often require the cooperation of all those parties. Furthermore,
what is required of any particular party on any particular occasion will be tied up in
complex ways with the responses of the other parties. For example, in a case where the
wrongdoer is quick to recognize her obligation to make amends and willing to satisfy
it, the victim is more likely to find himself obliged to reconcile, and there may be no
need for third parties to get involved at all. But in a case where the wrongdoer is
recalcitrant, third parties are more likely to be obliged to come to the victim’s aid.
This brings me the final advantage I see for a reconciliation theory of corrective
justice. It can distinguish between more and less ideal resolutions of wrongdoing. For
the Aristotelian model, a case in which compensation is voluntarily offered by a
wrongdoer and a case in which compensation is wrested from her by an authority
figure are equivalent in justice. The debt is paid in both cases. But a reconciliation
theory can account for why the former resolution is preferable, from the point of view
of justice, to the latter. More of what the wrongdoer owes to the victim has been made
good. When the wrongdoer makes amends voluntarily, both the victim and the
community have better reason to renew trust and normalize relations with her.
According to the reconciliation model, the correction of wrongdoing is frequently a
matter of degree. Fully voluntary amends are better than amends that result from
social pressure are better than compensation that is compelled by force. There are
cases where harms can be fully repaired and cases where lingering damage to bodies or
psyches or relationships is irreparable. This conclusion seems only fitting once we have
a realistic, multifaceted understanding of what is wrong with wrongdoing.
 TORT PROCESSES AND RELATIONAL REPAIR 245

VI. From Morality to Law

What I have presented here is a reconciliation theory of corrective justice that has
focused on moral wrongdoing. I would now like to move on to the question of what—
if anything—this conception might have to offer the study of tort law. There are two
possible ways of approaching the question, only one of which I pursue in any depth.
First, one might develop a reconciliation view of tort law that follows the pattern of
Murphy’s reconciliation theory of corrective justice in political contexts, which
I mentioned at the end of Section III. Murphy argues that corrective justice in the
aftermath of repressive regimes and civil conflict requires the reconciliation of dis-
tinctively political relationships, that is, relationships that people stand in to one
another as fellow citizens. These relationships are damaged by injustices such as
politically motivated human rights abuses and the systematic violation of the rule of
law. In order to correct such injustices, these political relationships must be repaired.
Similarly, a tort theorist who is attracted to the idea of a reconciliation theory might
argue that there is a distinctive sort of legal relationship that is violated by tortious
conduct and repaired through the civil legal process.25 Precisely what the defining
norms of that legal relationship are and how the tort process repairs that relationship
would need to be articulated.
Alternatively, one might keep the moral relationship that I described above at the
center of the account of corrective justice, rather than replacing it with some other
relationship. This is the approach I explore in the remainder of the chapter. On this
version, any wrongdoing or harm-causing potentially threatens moral relationships
and so raises issues of corrective justice. The wrongdoer or harm-causer is required to
repair or shore up any damaged or threatened relationships. Importantly, society has
different methods of enforcing such duties of corrective justice. Much of the time we
enforce these duties merely with social pressure. A friend who betrays his friend’s
secret is subjected to moral criticism or social withdrawal until he makes amends. In
more serious cases, duties of corrective justice may be enforced by the state through
criminal or tort processes. We do not want the state to become involved in the
correction of all cases of wrongdoing or harm-causing because such intervention
brings along serious costs, including intrusions on privacy, which are not justifiable
in every case. For those cases of harm-causing or wrongdoing that have also been
designated as torts, however, the civil legal process can play a role in achieving
corrective justice.
But how exactly do tort processes contribute to corrective justice, in the morally rich
sense of corrective justice I have been defending? To pursue this question further, let
us consider an example of a tort. Suppose Andy invites all of his colleagues, their
spouses and children to a party at his home. Despite knowing that his dog has reacted

25
I take it this would be distinct from Murphy’s political relationship since the question of whether the
other person is a fellow citizen is not relevant to torts.
246 LINDA RADZIK

aggressively to children in the past, Andy lets the dog roam among the guests. Little
Janie tries to pet the dog and is bitten in the face. Through his negligence, Andy has
both wronged and harmed Janie. He has failed to take the reasonable care that was due
to her as a fellow person equal in value to himself. Janie, her parents, and the other
guests have reason to regard him with resentment, indignation, and distrust. Janie has
suffered pain and fear. Her facial scars may negatively affect her future. She may
develop a phobia of dogs. Janie and her parents have the right to demand an apology,
compensation for her medical costs, and other, less tangible forms of reparation. The
other party guests too have a right to be indignant, to demand expressions of remorse,
and to socially pressure Andy to make amends. Additionally, Janie’s family can sue.
This case is based on one that happened among my circle of friends. The Andy
in that case immediately regretted his negligence, apologized sincerely, and offered
many different forms of amends, including euthanizing the dog, which was a beloved
family pet. Andy also asked the child’s family to sue him so that he could draw upon
his insurance in order to be able to offer them a larger sum of money, which could
better symbolize his remorse. However, Janie’s family declined to sue Andy because
they accepted his apology and other gestures. They accepted money from him for the
modest medical costs that were not covered by their own insurance and were satisfied.
In my opinion, this example provides a highly satisfactory form of corrective justice.
The wrongdoer made amends voluntarily and did his best to provide forms of amends
that were responsive and proportional to the reasonable claims of his victim and his
community (here, his other guests). Importantly, this resolution is also one in which
tort law plays no significant role. Had Janie’s medical bills required a larger sum of
money, and were it the case that Andy could only secure such a sum if the family
initiated a lawsuit, the satisfactoriness of the resolution may not have been diminished.
On the other hand, the trauma of being repeatedly forced to account for their behavior
within an adversarial legal system could have had its own damaging effects on Andy,
Janie, Janie’s parents, and their relations with one another. Indeed, the family’s
decision not to sue considered those possible costs. The degree of relational repair
achieved would likely have been quite significantly diminished had Andy been forced
into a tort process against his will, rather than freely recommending such a step to
Janie’s family. People generally react badly to having the coercive power of the state
called against them. Indeed, it might make them even less likely to acknowledge their
responsibilities to their victims and communities.
Am I saying, then, that tort processes undermine corrective justice? No. I do believe
that tort processes pose obstacles to some aspects of an ideal resolution of wrongdoing
or harm-causing. But there are many cases where the ideal is simply not available, and
not likely to become available. In these cases, tort processes can help achieve what we
might call a tragic form of corrective justice—they help secure some elements of justice
when full or ideal justice is unlikely.
Suppose Andy were recalcitrant. Suppose he refused to acknowledge his responsi-
bility or make amends to Janie and her family. At least through a tort process, they
could receive money to cover her medical bills. Scott Hershovitz has argued that if we
 TORT PROCESSES AND RELATIONAL REPAIR 247

look at the tort process as a whole and not just tort judgments, we see that they provide
victims with less tangible forms of corrective remedies as well.26 Through the pro-
ceedings, Janie and her family may finally receive an explanation of how this wrong
came about—what the dog’s history was, why it was aggressive to children, and what
Andy was thinking in letting the dog roam among children. This sort of information
may help Janie work through her fear of dogs. It might help her parents manage their
feelings of anger so that, even if they do not forgive Andy, they can move on.
If the court awards Janie compensation for pain and suffering, or if they award
punitive damages, they send her powerful messages that counteract (though they do
not erase) the insults and threats that may be suggested both by Andy’s negligence
and his subsequent refusal to make satisfactory amends to her.27 The court tells
Janie that it recognizes her status as a valuable person who deserves to be treated
with equal respect and due care. It tells her that the court takes her side in her conflict
with Andy.
In the case of Andy and Janie, Andy’s fault is clear and the family’s demands are
reasonable. Of course, many tort cases are not so straightforward. A large part of a tort
process’s contribution to corrective justice is in determining fault. Furthermore, the
state plays a role in tort cases similar to one I ascribed to third parties in moral cases—
that of moderating victims’ demands for reparation.
We should also note, the state itself has a relationship with Janie that would be
damaged by its failure to come to her aid in her conflict with a recalcitrant Andy.
Different theorists give different accounts of this relationship, but one that seems
compelling here is Locke’s.28 Insofar as the state limits the individual’s right to secure
corrective justice by aggressive means (i.e., by using force to punish wrongdoers or
seize compensation), the state has a duty to perform these tasks for her where they are
justified. I imagine there are many tort cases where victims are much more concerned
with what I describe as their relationship with the community than their relationship
with the wrongdoer or harm-causer. That is, they are more concerned that everyone be
held to the same set of rules and that everyone receive the equal protection of the law
than they are that the particular transgressor change his attitude.
So, while entering into a lawsuit may make it less likely that Janie’s moral relation-
ship with Andy will be repaired, the tort process may help maintain her relationship
with the state and strengthen her sense of self-respect. This interpretation of tort law is
now approaching the civil recourse view. What tort law provides for victims is not fully
fledged corrective justice, but rather recourse to civil support in securing some forms
of reparation.29 As other theorists have already noted, civil recourse is not so much a

26
Scott Hershovitz, “Harry Potter and the Trouble with Tort Theory,” 63 Stan. L. Rev. 1 (2010).
27
Civil recourse theorists have pointed out that corrective justice theories of tort law generally have trouble
accounting for civil courts’ use of punitive damages and compensation for pain and suffering. Cf. John
C. P. Goldberg, “Twentieth-Century Tort Theory,” 91 Geo. L.J. 3 (2003), 576. My version of corrective justice
does not encounter this objection. Thanks to John Oberdiek for pointing this out.
28
John Locke, Second Treatise of Civil Government (Buffalo, NY: Prometheus Books, 1986).
29
Benjamin C. Zipursky, “Civil Recourse, Not Corrective Justice,” 91 Geo L.J. 2 (2003).
248 LINDA RADZIK

rejection of the corrective justice view but a modification of it.30 What a reconciliation
account provides is a fuller explanation of why tort processes cannot (usually) fully
satisfy the claims of corrective justice. It is because what stands in need of repair
includes many things that cannot be achieved through the application of a legal
system.
Tort processes cannot address all forms of wronging or harming, or satisfy all rea-
sonable demands for amends, nor would we want them to attempt to do so. The state
can get involved when Andy fails to properly control his dog at the party, but not when
he negligently leaves his new colleague off the guest list. We do not want the state
mixed up in all aspects of social life. Nor do we want the state compelling apologies.
For one thing, a compelled apology would lose much of its value. But more import-
antly, the state must allow Andy the liberty to feel and express morally problematic
attitudes. The state can address his behavior, by forcing him to pay material repar-
ations to Janie. It can also send its own morally charged messages about Janie’s value as
a person and about Andy’s negligence. But it may not attempt to use its coercive power
to morally reform Andy’s character.
The fact that the reach of tort courts is limited in ways such as these might lead us to
reject the idea that tort processes properly aim at the morally rich form of corrective
justice that I have described in this chapter. But I think that would be too quick. To say
that tort law has a limited role to play in the pursuit of a robust corrective justice is not
to deny that this form of justice helps explain and justify tort law’s role in society. We
can instead see tort law, criminal law, and practices of moral criticism and informal
social pressure as a set of tools that society uses in its pursuit of corrective justice.
Sometimes one tool is more appropriate, sometimes another is.

VII. Conclusion
In this chapter, I have presented and defended a conception of corrective justice that
suggests that wrongs have been righted when moral relationships among affected
parties have been reconciled. I have argued that this conception is to be preferred to
compensation and to retribution because it is based on a more complete understand-
ing of what stands in need of correction in the aftermath of wrongdoing and harm-
causing. I have not denied that a principle of compensation is a part of justice. Instead,
I have supplemented that principle by pointing to the consequences of wrongdoing
and harm-causing that cannot be resolved merely through the transfer of goods to the
victim. I have argued that apologies, explanations, and gestures of respect can be
required by justice in many cases. Which measures are required in a particular case
will depend on which moral relationships were affected, how they were affected, and
what the parties take to be good reasons for normalizing those relationships.

30
Hershovitz, “Corrective Justice for Civil Recourse Theorists” (note 8).
 TORT PROCESSES AND RELATIONAL REPAIR 249

Another advantage of a reconciliation theory of corrective justice is that it discour-

ages an overly simplistic picture of the work that righting wrongs requires. The
problems that wrongdoing and harm-causing create may involve a number of different
parties (not just victims and wrongdoers, but also third parties). So, controlling the
ongoing, harmful effects of past deeds may require the participation of many parties as
well. This framework allows us to acknowledge that an institution such as the system
of tort law can contribute to corrective justice without pretending that it is capable of
securing corrective justice.
Although a reconciliation theory presents an ideal of justice (some might say, an
unrealistically high ideal of justice), it is an ideal that is amenable to approximation. It
may well be that wrongdoing is hardly ever fully corrected, at least in cases of any
severity or complexity. But we can distinguish more and less just resolutions by
thinking about which relationships have been improved and to what degree. So,
even if we conclude that taking a case to civil court makes some aspects of reconcili-
ation less likely (such as the rebuilding of goodwill and trust between the plaintiff and
defendant), we need not conclude that tort law is at odds with corrective justice.
Instead, we may conclude that tort law helps correct other aspects of the wrong (e.g.,
through the delivery of material compensation, the repair or preservation of the
relationship between the victim and the community, and support for the victim’s
sense of self-respect). Civil courts may only ever offer us non-ideal resolutions and
approximations of corrective justice. But, in a non-ideal world, that can count as
success.
Should we interpret tort law, with its distinctive processes, institutions and history,
as aiming at, or as providing civil recourse to, the reconciliation of relationships? I do
not know. Answering that question would require much greater knowledge of torts
than I have. So let me conclude with a modest recommendation: as we continue to
explore how tort law is related to justice, let us not pretend that the anemic, Aristo-
telian conception of corrective justice is our only option.31

31
I would like to thank the other participants in the conference on the Philosophical Foundations of Tort
Law, who were uniformly friendly and helpful, in addition to being impressive scholars. Special thanks go to
the conference organizer and editor, John Oberdiek, as well as to Benjamin Zipursky, Scott Hershovitz, and
Colleen Murphy who provided comments on earlier versions of this chapter.
 12
Tort Liability and Taking
Responsibility
David Enoch*

I. Introduction
I have never been much of a fan of corrective justice as a normative vindication of tort
law, and here’s why: we can easily imagine a no-tort law system that actually works.
Perhaps—I honestly don’t know—New Zealand is what I am imagining here. Or
perhaps it’s the New Zealand of the tort law theorists’ legend. But such a system is
certainly possible, a system where there are no special connections between tortfeasors
and “their” injured parties, where some other principles entirely govern the way the
costs of negligence (and other torts) are shouldered, and where there are no devastat-
ing instrumental outcomes (in terms of deterrence, for instance, or administrative
costs). Tell the New Zealand legend, then, as you like it most. Now let us ask—is there
anything missing there? Is there any reason to be less than fully happy, morally
speaking, with such a system? It seems to me that unless the answer—on intuitive,
not too theory-laden grounds—is “yes,” corrective justice theories of tort law do not
get off the ground as an attempt at a normative vindication of tort law. They may still
have something going for them, perhaps as an explanation of the underlying rationale
of our institutions.1 But in more purely normative terms, abstracting now from general

* An earlier version of this chapter was presented at the Rutgers School of Law (Camden) Philosophical
Foundations of the Law of Torts conference, in the fall of 2012, and in the Law and Philosophy Workshop at
the Hebrew University in the spring of 2013. I thank the participants for useful discussion and feedback. In
particular, I am grateful to Eric Claeys, Antony Duff, Kim Ferzan, Alon Harel, John Oberdiek, Hanoch
Sheinman, Victor Tadros, and Ben Zipursky.
1
It is clear that the attempt to offer such an explanation is central for the work of leading corrective justice
theorists, such as Weinrib, Coleman, and Ripstein. But, I want more (and I know I’m not alone in this)—we
want also a vindication of our current practices, and with this ambition in mind, the challenge in the text
stands. Some theorists seem to think that the line between a descriptive explanation and a normative
vindication (of tort law, or perhaps more generally) can somehow be blurred, perhaps all the way to non-
existence (see, for instance, John Goldberg and Benjamin Zipursky’s contribution to this volume). For reasons
that I cannot go into here, I remain unconvinced.
 TORT LIABILITY AND TAKING RESPONSIBILITY 251

conservative considerations about maintaining legal and other traditions, etc., if

nothing is missing in the best version of the New Zealand legend, corrective justice
is entirely without force.2
Notice how powerful this What’s Missing in New Zealand challenge is.3 If the
corrective justice theorist cannot cope with this challenge, her theory is out before
we start, as it were, before we even have to consider in detail the advantages and
disadvantages of alternative theories here. If nothing is missing in New Zealand,
corrective justice theory just doesn’t get off the ground.
Corrective justice theorists have not succeeded, it seems to me, in stepping up to the
What’s Missing in New Zealand challenge. From time to time things are said, of
course,4 but nothing compelling and general, it seems to me, has been said. If, for
instance, you believe in moral and legal luck, that causation is itself of intrinsic moral
significance, that we are responsible (indeed, even praiseworthy or blameworthy) for
the consequences of our actions rather than for the risks we create (whether or not
they materialize), you may think that a lot is missing in New Zealand. I do not believe
any of this,5 and this will be my departing point in this chapter (so assume I am right
on this, for the sake of argument). But you can think of the What’s Missing in New
Zealand challenge not as a challenge that assumes the rejection of legal luck, but rather
as further evidence for such rejection: if intuitively nothing seems to be missing in the
idealized New Zealand, then causation does not seem to be intrinsically important
after all. Now, corrective justice theorists sometimes emphasize that negligence is not
after all the paradigmatic tort economic analysis of law people like to say it is. Rather,
we should think of something like the traditional torts of theft or trespass, and ask
whether there is any consideration in favor of requiring that the thief return the stolen
goods to the owner. And here the answer seems to be “yes.” Indeed, many will be
happy to answer that something is missing in a New Zealand that does not even
include this much by way of corrective justice.6 But this will not help the corrective

2
Of course, this doesn’t mean that we should immediately abolish tort law. Perhaps there are consider-
ations in favor of (broadly speaking) the status quo. Perhaps people’s expectations give reasons of this sort. Or
perhaps the best version of the New Zealand legend is not available to us—either because of some pragmatic
concerns in general, or because of more path-dependent ones, given that we are already in a tort-law, perhaps
corrective-justice, system.
3
I am not assuming that the only way of stepping up to this challenge would be by offering an instrumental
or “functionalist” story. Pretty much any story can apply. But of course, as an answer to the What’s Missing in
New Zealand challenge, merely saying “Why, tort law is missing there!” is not good enough.
4
John Gardner, “What Is Tort Law For? Part I: The Place of Corrective Justice,” 30 Law and Philosophy 1
(2011), for instance, puts forward a related challenge (in terms of what tort law is for), but his answer (in terms
of the continuity thesis, according to which often compensation is the best still-available way of complying
with the reasons that grounded the breached duty in the first place) cannot answer the question What’s missing
in New Zealand.
5
See David Enoch and Andrei Marmor, “The Case against Moral Luck,” 26 Law and Philosophy 405 (2007);
David Enoch, “Luck between Morality, Law, and Justice,” 9 Theoretical Inquiries in Law 23 (2008).
6
For a related point, see Jules Coleman and Gabriel Mendlow, “Theories of Tort Law,” in Edward N. Zalta
(ed.), The Stanford Encyclopedia of Philosophy (Fall 2010 edition) (<http://plato.stanford.edu/archives/
fall2010/entries/tort-theories/> accessed October 28, 2013), section 3.1.
252 DAVID ENOCH

justice theorist, it seems to me, at least if he is after an attempt to offer a vindicating

normative account of tort law. The reason is that putting history to one side, nowadays
tort law is primarily the law of negligence, and so this is where the vindicating battle is
to be fought. If nothing is missing in New Zealand when it comes to negligence law,
corrective justice theory cannot vindicate modern-day tort law. And anyway, I am
happy to restrict the discussion that follows to just tort law as applied to negligence.
In this chapter, I tentatively offer another possible way of coping with the What’s
Missing in New Zealand challenge. Intuitively, the thought is that tort liability can be a
way of taking responsibility, in one important sense of this loaded phrase. What is
missing in New Zealand, on this suggestion, is the tortfeasor’s taking responsibility for
her actions.7 Now, there is more than one thing that may be meant by such “taking
responsibility,” and so quite a bit of what is to come will involve clarifying the sense of
“taking responsibility” I am after here.
My response to the What’s Missing In New Zealand challenge is tentative in two
ways: first, I have to confess I’m not sure it works. It seems to me to capture some
important normative intuitions in the vicinity, but I am not sure that at the end of the
day it actually works. Second, while I think such considerations can justify something
like a tort law practice, I don’t think they must do so. This is because—as I explain in
Section VI—while tort law may be a way of facilitating taking responsibility (and this
may justify it, to an extent), there may be other ways of doing that. If the line of
thought developed here works, then it suffices to show just that corrective justice
theory is not out before we start. But it does not suffice to show that it wins the day.
So the progress such considerations can achieve is somewhat limited.
In Section II, I do a little more to flesh out the intuitive response I will pursue to the
What’s Missing In New Zealand challenge. Then, in Sections III and IV I leave tort law
behind, in order to fill in the details of the kind of taking responsibility I have in mind
here. The phenomenon, I think, is much wider than tort law or corrective justice in
general, and should be discussed in this wider context. In Section V I return to tort law,
explaining how tort liability may be related to taking responsibility. In the following
section, I note the limitations of the conclusion of Section V.

II. New Zealand and Apologies8

Recall how we are playing the game: we are to imagine the best (realistic) version of the
New Zealand challenge. Accidents happen, then, at roughly the rates they happen

7
Jules Coleman and Arthur Ripstein, “Mischief and Misfortune,” 41 McGill L. J. 91 (1995) also talk in terms
of responsibility as the idea that a New Zealand system does not seem to accommodate—but as details of what
I mean by taking responsibility emerge below, it will become clear that their point is importantly different from
mine.
8
As far as I know, the line of thought initially developed in this section is original. But a focus on apologies
in a related context can be found also in Antony Duff ’s and Linda Radzik’s contributions to this volume (see
Chapters 10 and 11).
 TORT LIABILITY AND TAKING RESPONSIBILITY 253

elsewhere. People are harmed and compensated. But they are not compensated by
“their” harmers. Rather, all risk-creators contribute to a general pool (perhaps roughly
according to the magnitude of the risk they create, regardless of whether or not it
materializes), and those harmed are then compensated from this general pool. More
radical versions of no-tort law systems can be thought of (for instance, where there is
no realm of compensation for harms that is even partly autonomous vis-à-vis the
general distributive justice mechanisms), but for now, this will do. About such a New
Zealand, then, we are asking whether something is missing there. And in order to show
that something is indeed missing, all that has to be done, I think, is to show that this
imagined New Zealand can be improved upon.
Imagine, then, an alternative system, where everything is pretty much as it is in New
Zealand, except that harmers are expected to apologize to those they have harmed;
nothing too grand or monumental, or even formal. It is just that if your negligence
causes someone harm, while you are not expected to pay them (they get their
compensation from the pool, after all), you are expected to give them a call, perhaps
drive to the hospital, and say something like, “I’m so sorry you have to go through this
agony, which I caused.” Or, perhaps you are expected to bear the costs of the flowers
sent to the hospital—the pool will not pay for that.
Now, there are many possible ways of filling in the details—in particular, what
exactly it means that you’re expected to apologize or to pay for the flowers. It could just
be a matter of a social expectation, one the violation of which carries no more by way
of sanction than the troubled opinion of others. Or it could be somehow a matter
of backed-by-sanction legal obligation—and still more details have to be filled in. But
for now I think we can do without such further details.
The question I want to ask is whether the alternative system—New Zealand plus
personal apology—has anything morally desirable that the New Zealand system lacks?
Assuming the two systems are alike in all other respect, is New Zealand plus apology
better than New Zealand? The answer that seems intuitively plausible to me is “yes.”
This, in any way, is the intuition I want to inquire into and develop here.
Assuming you share this intuition—that other things being equal, New Zealand plus
apology is better than New Zealand—why is this so? What is present in the former
that is absent in the latter? The answer, I think, is very naturally put in terms of
taking responsibility. As far as paying for the hospital bills is concerned—the pool
takes care of that. But if you face up to the person you injured by your negligence and
apologize, there is a sense in which you take responsibility for your actions, in a way
that seems morally important and that is absent in the New Zealand without apology
system. But this may sound suspicious to you. Isn’t a part of the point of the New
Zealand system to acknowledge that we are responsible for the risks we bring about,
rather than for their (somewhat random) realizations? If you and I both create a
similar risk (say, by driving), and if your risk results in harm and mine does not, it isn’t
clear that there is a morally coherent sense in which you’re more responsible9 than

9
For the relevant distinction between being more responsible and being responsible for more things, see
Michael Zimmerman, “Taking Luck Seriously,” 99 The Journal of Philosophy 553 (2002).
254 DAVID ENOCH

I am, or—therefore—that it would be appropriate to expect you rather than me to

apologize. Anyway, such suspicions—about moral and legal luck, perhaps—seem
to underlie the move to a New Zealand pool system in the first place.10 But if you’re
not more responsible than I am, how can you take responsibility, and even if you can,
why should you?

III. Taking Responsibility: The Intuitive

Idea, and Some Examples11
We need more examples, I think.
You hit a pedestrian, causing him serious harm. But the accident was not your
fault (perhaps it was the pedestrian’s, or someone else’s, or no one’s fault at all). And
you know as much—we can stipulate that you are as certain of this as of anything.
Now, I do not believe in moral luck, and in particular I do not believe that the mere
causal story here—the fact that your agency was a part of the causal chain of events
that led to the injury—matters when it comes to your moral record. You are not, it
seems to me, morally responsible for the injury. Suppose—for the sake of argument—
that my reasons for so believing12 are at least prima facie compelling. Still, Bernard
Williams was on to something.13 While we should all feel bad for the fate of the injured
pedestrian, you, it seems, should feel that extra bit of agent regret. If you do not, we will
tend to judge you unfavorably, and there will be a distinctly moral flavor to this
judging (“I understand it wasn’t her fault, but still, what kind of person just walks away
from such an event and proceeds with her life as if it was nothing to her?”). At the very
least, we do not think that this kind of behavior is the uniquely rational response to the
situation, we think that it is, at the very least, rationally and morally permissible to feel
agent regret.
Your teenage son commits a crime, causing harm to person and property. You
are not, let us suppose, directly responsible for the crime in any straightforward way—
it is not as if you put him up to it, or even drove him to this kind of thing by your
poor parenting. Parenting too, after all, is a percentage game, and this time, you lost.
Now, I am an individualist about moral responsibility—I believe that you cannot be
morally responsible for something simply in virtue of being in a certain relationship
with someone else who is morally responsible for that thing. In morality—perhaps
unlike the law—real vicarious liability is impossible. Suppose—for the sake of

10
Jeremy Waldron, “Moments of Carelessness and Massive Loss,” in David G. Owen (ed.), Philosophical
Foundations of Tort Law (Oxford: Clarendon Press, 1995).
11
This and the next two sections draw on David Enoch, “Being Responsible, Taking Responsibility, and
Penumbral Agency,” in Ulrike Heuer and Gerlad Lang (eds.), Luck, Value, and Commitment: Themes from the
Ethics of Bernard Williams (Oxford: Oxford University Press, 2012), 95–132.
12
Some of which are in Enoch and Marmor, “The Case against Moral Luck” (note 5).
13
See Bernard Williams, “Moral Luck,” in Moral Luck (Cambridge: Cambridge University Press, 1981),
20–39.
 TORT LIABILITY AND TAKING RESPONSIBILITY 255

argument—that my reasons for this individualism are at least prima facie compelling.
Still, we would judge unfavorably a parent who neglects to—in some sense—take
responsibility for her teenage son’s behavior, perhaps, for instance, by apologizing for
him, or some such. If facing hard questions, you settle for noting (correctly) that the
relevant action was not yours, there seems to be something amiss—I would say,
morally amiss—with your so doing. It is not that what you’re saying is wrong. The
action, after all, was not yours, and you were not responsible for it. It is that your saying
so is somehow problematic. At the very least, we do not think that this is the only
rational response here. There is something to be said for your taking responsibility in
some sense for your teenage son’s action. Or so, at least, it seems to me.
Your country has recently been engaging in some morally problematic projects and
actions, perhaps more so than the extent to which this is true of just about any country
at any given time. You are not—let us suppose—directly responsible for any of this.
You have not yourself taken active part in the problematic activities; you haven’t even
voted for the people perpetuating it. You may have even gone to a couple of demon-
strations protesting against it. Once again, I am an individualist when it comes
to moral responsibility: I certainly do not believe that merely in virtue of being a
member in a collective, or a citizen of a country, your moral record is automatic-
ally stained when the collective’s or the country’s is.14 Suppose—for the sake of
argument—that my reasons for this individualism are at least prima facie compelling.
Still, if the opportunity arises when moral charges are brought against your country,
there would be something wrong, it seems to me, if all you did was to (correctly) point
out that these actions are not yours. To repeat, it is not that this would be false—these
are indeed not your actions, and you are not responsible for them. Rather, it’s that your
response should, at the very least, be more complicated than this. At least if some other
background conditions are in place, you should, I think, take responsibility for your
country’s actions, and, say, attempt to justify them, or excuse them, or apologize
for them, or some such. If you refuse to do any of this, and instead just settle for noting
that the relevant action is not yours, there is, I think, something morally amiss with
your response. At the very least, this is not the only acceptable response. There is
something to be said for taking responsibility for one’s country’s actions.
Of course, nothing here is obvious or uncontroversial. You may not find my claims
about the examples plausible. Or you may agree with me on these points, but differ on
the suggestion that these three kinds of cases are instances of an interestingly uniform
normative phenomenon. But in what follows I want to offer an initial and general
analysis of the phenomenon present here, assuming—as seems plausible to me—that

14
Notice that the issue relevant here is not when (if ever) a collective is responsible for anything. The issue
is, rather: given that the collective is responsible, what does it take for a (personally uninvolved) individual to
be responsible partly in virtue of being a member of a collective that is responsible for the relevant thing. In the
terms sometimes used in this literature (see, for instance, Marion Smiley, “Collective Responsibility,” in
Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2010 edition) (<http://plato.stanford.
edu/entries/collective-responsibility/> accessed October 28, 2013), section 4) the issue is that of the distribution
of collective responsibility, not that of its existence or emergence.
256 DAVID ENOCH

there is a uniform phenomenon here. Intuitively, in all these cases the thing for which
responsibility is in question—the consequences of your actions, the actions of your
son, or of your country—is not within the core of your agency. But it is not too far
from it either. As I will put things, we are here in the area of your penumbral agency.
The intuitive thought that I will make much of—and which, to an extent, I borrow
from Susan Wolf—is the following.15 Some things are within the core scope of your
agency. Your actions, perhaps, or your intentions, are such that you are responsible for
them. Furthermore, although it is often up to you whether to perform an action, or
form an intention, and although their being up to you may be importantly related to
your being responsible for them, your being responsible for them is not up to you.
Once you perform the action, or form the intention, and assuming the other necessary
conditions are in place, you are already responsible. Your being responsible does not
depend on any act of will (except, that is, insofar as the action or intention or whatever
it is you are responsible for so depends).
Other things lie outside the scope of your agency. The motion of the planets, say, or
the (natural) length of your nose, or the actions of some far away people to whom
you stand in no special relation are all things for which you are not responsible. If
the topic comes up, if someone treats you as if you were in some way responsible
for these things, then saying that you are not, that these are not actions of yours is
precisely the right response on your behalf. In these cases, unlike the three cases above,
nothing else is called for—you are just not responsible, and that is the end of
the matter. Furthermore, there is nothing you can do in order to change this. Even
if you want to become responsible for them, you cannot—it is not within your power
to take responsibility for them.
But there is a third group of cases. These are cases of things that lie in the penumbra
of your agency. In such cases—arguably exemplified by the three examples above—
you are not responsible for the relevant thing (the uncontrolled consequences of your
actions, your child’s actions, your country’s actions). But in these cases, an act of will
can make all the difference. In these cases, you can change the situation with regard to
your being responsible. You can, that is, take responsibility, and thereby become
responsible.
Furthermore, taking responsibility is a kind of a (possibly mental) action,16 and so it
may have one of the moral statuses that actions often have. In particular, it can be the

15
This is the guiding thought in Susan Wolf, “The Moral of Moral Luck,” 31 Philosophical Exchange 4
(2000), 121; reprinted in Cheshire Calhoun (ed.), Setting the Moral Compass: Essays by Women Philosophers
(Oxford: Oxford University Press, 2003). For some important differences between Wolf ’s view and mine, see
Enoch, “Being Responsible, Taking Responsibility, and Penumbral Agency,” supra, note 11.
16
This is one major way in which my use of “taking responsibility” differs dramatically from Fischer and
Ravizza’s. See, for instance, John Martin Fischer and Mark Ravizza, “Précis of Responsibility and Control:
A Theory of Moral Responsibility,” 61 Philosophy and Phenomenological Research 441 (2000), 443, and the
references there. Note that despite explicitly saying that taking responsibility is, as they understand it, not an
action (but rather the having of certain dispositional beliefs and having appropriately acquired them), they still
think about it as a way of making an action one’s own. But the metaphors here are somewhat misleading,
because “making one’s own” very much suggests an action.
 TORT LIABILITY AND TAKING RESPONSIBILITY 257

case that you ought to take responsibility. And sometimes, this is precisely the case.
Indeed, I think it is the case in the examples I started with. You are not, I think,
responsible for the injury to the pedestrian. But you ought to take responsibility. And
once you do, you will be responsible, partly in virtue of your act of will, your having
taken responsibility. Only partly, because the act of will, though necessary, is not
sufficient—it is also necessary for the case to antecedently lie in the penumbra of your
agency. If it does not, you cannot bring about a change in your responsibility simply by
attempting to take responsibility. But noticing the possibility of a duty to take
responsibility, I want to argue, opens up some room in logical space for another way
of accommodating our negative judgment of the driver who just proceeds as if not
much has happened. Her moral flaw is not constituted by her being antecedently
responsible for the pedestrian’s injury and failing to respond appropriately (she is, ex
hypothesis, not responsible). Rather, it lies in her violation of her moral duty to take
responsibility.17
Similarly, I would say, for the other two examples. You are not, I think, responsible
for your teenage son’s crime. But you are—in certain circumstances, at least—under a
moral duty to take responsibility for it. If you do, you are then responsible for it, partly
in virtue of having taken responsibility for it.18 If you do not, you are not responsible
(you did not, after all, take responsibility for it, and his action is not within your core
agency). So you are off the hook—off this hook, that is. But you are still acting in a
morally problematic way, because you are violating your moral duty to take respon-
sibility for your son’s actions. And this is what explains our negative evaluation of the
parent who in the face of such circumstances proceeds to just (rightly) note that the
teenager’s actions are not his own.
And you see how it goes. You are not responsible for your country’s actions. But you
are—in certain circumstances—under a moral duty to take responsibility for them.
If you do, you are then responsible for them, partly in virtue of having taken respon-
sibility for them. If you do not, you are not responsible for your country’s actions, but
you are in violation of your moral duty to take responsibility for them. And you may
very well—presumably depending on some other considerations—be responsible for
that. And this is what explains our negative evaluation of the citizen who proceeds in
such a case simply to (rightly) note that her country’s actions are not her own.
This kind of thing sounds to me like the right thing to say—I think that the last three
paragraphs capture rather nicely the phenomenology of these cases. In all of them, it
seems to me that being responsible is not something we find ourselves with, but rather
something that we do, there is something distinctively active about becoming respon-
sible in these cases. This, then, is what it seems to me we want to say about the three
cases, at least in outline. But of course, you can’t always get what you want. In order

17
There will be a qualification of this claim below.
18
It does not follow that your son is any less responsible for his action. Here as everywhere else we should
resist the temptation to think of responsibility as a given magnitude, to be distributed between different agents.
It is not as if assigning responsibility is a zero-sum game.
258 DAVID ENOCH

to see whether we can responsibly say such things, we need many more details. In
particular, we need an account of what taking responsibility consists in, and an account
of what the responsibility thus taken is.

IV. The Taking and the Responsibility

Sometimes, when we use the locution “taking responsibility” all we mean by it is
something epistemic in nature. If I say “It’s time that you start taking responsibility for
your actions!,” what I typically have in mind is not something about penumbral agency
at all. What I typically have in mind is that it is time that you acknowledge or recognize
the relevant responsibility facts, namely, that you are responsible for your actions
(in general, or perhaps for some particular actions). When “taking responsibility” is
thus used, the responsibility facts in no way depend on the taking. Rather, they are
there all along, and the taking is thought of as the appropriate response to the relevant
responsibility facts. In this sense, then, taking responsibility is analogous to acknow-
ledging a duty that is already there. And there is, of course, nothing wrong with using
“taking responsibility” in this way. But if you share my intuition that sometimes we
can take responsibility even in cases in which (as we know) we are not responsible
independently of and prior to the taking, then you already agree that this epistemic
sense of “taking responsibility” is not the only relevant one.
Another way of using this phrase is the way I want to focus on. This is the way
in which taking responsibility is analogous not to acknowledging a duty that is
already there independently of the acknowledging, but rather to undertaking a duty (or
obligation, or commitment). In this sense, when I take responsibility for something—
and when the other necessary conditions are in place—I thereby become responsible.
Having taken responsibility (in this sense) I am responsible, and this is so partly in
virtue of my having taken responsibility. Think of promises. When I promise to meet
you for lunch, typically the promise does not amount to acknowledging a duty (to
meet you for lunch) that was there anyway, independently of the promise. There was
no such duty independently of and prior to the making of the promise.19 Rather, the
making of the promise created the obligation,20 and now I am under a moral duty to
meet you for lunch partly in virtue of having made that promise.
Let me exploit the promise analogy a little further. When making a promise (say,
to meet you for lunch), typically—and as already noted—the normative situation will

19
Depending on your theory of promises, and also, I would say, on your theory of reason-giving in general,
there may have been a duty there independently of the making of the promise, namely, my duty to keep my
promises, and so perhaps also my duty to-meet-you-for-lunch-if-I-so-promise. See David Enoch “Giving
Practical Reasons,” 11(4) Philosophers’ Imprint (March 2011) (<http://quod.lib.umich.edu/cgi/t/text/text-idx?
c=phimp;view=toc;idno=3521354.0011.004> accessed October 28, 2013). But this doesn’t matter for our
purposes. All that matters here is that the duty to meet you for lunch did not exist independently of, and
prior to, the making of the promise.
20
Throughout this chapter I am using “duty” and “obligation” interchangeably.
 TORT LIABILITY AND TAKING RESPONSIBILITY 259

be as follows: independently of, and prior to the promise, I am not morally required to
meet you for lunch. The making of the promise, though, changes this, and now I am
morally required to meet you for lunch, partly in virtue of having promised to do so.
But let me note a further interesting feature of promises—one that is not, I think, often
noticed. The making of a promise is itself an action, and so we may wonder about
its moral status. In particular, I may be morally required to make a promise. And this
is consistent with my not being (prior to making the promise) morally required
to perform the action that I am morally required to promise to perform. Here is a
plausible, though somewhat melodramatic, example: a single-parent friend of mine is
diagnosed with a severe illness. He is extremely anxious about who will take care of his
daughter if he dies. Suppose that we are not close enough for it to be morally required
that I take care of her. Indeed, suppose that if my friend dies without anyone having
committed to taking care of his daughter, still one or two closer friends will step up to
the plate. But suppose further that none of them will promise my friend now that they
will. By promising, I can significantly alleviate my friend’s anxiety. It seems to me that
there are ways of filling in the details of the example such that I will be morally
required to promise my friend to take care of his daughter if he dies. This, even though
I am not morally required—independently of and prior to the promise—to take care of
her if he dies. Of course, if I make the promise and the friend later dies, I am then
morally required to take care of his daughter, partly in virtue of having made the
promise. If I neither make the promise nor take care of the child, I am not in violation
of my moral duty to take care of the child (I don’t have that duty). But I am in violation
of my moral duty to promise to take care of her—a duty I can be in violation of even if
I do proceed to take care of the child. More generally, it is one thing to have a duty to
ф, quite another to have a duty to promise to ф. The normative considerations that
apply to one of these need not apply to the other, certainly not as a matter of necessity.
And there is no general formal reason to think that you cannot have a duty to promise
to ф without also having a duty to ф (of course, if you act in accordance with your duty
to promise to ф, you then have a duty to ф; this, after all, is precisely my point).21
It is precisely the analogy with this normative structure that I want to exploit. In the
cases of taking responsibility of the kind I am interested in, I want to say, there is no
responsibility (independently of and prior to the taking); there is a duty to take
responsibility; if you take responsibility, say, for your child’s actions, you thereby
become responsible, partly in virtue of your having taken responsibility; and if you do
not take responsibility, you are not responsible for the child’s actions, but you are in
violation of your duty to take responsibility, and perhaps also responsible for that. And
now that we know that there is no formal problem with this normative structure when
it comes to promises, we should not be too worried about this structure in the case of
responsibility either.

21
I was surprised to find out that many seem to resist the point in the text here—which seems to me rather
obvious. I do more to defend it in Enoch, “Being Responsible, Taking Responsibility, and Penumbral Agency”
(note 11).
260 DAVID ENOCH

Before discussing the different senses of “responsibility” and characterizing the one
I am interested in, I want to spend some time on the technicalities here—how it is that
we take, or fail to take, responsibility.
The first thing to note here is that while we can certainly take responsibility
explicitly (“I hereby take responsibility for my son’s actions”), we much more often
do so implicitly. And the many ways in which we can do so implicitly differ widely
from one context to another. Here too the analogy with promising is telling. There is
hardly any way in which we cannot—in some context—make a promise. Sometime,
one’s intonation can make the difference—for instance, sometime the tone, and maybe
also the body language, with which a simple “don’t worry” is uttered suffices for it
to qualify as a promise, as genuinely duty-creating as the most solemn “I hereby
promise.” Indeed, sometimes there is no need for any text at all, and the promise is just
implied by a certain behavior. Indeed, sometimes a promise can be implied by the
initiation of a chain of actions that constitute keeping that promise.
Getting back to taking responsibility, then: it too, as I have already said, can be done
implicitly. And just as with promises, there is hardly any limit as to what can qualify
as an implicit way of taking responsibility—in the appropriate circumstances, a
meaningful look, a body gesture, a text seemingly doing some other thing can amount
to a taking of responsibility. And here too, sometimes one can implicitly take respon-
sibility by performing the first part of what it would take to live up to the responsibility
just taken. Below I argue that central ways of living up to the responsibility taken
include things like offering a justification for the relevant action, or an excuse for it, or
an apology. What I want to note here, though, is that doing such things can also
constitute implicit ways of taking responsibility, as well as the first step towards living
up to the responsibility thus taken. So, for instance, by offering an apology for your
child’s behavior you can at once both take responsibility for it, become responsible for
it, and then go a long way towards acting in the way one who is so responsible should.
This complicates things, of course, because it makes it harder to distinguish between
the hypothesis that one is—notwithstanding what I took as a premise earlier on—quite
simply responsible for one’s country’s actions (say), and the hypothesis that one is not,
but that one is under a moral duty to take responsibility. Once we see that on the latter
hypothesis the very same act can serve both as the taking of the responsibility and
as partly discharging it, it may be hard to see what the difference between the two is.
Does it matter, after all, if we go for a theory that judges an act of will (taking
responsibility) redundant, unnecessary for responsibility, or for a theory that considers
such an act of will necessary for responsibility in these cases, but is willing to read this
act of will into just about anything, including the very action that (both theories agree)
is a major part of what is required of someone who is already responsible? But
differences between the two types of theory remain. For the two hypotheses differ
regarding what it is that is wrong about the relevant agent’s morally faulty behavior.
On the first hypothesis, what is wrong in the behavior is that if fails to respond in the
appropriate way to the will-independent responsibility facts. On the hypothesis sug-
gested here, though, the flaw is the failure to live up to the duty to take responsibility.
 TORT LIABILITY AND TAKING RESPONSIBILITY 261

To see this, think of two citizens, Proud and Detached, who both refuse to take
responsibility for their respective countries’ actions—in either my sense, of refusing
to assume the responsibility, or in the competing sense, of just refusing to acknowledge
that they are already responsible and respond accordingly; and suppose that these
two citizens are alike in all relevant respects (their countries are equally democratic,
the wrongs committed by the country equally grievous, both citizens enjoy equally the
benefits that their countries offer them, and so on), except one: Proud takes pride in
the achievements (moral and otherwise) of his country. Detached, on the other hand,
feels detached—she can be happy, of course, for some of her compatriots when they do
well, even when they succeed in what may be considered national projects. But, the
national success itself is nothing to her, as it were. It seems to me clear that when we
judge Proud’s and Detached’s refusal to take or acknowledge responsibility, we want to
distinguish between them. There is something much more disturbing about Proud’s
attitude here than about Detached’s. Of course, on our assumptions—that this too
is a case where on the suggested theory Detached is under a moral duty to take
responsibility—Detached’s behavior is not morally ideal. She too violated a moral
requirement that applies to her. But something worse—or at the very least something
significantly different—seems to be going on in Proud’s case. Having taken pride in his
country’s achievements, shouldn’t he now acknowledge the responsibility for its
wrongdoings as well? If we understand what is going on here in terms of the duty
to take responsibility, and if we understand responsibility here as partly due to an act
of will (that is, the taking), we have a ready explanation of the phenomenon here.
By taking pride in his country’s achievements, Proud has already (implicitly) taken
responsibility for his country’s actions more generally, he has already incorporated this
activity from the penumbra of his agency into his agency. So, when he now says that he
is not responsible for his country’s wrongdoing, what he says is strictly speaking false:
having taken responsibility, he is now responsible. So by his denial, he in effect evades
the responsibility that is already there, he fails to respond to it correctly. Detached, on
the other hand, has not taken responsibility, and so she is not responsible. True,
she fails to take responsibility, as we are assuming that she should. But she does not
evade a responsibility that is already there. And there may be a difference in the degree
of severity (in general, or more plausibly, in the specific circumstances) between the
two wrongs. At least, on the suggested understanding of taking responsibility, Proud’s
and Detached’s behavior are different in a normatively significant way. And this is
precisely the result we were looking for. On the competing suggestion currently on
the table, however, it is hard to come up with a similarly clean explanation of the
phenomena. On that account, remember, Proud and Detached are equally responsible
for their country’s actions—their being so responsible is not up to them in any relevant
sense. Why is it, then, that we want to distinguish between the flaws in their behavior
when they claim not to be responsible for their country’s wrongdoing? How is it that
Proud’s having taken pride in his country’s achievements makes a difference here, if it
is not within Proud’s powers to change the relevant responsibility facts? I am not
saying that nothing can be said here. Perhaps there are other explanations of the
262 DAVID ENOCH

asymmetry between Proud and Detached—perhaps, for instance, on top of his other
moral problems (also present in Detached), Proud is also hypocritical. But the most
natural, unified explanation seems to be the former one, in terms of Proud’s having
exercised—by taking pride in his country’s achievement—his normative power to
make himself responsible for its wrongdoings.
We are now also in a position to solve a puzzle mentioned earlier. Prior to and
independently of taking responsibility for your teenage son’s action, we are here assuming,
you are not responsible for his action. And yet, earlier I suggested that in some circum-
stances, if all you do to respond to certain relevant reactions is to say something like, “But it
is not my action; I am not responsible for it and its consequences,” you behave in a morally
suspicious way. Why is this so, given that what you say is strictly speaking true? The
answer, it seems to me, is that sometimes by uttering this truth, you will be conveying your
unwillingness to take responsibility, as you should. Hence, the wrongness.
But what, it is at last time to say, exactly is this responsibility that you are taking?
What is the nature of this moral status, such that it can be changed by an act of will
in the described way? And what other implications follow from your being, or your
failing to be, responsible for something in the sense relevant here?
The term “responsibility” and related terms are multiply ambiguous, and so dis-
tinctions are in order. The sense of responsibility relevant here is of course moral, not
legal. Legal responsibility will become relevant again only in Section V. Likewise, the
sense of “responsible” in which it is used to attribute a property to an agent (distin-
guishing between those who are who are not responsible for their actions, for instance)
is not the one relevant here. Also irrelevant is the use of “responsibility” as standing for
a purely causal relation, as in “the weather was responsible for the quality of the
game”—a sense that can also be applied to agents, but that is still very different from
moral responsibility. There is also the forward-looking sense in which to say that you
are responsible for something is, roughly, to say that you are under a duty to bring
it about, or to look after it, etc.22 But we are here interested in backward-looking
responsibility, the kind of responsibility your son may bear to his (past) crime, and
that arguably you may bear to it as well, once you take responsibility for it.
A natural alternative to all of these notions of responsibility—and a very natural
attempt at capturing backward-looking responsibility—is an understanding of respon-
sibility that ties it very closely to blameworthiness and praiseworthiness. Thus, you
may think of responsibility in this sense as whatever it is that has to be added to
wrongness to get to blameworthiness, or you may think of responsibility in this sense
as that which is common to both praiseworthiness and blameworthiness. But this too
is not the sense I am after. I do not think that you are responsible—in a way closely
related to blameworthiness—for your son’s action, or (more controversially) for the
consequences of your actions, as in the example above. Nor do I think that by an act of
will you can make yourself blameworthy for something you are not.

22
See the closely related “task-responsibility” in Kurt Baier, “Types of Responsibility,” in Peter A. French
(ed.), The Spectrum of Responsibility (New York: St. Martin’s Press, 1972), 117–22.
 TORT LIABILITY AND TAKING RESPONSIBILITY 263

At this point you may be worried that no sense of “responsibility” remains for my
account of taking responsibility to apply to. Let me shift to a more positive mode, then,
and present the kind of thing I have in mind by “responsibility” here. I will use as my
starting point Marina Oshana’s understanding of responsibility as accountability.23
On this account:
. . . the judgment “X is responsible for Y” will express the beliefs “X did Y (or is Y),
X possesses the qualities of a responsible agent, and X is accountable or answerable for Y.

And:
“X is accountable for Y” can be unpacked as “It is appropriate that X explain her intentions
in doing (or being) Y.”24

The main thought here is rather simple: responsibility is sometimes not about what is
attributable to whom, exactly, but about what kinds of explanations are appropriate,
and in particular, whether it is appropriate to expect someone to explain an action, or
some such thing.
I think that we can improve somewhat on Oshana’s way of putting things.25 First,
Oshana speaks in terms of an explanation being appropriate. And at times26 it sounds
as if the question—whether an explanation is appropriately expected—is almost
epistemological, depending on whether someone is likely to have an explanation
available to her, or some such. But this is not, it seems to me, the right understanding
of the appropriateness of explanation here (as I think Oshana herself understands
elsewhere).27 Furthermore, Oshana’s account neglects the directionality of the phe-
nomenon here. Her account of accountability does not include—not explicitly,
anyway—anything about the explanation being not just appropriate but owed, and
furthermore owed to someone in particular. But it seems to me that the kind of
responsibility in the vicinity of what Oshana puts her finger on has precisely these
features.
Because of these facts—the relevant appropriateness is not epistemic; the thing
which is appropriate may be owed; and it may be owed to someone—it seems to me

23
Marina A.L. Oshana, “Ascriptions of Responsibility,” 34 American Philosophical Quarterly 71 (1997).
Oshana develops here themes she finds in Baier, and possibly elsewhere too. In more legal contexts, Antony
Duff ’s account of criminal responsibility is also very similar, at least in its underlying intuitions. I am not
relying more heavily here on Duff because it seems to me that when it comes to criminal responsibility, the ties
to blameworthiness are much closer than in the case of the responsibility-taking I am highlighting. And I do
not accept a theory of blameworthiness that ties it so closely to the appropriateness of explanations or
some such.
24
Oshana, “Ascriptions of Responsibility” at 77.
25
The quote above includes two parts that are irrelevant for my purposes: because I am assuming that in all
the relevant cases the relevant agent satisfies the conditions for general moral responsibility (as an ability), we
do not need to worry about that. Also, because I do not attempt to answer here the question of what categories
of things we can be responsible for, I do not need to worry about the “doing or being” in Oshana’s
characterization.
26
Oshana, “Ascriptions of Responsibility ” at 76.
27
Oshana, “Ascriptions of Responsibility ” at 77.
264 DAVID ENOCH

more fitting here to speak of an apology rather than an explanation. The sense of
responsibility I am interested in, and the sense in which you can sometimes make it the
case that you are responsible for something by taking responsibility for it, is the sense
that makes an apology appropriate. Thus, somewhat roughly: A is responsible (in this
sense of responsibility) for X to P if and only if, if X is wrong or otherwise morally
problematic, A owes an apology to P on account of X. This does not mean that
Oshana’s explanations are not relevant here. Apologies, I think, are the paradigmatic
example of a wider phenomenon, one that may very well include Oshana’s explan-
ations (again, understood non-epistemically). Other related responses may include
offering a justification, or an excuse. And thus we get, still somewhat roughly: A is
responsible (in this sense) for X to P if and only if, if X is wrong or otherwise morally
problematic, A owes it to P to justify X (or the relevant X-related thing), or to offer an
excuse for it, or to apologize for it, or to explain it, or something of this sort.28
Let me emphasize that it is in no way a part of what I am arguing for that the
understanding of responsibility I tried to develop above is the right understanding of
responsibility. This is also why I settle for an if and only if claim above: it is consistent
with my use of this understanding of responsibility that the bi-conditional is not
explanatorily basic, that there is a more basic truth here about the nature of respon-
sibility that explains the bi-conditional. Indeed, it is even possible (and consistent with
my use of the bi-conditional) that in these cases one owes an apology, say, because one
is responsible in some more basic way. All I will be relying on is that there is a sense of
responsibility for which this bi-conditional holds. And it is in this sense of responsi-
bility that you can—and sometimes should—take responsibility.
Let us return now to the three cases above. You do not, it seems to me, owe an
apology for your country’s actions (to the victims, say), simply in virtue of your
citizenship. Something else is needed. But you may be required to take responsibility.
And if you do take responsibility, it will then be, under certain conditions, appropriate
to expect you to offer a justification for your country’s actions, or an excuse, or in the
absence of either perhaps an apology. You owe such a response because you are now
responsible, in this sense, for your country’s relevant actions. And you are now
responsible for them because you have taken responsibility for them, as you may
have been morally required to do. Precisely what it is that living up to the responsi-
bility taken will amount to may depend on the details—for some, perhaps an apology
will be in order; for others, perhaps, just an explanation coupled with some form of
disassociation. But something by way of living up to the responsibility taken will be
relevant in each such case. Similarly, it seems to me, for the other two cases. Even if
your driving was impeccable, you should take responsibility for the consequences of
your action. This does not mean that you should think of yourself as blameworthy for

28
We need a somewhat amended version to deal with cases where there is no natural candidate to play P’s
role—if the accident victim is dead, and has no relatives, etc. But such amendments can be introduced in the
natural way. For more details (and for discussions of some more objections), again see Enoch, “Being
Responsible, Taking Responsibility, and Penumbral Agency” (note 11).
 TORT LIABILITY AND TAKING RESPONSIBILITY 265

them—you are not, and so you should not (you are not morally required to have this
false belief, after all). But if you take responsibility, you then become responsible, and
then you may owe “your” injured pedestrian some rudimentary form of explanation.
You may, for instance, owe it to him to present evidence for your good driving, and
also to offer your sympathy for his loss. And on my account, you do not owe him these
things independently of taking responsibility for his injury (or perhaps for the
consequences of your actions in general). If you do not take responsibility, you are
not responsible. But you are in violation of your duty to take responsibility. And
similarly, mutatis mutandis, for the parent case as well.
I need to discuss a complication here.29 I suggested that the right thing to say about
the country case, for instance, is that you are not responsible, in this sense, for its
actions, but that because of the nature of the relation between you and your country,
you’re morally required to take responsibility, and then to apologize, or some such.
And I have also noted in the previous section that apologizing may amount both to
an implicit way of taking responsibility, and to living up to the responsibility thus
assumed. But an alternative suggestion is that the nature of the relation between
you and your country directly makes an apology (or some such) appropriate. In this
accountability sense, the argument goes, you are responsible for your country’s
actions, regardless of whether or not you take responsibility. Now, I have been
assuming that people are not responsible for their country’s actions merely in virtue
of their citizenship. But it may now be argued that what made this claim plausible was
a different sense of “responsibility.” Perhaps mere citizenship is not enough for blame-
related responsibility. But why think that it is not sufficient for accountability or
apology-related responsibility either? And if so, the act of taking responsibility
becomes entirely redundant.
I want to acknowledge that this alternative account is not without plausibility. And
the comparative appeal of the two accounts may differ from one case to another. (For
instance, this alternative account seems more plausible when applied to the driver case
than when applied to the country case, I think; I return to this point below.) But two
kinds of consideration count against the alternative account as a global one. First, why
would anyone be expected to apologize, for instance, for actions that are not his or
hers, and that are not under his or her control? To me, this seems already problematic.
I acknowledge, of course, that this is hardly a conclusive or uncontroversial reason
to reject the alternative account. But the guiding intuition underlying my suggested
account still applies: according to my suggestion, you cannot be responsible, in
this sense, for anything outside the core scope of your agency unless by an act of
will you make yourself so responsible. This seems to me to be a feature of a theory
of responsibility worth maintaining. Second, as I suggested above, there seems to
be something active about the responsibility relevant in these cases, and this is a point
nicely captured by my suggested account, but not by the competing account. This was

29
I thank Peter Cane, Andrei Marmor, and Gary Watson for relevant objections.
266 DAVID ENOCH

a part of the point of introducing the comparison of Proud and Detached. If you
agree that there is a difference between them regarding the extent to which they
are responsible for their country’s wrongdoings, then you already agree that the
alternative account—one that considers the acts of will of the relevant agents redun-
dant here—is not the full story of responsibility in such cases. And this, I think, is all
that I need.

V. Tort Liability and Taking Responsibility

It is time, then, to get back to the law of torts, and indeed, to what’s missing in
New Zealand. Recall that I suggested (in Section II) that a New Zealand plus apology
system is morally better (other things being equal) than a mere New Zealand system.
The former answers, I suggested, and the latter does not, to some intuitive concerns
naturally put in terms of taking responsibility for the consequences of our actions. And
so we are now in a position to say more. The kind of responsibility involved, I suggest,
is the one described in the previous section. Furthermore, it is the kind of responsi-
bility that may be taken—in the sense above, that is, one where, by an act of will,
you can make yourself responsible. And offering an apology may be both a way of
implicitly taking responsibility (thereby making yourself responsible), and the main
thing that is needed in order to live up to the responsibility taken.
Still, tort liability has not been vindicated, of course. An understanding of taking
responsibility, even insisting on the moral importance of (sometimes) taking respon-
sibility for the consequences of one’s actions, is one thing; a legal regime resembling
the tort law of negligence is of course quite another. But perhaps this gap can be
bridged. The way forward is, I think, to think about why it is that we have the moral
power—and sometimes the duty—to take responsibility (that is, to make it the case
that we are responsible for something for which we were not, independently of
the taking, responsible). A helpful way of thinking about such questions—why do
we have a specific kind of normative power—is using the heuristic of a moral engineer.
For instance, it is rather clear, at least in outline, why a good moral engineer—who has
to decide what normative powers to endow us with—would endow us with the power
to bind ourselves by promises. Why would she also endow us with the power to take
responsibility?
The power to take responsibility for things that lie outside the scope of one’s core
agency is, it seems to me, a constitutive part of valuable relationships and ways of
thinking of ourselves. As Gary Watson—following Joseph Raz—puts the point: nor-
mative powers “can only be justified if the creation of such special relationships
between people is held to be valuable.”30 Thus, without the power—and sometimes
also the duty—to take responsibility for one’s children’s actions, the nature of

30
Gary Watson (2009), “Promises, Reasons, and Normative Powers,” in David Sobel and Steven Wall (eds.)
Reasons for Action (Cambridge: Cambridge University Press, 2009).
 TORT LIABILITY AND TAKING RESPONSIBILITY 267

parenthood would have been significantly different, and not, it seems to me, for the
better. Similarly for citizenship and the power to take responsibility for one’s country’s
action (assuming, that is, that citizenship is indeed a valuable relationship, an assump-
tion that seems plausible but far from obvious). And agency would have been different
had it been “pure” agency, utterly independent of the consequences our intentions and
the like have in the real world. And though there may be value in sometimes thinking
of ourselves and others as pure agents, still it seems clear that there is value in not
constantly thinking of ourselves in this way. The power to take responsibility for
the consequences of our actions seems necessary for our ability to think of ourselves as
such impure agents.
So much, then, for why it is that we have the normative power to take responsibility.
How about, though, the duty to take responsibility? This too is going to be sketchy, but
the intuitive idea is clear enough: there are ways in which we are morally required to
think of ourselves. In some—perhaps most—circumstances a parent should think of
himself as a parent, he should take this true description of himself as important,
normatively relevant, reason-giving. One of the things thinking of oneself as a parent
(in this normatively rich way) includes is the willingness, even the determination, to
incorporate things like one’s children’s actions into one’s agency. If you do not do that
(in the appropriate circumstances), then you fail to live up to the duty to think of
yourself as a parent. Of course, we cannot force you to think of yourself as a parent.
And if you do not, events like your children’s actions that lie in the penumbra of your
agency will not be ones for which you are responsible (in the sense explained in the
previous section). So the mere fact that these are actions of your children does not
suffice for your being responsible. But if you do not incorporate them into your
agency by taking responsibility for them, you are not thinking of yourself as a parent
(in the normatively rich way needed here). And—being a parent—you should think of
yourself as a parent. Hence, the moral failure. Similarly, you (arguably, and in certain,
but certainly not all, circumstances) ought to think of yourself as your country’s citizen.
And one of the things this means is that you should incorporate its actions—lying in
your penumbral agency—into your agency by taking responsibility for them. But if
you do not think of yourself as your country’s citizen, you are not responsible for its
wrongdoings. It is just that if the circumstances are such that you should think of
yourself as a citizen, the moral failure is precisely in failing to do so, in failing to take
responsibility for your country’s actions.
The case of responsibility for the consequences of your actions—the one directly
relevant to the attempted vindication of the law of torts—is less clear, I think, but is
still plausibly analogous to the two other cases. For arguably, we should take interest
in the world,31 not just in our will or intention, and so we should think of ourselves
as creatures whose intentions and bodily movements have effects in the world. So we
should incorporate the consequences of our actions (some of them, that is, in some

31
This is a central point in Wolf, “The Moral of Moral Luck” (note 14). For a related point, see Robert
Merrihew Adams, “Involuntary Sins,” 94 Philosophical Review 3 (1985), 16.
268 DAVID ENOCH

circumstances)—events lying in our penumbral agency—into our agency by taking

responsibility for them. Furthermore, perhaps we are also morally required to think
of ourselves as limited, as creatures who have some, but far from full, control over
the consequences of their actions in the world.32 If so, we have this further reason to
incorporate the unintended and unforeseen consequences of our actions into our
agency. Indeed, it may be argued that we—all of us, pretty much—already did. For
much like Proud—and much unlike Detached—we do take pride in and claim credit
for the consequences of our actions and projects when they turn out well, as we
intended them, and sometimes even when they turn out surprisingly well, in ways we
did not intend or foresee that they would. If in the political case it is plausible to think
of Proud’s taking pride in his country’s achievements as implicitly incorporating its
actions into his agency and taking responsibility for them, a similar point can be made
about the relation between each and every one of us and the consequences of his or her
actions.
The general suggestion, then, is that we have the (limited) power to incorporate
certain things into our agency by (perhaps among other things) taking responsibility
for them, and sometimes—when doing so is called for by a way we ought to think of
ourselves in this normatively rich kind of way—this is precisely what we should do.
All of this is, of course, very sketchy and somewhat impressionistic—I go through
more (but not enough) details elsewhere.33 But I hope enough has been said to start to
see the relevance of all of this to the law of torts. For tort liability may be thought of as
an institutionalized way of taking responsibility.
One way in which this may be so is instrumental. If the existence of the normative
power to take responsibility and its exercise really do serve some value, then there is
some reason to plan legal arrangements that will encourage the exercise of this power.
To the extent, then, that there is something to be said for us taking responsibility for
the consequences of our actions, there is something to be said for legal arrangements
that encourage our doing so, that foster a taking-responsibility culture, so to speak.
And perhaps something like tort liability is of this kind. If so, we have here an initial
instrumental justification of tort liability.
But much more interesting things can be said here. The relation between tort
liability and taking responsibility may be much more direct than merely instrumental.
Perhaps tort liability can just be a part of the taking responsibility story. I can think of
two ways in which this may be so. The first follows up on a point already mentioned:
seeing that we are all, when it comes to the consequences of our actions, more like
Proud than like Detached, perhaps we are best seen as having already taken respon-
sibility for the consequences of our actions (within some constraints, no doubt). If so,
tort liability may be thought of as a part of what living up to the responsibility already
taken consists in. Indeed, recall my speaking in Section II about the system that is like
New Zealand, except that people are expected to apologize. In Section II, I suggested

32
A conversation with Bob Kane helped me to see this.
33
Enoch, “Being Responsible, Taking Responsibility, and Penumbral Agency” (note 11).
 TORT LIABILITY AND TAKING RESPONSIBILITY 269

leaving the details of this expectation unspecified. But one way of thinking of tort
liability as a part of the taking-responsibility story is precisely as a way of fleshing out
such expectation in an institutional setting (though not for an apology precisely, of
course). Another possibility here is to think of the political decision (to the extent that
there is one) to have tort liability as a part of our legal system, or indeed not to shift to a
New Zealand system, as a collective act of taking responsibility. Thus, it amounts to
our collective decision to be, with regard to the consequences of our actions, more like
Proud than like Detached. To an extent, these two possibilities need not compete—tort
liability may be both a way of living up to the responsibility each of has already
(implicitly) taken for the consequences of her actions,34 and the political decision to
have (or maintain) tort liability may be a collective act of taking responsibility.
This way of thinking of things may have interesting implications. For instance, we
may think about legal liability for the other examples I have been using, and for others
as well, in similar terms. Perhaps, for instance, some instances of collective or vicarious
legal liability are best thought of along such lines. Or perhaps the way to make
normative progress on some of them (do we want tort liability for parents for the
actions of their teenage children?) is to think about it in the terms of taking respon-
sibility (as opposed to being responsible independently of any voluntary act of taking
responsibility), and of the values, if any, for the realization of which such a power—
and perhaps even more to the point here, the duty—to take responsibility is necessary.
Also, thinking about tort liability in terms of taking responsibility (in the sense these
words have been given above) may shed some light on the intuitive distinction here
between tort liability and criminal liability. In the context of the moral luck debate, the
point is sometimes made that while, say, a difference in the criminal punishment for
attempts and complete offenses raises intuitive objections to luck, in the context of tort
liability no such objections apply, at least not as forcefully. But, as I have argued
elsewhere,35 as far as the moral luck arguments and intuitions go, it seems to me that
the criminal case and the tort case are actually quite alike. But now we can say more
here: perhaps in terms of the responsibility that is already there, the tort case and the
criminal case are not that different. But in terms of the appropriateness of taking
responsibility, and the consequences of the responsibility having been taken, they may
be very different. And we know why, too—criminal liability is closely related to moral
blameworthiness. The precise nature of that connection is a controversial matter, of
course, but the thought that some such relation is present shouldn’t be controversial, it
seems to me. But, as emphasized in Section IV, it is not in the sense of responsibility
that is close to blameworthiness that responsibility may be taken. Perhaps sensitivity to

34
As Hanoch Sheinman emphasized to me, there is a sense, of course, in which any tort case that makes it
to the courts is one where the (purported) tortfeasor refuses to take responsibility. But as the text makes clear,
this is not relevant here. If I am right, all of us have already taken responsibility for the consequences of our
actions, so that all that remains now (after the tortious event) is to acknowledge responsibility that is arguably
already there.
35
See Enoch, “Luck between Morality, Law, and Justice” (note 5). See also Waldron, “Moments of
Carelessness and Massive Loss” (note 10).
270 DAVID ENOCH

luck can be justified when it comes to tort law because of considerations having to
do with taking responsibility, but not when it comes to criminal law because taking
responsibility of the kind discussed here is just not relevant there, or at least not in the
same way.36
In these ways, then, the taking-responsibility story may be productive both in
thinking about some common themes across different contexts, and in being sensitive
to important differences between contexts.

VI. Is This Enough?

Let us recap. Taking responsibility is sometimes of value. Indeed, I have argued,
sometimes we have a moral duty to take responsibility, thereby making ourselves
responsible in the sense making an apology, for instance, appropriate. And so, even
though we are not, assuming for the sake of argument, in any sense more responsible
for the risks we create that are actualized than for the risks we create that are not, still
an important difference remains—not in terms of how responsible we are for the
relevant consequences, but in terms of whether we have a duty to take responsibility
for them. A New Zealand system does not incorporate taking-responsibility elements
in a way that a tort system may—this is what is missing in New Zealand. And this
may count at least somewhat in favor of a tort-liability system. I do not have a good
enough feel for what is meant by “corrective justice.” To the extent that I do, though,
this line of thought seems to me to capture something in the vicinity of corrective
justice intuitions.
But even if tort liability can be a way of incorporating the significance of taking
responsibility into our institutional ways of dealing with the costs of negligence, it in
no way follows that this is the only way of doing so. Perhaps, in other words,
something is missing in New Zealand, and that something is not missing in a tort-
liability system, but neither is it missing in some other systems; systems that don’t
include anything like tort liability? If this is a possibility, there is a crucial further step
for the vindication of tort liability: it has to be shown that it does better—in terms of
the values in the vicinity of taking responsibility, or more generally—than other, not
quite New Zealand systems that utilize some other ways of engaging with the values of
taking responsibility.37

36
Relatedly, the question whether parents should be liable in torts for the actions of their teenage children
is, I take it, a legitimate question, but the analogous question with regard to criminal liability is not. And
similarly, I think, for an uninvolved citizen’s responsibility for her country’s actions. This confirms, I think, the
diagnosis in the text.
37
When I introduced the What’s Missing in New Zealand challenge, I started with something like “imagine
your favorite version of the New Zealand legend.” But it may now be thought that if taking responsibility is
indeed of value, and if there are non-tort-liability ways of accommodating that, then the best version of the
New Zealand legend will accommodate the relevant values. If so, something will be missing in the actual New
Zealand, but not in the best, legendary, New Zealand. Needless to say, I’m OK with this way of putting things.
 TORT LIABILITY AND TAKING RESPONSIBILITY 271

I don’t know whether there are such other ways, but I don’t see any reason to think
that there aren’t. Perhaps there are some possible ways of institutionalizing a require-
ment to apologize, or to send the flowers to the hospital, that do not involve all
the messy stuff that is a part of tort litigation, or out of court settlements, or even
insurance. Indeed, if this is really the only thing that is (intrinsically) missing in New
Zealand, then there seems to be something a little primitive about the thought that the
only way to improve, in this respect, on a New Zealand system is by going for a full-
scale tort-liability system. And if this is right—if there are other, better, indeed less
pricey ways of accommodating and engaging with the values relevant to taking
responsibility, the story I have been telling may after all be best read not as a
vindication of tort liability, but as a debunking explanation of its appeal. We find it
appealing because we find taking responsibility appealing, and we’ve not been very
imaginative about other possible ways of accommodating that.
Or perhaps a part of the lesson we can learn here is not about whether or not having
a tort-liability system is desirable, but more specifically (and perhaps, to an extent,
more interestingly) what characterizations we want our tort system to have, assuming
we will continue to have one. In particular, if the main point of having a tort system,
rather than a New Zealand system, is institutionalizing taking responsibility and living
up to the responsibility taken, then perhaps we want our tort system to be more
personal, as it were? After all, if there’s something to be said for an apology or some
such in a case where my negligence causes you harm, and does not merely impose a
risk on you, then it is not desirable (and perhaps not even coherent) to allow my
insurance to take care of that as well.
 13
Exploring the Relationship between
Consent, Assumption of Risk, and
Victim Negligence
Kenneth W. Simons*

I. Introduction
What is the relationship between the following three legal doctrines?
• Consent as a defense to an intentional tort (henceforth, “consent (IT)”);
• Assumption of risk (AR) as a defense to negligence;
• Contributory negligence (CN), i.e., a victim’s failure to use reasonable care with
respect to his own safety, as a defense to negligence and to an intentional tort.
Should these doctrines be no defense, a partial defense, or a full defense? If it depends,
on what does it depend?
Consider five examples.
1. The Death of Houdini
Harry Houdini, the extraordinary magician and escape artist, performed numerous feats
of physical endurance for his various audiences, some of them quite astonishing. One of
his more prosaic feats turned out to be fatal.
Houdini was known to dare others to punch him as hard as they could in the stomach.
On October 20, 1926, after a public performance, a college student met him in his dressing
room, took up the dare and delivered several strong blows to his stomach. Those blows
may well have caused his death a short time thereafter.1

* I am grateful to the participants at the Conference on the Philosophical Foundations of the Law of Torts,
sponsored by Rutgers Institute of Law and Philosophy, November 2012, for their helpful reactions. Special
thanks to Kim Ferzan, who commented on the chapter at the conference, and to Michelle Dempsey.
1
The circumstances of Houdini’s death are hotly contested. It is not clear whether the punches to his
stomach were a cause of his death, nor whether the appendicitis that was diagnosed soon thereafter, and
which led to the peritonitis that led to his death, would have caused his death apart from the blows. Also in
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 273

If Houdini’s estate had sued the student for the intentional tort of battery, the student
would undoubtedly have relied on the defense of consent. Should this be either a
partial or complete defense? Should it be a defense only if it was unreasonable, rather
than reasonable, for Houdini to agree to be punched in the stomach?
2. Jane, a Pedestrian Fleeing from Danger
An armed robber is chasing Jane. Her only escape route is across a street busy with
vehicular traffic. Rushing into the traffic, she is struck by the car of speeding driver Jill.

If Jane sues Jill, should Jill be able to rely on the defense that Jane chose the risk of
injury from traffic over the risk of injury from the robber? Should this be a (partial or
complete) defense only if it was unreasonable, rather than reasonable, for Jane to act as
she did?
3. Yin, a Pedestrian in a Hurry
Yin is running late for the closing time of his child’s daycare center. So is Yang. Yin hurries
across a busy street. Yang drives above the speed limit. Yang’s car strikes Yin.

If Yin sues Yang, should his behavior be a (partial or complete) defense only if it was
unreasonable, rather than reasonable, for Yin to act as he did?
4. Pascal, a Passenger in a Very Great Hurry
Rascal offers his friend Pascal a ride to the airport. At the start of the drive, Pascal
mentions that he is running very late for his plane, and strongly urges Rascal to speed
twenty miles per hour over the speed limit. Rascal does so. In the ensuing crash, Pascal is
badly injured.

If Pascal sues Rascal, should Pascal’s behavior be irrelevant to liability, a partial

defense, or a complete defense?
5. Scott, the Extreme Skater
Dot, the operator of a public skating rink, improperly prepares the ice, leaving it very rough,
and, thus, unusually dangerous for most skaters. She puts up a sign: “Dangerous rough ice.
Sorry. Skate at your own risk. Will be fixed by tomorrow.” Scott, who is preparing for an
outdoor race on ice in very similar conditions, decides to skate around the rink. Due to the
rough ice, he falls and suffers injury.2

dispute is whether the punches were truly consensual: by some accounts, the person who punched Houdini
approached him so suddenly that he did not have time to tighten his abdominal muscles in preparation for the
blows. See “The Life and Death of Harry Houdini,” <http://www.npr.org/templates/story/story.php?storyId=
1485266> accessed October 28, 2013; “Obituary: Harry Houdini Dies After Operations,” <http://www.nytimes.
com/learning/general/onthisday/bday/0324.html> accessed October 28, 2013; “October 31, 1926: Death
Proves Inescapable for Even Houdini,” <http://blogs.smithsonianmag.com/aroundthemall/2011/10/october-
31-1926-death-proves-inescapable-for-even-houdini/> accessed October 28, 2013. For purposes of this chap-
ter, please assume that Houdini did agree to be punched in the stomach, in just the way that he was actually
punched, and that the punches caused his death.
2
The example is loosely based on the case of Meistrich v Casino Arena Attractions, Inc., 31 N.J. 44, 155 A.2d
90 (1959). See Kenneth W. Simons, “Reflections on Assumption of Risk,” 50 UCLA L. Rev. 481 (2002), 490.
274 KENNETH W . SIMONS

If Scott sues Dot, should his behavior be a (partial or complete) defense only if it was
unreasonable, rather than reasonable, for him to continue to skate?
The most defensible answers, as we shall see, are as follows:
(1) Houdini: Consent or AR is a complete defense; if consent is found, the reason-
ableness of plaintiff ’s conduct is irrelevant.
(2) Jane: No consent or AR defense. Reasonableness is relevant, but plaintiff ’s
conduct is reasonable, so damages should not be reduced.
(3) Yin: No consent or AR defense. Reasonableness is relevant, and plaintiff ’s
conduct is unreasonable, so damages may be reduced.
(4) Pascal: Consent or AR is a complete defense.
(5) Scott: Consent or AR is a complete defense.
To explain this pattern of answers, we need to explore some fundamental questions
about the concepts of consent and fault, their respective rationales, and their inter-
relationship. Section II briefly reviews the legal background. Section III explores the
phenomenology and structure of consent, especially the following question: when a
person legally consents to conduct X, and knows that there is a significant risk that
Y will occur (where Y is conduct similar to X or is a harmful result of X), does he
therefore also legally consent to Y? This is a surprisingly common situation, yet there
has been surprisingly little discussion about how to resolve it. Often, we will see, by
legally consenting to X, a person is properly treated as legally consenting to the
“package” of [X plus the risk of Y]. But whether she is properly deemed to accept
such a package deal depends, crucially, on the nature of the tort in question.
Section IV carefully examines whether and how consent (IT) and AR can be
distinguished. Why is the reasonableness or unreasonableness of the victim’s conduct
relevant in some cases but not in others? Why does reasonableness seem more relevant
when the victim decides to encounter a negligently-created risk—cases (2), (3), (4),
and (5)—than when the victim decides to permit a physical contact or intrusion that
would otherwise be an intentional tort—case (1)? Whether Harry Houdini acted
prudently or foolishly in allowing someone to punch him in the stomach seems
irrelevant to his estate’s claim for recovery. If he genuinely consented, surely recovery
should be precluded, even if his decision was reasonable. Doctrinally, this first puzzle
resolves into the question: why is AR such a disfavored defense today? We will see that
treating AR and consent (IT) differently is sometimes justified, but at the same time,
their shared underlying rationale too often is ignored.
Two preliminary notes: first, the term “consent” is notoriously ambiguous. We must
distinguish between the minimal concept of assent, on the one hand, which is, roughly,
a preference or willingness that the conduct occur; and the more robust concept of
legally binding consent, on the other hand, which is assent given with sufficient levels
of knowledge, competence, and freedom of choice.3 In this chapter, the context should

3
See Peter Westen, The Logic of Consent (Burlington, VT: Ashgate, 2004); Kenneth Simons, “Book Review:
The Conceptual Structure of Consent in Criminal Law,” 9 Buff. Crim. L. Rev. 577 (2006), 580–1.
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 275

clarify which concept is at issue. Second, the chapter explores the range of contexts in
which consent (IT) or AR should be a defense, but, for reasons of space, does not
explore the situations in which consent is legitimately ignored or overridden for
compelling paternalistic or other reasons.

II. Legal Background

The basic story of the evolution of contributory negligence (CN) from complete to
partial defense is well known. Throughout most of the twentieth century, both CN and
AR served as complete defenses. But widespread discomfort with the harshness of
these rules caused almost all Anglo-American jurisdictions to reject the all-or-nothing
CN rule with rules of comparative responsibility that permit many plaintiffs who
would previously have been barred to recover a portion of their damages.4
However, rejecting the all-or-nothing remedial consequence of CN leaves unre-
solved the status of AR. Most American jurisdictions also abolished AR, “merging” it
into the new comparative responsibility rules. If the plaintiff acted unreasonably,
he might obtain partial recovery; if he acted reasonably, he would obtain full recovery.
Some Anglo-American jurisdictions continue to recognize AR, but they usually interpret
it extremely narrowly. In England and Canada, for example, AR is usually taken to
require that the plaintiff agreed, not just to accept the physical risks of injury, but
to accept the legal risks. The defendant must show that the plaintiff specifically
intended to waive his legal right to sue.
It is understandable that so many courts have merged AR into comparative fault, or
have interpreted AR very narrowly: a broad interpretation of AR leads to patently
unjust results. For example, in cases where the victim reasonably chooses either to
rescue herself (recall Jane) or to rescue another, a broad interpretation of AR results in
a complete denial of liability.5 In the early twentieth century, especially, broad inter-
pretations of AR frequently led to unconscionable results for employees: if an
employee had the temerity to complain about an unsafe condition in his workplace,

4
To be sure, even after the advent of comparative fault, many jurisdictions continue to employ an all-or-
nothing rule when the victim is negligent but the injurer commits an intentional tort; this categorical rule
ignores victim fault and permits full recovery. However, some jurisdictions have relaxed this rule, permitting
comparative apportionment between an intentional tort and negligent conduct, especially when the negligent
conduct is committed by a co-defendant rather than by the plaintiff, and especially when the “intentional” tort
does not express a serious type of fault.
5
For a recent example of such an unjust result, see Duda v Phatty McGees, Inc., 2008 S.D. 115, 758 N.W.2d
754 (2008), concluding that it was a question for the jury whether a customer at a bar assumed the risk of the
bar’s negligently inadequate security when he stepped in front of a friend to protect him from a broken beer
bottle during a fist fight. Here is the court’s dubious reasoning:
He deliberately placed himself between his friend and two assailants. Heroism can exact a high cost.
Coming to the aid of an outnumbered friend in a fight, though understandable, and even commendable,
still reflects a conscious decision to inject oneself into a volatile and dangerous situation.
Duda, 758 N.W.2d at 759.
276 KENNETH W . SIMONS

his complaint underscored his awareness of the risk and thereby made the defense of
AR easy to establish.
Thus, AR has had a bad rap for good reason: the versions of AR that many courts
traditionally endorsed are unduly broad. But narrower versions of AR, versions that
express a genuine consensual rationale for precluding recovery, are much more
justifiable. In my own writing, I have endorsed a narrow “full preference” approach
to AR and consent (IT).6 Under this approach, AR would be a bar only if P preferred
the option that D negligently offered to the option that P would have confronted if
D had not been negligent. Pascal and Scott are thus valid cases for AR: Pascal preferred
that Rascal drive at a high speed rather than at a safe speed; Scott preferred skating
on the negligently-prepared rough ice to skating on properly prepared ice.
It is not surprising, then, that AR is currently disfavored. What is surprising is
the failure of courts and academics to reconcile the feeble condition of AR with the
robust health of consent to an intentional tort. The feeble state of AR is due in
significant part to the modern view that a victim’s conduct should reduce his recovery
only if that conduct was unreasonable. Yet even courts that endorse this view do not
extend it to consent (IT). Courts uniformly hold that an actor need not show that the
victim acted unreasonably in order to show that the victim validly consented to an
intentional tort, whether the tort is battery, false imprisonment, trespass to land,
trespass to chattels, or invasion of privacy. Harry Houdini’s consent to a punch in
the stomach (a battery) is a complete defense, even if a jury would conclude that his
decision—to expose himself to punches in the stomach, with the accompanying risk of
injury, for the sake of a modest benefit to his career—was shrewd rather than foolish.
The same is true of Houdini’s consent to what would otherwise be a false
imprisonment—by agreeing to be confined and shackled in his famous Water Torture
Cell,7 for example. Similarly, when a patient genuinely consents to cosmetic surgery, or
to surgical option A rather than B, she cannot recover, regardless of whether the fact-
finder would conclude that a reasonable person would not agree to that procedure.
Just as consent (IT) can bar recovery even though, in consenting, the party acted
reasonably, obviously a party can fail to consent (IT) even though that failure might be
viewed as unreasonable. (Suppose she refuses life-saving medical treatment for no
good reason.) Moreover, a victim’s conduct can be unreasonable quite apart from the
connection of that conduct to consent or nonconsent. Inadvertently walking into a
road full of traffic is almost always unreasonable conduct, whether or not the pedes-
trian would have consented to the risks had he adverted to them. In sum, consent and
unreasonable conduct are distinct concepts, and provide distinct reasons for denying
or reducing tort recovery: some fact patterns instantiate both, but others instantiate
only one or the other.

6
See Kenneth Simons, “Assumption of Risk and Consent in the Law of Torts: A Theory of Full Preference,”
67 B.U. L. Rev. 213 (1987); Simons, “Reflections on Assumption of Risk” (note 2).
7
Houdini would be suspended upside-down in a locked glass-and-steel cabinet full of water. The act
required him to hold his breath for more than three minutes.
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 277

Below, we will examine more carefully why AR is so disfavored relative to consent

(IT). But first, we need to better understand what consent, in either guise, is.

III. The Phenomenology and Structure of Consent

Philosophers and legal academics have vigorously disputed the correct phenomenology
of consent as a moral and legal matter.8 Is consent a subjective mental state? An act of
willing? An attitude of acquiescence? Indifference? Is it a preference? A communication?
They also disagree about what kinds of mistake vitiate consent, and, more affirmatively,
about how closely the consenting actor’s belief, desire, attitude, or communication about
the conduct, act, or result assented to must match the conduct (or other consent object)
that actually occurs, in order to count as morally or legally adequate consent.
The bare outlines of tort doctrine here are relatively clear, though the generality with
which the doctrine is stated obscures important difficulties. Consent is “willingness in
fact” that tortious conduct occur.9 The actor’s assent does not count as legal consent
unless it is sufficiently voluntary and knowing. Tort law adopts a subjective conception
of consent, though it also recognizes “apparent” consent (where the defendant reason-
ably believed that the plaintiff subjectively consented, even if the plaintiff did not).10
I believe that the moral conception of consent defended by Larry Alexander best
explains and justifies tort law’s conception: consent is a subjective state of mind, a
conscious forgoing of moral objections to otherwise wrongful conduct.11 But what
kind of choice, preference, or acquiescence should be necessary for purposes of tort
law? The answer here will often differ from the answer appropriate to the moral
domain. In the moral sphere, for example, it might be correct to describe an interaction
between A and B as nonconsensual whenever B knows of a fact unknown to A that
would cause A not to consent and B fails to disclose that fact. But, in tort law, mistake or
fraud vitiates consent only if it pertains to an “essential” rather than a “collateral”
matter.12 Thus, a misrepresentation of motives ordinarily will not vitiate consent.
Suppose B lies and says, “Yes, I really love you,” as an inducement to A to consent to

8
See, e.g., Larry Alexander, “The Moral Magic of Consent (II),” 2 Legal Theory 165 (1996), 166; Heidi
M. Hurd, “The Moral Magic of Consent,” 2 Legal Theory 121 (1996); Alan Wertheimer, Consent to Sexual
Relations (Cambridge: Cambridge University Press, 2004), Chapter 7 (defending a performative account);
Westen, The Logic of Consent (note 3).
9
Restatement (Second) of Torts, } 892(1).
10
Is an external communication of consent to defendant not only sufficient but also necessary to preclude
liability? This is a separate and unresolved question. For an argument that tort liability is sometimes justified
notwithstanding the victim’s private, uncommunicated consent, see Simons, “Consent and Assumption of Risk
in Tort and Criminal Law,” in Unravelling Tort and Crime (Cambridge University Press 2014) (forthcoming).
11
Alexander, “The Moral Magic of Consent (II)” (note 8) at 166.
12
In this respect, tort doctrine bears some similarity to criminal law rape doctrine, under which “fraud in
the factum” (about the nature of the sexual act) vitiates consent but “fraud in the inducement” (e.g., about the
defendant’s motives) does not.
278 KENNETH W . SIMONS

sexual intercourse. This, without more, does not demonstrate that A has not legally
consented.
Moreover, whatever we decide is the most appropriate mental state, attitude, or
communication to satisfy legal consent standards, a critical additional issue is how
closely that mental state must “match” the world. If A agrees with B to play a game of
“touch football,” what kinds of “touches” have they consented to? Only those contacts
that are necessary to tag the other player? Or, in addition, incidental bumps while
running towards another player? Hard tackles?13
To answer the matching question, it might seem that we should require either a
precise meeting of the minds between the victim and injurer, or at least a description
by the victim of the conduct consented to that precisely matches the conduct that
actually occurs. But these requirements are too demanding. Consider the criterion
suggested by the Restatement (Second) of Torts:
In order to be effective, the consent must be to the particular conduct of the actor, or to
substantially the same conduct. Thus consent to a fight with fists is not consent to an act of
a very different character, such as biting off a finger, stabbing with a knife, or using brass
knuckles. Very often the question whether the particular conduct is within the scope of the
consent given becomes a question of degree. Minor differences in degree or extent, such as
the fact that the force exerted by the actor in delivering a blow is slightly greater than
would ordinarily have been contemplated, usually will not be held to exceed the consent,
although a much greater force would clearly exceed it.14

But is this criterion too lax? Why not limit legal consent to where P assents to
precisely the conduct that occurs? Why should we treat P as consenting to any
deviation from what he expected to occur, even an “insubstantial” deviation? Diffi-
culties of proof are a partial answer, but they are not the complete explanation. Even if
we were absolutely certain that Houdini expected only an average punch in the
stomach, we should not permit him to recover if the punch exerted twenty per cent
more force than that. One reason is the unfairness of expecting the puncher to
calibrate his force that closely, which might be very difficult or impossible to do.
Another is the concern that insisting on a very precise match between the parties’
expectations and their conduct would undermine the autonomy and efficiency values
that the social practice of legal consent serves.
And this suggests a broader point. Consent in the law very often amounts to a
“package deal” in the following sense: P consents to X [the species of physical conduct
that would, absent consent, amount to the relevant tort], aware that there is some risk
that D will deviate from X in an insubstantial way, and some risk of an untoward
consequence of X occurring. Call such a deviation or consequence Y. P rarely consents

13
As Hurd explains, whether D has consented to an act X or a consequence Y is a de dicto rather than de re
question: it depends on the description of X and Y. Hurd, “The Moral Magic of Consent” (note 8) at 126–7.
14
Restatement (Second) of Torts, } 892A, comment c (emphasis added). The black letter of this provision
states: “To be effective, consent must be . . . to the particular conduct, or to substantially the same conduct.”
Restatement (Second) of Torts at } 892A(2)(b).
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 279

to Y in the fuller sense that he consents to X. Although he might prefer X to all

alternatives, or might acquiesce to X occurring, considered by itself, often he will not
have a comparably favorable or willing attitude towards Y.15 So when we say that P, by
voluntarily and knowingly consenting to X, thereby consents to Y, often we mean only
that the relationship of X and Y is sufficiently close, and well enough understood by P,
that it is just to deny P recovery for Y as well as for X. By consenting to X, P is deemed
to consent to the package of [X plus the risk of Y]. And this, in turn, means that P has
no legal complaint about suffering result Y if the known risk of Y is realized.
Thus, when Houdini agrees to be punched in the stomach, he undoubtedly knows
that the punch might be a bit more forceful than average, and also undoubtedly knows
that by permitting a punch he is therefore risking a stomach injury. What he actually
assents to (and legally consents to) is a punch. He assents (in a different and weaker
sense) to the risk of a slightly more forceful punch or of a stomach injury. And he
thereby does not have, and should not have, any valid legal complaint if the more
forceful punch or the injury actually occurs. We thus reach a somewhat surprising
conclusion: many consent (IT) cases actually involve consent or AR with respect to a
risk of harm. That is, when P legally consents (IT) to a touching or confinement, often
P should also be deemed to have legally assumed or consented to a risk (typically only
a small risk) of a slightly different touching or confinement, or of further physical
harm. Ordinarily, however, an actor who consents to a physical contact should not be
deemed to consent to a very high probability of physical harm.16 When Houdini tells a
stranger, “Go ahead, punch me as hard as you can,” it would not be plausible (absent
further facts) to interpret this as legally valid consent for the stranger to use hidden
brass knuckles or a knife.
The significance of the package deal expansion of consent from X to Y is not widely
appreciated. We like to think that consent requires assent to precisely the conduct that
P confronted or precisely the consequence that befell him. Yet actors are often (and
justifiably) deemed to have legally consented, even where such assent is lacking.
Suppose I prefer to ride the subway without any risk of physical contact. And
I vociferously announce this preference to other riders. Despite the absence of my
actual or even apparent consent, I will be deemed to have consented to the minor
physical contacts that predictably occur in this context.17 Similarly, I might prefer to

15
Peter Westen argues that “consent to the risk of Y” does not exist unless P consented to (a certainty of )
Y itself occurring. Westen, The Logic of Consent (note 3) at 280–4, as discussed in Simons, “Book Review: The
Conceptual Structure of Consent in Criminal Law,” (note 3) at 621–9. But that is a very strict interpretation of
what constitutes consent, much stricter than courts and commentators ordinarily employ. And it is quite
coherent and defensible to view someone as consenting to a risk of Y even though he did not or would not
consent to a high probability or certainty of Y occurring. Nevertheless, Westen’s argument does underscore the
importance of providing an adequate explanation of when it is justifiable to hold that P’s consent to X, knowing
that thereby he is creating a risk of Y, also should be treated as precluding recovery for Y if it occurs.
16
This obviously depends on the context. In medical operations, patients often do consent to a high risk of a
seriously harmful side effect.
17
Courts sometimes employ the category “implied-in-law” consent for this type of case. The category is
more widespread than generally appreciated. See Westen, The Logic of Consent (note 3) at 272–8, 322
280 KENNETH W . SIMONS

play the quarterback position in football without ever being touched or without any
risk of injury; nonetheless, if I choose to play, I will be deemed to have consented to the
package of risks inherent in the game. In short, in many of the examples that we have
been considering, especially when Y is a harmful consequence of X, P assents to
X despite the risk of Y—for P is often quite unwilling that Y occur and, indeed, wishes
that he could avoid any risk of Y. Nevertheless, as a matter of law, his actual assent to
X is deemed to amount to a legal consent to the risk of Y (and, thus, constitutes legal
consent to Y, should Y occur). From a broader perspective, P’s consent to conduct
X often entirely vitiates the wrongfulness of that conduct, while his consent to the risk
of harmful result Y only rarely (if ever) eliminates the harmful quality of that result. So,
if consent justifiably precludes recovery in the latter scenario, it will do so because of
the package deal argument (or a comparably strong justification).
However, not all tort cases in which the issue of consent arises have this package
deal structure. In many intentional tort cases, the victim is entitled to a damages
remedy without any proof that he suffered physical or even emotional harm as a result
of the tortious invasion.18 Thus, if P is intentionally assaulted or confined without his
consent, that conduct alone (X) warrants a damage remedy; P need not prove a
resulting physical or emotional harm (Y).
What, then, is the requisite connection between X and the risk of Y that justifies
treating legally adequate consent to X (that also creates a risk of Y) as also amounting
to legally adequate consent to Y, if Y occurs? The answer depends crucially on
context—i.e., the nature of the interests furthered or protected by the particular legal
right as well as the nature of the activity in question. Obviously, the necessary and
sufficient conditions for consent in criminal law, contract law, and tort law differ.19
Within tort law, the requirements for consent to medical treatment include duties to
inform the patient of the risks of treatment, duties that do not exist in most other
contexts in which actors consent to physical contacts. And conduct that suffices as
consent for purposes of a false imprisonment claim might well not suffice for purposes
of a battery claim. If store personnel ask a suspected shoplifter to come with them to a
back room, where they suddenly reach into her pocket to search for stolen items, it is
likely that in choosing to accompany them, she will be deemed to have legally consented
to what otherwise would be an unlawful confinement. However, it is unlikely that her
conduct will be treated as consent to what otherwise would be a battery.
Finally, the proper remedy for conduct that is tortious because of lack of consent
depends on why consent is required and precisely how the consent offered was deficient.
As we have seen, a remedy can be awarded for an assault or false imprisonment even

(discussing the concept of “constructive consent”), as discussed in Simons, “Book Review: The Conceptual
Structure of Consent in Criminal Law” (note 3) at 616–21.
18
See Dan B. Dobbs, Paul T. Hayden, and Ellen M. Bublick, The Law of Torts (Eagan, MN: West
Publishing, 2011), } 47.
19
For example, in many jurisdictions consent to an illegal fight precludes tort liability but not criminal
liability.
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 281

absent proof of resulting physical or emotional harm. What if the defendant never even
sought the victim’s consent, but the victim would have consented? In this scenario, too, a
tort remedy to address the dignitary injury is sometimes appropriate. However, in the
analogous context of breach of the duty of informed consent—where medical personnel
fail to provide adequate information to a patient about the risks of a proposed medical
treatment—the standard American approach is to deny any remedy unless the patient
can show that he (or, in many jurisdictions, a reasonable patient) would have decided
differently and thus would have avoided the physical harm he suffered from the
treatment. This requirement, to show “decision causation,” is defensible, though only
if the failure to obtain informed consent is properly conceptualized as a negligence claim
and only if we are justified in requiring proof of physical harm, not just dignitary injury,
for this type of negligence claim.

IV. Comparing Consent (IT) with AR

Let us now take a closer look at the issues identified in the introduction, especially the
surprisingly narrow scope of AR. This section will first analyze paradigm instances of
consent (IT) and AR. Next, we will identify an “apples and oranges” difficulty in
comparing the two doctrines: the doctrines can apply to different types of tortious
wrongs. We will then consider, and reject, the possible criterion that consent (IT)
applies only when P believes that the harm or the tortious conduct is highly
probable, while AR applies only when he believes that these are unlikely to occur.
Finally, the concluding subsection will explain that factual differences between
paradigm consent (IT) and AR scenarios partially justify their different treatment,
but will also suggest that the consensual rationale underlying many AR cases is too
often neglected.20

A. Distinguishing paradigm instances of consent (IT) and AR

Consider first some paradigm cases over which there is no disagreement. The Houdini
case is a paradigm instance of consent (IT). Absent Houdini’s agreement to be
punched, the conduct of the other would constitute an intentional battery (a purpose-
ful or knowing touching); and, at the time Houdini consents, he believes that the other
is very likely to engage in such conduct.21 Pascal is a paradigm instance of AR. Absent
Pascal’s agreement to Rascal’s driving at a high rate of speed, defendant Rascal’s
conduct would be considered negligent towards plaintiff Pascal; and, when Pascal

20
For some prior analyses of the issue, see Simons, “Assumption of Risk and Consent in the Law of Torts:
A Theory of Full Preference” (note 6) at 248–58 (and sources cited therein); Simons, “Reflections on
Assumption of Risk,” (note 2) at 518–25.
21
To simplify matters, assume that the student asked Houdini if he could punch Houdini in the stomach
and Houdini said yes.
282 KENNETH W . SIMONS

agrees to Rascal’s dangerous conduct, Pascal believes that the relevant risk of physical
harm that Rascal has created or is about to create is only a possibility of harm, not a
very high likelihood.
But it is unclear which features make these cases “paradigm.” For there are actually
three distinctions at play here: (1) between a low and high probability that the relevant
(otherwise tortious) conduct will occur; (2) between a low and high probability that
harm, offense, or some other legally compensable result of that conduct will occur; and
(3) between legally valid consent to a tort of negligence22 and legally valid consent to
an intentional tort. Which factor matters? The characterization AR could mean any or
all of the following:
(1) The consenting actor believed there was a low rather than high probability
that the other would act tortiously (either negligently or by committing an
intentional tort);
(2) The consenting actor believed there was a low rather than high probability that
harm (or some other compensable injury) would result from the other’s tort; or
(3) The actor consented to otherwise negligent conduct rather than to an inten-
tional tort.
Insofar as many jurisdictions interpret AR much more narrowly than consent (IT),
this is not merely a terminological dispute. Suppose, at the time of his alleged consent,
plaintiff agrees to accept a small risk that a later intentional battery will occur. This
counts as AR under (1) but not under (3). It is also a quite common scenario, for it
frequently arises in sporting and recreational activities. Consider the physical contacts
that often occur in soccer, baseball, and many other sports. Should the narrow (or
non-existent) defense of AR apply or the apparently broader defense of consent (IT)
instead?
Moreover, although (2) is more often satisfied in negligence cases than in inten-
tional tort cases, this relationship is contingent. Sometimes the relevant tortious
conduct is at worst negligent, yet the plaintiff believes that the tort is quite likely to
result in harm. This counts as AR under (3) but not under (1) or (2). (Suppose a
suicidal pedestrian waits for a speeding car and then leaps into the car’s path.)23
We obtain little guidance on the question from actual judicial practice. Most courts,
insofar as they have explicitly addressed the issue, endorse (3), rather than (1) or (2), as
the criterion for distinguishing AR from consent (IT).24 At the same time, courts also
frequently address the issue of whether and when participants in recreational and

22
AR rather than consent is also the terminology typically used when the plaintiff accepts the risks arising
from torts of recklessness or strict liability.
23
Moreover, in some rescue cases, the defendant’s negligence prompts plaintiff to try to save his own life or
the life of another in a manner that he realizes is quite likely to lead to self-injury. However, although some
courts treat a choice to rescue as an instance of either consent (IT) or AR, I believe we have good reason to treat
most rescues as not falling within either category.
24
“Consent to conduct that is merely negligent, creating an unreasonable risk of harm, is commonly called
‘assumption of risk’. ” Restatement (Second) of Torts, } 892 (1965), comment a.
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 283

sporting activities are barred by consent, and in this context some use the language of
“consent” while others prefer the language of “assumption of risk.” For purposes of
clarity, it would certainly be useful if we had distinct terms for these different
concepts.25
How should we distinguish AR from consent (IT)? Or should they be distinguished
at all for legal purposes? We shall see that (3) is not, by itself, a satisfactory distinction
because it compares apples to oranges. I will then show that (1) and (2), although
coherent distinctions, do not justify treating AR differently from consent (IT) for
purposes of applying consent principles, though they sometimes justify differential
judgments of comparative fault. Finally, we will see that paradigm cases of AR and
consent (IT) are indeed often distinguishable, but only because of contingent factual
differences that often attend such paradigm cases, not because they differ in under-
lying rationale.

B. The “apples and oranges” problem

The usual manner of comparing AR and consent (IT) poses a serious and insufficiently
appreciated problem: when the comparison takes the form of version (3) above, we are
comparing apples and oranges.26 Intentional torts such as battery do not simply
protect against the intentional invasion of the same interests whose unintentional
invasion is protected by negligence. Rather, the intentional torts protect some rights
and interests not protected by negligence law at all—for example, the right to auton-
omy with respect to one’s bodily integrity, which is violated by a nonconsensual
touching, and the right to freedom of movement, which is violated by a nonconsensual
confinement. If an actor unintentionally trips into someone, touching him without his
consent, her conduct is neither a battery, a viable negligence claim, nor any other
tort.27 If an actor carelessly but unknowingly locks someone in a room, confining him
without his consent, her conduct is neither false imprisonment nor any other tort.
Thus, although it is commonly assumed that intentional torts reflect a more egregious
type of fault than torts of negligence, this assumption is not invariably justified.
More precisely, it is only justified insofar as the intentional tort in question implicates
precisely the same right or interest as the corresponding tort of negligence.28 Thus, the

25
The following terminology might be suitable for the three categories:
(1)
“consented to [the specified tortious] conduct” or “was willing that D engage in [that conduct]”;
(2)
“assumed the risk of harm [or offense, etc.] from [the specified conduct]”; and
(3)
“consented to D’s negligence” or “to [what otherwise would be] D’s intentional tort.”
26
See Kenneth Simons, “A Restatement (Third) of Intentional Torts?,” 48 Ariz. L. Rev. 1061 (2006),
1080–3.
27
Negligence liability requires physical harm; but one can commit a battery without causing physical harm.
28
Indeed, it is not always true even in this situation. For example, knowingly causing physical harm to
another is not always more blameworthy or a more serious type of fault than negligently causing physical harm
to another. The first is sometimes justified by necessity or self-defense. But the second is by definition
unjustifiable, given the usual understanding of negligence as an all-things-considered judgment of fault.
284 KENNETH W . SIMONS

assumption is warranted in the following situation. Negligently causing physical harm is

one type of fault. Intentionally (i.e., knowingly or purposefully) causing the same type
and degree of physical harm ordinarily is, indeed, a more serious form of fault. And a
similar hierarchy of fault ordinarily obtains when we compare intentionally or recklessly
causing emotional distress with negligently causing the same type and degree of
emotional distress.
Yet there is no recognized intentional tort in Anglo-American law that exactly
corresponds to the general duty not to negligently cause physical harm (or to any other
duties not to negligently harm others). The tort of harmful battery, i.e., battery that
causes physical harm, is the best candidate. But it does not fully correspond. Battery, of
course, requires a physical contact. So, even in those jurisdictions in which harmful
battery requires an intention to cause physical harm,29 the tort also requires that such
harm be caused by means of the defendant’s contacting the plaintiff ’s person (or
something closely connected with his person). A defendant does not commit a battery
if he purposely or knowingly causes physical harm to the plaintiff by some other
means. If Katie abruptly breaks up with her boyfriend Tom, knowing that, given his
sensitive disposition, this will cause him serious emotional distress and resulting
physical illness, she has not committed a battery.30

C. A possible criterion: Plaintiff ’s belief about

the probability of harm
Nevertheless, it is illuminating to consider how the distinction between AR and
consent (IT) would play out if the law did recognize a general intentional tort of
“intentionally (i.e., purposely or knowingly) causing physical harm,”31 corresponding
to the well-recognized general tort of negligently causing physical harm. Indeed,
suppose that all the intentional torts were simply “higher culpability” versions of
corresponding torts of negligence: intentionally inflicting severe emotional distress
would then be the more culpable variant of negligently inflicting severe emotional
distress, and so forth.
In this imaginary legal world, the objects of AR and of consent (IT) would both be
apples, rather than an apple and an orange. But the consensual object of AR would be a
smaller apple (otherwise tortious conduct that negligently risks harm), while the object
of consent (IT) would be a larger one (otherwise tortious conduct by which the actor
intends to cause harm). How, in this world, would AR and consent (IT) operate and

29
American jurisdictions are split on this question. Some require only the single intent to cause a contact
that is unpermitted; others require, in addition, either the intent to cause physical harm or the intent to cause
offense.
30
Or, if a malicious or indifferent doctor fails to prescribe a medicine for P’s severe illness that proper
medical care requires, knowing that her omission will cause P serious physical pain and suffering, the doctor is
not liable for a battery.
31
See Restatement (Third) of Torts, Physical Harm, } 5 (“An actor who intentionally causes physical harm
is subject to liability for that harm.”).
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 285

how would they relate to each other? The question is instructive. Answering it should
help us to determine whether the complexity of the relationship between AR and
consent (IT) is due to the complexity and incommensurability of extant negligence
and intentional tort doctrines or instead is due to something special about how the
concept of consent operates when the object of that consent is a tort reflecting a lower
rather than higher degree of fault. The former, we will see, is closer to the truth.
Thus, compare two situations: in one, marked as (a), V1 consents to D1 acting in a
way that each knows is very likely to cause a minor personal injury to V1. In the other,
marked as (b), V2 consents to D2 acting in a way that each knows creates a small risk
(but not a high likelihood) of causing a minor personal injury to V2.32
Scenario 1: Mutual Boxing and Tennis Matches

(a) Victor agrees to a friendly boxing match with Dennis. Both know that Dennis will land
blows that are very likely to cause Victor at least minor physical harm. Victor is injured by
one of Dennis’s blows.
(b) Vicky agrees to a friendly tennis match with Debby. Both know that in the course of the
match, there is a small chance that Debby will cause Vicky at least minor physical harm
by accidentally striking Vicky with a tennis ball that Debby hits during play. Vicky is
injured in this manner.

Is consent more likely to bar recovery in Scenario 1(a) (where the conduct is otherwise
an intentional tort) than in Scenario 1(b) (where the conduct is otherwise at worst a
tort of negligence)?33 In these examples, there seems to be no relevant difference. If the
assent to each match is equally voluntary and the two victims are equally knowledge-
able about the degree of risk, then consent should be equally effective.
At the same time, analogous unilateral scenarios would be equally nonconsensual:
Scenario 2: Unilateral Boxing and Unilateral Striking of Tennis Ball

(a) Dennis suddenly approaches Victor, a stranger, and lands several punches to Victor’s
face. (Both again know that this is likely to cause some physical harm.) Victor is injured
by one of Dennis’s blows.
(b) Debby suddenly approaches Vicky, a stranger, and hits a tennis ball in her direction.
(Both again know that this poses a small risk of causing some physical harm.) Vicky is
injured.

32
These two scenarios focus on factor (2) above. Analogous scenarios could be created for factor (1).
Moreover, each of the two scenarios involves knowledge. One could formulate analogous consensual scenarios
involving purpose to cause (a) minor physical injury or merely (b) a risk of minor physical injury, e.g., cases of
sadomasochistic sexual conduct. The latter scenarios will arise less often, so I focus instead on scenarios
involving knowledge that one will or might cause harm.
33
In some jurisdictions, Debby would, absent consent, still be liable for a battery, if she believes it very likely
that one of her shots will hit Vicky. But for purposes of this discussion, we are assuming that battery requires
purposely or knowingly causing harm, a level of fault that Debby lacks.
286 KENNETH W . SIMONS

This second set of scenarios still yields no difference between (a) consent (IT) and (b)
AR. Both victims are clearly entitled to recover in tort.
For our next variation, suppose that the injurer again acts unilaterally, but there-
after, the victim has an opportunity to choose whether to confront either the high or
low risk of harm. (In the prior scenario, the victim had no such opportunity.)
Scenario 3: Threat to Punch or to Strike with a Ball

(a) Dennis suddenly approaches Victor and threatens to punch him in the face. Victor says
nothing. Dennis makes good on his threat, injuring Victor. (Both again know that the
threatened action is likely to cause some physical harm.)34
(b) Debby suddenly approaches Vicky and threatens to hit a tennis ball in her direction.
Vicky says nothing. Debby makes good on her threat by hitting the ball, which injures
Vicky. (Both again know that the threatened action poses a small risk of causing some
physical harm.)

Here, too, there seems to be no relevant difference that affects whether the victim
should be able to recover in full. Neither victim has actually (or apparently) consented
to the tortious conduct of the other. And neither victim is negligent.
But consider yet another variation that might justify different results in the two
scenarios—namely, “rescue” scenarios that differ greatly in the probability that the
victim will suffer harm.
Scenario 4: Rescuing a Smartphone

(a) After crossing the road, Victor realizes that he dropped his smartphone in the middle of
the road. He decides to retrieve it, even though he sees that driver Dennis is speeding
nearby and is very likely to run into him, causing him injury. Dennis recognizes this as
well. Dennis’s car hits Victor, causing injury.
(b) After crossing the road, Vicky realizes that she dropped her smartphone in the middle of
the road. She decides to retrieve it, even though she sees driver Debby speeding half a
block away. Vicky recognizes a very small chance that Debby will not be able to stop in
time and will cause her injury. Debby’s car hits Vicky, causing injury.

In this scenario, a jury is justified in finding Victor contributorily negligent, but is not
so clearly justified in finding Vicky to be contributorily negligent. The obvious

34
See Restatement (Second) of Torts, (1965), } 892, illustration 4:
In the course of a quarrel, A threatens to punch B in the nose. B says nothing but stands his ground.
A punches B in the nose. A is not justified upon the basis of apparent consent.
A cannot successfully invoke either B’s actual or “apparent” consent. Apparent consent is lacking because a
reasonable person in A’s shoes would not conclude that B actually consented.
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 287

difference between the cases is that an actor’s modest interest in saving his property is
insufficient to outweigh a high risk of self-injury but might suffice to outweigh a much
lower risk of the same harm.35
What is telling, however, is that the only scenarios in which the proper legal
treatment of (a) differs from the proper treatment of (b) are those in which a
substantial decrease in the probability of risking harm to oneself changes the actor’s
conduct from unreasonable to reasonable (or at least from highly unreasonable to
mildly unreasonable). But there is no reason why we should treat such a decrease in
perceived probability of self-harm, without more, as transforming an actor’s conduct
from consensual to nonconsensual (or vice versa). Thus, in Scenario 4, neither victim
consents to the driver’s speeding.
Reconsider the five introductory examples. In all of the examples, an increase or
decrease in the probability of self-harm does not change whether the victim consented,
though in some it might change whether, and to what extent, the victim’s conduct was
unreasonable.36

D. Factual differences between paradigm consent (IT)

and AR scenarios
If the arguments thus far are correct, a consensual rationale underlies many cases
conventionally categorized as AR as well as cases of consent (IT), and that rationale is
not undermined simply because the consenting party believes the tortious conduct or
the resulting harm is improbable rather than probable.
Why, then, have so many courts flatly rejected AR in recent years? This section
addresses one set of reasons: quite often in paradigm AR scenarios, the consensual
rationale is factually inapplicable, while in paradigm consent (IT) scenarios, that rationale
is much more likely to apply—either because the two parties mutually benefit from the
interaction or because the allegedly tortious party justifiably relies on the other’s assent.37

35
On the other hand, Dennis is more culpable than Debby insofar as the risk of harm he believes that he is
posing is higher than the risk of harm that she believes she is posing. That militates in favor of Dennis’s victim
obtaining a larger portion of his damages than Debby’s victim obtains (under comparative apportionment).
We could isolate the legal significance of these contrasting features, of the degree of fault of the victim vs. the
degree of fault of the injurer, by modifying the scenarios. Suppose Victor believes that he is very likely to be
injured, but Dennis believes that he is not very likely to injure (or vice versa). I will spare the exasperated reader
the burden of reading four (or eight) more scenarios.
36
In all of the cases from the introduction, a difference in the probability of harm might affect whether (or
the extent to which) the victim is acting unreasonably and, thus, whether he should, on a comparative fault
assessment, obtain a lesser recovery for the harm. (In Jane’s case, the probabilities are unlikely to matter unless
she knows she is very likely to die from running into the traffic.) Of course, the reason why the actor is choosing
a risky option is highly relevant. If Jane were running across a busy street in order to avoid being tagged in a
game of touch football, she could easily be found CN.
37
Another important reason is that courts sometimes preclude tort liability by characterizing AR cases as
cases falling within “no duty” or “limited duty” rules. Thus, instead of saying that Scott assumes the risk, they
might say that Dot owes only the limited duty to warn him of the risks. See Simons, “Reflections on
Assumption of Risk” (note 2) at 497–503.
288 KENNETH W . SIMONS

It is perfectly defensible to take these factual differences into account, but this is
quite consistent with the point that the same consensual rationale underlies those cases
in which courts properly recognize AR and those in which they properly recognize
consent (IT).
Situations in which a person consents to a potentially harmful battery very often
involve mutual benefit—for example, medical treatments, other physical therapies,
sexual relations, contact sports, and recreational activities. Many other consent (IT)
cases at least involve reliance: D engages in a potentially dangerous activity only after
discovering or reassuring himself that P assents. But many AR scenarios involve no
such mutual benefit, so it is much less clear that the consent is sufficiently voluntary.
Often, D acts, then sets the stage for P to choose whether or not to accept the risk that
D has created. But at that point, P’s choices might be so constrained that it is unjust to
treat him as consenting to the dangerous condition. (This is often true of unsafe
conditions that employees face in the workplace.) Similarly, many AR scenarios
contain no reliance by D, so D has no valid complaint that recovery would be unjust
on that ground. Often, D is not in a position even to know whether P does accept the
risk. And even if he does know, D might not be in a position to minimize or eliminate
the risk. On the other hand, in those AR scenarios where both parties do benefit from
the risky activity, or where D relies on P’s willingness to run the risk, the case for
denying recovery is much stronger. (Recall Pascal and Scott,38 from the Introduction.)
In many consent (IT) cases, moreover, the parties do not simply mutually benefit
from a particular type of invasive or risky activity: they specifically choose to engage in
an activity at a particular level, or characterized by a particular quality, of risk. Houdini
can choose to be punched once or many times, or to be pinched rather than punched.
A group of friends can play touch, instead of tackle, football. This power to choose the
“rules of engagement,” to adjust the risk levels to the preferences of the other or others,
differentiates typical consent (IT) from typical AR scenarios. It is thus both understand-
able and defensible that courts more readily exclude liability in consent (IT) scenarios.
Consider a related point. Apparent consent (IT) is just as complete a defense as actual
consent (IT). Then is apparent AR similarly a defense in jurisdictions that recognize AR?
Suppose a medical patient tells his doctor that he understands the nature of the operation
that she is proposing, but, actually, he is afraid to admit that he does not. Apparent consent
(IT) will protect the doctor from tort liability because she has a reasonable basis for
believing that P actually consented. Does apparent AR similarly protect D when P does
not actually agree to accept a risk but D reasonably believes that he does?
The doctrinal answer is unclear,39 but I believe the dearth of such cases is simply due
to the paucity of occasions in which D justifiably relies on P’s apparent AR. If such an

38
Scott would be an example of reliance if the rink operator considered closing down the rink but left it
open to accommodate Scott.
39
Simons, “Reflections on Assumption of Risk” (note 2) at 251–2; see Restatement (Third) of Torts:
Apportionment of Liability, } 3, comment c (“Whether the defendant reasonably believes that the plaintiff is
aware of a risk and voluntarily undertakes it may be relevant to whether the defendant acted reasonably.”).
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 289

occasion did arise, however, then apparent AR is as defensible a doctrine as apparent

consent (IT). Thus, in the Scott example from the introduction, suppose Dot, the rink
owner, reasonably believes that Scott fully understands the rough condition of the ice
(based on a conversation between them), but Scott, in fact, does not appreciate the risk.
If Dot therefore permits Scott to continue to skate, she should be able to rely on
apparent AR to preclude Scott’s recovery for his resulting injuries.
Indeed, in one important category of cases, apparent AR is implicitly recognized.
When individuals agree to play a sport or engage in a recreational activity involving
physical contact, courts often apply a special set of rules. They frequently deny
recovery for ordinary negligence and impose only a limited duty not to recklessly or
intentionally cause harm by creating risks that are not inherent in the sport or activity.
To some extent, this set of rules reflects the risks to which the parties actually do and
do not consent. But the rules require no individual inquiry into how knowledgeable
and voluntary the plaintiff ’s choice to accept the risk was. So they might implicitly
reflect the view that a reasonable person in the defendant’s shoes would believe that
the plaintiff consented, even if a particular plaintiff did not consent. However,
I concede that the rules are wider in scope than this: they reflect the view that
enthusiastic and fervent participation in sports and games is socially valuable and
thus justifies making participants responsible for the inherent risks of such activities,
quite apart from whether the participants actually or even apparently consent.
Finally, if a jurisdiction interprets AR according to the narrow “full preference”
approach that I endorse,40 AR can easily be reconciled with consent (IT). In many AR
cases, D has breached her duty to P to provide a less risky option, but that option is not
available to P at the time he encounters the risk. Instead, P is faced with a difficult
and unfair choice: either engage in a desired activity yet confront the negligently
created risk or do not engage in the activity at all. The third option—engaging in the
activity while facing a lesser risk—is usually not open to P. (Skaters in the Scott
scenario are faced with the decision whether to skate on dangerous ice or wait until
tomorrow to skate; they lack the third option to skate now on safe ice.) By contrast, in a
typical consent (IT) case, if D offers P the choice of a dangerous interaction or no
interaction, there is no unfairness in asking P to make that choice because he can
readily avoid the interaction altogether. If the student who approached Houdini
proposed punching him in the stomach with brass knuckles, it would have been
easy enough for Houdini to decline. And normally, unlike an actor who is negligent
towards P, an actor who is alleged to be an intentional tortfeasor has no duty to offer a
third, less dangerous option.41

40
See Simons, “Assumption of Risk and Consent” (note 6).
41
Normally this is true, but not always. Suppose P goes to an emergency room, the only doctor on duty is
drunk, and P assents to being treated by him, given the dire circumstances. P does of course have a negligence
claim if the doctor provides substandard care that results in physical harm. If the doctor treats him with
reasonable care but the operation causes a harmful side effect, it would be plausible to permit a battery claim,
because P did not consent to being treating by an intoxicated doctor, and the doctor had a duty to provide
medical care in a sober condition.
290 KENNETH W . SIMONS

V. Conclusion
This inquiry into the relationship between consent (IT), AR, and CN has offered some
specific lessons, and has also unearthed some broader themes. The lessons include the
following:
• Often, when P consents to D’s engaging in conduct X, knowing that D might
engage in somewhat different conduct Y or might create a risk of harm Y, P is
deemed to legally consent to Y. Whether it is justifiable to hold P to this kind of
“package deal” depends on the nature of the rights that the particular tort
protects.
• Many consent (IT) cases therefore actually involve AR with respect to a risk of
harm, in the following sense: the actor is deemed, as a matter of law, to have no
basis for complaint when the harm occurs, even if the actor did not assent to that
risk of harm, considered by itself.
• Whether AR is normatively distinguishable from consent (IT) depends on how one
distinguishes them—according to the nature of the tort consented to, the likelihood
that the tort will occur, or the likelihood that the tort will result in harm.
• Comparing the interests protected by the right not to be negligently harmed with
the interests protected by the various intentional torts compares apples to oranges.
• Consent (IT) and AR (when appropriately narrowed) are equally valid bases for
precluding recovery.
• It is justifiable to invoke consent (IT) more often than AR, not because the
doctrines differ in principle, but because the scenarios in which each doctrine
commonly arises often differ factually in relevant ways. In paradigm consent
(IT) scenarios, the two parties mutually benefit from the interaction or the
allegedly tortious party justifiably relies on the other’s assent. These factual
features support the consensual rationale underlying both doctrines; but they
are often lacking in AR scenarios.
The analysis also has two broader implications. The first concerns the role of reason-
ableness and consent in tort doctrine and theory. Reasonableness is not all that
matters, nor all that should matter. To be sure, comparative fault principles properly
give significant weight to the fault of the various parties, judged by whether they
departed from a standard of reasonable care. But principles of reasonableness are
often legitimately trumped by principles of consent, which may take the form of a
complete defense or instead operate as a crucial component of a no-duty or limited-
duty rule.
The second broader implication is that Anglo-American tort law is pluralistic. The
rights and interests protected by tort law, and the manner by which they are protected,
are varied. A single, unified hierarchy cannot explain these phenomena. That is one
reason why it is so difficult to compare AR and consent (IT).
 CONSENT , ASSUMPTION OF RISK , AND VICTIM NEGLIGENCE 291

For those seeking theoretical and doctrinal simplicity and purity, pluralism is an
unfortunate state of affairs. But it is the only authentic explanation of the complexities
and tensions within Anglo-American tort law. And I see more reason to celebrate than
to regret that heterogeneity.
 14
Strict Liability Wrongs
Gregory C. Keating*

I. Introduction
American tort scholarship is split between two competing conceptions of tort liability.
One conception is economic; the other, for lack of a better word, is moral.1 The
economic conception models tort law on the market. It conceives of liability rules as
prices, of torts—especially accidents—as costs, and it understands tort law as an
institution that addresses a pervasive form of market failure. Torts take place mostly
among strangers. They are involuntary transactions among parties who do not have
bargaining or market relations with one another. The costs that torts inflict are
therefore externalities. The role of liability rules is to correct externalities by forcing
injurers to take their costs into account and adjust their behavior accordingly. The
larger role of the law of torts is to take both the costs of accidents and the costs of their
prevention into account, so that victims and injurers alike are induced to minimize the
combined costs of accidents and their prevention, thereby maximizing wealth.2

* I am grateful to audiences at Rutgers and the Toronto Legal Theory Workshop for comments and
criticisms. Particular thanks are owed to John Gardner, John Goldberg, Arthur Ripstein, Ken Simons, Rebecca
Stone, Victor Tadros and Ernest Weinrib for comments. I am indebted to Jacob Agi, Nicole Creamer, Kevin
Crow, and Misa Scharfen for research assistance.
1
The word “moral” is not ideal. The economic conception embodies a morality of sorts, namely, the morality
of putting resources to their highest use. This morality of minimizing waste and maximizing wealth can itself be
understood as the best way to promote human welfare. See Louis Kaplow and Steven Shavell, Fairness versus
Welfare (Cambridge, MA: Harvard University Press, 2002). To the extent that the economic conception warrants
the label “non-moral” it does so because it recasts the apparently moral judgments of tort law as purely
instrumental judgments of rational prudence. For example, it recasts “reasonable care” as rational care. Reason-
ableness is intrinsically moral. Rationality is not. See Gregory C. Keating, “Reasonableness and Rationality in
Negligence Theory,” 48 Stan. L. Rev. 311 (1996). For its part, the “moral” conception earns that label by taking the
moral concepts at the heart of tort law—reasonableness, right, wrong, obligation, duty, breach, harm, responsi-
bility, corrective justice and so on—at face value and seeking to account for tort law in those (moral) terms.
2
“Aside from the requirements of justice, I take it as axiomatic that the principal function of accident law is
to reduce the sum of the costs of accidents and their prevention.” Guido Calabresi, The Costs of Accidents (New
Haven, CT: Yale University Press, 1970), 26. Calabresi’s nod to questions of justice is often omitted by
economic theorists of tort, but his criterion is widely accepted as “axiomatic.”
 STRICT LIABILITY WRONGS 293

Whatever its faults, the economic conception nurtures a habitat in which strict
liability is at home. To be sure, from an economic point of view, strict liability is by no
means necessarily superior to negligence. By virtue of the Hand Formula’s apparent
embrace of cost-benefit analysis, negligence liability has the immediate attraction of
placing economic thinking at the center of accident law. But strict liability has an
equally powerful and immediate attraction: by building the costs of accidents into
liability rules, it makes liability rules conform more closely to market prices. And by
making injurers bear the costs of all the accidents attributable to their activities—not
just the accidents that should have been avoided—strict liability induces injurers not
only to exercise efficient care, but also to undertake their activities at efficient levels.
Whereas negligence liability modeled on an economic reading of the Hand Formula
replicates the cost-benefit thinking that is the hallmark of market rationality, strict
liability replicates the price system itself. Economic analysis thus makes a home for
both of the major forms of tort liability.
The moral conception, curiously, fares worse on this score. Prominent moral
theorists conceive of strict liability as a kind of “pay as you go” form of liability.3
“Paying as you go” is not a matter of making reparation for harm wrongly done; it is
merely a matter of bearing the costs of your activities. Paying as you go for the harm
that you do is, therefore, a matter of mimicking the market and putting resources—
including other people’s lives and property—to their highest uses. Strict liability is the
price system in liability form. So conceived, strict liability is not a matter of duty,
breach, right, wrong, responsibility, and obligation. If torts is a law of wrongs, strict
liability, so conceived, is a foreigner in its midst. Agents who inflict efficient injury on
others—who pay the proper price for the injuries that they inflict and walk away
richer—do no wrong. On the contrary, they do the right thing.
Moral theorists of torts are generally committed to the idea that torts are wrongs.
Moral conceptions of torts, therefore, take it as fixed that there are no justified torts.
When we label conduct tortious, we are saying that the conduct is unjustified and, in
that sense, wrong. Strict liability is a familiar form of liability in tort. One might,
therefore, expect moral theorists to resist the economic appropriation of strict liability
on the ground that it treats tortious conduct as justified conduct, and to offer an
alternative account. Instead, moral theorists have tended to embrace the idea that strict
liability as generally understood does not involve wronging. Moved by this perception
moral theorists have sought either to fit the round peg of strict liability into the square
hole of negligence liability, or to expel strict liability from the law of torts,4 Jules

3
Arthur Ripstein has used this expression at the Toronto workshop.
4
This is hardly the only view of strict liability taken by moral theorists. Ernest Weinrib, for example, takes
the view that strict liability is rejected by the law of torts, and rightly so, because it is incompatible with the
equal freedom of the parties. He writes “Under strict liability, the plaintiff ’s person and property are a
sacrosanct domain of autonomy, within which the plaintiff is entitled to freedom from interference by anyone
else. But strict liability protects the plaintiff ’s rights without allowing room for an intelligible conception of the
defendant’s duty. A duty must be operative at the time of the act that the duty is supposed to govern. Under
strict liability, however, the actor’s duty not to do the harm-causing act need not appear until the moment of
294 GREGORY C . KEATING

Coleman illustrates the first approach when he characterizes strict liability as the
imposition of a duty to harm, full stop. This conceptualization of strict liability
makes it a variation of negligence liability. Negligence liability imposes a duty not to
harm through the failure to exercise due care.5 John Goldberg and Ben Zipursky
illustrate the strategy of expulsion when they write:
We must acknowledge that our insistence that tort law is a law of wrongs puts us in a
difficult spot when it comes to explaining the presence of common law strict liability for
abnormally dangerous activities. After all . . . the rationale that seems to prevail in this
domain is that liability should attach to activities that are not wrongful in and of
themselves, and without regard to whether they are undertaken in a wrongful (i.e.,
careless) manner. How does liability imposed in these terms fit within a law that is
supposed to be all about defining wrongs and providing victims of wrongs with recourse?
The short answer is that it does not.6

This admirably candid answer encapsulates the predicament of most contemporary

moral theorists of tort. Strict liability is an embarrassment to their theories.
These responses should give us pause. Strict liability is an important form of tort
liability and perhaps especially so in the past half-century. Few opinions, if any, are
more important to twentieth century tort law than Roger Traynor’s dissent in Escola v
Coca Cola Bottling Co of Fresno.7 Traynor’s brief for the imposition of strict liability on
defective products is the chestnut from which }402A of the Second Restatement—and,
indeed, the might oak of modern products liability law itself—grew. Yet contemporary
moral theorists of tort are, at best, uncomfortable with the strict theory of responsi-
bility that it espouses. There is, moreover, an irony here. One of the principal—and
persuasive—arguments that moral theorists of tort have made for their view and
against the economic view is that only a moral conception can make sense of the
central, intrinsically moral, concepts of tort law, concepts such as duty, breach, right,
wrong, obligation, responsibility, and justice. Broadly speaking, the moral theorists of
torts have tasked their economic counterparts with being unable to explain the law
of torts from the internal point of view of the lawyer or citizen concerned with the
law of torts as a guide to action. From the internal point of view of these actors, the law
of torts is a law of obligations. When it comes to strict liability, however, most moral
theorists find themselves either confessing that they cannot make sense of it as a form
of tortious wrong or shoehorning it into the mold of negligence liability. Any account
of tort law that specifies the constitutive features of tortious wrongs in a way that

injury. Only retrospectively through the fortuity of harm does it then turn out that defendant’s act was a
wrong.” Ernest Weinrib, The Idea of Private Law (Oxford: Oxford University Press, 1995), 179. Professor
Weinrib’s view deserves more attention than I can give it here, but I believe that its force, such as it is, depends
on understanding strict liability in the essentially causal terms that Richard Epstein proposes. See, e.g., Richard
Epstein, “A Theory of Strict Liability,” 2 Journal of Legal Studies 151 (1973). Epstein’s conception does not, in
my view, track our law of torts.
5
See note 24, and accompanying text.
6
John C.P. Goldberg and Benjamin C. Zipursky, Torts (Oxford: Oxford University Press, 2010), 267.
7
150 P.2d 436 (1944).
 STRICT LIABILITY WRONGS 295

cannot acknowledge—or properly characterize—strict liability in tort is, for that

reason alone, seriously defective as an account of tort law from within its practice.
To be sure, this interpretive failure of the dominant moral conceptions of tort does
not loom so large at the moment. There is much less strict liability in the law of torts
than there was thirty-five years ago. In the late 1970s, strict products liability was both
strict and on the rise; the idea of enterprise liability was deeply rooted in the legal
academy and expanding into the law of automobile accidents; the environmental
movement seemed poised to expand both common law and statutory strict liabilities;
and negligence liability appeared to be in retreat throughout the law of torts.8 Now,
however, enterprise liability is on the wane, and the twenty-first century surprise has
been the “unexpected persistence of negligence liability.”9 Yet even if tort reform has
put strict enterprise liability on the defensive, it is a mistake to assert that strict liability
is a part of the law of torts in name only, as Goldberg and Zipursky do.10 For one thing,
enterprise liability figures prominently in the history of tort law. Just as negligence
liability only appeared to be dying during most of the twentieth century, it is probably
the case that strict liability only appears to be dying now. Some unforeseen historical
turn may well induce its revival.
More importantly, perhaps, two basic forms of strict liability are too deeply embed-
ded in our law of torts to be purged entirely from the field. One form of strict liability is
harm-based. The other form is based on an impermissible interference with an
autonomy right. The first, harm-based, form is found in corners of accident law and
in nuisance law. The second, autonomy rights-based, form is found mostly in inten-
tional wrongs to the person or to property. Different meanings of the slippery terms
“fault” and “strict” are relevant in these two contexts. Harm-based strict liabilities are
strict in that they are imposed on justified conduct. Rights-based strict liabilities are
strict in that they are imposed on innocent conduct, conduct that—from a moral point
of view—we would be inclined to excuse. Moral theorists of tort need to come to grips
with both of these forms. To distinguish their views from the economic view—and to
register the fact that torts always involve unjustified conduct—moral theorists need to
explain why strict liability torts are wrongs. My aim in this chapter is to do just that.

8
See G. Edward White, Tort Law in America (New York: Oxford University Press, 1980), 168–72
(describing the triumph of enterprise liability in early products liability law).
9
G. Edward White, “The Unexpected Persistence of Negligence, 1980–2000,” 54 Vand. L. Rev. 1337
(2001), 1344–6. It would be unwise, however, to dismiss entirely the continued presence of enterprise liability.
It is hard to see the tobacco litigation of the 1990s, for instance, as the expression of anything but enterprise
liability ideas.
10
Goldberg and Zipursky write, for instance, that “[a]lthough by convention, strict liability for abnormally
dangerous activities clearly is part of what lawyers define as ‘tort law,’ strictly speaking it does not belong in this
department”; Goldberg and Zipursky, Torts (note 5) at 267.
296 GREGORY C . KEATING

I shall argue that strict liability torts are wrongs because they involve violations of
rights and that they delineate two distinctive domains of wrongful conduct. One
domain—the territory of “harm-based” strict liabilities—involves the distinctive
wrong of harming-without-repairing. The other domain—the territory of “sovereignty
torts”—involves the distinctive wrong of violating fundamental powers of control over
one’s own person and property. These two forms of strict liability are unified by the
fact that they defend autonomy. Physical harm is an assault on one of the basic
conditions of human agency. And sovereignty torts involve the violation of core
autonomy rights.

II. Harm-Based Strict Liability

Harm-based strict liability is illustrated most clearly, perhaps, by the modern Ameri-
can law of intentional nuisance when it distinguishes between unreasonable conduct
and unreasonable harm—and imposes liability for the infliction of unreasonable
harm. Unreasonable conduct is conduct that inflicts injury unjustifiably. Negligent
conduct, or faulty conduct, is unreasonable conduct: it exposes others to a risk of
harm that ought not to have been imposed. By contrast, unreasonable harm is harm
that should not go unrepaired by the party responsible for its infliction, even though
that harm issued from conduct that was beyond reproach. Unreasonable harm is
harm for which an actor is strictly liable. Nuisance law imposes liability for the
infliction of unreasonable harm when, as in the famous case of Boomer v Atlantic
Cement, it holds that damages should be paid for an unreasonable interference with
plaintiffs’ rights to the reasonable use of their property, even though the conduct
responsible for that interference is justified and ought to be continued. By revising the
normal remedy for the wrong of nuisance in New York—from an injunction as a
matter of right to damages as a matter of right and injunctive relief only on a showing
of unreasonable conduct—Boomer revises the underlying right. Reasonable conduct
resulting in unreasonable interference with another’s use and enjoyment of land is
wrong only if the party inflicting the interference fails to make reparation for the
harm that he or she does. Reparation transforms unreasonable harm into reasonable
harm and fairly reconciles competing, equal rights to the use and enjoyment of land.
Before Boomer—when injunctive relief was available as a matter of right when
plaintiff could show that defendant’s activity unreasonably interfered with plaintiff ’s
right to the reasonable use and enjoyment of its land—the primary duty in this corner
of nuisance law was a duty to do no harm, full stop. After Boomer—when money
damages, and only money damages are available as a matter of right—the primary duty
in this corner of the law is now a duty not to inflict reasonable harm without repairing
that harm. Other entrenched instantiations of this primary duty include: private
necessity cases such as Vincent v Lake Erie; liability for abnormally dangerous activities;
 STRICT LIABILITY WRONGS 297

liability for manufacturing defects in products liability law; and the liability of masters
for the torts committed by their servants within the scope of their employment.
The obligation imposed by all of these doctrines is an obligation not to harm without
repairing even if, there is no fault in the infliction of the harm itself. The reciprocal right
is a right to have any physical harm done to you undone by the party responsible for its
infliction. All of these harm-based liabilities are strict in that they impose liability on
conduct that is not wrong. Stated affirmatively, these strict liabilities impose liability on
conduct that is justified, on inflictions of harm that are reasonable. These liabilities do
not condemn the infliction of harm, per se. They condemn harming without repairing.
Their target is not unreasonable conduct, but unreasonable harm. Negligence liability,
of course, predicates liability on conduct that is unjustified, on conduct that is
unreasonable—because it does not show due regard for the property and physical
integrity of those that it harms.

III. Right-Based Strict Liability

Autonomy-right based strict liabilities impose liability not on justified inflictions of
harm, but on boundary crossings that may both do no harm and be entirely free of
fault.11 This form of strict liability is epitomized by the torts of conversion and trespass
and by some batteries. Here, the wrong is the violation of a right that assigns a power
of control over some physical object or, in the case of battery, control over some
subject. The law’s specification of these powers of control gives rise to a form of strict
liability predicted on the voluntary, but impermissible, violation of that right. If you
enter my land or appropriate my pen without my permission, you have violated my
right of exclusive control over these objects, even if your entry is entirely reasonable
and justified. The wrong consists in the failure to respect the right. Fault is simply
irrelevant. If you enter my property under an entirely reasonable and innocent
misapprehension of where the boundary between my property and yours lies, you
trespass. You need not even know that you have entered my property without my
permission, much less intend the wrong of entering my property without permission.
You need intend no wrong and you need do no harm other than the harm of violating
my right of control over my person or my property. Indeed, you may benefit me—say,
by trimming, topping, and cleaning my trees of bagworms.12 So, too, you may trespass
if you are innocently and reasonably mistaken about which building you have per-
mission to use as a set for your movie,13 or if you are simply mistaken about the scope
of the permission that I have given you.14 And you may convert my chattel simply by

11
Two useful terms for these torts are “sovereignty-based” and “trespassory.” The former calls attention to
the connection between these torts and autonomy. The latter calls attention to the fact that these torts involve
the impermissible crossing of boundaries, not the infliction of injury.
12
Longenecker v Zimmerman, 175 Kan. 719, 267 P.2d 543 (1954).
13
Bigelow v RKO Radio Pictures, 327 U.S. 251, 66 S. Ct. 574, 90 L. Ed. 652 (1946).
14
Cleveland Park Club v Perry, 165 A.2d 485, 488 (Mun.App.D.C.1960).
298 GREGORY C . KEATING

exercising dominion over it in a way which is “in denial of or inconsistent with [my]
rights therein.” You do not need to intend to deny my rights, or even know that
you are.15 Liability for battery may, likewise, be predicated on innocent intentional
touchings,16 and even on touchings that benefit those who are touched without their
consent.17
There are two lessons here. One is that intentional torts do not categorically represent
a more egregious form of fault than negligent torts. Paradigmatic intentional wrongs, to
be sure, do have this character. They are governed by the aim of inflicting harm on the
victim in violation of her rights. The wrongdoing involved is worse than negligent
wrongdoing. Negligence involves only objectively insufficient regard for the victim.
Paradigmatic intentional wrongs involve both intended harm and knowing disregard of
the victim’s rights. They express “contempt for [the victim], often more unbearable
than the harm itself.”18 The law of intentional torts, however, is not restricted to wrongs
whose commission entails the expression of an intense and objectionable disregard for
the rights of their victims. Batteries, trespasses, and conversions can all be committed
without intending either the wrongs or the harms involved. The intent need reach only
as far as the action which impermissibly crosses a protected boundary and no further.
This is because the rights in question are autonomy rights. They are rights to exclusive
control over one’s person and one’s property, real and moveable. Liability for violation
of a right of exclusive control is strict for the simple reason that the right itself would be
fatally compromised by tolerating all reasonable (or justified) boundary crossings
without regard to whether consent was given to those crossings.19 Those who hold
the relevant rights are entitled to forbid even reasonable boundary crossings, and they
are presumptively wronged whenever their boundaries are crossed without their
permission by the intentional actions of others. The rights in question thus give rise
to stringent “duties to succeed” on the part of others.20 In this class of cases, the
strictness of liability in tort is a reflex of rights of control.

15
Zaslow v Kroenert, 29 Cal. 2d 541, 176 P.2d 1 (1946).
16
E.g., Vosburg v Putney, 80 Wis. 523, 50 N.W. 403 (1891); White v University of Idaho, 118 Idaho 400, 797
P.2d 108 (1990).
17
E.g., Mohr v Williams, 95 Minn. 261, 104 N.W. 12 (1905). The touching exceeded the scope of the
consent given.
18
Jean-Jacques Rousseau, The Social Contract and the Discourses (trans. G.D.H. Cole, New York: Dutton,
1979), 82. The quotation is from the Discourse on the Origin of Inequality. I have altered the translation slightly.
19
“ . . . [S]trict liability for trespass—to the person or to property—is morally demanded. With respect to
battery, for instance, we surely cannot adopt the view that people are at liberty to touch each other without
consent, as long as that touching be not angry, hostile, unordinary or even unreasonable. Why should one have
to put up with being intentionally touched just because that form of touching is ordinary or thought reasonable
by a judge, for instance? As Cardozo . . . said, ‘Every human being of adult years and sound mind has a right to
determine what shall be done with his own body.’ ” Allan Beever, “The Form of Liability in the Torts of
Trespass,” 40 Comm. L. World Rev. 378 (2011), 392 (citing Schloendorff v Society of New York Hospital, 211
N.Y. 125,105 N.E. 92, 93 (1914)).
20
The concept of “duties to succeed” is developed in John Gardner, “Obligations and Outcomes in the Law
of Torts,” in Peter Cane and John Gardner (eds.), Relating to Responsibility: Essays in Honour of Tony Honoré
on his 80th Birthday (Oxford: Hart Publishing, 2001).
 STRICT LIABILITY WRONGS 299

The difficult normative questions raised by these torts are mostly questions of
responsibility. Why, exactly, is it permissible to hold someone accountable for inno-
cent conduct that they could not have been reasonably expected to avoid? For the
moment, I want to defer consideration of this question and comment on the relation
of these torts to what moral theorists of torts call conduct-based wrongs. In cases
where liability is strict and sovereignty-based, the wrong committed is conduct-based
in only the most attenuated sense of the term. In negligence—the canonical example of
a conduct-based wrong—liability is predicated on the wrongfulness of the defendant’s
conduct; the wrongfulness of the conduct (the breach of duty) does the work and
triggers liability. In “sovereignty” or “trespassory” torts, it is the violation of the
plaintiff ’s right that does the work and triggers liability. The “duty” is a duty not to
violate the right, and any intentional action that violates the right is therefore wrong-
ful. Viewed in isolation from the right, the conduct may be innocent and even justified.
The defendant doctor in Mohr v Williams, for example, benefited the plaintiff by
curing her disease.21 The wrongs committed in cases like Mohr are wrongs only
because they violate rights. They do harm and express disrespect only insofar as
they disregard important autonomy rights.
Derek Parfit’s distinction among belief-relative, evidence-relative, and fact-relative
standpoints may help to isolate the limited sense in which the conduct at issue is
wrong.22 Innocent batteries, trespasses, and conversions are wrongs relative to the
facts (what is actually the case) because they violate rights that their victims actually
possess. They are not wrongs relative to the beliefs of the persons who commit them
because the persons who commit them believe themselves to be acting with due
regard for everyone’s rights. Nor are innocent batteries, trespasses, and conversions
wrongs relative to the evidence available to those who commit them. These tortfeasors
reasonably believe that they have consent, are on their own property, or own the
chattel in question. Negligence wrongs, by contrast, are generally wrong relative to the
evidence. Those who commit them should have known that they were not exercising
reasonable care. We are most at fault (most blameworthy) when we commit wrongs
that we believe to be wrong, and least at fault when we commit wrongs that, in fact, are
wrongs, even though we did not believe them to be wrongs and should not, in fact,
have determined that they were wrongs on the evidence available to us. The paradox of
strict liability in intentional torts surprises us by showing that a class of wrongs that we
thought consisted exclusively of belief-relative wrongs—and thus of wrongs involving
the highest degree of fault in Parfit’s pecking order—in fact includes prominent
examples of wrongs that are merely fact-relative wrongs.
Liability without fault in the standard negligence sense—that is, liability in circum-
stances where the wrongdoer’s conduct is not wrong relative to the reasonable
appraisal of the relevant evidence—raises difficult questions of responsibility.
“Ought” implies “can” and responsibility that is relative only to the facts as they

21
Mohr v Williams, 95 Minn. 261, 104 N.W. 12 (1905).
22
Derek Parfit, On What Matters (New York: Oxford University Press, 2011), 150–3.
300 GREGORY C . KEATING

are—and not to the evidence as it appears to be—is therefore troubling. For our
purposes, however, the important point is that trespassory wrongs are a well-estab-
lished and robust domain of tort law. Contrary to conventional wisdom, trespassory
wrongs need not be conduct-based wrongs in anything but the most attenuated and
misleading sense. The only respect in which the conduct involved in a trespassory tort
must be wrong is that it violates a right. Canonical conduct-based wrongs, as correct-
ive justice and civil recourse theorists conceive them, are wrongs where the wrong-
doers’ conduct evinces a failure to appraise the pertinent evidence in an objectively
justifiable way and act accordingly. Conversely, some canonical trespassory wrongs are
strict liability wrongs in the sense that they impose liability on morally blameless
conduct. These intentional torts are, therefore, important counter-examples to both
the claim that tort is a law of fault-based wrongs and to the claim that tort is a law of
conduct-based wrongs. These wrongs, moreover, are deeply entrenched in our law.
We can live without these wrongs only if we are prepared to compromise the
autonomy rights that they defend.
Harm-based strict liabilities of the sort found in intentional nuisance law, abnor-
mally dangerous activity law, and some parts of products liability law are different.
These strict liabilities embody a more robust form of wrongfulness. These are liabilities
where the wrongfulness of the defendant’s conduct lies in harming-without-repairing.
The distinctive character and morality of these strict liabilities is also obscured by
contemporary tort theory, with its preference for fault liability and its insistence that
tort is a domain of conduct-based wrongs. Indeed, contemporary moral theorists of
tort show either the inclination to expel these torts from the law of torts proper or to
absorb them into the architecture of negligence wrongs. Goldberg and Zipursky
exhibit the first inclination, claiming that it is merely a matter of convention that we
call this kind of liability a tort. When harm issues from an abnormally dangerous
activity, and strict liability is imposed on the party engaging in that activity, neither the
activity in which the injurer is engaged nor their conduct in inflicting the injury are
wrong.23 Tort, properly understood, is a law of wrongs. Harm-based strict liabilities do
not involve wrongs and are therefore not really torts. Jules Coleman assimilates this
form of strict liability to negligence, by casting the duty involved as a duty to do no
harm, full stop.24 Neither Goldberg and Zipursky’s expulsion of harm-based strict
liability from tort law proper nor Coleman’s absorption of such liability into negli-
gence does justice to this form of liability. The wrong involved is a conditional one,
and its morality is a matter of both commutative and corrective justice.

23
See notes 5 & 7 and accompanying text.
24
Jules Coleman, “Facts, Fictions, and the Grounds of Law,” in Campbell (ed.), Law and Social Justice
(Cambridge, MA: MIT Press, 2005), 329; see also Coleman, Practice of Principle (New York: Oxford University
Press, 2001), 35, n. 19 (“The concept of a duty in tort law is central both to strict and fault liability. In strict
liability, the generic form of the duty is a ‘duty not to harm someone,’ while in fault, the generic form of a duty
is a ‘duty not to harm someone negligently or carelessly’. ”).
 STRICT LIABILITY WRONGS 301

A. Strict liability as a conditional wrong

Harm-based strict liabilities occupy a space that negligence liability tends to eclipse. In
negligence, the obligation to repair arises from the infliction of harm that should have
been avoided. In strict liability, the obligation to repair arises from the infliction of
harm that should not have been avoided, but which should be borne by the party
responsible for its infliction. In the circumstances where they govern, harm-based
strict liabilities assert that it is wrong for an actor to do harm without stepping forward
and making good the harm done. The primary duty that harm-based strict liability
institutes is not a duty not to harm; it is a duty to harm only through reasonable,
justified conduct, and to make reparation for any harm done even though due care has
been exercised. Failing to make reparation evinces insufficient regard for the rights of
the person harmed, even though the conduct responsible for inflicting that harm is
beyond reproach.
A conduct-based wrong is one where a right is violated when an agent fails to
conform her conduct to the standard required by the law. Fault liability is a canonical
illustration; it predicates responsibility for physical injury on the judgment that the
defendant failed to conform her conduct to the standard of reasonable care.25 Con-
duct-based wrongs express what I shall call primary criticism of conduct.26 The law
lodges its criticism against the infliction of harm in the first instance on the ground
that the conduct responsible for the harm was wrong. The harm, therefore, should
never have occurred. Strict liability, by contrast, predicates responsibility on the
judgment that the conduct at issue was justified (or reasonable) in inflicting injury,
but unjustified (or unreasonable) in failing to repair the injury done. This is secondary
criticism of conduct. The law lodges its criticism against harming-justifiably-without-
repairing. This kind of strict liability identifies a conditional wrong. It circumscribes a
domain within which the infliction of harm is justifiable, but only on two conditions:
(1) the conduct inflicting injury is justified or reasonable; and (2) reparation is made
for physical harm done by that reasonable conduct.
Conditional privilege in the law of private necessity—the doctrine of Vincent v Lake
Erie27—illustrates the distinction between primary and secondary criticisms of con-
duct. The defendant ship owner’s conduct in lashing its ship to, and damaging, the
plaintiff ’s dock was reasonable, not unreasonable—right, not wrong. Counting the
interests of the two parties as equal, the dock owner’s rights in its property must yield
to avoid greater harm to the ship owner. The defendant was therefore privileged to use

25
The fundamental question in negligence law is whether conduct falls below “a standard established by the
law for the protection of others against unreasonable harm.” Negligence law fixes that standard by the conduct
of a “reasonable person in the circumstances.” Restatement (Second) of Torts } 283 (1975); see also, e.g., Ladd v
County of San Mateo, 12 Cal. 4th 913, 917, 911 P.2d 496 (1996).
26
I owe this term to Lewis Sargentich. Robert Keeton’s contrast between “fault” and “conditional fault” also
describes the distinction drawn in the text. See Robert E. Keeton, “Conditional Fault in the Law of Torts,” 72
Harv. L. Rev. 401 (1959).
27
Vincent v Lake Erie Transp. Co., 109 Minn. 456, 124 N.W. 221 (1910).
302 GREGORY C . KEATING

the dock to save its ship from destruction at the hands of the storm, even if using the
dock involved damaging the dock. The defendant’s privilege28 to trespass was not
conditioned on doing no harm to the dock, a requirement that would have been
impossible to meet in the circumstances. The defendant’s privilege was conditioned on
making reparation for any harm done to the dock, even though that harm was done
rightly and not wrongly. The wrong in Vincent lay not in the defendant’s doing damage
to the dock, but in the defendant’s wrongful (or unreasonable) failure to step forward
and volunteer in the aftermath of the storm to make good the damage done the dock.
In refusing to repair the dock, the ship owner was not reasonably inflicting lesser harm
to avoid greater harm; he was simply benefitting himself by harming another. The
infliction of harm was therefore justified; the failure to repair was not. Put differently,
Vincent’s strict liability is liability for unreasonable harm, not liability for unreasonable
conduct. In Vincent, making reparation for the harm done by docking prevents the
injustice of shifting the cost of the ship’s salvation from the ship owner, who profits
from it, onto the dock owner, who does not. The imposition of liability on the ship
owner, for failing to make such reparation, rights the wrong of shifting the cost of the
ship’s salvation onto the dock owner, whose property was the instrument of that
salvation. The wrong in strict liability is thus “harming justifiably but unjustifiably
failing to repair the harm justifiably done.”29 Generalizing from Vincent, we can say
that the wrong involved in harm-based strict liabilities is both akin to the wrong in
restitution and subtly different. In restitution cases, the basic wrong consists of
retaining a benefit whose retention unjustly enriches its recipient at the expense of
the party conferring the benefit. In harm-based strict liabilities the basic wrong
consists of benefitting from the reasonable infliction of harm on another. This wrong
has two aspects. On the one hand, the victim is harmed—a means which is supposed to
be subject to their discretionary control is damaged and their capacity to exercise their
will effectively is thereby impaired. On the other hand, unless reparation is made the
victim is left in an impaired condition simply so that her injurer may benefit—the

28
Taxonomically, this is a complicated matter. In Hohfeldian terms, the ship’s privilege to enter is a right:
the ship is entitled to enter, and the dock owner is under a duty not to resist. See Francis H. Bohlen,
“Incomplete Privilege to Inflict Intentional Invasions of Interests in of Property and Personality,” 39 Harv.
L. Rev. 307 (1926). This privilege is also a power in Hohfeld’s terms, because it enables the ship owner to alter
its relations with the dock owner without the dock owner’s permission, as long as the ship enters the dock
owner’s property for certain purposes (to save its own property), and conducts itself in certain ways (only does
what is necessary to save its own property). Along with Robert E. Keeton, “Conditional Fault in the Law of
Torts,” 72 Harv. L. Rev. 401 (1959), Bohlen’s article is a classic statement of the idea of strict liability I am
developing in this chapter. Similar positions have also been reached by others. See, e.g., Howard Klepper,
“Torts of Necessity: A Moral Theory of Compensation,” 9 L. & Phil. 223, 239 (1990) (“The need to compensate
in the necessity cases is best explained by the wrongfulness of knowingly benefitting oneself by transferring a
loss to another, however reasonably, and then letting the loss lie with one’s unwitting benefactor. Such a
transfer of the loss or risk is wrongful in that it does not allow the innocent party to freely choose the risks she is
willing to undertake.”).
29
Vincent is thus a clear counter-example to the claims of some prominent tort scholars that strict liability
involves a duty not to do harm, full stop. Jules Coleman and John Gardner hold views of this kind. See note 22;
see also Gardner, “Obligations and Outcomes in the Law of Torts” (note 18) at 111.
 STRICT LIABILITY WRONGS 303

injurer benefits through harming the victim. The role of reparation is to undo that
wrong by erasing the harm. When reparation is made the injurer no longer benefits
through harming the victim. The flip side of this coin is that the victim is no longer
forced to suffer harm merely so that the injurer may benefit. Structurally, harm-based
strict liability in tort resembles eminent domain in public law. Eminent domain law
holds that it is permissible for the government to take property for public use only if
the government pays just compensation to those whose property it takes. This is a two-
part criterion. First, the taking must be justified; that is, it must be for a public use.
Second, compensation must be paid for the property taken. Strict liability in tort has a
parallel structure.30 In negligence, the defendant’s primary conduct determines liabil-
ity, and it does so only when that conduct is wrongful. In strict liability, the defendant’s
conduct triggers liability when the defendant’s failure to step forward and repair the
harm faultlessly inflicted is wrongful. Strict liability asserts that the costs of necessary
or justified harms should be borne by those who benefit from their infliction, and not
by those whose misfortune it is to find themselves in the path of someone else’s pursuit
of his or her own benefit, however reasonable that pursuit may be.
Harm-based strict liability thus involves both fairness or justice, and wrong or rights
violation. To say that it is unfair for an injurer to thrust the cost of its activities onto a
victim is not the same as saying that the victim’s right is violated by so doing. It may,
for example, be unfair for me to rebuild my house and block the passage of air through
your chimney. The loss to you may be great and the gain to me may be trivial. Unless
you have a right to that passage, however, what I have done is not a legal wrong.31
Strict liability is thus justified both by the principle of fairness that those who benefit
from inflicting harm on others should also shoulder the cost of that harm and by the
further claim that the harm done is the invasion of a right so that failure to make
reparation for harm done would be a wrong. Strict liability asserts that an injurer
subject to a regime of strict liability does wrong when the injurer fails to step forward
and repair harm rightly inflicted, and it makes this assertion because leaving the cost of
the harm on the victim who suffers it shows insufficient respect for the victim’s
rights—rights of property in the case of both nuisance and the conditional privilege
of private necessity.
Vincent is once again illustrative. Not only would it be unfair for the ship owner to
shift the cost of saving its ship off onto the dock owner, but it would also violate the
dock owner’s property rights. The dock owner’s right to exclude the ship must yield

30
This “private eminent domain” conception of strict liability may make its first appearance in American
tort theory in the writings (some famous and some obscure) of Oliver Wendell Holmes. These writings are
cited and discussed in Thomas C. Grey, “Accidental Torts,” 54 Vand. L. Rev. 1225 (2001), 1275–81, and at
greater length in his unpublished manuscript, Holmes on Torts (on file with author). Two other classic
statements are Francis Bohlen, “Incomplete Privilege to Inflict Intentional Invasions of Interests of Property
and Personality,” 39 Harv. L. Rev. 307 (1926) and Robert E. Keeton, “Conditional Fault in the Law of Torts,” 72
Harv. L. Rev. 401 (1959).
31
Bryant v Lefever, 4 C.P.D. 172 (1878–79).
304 GREGORY C . KEATING

to the dire emergency—the “necessity”—in which the ship found itself.32 But there is
no reason why the dock owner’s right to the integrity of his property should also be
extinguished. Saving the ship requires damaging the dock, but it does not require that
the cost of saving the ship be shifted onto the owner of the dock instead of being borne
by the ship owner who profits from doing that damage. Harm-based strict liabilities
thus define a particular class of conditional wrongs where the law lodges its criticism
against a defendant’s secondary failure to repair, not against a defendant’s primary,
injury inflicting conduct.33

B. Correcting corrective justice theory

Corrective-justice theory and civil recourse theory both conceive of tort as a domain of
conduct-based wrongs.34 This insistence on wrongful primary conduct as essential to
tort law prevents corrective-justice theory and civil recourse theory from giving an
adequate account of strict liability in tort.35 Strict liability is not predicated on the
assertion that the defendant should have behaved differently and not harmed the
plaintiff.36 Strict liability, moreover, is common enough that we can reasonably insist
that an adequate theory of tort should be able to explain and justify its existence as an
alternative to negligence. Contemporary moral theories of tort flunk this test. Jules
Coleman’s sophisticated and powerful version of corrective-justice theory, for
example, fails to give an adequate account of strict liability for two reasons. First,
Coleman conceives of substantial portions of strict liability as lying outside the core of
tort that corrective justice succeeds in explaining. He notes that corrective-justice
theory “does not explain” various features of tort law, “for example, vicarious liability
or perhaps product liability.”37 Coleman also excludes products liability from the core
of tort and tentatively suggests that it should be understood not in terms of corrective

32
“The situation was one in which the ordinary rules regulating property rights were suspended by forces
beyond human control . . . ” Vincent (note 25), 124 N.W. at 221.
33
It is possible to construe the concept of a conduct-based wrong in a way which obliterates the distinction
between primary and secondary criticism of conduct. Asserting, say, that any conduct which violates a right is
wrongful conduct obliterates the distinction. Some tort scholars, including, perhaps, Ernest Weinrib and
Arthur Ripstein, may hew to such a conception. See Ernest Weinrib, The Idea of Private Law (Oxford: Oxford
University Press, 1995). The fundamental reason to reject this understanding of “conduct-based wrong” is that
obliterating the distinction between primary and secondary criticism of conduct impairs our ability to
understand strict liability in tort. We want categories which enable comprehension instead of frustrating it.
34
For corrective justice theory, see Jules Coleman, “The Practice of Corrective Justice,” in David G. Owen
(ed.), Philosophical Foundations of Tort Law (Oxford: Oxford University Press, 1995), 56–7 and Weinrib, The
Idea of Private Law (note 31) at 140–2, 197–8. For civil recourse theory, see Goldberg and Zipursky (notes 5 & 7).
35
This is ironic, because strict liability “duties” are the only primary duties that might be plausibly
described as corrective; they involve obligations not to harm without repairing.
36
The identification of tort with conduct-based wrongs is not particular to Coleman or Goldberg and
Zipursky. Ernest Weinrib holds the same kind of view, a fact vividly illustrated by his criticisms of strict liability
as a norm of conduct that condemns “any penetration of the plaintiff ’s space.” Ernest Weinrib, The Idea of
Private Law (note 31) at 177.
37
Coleman, Practice of Principle (note 22) at 36.
 STRICT LIABILITY WRONGS 305

justice, but in terms of rational bargaining.38 Both the concession and the exclusion are
troubling. A theory of tort that can explain its domains of strict liability is interpret-
ively superior to a theory that cannot. Coleman’s theory is vulnerable precisely because
it cannot. The absorption of product liability into tort, moreover, is the most import-
ant development in twentieth-century tort law. An adequate theory of tort ought to be
able to account for that development.
Second, Coleman’s account of the nature of strict liability goes wrong because it
models strict liability on negligence. According to Coleman, negligence liability imposes
a duty to exercise reasonable care to avoid inflicting physical harm on others and strict
liability imposes a duty not to inflict physical harm full stop. This turns harm-based
strict liability upside down. Harm-based strict liability applies to harms that should not
have been avoided. It does not criticize the infliction of harm; it criticizes the failure to
repair harm reasonably inflicted. Harm-based strict liability asserts that the costs of
justified harms should be borne by those who benefit from their infliction, not by those
whose misfortune it was to find themselves in the path of a justified imposition of harm.
Recall, too, that a theory of strict liability in tort must make room for trespassory, or
sovereignty-based, strict liabilities. The wrong committed in a sovereignty-based tort is
conduct-based in only the most attenuated sense of the term. It is the violation of the
plaintiff ’s right that does the work and triggers liability. The duty is a duty not to
violate the right, and conduct that violates the right is wrongful only because it violates
the right. The conduct may be otherwise blameless, reasonable, and even beneficial.
Sovereignty-based torts, moreover, are clear counterexamples to Coleman’s particular
thesis that tort is a law of wrongful losses. Trespassory torts may be committed without
inflicting wrongful loss because the rights at issue are powers of control, not protec-
tions against harm. The distinguishing features of both harm-based and autonomy
right-based strict liabilities are obscured and distorted by calling strict liability torts
“conduct-based wrongs.” Strict liability wrongs are based not on wrongful conduct in
the normal sense of that term, but on either secondary failings of conduct—on
conditional, not primary fault39—or on violating certain powers of control.

C. The morality of strict liability: Corrective and

commutative justice
So far, this chapter has distinguished harm-based strict liability wrongs from negli-
gence-based ones by asserting that strict liability wrongs involve justified conduct
inflicting harm, whereas negligent torts do not. And it has argued that strict liability
wrongs involve both a wrong, that is, the violation of a right, and unfairness in failing
to bear the cost of repairing harm reasonably inflicted. But the chapter has not said
just what the legal wrong is or when it occurs. The wrong, in brief, lies in failing to

38
Coleman, Risks and Wrongs (Oxford: Oxford University Press, 2002), 417–29.
39
See Keeton, “Conditional Fault in the Law of Torts” (note 24).
306 GREGORY C . KEATING

repair the harm when the actor’s responsibility for that harm becomes clear, and the
obligation to repair is part of the primary norm governing the conduct in question. To
put the matter in corrective justice terms: the wrongful interaction is the interaction
that takes place when the injurer fails to step forward and repair harm for which the
injurer is all things considered responsible because the harm may fairly be charged to
the injurer’s activity.40 The moral wrong involved in harm-based strict liabilities lies in
benefitting from the reasonable infliction of harm on another in violation of her right
and at her expense. The legal wrong is committed when a reasonable injurer would
have recognized and acted upon its responsibility to repair. The lawsuit therefore
enforces a preexisting duty of repair just as a negligence suit does, albeit a different pre-
existing duty of repair.
This account of the morality of harm-based strict liability wrongs means that strict
liability is corrective in a fundamental way: it undoes interactions that involve one
person who benefits herself by reasonably harming another person. In this sense, there
is no logical incompatibility between strict liability and corrective justice. Quite the
opposite in fact: strict liability in tort embodies corrective justice. However, it is equally
important to see that strict liability also goes beyond corrective justice because it rests
on a principle of fairness as well as on a conception of wrongdoing. To see this point, it
helps to consider, very briefly, harm-based strict liability in full flower. Harm-based
strict liabilities blossom most fully in enterprise liability, either within or beyond tort.41
Enterprise liability applies to activities as opposed to individual actions. In the law of
vicarious liability—its most important common law application—enterprise liability
applies to firms, through the risks created by their employees in the course of pursuing
the firm’s business. In its most important administrative application, worker’s com-
pensation, enterprise liability applies to the activities of firms insofar as those activities
occasion harm to their employees. In a nutshell, enterprise liability asserts that the
costs of accidents should: (a) be imposed on the enterprises responsible for their
infliction and (b) be dispersed among all those within the enterprise—that is, all
those who benefit from risk impositions that result in enterprise related harm. Within
tort, strict enterprise liability42 is found primarily in vicarious liability, abnormally
dangerous activity liability, and pockets of products liability. While it has surely
retreated, it is a basic form of modern tort liability and it flourished throughout the

40
Whether a particular harm may be fairly charged to an injurer’s activity is often at the heart of
strict liability litigation. See, e.g., Ira S. Bushey & Sons, Inc. v United States, 398 F.2d 167 (2nd Cir. 1968)
(Friendly, J.).
41
The fairness case for enterprise liability beyond tort is powerfully made in Jeremy Waldron, “Moments of
Carelessness and Massive Loss,” in David G. Owen (ed.), Philosophical Foundations of Tort Law (Oxford:
Oxford University Press, 1995), 387.
42
The proposition that the vicarious liability of a master for the torts of its servants is strict is a contestable
one (see Weinrib, The Idea of Private Law (note 31) at 185–7). There is little doubt, however, that it is the
dominant understanding of the doctrine in American law, even by those who are not proponents of strict
liability. See, e.g., Konradi v United States, 919 F.2d 1207, 1210 (7th Cir. 1990) (“The liability an employer for
torts committed by its employees—without any fault on his part—when they are acting within the scope of
their employment, the liability that the law calls ‘respondeat superior,’ is a form of strict liability.”).
 STRICT LIABILITY WRONGS 307

twentieth century. The case for it is most commonly made on economic grounds, but
the rhetoric of enterprise liability draws heavily on fairness.
The fairness case for enterprise liability is epitomized by saying that it distributes the
costs of accidents across those who benefit from the underlying risks. This slogan can
be unpacked into three components. The first of these is fairness to victims. It is unfair
to concentrate the costs of characteristic risk on those who simply happen to suffer
injury at the hands of such risk when those costs might be absorbed by those who
impose the characteristic risk. Fairness prescribes proportionality of burden and
benefit. Victims who are strangers to the enterprise derive no benefit from it; therefore,
it is unfair to ask them to bear a substantial loss when that loss might be dispersed
across those who participate in the enterprise and, therefore, do benefit from it.
Victims who are themselves participants in an enterprise share in its benefits, but
not in proportion to the detriment that they suffer when they are physically harmed by
the enterprise. Here, enterprise liability is fairer than negligence. It disperses the costs
of enterprise-related accidents and distributes them within the enterprise, so that each
bears a proportionate share.
Second, enterprise liability is fair to injurers because it simply asks them to accept
the costs of their choices. Those who create characteristic risks do so for their own
advantage, fully expecting to reap the benefits that accrue from imposing those risks. If
those who impose characteristic risks choose wisely—if they put others at risk only
when they stand to gain more than those that they put in peril stand to lose—even
under enterprise liability, they will normally benefit from the characteristic risks that
they impose. If they do not, they have only their poor judgment to blame and society,
as a whole, has reason to penalize their choices. Consider the facts of Ira S. Bushey and
Sons v United States43 where a drunken sailor, returning from shore leave, flooded the
dry-dock in which his ship was berthed by spinning the wheels that controlled the
dock’s valves. In upholding the imposition of liability, Judge Friendly asserted, in a
nutshell, that the Coast Guard lets its sailors loose on shore leave for its own benefit (as
well as for its sailors) and it reaped the rewards of their shore leave. It, therefore, had to
take the bitter with the sweet. If the costs of shore leave are greater than the benefits,
the Coast Guard has reason to reconsider the practice, and society has reason to
discourage it.
To be sure, if vicarious liability is ensconced in our legal system, enterprise liability
is embattled and it has receded in our law, rightly or wrongly. The merits of enterprise
liability, however, are not our concern. For our purposes, the point is that the idea of
fairness at work in harm-based strict liabilities flowers in enterprise liability and that
flowering sheds light on the idea itself. For our purposes, the lesson is that strict
liability embodies commutative as well as corrective justice. Corrective justice is
transaction centered; it is, as Aristotle said, “justice as rectification.” Corrective justice
is justice between parties to a wrong; it rights wrongs. Harm-based strict liabilities are

43
Bushey (note 39), 398 F.2d at 167. Compare id. and Taber v Maine, 45 F.3d 598 (2d Cir. 1995)
(Calabresi, J.).
308 GREGORY C . KEATING

corrective insofar as they undo wrongs whose essence lies in benefitting at the expense
of another person by harming that other person. But harm-based strict liabilities are
also commutative; they involve the proportional alignment of burden and benefit
across a plurality of persons. Strict liability, in the sense relevant here, imposes a
condition on conduct and thereby seeks not just to correct harm wrongly done, but to
prescribe the proper alignment of burden and benefit as far as the activities that it
governs are concerned.
Commutative justice is concerned with the proportional alignment of burdens and
benefits. It is, for Aristotle, distinct from both corrective justice (which is concerned
with the rectification of wrongs) and distributive justice (which distributes goods on
the basis of status or virtue).44 Commutative justice distributes in accordance with
benefit. Enterprise liability is fundamentally commutative because it holds that acci-
dental losses should be borne in accordance with the benefit that people derive from
the enterprise or activity in question. Like corrective justice—and unlike distributive
justice as Aristotle understands it—commutative justice is concerned with conduct,
with what has been done, with who has suffered, and with who has gained. Unlike
corrective justice, however, commutative justice does not just seek to right a wrong. It
seeks to bring into existence a state of the world in which burdens and benefits are
aligned. Harm-based strict liabilities attempt to do just that. They are primary con-
duct-governing norms of conduct only insofar as they prescribe that those who inflict
harm reasonably on others must repair the harm that they reasonably inflict in order
to align burden and benefit fairly.
In their most distinctive, if embattled, incarnation, strict liability wrongs are thus
distinctive in two ways. First, they are conditional wrongs. The wrong involved is
“harming without repairing.” Second, they involve both corrective and commutative
justice. They involve corrective justice because they right the wrong of harming
another and thereby benefitting unjustly at that person’s expense. They involve
commutative justice because they seek to prescribe a social world in which the burdens
and benefits of reasonable risk impositions are fairly distributed. The basic lesson of
autonomy rights-based forms of strict liability reinforces the lessons of harm-based
strict liabilities in an important way. The wrong in trespassory forms of strict liability
lies only in the invasion of the right. The conduct involved may be otherwise entirely
innocent, entirely justified, and even beneficial. Strict liabilities are not conduct-based
wrongs in the sense that corrective justice and civil recourse theorists use that term.

D. Defenses and limits

My principal concern in this chapter is with strict liability as the form of liability that
attaches to defendants in strict liability cases. Strict liability regimes, of course, involve

44
Aristotle, Nicomachean Ethics, bk. V, ch. 5. The appropriation of the term commutative justice to defend
forms of nonfault liability is illustrated by Joel Feinberg, Doing and Deserving (Princeton, NJ: Princeton
University Press, 1970), 221. It is a fair question whether this use of the term is faithful to Aristotle’s usage.
 STRICT LIABILITY WRONGS 309

defenses and limits on liability as well as basic forms of liability. Full consideration of
these is beyond the scope of this chapter. A brief discussion of limits and defenses may
nonetheless be useful. One of the great worries associated with strict liability is the
worry that such liability has no limits. This, indeed, is the worry that Guido Calabresi
and Jon Hirshchoff mention in the first paragraph of their seminal paper on strict
liability, and it is inextricably bound up with the question of what defenses to strict
liability should be recognized.45
Generally speaking, the boundaries of strict liability are fixed in two ways. When
harm-based strict liabilities are at issue, the liability is normally liability for the
characteristic risks of the activity. In Escola, for example, the risk that materializes is
a characteristic risk of the defendant’s activity. They chance that a particular Coke
bottle will fail to conform to the manufacturer’s own intentions for the product—and
explode unexpectedly because the glass is flawed or the bottle is overpressurized—is a
risk of manufacturing Coke bottles, a risk whose incidence is determined by the quality
of the inputs Coke chooses and the care with which it manufactures, tests and markets
the product. In many cases, to be sure, the identification and attribution of character-
istic risk is difficult. The availability of strict liability as an alternative to fault liability
often depends on whether non-fault criteria for charging accidents to activities can be
devised. Practically important as this problem is, however, it is not one that we can
address here. For our purposes, it will do to say that strict liability is liability for the
characteristic risks of an activity and to point to the example of Escola to illustrate that
idea. When rights-based liabilities are at issue, the limit of the liability is determined in
the first instance by the character of the right. The proper scope of the power to
exclude others from one’s property determines the reach of the tort of trespass.
The arguments of this chapter contribute to thinking about the scope of strict
liability in only two ways. With respect to harm-based activities, the relevant point
is that strict liability is liability for harm that should not be avoided, yet which should
be repaired because otherwise that party will benefit itself by imposing harm on others.
When we consider the choice between strict liability and negligence, the question to
which this point directs our attention is the question of whether the harms caused by
an activity once reasonable care is exercised are harms that those who suffer them may
plausibly assert should be borne by those who benefit from their infliction. If so, strict
liability is a live alternative to negligence. An all things considered case for preferring
strict liability over negligence, however, must rest on a broader set of considerations.
The primary burden of this paper has been to try to show what the distinctive morality
of strict liability is, not when strict liability is, all things considered, justified. With
respect to autonomy rights-based strict liability, the lesson of this chapter is that strict
liability is appropriate when the autonomy guarded by the right in question would be
compromised by the recognition of the absence of fault as an excuse for failing to
comply with the right.

45
Guido Calabresi and Jon. T. Hirschoff, “Toward a Test for Strict Liability in Torts,” 81 Yale L.J. 1055
(1972).
310 GREGORY C . KEATING

The ideas at work in strict liability may shed a bit more light on the defenses to strict
liability. Prior to the adoption of comparative negligence, the defense of assumption of
risk was commonly associated with strict liability, and rightly recognized as a kind
of victim strict liability.46 The connection between assumption of risk—properly
applied—and strict liability is a strong one. Assumption of risk involves the assertion
of the victim’s agency, in the pursuit of the victim’s purposes, with respect to harm.
Physical harm (the form or harm with which tort law is preoccupied) is the impair-
ment of (normal) physical agency. The assertion of agency with respect to harm can be
transform a victim’s relation to harm.
Consider an admittedly extreme example. High-altitude mountaineering (e.g.,
summiting K2) exposes its practitioners to a significant risk of long-term memory
loss. For one person to put another in a hyperbaric chamber and simulate the low
oxygen, low barometric pressure of the “death zone” above 26,000 feet would be an
infliction of unacceptable risk. For even if the victim were acclimated to the conditions,
the exposure to air that thin would put them at risk of significant memory loss—at risk
of serious impairment. For someone who chooses to pursue high-altitude mountain-
eering matters are different. They incorporate the risk of serious harm (memory loss)
into a larger narrative; they yoke exposure to that risk to the pursuit of a value they
consider more important, and they encounter that risk in the course of exercising
various capacities of agency very fully. To put it differently, they identify with the
pursuit of a valuable end which can only be realized by encountering this particular
risk of harm.
Both the exercise of agency and the exposure to harm involved in summiting K2 are
extreme, but the same phenomenon occurs in less extreme cases to which we do apply
the doctrine of assumption of risk. In our era, sports are the principal habitat of
assumption of the risk and sports are activities whose distinctive values can often only
be realized by encountering above-normal risks of harm. A skier confronting an expert
slope wants to encounter the steepness, the iciness and the bumpiness that make it
difficult to ski. The capacities that she wishes to exercise, the competence that she
hopes to display, and the value that she seeks to realize can only be realized by
encountering a heightened risk of falling and injuring herself. Her identification
with that project transforms her relation to the harm she risks. Harm chosen in the
pursuit of value is not harm imposed and suffered.

E. Embracing strict liability

Rights-based and harm-based forms of strict liability may both ebb and flow, but they
are entrenched forms of tort liability. Trespass, battery, and conversion are neither
going away nor on the path to becoming fault-based forms of liability. They are here to

46
This is very clearly recognized and powerfully explained in economic terms in Guido Calabresi and Jon
T. Hirschoff, “Toward a Test for Strict Liability in Torts,” 81 Yale L.J. 1055 (1972).
 STRICT LIABILITY WRONGS 311

stay, and they will stay strict; the rights that they protect require this strictness.
Harm-based strict liabilities ebb and flow more dramatically, but they are not about
to disappear from our law of torts altogether. Vicarious liability is ineradicable, and
intentional nuisance is as entrenched as trespass or conversion. Harm-based strict
liabilities are, therefore, likely to persist. Moral theories of tort must therefore make
room for both sovereignty-based and harm-based strict liabilities and abandon the
quixotic project of purging them from the law of torts. Making room for these two
forms of strict liability requires accepting two ideas. The first is that one can commit a
tortious wrong simply by violating an autonomy right—by interfering with someone’s
control over their self or their property in ways that that person has not authorized.
The second is that there is a form of tortious wronging which consists of harming
without repairing. The ship owner in Vincent v Lake Erie committed this very wrong.
He was justified in lashing his ship to plaintiff ’s dock even though he was certain to do
damage to the dock. He was justified in inflicting lesser harm on the dock to avoid
greater harm to his ship. But he was not justified in failing to make reparation for the
harm he had done. In failing to make reparation for the harm he had done, the ship
owner was simply benefitting himself by harming someone else. By failing to repair the
damage to the dock, the ship owner both denied and impaired the dock owner’s
autonomy. He denied that autonomy by depriving the dock owner of the power to
choose the uses to which its resources would be put; he impaired it by leaving the dock
owner’s damaged. The ship owner was therefore privileged to inflict harm but obli-
gated to repair the harm that he inflicted. He had a duty not to harm without repairing.
 15
Normative Theories of
Punitive Damages:
The Case of Deterrence
Anthony J. Sebok*

I. Introduction
Debate over punitive damages can fall into three categories, albeit with overlapping
boundaries. First, one can try to define the appropriate legal standards or rules for
punitive damages in their specific legal system. Comparative study of punitive dam-
ages is merely an extension of this category. After defining the law properly in more
than one jurisdiction, one can then compare the law of punitive damages amongst
jurisdictions, whether between states within a nation or between other nations.1
Second, one can engage in empirical studies of punitive damages. One can try to
determine how often they are awarded, in which type of lawsuit they are awarded, how
much are the awards, how often plaintiffs win awards, etc. This is difficult but crucial
work, since public perceptions and misperceptions about the “real world” effects of the
existing law of punitive damages drives both legislative and interpretive alteration of
the law.2
Third, one can attempt a normative evaluation of punitive damages. One can offer
an argument for or against them, or for them only under certain circumstances, based

*
I am grateful to Mark Geistfeld, Steve Landsman, Robert Rabin, Peter Schuck, Cathy Sharkey, and Larry
Solan for their comments and suggestions. All mistakes are my responsibility alone.
1
See Linda L. Schlueter and Kenneth R. Redden, Punitive Damages 4th ed. (LEXIS, 2000); James
D. Ghirardi and John J. Kircher, Punitive Damages Law & Practice (West, 1997). Punitive damages have
been retained in some common law nations but are technically not available in civilian legal systems. See John
Y. Gotanda, “Charting Developments Concerning Punitive Damages: Is the Tide Changing?,” 45 Colum.
J. Transnat’l L. 507 (2007).
2
Empirical research about punitive damages can be misused. This was illustrated in Exxon Shipping Co. v
Baker 554 US 471, 128 S Ct 2605 (2008). After the decision appeared, some of the authors of studies upon
which the court relied in that case disputed its interpretation of their work. See Theodore Eisenberg, Michael
Heise, and Martin T. Wells, “Variability in Punitive Damages: An Empirical Assessment of the US Supreme
Court’s Decision in Exxon Shipping Co. v. Baker,” 166 J. Institutional & Theoretical Economics 5 (2010).
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 313

on principles of economics, moral philosophy, politics, or even theology. Normative

evaluation of this sort can be rooted either in one’s own substantive commitments or,
in theory, it could be based on a more arm’s-length evaluation based on the norms that
instantiate one’s legal system regardless of whether one actually approves of those
norms. What distinguishes the first category of research from the third—although they
overlap, of course—is that the latter is predominantly normative, while the former is
predominantly positive.
Research connected to the first category has dominated American legal scholarship,
followed by research related to the second category and only then have we seen some
tentative efforts in the third category. This chapter is a companion piece to an earlier
article where I argued that punitive damages are best understood as a form of private
retribution.3 My argument is only one variation of a family of arguments that take as
their starting point that punitive damages (at least as currently conceived in the
common law) are consistent with corrective justice or civil recourse theories of private
law.4 These arguments, in turn, belong to a larger family of arguments that justify
punitive damages because they are a form of retribution either required by, or
consistent with, the common law.5 There is considerable disagreement amongst the
various theories of punitive damages based on corrective justice, civil recourse, and
retributivism, but this chapter will not examine those disagreements, leaving that
project to another time. Nor will this chapter even focus on providing a positive
argument for these families of theories, which is not to say that I think that their case
has been made persuasively by any of their advocates, myself included. This chapter
will have a more limited goal, which is to argue against the other families of normative
arguments for punitive damages, which I call Deterrence Theory. My strategy in this
chapter is to argue that Deterrence Theory fails as a normative theory of punitive
damages on its own terms; that is, that if the goal of the common law is to maximize
social welfare through deterrence, then punitive damages as they exist in the common
law cannot be justified. Clearly, this chapter, even by its own terms, does not offer a
strong and direct argument for an alternative theory of punitive damages in the
common law. However, as I will argue in the conclusion, there are lessons that can
be drawn from the failure of Deterrence Theory to justify punitive damages that may
offer tentative clues as to why retributive theories of punitive damages might be
adequate to the job of justifying the practice.

3
Anthony J. Sebok, “Punitive Damages: From Myth to Theory,” 92 Iowa L. Rev. 957 (2007).
4
See, e.g., Benjamin C. Zipursky, “A Theory of Punitive Damages,” 84 Tex. L. Rev. 105 (2005); Curtis
Bridgeman, “Note, Corrective Justice in Contract Law: Is There a Case for Punitive Damages?,” 56 Vand.
L. Rev. 237 (2003).
5
See, e.g., Marc Galanter and David Luban, “Poetic Justice: Punitive Damages and Legal Pluralism,” 42 Am.
U. L. Rev. 1393 (1993); Thomas B. Colby, “Beyond the Multiple Punishment Problem: Punitive Damages as
Punishment for Individual, Private Wrongs,” 87 Minn. L. Rev. 583 (2003); Dan Markel, “Retributive Damages:
A Theory of Punitive Damages as Intermediate Sanction,” 94 Cornell L. Rev. 239 (2009).
314 ANTHONY J . SEBOK

II. The Law of Punitive Damages

Before I take up the normative theories that have been offered, it is worth pausing to
note that the debate over the positive law of punitive damages in the United States has
been complicated by an obsession with constitutional law and that this complication
may have had the effect of suppressing a more explicitly normative discussion of
punitive damages in the United States. It is important to understand the relationship
between the US Supreme Court’s constitutional theory of punitive damages and the
various common law theories that the states have always been using to set out the
grounds and limits of punitive damages in their own jurisdictions.6
Any substantive law of punitive damages possesses a certain structure, much like
any other liability rule. A law of punitive damages must describe why a court will allow
punitive damages to be awarded; when the court will allow punitive damages to be
awarded; and how punitive damages will be awarded. The answer to the why question
informs the fact-finder’s decision whether to award punitive damages at all and, if the
decision to award is positive, how large to make the award. For example, the decision
to award might be different, depending on the facts of a case, if the rationale for
punitive damages is to secure “restitution” for the theft of the plaintiff ’s rights or
freedom as opposed to providing general deterrence to prevent actors similarly
situated from engaging in the same antisocial conduct. The “when” question refers
to two things: case type, and defendant’s culpability. The “how” question refers to the
evidence and tools that the fact-finder is supposed to use in deciding whether to award
punitive damages and, if the decision is positive, the amount of those damages.
Inevitably the “how” question is intertwined with the “why” and “when” questions.
If a jurisdiction takes the position that a defendant may properly be asked to pay
punitive damages if they acted “wantonly,” then the fact-finder will have to access
evidence that either supports or rebuts the claim that the defendant acted wantonly. If
a jurisdiction takes the position that one purpose of punitive damages is to punish a
defendant for the injuries that he caused to non-parties in addition to the injuries that
he was proven to have caused the plaintiff, then the fact-finder will have to access
evidence that either supports or rebuts the claim that the defendant’s acts injured non-
parties.
Since independence from England, when state courts confronted their own law of
punitive damages, they were essentially working through a problem in the common
law of torts.7 Occasionally the state courts would explicitly or implicitly rely on specific
normative theories about punitive damages in order to answer the questions described
above. Sometimes that theorizing attempted to integrate punitive damages into the

6
Four states allow punitive damages only when authorized by statute and numerous states have imposed
statutory limitations and/or procedures on the awarding of punitive damages. See Exxon 554 US (note 2) at
495. Certain federal statutes authorize punitive damages as well.
7
But see note 6.
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 315

fabric of the common law, and sometimes judges rejected punitive damages as
normatively incoherent (or worse).8 In contrast, until recently, the United States
Supreme Court rarely, if ever, considered the normative foundations of punitive
damages, since the issue of punitive damages rarely presented itself in federal court,
and even more rarely as a matter of original jurisdiction.9 This changed in the late
1980s in the United States with a series of cases reviewing the constitutional permis-
sibility of punitive damages.10
In this chapter, I will not summarize the court’s current understanding of how the
Fourteenth Amendment constrains the substantive law of punitive damages. That has
been the subject of substantial analysis.11 Instead, I want to point out the relationship
between the constitutional question, which was the only explicit topic of seven of its
nine major decisions since 1989, and the subject of this chapter, which is the normative
theory behind punitive damages. As it developed its modern due process jurispru-
dence the court insisted that it would not impose its own substantive theory of
punitive damages on the states, but would set out the outer boundaries beyond
which no common law theory could go. The court’s assumption has always been,
therefore, that within the outer limits of the Constitution, the states are free to adopt a
wide range of legal principles relating to punitive damages.
It is easy to see why the court’s formal respect for state common law was difficult to
maintain. It is not clear how the US Supreme Court could identify unconstitutional
reasons for awarding punitive damages without evaluating the content of the reasons
proffered by any state’s common law theory of punitive damages. The US Supreme
Court has dealt with this challenge with two strategies. First, it set out the minimum set
of reasons which, from a constitutional point of view, no state can be permitted to use as
grounds for punitive damages, and second, in its most recent case, concerning federal
maritime law, it set out grounds for punitive liability in federal common law, thus giving
a hint as to what it thinks is an attractive substantive theory of punitive damages.
The US Supreme Court has established certain conditions which a state punitive
damages must meet if it is to survive appellate review under the federal constitution.12
To satisfy the Due Process Clause of the Fourteenth Amendment, an award must
reflect the reprehensibility of the defendant’s conduct, the compensatory damages

8
“The idea is wrong. It is a monstrous heresy. It is an unsightly and an unhealthy excrescence, deforming
the symmetry of the body of law.” Fay v Parker 53 NH 342, (NH S Ct, 1873), 382.
9
Cases involving the substantive law of punitive damages come before the federal courts only as a matter
of federal common law or diversity jurisdiction. See, e.g., The Amiable Nancy 16 US 546 (1818) (admiralty law);
Day v Woodworth 54 US 363 (1852) (diversity).
10
See Martin H. Redish and Andrew L. Mathews, “Why Punitive Damages Are Unconstitutional,” 53
Emory L.J. 1 (2004) (reviewing this history).
11
See Jill Wieber Lens, “Punishing for the Injury: Tort Law’s Influence in Defining the Constitutional
Limitations on Punitive Damage Awards,” 39 Hofstra L. Rev. 595 (2011), 608–34; Dan Markel, “How Should
Punitive Damages Work?,” 157 U. Pa. L. Rev. 1383 (2009), 1390–8 (reviewing recent literature).
12
This and the following sentences are drawn from Anthony J. Sebok, “The U.S. Supreme Court’s Theory of
Common Law Punitive Damages: An Inauspicious Start,” in Lotte Meurkens and Emily Nordin (eds.), The
Power of Punitive Damages: Is Europe Missing Out? (Cambridge/Antwerp/Portland: Intersentia, 2012), 143–4.
316 ANTHONY J . SEBOK

awarded by the fact-finder, and, when possible, comparable civil and criminal penal-
ties for similar acts by the defendant. The reprehensibility of the defendant’s act may
be based on the risk of harm or harm imposed by the defendant on others but the
defendant may not be punished for harming others. Although this list is supposed to
be modestly neutral with regard to the “why” question of punitive damages (naming
only the most anodyne rationales, “retribution” and “deterrence”), it is actually quite
aggressive. By limiting the answers to the “how” question in the way it does, the court
excludes many normative conceptions of punitive damages.
It would be perfectly reasonable for a theorist to criticize the court for imposing its
own narrow normative conception of punitive damages when a more pluralistic
approach would be in order.13 This was Justice Scalia’s objection as early as 1996.
Since the Constitution does not set out any limits to common law punitive damages,
the court’s constitutional arguments were just a pretext for imposing its own substan-
tive normative theory of punitive damages on the states.14
Approaching the court’s case law in this spirit, the theory of punitive damages
suggested by its one explicitly substantive decision does not look very attractive. In
Exxon, the US Supreme Court set out its theory of the common law of punitive
damages because federal maritime law governed the case, which is predominantly
federal common law. In addition to the minimum constraints it had set out earlier, the
court held there to be a new constraint: the amount of punitive damages must be,
except in special cases, based on a 1:1 ratio with compensatory damages.15 This was to
achieve two dimensions of fairness: consistency and predictability.16 The ratio was
based on the historical relationship between median punitive damage awards and
median compensatory.17 The court noted that its theory kept the door open to higher
awards in some circumstances, especially when the plaintiff experienced only modest
economic harm.18
This theory of common law punitive damages leaves many questions unanswered.
The first is why it emphasizes predictability, given that the common law of damages is
rife with unpredictability. In the United States especially, final compensatory damages
judgments are highly unpredictable for many reasons. Ex ante the actual accident, the

13
See, e.g., Catherine M. Sharkey, “The Exxon Valdez Litigation Marathon: A Window on Punitive
Damages,” 7 U. St. Thomas L.J. 25 (2010), 43 (observing that “what is perhaps most astonishing . . . is how
the Court” has “seize[d] the reins to define the legitimate state interest in punitive damages”); Tracy
A. Thomas, “Proportionality and the Supreme Court’s Jurisprudence of Remedies,” 59 Hastings L.J. 73
(2007), 76 (describing the court’s evolving doctrine as judicial activism).
14
BMW of North America, Inc. v Gore 517 US 559 (1996), 600–2 (Scalia, J., dissenting).
15
Exxon, 554 US (note 2) at 512. Earlier, the court suggested that the ratio between punitive damages and
compensatory damages should be in the “single digits” but it had not indicated that it viewed a ratio of 1:1 as
normatively attractive or required. State Farm Mut. Auto. Ins. Co. v Campbell 538 US 408 (2003), 425.
16
See Exxon, 554 US (note 2) at 497–502 (“Courts of law are concerned with fairness as consistency . . . ”
and “a penalty should be reasonably predictable in its severity, so that even the [theoretical bad actor] can look
ahead with some ability to know what the stakes are in choosing one course of action over another.”).
17 18
Exxon at 512. Exxon at 512.
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 317

scale of the actual injury resulting from the wrong can vary depending on moral luck.19
A defendant may strike a poor person with no income or a wealthy banker. The same
risk may result in instant death, with limited pain and suffering, or a grievous
permanent but non-fatal injury, which in the case of a young person, may result in
many decades of compensable pain. Ex ante the final judgment but ex post the
accident, the amount awarded can vary depending on the state of the evidence, the
skill of the lawyers, the presence of collateral sources (in states where the collateral
source rule has been dropped or amended) and of course, the subjective judgments of
the fact-finder—which may be a jury.20
But, of course, one can say that punitive damages are different from compensatory
damages in the eyes of the common law. Their main rationale is “retribution and
deterrence,” not compensation. But even taking this difference into account, it is not
clear why predictability—which entails making punitive damages the product of a
known and invariable multiplier—is the most important (or even a relevant) value. If
one were to take seriously the goal of private retribution, it is not clear why any fixed
ratio should be the law’s starting point, given that the project of tailoring punishment
to “fit the crime” will necessarily involve the fact-finder’s response to context-specific
facts that reflect the defendant’s contempt and indifference to the plaintiff ’s rights.21
And if one were to take seriously the idea that punitive damages are designed to force
the defendant to pay for “societal” damages, then it is not clear why the degree of
punishment should be tethered to the actual amount of damages awarded at trial.22
Finally, the most important rationale that the US Supreme Court excludes under
its common law theory of punitive damages is the view that punitive damages should
be designed to promote deterrence. As first set out by Polinsky and Shavell, punitive
damages perfect the cost-internalization strategy expressed in the Hand Test.23 The
Hand Test must be extended to permit supra-compensatory awards because of the
problem of under-enforcement, as not all inefficient acts result in liability equal to
the social loss they cause. This is for multiple reasons, although Polinsky and Shavell
emphasize just two, under-detection and under-compensation. It might be the case
that under-enforcement is the result of other imperfections as well, such as false
negatives in adjudication and the cost of adjudication. In any event, as illustrated by
Judge Posner in Mathias v Accor Economy Lodging Inc., deterrence may, on occasion,

19
See, e.g., Jeremy Waldron, “Moments of Carelessness and Massive Loss,” in David G. Owen (ed.),
Philosophical Foundations of Tort Law (Oxford: Oxford University Press, 1995), 387.
20
For a comprehensive examination of how bias can affect tort awards, see Martha Chamallas and Jennifer
B. Wriggins, The Measure of Injury: Race, Gender, and Tort Law (New York and London: New York University
Press, 2010).
21
See Keith N. Hylton, “Punitive Damages and the Economic Theory of Penalties,” 87 Geo. L.J. 421 (1998),
454.
22
See Catherine M. Sharkey, “Punitive Damages as Societal Damages,” 113 Yale L.J. 347 (2003), 367–70.
23
A. Mitchell Polinsky and Steven Shavell, “Punitive Damages: An Economic Analysis,” 111 Harv. L. Rev.
869 (1998). For a good review of the development of Deterrence Theory since 1998, see Robert J. Rhee, “A
Financial Economic Theory of Punitive Damages,” 111 Mich. L. Rev. 33 (2012), 52–8.
318 ANTHONY J . SEBOK

demand a ratio much higher than 1:1, even absent any extraordinary conditions such
as malice or very minor compensatory damages.24

III. Theories of Punitive Damages

Various surveys of the purpose of punitive damages produce long and somewhat
imprecise lists. I am guilty of this myself—in a 2002 article reviewing the historical
functions of punitive damages according to nineteenth and early twentieth-century
judicial and legal documents, I stated that judicial opinions reflected six distinct “pur-
poses” of punitive damages: compensation for emotional suffering, compensation for
insult, personal vindication, vindication of the state, punishment to set an example, and
punishment to deter.25 David Owen, in 1994, suggested that there were five possible
purposes: education, retribution, deterrence, compensation, and law enforcement.26
A decade earlier Ellis Dorsey suggested seven: punishing the defendant, deterring the
defendant from repeating the offense, deterring others from committing an offense,
preserving the peace, inducing private law enforcement, compensating victims for
otherwise uncompensable losses, and paying the plaintiff ’s attorneys’ fees.27
While there is some historical and practical truth to the claim that punitive damages
helped plaintiffs maximize their compensation, it is not much of an argument for
punitive damages.28 Punitive damages become a second-best solution to a problem
that should be, and in some cases has been, solved by expanding access to the courts
and the tort system. To the extent that punitive damages replace emotional distress,
there is no reason why, in an ideal world, those losses cannot be named under their
own heading if they are a compensable wrongful loss. This may be, in fact, what
happened, according to the Supreme Court in Cooper Indus. v Leatherman Tool
Group, Inc.29 To the extent that punitive damages make up the cost of attorneys’
fees, the tort system should be reformed so as to deal with the fact that the American
rule fails to fully correct all wrongful losses. But it is an odd defense of punitive
damages, especially since punitive damages are awarded so rarely, that they could not
be said to provide much relief to victorious plaintiffs dependent on contingency fee
attorneys. Similarly, saying that a “purpose” of punitive damages is the education
of the wrongdoer, keeping the peace, or incentivizing private law enforcement seems

24
347 F 3d 672 (US Ct of Apps (7th Cir), 2003) (37:1 ratio where hotel refused to prevent bedbug
infestation).
25
See Anthony J. Sebok, “What Did Punitive Damages Do?” 78 Chi.-Kent L. Rev. 163 (2003), 197.
26
David G. Owen, “A Punitive Damages Overview: Functions, Problems and Reform,” 39 Vill. L. Rev. 363
(1994), 373.
27
Dorsey D. Ellis, Jr., “Fairness and Efficiency in the Law of Punitive Damages,” 56 So. Cal. L. Rev. 1 (1982), 3.
28
This criticism is directed against those who see punitive damages as a supplement to the conventional
menu of compensatory damage awards (non-economic damages, especially). It is not meant to address the
more sophisticated argument that punitive damages provide compensation for wrongful imposition of risk. See
Mark Geistfeld, “Punitive Damages, Retribution, and Due Process,” 81 S. Cal. L. Rev. 263 (2008), 279–80.
29
532 US 424 (2001), 437 n. 11.
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 319

to trade on the weakness of other parts of the legal system which, were they working
well, should directly take on functions that the tort system is not well suited to
perform.
This leaves three (or four) concepts: punishment, deterrence, and retribution (and/
or vindication).30 These terms have a tendency to be used in overlapping ways, and so
it is important to define them clearly so that we can “translate” the use of a term by a
court or a scholar into a language that allows us to find its partner in the vocabulary of
another court or scholar. The easiest term to pick out is “deterrence.” It has a very
specific set of meanings, although many subsidiary controversies remain among those
who wish to advocate for a particular conception of deterrence. At a very minimum,
deterrence justifies itself by reference to the practical consequences it will produce in
the future. Usually the consequences which underwrite deterrence are welfarist, but
not always. One could imagine a deterrence argument being based entirely on non-
welfarist consequences, such as threatening someone with a sanction in order to
prevent them from breaking a promise they had made; not because of the effect of
the broken promise on social welfare, but because one simply believes that promises
ought not to be broken. A more subtle and difficult question is whether the conse-
quence sought by deterrence is a change in the behavior of the person who suffers the
sanction (specific deterrence) or in anyone who may be in the position of the person
who suffered the sanction (general deterrence). Criminal law is generally justified on
the basis of general deterrence, and it is hard to see why punitive damages would not
also be justified under general deterrence as well.
Of the remaining terms, there is an easy and a hard set of distinctions to draw. The
easy distinction is between punishment and retribution (and/or vindication). The term
“punishment” is not a useful concept for explaining the purpose of punitive damages
since the word “punitive” is a synonym for punishment. Punishment, along with
compensation and restitution, justifies the practice of awarding damages (as opposed
to some other judicial response, such as a declaratory or injunctive command). That is
why I argued above that the effort to justify punitive damages in terms of making up
for some missing compensation is embarrassed by the obvious question as to why the

30
There is a sense in which retribution and vindication are quite similar. Both ground the vulnerability of
the defendant to hard treatment on the defendant’s wrongful act. Markel draws a distinction between the two
concepts based on two distinctions. First, in retribution the defendant is vulnerable to hard treatment by the
state, while in vindication the vulnerability is to a private citizen, and second, for that reason, vindication is
more likely to produce over-punishment. See Markel, “Retributive Damages: A Theory of Punitive Damages as
Intermediate Sanction” (note 5) at 254–7. I am not sure that these distinctions are illuminating. Markel’s claim
that vindication is inferior to retribution because it is “primarily a victim vindication account, and only
secondarily as an account about vindicating the public’s interest in retributive justice” just begs the question he
claims to be answering. Markel, “Retributive Damages” at 254 n. 62. If retribution means public vindication—
which is not obvious—then Markel’s account runs into a variation of the criticism I am making against
Deterrence Theory, which is that tort law seems especially ill-equipped to secure the ends which justify it. That
is, it is almost certainly the case that administrative proceedings will secure the retribution owed by the
defendant to society better than common law litigation. The problem with Markel’s theory is not that he sees
punitive damages as an intermediate sanction justified by the defendant’s wrongful conduct. It is that he
cannot explain why this sanction is part of private law, as opposed to public law.
320 ANTHONY J . SEBOK

tort system does not simply reform its compensatory damages law. Similarly, argu-
ments that punitive damages are justified because they are a form of restitution, while
intriguing, shift the very question of punitive damages to a new ground.31 Rather than
defend the choice to punish the defendant, those who see punitive damages as
restitution claim that the ground of the remedy is the wrongful status of certain
gains in the hands of the defendant; the purpose of punitive damages is to “return”
these gains to their rightful owner, the victim. Restitution-based theories have an
additional burden that retributivists do not: they must explain why these damages are
conditioned in the common law upon proof of a certain culpable mental states, namely
recklessness and intention, which are not otherwise conditions for extra-compensa-
tory damages in the law of restitution.32
If punishment is an adjective that describes a type of remedy—that is, monetary
awards designed to punish, not compensate or restore—then deterrence and retribu-
tion (with its variant, vindication) are “functions of punishment.”33 Deterrence and
retribution are the functions of punishment identified and debated by the two major
“camps” in mainstream Anglo-American criminal law theory, and they represent the
two poles around which modern penal theory is organized and debated.34 Therefore,
while the distinction (whether real or illusory) between retribution and vindication is
important, I will treat them both under the rubric of retribution for simplicity’s sake,
and thereby track the debate over punishment as it is conducted in criminal law.
Assuming that retribution and deterrence are the chief functions invoked by courts
and scholars for punitive damages, it is worth noting that, among criminal law
scholars, the parallel debate is really between “consequentialist theory of punishment”
and the “deontological theory of punishment,” since these theories of punishment are
invoked to support the choice of one function over another (retribution vs. deter-
rence).35 Thus, the debate between retribution and deterrence is over the justification
of punishment. The main point to be taken away from this discussion is that it makes
no sense to say that a function of punitive damages is either deterrence or punishment
(or a blend of the two); the debate is over the best justification of punishment, and
what functions are entailed by the answer to that question.36

31
See David G. Owen, “Response, Aggravating Punitive Damages,” 158 U. Pa. L. Rev. PENNumbra 181
(2010) (<http://www.pennlawreview.com/responses/index.php?id=85> accessed October 28, 2013); Ernest
J. Weinrib, “Punishment and Disgorgement as Contract Remedies,” 78 Chi.-Kent L. Rev. 55 (2003); Arthur
Ripstein, Equality, Responsibility and the Law (Cambridge: Cambridge University Press, 1999), 150–1.
32
See Hanoch Dagan, The Law and Ethics of Restitution (Cambridge: Cambridge University Press, 2004),
214.
33
See Kyron Huigens, “The Dead End Of Deterrence, And Beyond,” 41 Wm. & Mary L. Rev. 943 (2000),
945.
34
Huigens, “The Dead End Of Deterrence” at 952.
35
Huigens, “The Dead End Of Deterrence” at 945.
36
“The term ‘deterrence theory of punishment’ is a misnomer, as is the term ‘retributive theory of
punishment’ . . . both deterrence and retribution are merely functions of punishment, as are incapacitation,
rehabilitation, and so on. Whether any of these functions has a justifying effect depends on whether a moral or
ethical theory attributes a justifying effect to it.” Huigens, “The Dead End Of Deterrence, And Beyond”.
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 321

IV. The Incoherence of Deterrence Theory

Deterrence theories of punishment are varied, but they depend ultimately on some
form of consequentialist moral theory. The test for deterrence theories of punitive
damages is how well they fit and justify the practices currently found in the common
law. Under consequentialism, punishment is justified by the promotion of social
welfare through the deterrence of harm.37 The deterrence theory of punishment,
which was articulated in its modern form by Bentham, shares its premises with law
and economics, although for the most part law and economics is satisfied to explain
and justify how private law adjudicates claims for compensatory damages.38 Just as the
rise of law and economics in the twentieth century saw a reaction in the form of
corrective justice theory in torts, the deterrence theory of punishment—which saw a
tremendous outpouring of creative scholarly support in the twentieth century—has
been met with a reaction by various non-consequentialist criminal law scholars, such
as retributivists and virtue ethicists.39
The deterrence theory of punishment is indifferent to the motives of the person who is
the object of punishment or her character. Fault and purpose are irrelevant at a primary
level; neither are sufficient or necessary conditions of punishment. In theory, the prospect-
ive orientation of consequentialism means that the justification for punishment does not
even have to be connected with an actual crime.40 This last concession has been the trump
card for many critics of the deterrence theory of punishment. They point out that if
consequentialism, under the right circumstances, requires the punishment of the innocent,
then deterrence has no limits and is not only morally obnoxious but also politically
dangerous because the state’s power will be limitless.41 Needless to say, consequentialists
have responded to the retributivists with a variety of arguments. Some are purely negative
and argue that, whatever risks deterrence poses, it has the virtue of candor and transpar-
ency, in contrast with retributivism, which also relies upon the punishment of the innocent,
but simply pretends not to.42 The more interesting line or argument for purposes of

37
See John Rawls, “Two Concepts of Rules,” 64 Phil. Rev. 3 (1955), reprinted in H.B. Acton (ed.), The
Philosophy of Punishment: A Collection of Papers (London: St Martin’s Press, 1969), at 105, 107 (“[i]f
punishment can be shown to promote effectively the interest of society it is justifiable, otherwise is not”).
38
But see William S. Dodge, “The Case for Punitive Damages in Contracts,” 48 Duke L.J. 629 (1999).
39
See David Dolinko, “Three Mistakes of Retributivism,” 39 UCLA L. Rev. 1623 (1992), 1623 (“It is widely
acknowledged that retributivism, once treated as an irrational vestige of benighted times, has enjoyed in recent
years so vigorous a revival that it can fairly be regarded today as the leading philosophical justification of the
institution of criminal punishment.”).
40
See Rawls, “Two Concepts of Rules” (note 37) at 107 (“What we may call the utilitarian view holds that
on the principle that bygones are bygones and that only future consequences are material to present decisions,
punishment is justifiable only by reference to the probable consequences of maintaining it as one of the devices
of the social order.”).
41
Rawls, “Two Concepts of Rules” at 112. This criticism is resonant of the claim, attributed to Dostoevsky,
that “[i]f God does not exist, everything is permitted.”
42
See Dolinko, “Three Mistakes of Retributivism” (note 39) at 1632–3; Polinsky and Shavell, “Punitive
Damages: An Economic Analysis” (note 23) at 1270 (“in any realistic situation—in which some punishment of
the innocent is unavoidable—retributivist punishment will in fact involve punishing the innocent”).
322 ANTHONY J . SEBOK

this chapter is the effort by defenders of deterrence to develop “mixed theories” of

deterrence where punishment is carefully shaped by side-constraints that either recognize
non-consequentialist rights as coordinate with utility or are second-order strategies for
maximizing social utility, properly measured.43 The literature on the accommodation of
retribution and other non-consequentialist norms by deterrence theory is vast, and this is
not the place to provide a review of its history, nor an evaluation of whom, if anyone, has
the better of the debate.44
The critic of the deterrence theory of punitive damages does not have to attack
utilitarianism or the deterrence theory in criminal punishment. One can reject the
deterrence theory of punitive damages and be agnostic on all these other questions.
This is for numerous reasons. First, it might be the case that overall social utility is
promoted by a system that based its criminal sanctions and private law remedies on
theories of retribution, compensation, or restitution. Or one might think that deter-
rence could justify punishment, but only in a subset of cases—that is, those cases of
criminal wrongdoing. Finally, one might think that deterrence could justify punish-
ment in any type of wrong, but only through a system of punishment that is centrally
designed and administered by the state. This is a minimalist critique to punitive
damages, since it would leave untouched deterrence-based arguments for criminal
sanctions and civil fines. This is the position of this chapter.
One thing that one notices as one reviews the voluminous literature defending
deterrence as an end in criminal law is how quickly deterrence’s defenders concede
that the success of punishment (from a consequentialist point of view) depends on an
extraordinary complex system of signals from the state. For example, Robinson and
Darley argue that the criminal law cannot effectively deter wrongdoing unless it
reflects popular intuitions about desert.45 Given that the criminal law sometimes
punishes people who have made mistakes of fact, it is hard to know how to square
this with so-called popular intuitions about desert, unless one retreats, as does Hart, to
a complex theory of excusing conditions in law.46 It may be the case that this is
sufficient to render the deterrent theory of criminal punishment both comprehensive
and attractive, or it may be the case that even further refinements about the conditions
under which punishment will successfully guide behavior away from criminal acts are
necessary.47 This will require persons trained in the techniques of legal reasoning to
conceive of and employ these special terms and concepts. But the complexity of the
system comes from other directions as well. Punishment works only if the sanctions

43
See Huigens, “The Dead End Of Deterrence, And Beyond” (note 33) at 958–61.
44
See H.L.A. Hart, “Prolegomenon to the Principles of Punishment,” in H.L.A. Hart, Punishment and
Responsibility: Essays in the Philosophy of Law (Oxford: Oxford University Press, 1968), 1, 5–11.
45
Paul H. Robinson and John M. Darley, “The Utility of Desert,” 91 Nw. U. L. Rev. 453 (1997), 477–8.
46
See H.L.A. Hart, “Negligence, Mens Rea and Criminal Responsibility,” in H.L.A. Hart, Punishment and
Responsibility: Essays in the Philosophy of Law (Oxford: Oxford University Press, 1968), 152.
47
See Huigens, “The Dead End Of Deterrence, And Beyond” (note 33) at 987 (citing Dan Kahan, “New
Path of Deterrence,” 95 Mich. L. Rev. 2477 (1977). For Kahan, “deterrence works, not only on our instru-
mental, means-ends reasoning, but also on our choices of ends and on our motivations.”).
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 323

imposed are neither too mild nor too severe because over-deterrence may be just as
costly to social welfare as under-deterrence.48 This will require, at a minimum, massive
data gathering, testing, constant reviews and readjustment, and, if behavioral law and
economics scholars are to be believed, the integration of complex social psychology
literature dealing with features of human behavior which are not intuitive and which
the average lay person may completely misapprehend.
The foregoing is not necessarily a reason to reject deterrence-based punishment in
criminal law, but it suggests that the continued support for deterrence as a justification
in the criminal justice system must coexist with an implicit, if not explicit, under-
standing that it is based on conditions which are artificial and which require significant
coordination to maintain.49 This suggests that a deterrence-based theory of punish-
ment in the common law could not achieve the legitimacy that a deterrence-based
theory of criminal punishment has among those who are generally supportive of
consequentialist theories of law. This is for two reasons. First, as we have seen, a
sophisticated deterrence-based theory of punishment will need to employ concepts
such as excuse and justification which are neither purely a matter of empirical fact nor
matters of convention. They will depend on specialized judgments made by fact-
finders guided by norms which are themselves controversial (but settled as a matter
of law). Second, a sophisticated deterrence-based theory of punishment will need to
employ not only a variety of facts about the probability and value of certain actions
and outcomes, but also the likely expected impact of those values on future actors.
While these facts may not be inherently controversial—they may be based on evidence
available to anyone—unless the facts are shared broadly and used consistently, the
individual punishment decisions may result in what in the punitive damages context
has been called “predictably incoherent judgments,” thus destroying the deterrence
signal to future actors.50 These risks are mitigated (perhaps not enough in the eyes of
some critics) in criminal law and in civil fines because the values (both normative and
empirical) used to produce punishment are selected by the state. This confers legit-
imacy at two levels—because they are the product of coordinate decision-making, the
values provide each decision-maker with a shared input (albeit, somewhat arbitrarily
derived) and because they are the product of the state as decision-maker, the values
can be defended as democratically legitimate—even if they are, in the end, somewhat
arbitrarily derived. Deterrence-based punishment in criminal law is a product of
coordination made possible by the sovereign, and it is the deterrence produced by
the sovereign, which, in the eyes of a consequentialist, legitimates the punishment.
The obvious objection to my argument is that there is no reason why deterrence-
based punishment cannot operate in the context of the common law if it were to

48
Polinsky and Shavell, “Punitive Damages: An Economic Analysis” (note 23) at 1263–4.
49
Public enthusiasm for deterrence as a justification for punishment persists despite evidence that our
current policy tools cannot achieve significant welfare gains through deterrence-based approaches. See John
J. Donohue, III and Justin Wolfers, “Estimating the Impact of the Death Penalty on Murder,” 11 Am. L. &
Econ. Rev. 249 (2009) (casting doubt on studies demonstrating the marginal deterrent effect of executions).
50
Cass R. Sunstein et al., “Predictably Incoherent Judgments,” 54 Stan. L. Rev. 1153 (2002).
324 ANTHONY J . SEBOK

somehow model itself after deterrence-based punishment in criminal law.51 After all,
all that would be necessary for deterrence-based punishment to work in the common
law would be a set of easily identifiable norms accessible to the fact-finder and a “good
enough” sense of how punishment decisions will influence future actors. To make the
objection even more pointed, one might point out that this is all that is required for the
Hand Test to work. I have already stated that my argument against deterrence theories
of punitive damages does not entail rejecting other forms of consequentialism, includ-
ing deterrence theories for compensation in tort and contract. If I concede that
deterrence might be justifiable in the determination of compensatory damages, then
I must concede that it could justify punitive damages.
There are two lines of argument against this response, one practical and one
conceptual. The practical argument is while there may be some circumstances where
the decentralized processes of the common law produce efficient outcomes—which is,
after all, central to Posner’s account of the development of the common law of torts—
the circumstances are in a context quite different from those in which common law
remedies are chosen in order to punish, as opposed to compensate.52 In most private
law cases, while there may be knotty questions of fact, which may often invite overt or
covert value judgments by the fact-finder, the fact-finder is not actually in a position to
exercise “strong” discretion over the remedy.53 Formally speaking, the remedy in a
hundred medical malpractice cases is always the same: compensation for the legally
cognizable injuries caused by the defendant’s negligence. The formal categories are not
up to the fact-finder to decide, even if the application of the formal categories might
vary dramatically in similar cases because of variations in the evidence and how fact-
finders work with the evidence. Assuming it was even possible to analogize a common
law jury to a sovereign promulgating and applying laws designed to promote obedi-
ence, the jury’s judgments over when and how to sanction conduct would involve
strong discretion. Legislators employ strong discretion when drafting penal statutes,
policies, and penalties—whether to fine, how much to fine, and with what levels of
increase due to aggravating factors; whether to imprison, how long to imprison, and
with what levels of increase due to aggravating factors; and so on.54 This simply cannot
legitimately be done in a decentralized process—which is perhaps why Bentham, one

51
This is Markel’s position.
52
See, e.g., Richard A. Posner, “A Theory of Negligence,” 1 J. Leg. Stud. 29 (1972), 31–2. As John Goldberg
has put it, “[i]ndeed, Landes and Posner go so far as to claim that tort doctrine—at least as it existed in the first
half of the Twentieth Century—tracks nearly perfectly the rules that would have been chosen by an efficiency
analyst who knew nothing of the state of existing law.” John C.P. Goldberg, “Twentieth-Century Tort Theory,”
91 Geo. L.J. 513 (2002), 547.
53
See Ronald Dworkin, Taking Rights Seriously (Cambridge, MA: Harvard University Press, 1977), 33–4
(strong discretion, in contrast with “weak” discretion, describes the power of a legal actor to choose among
norms and not just apply a norm as an agent of another actor).
54
The challenge for a decentralized system becomes even more apparent if one divides laws designed to
promote obedience into “rules of conduct” and “rules of decision.” See Meir Dan-Cohen, “Decision Rules and
Conduct Rules: On Acoustic Separation in Criminal Law,” 97 Harv. L. Rev. 625 (1984).
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 325

of the earliest architects of deterrence theories of punishment, strongly preferred

legislation over the common law.55
The conceptual argument goes to the heart of my objection to the deterrence-based
theory of punitive damages. Rather than argue that common law juries exercise the
same sort of judgment as a sovereign designing a deterrence-based set of commands
(like a legislature writing a criminal code), the proponent of a deterrence-based theory
of punitive damages might object that I have defined the relevant type of judgment too
narrowly. Fact-finders who issue remedies in tort and contract cases may not be
primarily directing their thoughts toward future actors, but they are issuing judg-
ments, the result of which is that the behavior of future actors is incentivized.
Attention to the subtleties of institutional design in a deterrence-based system of
commands is most relevant if deterrence is based on the traditional conception of
mens rea.56 Bentham had a conception of mens rea that would be considered by most
modern law and economics scholars (as well as Hart) to be artificially narrow: he
believed that unintentional wrongdoing, including negligence, could not be deterred.57
Bentham’s conceptual separation between states of mind that were the proper subject
of public law commands and those that were not, did not necessarily entail a rejection
of the common law. As Holmes observed, the legitimacy of the common law did not
vanish once it was understood that common law liability was not based on identifying
culpable mental states similar to the mens rea of criminal law.58 All that vanished was
the confusion that the common law was somehow about punishment.59
Modern deterrence theorists deemphasize the significance of culpable mental states
(like mens rea in criminal law) in part because they reject the idea that the primary
value of sanctioning is to deter someone from acting wrongfully on purpose. Not only
do they reject any special moral or normative significance between intentional acts and
negligence (or strict liability), they also take a purely functional view of punishment.60
Punishment is indistinguishable from liability because, from a deterrence point of
view, there is no necessary difference amongst grounds of legal responsibility:
It is important to remember, too, that “negligence” as it is used [in deterrence theory] is a
term of art that covers more than its usual legal meaning. That is, in economic analyses of
law, “negligence” is used to refer to any regime in which defendants are legally liable for

55
See Lindsay Farmer, “Codifying Crime, Finding Government: Reconstructing the English Codification
Debate: The Criminal Law Commissioners 1833–45,” 18 Law & Hist. Rev. 397 (2000), 409.
56
Huigens, “The Dead End Of Deterrence, And Beyond” (note 33) at 956.
57
Jeremy Bentham, An Introduction to the Principles of Morals and Legislation (Hafner Publishing Co.,
1970), 172–5.
58
See Oliver Wendell Holmes, The Common Law, Lecture III: Trespass and Negligence.
59
Holmes argued that early law was characterized by an extreme concern for subjective blameworthiness,
or immorality, on the part of the defendant. Oliver Wendell Holmes, Primitive Notions in Modern Law (pts.
1 and 2), 10 Am. L. Rev. 422 (1876), 11 Am. L. Rev. 641 (1877).
60
See Huigens, “The Dead End Of Deterrence, And Beyond,” (note 33) at 956 (“If the promotion of social
welfare is the justifying purpose of punishment, then punishment need not turn on any particular aspect of the
wrongdoer or her wrongdoing. Consequentialism authorizes not only nonintentional criminal liability, such as
negligence, but also nonfault, or strict, criminal liability.”).
326 ANTHONY J . SEBOK

the harm they cause if, but only if, some aspect of their behavior is judged by a court to fall
short of some socially desirable level.61

Craswell uses the label “negligence” to cover any legal conclusion that the defendant
should pay the plaintiff. This “one size fits all” labeling is reminiscent of the way that
the term “trespass” was stretched to fit utterly dissimilar cases in early English tort law
which was, as Holmes noted, woodenly fixated on connecting liability to a culpable
mental state.62 Polinsky and Shavell recognize this in their theory of punitive damages.
They do not describe the grounds of liability in tort in terms of distinct categories of
strict liability, negligence, and intentional tort. Whether the defendant is legally bound
to pay the plaintiff is based on a unitary ground of liability, the principle that “to
achieve appropriate deterrence, injurers should be made to pay for the harm their
conduct generates, not less, not more.”63 The subjective mental state of the actor is
irrelevant to the legal question of whether to punish. In practical terms, this means that
sometimes defendants who, under the language of common law torts, acted neither
intentionally nor recklessly, will be subject to punitive damages while others who did
act not only intentionally but also reprehensibly, will not be subject to punitive
damages at all.64
There is, of course, a breathtaking lack of fit between what common law courts say
about the relationship between mens rea and punitive damages and what deterrence
theorists say is the true relationship. Any survey of published judicial opinions and
pattern jury instructions reveals a laser-like focus on intentional and reckless states of
minds.65 Whether deterrence theory should be discounted simply because it seems to
fail as an interpretation of current practice is a methodological dispute best left for
another time.66 My point here is that there is a relationship between the so-called
traditional conception of mens rea and punishment that, even if no longer relevant to a
modern tort law concerned mostly with negligence, is still relevant to tort law when it
is concerned with punitive damages. Despite sophisticated arguments over the precise
meaning of mens rea, modern punishment theory has never abandoned its centrality nor
its traditional “intentionalist” conception.67 This is true even for modern deterrence
punishment theory, which has, with great creativity, attempted to build intention back

61
Richard Craswell, “Deterrence and Damages: The Multiplier Principle and its Alternatives,” 97 Mich.
L. Rev. 2185 (1999), 2217.
62
See Charles O. Gregory, “Trespass to Negligence to Absolute Liability,” 37 Va. L. Rev. 359 (1951).
63
Polinsky and Shavell, “Punitive Damages: An Economic Analysis” (note 23) at 873.
64
Polinsky and Shavell, “Punitive Damages: An Economic Analysis” (note 23) at 905–6 (reprehensibility
“generally . . . should not . . . affect the imposition of punitive damages”).
65
Georgia is typical: punitive damages are authorized when “the defendant’s actions showed willful
misconduct, malice, fraud, wantonness, oppression, or that entire want of care which would raise the
presumption of conscious indifference to the consequences.” Ga. Code Ann. }51-12-5.1(b) (2010).
66
See Benjamin C. Zipursky, “Pragmatic Conceptualism,” 6 Legal Theory 457 (2000), 466 (a legal theory is
inadequate if it fails to account for and discuss the concepts and structure that are actually employed by the
legal actors within that area of law).
67
See Huigens, “The Dead End Of Deterrence, And Beyond” (note 33) at 945 (“Fault-also known as desert,
culpability, or blameworthiness is the distinctive feature of the criminal law.”).
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 327

into its theory.68 This is for a reason that tort theory cannot ignore: Judgments by courts
designed to induce others to conform their behavior (which is the linkage between
making the current defendant suffer a sanction and future actors fearing the sanction)
have to make sense to those to whom they are addressed; they have to be comprehensible
as punishments and not just prices. Although one could, functionally, stipulate that a
punishment is a price that exceeds the value received by the culpable defendant, this
would simply drain the former of its expressive content and reduce it into a variation of
the latter.69 The deterrent effect of a sanction, if it is to actually work to affect the choices
of future actors, needs the expressive content of an injunction—that is, it has to be
communicated as a mandatory and not optional practical statement. Recent behavioral
law and economic theory supports this point.70
Modern deterrence theory does not need to go further than describing liability
judgments in torts as prices for its account of the common law to work, and in the vast
majority of cases it covers, liability judgments might plausibly be described as prices,
since they are compensatory judgments. The dispute between corrective justice and
law and economics scholars therefore takes place on another level than description
of the form of the liability judgment. Both agree, referring to the example above,
that under the properly described rule of medical malpractice, a defendant who is found
liable must compensate the plaintiff by giving to her money equal to her legally
compensable losses. Where deterrence theorists of punitive damages go off the rails, so
to speak, is in their understanding of what it means for a tort judgment to punish. They
need not, and therefore do not, distinguish between a decision to punish from a decision
to compensate. The ground for a decision to punish in tort is the same as the decision to
compensate. They must treat the rule under which the decision is made as the same.
Conflating the reasons for punishment and compensation mis-characterizes what
occurs when common law courts turn to the question of punishment in torts cases. As
the United States Supreme Court perceptively noted, punitive damages are similar to
penalties in criminal law—and not compensatory damages in a torts or contracts
case—because the latter are based on “concrete” facts while the former “is not really a
‘fact’ ‘tried’ by the jury.”71 If all that would be needed for the tort system to achieve

68
See, e.g., Robert Cooter, “Prices and Sanctions,” 84 Colum. L. Rev. 1523 (1984), 1524–6; Jeffrey S. Parker,
“The Economics of Mens Rea,” 79 Va. L. Rev. 741 (1993), 741; Steven Shavell, “Criminal Law and the Optimal
Use of Nonmonetary Sanctions as a Deterrent,” 85 Colum. L. Rev. 1232 (1985), 1247.
69
See Thomas Hobbes, in Leviathan Richard Tuck (ed.), (Cambridge: Cambridge University Press, 1991),
215 (“Seventhly, If the harm inflicted be lesse than the benefit, or contentment that naturally followeth the
crime committed, that harm is not within the definition; and is rather the Price, or Redemption, than the
Punishment of a Crime . . . ”).
70
See Dan M. Kahan, “Between Economics and Sociology: The New Path of Deterrence,” 95 Mich. L. Rev.
2477 (1997), 2479–89; Dan M. Kahan, “What Do Alternative Sanctions Mean?,” 63 U. Chi. L. Rev. 591 (1996);
Kenneth G. Dau-Schmidt, “An Economic Analysis of the Criminal Law as a Preference-Shaping Policy,” 1990
Duke L.J. 1. (1990).
71
Cooper Indus., Inc. v Leatherman Tool Group, Inc. 532 US 424 (2001), 429 (the Seventh Amendment did
not apply to punitive damage awards, since punitive damage awards were not findings of fact “found” by a
jury), 437 (a fact is “concrete” if it “presents a question of historical or predictive fact” (quoting Gasperini v
Center for Humanities 518 US 415 (1996)), 459 (Scalia, J., dissenting)) (citations omitted).
328 ANTHONY J . SEBOK

optimal deterrence is the accurate application of liability rules provided by the court to
facts provided by the parties, then the typical negligence or strict liability trial would
not involve strong discretion (at least by the fact-finder). The same could even be said
about the typical trial over an allegation of intentional tort, because in theory, the fact-
finder is only finding a “concrete” fact about the defendant’s mental state. As Polinsky
and Shavell concede, this concrete fact is simply empirical information about human
behavior that informs the design of the penalty needed to deter future similar conduct,
information which is the same as that used to determine whether (and how much) to
award compensatory damages.
Deterrence Theory cannot explain the Supreme Court’s assertion that punitive
damages do not require the same sort of fact-finding that is done in the rest of the
tort suit. This awkwardness only grows when one realizes that the punitive damages
portion of a case is nothing like the rest of the “typical” liability case, even if it is
conducted in its midst. For example, whether punishment will occur is an act of
discretion by the victim; it is not a necessary entailment of the victim seeking redress.72
Further, and more important, the decision whether to award punitive damages is not a
mechanical process of applying the law to concrete facts. Fact-finders are not required
by the law to award punitive damages even if they find the requisite mental states of
intent or recklessness, and notwithstanding (or in spite of) the plaintiff ’s demand that
the defendant be punished if the “concrete” facts supporting compensation are proven.
The fact-finder’s reasons for choosing to punish defendants who are vulnerable to be
punished are not guided by anything more than the judge’s instruction that they are
permitted to punish and, of course, their own moral judgments about whether to
punish. This is precisely the sort of legislative power that sovereigns typically exercise.
Packed into what appears to be a “black box” of jury deliberations are decisions that
reflect, however unconsciously and crudely, intuitions about desert, excuses, mercy,
and (of course) the effect of the punishment decision on future actors. The same can be
said for the decision of how much to punish. This too is a legislative decision
analogous to the sentencing decision typically made with so much strurm und drang
by legislatures.73

V. Conclusion
Let me be clear about my argument against Deterrence Theory. It is not just that we
have no reason to believe that the common law system of punitive damages does what
it is designed to do—that is, it is not just that we have no reason to believe that a tort

72
This is one reason that civil recourse theorists find Deterrence Theory implausible given the actual
practices of private law. See Zipursky, “A Theory of Punitive Damages” (note 4) at 153–4; Sebok, “Punitive
Damages: From Myth to Theory” (note 3) at 1024–5.
73
Kate Stith, “The Arc of the Pendulum: Judges, Prosecutors, and the Exercise of Discretion,” 117 Yale L.J.
1420 (2008).
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 329

system with only compensatory damages deters inefficient conduct less effectively
than the current practice of punitive damages. Others have made that argument
persuasively.74 The debate over the Supreme Court’s misuse of statistics in Exxon
shows only that the court’s rationale for demanding that punitive damages conform to
a certain definition of “predictability” is without foundation, not that the deterrence
arguments for punitive damages based on ratios far exceeding 1:1 are valid.75 Given
that there is no reason to believe that the common law system of punitive damages
plays a significant role in deterring inefficient conduct, why do so many deterrence-
oriented torts theorists still embrace punitive damages? Perhaps they have an attitude
similar to that adopted by some deterrence-oriented citizens when faced with studies
that show that the death penalty (especially as currently imposed) is far too unpre-
dictable and infrequent to serve as a serious engine of deterrence.76 A deterrence-
oriented person might nonetheless view the death penalty as worth preserving for two
reasons. First, it might occasionally (and quite accidentally) deter, and something is
better than nothing. Second, those upon whom the death penalty falls have no reason
to complain since they have made themselves vulnerable to harsh treatment by their
wrongdoing.77 For someone opposed to the death penalty on non-consequentialist
grounds, this attitude of resigned tolerance towards the death penalty is immoral, but
it is not obviously illogical or immoral from within the confines of deterrence theory.
I think that punitive damages are different from capital punishment (or any criminal
punishment) in this respect: Unless a deterrence-oriented advocate of the tort system
actually believes the common law system of punitive damages really deters, building it
into the private law is not just a matter of taste or discretion. It is illegitimate; it is an
indefensible exercise of state power.
My argument in Section IV suggests why one cannot make the same sort of
“it’s better than nothing” argument for punitive damages that one could make if
one were a committed deterrence theorist confronted with the objection that
capital punishment is too crude (and muted) a signaling device to secure much
in the way of deterrence. The argument is simply this: Criminal penalties
designed to deter that are inefficient or even ineffective are a matter of regret for
those who design the system, but they are not illegitimate simply because they are

74
See, e.g., Cass R. Sunstein et al., Punitive Damages: How Juries Decide (Chicago, IL: University of Chicago
Press, 2002).
75
See Sharkey, “The Exxon Valdez Litigation Marathon: A Window on Punitive Damages” (note 13) at 41
(“the single point of agreement” amongst those who study punitive damages from an empirical point of view
“seems to be that the Court erred”).
76
See, e.g., John J. Donohue and Justin Wolfers, “Uses and Abuses of Empirical Evidence in the Death
Penalty Debate,” 58 Stan. L. Rev. 791 (2005).
77
This minimizes the risk that the system of imposing capital punishment produces no false positives.
Markel is quite sensitive to the seriousness of the question of false positives in punitive damages, and
recognizes that, while the moral stakes are different, the principle is the same: a punitive damages system
cannot be justified (especially under Deterrence Theory) if it is insensitive to the risk of false positives. See
Markel, “Retributive Damages: A Theory of Punitive Damages as Intermediate Sanction” (note 5) at 247, 266.
330 ANTHONY J . SEBOK

inefficient or ineffective. This is not to say that there could not be some level of
failure in design at which a deterrence system of criminal punishment becomes so
random and ineffective that it not only fails to qualify as a system of punishment
under its own terms anymore, but it would even violate moral and constitutional
limits on procedural justice and due process. But short of that, a “fair” system of
criminal law that takes the guilty and punishes them in order to produce deterrence
in others, which is incompetently designed to produce deterrence in others, may be
an embarrassment from the Benthamite perspective, but it is a legitimate embar-
rassment from the Benthamite perspective.
The legitimacy comes from the exercise of the sovereign power of the state to choose
penalties, which does not vary with the empirical success of the exercise of that power
or the technical competency of those exercising that power.78 The process of attempt-
ing to secure optimal deterrence for society is a public one; whether for good or for ill,
voters can select the public officials who will design the machinery of punishment
based on public-regarding reasons and information selected and presented according
to publicly approved procedures.79
None of this is true of punitive damages in the common law. As noted by civil
recourse theory, punitive damages instantiate several essential features of the general
practice of tort litigation and tort claiming, and these features are intrinsically hostile
to the those features of criminal law that provide legitimacy to the “it’s better than
nothing” argument adopted by a deterrence theorist hoping to defend criminal
penalties which are demonstrably inefficient or ineffective.80 As mentioned above, a
claim for punitive damages, like any claim in tort, is the exercise of a right to redress
which the state permits but does not require.81 Not only is the decision whether to
punish a personal choice—not one governed by public reason—but whether to credit
the plaintiff ’s judgment over how to punish is completely within the fact-finder’s
discretion, subject only to the constraints of state common law, state constitutional,
and federal constitutional boundaries—none of which necessarily require consider-
ations of the sort of technical questions asked by criminologists. This combination of
private authority on the part of the person deciding whether to ask for punishment
with strong discretion on the part of the persons deciding why and how to punish is
unlike the formal procedures by which public penalties are determined.

78
Again, it may be the case that at some point incapacity to design a deterrence-driven system of public
punishment may result in the loss of legitimacy, but it is hard to know when that point is reached.
79
The influence of public values on the penal system adopted in different states in the United States is quite
striking. See Vanessa Barker, The Politics of Imprisonment: How the Democratic Process Shapes the Way
America Punishes Offenders (New York: Oxford University Press, 2009).
80
See, e.g., Benjamin C. Zipursky, “Palsgraf, Punitive Damages, and Preemption,” 125 Harv. L. Rev. 1757
(2012), 1760; Andrew S. Gold, “The Taxonomy Of Civil Recourse,” 39 Fla. St. U. L. Rev. 65 (2011); John C.
P. Goldberg and Benjamin C. Zipursky, “Torts as Wrongs,” 88 Tex. L. Rev. 917 (2010), 961–3; Zipursky, “A
Theory of Punitive Damages” (note 4); Sebok, “Punitive Damages: From Myth to Theory,” (note 3) at 1024–5.
81
See Anthony J. Sebok, “What is Wrong About Wrongdoing?,” 39 Fla. St. U. L. Rev. 209 (2011), 220–1.
 NORMATIVE THEORIES OF PUNITIVE DAMAGES : THE CASE OF DETERRENCE 331

If there is a role for punitive damages to play in common law adjudication, it cannot
be built by analogy to public law punishment, since the reasons for tolerating the
failures of deterrence in public law (which are themselves a topic of constant debate
and reform) are simply absent in the private law. This is, as I have said, not just a
matter of whether Deterrence Theory can deliver what it promises, but it is a matter of
whether its failure can be excused under an adequate theory of legitimate legal
authority.
 PART III

DISTRIBUTIVE
JUSTICE
IN TORT LAW
 16
What is Tort Law For?
Part 2. The Place of
Distributive Justice
John Gardner*

I. Distributive Justice in a Corrective Context

[P]rinciples of distributive justice designed to cover the distribution of the benefits
and burdens of economic activity among individuals in a society . . . have been the
dominant source of Anglo-American debate about distributive justice over the last
four decades.1

So they have. It is hardly surprising, then, that when we think about distributive justice
and the law of torts, the issues that first spring to mind concern tort law’s impact on,
and sensitivity to, the distribution of resources (or wherewithal) across the wider
population. Should tort damages be calculated on the basis of the plaintiff ’s lost
earnings even where those earnings were obscenely high or obscenely low? Shouldn’t
there be a ceiling and a floor? Even with a ceiling and a floor, doesn’t tort law create
incentives to divert risks onto the disadvantaged, thereby compounding their disad-
vantage? Moving our attention from disparities among plaintiffs to disparities among
defendants, shouldn’t the precautions against accidents expected of the well-heeled be
greater than those expected of humbler folk? At any rate, shouldn’t the relative burden
of meeting the costs of safety, in the defendant’s particular circumstances, be relevant

* A very early and sketchy version of this chapter was presented as a paper at the “Obligations V”
conference in Oxford in July 2010. Later drafts were discussed at Rutgers and Harvard. I am grateful to the
many people who made valuable interventions on these occasions, and in particular to Peter Cane, David
Enoch, Dick Fallon, Andrew Gold, John Goldberg, Frances Kamm, Paul McMahon, John Oberdiek, Tim
Scanlon, Ken Simons, Jenny Steele, and Victor Tadros. Thanks also to Jeremy Farris, Ori Herstein, Sandy Steel,
and the students of Harvard’s Phil 277 course (Spring 2013) for written comments.
1
Julian Lamont and Christi Favor, “Distributive Justice,” in Edward N Zalta (ed.), The Stanford Encyclo-
pedia of Philosophy (Stanford, CA: Stanford, Fall ed. 2008), available at: <http://plato.stanford.edu/archives/
fall2008/entries/justice-distributive/> accessed October 28, 2013.
336 JOHN GARDNER

to the stringency of the defendant’s tort law duties? In general, shouldn’t the choices of
those with more limited options be treated differently by the law of torts, and by the
law generally, from the choices of those better furnished with alternatives?
Putting such questions, of (what we might call) “socio-economic justice,” center
stage, notes Tsachi Keren-Paz, are at least “two streams of scholarship” about the law
of torts:
The first stream attempts to defend the relevance and legitimacy of using private law for
(limited) redistributive purposes. The other . . . is the one trying to employ tort law
progressively, with an ambition to be sensitive to . . . the interests of disadvantaged groups
in society.2

In his excellent book on distributive justice in tort law, Keren-Paz assesses and
augments both streams of scholarship. In this essay, by contrast, I will not assess or
augment either of them. That is not because I disagree with them. On the contrary:
I believe that the questions listed in the previous paragraph are pressing. If it turns out
to be true that the law of torts has what Keren-Paz calls a “regressive bias,”3 i.e., that it
has worse impacts on the less well-off, we ought to be striving to put that right, subject
to the obvious imperative to make sure that we don’t end up doing more harm than
good in the attempt.
I mention this imperative, even though it is obvious, because it is easily forgotten by
those with a zeal for reform. Eliminating one set of regressive allocations does not
always help the cause of securing progressive allocations across the board. The
temptation to phase out regressive non-means tested distributions of public money,
such as universal childcare benefits, higher education and transport subsidies, and
blanket legal aid for criminal defendants, has already done great harm to the political
sustainability of the (generally progressive) welfare state in Europe. Denying the
middle classes a direct return on the welfare benefit system weakens the broad
consensus in its favor and hence opens the way to dismantling it, as today’s right-
wing politicians have gleefully come to realize. That is one reason why they like to
brand non-means tested benefits and subsidies as “regressive”: so that they can dupe
unsuspecting progressive voters, and even progressive politicians, into an unholy
alliance that will not have progressive consequences in the longer run. Before we
support overhauling or abolishing a regressive law of torts, we had better be sure that
we are not likewise playing into the hands of those who would like to see less
protection for the vulnerable.
But that, to repeat, will not be my topic here. I will not be concerned with the
desirability of changing the law of torts to cure or alleviate, or otherwise to respond to,
distributive injustices that could equally exist quite apart from the law of torts, such as
those that concern Keren-Paz. I will be concerned, rather, with problems of distribu-
tive justice that come into being only because the law of torts exists. And I will be less

2
Taschi Keren-Paz, Torts, Egalitarianism and Distributive Justice (Aldershot: Ashgate, 2007), 2.
3
Keren-Paz, Torts, Egalitarianism and Distributive Justice at 67.
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 337

interested in working out how these problems should be solved than in establishing
that they are there: that tort law creates, and cannot avoid dealing with, distributive
problems of its own.
This may have incidental implications for Keren-Paz’s project. One of the possible
ways in which we may do more harm than good in reforming tort law, if I am right, is
in rushing to mitigate tort law’s “regressive bias” without regard to the more special-
ized distributive tasks that are thrown up by tort law itself. Again, I leave it to others to
assess how much importance to attach to these more specialized distributive tasks.
My role will be limited to sketching out what they (or some of them) are.
In carving out this role for myself I am reacting mainly to those, led by Ernest
Weinrib, who regard all considerations of distributive justice as “extrinsic” or “alien”
to the law of torts. “Corrective justice is the form of the private law relationship,”4
claims Weinrib, and there can be no “combining [of] distributive and corrective
considerations within a single relationship.”5
In an earlier essay,6 of which this chapter is the promised sequel, I offered some
support to the first of these two claims. I argued that “any complete explanation of tort
law—whatever other considerations it may invoke—cannot but invoke considerations
of corrective justice.”7 The reason I gave was this: Some legal norms central to the law
of torts are themselves norms of corrective justice, and it follows, I argued, that they
can be assessed only in the light of their contribution to the doing of corrective justice.
This was, in outline, my argument:
[T]he law of torts cannot include a sound norm of corrective justice without there also
being a moral norm of corrective justice that the legal norm of corrective justice helps to
constitute. And once there is such a moral norm of corrective justice, the law of torts
cannot be [fully] justified without pointing to the role that the law of torts plays in securing
conformity with that very same moral norm.8

“Cannot be fully justified without” does not, however, entail “can be fully justified by.”
So, there is nothing here to lend credence to Weinrib’s second claim, the claim that the
“immanence of corrective justice in tort law”9 tells against “the introduction of
distributive considerations” into tort law thinking.10
In what follows, I will resist this second claim. Indeed, I will defend an almost
diametrically opposed claim: that certain questions of distributive justice are central to
the law of torts, and cannot but be faced by those who administer and develop it,
precisely because the law of torts is a site of corrective justice.

4
Ernest Weinrib, The Idea of Private Law (Cambridge, MA: Harvard University Press, 1995), 75.
5
Weinrib, The Idea of Private Law at 163.
6
John Gardner, “What is Tort Law For? Part 1. The Place of Corrective Justice,” Law and Philosophy 30
(2011), 1.
7
Gardner, “What is Tort Law For? Part 1” at 6.
8
Gardner, “What is Tort Law For? Part 1” at 25.
9
Weinrib, The Idea of Private Law (note 4) at 171.
10
Weinrib, The Idea of Private Law (note 4) at 74.
338 JOHN GARDNER

Those with pigeonholing instincts may be tempted to label this a “mixed” or

“pluralistic” explanation of tort law. Since I think that every pro and every con of
every action or practice counts in its assessment, and that no amount of theorizing can
properly eradicate the ultimate diversity of pros and cons, I am hardly in a position to
object to these as designations of my wider outlook on life.11 At the same time, the
interplay of corrective and distributive concerns that I will be investigating here is not
well described as a mere mixture, or a mere plurality. I endorse (and regard what
follows as helping to develop and finesse) Peter Cane’s thesis that “corrective justice
provides the structure of tort law within which distributive justice operates.”12 In my
version of this thesis,13 as we will see, the place of corrective justice in tort law enjoys
some kind of explanatory priority. So, “What is Tort Law For? Part 2” will presuppose
and rely on the main findings of “What is Tort Law For? Part 1” (hereinafter simply
“Part 1”). Yet the implication is not that the pursuit of distributive justice is the pursuit
of a goal extrinsic to tort law in the sense given to that expression by Weinrib. The
specialized distributive goals for tort law that we will be studying here are not goals
“independent of . . . the law that they [help to] justify.”14 Nor are they goals the
achievement of which is “socially desirable quite apart from tort law.”15 They are
distributive goals that the specifically corrective context of tort law brings into being,
and the pursuit of which apart from that context (or something very like it) would be
unimaginable, perhaps even unintelligible.

II. The Distribution of Correction

As I emphasized at the end of Part 1, there are rights to and duties of corrective justice
that exist independently of the law, and independently of any other kind of use,
observance, recognition, or adoption by anyone. They exist (as I put it) “in the raw
morality of trips to the beach, students in trouble, and disappointed children.”16 In the
situations I had in mind, a moral duty owed to another person, a rightholder, goes
unperformed. Even though the time for performing the duty is past, the reasons why
that (“primary”) duty to the rightholder existed (as well as the reason constituted by
the fact that it was a duty) still exert their force as reasons for some fallback action,

11
Even for Gardner, “[p]resumably, coherence would count towards soundness,” hopes Weinrib in the
conclusion of his Corrective Justice (Oxford: Oxford University Press, 2012), 336. Not so. Any justification has
to be coherent in the thin sense of intelligible. But Weinribian unity (or Dworkinian integrity) is not, in my
eyes, any kind of plus. Reality, including moral reality, is fragmentary.
12
Peter Cane, “Distributive Justice in Tort Law,” New Zealand L. Rev. 401 (2001), 413.
13
Which may well differ from Cane’s. For him “corrective justice is a ‘formal’ principle whereas distributive
justice is a ‘material’ principle,” Cane, “Distributive Justice in Tort Law” at 416. I have argued that there are no
formal principles of justice in John Gardner, Law as a Leap of Faith (Oxford: Oxford University Press, 2012),
Chapter 10.
14
Weinrib, The Idea of Private Law (note 4) at 4.
15
Weinrib, The Idea of Private Law (note 4) at 4.
16
Gardner, “What is Tort Law For? Part 1” (note 6) at 50.
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 339

which is the subject of a new (“secondary”) duty to the same rightholder. By perform-
ing the secondary duty—say, providing a new treat in substitution for a missed
outing—one reduces the deficit in one’s reason-conformity that was left by one’s
nonperformance. I called this thesis, the thesis that the secondary duty exists for the
reasons that were left unsatisfied by the nonperformance of the primary duty, the
“continuity thesis.”
By their nature, raw moral rights and duties are not allocated by anybody. They
exist, as I said, irrespective of their use, observance, recognition, or adoption. There is
therefore no question of anyone’s having such rights and duties either justly or
unjustly. Even when they are raw duties of justice it is neither just nor unjust that
they are the duties of justice one happens to have. So there is in the context of morality
in the raw nothing analogous to the problem that, in the context of legal decision-
making, Guido Calabresi and Douglas Melamed call “the problem of entitlement,” the
problem of how rights and duties are to be distributed as between “two or more people,
or two or more groups of people” with “conflicting interests.”17 It makes no sense to
ask what distribution of raw moral rights and duties would be just, even when the
interests that they serve conflict, because raw moral rights and duties are incapable of
being distributed. They come and go with the reasons for and against their existence,
and irrespective (to repeat one more time) of their use, observance, recognition, or
adoption. In this context, and to this extent, Weinrib is quite right to say that
“corrective justice operates on entitlements without addressing the justice of the
underlying distribution.”18 With morality in the raw, there is no such thing as the
justice of the underlying distribution, because nothing relevant has been distributed.19
Things are, however, very different with the law. Legal rights and legal duties, as
Calabresi and Melamed rightly assume, are respectively conferred and imposed by
someone. Calabresi and Melamed focus on “decisions” to confer and impose them.
This is too narrow. There are also many accidental conferrals and impositions of legal
rights and duties, notably by custom in foro and in the tacit premises of judicial
argument.20 Nevertheless, Calabresi and Melamed are right to think that conferral and
imposition (whether intentional or accidental) is how legal rights and duties respect-
ively come into existence. Rights and duties must be used, observed, recognized, or
adopted by someone in order to be part of the law. I will call the process of making
them part of the law, intentionally or accidentally, their “institutionalization.” In Part 1
I concentrated on the role that institutionalization can play, when it is done well, in
augmenting and refining the raw morality of corrective justice, in particular by

17
Guido Calabresi and Douglas Melamed, “Property Rules, Liability Rules, and Inalienability: One View of
the Cathedral,” 85 Harv. L. Rev. 1089 (1972), 1090.
18
Weinrib, The Idea of Private Law (note 4) at 80.
19
Which should not be mistaken for the proposition that there are no raw moral rights and duties of
distributive justice. There most certainly are. They bear on the distribution of things other than raw moral
rights and duties. For a compelling defense of their existence against “constructivist” doubters, see G.A. Cohen,
Rescuing Justice and Equality (Cambridge, MA: Harvard University Press, 2008).
20
See Gardner, Law as a Leap of Faith (note 13) at Chapter 3.
340 JOHN GARDNER

“determining at least some of its applications.”21 (The implication being that the raw
moral position is no longer the whole moral position.) Here I want to focus, instead,
on one sub-question that arises when we ask whether the institutionalization of
corrective justice has been “done well,” namely whether it has been done justly. Unlike
some people I know,22 I do not think this is the only sub-question that arises. The
institutionalization of corrective justice also needs to be done prudently, sensitively,
humanely, efficiently, honestly, and so forth, sometimes at the expense of its justice.
For present purposes, however, it is the question of justice that interests us. And this
question of justice, it seems to me, is not a question of corrective justice. The relevant
question of corrective justice has ex hypothesi already been answered. There is a moral
duty of corrective justice in the neighborhood, and it is now a candidate for institu-
tionalization. The relevant question of justice now is: How do we allocate the institu-
tionalization? How do we distribute, across the vast range of candidate wrongdoers
and candidate persons wronged, the numerous possible sets of legal arrangements to
support the doing of corrective justice as between them?
Here is how the question arises, more specifically, in the law of torts. It is part of the
nature of a tort that designating some wrong as a tort—classifying it as a legal wrong
under the “tort” heading—entails creating a legal right to corrective justice in favor of
those who are wronged.23 This legal right is a complex one. Its incidents include not
only the wrongdoer’s legal duty to repair, but also a largely undirected24 legal power
for the person wronged to determine whether that legal duty is concretized and
enforced through the courts, with a consequent duty on the courts to assist, when
that power is validly exercised by the issue of proceedings.25 When this right is
conferred, public authority (the authority of the court) is put at the disposal of the
wronged person. When the rule of law prevails, moreover, the authority is laid on
partly at public expense (in that the court does not recoup its full running costs from
its users, and may also arrange for other user costs to be subsidized). The wronged
person, in short, is given a right not only against the wrongdoer but also against the
court, a right to conscript the court (and its officers) in his or her quest for corrective
justice against the wrongdoer.
In deciding whether something should be a tort, then, it is never enough to conclude
that it is a wrong calling for repair. It is not even enough to conclude that it should
be recognized by the law as a wrong calling for repair. The question that must be

21
Gardner, “What is Tort Law For? Part 1” (note 6) at 18.
22
E.g., John Finnis, Natural Law and Natural Rights (Oxford: Clarendon Press, 1980), Chapter 7.
23
This is a point of law. I defended it as a valid one in: John Gardner, “Torts and Other Wrongs,” 39 Fla. St.
U. L. Rev. 43 (2011).
24
An undirected legal power is a legal power that is not coupled with legal duties regulating its exercise. See
Joseph Raz, “The Inner Logic of the Law,” in Ethics in the Public Domain, rev. ed. (Oxford: Oxford University
Press, 1995), 238 at 241ff.
25
This complex legal power is rightly emphasized (although sometimes wrongly overemphasized) by John
Goldberg and Ben Zipursky in their joint and solo writings on tort law, a mature conspectus of which is: John
C.P. Goldberg and Benjamin C. Zipursky, “Torts as Wrongs,” 88 Tex. L. Rev. 917 (2010).
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 341

confronted, in addition, is whether the law should give it this kind of recognition—the
tort law kind of recognition—complete with its generous terms for power-sharing and
cost-sharing as between the aggrieved party and the legal system. That question is a
question of distributive justice. The law is selecting some people for a measure of
official support in their personal affairs that most other clients of the welfare state can
only dream of. Even among those who have been wronged, not all can possibly enjoy
this level of support in putting things right, and the question is always live of who
should be the privileged ones who qualify for it. That mirrors the question we ask
when we ask who should get the latest cancer drugs, or who should get sent on the
most effective offender-rehabilitation courses. Even for those who say that “user pays”
is the best (most just) solution, there is no doubt that such problems of how to allocate
scarce medical and social services fall under the heading of distributive justice.26 Why
doubt that the same is true regarding the allocation of scarce judicial services? The
difference lies not in the character of the allocation. The difference lies in what is being
allocated. What is being allocated, when we allocate rights to issue proceedings in torts,
is access to a special apparatus for the doing of justice in another form, viz. corrective
justice. It is one question whether corrective justice is being dispensed in the courts. It
is a further question whether the system justly distributes access to the corrective
justice it dispenses.
My locating the courts among the diverse public services of the welfare state might
be taken to suggest that I have come back round, in spite of myself, to focusing on the
overall “distribution of the benefits and burdens of economic activity among individ-
uals in a society.”27 You may think, in particular, that I am assuming a scarcity of
publicly provided legal services that arises only from the cost of supplying them in
competition with other calls on public funds. But that is a simplification. The potential
economic strain of funding the courts and access to them clearly cannot be ignored.
But even if money were no object, we should still be sparing in handing out legal rights
because of the cultural costs of excessive juridification, i.e., of turning too much of our
lives over to the law.28 The rule of law favors access to justice, but it also favors the
existence of non-juridified space in which people can readily steer clear of the law.29
So, even if there were plentiful funding, we would still be forced by the rule of law to
confront the question of how to distribute legal rights and duties as between many
potential plaintiffs and defendants, respect for whose moral rights and duties could

26
Occasionally talk of “distributive justice” is taken to conceal a leaning towards redistribution. But it is
better to follow Nozick in treating redistributive and non-redistributive norms as competing within the space
of distributive justice: Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), 153ff. See my
discussion in Gardner, “What is Tort Law For? Part 1” (note 6) at 11–13.
27
Lamont and Favor, “Distributive Justice” (note 1).
28
For valuable critical reflections on juridification, see the essays in Gunther Teubner (ed.), Juridification of
Social Spheres (Berlin: de Gruyter, 1987).
29
I discuss other implications for the law of torts of this desideratum of the rule of law in John Gardner,
“Some Rule-of-Law Anxieties about Strict Liability in Private Law,” in Lisa Austin and Dennis Klimchuk (eds.),
Private Law and the Rule of Law (Oxford: Oxford Univeristy Press, 2014).
342 JOHN GARDNER

potentially be well-served by their institutionalization, and the institutionalization of

which would not be ruled out on other grounds,30 but not all of whose moral rights
and duties it would be advisable, even if it were humanly possible, to institutionalize
simultaneously.
It might be thought that such a distributive question can properly be live for
legislators, but not for judges. Judges in tort cases should only do “justice between
the parties.” But where the rule of law prevails, doing justice between the parties, in the
relevant sense, cannot but entail consideration of whether the plaintiff belongs to a
class of people who should enjoy a right to proceed in tort against the defendant.
Under the rule of law judges must decide cases according to law, which means
(minimally) that they must not separate the rule from the ruling, either by declaring
what the rule is or will henceforth be while declining to apply it to the case in hand, or
by denying that there is a rule.31 It follows that no judge may rule in favor of any
plaintiff except by locating the plaintiff within a class of imaginable plaintiffs who
would, according to the judge, be entitled to the same ruling. To determine which class
this is, it is not enough for judges to settle whether the plaintiff has been wronged by
the defendant and whether corrective justice could now be done. They must also
confront the question of whether corrective justice should be done with the aid of tort
law, which is only one possible mechanism among many (indeed, among many found
in the law). And that cannot but be confronted as a distributive question: How do we
dole out the right to proceed in tort among various candidate classes? Do we ask which
class deserves the right most, which has most to gain from it, which is least likely to
abuse the right, which is best-placed to make use of it for the public good, which needs
it most, which can have it at least cost to the upholding of legal certainty, or what?
In England, to be sure, it took primary legislation to confer adequate tort law rights
on trespassers injured by the negligence of the occupiers of land.32 But there would
have been nothing improper or even unusual about such rights having been developed
at common law by the following line of thought: The fact that trespassers are
themselves tortfeasors has been disproportionately visited upon them by too often
denying them causes of action in tort for wrongs committed against them by occu-
piers; so, some of the cases denying them such causes of action should be overruled or
distinguished, reining in an over-harsh version of the ex turpi causa non oritur actio
defense. Notice that it was a closely analogous line of thought by which the modern
law of negligence came of age in Donoghue v Stevenson:33 The fact that retail con-
sumers are not in privity of contract with manufacturers and wholesalers (argued the
majority in the House of Lords) has been disproportionately visited upon them by too
often denying them causes of action in tort for wrongs committed against them by

30
I am thinking of the harm principle (a norm of toleration, not of justice) and similar moral norms that
place independent restrictions on the use of law as an instrument for improving moral conformity. See Joseph
Raz, “Autonomy, Toleration, and the Harm Principle,” in Ruth Gavison (ed.), Issues in Contemporary Legal
Philosophy (Oxford: Clarendon Press, 1987).
31
I discuss this requirement further in Gardner, Law as a Leap of Faith (note 12), Chapter 8.
32 33
Occupiers’ Liability Act 1984. Donoghue v Stevenson, [1932] AC 562.
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 343

manufacturers and wholesalers; some of the cases denying them such causes of action
should thus be overruled or distinguished, reining in an over-harsh version of the
doctrine of privity of contract. The “disproportion” in both examples is clearly a
distributive one. It means something like “comparatively undeserved.” Plaintiffs of a
certain class (trespassers and strangers to contracts) have been given undeservedly
little in the distribution of causes of action in tort as compared with plaintiffs of other
classes (lawful visitors and parties to contracts respectively). This shows how, in
attempting corrective justice between the parties in a tort case, judges may also
(often inexplicitly) be attempting distributive justice between classes of parties in the
allocation of access to tort law’s apparatus for doing corrective justice. It shows,
indeed, that whenever there is a question before them of which acts ought to be
classed as torts, judges cannot avoid attempting such distributive justice.
This reveals that the scarcity of law as a public resource is not the only reason, and
may not even be the most important reason, why attempts at distributive justice are
inevitable in the law of torts. Attempts at distributive justice are also inevitable because
the law of torts is part of the common law, and, in the common law, the standard way for
judges to develop the law is by making comparisons between different classes of plaintiffs
and between different classes of defendants. “People of class P have an established cause
of action against people of class D,” argues a plaintiff. “Is the difference between people of
class P and people like me, call us class P´, really so great that we should have no cause of
action against people of class D at all?” Or: “People like me, in class P, have an established
cause of action against people of class D. Is the difference between people of class D and
people of class D´ really so great that we should have no cause of action against people of
class D´ at all?”
It may have been the pervasiveness of this kind of argument in the common law that
led H.L.A. Hart to the famous but mistaken thesis that “we have, in the bare notion of
applying a general rule of law, the germ at least of justice.”34 In the cases I am thinking
of, the judges are certainly generalizing, but they are not merely applying a rule. They
are forging a new rule by generalizing from an existing one and doing so on the ground
that, in their view, it would be unjust for one class of persons to enjoy recourse to tort
law (recourse taken to be justified) that is denied to a neighboring class. Legislatures,
unlike judges, are not constrained to work in this way. They may create new causes of
action without building on existing ones. But that does not show that they are not
attempting a just distribution of tort law rights and duties. It only shows that not all
distributors proceed by comparing what they are asked to bestow with what has
already been bestowed.
When courts or legislatures recognize new causes of action in tort, or extend
existing causes of action, they are distributing legal rights and duties to new classes
of potential plaintiffs and potential defendants. In the first place, they are distributing

34
H.L.A. Hart, The Concept of Law (Oxford: Clarendon Press, 1961), 202. For a thorough critique of this
remark, see David Lyons, “On Formal Justice,” 58 Cornell L. Rev. 833 (1972). See also Gardner, Law as a Leap
of Faith (note 13), Chapter 10.
344 JOHN GARDNER

new primary legal duties, breaches of which will count not just as legally recognized
wrongs but more specifically as torts, violating not just any legal rights but legal rights
in the law of torts. In placing these legal rights in the law of torts—in making the
breaches of primary duty tortious—courts and legislatures also unavoidably distribute
associated secondary duties. These are legal duties of corrective justice—to be more
exact, duties of repair—that arise from breach of the primary legal duties and are owed
to the same rightholder. As already noted, these duties are bundled with generous
powers on the part of the rightholder to concretize and enforce them through the
courts. Nevertheless, what is distributed remains something irreducibly corrective.
And that lends a certain explanatory priority to corrective over distributive justice in
what Weinrib might call the “immanent rationality” of tort law.35 We need to grasp
the essentially corrective ingredient in tort law in order to grasp the whole package
deal, the structured normative arrangement, that tort law is in the distinctive
business of distributing—a deal, indeed, that would not even be up for distribution
without tort law.
Calabresi and Melamed are sometimes remembered as having denied this. They are
remembered as having assigned to the law the task of distributing only “the set of
initial entitlements,”36 understood (in the context and idiom of tort law) as the set
of primary duties, breach of which constitutes a tort. The secondary (corrective) duties
of tort law are sometimes imagined to have been, for Calabresi and Melamed,
automatic implications of the initial entitlements. In other words, Calabresi and
Melamed are often associated with something like my “continuity thesis.” Doing
corrective justice is responding belatedly to the reasons that one should have
responded to in the first place, in not committing the tort. In combination with the
Calabresi and Melamed view that these first-place reasons are reasons for the alloca-
tion of “initial entitlements,” this lures one into what might be called the “deflationary
view” of corrective justice. Corrective justice is revealed not to be a distinct form of
justice, but merely distributive justice redone following a disruptive intervention (that
was not itself licensed by distributive justice). And from here it is a short step to what,
in Part 1, I called “a perennial student objection to tort law,”37 namely that it cannot be
just to restore a distribution that was not itself just.
We should not be so quick to sign up to this perennial student objection to tort law,
for the deflationary view of corrective justice is false. It is false in raw morality, as we
already saw, because in raw morality there is no question of distributing initial
entitlements and so no question of restoring their initial distribution. And it is also
false, we can now add, in the law. In the law, quite apart from the question of which
primary duties (= which “initial entitlements”) to recognize, there is the further
question of how to deal with the breach of those legally recognized primary duties

35
Weinrib, The Idea of Private Law (note 4) at 206.
36
Calabresi and Melamed, “Property Rules, Liability Rules, and Inalienability” (note 17) at 1097.
37
Gardner, “What is Tort Law For? Part 1” (note 6) at 15. For a properly worked-out version of the
objection, see James Nickel, “Justice in Compensation,” 18 Wm. & Mary L. Rev. 379 (1976), 381ff.
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 345

and, in particular, whether to institutionalize a secondary duty of corrective justice. If

one locates the primary duties in the law of torts, as we saw, one inevitably includes a
secondary duty of corrective justice as a part of the “tort law” package. But one need
not use the law of torts, and, more generally, one need not grant a secondary duty of
corrective justice. One may choose a different (non-corrective) legal response to some
legal wrongs or, indeed, no legal response at all. So, there are always two questions for
the court or legislature: which “initial entitlements” to include in the law and how to
respond—correctively or otherwise—to their violation.
It follows that the corrective duties are not mere automatic implications of the initial
entitlements, whether by the logic of the continuity thesis or otherwise. They are
distinct entitlements that also need to be distributed by the law. And this, indeed, is
what Calabresi and Melamed say. They emphasize “the [twin] problems of selecting
the initial entitlements and the modes of protecting these entitlements.”38 Indeed, one
way to read their famous article is as a critique of the view that corrective remedies are
automatically in order, i.e., that they just follow without further ado from the disrup-
tion of initial entitlements. There is always the further question, when initial entitle-
ments are disrupted, of whether and how the law should respond to the disruption.
As Calabresi and Melamed express the question:
Why . . . cannot society limit itself to the property rule? To do this it would need only to
protect and enforce the initial entitlements from all attacks, perhaps through criminal
sanctions, and to enforce voluntary contracts for their transfer. Why do we need liability
rules at all?39

Calabresi and Melamed have their own answer. They see many advantages in choosing
the tort law route for protecting property rights. It follows that they reject the idea that
there are no other routes. In protecting property rights by law, they notice, having a
tort of trespass is but one option among many. Indeed, even when it comes to
recognizing duties of non-trespass in the law, having a tort of trespass is but one
option among many. The tort lawyer’s slogan “ubi ius, ibi remedium” may serve well as
a recommendation, but it bespeaks no rational inevitability.40
Does all this cast doubt on the relevance to the law of torts of the continuity thesis,
which does bespeak a kind of rational inevitability? Are we quietly backing away from
the conclusions of Part 1? Not at all. It is one question why, as rational beings, we
would want to do corrective justice, and want to see corrective justice done, and the
doing of it supported. The continuity thesis helps us to see why. But it is another
question when and how, if at all, we should actually support the doing of corrective
justice. Here, we may have cause to reflect on the desirability of using the law (or other
similar institutional systems) as a method of support. And, as I have explained, we

38
Calabresi and Melamed, “Property Rules, Liability Rules, and Inalienability” (note 17) at 1089 (emphasis
added).
39
Calabresi and Melamed, “Property Rules, Liability Rules, and Inalienability” (note 17) at 1106.
40
On the interpretation of the slogan, see Ted Sampsell-Jones, “The Myth of Ashby v White,” 8 University of
St. Thomas L. J. 40 (2010).
346 JOHN GARDNER

cannot but encounter that question of desirability as, at least partly, a question of
distributive justice. Legal support for the doing of corrective justice is a scarce good.
What Rawls calls “the circumstances of justice” obtain in respect of it.41 Moreover the
courts constantly face the question, in common law systems, of why one class of
plaintiffs should have a cause of action while another nearby class does not, or why one
class of defendants should be suable while another nearby class is not. Which plaintiff-
classes should be supported as against which defendant-classes? Which corrective
justice is to be chosen for legal recognition? That is clearly a distributive problem.

III. Distributive Justice between the Parties

Not all questions of justice arising in tort cases are questions of justice between the
parties to those cases. As we have seen, there is also the ever-present question of how to
distribute, among imaginable classes of potential parties, tort law’s special apparatus
for doing or helping to do justice between them. That, I have claimed, is a question of
distributive justice. This claim still leaves open, however, the possibility that justice
between the parties, the justice that tort law’s special apparatus is there to facilitate, is
always and only corrective. Weinrib argues that it is. As promised in Part 1,42 I will
argue the opposite. Corrective justice is always justice between the parties, but justice
between the parties is not always corrective. Some of it is distributive. And such
“localized” distributive justice (as it is known43) has a key role to play, I will argue,
in the doing of justice between the parties in tort law.
Problems of localized distributive justice arise in raw morality as readily as they do
in the law. Here is an example that could be considered under either heading:
Having lured Prey to a remote abandoned factory, Hunter engineers a situation such that
Prey must blind Hunter if she is to avoid being blinded by Hunter. As Hunter planned, the
two of them are now suspended above a tank filled with an eyesight-destroying chemical.
The only way Prey has to stop Hunter pushing Prey into the tank is for Prey to push
Hunter into the tank instead.

Indeed, Hunter’s plunge will lift Prey clear, and vice versa. Thanks to Hunter there is
now, a scarcity of future eyesight as between the two of them. They are competing,
winner takes all, for the future capacity to see. Who gets to blind whom and thereby
keep her own sight?

41
John Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971), 110ff. The circum-
stances in question are “moderate scarcity” and “conflict of interests.” I should stress that I do not follow Rawls
in thinking, if this is what he thinks, that questions of distributive justice are never forced on us outside these
circumstances. I agree, however, that they are always forced on us in these circumstances. See Gardner, Law as
a Leap of Faith (note 13) at 264–7.
42
Gardner, “What is Tort Law For? Part 1” (note 6) at 12.
43
Following Stephen Perry, “The Moral Foundations of Tort Law,” 77 Iowa L. Rev. 449 (1991), 461.
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 347

The problem has many aspects, and morality gives us more than one way to think
about it. But one way that morality gives us to think about it is as a problem of justice.
Albeit in respect of a single interaction, Hunter has contrived the circumstances of
justice as between herself and Prey. There is no escaping the need for allocation as
between the two of them. So, what form of justice is called for? To borrow Weinrib’s
own criteria of classification, it is clearly not yet a problem about “what the doer of
harm owes to the sufferer of harm.”44 Nobody has yet been harmed. The problem does
not yet have “the shape of corrective justice.”45 Rather, it is a problem having “the
shape of distributive justice,” a problem about how to “divid[e] a benefit or burden
among a group.”46 It matters not that the only possible division of the relevant burden
(loss of sight) as between Prey and Hunter is an all-or-nothing division (because the
burden cannot be shared). Nor does it matter that Prey and Hunter together constitute
a group with only two members. Clearly, neither the size of the group nor the share-
ability of the burden makes a difference to the form of justice that falls to be done.
According to a popular view, justice forbids Hunter from blinding Prey but it does
not forbid Prey from blinding Hunter. And that is so because it was Hunter who made
it the case, by her wrongdoing, that one of them has to be blinded. Here, tweaked to
eliminate some distracting specificity, is Jeff McMahan’s nice formulation of the
relevant distributive norm:
[I]n cases in which a person’s [wrongful] action . . . has made it inevitable that someone
must suffer harm, it is normally permissible, as a matter of justice, to ensure that it is the
[wrongdoer] who is harmed rather than allowing the costs of his wrongful action to be
imposed on the [other(s) on whom they might instead have fallen].47

Let’s call this the “responsibility” norm of distributive justice, meaning thereby to
emphasize the fact that the norm makes a distribution of costs turn on their causal
attribution. The causal element of the responsibility norm, like many other elements of
it, calls for further elaboration, which I will not offer here. I will also play down some
doubts I have about the responsibility norm’s soundness, at least as a raw moral norm.
All I will do here is treat the norm as sound for the purpose of argument.
The main case for doing so is that, even in its indeterminacy, one can readily see the
potential application of the responsibility norm to tort litigation.48 Like Hunter and
Prey, Plaintiff and Defendant are caught up in a zero-sum situation in which each wins
if and only if the other loses. If Plaintiff is in the right, it was Defendant who, by his
wrongdoing, put them in that zero-sum situation. By committing the tort, Defendant
“made it inevitable” that one of them must come out of the conflict a loser. So, one
might well think: that is why it is permissible for Plaintiff to insist on damages, and
why the court is required, if Plaintiff insists, to award them. Tort litigation, one might

44
Weinrib, The Idea of Private Law (note 4) at 73.
45 46
Weinrib, The Idea of Private Law (note 4) at 73. Weinrib, The Idea of Private Law (note 4) at 73.
47
Jeff McMahan, “Self-Defense and the Problem of the Innocent Attacker,” 104 Ethics 252 (1994), 259.
48
As McMahan also notes: “Self-Defense and the Problem of the Innocent Attacker” (note 47) at 279. See
also Perry, “The Moral Foundations of Tort Law” (note 43) at 499.
348 JOHN GARDNER

think, is an occasion for doing localized distributive justice, and the responsibility
norm regulates it, and (if sound) justifies its characteristic features.
The objection to this view is well known. Tort litigation is not conducted in an
abandoned factory. When Hunter confronts Prey, the two of them are cut off from
civilization. There are only two candidates for blinding. But back in ordinary life, there
are many other potential loss-bearers around apart from Plaintiff and Defendant. By
handing the problem over to litigation, we have taken a preemptive step to localize, as
opposed to socializing, the problem. We are turning what might have been a zero-sum
interaction among many (played out, say, through general taxation and social insur-
ance) into a zero-sum interaction between just two litigants. Why? There is nothing in
the relevant distributive norm to explain it. As Stephen Perry says:
the localized nature of the distributive scheme is arbitrary and unjustified; there is no basis
for limiting the group of potential loss-bearers to the injurer and the victim alone.49

I hasten to add that Perry does not mean that there is no basis for the localization full
stop. He only means that there is no basis, in whatever norm we use to do localized
distributive justice, to explain why it is only localized distributive justice that we are
doing. Perhaps, the responsibility norm, applied across the wider population, would
still end up casting Defendant as the loser and Plaintiff as the winner. The question is
only why we are not applying it across the wider population, but instead applying it as
between Plaintiff and Defendant alone. The answer, as Perry notes, can only be for
some “extraneous reasons,”50 i.e., reasons not provided by the responsibility norm
itself.
When we hear talk of “extraneous reasons,” our Weinribian alarm bells may ring.
We may imagine economists queuing up to show that localizing a conflict, even with
generous social support through the judicial system, is cheaper than socializing it
through the taxation and welfare benefit systems.51 And maybe it is. And, surely, that
would be relevant to our thinking about institutional arrangements for responding
to the conflict. But it is premature to be thinking like that. We are leaping over the
most obvious “extraneous reasons” that would support the localizing of the problem
of distributive justice before us, namely: reasons to do (and to support the doing of)
corrective justice. These reasons are extraneous, not of course relative to the law of
torts, but relative to the responsibility norm, which is a norm of distributive justice.
They are the same reasons why Defendant should not have done what he did to
Plaintiff, reasons why his doing it counted as a wrong against Plaintiff. Since the wrong
has been done, at least some of these reasons have been left unsatisfied. They now exert
their force as reasons for Defendant to do right by Plaintiff in some fallback way, as
well as can now be achieved. They are reasons for corrective justice to be done, reasons
inter alia to pay reparative damages, and they explain (together with some institutional

49
Perry, “The Moral Foundations of Tort Law” (note 43) at 471.
50
Perry, “The Moral Foundations of Tort Law” (note 43) at 468.
51
See, e.g., Richard Posner, “A Theory of Negligence Law,” 1 Journal of Legal Studies 29 (1972), 48–9.
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 349

facts) why Plaintiff and Defendant are caught up in the zero-sum conflict that is tort
litigation. So, they also explain why we might be confronted with questions of
distributive justice that are already pre-localized and already assume the context of a
bilateral zero-sum conflict. To quote Cane again, “corrective justice provides the
structure of tort law within which distributive justice operates.”52 Once again, it is
corrective justice that has the explanatory priority.
The responsibility of wrongdoers is, surely, already central to corrective justice. So,
one might wonder why, once we are admittedly doing corrective justice between the
parties, we should be drawn into applying the responsibility norm of distributive
justice between the parties as well, albeit only as a subsidiary concern. Isn’t that just
duplicative? Far from it. One important reason why is that, unlike the zero-sum
conflict over the distribution of future eyesight between Hunter and Prey, the zero-
sum conflict over the distribution of losses between Plaintiff and Defendant is not
necessarily a winner-takes-all conflict. Once we are proposing to deal with the losses
by an award of money damages—and we always are in the law of torts—the losses can
be shared between Defendant and Plaintiff. Determining how to effect such sharing in
particular cases, it seems to me, is the main function of several doctrines of the law of
torts: notably, those of mitigation and remoteness of damage and the modernized law
of contributory negligence. Interpreted as devices to effect loss-sharing, these doctrines
lack a corrective-justice rationale. Corrective justice, as Aristotle explains, knows only
addition and subtraction. It has no room for division, which is the business of
distributive justice.53 True, our responsibility norm, formulated by McMahan with
all-or-nothing conflicts in mind, does not yet attend to cases of shareable loss.
Tweaking it to do so would, however, make it even more conspicuously a norm of
distributive justice—a norm for “dividing a benefit or burden among a group.”54 And
it would, therefore, reveal even more clearly the distinct role that considerations of
localized distributive justice have to play in tort adjudication.
You may say that I have not shown that these considerations should be playing any
role in tort adjudication. You may say that if mitigation and remoteness of damage and
contributory negligence are doctrines that exist to do distributive justice between the
parties, so much the worse for them. They are alien doctrines that do not belong in tort
law. But it seems to me, on the contrary, that tort law cannot properly abdicate
responsibility for tackling the local conflicts that are turned into zero-sum, and
hence distributive, conflicts by tort law itself. Tort law offers an apparatus for the
doing of corrective justice—litigation—that creates the circumstances of localized
distributive justice. The courts may as a consequence be faced with collisions, perhaps
endemic collisions, between sound corrective norms and sound distributive norms.

52
Cane, “Distributive Justice in Tort Law” (note 12).
53
EN 1131b12–15, 1132a1–6. For discussion, see Gardner, “What is Tort Law For? Part 1” (note 6) at 9.
54
What needs to be added, most conspicuously, is some kind of proportionality condition. As stated by
McMahan, the norm would allow us to dump everything on the wrongdoer, e.g., to ruin him in order to avoid
each of us paying a penny. For discussion, see McMahan, “Self-Defense and the Problem of the Innocent
Attacker,” (note 47) at 261–4.
350 JOHN GARDNER

Different legal systems, and different judges and courts of the same legal system, may
reasonably deal with the collisions in different ways. But all must grapple with the fact
that the collisions are theirs to deal with, built into the fabric of the law of torts, not
imposed upon it from without, with no other place to go. That is because, as John
Finnis explains:
[W]hether the subject-matter of [an] act of adjudication be a problem of distributive or
[corrective] justice, the act of adjudication itself is always a matter for distributive justice.
For the submission of an issue to the judge itself creates a kind of common subject-matter,
the lis inter partes, which must be allocated between parties, the gain of one party being the
loss of the other.55

IV. On “Risk-Distributive” Justice

I have explored two endogenous distributive aspects of tort law. Both have sometimes
been treated as raising problems about the distribution of risk. Thinking of the
problems that I traversed in Section III above, Tony Honoré writes:
[T]he person who, in a situation of uncertainty, has a degree of control over how it will
turn out, and who stands to gain if it goes in his favour, must bear the risk that it will turn
out to harm another.56

And thinking of my Section II topic, he says, influentially:57

To justify [the social or legal upholding of] corrective justice involves appealing at a certain
stage to the just distribution of risk in a society. Corrective justice is a genuine form of
justice only [when and] because the just distribution of risks requires people of full
capacity to bear the risk of being held responsible for harming others by their conduct.58

I share these thoughts, but I resist the framing of them as thoughts about risk. Let me
end by explaining why, focusing my attention on the second of the two Honoré
quotations above.
There is no doubt that when the law of torts is effective in its role as an apparatus for
the doing of corrective justice—as it has to be, if it is to be justified59—it has an impact

55
Finnis, Natural Law and Natural Rights (note 22) at 179. Finnis says “commutative” where I have
inserted “corrective.” See also Gardner, Law as a Leap of Faith (note 13), Chapter 10.
56
Tony Honoré, “The Morality of Tort Law: Questions and Answers,” in Responsibility and Fault (Oxford:
Hart Publishing, 1999), 81.
57
Among those conspicuously influenced by Honoré on this point are Arthur Ripstein, “Private Law and
Private Narratives,” in Peter Cane and John Gardner (eds.), Relating to Responsibility (Oxford: Hart Publish-
ing, 1981) and Jenny Steele, Risks and Legal Theory (Oxford: Hart Publishing, 2004), 87ff.
58
Honoré, “The Morality of Tort Law” (note 56) at 80. I add the words in the first set of square brackets
because Honoré makes clear that he is not thinking of the raw morality of corrective justice so much as its social
and legal implementations. I add the words in the second set of square brackets to reduce the length of the
quotation, which in the original goes on to endorse a concrete “when” proposal that need not detain us here.
59
Gardner, “What is Tort Law For? Part 1” (note 6) at 17–22.
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 351

on the social distribution of risks. And that impact itself calls for justification, which
inevitably raises questions of distributive justice. But, before we just let that statement
stand, we need to be careful to distinguish various impacts that might be described as
impacts on the “social distribution of risk.” There is (1) tort law’s impact on the way in
which people create and avoid risks to themselves and others in their everyday
activities, in the light of any legal risk to themselves that they thereby create or
avoid. And, then, there is (2) the legal risk itself. The expression “the legal risk” in
turn can be unpacked in at least two ways. There is (2a) the risk that one will not have
the law on one’s side, and there is (2b) the risk that one will lose in litigation (or
litigation-averting negotiation) to someone who has the law on their side.
When judges say that potential tort defendants of some class must “bear the risk of
loss”60 as against potential tort plaintiffs of some class they are not, one hopes,
claiming for themselves the magical ability to make it the case that, when tortious
interactions occur between people of these two classes, the person in the defendant
class will henceforth be the only one who ever gets hurt. They are not affecting to
neutralize the type (1) risks faced by potential plaintiffs, let alone to turn the same risks
back against the potential defendants. Imagine a judge who says: “The risk is hereby
shifted. From now on, people who fall to their deaths down open mineshafts will only
ever fall to their deaths down their own open mineshafts; and from now on people
injured by bad driving will only ever be injured by their own bad driving.” That is just
plain silly. The courts can undoubtedly have an effect of some kind on the distribution
of type (1) risks, both by altering incentives to create them and by arranging redress
when they materialize (which can never be quite the same as their not having
materialized to begin with).61
But the only way in which the courts can alter the incentives or arrange the redress,
so as to have some impact on the distribution of type (1) risks, is by redistributing type
(2) risks. Even the type (2b) legal risks they can only really adjust by redistributing the
type (2a) legal risks associated with the activities of the defendant class. They can give a
certain class of plaintiffs the legal right to reparative damages in respect of certain
losses suffered at the hands of a certain class of defendants, but they can do little to
ensure that the right is exercised (more precisely, that the power to seek an award of
those damages is exercised in cases in which there is a right to the award). So, the only
risks of loss, which are such that the courts can literally make the potential defendant
or potential plaintiff bear them, are the type (2a) legal risks, also known (more
straightforwardly) as the legal rights and duties of the classes of plaintiffs and defend-
ants involved. Making the plaintiff class bear the risk simply means denying members
of that class a legal right to corrective justice, tort law style, as against those in the
defendant class; making the defendant class bear the legal risk means giving the
plaintiff class that legal right.

60
Honoré, “The Morality of Tort Law” (note 56) at 79.
61
Gardner, “What is Tort Law For? Part 1” (note 6) at 34–5.
352 JOHN GARDNER

There is a puzzle about the use of the language of risk, here. When judges say that a
certain class of people must “bear the risk of loss” they are usually determining the
distribution of the relevant legal rights and duties. So you may wonder why they
present that distribution as merely a legal risk rather than (now) a legal certainty. The
answer is that they are typically thinking of how agents (potential plaintiffs and
defendants) will factor the law into their thinking before they engage in the activities
that give rise to (or don’t give rise to) the relevant legal rights and duties. Let it be as
certain as you like, at that point, whether a potential plaintiff will have a right to
corrective justice, tort law style, if a certain loss materializes; there remains the
uncertainty as to whether the loss will indeed materialize. That is where the element
of legal risk comes in. The distribution of this legal risk as between different classes of
people is the same distribution—now viewed, ex ante, as one factor in thinking about
one’s choice of future activities—as the distribution of legal rights and duties as
between those classes of people. Nothing more, nothing less.
We can now glean quite a few reasons why, outside of judicial rhetoric, it is not
revealing and can be misleading to classify this as the “distribution of risk in a society.”
First, replacing talk of the social distribution of legal rights and duties with talk of
the social distribution of risks is not perspicuous. It adds an extra layer of complication
to an already complicated subject matter. Eventually, as we just witnessed, the extra
complications need to be analyzed out, and the more basic discourse of legal rights and
duties has to be restored in order to make clear which risks, exactly, we are talking
about.
Second, attributing to tort law an ability to redistribute risks of loss invites a
hubristic and almost comical view of the influence of law. Even when tort law is
effective enough to have a role in our lives, tort law’s distribution of the right to
corrective justice has, at best, a highly contingent connection with the risk of suffering
tortious loss (never mind loss more generally) that we face as we go about our daily
lives. The reception of tort law into the thinking of risk-takers is doubtless patchy. And
many other risk-affecting mechanisms and forces are also in play. Those who walk
home late at night may be at extra risk of a tortious injury; but fortunately there are
often insurers or public agencies which are prepared to cover the associated healthcare
costs without regard to whether the person who, in tort law, will “bear the risk of loss,”
can be found. So tort law is not the only (and in some societies may not even be the
main) institutional distributor of the risk of tortious losses, never mind losses more
generally.
Third, risk-distributive justice is not easily kept distinct from the rest of distributive
justice. Risk, at its most general, is simply the probability of something unwelcome
coming to pass. Since the fact that something welcome does not come to pass is itself
unwelcome, the scope of “risk-distributive justice” seems to be, in one natural inter-
pretation, indistinguishable from the scope of distributive justice in its entirety. All
welcome and unwelcome things that are capable of being distributed are capable of
being justly or unjustly distributed. Presumably, we want to work out first (“stage 1”)
who should be allocated how much of which of these things—and by whom. The
 WHAT IS TORT LAW FOR ? PART 2. THE PLACE OF DISTRIBUTIVE JUSTICE 353

probabilities of those people actually getting these allocations are, presumably, to be

factored in later (“stage 2”), when we try to implement our stage 1 conclusions. Why
begin by foregrounding the probability question by thinking of the thing to be
allocated as the probability of some other unwelcome allocation? Why introduce the
uncertainty of distributive success in stage 1, when what we are trying to establish at
that stage is what would count as success? True, the uncertainty of something
unwelcome coming to pass might sometimes be unwelcome in itself, an extra curse.
But how would one even begin to think about distributing the uncertainty in isolation
from the distribution of the things to which it attaches? It is hard to think of risk as
such (as distinct from particular risks) as something that is up for distribution. At any
rate, it is hard to think that way with any clarity.
Talk of “risk” may seem to offer a common currency into which we can convert all
the various unwelcome possibilities we face in life. In that respect, it fosters a
dangerous illusion and encourages reductive theorizing. It conceals the irreducible
diversity of things we should care about and, therefore, of things we should care about
the allocation of. Our discussion in this essay of the distribution of rights and duties
through the law of torts was intended to bring this out. When a wrong is recognized
as a tort, as we saw, the law thereby effects a new distribution of the legal right
to corrective justice in its tort law form. That means a new distribution of the power
to commence proceedings, a new distribution of the court’s duty to award damages to
successful plaintiffs, and a new distribution of the duty to pay such damages on the
part of defendants. At least these three valuable things are being distributed, in one fell
swoop, as incidents of a single legal right. Each is valuable in its own way.
It is arguable that sometimes the three should be split up and distributed by
different mechanisms and to different people. Sometimes, indeed, they are. But in
tort law they are distributed as a package deal—as aspects of a single specialized
institutional apparatus for the doing of corrective justice. Thinking of this distribution
as a distribution of risk—to come to our fourth and final objection to Honoré’s
characterization—obscures the moral distinctiveness of what is being distributed. It
obscures the fact that what is being distributed is, in Weinrib’s words, “the bipolar link
between the parties that characterizes the doctrines and institutions of private law.”62
And, consequently, it obscures the fact that when the distribution of that bipolar link is
effected, yet further localized problems of distributive justice between the parties to it
are created. Talk of “risk-distributive justice” obscures the fact, in short, that tort law,
understood as an institution of corrective justice, does not merely meddle in the
various subject matters or “currencies”63 of distributive justice, but also, in more
than one way, helps to constitute them.

62
Weinrib, The Idea of Private Law (note 4) at 76.
63
It was G.A. Cohen who coined the apt expression “the currency of justice,” although he also gave succor,
alas, to the view that (social) distributive justice has only one currency, into which every distribuand must be
converted. See G.A. Cohen, “On the Currency of Egalitarian Justice,” 99 Ethics 906 (1989).
 17
Tort Law and Distributive Justice
Hanoch Sheinman*

The judge is intended to be a sort of living embodiment of the just.

The just in distributions must accord with some sort of worth but what they call
worth is not the same.
The judge restores equality, as though a line had been cut into unequal parts, and he
removed from the larger part the amount by which it exceeds the half of the line, and
added this amount to the smaller part. And when the whole has been halved, then they
say that each person has what is properly his own, when he has got an equal share.
Aristotle

I. Introduction
This chapter discusses the place of distributive justice in the normative theory of tort
law, although much of the discussion will apply to civil law generally. My goal is not to
discuss the distributively just and unjust effects of tort law, but to identify and describe
the distributive justice of tort law in particular, that morality of distribution that
cannot fail to apply to this area of law, and therefore bear on its justification.1 But in
characterizing tort law’s distributive justice, the chapter also characterizes its correct-
ive justice. Indeed, the chapter’s central claim is that tort law’s distributive justice just
is its corrective justice. Thus the chapter argues that tort law’s justice is fully distribu-
tive as well as fully corrective.

* I would like to thank David Enoch, John Goldberg, Greg Keating, and John Oberdiek for their helpful
comments during the presentation of an early version of this chapter. Many thanks to my colleagues Ori
Aronson, Jacob Nussim and Ziv Bohorer for pressing me to clarify the discussion. For useful questions thanks
also to Michal Alberstein, Tsilly Dagan, Yoed Halbersberg, and Arie Reich. My debt to John Gardner should be
clear from the discussion.
1
The problem is similar to the one addressed in John Gardner, “What is Tort Law For? Part 2. The Place of
Distributive Justice,” Chapter 16 of this volume, and in Peter Cane, “Distributive Justice in Tort Law,” 4 New
Zealand L. Rev. 401 (2001), 413.
 TORT LAW AND DISTRIBUTIVE JUSTICE 355

In arguing for these claims, the chapter offers an unorthodox account of the Aristo-
telian distinction between the two forms of justice. The chapter does not reject the very
distinction, but it does reject the common assumption that the forms are mutually
exclusive. The thought is that corrective justice is a distinct category of distributive
justice, and that tort law’s justice is a distinctly corrective principle of distributive
justice.
The chapter lends some support and shape to the claim that corrective justice enjoys
certain interesting priorities over distributive justice in relation to tort law. First is the
broadly conceptual claim that tort law’s justice can be said to be corrective but not
distributive justice as such. More important is the normative-ethical claim that
corrective justice enjoys some priority in the justification of tort law, not so much
over distributive justice per se as over non-corrective principles of distribution. As an
institution whose job it is to do corrective justice, it is hard to see how tort law can
escape the demands of corrective justice. It is not clear that the same holds for any
particular non-corrective principle of distribution. To each according to her needs,
abilites, or virtues, for example.
The chapter also raises the problem of reconciling the priority of corrective justice
with the platitude that our modern tort law is not simply a corrective justice system but
rather a mixed system that involves non-corrective redistributive mechanisms, such as
liability insurance. The discussion resists one natural way to achieve such reconcili-
ation, recommends another, and illustrates its possible implications.
The emerging picture is neither as revisionary nor as traditional as it might
seem. Much of the claims are old wine, new wineskins. This much is true of the
claim that tort law’s justice is distinctly corrective. Here the discussion will mostly
exploit the familiar insights of others. The emerging picture goes beyond these
insights by insisting on and characterizing the genuinely distributive nature of tort
law’s justice. Here my discussion breaks with some interesting claims. Most
obviously, it conflicts with the claim that tort law faces no special problems of
distribution and has no special principles of distributive justice. But it also conflicts with
the weaker claim that tort law’s distributive justice is different from, and derivative of,
its first and corrective justice, that tort law’s special problems or principles of distribu-
tion are incidental to, or derivative of, its special problems or principles of correction.
Section II (Background) provides some context for the discussion that follows.
Section III (The Aristotelian Distinction) discusses Aristotle’s classic account and
explains the need for reconstruction. Section IV (An Alternative Account) offers
such reconstruction. Section V (Reparative Justice in Tort Law) uses the account to
show how tort law’s distinct principle of justice is fully distributive as well as fully
corrective. Section VI (The Priority of Corrective Justice) says how the discussion can
and cannot support the priority of corrective over distributive justice in relation to tort
law. Section VII (Similar Views) compares the account with recent views about tort
law’s distributive justice. Section VIII (The Mixed Tort Law) entertains the possibility
of reconciling the priority of corrective justice in torts with non-corrective redistribu-
tive mechanisms, such as liability insurance.
356 HANOCH SHEINMAN

II. Background
Aristotle famously distinguished between distributive and corrective justice. In gen-
eral, both concern the allocation of goods among people, but while distributive justice
requires “geometric” allocation in accord with the relative merits of the parties,
corrective justice requires “arithmetic” allocation back relative to some past wrongful
interaction. Aristotle presents the two categories (“forms”) of justice as mutually
exclusive. A principle that belongs to one category cannot also belong to the other.
Aristotle’s discussion of the forms of justice has become tremendously influential in
the history of ethics.2 That it has proved particularly influential in legal theory is not
surprising. Aristotle clearly thought of justice as particularly applicable in the legal
context, identifying the judge as its personal representative. More to the point,
contemporary theorists use the Aristotelian distinction to make interesting claims
about civil law, including claims about the relative place of distributive and corrective
justice in the constitution, foundation, or justification of tort law.
It is often argued that corrective justice is the form or foundation of tort law, and
that the place of distributive justice is at most subordinate.3 This view implies the
somewhat weaker view that tort law’s first principle of justice—in short, tort law’s
justice—is corrective rather than distributive. And it is sometimes argued that the
priority goes the other way, that tort law’s foundation or foundational justice is
distributive rather than corrective.4
Proponents of such views seem to share the Aristotelian assumption that corrective
and distributive justice are mutually exclusive categories of justice that do not intersect

2
See, e.g., Izhak Englard, Corrective and Distributive Justice: From Aristotle to Modern Times (Oxford:
Oxford University Press, 2009).
3
See, e.g., Ernest Weinrib, The Idea of Private Law (Cambridge, MA: Harvard University Press, 1996). See
e.g. the claim that “[c]orrective justice is the form of the private law relationship” (p. 75); cf. also Corrective
Justice (Oxford: Oxford University Press, 2012). See e.g. the claim that “[c]orrective justice is the term given to
the relational structure of reasoning in private law” (p. 2). See also Martin Stone, “The Significance of Doing
and Suffering,” in Gerald Postema (ed.), Philosophy and the Law of Torts (Cambridge: Cambridge University
Press, 2001), 131–82. See e.g. the claim that corrective justice “can serve to identify the aim of tort law and thus
provide a way of grasping its practical unity” (p. 133). Cf. also Jules Coleman, The Practice of Principle (Oxford,
2001), e.g. the claim that “tort law is best explained by corrective justice” (p. 9) and Stephen Perry, “On the
Relationship Between Corrective and Distributive Justice,” in Jeremy Horder (ed.) Oxford Essays in Jurispru-
dence: Fourth Series (Oxford: Oxford University Press, 2000), 238–62, e.g. the comment that corrective justice
“constitutes the normative foundation of tort law” (p. 262). Cf. also John Goldberg and Benjamin Zipursky,
“Tort Law and Responsibility” (Chapter 1 of this volume), who understand tort law’s main moral principle in
terms of responsibility rather than corrective justice. It is also reasonable to maintain that the moral
responsibility at issue in tort law is responsibility in corrective justice.
4
See, e.g., Hanoch Dagan, “The Distributive Foundations of Corrective Justice,” 98 Mich. L. Rev. 138
(1999); Gregory C. Keating, “Distributive and Corrective Justice in the Tort Law of Accidents,” 74 S. Cal.
L. Rev. 193 (2000). See also Jeremy Waldron, “Moments of Carelessness and Massive Loss,” in David G. Owen
(ed.), The Philosophical Foundations of Tort Law (Oxford: Oxford University Press, 1997), 387–408. Cf. also
Tsachi Karen-Paz, “Egalitarianism as Justification: Why and How Should Egalitarian Considerations Reshape
the Standard of Care in Negligence Law?” 4 Theoretical Inq. L. 275 (2003), Art. 5, which implicitly rejects the
evaluative priority of corrective over distributive justice in tort law.
 TORT LAW AND DISTRIBUTIVE JUSTICE 357

and so that neither category subsumes the other. To the extent that a problem,
principle, or justification is corrective, it is not also distributive, and vice versa.
Again, a principle of one kind can never apply as a principle of the other as well.
This assumption rules out the following possibilities as to the relation between the two
categories, as shown in the following three figures:
What really matters for the discussion is that the assumption rules out the possi-
bility depicted in Figure 17.2, in which corrective justice is a special case of distributive.
But for simplicity, we can focus on the stronger assumption that the forms (kinds,
categories) of justice are mutually exclusive. This assumption rules out the possibility

Distributive Corrective
Justice Justice

Figure 17.1 Distributive and corrective justice intersect

Corrective
Justice

Distributive Justice

Figure 17.2 Distributive justice subsumes corrective

Distributive
Justice

Corrective Justice

Figure 17.3 Corrective justice subsumes distributive

358 HANOCH SHEINMAN

depicted in Figure 17.1 and therefore those depicted in Figures 17.2–17.3 as well.
Taking a leaf from Ernest Weinrib, I will describe it as the assumption about the
categorical nature of the distinction: the two forms of justice are categorically distinct
(mutually exclusive). It follows that neither is a subclass of the other.5
Now it is important to keep the assumption about the categorical nature of the
distinction between principles of distributive and corrective justice separate from a
further view that Weinrib and other friends of the relevant assumption seem to hold,
namely the view that a single institution cannot owe its justification to some combination
of corrective and distributive principles.6 Clearly, the assumption about the categorical
nature of our distinction does not have this questionable implication. On the face of it,
principles belonging to mutually exclusive categories of justice can apply to the same
institution. And a single institution might well depend for its justification on con-
formity to such combination of principles.
Contemporary theorists who accept the possibility of such mixed justification of tort
law share the assumption about the categorical nature of the relevant distinction.
According to Stephen Perry for example “the two principles have different normative
functions. One is concerned with the just distributions of resources, while the other is
concerned with remedying harmful interactions between persons.”7 John Gardner
similarly claims that, while distributive and corrective norms both regulate the allo-
cation of goods, they do so in distinct ways:
Norms of distributive justice regulate the allocation of goods among people together with
the grounds of such allocations (“division”). Norms of corrective justice regulate the
allocation of goods back form one person to another together with the grounds of such
allocations back (“addition and subtraction”).8

Principles of corrective and distributive justice might well bear on the justification of a
single institution, but they can only do so in their own distinct corrective or distribu-
tive ways. Principles of correction (distribution) cannot apply to or justify anything in
the distinctly distributive (corrective) way.
The assumption about the categorical nature of the distinction between the forms of
justice has implications for views about their relative place in the justification of tort
law. In particular, it has implications for the common view that the form, foundation,
or first principle of tort law is corrective justice, as well as for its weaker corollary that
tort law’s first principle of justice—in short, tort law’s justice—is corrective. Given the
assumption, the common view implies that tort law’s first principle or justice is not
also distributive. The twofold implication that tort law’s foundation, or justice, is

5
Weinrib, Private Law at 72 and Corrective Justice at 18–19, 269 (note 3).
6
Cf. e.g. his comment in Private Law that “any given relationship cannot rest on a combination of
corrective and distributive considerations” at 73.
7
Perry, “On the Relationship between Corrective and Distributive Justice” at 262.
8
Gardner, “Corrective Justice” at 17 “What is Tort Law For? The Place of Corrective Justice,” Law and
Philosophy 30 (2011), 1 at 17.
 TORT LAW AND DISTRIBUTIVE JUSTICE 359

corrective justice and not distributive justice warrants the claim that, when it comes to
tort law, corrective justice enjoys priority over distributive justice.
Peter Cane and John Gardner have recently offered plausible versions of this
priority claim, arguing that tort law’s own problems of distributive justice are inci-
dental to its more fundamental problems of corrective justice. Taking the categorical
nature of the distinction for granted, this claim implies that tort law’s most funda-
mental problem of justice is not also distributive.
This chapter offers an alternative account of the special relation between tort law and
distributive justice in which its distinctive problem, function, or principle of distribution is
not incidental to its distinctive problem, function, or principle of correction. Nor is it
more fundamental. Rather, tort law’s most fundamental principle of distribution just is
its most fundamental principle of correction. Tort law’s distributive justice is correct-
ive justice.9
The basic thought behind this claim is that corrective justice is a distributive justice.
The claim is not offered as a stipulation, but as a largely substantive claim about the
kind of corrective justice at issue in civil law in particular; it certainly allows for
the possibility of related nondistributive principles of corrective justice. The idea is that
the corrective justice principles that can be plausibly said to have special application,
or particularly fundamental place, in the working and justification of civil law are
principles of distributive justice. The claim is offered under a particularly natural sense
of “distributive justice,” one that remains agnostic about the relevant principle or
criterion of distribution (let alone about intricate controversies in contemporary
normative ethics). But this, I think, is how it should be.

III. The Aristotelian Distinction

What is the Aristotelian distinction between corrective and distributive justice? I share
the assumption that Aristotle’s original way of drawing the distinction contains much
that makes it particularly helpful for discussing civil law. At the same time, I believe
that Aristotle’s original definitions of distributive and corrective justice are just too
narrow for this purpose and require reconstruction. A major difficulty concerns
Aristotle’s highly specific comparative definition of distributive justice. Another diffi-
culty concerns his arithmetic definition of corrective justice. Once the first problem is
fixed, it is no longer clear why the corrective justice of civil law should not be thought
of as a special case of distributive justice.
In this section, I present a familiar picture of Aristotelian justice and raise some
problems about it. In the next section, I offer a reconstruction. Let me start by clarifying

9
The chapter claims that (1) tort law’s distributive justice is corrective (or is corrective justice) and that (2)
tort law’s corrective justice is distributive (or is a distributive justice, or is tort law’s distributive justice), but not
that (3) tort law’s corrective justice is distributive justice or that (4) tort law’s justice is distributive justice.
360 HANOCH SHEINMAN

that the Aristotelian justice at issue does not extend to Aristotle’s entire province of
justice; that province (his “general justice”) encompasses interpersonal ethics in its
entirety.10 The Aristotelian justice at issue in this chapter on the other hand is confined
to that particular subclass of interpersonal ethics that comprises distributive and cor-
rective justice (Aristotle’s “particular justice”). By “Aristotelian justice” I simply mean
that justice that includes all and only principles of distributive and corrective justice.
What is it that the forms of Aristotelian justice have in common and that sets them
apart from other interpersonal ethical norms? Aristotle does not provide a very clear
independent definition of Aristotelian justice,11 but perhaps we can glean the general
idea from his definitions of its two forms. Here then is one natural suggestion:
Aristotelian justice is concerned with forced competitive allocations of (positive or
negative) person-affecting goods among parties.12 The allocation is forced in that we
cannot avoid making it in some way (if only by omission); at least one party will end
up with some of the goods and all the goods will end up with at least some party. It is
competitive in that any allocation will favor at least one party over at least some other
party relative to some alternative. It does so by benefiting the former party rather than
the latter (or burdening the latter party rather than the former) in some way. The
“rather than” implies an alternative allocation that would favor the latter over the
former relative to the actual allocation.13
Now there is no mystery about our special interest in such multiparty allocations,
where at least two parties compete over benefits or burdens that could go either way.
But perhaps it is not too early to register a worry that will become clearer shortly. On
the face of it, defining Aristotelian justice in terms of multiparty allocation excludes
principles that apply in single-party problems of allocation. Suppose, however, that the
two forms of Aristotelian justice are unified by their concern with multiparty alloca-
tion. What is it that sets them apart?

A. Distributive justice
Aristotle defines distributive justice as “geometric equality” or “equality of ratios.”
Such equality presupposes the allocation of one or more goods between at least two
parties. (But there is no need for more than one good or two parties). It also
presupposes that each party merits a certain quantity of the goods under some

10
Aristotle, who often thought of justice in terms of moral virtues, defined general justice as “complete
virtue in relation to another” (NE bk V, Ch 2, }15). In this chapter I think about justice in terms of moral
principles. Aristotelian principles of general justice are those that regulate how we treat others.
11
He defines the relevant (“particular”) injustice as overreaching, which makes Aristotelian justice a matter
of avoiding overreaching.
12
Cf. also Gardner, “Corrective Justice.”
13
Notice that competition does not imply “winner takes all.” A compromise is just one solution to a
competitive problem of allocation. Our dinner forces a competitive problem of allocation: Any allocation will
favor one of us over the other relative to some alternative. Splitting the bill favors me relative to the alternative
in which you foot the entire bill and favors you relative to the converse alternative.
 TORT LAW AND DISTRIBUTIVE JUSTICE 361

principle of distribution. Geometric equality requires that the ratio between the
quantity had by, or allocated to, a party and the quantity she merits remain the
same for each party.14 Suppose that what is allocated between A and B is money and
that the relevant principle requires distribution in accord with need (A and B have
no money). Then the ratio between the amounts of money A and B receive should
also be the ratio between the amounts they merit. If A is twice as needy as B,
A should get twice the amount that B gets.15
Aristotle’s geometric definition of distributive justice is in one way very general and
in another, very narrow. The view clearly presupposes merit-determining principles of
distribution, but it tells us nothing about their content: “The just in distribution must
accord with some sort of worth but what they call worth is not the same.” What counts
as “worth” or makes it the case that the parties merit the quantities they do can vary
from one case to another. It can be some character trait (being virtuous), action
(murder), or relational property (being the student of Plato).
Notice that the merit-determining principles of distribution geometric equality pre-
supposes are noncomparative. They tell us how much of the relevant person-affecting
goods one merits by virtue of her own merits, regardless of how much others merit or get.
Thus in principle, they apply to single-party allocations. Take the principle that requires
distribution of help in accord with need. It would certainly support allocating some help to
its only potential recipient, provided she is in need. And on the plausible assumption that
the more meriting of some good one is the more good one merits, the principle would
support allocating more of the help to her the greater her needs.
The geometric definition of distributive justice is highly specific. It requires the ratio
between what we merit and what we get to be the same, regardless of whether we get
what we merit. Suppose we are equally meriting of some good (benefit or burden).
Insofar as geometric equality of ratios is concerned, there is absolutely no difference
between the case in which we both get exactly as much as we merit and the case in
which we both get twice or half as much. Or suppose that you already got twice as
much of the relevant good as you merit, and there is simply nothing we can do about
that. Conceived of as geometric equality, Aristotelian distributive justice requires that
I also get twice as much as I merit.
Now this is a strikingly peculiar definition of the generic form (concept, ideal,
category) of distributive justice. What is most striking about this definition is not that
it is controversial or ultimately indefensible, but that it is essentially comparative in
nature. Comparative justice demands that the quantity had by, or allotted to, each party
bear a certain relation to the quantity had by, or allotted to, each other party, where the
precise relation is itself a function of how meriting they all are. Here is how Shelly Kagan
illustrates the claim of comparative justice (or desert) when the good is wellbeing and the
meriting feature is virtue:

14
For simplicity I will assume that each party has the quantity allotted to her.
15
That is essentially the Aristotelian ratio view Shelly Kagan criticizes in The Geometry of Desert (New
York: Oxford University Press, 2012), Chapter 7.
362 HANOCH SHEINMAN

When I am as virtuous as you, then I should be doing as well as you, no matter how well
you are doing. If I am not, then there is something to be said in favor of improving my lot
to bring me up to where you are—regardless of where you are.16

Kagan has recently argued that Aristotle’s ratio view is an indefensible principle of
comparative justice.17 Moreover, the very idea of comparative justice is controversial.
It is at least arguable that there are simply no such principles, that the morality of
distribution is concerned exclusively with how much we merit and get and never with
how much we merit or get as compared to one another. Notice that we do not always
need to invoke a comparative principle of justice to reach comparative conclusions.
Given the plausible assumption that the more meriting you are the more you merit,
you should get more than me if you merit more than me. But both the merit-
determining principle and the assumption are applicable to the case in which you
are the only potential recipient.
Now this already provides a strong reason against taking geometric equality of ratios
as a working definition of distributive justice in the present context. Those who discuss
the relation between civil law and distributive justice would not typically want to
commit themselves to any particular view about the existence or content of compara-
tive justice. However, the main, and in my view, decisive reason to reject the geometric
definition of distributive justice is simply that it fails to include the principles of
distribution it presupposes, namely all noncomparative principles that determine the
relevant meriting features of the parties. An essentially comparative definition of this
sort can hardly be said to capture our general notion (“form”) of distributive justice.
And I see no special reason to adopt it in the legal context.

B. Corrective justice
Aristotle defines corrective justice as “arithmetic equality” (sometimes called “equality
of difference.)” It presupposes some directed wrongful interaction (“transaction”)
between two parties, namely a morally asymmetric interaction in which one party
does some sort of wrong to the other. The interaction effectively takes (“subtracts”)
some positive or negative good from one party and gives (“adds”) it to the other.
Arithmetic equality then requires taking the relevant good form the latter and giving it
back to the former, thereby restoring the pattern of its original distribution. In short,
arithmetic equality presupposes a bilateral directed wrongful interaction that effect-
ively transfers some good from one of its parties (the original holder) to the other (the

16
Kagan, “Geometry of Desert” (note 15) at 350. Of course, wellbeing is not something we can ever allot
directly, but the relevant principle can bear on how we should allot resources that affect the wellbeing of
parties.
17
Kagan shows how this view can lead to unacceptable recommendations. For example, it can urge us to
move a more virtuous person below what she merits in response to a less virtuous person being above what he
merits. “Geometry of Desert” (note 15) at 357.
 TORT LAW AND DISTRIBUTIVE JUSTICE 363

original receiver), and requires the bilateral retransfer of that very good back from its
receiver to the original holder. This is the point of Aristotle’s line example.18
The definition of corrective justice as arithmetic equality faces a familiar problem.
Arithmetic equality of difference requires retransferring back that which has already
been transferred forward, through the underlying wrongful interaction. This only
applies to cases in which that interaction involved taking some benefit (e.g. money)
or giving some burden (e.g. debt). Here the judge can restore equality by retransferring
the relevant good back to its original holder. But such cases are quite special, and do
not represent the great majority of cases in which civil law requires correction. On the
face of it, arithmetic equality of difference simply has no application in the typical case
of reparation in torts. Consider the case in which the wrongdoer (tortfeasor) causes the
victim some loss through breaching her duty of care in negligence. The loss in question
has never been transferred from the tortfeasor to the victim. Thus it cannot be
retransferred back to the wrongdoer. Clearly, transferring the loss to the wrongdoer
would not restore its original distribution between the parties (arithmetic equality).
It is tempting to think that the problem is technical. The problem, you might think,
is that the option of transferring back is not always available, so maybe we can solve it
by adding an “as far as possible” proviso (for example). But the familiar problem with
arithmetic equality is symptomatic of a deeper problem. The main difficulty with such
equality is not so much that it is frequently impossible but that it is not particularly
corrective. Consider a deviant version of Aristotle’s line example in which the wrongful
interaction has effectively transferred some benefit (say, money) from the wrongdoer
to the victim or burden (say, debt) from the victim to the wrongdoer (perhaps it has
done both of these things). If what matters in correction is arithmetic equality of
difference (restoring the original distribution of the relevant person-affecting conse-
quences), then corrective justice requires retransferring the wrongdoer’s loss back to
the victim and the victim’s gain back to the wrongdoer! Clearly, it does not. Aristotle
took this much for granted. But then arithmetic equality isn’t what really matters in
correction.

C. The relation between the forms

The Aristotelian definition of his forms of justice in terms of distinct equations clearly
suggests that they are mutually exclusive categories of his allocative justice. Principles
of both types might happen to apply to a single case, but the principles themselves
belong to mutually exclusive categories. The relation between the Aristotelian forms of
allocative justice looks like this:19

18
Imagine a line A----E--D--C----B where AD = DB and ED = DC. Aristotle (Bk V, Ch 4, } 8) writes:
The judge restores equality, as though a line [AB] had been cut into unequal parts [AC and CB], and he
removed from the larger part [AC] the amount [DC] by which it exceeds the half [AD] of the line [AB],
and added this amount [DC] to the smaller part [CB]. And when the whole [AB] has been halved [into
AD and DB], then they say that each person has what is properly his own, when he has got an equal share.
364 HANOCH SHEINMAN

Allocative Justice

Distributive Corrective
Justice Justice

Figure 17.4 Distributive and corrective justice are categorically distinct

Why isn’t corrective justice or allocation a special principle of distributive justice or

allocation under Aristotle’s account? One possibility is that although corrective justice
is relational principle (it reference to both parties to the underlying interaction), is
not a special case of comparative justice. Arguably, retransferring the relevant good
back to its original holder would result in geometric equality only if the original
distribution of the relevant good was geometrically equal.
Granted but I have already noted that the restriction of distributive justice to
comparative principles is unmotivated, at least in the present context. And this raises
the possibility that corrective justice can be understood as a special, relational thought
noncomparative principle of distribution. I now turn to explore this possibility.

IV. An Alternative Account

The Aristotelian distinction between the forms of justice has some tremendous
advantages in the civil law context. First, like the forms of justice, civil law is often
concerned with the allocation of person-affecting goods (benefits and burdens)
between multiple parties. Second, civil law is often taken by those who make,
administer, or practice it to be a legal institution whose job it is to mete out or
pursue justice in the allocation of such benefits (“gains”) and burdens (“losses”).
Third, it seems that insofar as we think about civil law as a distinct legal institution
with its own principle, we think about it as an institution of something like corrective
justice. These are all features of the Aristotelian distinction I want to preserve.
But as we have seen in the previous section, the particular equations Aristotle
uses to draw the distinction have some disadvantages in the present context.
The essentially comparative formula of “equality of ratio” makes for a peculiarly
specific conception of justice in distribution. The definition of corrective justice as
“equality of difference” makes it inapplicable when there is nothing to transfer
back. And the assumption about the categorical nature of the distinction between

19
This is just the simplest way to understand the relation in the Aristotelian picture. Properly understood, it is
compatible with the possibility that a single allocation will happen to achieve both types of equality; what it rules
out is the possibility of achieving one type of equality simply by achieving the other. In any event, all that is
important for my purposes is that the Aristotelian assumption that geometric and numeric equality are categor-
ically distinct (which makes possible non-geometric numeric equality and non-numeric geometric equality).
 TORT LAW AND DISTRIBUTIVE JUSTICE 365

the forms is hard to reconcile with their common concern with allocation. Granted:
you can have non-corrective principles of distribution and non-distributive prin-
ciples of correction. What is less clear is that you can have non-distributive
principles of correction in the allocation of goods. And arguably, the principle
that defines civil law’s function or ethical function is not just any principle of
correction; it is principle of correction in the allocation of goods. It is a genuine
principle of distributive justice. My goal in this section is to characterize a recon-
structed version of the distinction that preserves the main advantages of the more
familiar picture and at the same time brings out the genuine distributive nature of
civil law and justice.

A. Distributive justice
To make room for noncomparative principles of distribution we must first give up on
the Aristotelian assumption that distributive and corrective justice are confined to
multiparty allocations. We can still focus on cases of forced competitive multiparty
allocations. We can continue to assume that the relevant problems of justice involve
forced competitive allocation. What I no longer wish to assume is that the principles
that bear on the problems make reference to more than one party.
As I wish to understand it, justice in the distribution of goods is not a well-defined
principle; nor is it even a distinct form of justice. Rather, it is simply the variable set of
principles that bear on the right or desirable way to allocate goods between parties in
any given case, whichever they happen to be. These can include noncomparative
principles of distribution such as “To each according to her needs” as well as com-
parative principles such as Aristotle’s ratio view. Notice that while there is nothing
particularly geometric about this way of understanding distributive justice—it entails
neither Aristotelian equality of ratio nor any other principle of comparative justice—it
does comport with a broad understanding of Aristotle’s remark that “the just in
distributions must accord with some sort of worth.”20 We can read “(absolute)
worth” as “worth or relative worth.” Similarly we can say that distributive justice
requires that goods be allocated in accordance with the merits of the parties, where
“merits” should be read as “(absolute) merits or relative merits.” The primary notion
of a party’s merits remains as noncomparative (“absolute”) as it is in Aristotle’s
geometric equality, but the remark uses the term in a broader sense that includes
some relation between what the parties get (relation itself defined in terms of what they
merit in the primary sense). The resulting conception of distributive justice is com-
pletely agnostic about the existence and content of comparative justice.
The reconstructed definition also retains Aristotle’s largely formal conception of
merit (“what they call worth is not the same”). This indeed is a familiar feature of our
most general use of the terms “distributive justice” in moral and legal discourse. As

20
Bk V, Ch 2, } 7.
366 HANOCH SHEINMAN

I understand this notion, a party’s “merits” are simply those features that merit
(require or otherwise favor) her having or being allotted the relevant good or quantity
under some applicable ethical principle. She merits the good or its allocation to the
extent that she qualifies for it under such principle. The principles that determine the
meriting properties can be teleological or deontological. And the merits are not
confined to character traits (let alone virtues). They extend to relational properties
such as promissory or filial relations. And they are not confined to present-time
features of the situation such as (current) suffering or need. They extend to historical
features (past events and relations).
I want to retain the assumption that the goods the allocation of which is in question
must be person-affecting, namely good or bad to the recipients, but not the assump-
tion that they must include at least two goods or some divisible good; we can compete
over the distribution of a single indivisible good. Again, there is no assumption that the
goods in question are being assigned for the first time. The problem of whether and
how to redistribute an already assigned good is a problem of distributive justice.
Finally, there is nothing particularly political about the relevant goods or parties.
The problem of how to divide some good between two private persons is one of
distributive justice.
If you complain that this definition of distributive justice is rather indiscriminate,
you are probably right. One part of my defense is that this indiscrimination reflects a
perfectly familiar sense in which we use the language of distributive justice in moral
and legal discourse; the alternative is to use some controversial or rather narrow
theory of distributive justice (such as the ratio view). Another part of my defense is
that problems of distributive justice are still confined to the allocation of goods (and,
we can assume, to forced competitive allocation). Finally, part of my claim will
be that perhaps the most central issue about the relative place of distributive
and corrective justice in civil law has more to do with the latter than with the
former. The contrast that drives competing claims in this area is not so much
between corrective and distributive justice as it is between corrective and
non-corrective justice.

B. Corrective justice
The next order of business is to offer an account of corrective justice with an eye to civil
law. I believe that the account should retain two major Aristotelian insights. The first is
that the corrective justice of civil law (the justice Aristotle imagines the judge to mete out
in his line example) regulates the forced competitive allocation of person-affecting goods
between multiple (i.e., at least two) parties in the wake of some directed wrongful
interaction between two parties. While this insight does not yet make it “geometric”
under the essentially comparative Aristotelian scheme, it certainly does seem to make it
distributive in the ordinary sense of the word. As a principle regulating the allocation of
 TORT LAW AND DISTRIBUTIVE JUSTICE 367

goods, corrective justice may well qualify as distributive under the non-geometric
account of the previous subsection.
The other Aristotelian insight to retain is that the relevant principle is not just any
principle of distribution; it is a distinctly corrective or backward-looking one. It
presupposes and responds to some past wrongful interaction. At the same time,
I leave out the further assumption that backward-looking or corrective distribution
requires retransferring consequences of the underlying interaction back to their
original holders.
Since I take the idea of correction to be more or less basic, my basic character-
ization of corrective justice is going to be circular. Corrective justice requires
corrective backward-looking redistributive action or operation in the wake of
directed wrongful interactions. The operation is redistributive in that it requires
an active interference with the existing distribution of goods, namely the person-
affecting consequences the underlying interaction has had for its parties. The
requisite redistribution is backward-looking in being a direct response to the
underlying wrongful interaction. Corrective justice takes the wrongful interaction
itself as a reason or justification for intervention (or as a duty to intervene).
Elizabeth Anscombe captures the idea when she writes: “something that has
happened . . . is given as the grounds of an action or abstention that is good or
bad for the person at whom it is directed.”21 The idea of backward-looking
reasons, justification, or duties is not essentially distributive; if I offend your
feelings, I have a backward-looking reason to apologize, but there is no question
of any particular good changing hands. Yet there is nothing odd about backward-
looking reasons for redistribution. My claim is that corrective justice is a back-
ward-looking principle of redistribution in the Anscombian sense.
Not all backward-looking action or operation is corrective, however. Praising
someone is backward-looking, taking one’s past action as a reason to praise, but a
praiseworthy action gives us nothing to correct. When we take corrective action, we do
not simply take the past event as our reason for acting; we also see our present action
as something that counteracts the wrongfulness of that event (“right the wrong”).
Corrective redistribution is one that can be properly said to make the underlying
wrongful interaction, or its wrongful consequences, right again.22
What sort of redistribution do we have backward-looking reason or duty to
undertake under corrective justice? What sort of backward-looking redistribution
can be said to make the aftermath of the wrongful interaction right again? In this
subsection I characterize the underlying interaction corrective justice presupposes and
the problem of distribution to which it gives rise. In the next subsection I characterize
the distinctly corrective solution.

21
Elizabeth Anscombe, Intention (Cambridge, MA: Harvard University Press, 1963) } 20. Anscombe wrote
about “backward-looking motives.”
22
Part of the idea might be that the reasons for the corrective action reflect the reasons against the wrongful
interaction that gave rise to it. See Gardner’s continuity thesis in his “Corrective Justice.”
368 HANOCH SHEINMAN

Corrective justice presupposes a directed wrongful interaction in which one party

(the wrongdoing moral agent A) wrongs another party (the wronged moral patient P).
It concerns the distribution of the good or bad person-affecting consequences of that
interaction for its parties (A and P). Thus it is confined to the transferrable person-
affecting interparty consequences of the underlying interaction. What counts as a
consequence of the wrongful interaction is party given by the problem and party
determined by the solution (corrective justice).
So the wrongful interaction that gives rise to a problem of corrective justice also
gives rise to a problem of distributive justice: How to distribute the interparty
transferrable consequence of the underlying interaction between their potential
bearers? The potential bearers are all and only those to whom the relevant
consequences are transferrable. Thus the parties to the problem depends on our
options. Suppose that the interaction produces some gain or loss to one of the
parties to the original interaction. I am assuming that we always have the option of
doing nothing, thereby allocating the gain/loss to the party who already enjoys/
suffers it by omission (A or P). I’m also assuming that we can transfer the loss/gain
onto the other party. Given these assumptions, the underlying interaction cannot
fail to give rise to a problem of multiparty distribution of goods: it always requires
some forced competitive allocation of some good (gain/loss) between at least two
parties (A and P).
However, the bilateral nature of the underlying interaction of corrective justice (the
sheer fact that it involves exactly two parties) does not guarantee the bilateral nature of
distribution problem. A gain/loss that is transferrable to the other party (the one who
does not already enjoys/suffers it) might well be transferrable to some third party (C).
Suppose that we also have the option of transferring the gain/loss to C. Then we have a
three-party problem of distribution.

C. Redistribution in the right-making direction

Corrective justice presupposes some wrongful interaction between A and P with
transferrable person-affecting consequences for at least one of them. Corrective
justice tells us how to distributive these consequences between multiple parties,
which include at least A and P, but can include others. Corrective justice is a
distinctly backward-looking solution to this problem, for it takes the wrongful
interaction as itself the reason for the solution. It requires a distribution that can
appropriately be said to make the post-interactive situation right again (in some
sense I’m taking as unanalyzed). What sort of distribution would fit the description?
In general, one that is directly sensitive to the nature of the wrongful underlying
interaction in the right sort of way. But what sort of distributive solution would
reflect such sensitivity?
A partial answer is that to reflect the bilateral nature of the underlying interaction,
corrective allocation must itself be bilateral: it requires allocating the relevant goods
 TORT LAW AND DISTRIBUTIVE JUSTICE 369

between the parties to the underlying interaction and them alone, namely between
A and P. Thus allocating the relevant goods to third parties cannot satisfy the
demands of corrective justice. Another part of the answer is that to respond directly
to the wrongful interaction, corrective allocation must be redistributive: it requires
an active intervention with the existing distribution of the relevant goods. The
wrongful interaction has produced a distribution that is itself wrongful in some
way, simply on account of having been produced by that interaction. You cannot
correct a distribution without redistribution.
Redistributing the relevant consequences between the parties is not yet corrective,
however. Suppose that the relevant interaction has produced some gain for P or loss
for A. Surely there is nothing corrective about redistributing it to the other party. Far
from making things right again, such redistribution would make things even worse.
And as we have seen, the arithmetic principle that the relevant gain/loss be retrans-
ferred back to its original holder (thereby restoring the original pattern of their
interparty distribution) does not solve that problem. What seems to be needed is
some other principle that bridges the gap between correction and distribution.
The requisite principle must relate the “directions” of the corrective redistribution
and underlying wrongful interaction. To formulate the principle with clarity we need
to introduce three theoretical terms. Let us say that a directed wrongful interaction
between X and Y has the X to Y direction of wrongdoing when it involves X wronging
or doing wrong to Y. We now distinguish between two senses in which a redistribution
of some good can be said to take a particular direction. Let us say that a directed
redistribution of some person-affecting good/benefit/burden between X and Y has the
X to Y direction of good/benefit/burden-moving when it redistributes the good/benefit/
burden from X to Y. Finally, let us say that a directed redistribution of such good
between X and Y has the X to Y direction of burdening-by-benefiting when it burdens
Y by benefiting X to the same extent (equivalently: when it benefits X by burdening
Y to the same extent).
Notice that while a redistribution with the X to Y direction of benefit-moving has
the Y to X direction of burdening-by-benefiting (it burdens X by benefiting Y to the
extent of the benefit), a redistribution with the X to Y direction of burden-moving has
the opposite, X to Y direction of burdening-by-benefiting (it burdens Y by benefiting
X to the extent of the burden). The principle we need says that corrective redistribu-
tion reverses the underlying interaction’s direction of wrongdoing with its direction of
burdening-by-benefiting. Corrective redistribution in the aftermath of an interaction
with the X to Y direction of wrongdoing takes the Y to X direction of burdening-by-
benefiting.
Using “A” and “P” schematically for the wrongdoing agent and wronged patient
affords a simple statement of the relevant principle. Since directed wrongful inter-
actions have the A to P direction of wrongdoing, the principle says that corrective
redistribution takes the P to A direction of burdening-by-benefiting. Securing that
direction of burdening requires redistributing the good and bad interparty conse-
quences of the underlying interaction (gains and losses) in opposite directions. Gains
370 HANOCH SHEINMAN

should be redistributed from A to P (with the A to P direction of good-moving); losses,

from P to A (with the P to A direction of good-moving). To mark the fact that
redistribution in this direction of burdening tends to warrant the imagery of righting
the wrongful interaction or its consequences, I will sometimes call it the right-making
direction.
Recall the case in which the wrongful interaction transfers some gain from A to
P and loss from P to A. Unlike arithmetic equality of difference, the principle on offer
avoids the implication that corrective justice requires retransferring the said conse-
quences to the other party: such retransfer would have the wrong, A to P direction of
burdening-by-benefiting. It would fail to put things right again because it would fail to
reverse the interaction’s direction of wrongdoing with is direction of burdening. Now
someone might object that the principle still yields the wrong result. Redistribution
with the P to A direction of burdening would require transferring the relevant
consequences to the parties who already have them, which is impossible.
It is however possible to reply that the principle is undefined for such cases (as for
cases in which the wrongful interaction has no interparty person-affecting conse-
quences at all). Not all person-affecting consequences that wrongful interactions have
for their parties require correction, only those that benefit A or burden P. Let us say
that when a directed wrongful interaction between X and Y produces some person-
affecting consequence for one of them, that consequence has the X over Y direction of
favoring when it benefits X or burdens Y. Then we can say that the relevant principle is
only defined for consequences with the A over P direction of favoring.
We can think of various principles of corrective justice. Some principles require
A to take corrective responsibility for the relevant consequences of her wrongful
agency by effecting or facilitating the requisite redistribution. Others require the judge
to hold A correctively responsible and enforce the said redistribution. Still others
require the law to pursue corrective justice. Principles of the first two kinds are most
naturally thought of as deontic (imposing a requirement). Principles of the third kind
might well be telic (generating reasons).

E. Reconciling correction and redistribution

The reconstructed account does away with the assumption that the forms of justice are
categorically distinct and describes a kind of corrective justice that is fully distributive
as well as fully corrective. The most basic idea of corrective action in normative ethics
is that of backward-looking principles that identify past wrongdoing as a reason for
doing something about it now, something that can be properly said to correct or make
things right again. This idea is not particularly distributive. Reasons to express certain
reactive attitudes such as regret or blame are backward-looking and broadly corrective.
Reasons to apologize are clearly corrective. Such actions do not necessarily allocate
person-affecting goods. There is however nothing about the basic idea to exclude
genuinely corrective principles of distributive justice. The idea of backward-looking or
“historical” principles or theories of distributive justice is not unfamiliar. One example
 TORT LAW AND DISTRIBUTIVE JUSTICE 371

is the familiar view that takes the fact that some good has been acquired through
voluntary exchange to justify or favor its existing distribution. Another example is the
familiar argument that we should transfer opportunities from one group to another
as compensation for past wrongdoing. The theory of distributive justice behind
the argument is historical and corrective. It requires redistribution with the P to
A direction of burdening by virtue of some past interaction with the A to
P direction of wrongdoing.
The corrective justice that preoccupied Aristotle and continues to exercise contem-
porary civil law theorists is concerned with the allocation of person-affecting goods in
accord with merit (“some sort of worth”). The merits of the parties are relational and
interdependent; they are determined by the roles they played or didn’t play in the same
wrongful interaction and cannot be described independent of one another: A plays the
role of the agent who wrongs P; P plays the role of the patient who is wronged by A.
Relational principles or theories of distribution are not unfamiliar. To use an example
from the previous paragraph, the process through which a good is acquired is a
relational fact. All criteria of distribution that depend on social facts are implicitly
relational, to some extent.
So corrective justice is not just any principle of distribution. Unlike principles that
require distribution in accordance with need, suffering, or talent, it takes the wrongful
interaction itself to require doing something about its interparty transferrable conse-
quences, something that can be properly said to put things right. Unlike these other
principles, it requires that these consequences remain between P and A. And unlike these
other principles, it requires the redistribution to have the A to P direction of burdening.
The proposed account of the distinction does not baffle competing claims about the
relative priority of corrective and distributive justice in relation to civil law. Rather, it
requires their reinterpretation as claims about the relative priority of corrective and
non-corrective principles of distribution or justice. In the next section, I apply the
distributive account of corrective justice to tort law.

V. Reparative Justice in Tort Law

My distributive account of corrective justice was characterized with an eye to civil
law, that area of law that includes the law of torts, contracts, and restitution. It is
often argued that civil law’s first principle or principle of justice is corrective.
The claim is plausible. Civil law requires corrective action in the aftermath of civil
law wrongs (torts, breach of contract, unjust enrichment). This preoccupation also
seems to define civil law as a distinct area of law. What could possibly unify the
paradigmatic institutions of civil law if not some sort of backward-looking concern
with correction? There is nothing sacred about conceiving of the institutional
subject in this way, but as soon as you conceive of it in some other way you
begin to change the subject.23 Under the assumption about the categorical nature
of the key distinction, this corrective view implies that civil law’s justice is not
372 HANOCH SHEINMAN

distributive. The implication is not simply that civil law’s justice is not identical with
distributive justice as such. No one thinks that corrective justice exhausts distribu-
tive justice, and the claim that civil law’s distinct function is distributive justice as
such rings neither true nor helpful; clearly concern with distributive justice can
hardly set civil law apart from most other areas of law. Indeed the categorical nature
of the distinction forces the stronger implication that civil law’s justice, which is
corrective, is not also distributive.
If it turns out that civil law’s first principle of justice is basically the reconstructed
Aristotelian corrective justice I have been characterizing, then we are entitled to reject
this implication. For in that case the denial of the distributive nature of civil law’s
justice would fail to be true and, in any event, would fail to register an important truth
about the normative foundation of civil law, namely that it is an essentially distributive
institution whose function it is to solve special problems of distribution by way of
special redistributive action. In this section I will try to support this claim in relation to
the law of torts, which I take to be a paradigm case of civil law, along with the laws of
contracts and restitution. I will try to support the claim that tort law’s first principle of
justice is a special case of the distinctly distributive principle of corrective justice I have
described in the previous section. I will use the account to show how tort law’s justice is
both corrective and distributive.
Now on the one hand, it is the distributive nature of tort law’s justice that stands in
greater need of vindication. So for the most part, my discussion of the distinctly
corrective nature of tort law’s justice will rely on familiar arguments (and hence
depend on their cogency). On the other hand, I am also concerned to show how the
distributive nature of tort law’s justice is compatible with a certain priority of correct-
ive justice. Thus corrective and distributive aspects of tort law’s justice will be mixed in
the discussion that follows. This will reflect one of this chapter’s themes, that when it
comes to tort law, corrective and distributive justice are inseparable; tort law’s justice is
corrective-and-distributive throughout. Before discussing tort law, I briefly comment
about civil law generally.

A. Civil law justice

Remedial civil law rules require corrective redistribution in the wake of directed
wrongful civil law interactions. They seem to define duties, rights, and powers (in
short, relations) of corrective justice in the reconstructed Aristotelian sense. A civil law
wrong is a bilateral interaction with the A to P direction of wrongdoing (a tort, breach
of contract, unjust enrichment). But not every such interaction is actionable. To
amount to an actionable or complete civil law wrong, the interaction must have
transferrable person-affecting consequences for its parties. Remedial civil law rules

23
Cf. Martin Stone’s remark that in conceiving of tort law as “the legal response to the problem of
accidents,” one never gets “tort law into view at all.” Stone “Doing and Suffering”, p. 151.
 TORT LAW AND DISTRIBUTIVE JUSTICE 373

require the redistribution of these consequences between the parties in the opposite,
P to A direction of burdening-by-benefiting. What is distinctive about civil law liability
rules as such (if anything) is that they require this redistributive operation simply as
the appropriate response to the underlying wrongful interaction, quite regardless of
other reasons for or against it. When the interaction produces a gain for A, loss for P,
or both, the sanctioned civil law response is transfer with the A to P direction of good-
moving (“restitution”), transfer with the P to A direction of good-moving (“repar-
ation”), or both.
It may be objected that my account cannot explain restitution. It is commonly
argued that the law of restitution imposes duties to relinquish gains that are not
wrongful (as when I find that you have mistakenly transferred some money into my
account). But if these are duties of corrective justice at all, they are highly nonparadig-
matic ones. The typical argument for treating such duties as corrective is that they
presuppose an interaction what would be wrongful unless they are discharged. But if
an argument of this sort is cogent, the duty to keep promises or perform contracts
is a duty of corrective justice. I find this result counterintuitive. There is simply
nothing for promisors or contractors to correct by keeping their promises or contracts.
Things change as soon as the potentially wrongful interaction becomes wrongful in
actual fact owing to a failure to discharge the duty (relinquish the gain, keep the
contract). Potentially wrongful interaction can only give rise to potentially corrective
duties. Notice that the objection effectively collapses the corrective nature of the
relevant duties to their backward-looking or historical nature. The duties to relinquish
unmerited gains or keep contracts are backward-looking in the Anscombian sense of
identifying some historical fact (mistaken transfer, contract) as itself a reason or duty
to do something. But they are not fully corrective, because the agent has not done
anything she can correct or fail to correct.

B. What is harm reparation?

Some philosophers define corrective justice as the requirement to repair harm.24 And
in tort law, correction does seem reparative. But what is this requirement? A familiar
answer says that to repair P’s harm is to bring P to the position in which P was before
(or would have been apart from) the tort. But such restoration is often impossible
because the harm to P is irreversible. Suppose that A causes his neighbor P permanent
physical injury through his careless conduct. Suppose also that the harm leads to
financial losses that are associated with P’s condition (e.g. medical expenses). Bringing
P back to his original position is impossible. All that can be done is bringing P back
to his original position with respect to the collateral losses (as we might call them). To
accommodate this familiar point, we can retreat to the claim that tort law requires
bringing P to his original position “as far as possible.” But this still sounds too

24
See e.g. Perry, “Corrective and Distributive Justice.”
374 HANOCH SHEINMAN

strong. Tort law does not require restoring P’s position in every possible way. It
does not require A to apologize, work for P, or make some third party cover P’s
expenses.
An account that acknowledges the essentially distributive nature of corrective
justice offers a more informative yet simpler definition of reparation. Reparation of
harm in torts is essentially redistributive and can only take the form of transfer.
Physical injury is not transferrable. Therefore, it is not repairable. The collateral loss
of mitigating and coping with the injury on the other hand is financial and can be
transferred to A as debt. Harm reparation in tort requires the transfer of P’s collateral
(i.e., transferrable) tort-generated losses onto A.

C. Redistribution in the right-making direction

Tort law reparation also provides an occasion to compare the redistributive account
of corrective justice with arithmetic equality accounts. The redistributive account of
reparation retains two attractive features of the Aristotelian arithmetic account of
corrective justice: The assumption that (1) corrective justice operates on material,
person-affecting goods, gains and losses the parties to some wrongful interaction
enjoy or suffer in its wake; and the assumption that (2) doing or achieving corrective
justice is a matter of transferring these goods between these parties. The first
assumption explains why the interaction creates a problem of forced competitive
distribution of goods. The good is here to be enjoyed or suffered by at least some
party, but there is some alternative allocation under which it is enjoyed or suffered by
someone else. The second assumption begins to explain how the corrective solution
differs from others: it requires the relevant good to remain between the parties. And
it requires redistribution of the relevant good, an intervention in its existing distri-
bution. But these assumptions are not sufficient to explain the distinctly backward-
looking or corrective nature of reparation. Changing the interparty distribution of
the relevant good does not yet restore or correct anything. What explains the
backward-looking or corrective nature of corrective justice in the Aristotelian picture
is the idea of arithmetic equality, the assumption that (3) corrective redistribution
transfers back some positive or negative good that has been transferred through the
underlying interaction (in much the same way a refund does), thereby restoring the
original pattern of its distribution.
As I have noted, arithmetic equality view renders corrective justice all but inapplic-
able to reparation in torts. A in our previous example would be liable to repair P’s
collateral losses. Such reparation would require transferring P’s loss to A. But the loss
has never been transferred from A to P in the first place. A could not have had that loss
before the tortious encounter that created it. And the loss has remained in P’s
possession ever since (otherwise transferring it to A would not be an option now).
Therefore, transferring P’s loss to A would not transfer that loss back.
 TORT LAW AND DISTRIBUTIVE JUSTICE 375

The Aristotelian conception of corrective justice as arithmetic equality is still

influential. However, the impetus to reconcile it with tort reparation exerts pressure
to adopt some “normative” interpretation of the relevant goods, thereby giving up on
the attractive Aristotelian assumption (1). Gardner’s account of corrective justice
provides a recent illustration. The question of corrective justice on this account is
whether or how something that “has already shifted between the parties” to a trans-
action “should now be allocated back from one party to the other, reversing the
transaction that took place between them.”25 How should we understand this claim
in the context of reparation in torts? I assume that what reparation allocates from
P to A is some loss associated with the tort. But in what sense does it allocate the loss
back to A?
The claim seems to imply that P’s loss had been A’s prior to the tort: the tort has
shifted it from A to P, and should now be shifted back. But this is impossible, as it was
the tort that created the loss in the first place. Gardner’s idea might be that P’s loss
should be returned to A, who has been its rightful owner since its creation through the
tort. But since the material, person-affecting suffered loss the tort has produced lies
with P, the ownership in question must be a kind of normative (moral or legal)
ownership or liability. Thus the explanation of the sense in which reparation is
corrective presupposes something like the Pottery Barn principle: “You break it, You
own it.” Alas this principle does not require the returning or otherwise transferring
anything back from P back to A. We reach a dilemma. If the loss is material, then A has
never had or suffered it, and it cannot be allocated back to A. And if the loss is
normative (duty), then P has never had it, and it cannot be allocated from P. The
Pottery Barn principle simply reformulates the duty of reparation in terms of owner-
ship. It cannot explain reparation’s (i.e., its own) corrective or backward-looking
nature.
The account of the previous section provides the beginning of an explanation.
Reparative justice is backward-looking simply in that it takes the past tortious
interaction as itself a reason/duty to interfere with the existing distribution of the
loss. The corrective nature of reparation is hard to explain in non-circular terms.
Still reparation can be brought under a more general principle that requires us to
redistribute the person-affecting consequences of wrongful interactions in the P to
A direction of burdening-by-benefiting, namely in whichever direction of good-
moving would work to the detriment of A by working to the benefit of P. Why?
Because that would warrant the claim that the redistribution makes things right
again. Why? Because it was A who wronged P and not the other way round.

25
Gardner, “Corrective Justice” at 11.
376 HANOCH SHEINMAN

D. Corrective defenses
Tort law provides defenses that limit the scope of the tortfeasor’s (A’s) liability to
repair the victim’s (P’s) harm, including “proximate cause,” “contributory negligence,”
and “mitigation of damages.” Such doctrines are sensitive to the relative contribution
of the parties to the production of the wrongful interaction or its consequences.
Suppose for example that both parties are “at fault” or otherwise responsible for the
tortious interaction or P’s loss. Suppose that the tortfeasor A can be said to have been
responsible for seventy-five per cent of the loss while the victim P can be said to have
been responsible for the rest. Then the law may well take this as a reason to limit the
scope of A’s duty to repair to seventy-five per cent of the loss. Such doctrines seem
essentially distributive. They concern the question: How should we divide the loss
between the parties? (Notice that the rule in my own example seems essentially
comparative, reflecting something like the ratio view; but I have already noted that
this interpretation is optional.)26 If tort law’s first principle of justice (i.e., reparative
justice) is not distributive, these doctrines cannot be said to form part of tort law’s first
principle; they must be relegated to tort law’s second principle. But intuitively, these
doctrines are integral to tort law’s first and reparative principle of liability. The
distributive account of corrective justice comports with this intuition. It treats tort
law’s sensitivity to the respective contributions of the parties to the tortious interaction
or loss as internal to its corrective–reparative justice.
Reparative justice is a distinctly corrective, backward-looking and relational prin-
ciple of distribution. It takes the past tortious encounter between A and P as itself a
reason to redistribute its interparty consequences in the right-making direction. This
can at least begin to explain the sensitivity of liability rules to the respective contribu-
tions of the parties without relying on non-corrective principles of distribution. It can
do so in one of two ways, depending on the kind of case.
(1) Bi-directional wrongdoing. In one sort of case, the underlying interaction
comprises two sub-interactions with opposite directions of wrongdoing, calling for
two reparative transfers with opposite directions of burdening. The overall correctively
just response will require setting these transfers off with or without remainder. No
additional principle of distribution is required.
(2) Joint contribution to one-directional wrongdoing. In the more interesting sort of
case the underlying interaction has just one direction of wrongdoing, but the wronged
moral patient P nevertheless plays some morally significant agential role in the
production of the interaction or her own loss. For example, P can be partly responsible
or “at fault” for the loss by virtue of failing to take reasonable measures to protect
herself without doing some wrong to A. While P’s imprudent agential contribution is
not wrongful, it is still very much a feature of the underlying wrongful interaction, and

26
For example, the rule can also be said to reflect the twin noncomparative principles: (1) one’s being at
fault for some loss merits one’s suffering it and (2) one’s being more at fault merits one’s getting more of it.
 TORT LAW AND DISTRIBUTIVE JUSTICE 377

so can be taken to generate a backward-looking and relational normative factor: a

reason against redistributing twenty-five per cent of the loss in the right-making
direction (i.e., to A) or perhaps a permission not to repair that part (or immunity to
liability to repair it). The lesson is that corrective justice does not always take the form
of duties to correct; it sometimes takes the form of rights not to correct. Corrective
defenses delimit the consequences that are subject to positive duties of corrective
redistribution with the right-making direction.
It may now be objected that the said defenses and other tort law doctrines are
sensitive not simply to the respective contributions of the parties to their past encounter
but also to their respective needs, abilities, and other non-corrective features. This tends
to show that they reflect non-corrective principles of distributive justice.27 Therefore, so
goes the thought, they cannot be subsumed under a distinctly distributive principle of
corrective justice. Granted, there is more to tort liability than corrective redistribution,
including sensitivity to non-corrective principles of distribution. The thought is that
reparative–corrective redistribution is still tort law’s first principle of justice.
The institution I have discussed in this section is a rather idealized tort law, tort law
as it is traditionally conceived. Later I will relate the discussion to a more realistic tort
law, our mixed tort system. But first I want to say a word about the lessons of the
discussion so far for the priority of corrective justice.

VI. The Priority of Corrective Justice

If the foregoing discussion is basically right, tort law’s distributive justice is nothing
other than its corrective justice. What are the implications of this view for apparently
conflicting claims about the relative place of corrective and distributive justice in the
explanation or justification of tort law? It depends on how we understand the debate.
Insofar as these priority claims are taken to assume the categorical nature of the
relevant distinction, the account rejects them all, regardless of side; it rejects the debate.
However, the account lends some support to a priority of corrective justice that does
not assume its non-distributive nature.
One lesson of the discussion is the asymmetry between our most general concep-
tions of distributive and corrective justice. As we typically use this term in normative
discourse, “distributive justice” does not identify any well-defined principle of distri-
bution, but simply the principles that make some distribution just (whichever they
happen to be). Of course, we often use the expression in some narrower sense, focusing
on some particular subclass of such principles (e.g. political, deontological, comparative),
and the context of civil law is no exception. But at least when we ignore the Aristotelian
contrast with corrective justice, that context suggests no obvious or obviously useful

27
See e.g. Cane, “Distributive Justice” and Karen-Paz, “Egalitarianism as Justification.”
378 HANOCH SHEINMAN

such restriction. The language of “corrective justice” on the other hand does seem to
identify a minimally well-defined principle or family of principles. On any plausible
definition, principles of corrective justice are “backward-looking” in some interesting
sense that excludes most principles of justice or distribution.
The thought suggests that perhaps the issue is not priority as between corrective and
distributive justice, but as between corrective and non-corrective distributive justice.
Put another way, the pertinent contextual definition of distributive justice is not
independent of the concept of corrective justice. “Distributive justice” in this context
simply designates non-corrective principles of distribution (in the ordinary flexible
sense).
Notice that on this way of thinking, the relevant contrast is not between two well-
defined forms of justice, but between the well-defined form of corrective justice and its
anti-form, a residual category of principles of distribution as distinct from each other
as they are from principles of corrective justice. Still the redistributive account of tort
law’s justice vindicates the priority of corrective justice over the relevant set of
principles. For while tort law’s first principle of justice is distributive (as these other
principles are), it is a distinctly corrective such principle (as they are not). Tort law has
all manner of distributive effects that open it up to evaluation in light of various
principles of distributive justice (need, virtue, equality of opportunity). But this much
seems true of most every other legal institution, from criminal to tax to information
law. What does seem true of civil law, but not of most other institutions, is the
requirement of distinctly backward-looking, relational, and corrective redistribution,
one that is justified by some past directed wrongful interaction and is properly
sensitive to its morally relevant features (but not to other morally relevant features
of the case).
This represents an institution–internal justificatory priority of corrective justice.
Whether it can support a stronger justificatory priority is a difficult question that is
not special tort law. At a minimum, it is plausible to think that when an institution’s
first principle of justice is X and not Y, X will be more frequently applicable to the
institution’s operation, and therefore more relevant to the justification of the insti-
tution itself, than Y. It is also plausible to believe—but much harder to defend—that
institutions with their own “ethical function” (roughly, ethically acceptable or
valuable function) enjoy a certain justificatory autonomy. The basic idea is that of
division of ethical labor. The general thought can be broken into sufficiency and
necessity claims. First (sufficiency), an institution that performs its defining ethical
function sufficiently well is justified, even if this comes at some moral price, unless
that price is too large. Second (necessity), an institution that fails to perform its
defining ethical function sufficiently well is unjustified, unless the ethical price of its
abolition is too large.
Suppose that tort law’s ethical function is corrective justice. Then justificatory
autonomy makes it more significant to the justification of tort law than any other
 TORT LAW AND DISTRIBUTIVE JUSTICE 379

principle of distribution. Up to a point, if tort law is sufficiently good at doing or

achieving corrective justice, then it is justified (sufficiency). And up to a point, if tort
law is not sufficiently good at doing or achieving corrective justice, then it is unjustified
(necessity). No other principle of distributive justice can be said to enjoy such
justificatory priority.

VII. Similar Views

A distributive account of tort law’s corrective justice offers one way to reconcile tort
law’s corrective and distributive functions or principles. Recent accounts of tort law’s
distributive justice suggest another route to reconciliation.
Peter Cane and John Gardner claim that tort liability rules effectively distribute
bundles of reparative legal rights and responsibilities or relations—enforceable duties,
powers, and immunities—among members of different groups (e.g., doctors and
patients, landowners and tenants, employers and employees). Since the allocation of
such legal relations is always beneficial or burdensome to members of some groups
(and different allocations benefit/burden members of different groups), it is clearly
answerable to principles of justice, including non-corrective ones, such as allocation in
accord with suffering, need, or ability. However, since these relations are themselves
distinctly corrective, the problem/function of distributing them is incidental to that of
correcting tortious interactions or losses. Thus corrective justice enjoys a certain
explanatory priority over non-corrective distributive justice in tort law.28
A related view says that the way in which tort law or the courts allocate the loss
between the parties to the tort litigation is sensitive to principles of distribution,
including non-corrective such principles. However, what creates the problem and
restricts it to the parties is tort law’s duty of reparation. Here too tort law’s distributive
function is merely incidental to its corrective function.29
These claims seem true as far as they go. Tort law cannot fail to create and face
certain problems of distribution between members of different interest groups,
between the litigating parties, and, I would add, between the parties and others simply
by virtue of implementing its first and corrective principle of justice. So to the extent
that tort law’s response to such problems takes account of non-corrective principles, it

28
Cane writes that, while liability rules affect the distribution of benefits and burdens, the goods being
distributed are not “material goods” but reparative rights and responsibilities. He takes this to show that, in tort
law, corrective justice is logically prior to distributive. “Corrective justice provides the structure of tort law
within which distributive justice operates.” “Distributive Justice”, 412–13, 416. Gardner writes that what tort
law allocates “is access to a special apparatus for the doing of justice in another form, viz. corrective justice.”
“Distributive Justice”, } 2.
29
This view is adapted from Gardner, “Distributive Justice,” }3: reasons of corrective justice explain “why
we might be confronted with questions of distributive justice that are already pre-localized and already assume
the context of a bilateral zero-sum conflict . . . Once again, it is corrective justice that has the explanatory
priority.”
380 HANOCH SHEINMAN

is incidental to tort law’s corrective response. But this admittedly accurate claim makes
no mention of a tort law principle of distributive justice and leaves out the genuinely
distributive nature of its first and corrective principle.
In the previous sections, I have tried to characterize tort law’s first principle of
distributive justice. Tort law’s distributive justice on this view is not a by-product of its
first and corrective principle of justice; it just is that principle.

VIII. The Mixed Tort Law

My characterization of tort law’s justice assumed a highly idealized conception of the
subject matter. This idealization is a standard criticism of the view that tort law’s first
principle of justice is corrective. It is now a platitude that our tort law is a “mixed
system.” For our purposes, the key observation is that major tort law mechanisms are
now governed by non-corrective principles of distribution. The question is whether my
claim about the distinctly corrective nature of tort law’s justice and tentative remarks
in favor of the priority of corrective justice are compatible with this observation. I raise
the problem by retelling a brief story. I then examine a possible response. I close by
making a tentative proposal.

A. The story of tort law

Part 1. At the beginning there was a tort law system that pursues corrective justice (i.e.,
corrective redistribution) single-mindedly (the pure tort system). It is not surprising
that such single-minded pursuit has some unhappy distributive effects. After all, what
makes corrective justice a distinct principle of distribution also makes it indifferent
to many ethically relevant features of the distribution problem, features that bear on
the justice of our response. Features that are not directly related to the underlying
wrongful interaction itself, including suffering to be alleviated, needs to be satisfied,
and harm to be prevented are simply irrelevant. The single-minded and quite
successful achievement of corrective justice, so the story goes, comes at a real price
in the currency of distributive justice. It is possible to maintain that implementing
corrective justice can lead to major distributive injustice. Suppose that A’s moment-
ary tortious carelessness results in P’s massive transferrable loss. It is possible to
maintain that corrective justice would require shifting that loss to A. That would
typically be a major distributive injustice. As Jeremy Waldron complains: “It is hard
to explain why this is a fair price of a moment’s carelessness.”30 But often the price of
our system’s pursuit of corrective justice lies simply in the failure to satisfy important

30
Waldron, “Moments of Carelessness,” p. 388.
 TORT LAW AND DISTRIBUTIVE JUSTICE 381

non-corrective principles of distribution. The crucial point is that the purely cor-
rective system has major distributive side effects.
Part 2. In the second part of the story, mass-redistributive mechanisms develop in
response to the said side effects. Their main function (as we might say) is to ameliorate
the distributive side effects of doing corrective justice. We can focus on one notable
such mechanism, liability insurance.31 When the tortfeasor A is insured against
liability in tort, the tort victim’s (P’s) loss is transferred not to P but to his insurance
company C, which spreads it in small portions across its policyholders D. The upshot
of the insurance turn is a mixed modern tort law system that comprises the corrective
justice system of old plus the non-corrective redistributive insurance mechanism.
The end.
The story raises a challenge for the proposed account of tort law’s first principle of
justice and for any priority of corrective justice that it can be said to support. For a
large number of cases, insurance has replaced corrective redistribution of P’s loss to A
with non-corrective redistribution of that loss to third parties C and fourth parties D.
Such redistribution is not bilateral (the loss is not redistributed between A and P).
In particular, it does not have the P to A direction of burdening. So it is no longer
distinctly corrective under the proposed account.

B. Implementing corrective justice?

It is natural to reply that insurance simply implements corrective justice. Since C is
a contractually authorized agent of the tortfeasor A, C can discharge A’s corrective
justice liability on his behalf.32 The response underestimates the distance between
the complex multilateral insurance operation (the transfer of P’s loss to C and then
to D) and the bilateral corrective transfer it replaces. One difficulty is that the
transfer of P’s loss to C is twice removed from its transfer from to A. First, C is no
longer a party to the interaction with P. And second, it isn’t a moral agent in the
primary sense.

31
Tort experts claim that we cannot understand tort law without understanding insurance. Here is a recent
statement:
The tort liability and insurance systems are very much like the two suns in a binary star, dependent on each
other for their position in our legal system. For more than a century these two systems have influenced each
other’s course of development. Neither would be anything like what it is today if the other had not existed
and developed along with it. Today the two systems constantly interact, and almost no effort to understand
or reform one of them can take place without understanding the role played by the other.
Kenneth Abraham, The Liability Century: Insurance and Tort Law from the Progressive Era to 9/11 (Cam-
bridge, MA: Harvard University Press, 2008), 1.
32
Cf. Gardner, “Corrective Justice.”
382 HANOCH SHEINMAN

But the main difficulty with the response is that C is not the ultimate bearer of the
loss. In effect, C is a mechanism for transferring the loss onto the many policyholders
D, who were neither involved in the wrongful interaction with P nor authorized agents
of A. By this point, the imagery of bilateral redistribution between P and A so as to
make the situation right again no longer seems apt.

C. Priority in a mixed system

An alternative way to reconcile the distinctly corrective nature of tort law’s
first justice with insurance against tort liability starts by openly acknowledging
that mass redistrbution through insurance displaces rather than implements
corrective justice. Still, the non-corrective redistributive institution of liability insur-
ance presupposes liability to insure against. That is the liability the tortfeasor would
have if he were not insured. It is not essentially different from the liability she would
have under the pure tort law of old. That basic liability is still essentially corrective.
Therefore, the non-corrective redistributive part of the mixed system—and hence the
mixed system as a whole—presupposes its corrective part.33
These observations raise a more interesting priority claim, one that concerns the
justification of the system.34 The basic claim is that our mixed tort system—
understood as our substantive tort law together with the redistributive mechanisms
that presuppose that law—is justified only if our mixed system would be justified even
apart from the mechanisms. Put another way, the justification of the substantive
subsystem (i.e., its being independently justified) is a condition on that of the mixed
system (i.e., on the combined system’s being justified).
Since we are assuming that our substantive tort law is corrective and the mech-
anisms are not, we can claim that corrective justice enjoys the following priority in
the justification of our mixed tort law: The justification of the corrective tort
subsystem is a condition on that of the mixed system as a whole—the mixed system
is not justified unless it would be justified even apart from its non-corrective
redistributive subsystem. For completion, we should add that the converse does
not hold: It is not the case that the justification of our non-corrective redistributive
subsystem is a condition on that of the mixed system as a whole. (That is simply
because you cannot have the redistributive subsystem without the corrective one; the
former presupposes the latter.)
The entertained priority claim is not obviously correct, and I find it hard to provide
it with some deeper rationale.35 But suppose it is correct. Then it can have non-trivial

33
The converse claim cannot be made without threatening regress.
34
I am indebted to David Enoch for the suggestion.
35
An analogy might help. Our mixed criminal law system includes the substantive criminal law as well as
subsystems that presuppose it, for instance the institution of plea-bargaining. It is plausible to maintain that
 TORT LAW AND DISTRIBUTIVE JUSTICE 383

normative implications for our mixed tort law system. The key question is whether the
corrective justice component of our mixed tort law (the “substantive” law of torts) is
justified, apart from the non-corrective redistributive mechanisms that supplement it.
Would our tort law still be justified if we did not have liability insurance? This is
similar to asking whether the pure corrective system of our story is justified. The story
assumes that insurance ameliorates some significant distributive side effects of doing
corrective justice. If these side effects are so serious that our tort system would not be
justified without their amelioration, then it is not justified with that amelioration,
either.

IX. Conclusion
In this chapter I have offered an account of the distinction between corrective and
distributive justice with an eye to civil law. The account does away with the assump-
tion that they are mutually exclusive categories. Corrective justice in this account is a
principle of distributive justice.
Corrective justice concerns the distribution of the person-affecting consequences of
some directed wrongful interactions. What distinguishes corrective justice from other
principles of distribution is its backward-looking sensitivity to the nature of the
underlying interaction. Proper such sensitivity requires redistribution of the relevant
consequences the interaction had for its two parties between them alone in the
direction that would burden the wrongdoing agent by benefitting the wronged patient
(warranting the wrong-righting imagery). It does not require retransferring any of the
relevant consequences back to its original holder or restoring the original pattern of its
distribution (“arithmetic equality”).
This account explains how tort law reparation is fully distributive as well as fully
corrective. The wrongful interactions of tort law are harmful. Corrective justice
requires redistributing the victim’s transferrable losses to the tortfeasor and limits
the scope of the requisite redistribution by reference to the victim’s agential contribu-
tion to the tortious interaction or loss.
The account suggests that some of the more interesting claims about the relative
place of corrective and distributive justice in tort law can be understood as claims
about corrective and non-corrective principles of distribution. Corrective justice is
the only principle of distribution that can be said to be tort law’s first principle of
justice.
Finally, the chapter entertains the possibility of reconciling the priority of
corrective justice with the non-corrective redistributive mechanisms of our
mixed tort system, such as liability insurance. It claims that while liability

our mixed criminal law is unjustified unless our substantive criminal law is independently justified, that is to
say unless our system would be justified even apart from the plea bargaining subsystem.
384 HANOCH SHEINMAN

insurance displaces corrective liability in many cases, it presupposes such liability.

And it raises the possibility that the justification of the corrective component of
our tort system is a condition on that of our mixed system as a whole such that if
our tort system would not be justified without insurance, then it is not justified
with it, either.
 PART IV

SKEPTICAL
PERSPECTIVES
 18
Finding No Fault with Negligence
Heidi M. Hurd*

I. Introduction
According to the view embraced by most tort theorists who take the goal of accident
law to be corrective justice, there is no justice in imposing liability “strictly”—that
is, without a finding of fault on the part of persons whose actions cause injury to
others. When the costs of accidents are transferred to parties whose negligent failure to
appreciate the risks of their actions was a proximate cause of those accidents, tort law
is vindicated by the corrective justice that it achieves. I shall argue in this contribution,
however, that one cannot morally condemn people for failing to advert to risks, and
thus, that negligence liability fails to achieve corrective justice. Put differently, inas-
much as one cannot find fault with negligence, negligence liability is in fact a form of
strict liability. Those who embrace and advance a corrective justice theory of tort law
thus have three choices: (1) they must give up their conviction that tort law should be
in the business of correcting injustices; or (2) they must give up their conviction that
injustices occur only when persons cause legally recognized harms to others in
ways that are blameworthy; or (3) they must denounce negligence liability in tort
law and work towards the adoption of doctrinal requirements that genuinely map
civil liability onto conditions of fault—by, for example, requiring that defendants
be at least reckless (if not knowing or purposive) with regard to the harms their
actions cause.
It is the burden of this contribution to make this tripartite choice inescapable; that
is, to defend the conclusion that negligence liability is not fault-based liability. Let me
begin, however, by disentangling tempting misinterpretations of my project. First, it

* I am grateful to the participants in the 2012 Rutgers Tort Theory Conference for their illuminating
questions and comments during their discussion of this contribution, and particularly to Ken Simons for his
penetrating critique of its principal arguments. I am also grateful to Larry Alexander, Kimberly Ferzan, and
John Oberdiek for motivating this piece, and to Michael Moore for allowing me to here extend, stretch, and
sometimes alter arguments that he contributed to our earlier joint work on the blameworthiness of negligence
in criminal law.
388 HEIDI M . HURD

would be natural, but wrong, to assume that the thesis I am advancing is simply the
standard claim that Oliver Wendell Holmes made so eloquently in 1881 in The
Common Law,1 namely, that sometimes an objective standard of conduct of the sort
captured by “the reasonable person test” will result in the imposition of liability on
those who cannot meet such a standard. Certainly anyone genuinely concerned with
ensuring that tort liability tracks moral fault is embarrassed by tort law’s quite explicit
willingness to punt such a concern when the going gets tough—that is, when it gets
hard to determine whether someone could not, or simply did not, live up to our
expectations. As Peter Westin puts the point: “To assess an actor’s individual blame-
worthiness by idealized standards that make no allowance for traits over which he has
no control is to risk blaming the blameless.”2 But my project is not to suggest that
corrective justice theorists (merely) have to live with the embarrassment of border-
line cases. My claim is that, however “average the man,” and however much
“ordinary intelligence and prudence” he has, he is without moral blame if the
harm he causes another is a product of genuine inadvertence (however unreason-
able) to the riskiness of his own behavior. In other words, all cases of negligence are
cases of strict liability—not just purportedly exceptional, textbook cases like Vaughn
v Menlove3 that do the equivalent of imposing liability on foolish farmers whose
harmful behavior is a result of “the misfortune of not possessing the highest order of
intelligence.”4
Second, and relatedly, it might be supposed that my project is to persuade readers
that if tort law adopted a subjective standard of reasonableness when assessing the
blameworthiness of a defendant who has caused harm, complaints about negligence
liability would dissolve. But my claim is that however particularized our assessment of
a person’s choices becomes in light of the traits and talents she possesses, we will not be
able to defend the assignment of blame to anyone who failed to advert to a risk that
sadly materialized. In short, negligence liability would remain an embarrassment even
if we were to take a fully subjectivized approach to the question of whether a defendant
could have, and should have, done other than he did.
Finally, some might be puzzled by the fact that I characterize negligence as inad-
vertence to the unjustified risks that a defendant’s conduct imposes on others. Is it not
the case that in many negligence cases, defendants in fact adverted to the relevant risks
of their behavior, but simply failed to assign the right values to the variables captured

1
If, for instance, a man is born hasty and awkward, is always having accidents and hurting himself or his
neighbors, no doubt his congenital defects will be allowed for in the courts of Heaven, but his slips are no less
troublesome to his neighbors than if they sprang from guilty neglect. His neighbors accordingly require him, at
his proper peril, to come up to their standard, and the courts which they establish decline to take his personal
equation into account . . . The law considers, in other words, what would be blameworthy in the average man,
the man of ordinary intelligence and prudence, and determines liability by that.
Oliver Wendell Holmes, The Common Law (Boston, MA: Little, Brown, and Co., 1881), 108.
2
Peter Westen, “Individualizing the Reasonable Person in Criminal Law,” 2 Criminal Law and Philosophy
137 (2008), 143.
3 4
Vaughan v Menlove (1837) 132 ER 490. Vaughan v Menlove (1837) 132 ER 490, at 492.
 FINDING NO FAULT WITH NEGLIGENCE 389

by Learned Hand’s famous formula,5 or otherwise failed to weight the risks as a

reasonable person would weight them? My answer is twofold. When defendants act
with conscious awareness of the fact that they may be creating an unjustifiable risk,6
they can be thought culpable for the harms that causally ensue. Such cases do not
invite any moral problems, for they do not invite faultless liability. However, in some
cases, defendants may advert to relevant risks, but they may fail to advert to factors
that would give a reasonable person grounds for assigning greater value to the
probability that those risks will materialize or to the gravity of the harm that will
ensue, or they may over-weight the value of the burdens that would have to be
shouldered in order to avert such risks. I am happy to include such cases in the
category of negligence cases, and to thus count such cases as among those that involve
faultless wrongdoing.7
My thesis that carelessness is blameless should now be clear, even as it will likely
remain counter-intuitive. It follows, however, from the essential fact that those who
merely negligently cause harm do not choose to cause such harm, nor do such agents
choose their harmful actions in light of any conscious awareness of the risks to others
that attend those actions.8 Some may be quick to respond by jettisoning the connec-
tion between choice and blame, but before we pursue this move it is crucial to
appreciate why legal theorists of both utilitarian and rights-based persuasions appear
committed to, and often advance, the position that choice is an essential precondition
of moral blame and civil liability.
If the highest and best purpose of law is to deter socially harmful (cost-inefficient)
conduct and to encourage socially beneficial (cost-efficient or wealth-maximizing)
activities, then it would seem that it should attach penalties only to actions that are
chosen by actors in light of their risks. After all, the law cannot affect those who are
oblivious to its sanctions and rewards. Inasmuch as those who fail to advert to risks
also inevitably fail to advert to the legal consequences that will attach if those risks

5
“If the probability be called P; the injury, L; and the burden [of precautions], B; liability depends upon
whether B is less than L multiplied by P: i.e., whether B [is less than] PL.” United States v Carroll Towing Co.
159 F 2d 169 (US Ct of Apps (2nd Cir), 1947), 173.
6
I am borrowing here the language employed by the Model Penal Code to unpack the concept of
recklessness. American Law Institute, Model Penal Code Official Draft, 1962, } 2.02 (2)(c).
7
While I am employing the language of Learned Hand’s calculus of risk here, I am fully agnostic for the
purposes of this chapter about the appropriate characterization of the concept of negligence. Some tort scholars
take negligence to involve cost-inefficient risk-taking; others take it to constitute a violation of the standard set
by a hypothetically reasonable person; still others characterize it as action that externalizes costs and
internalizes benefits in ways that result in unjust enrichment; and there are those who have advanced further
construals that reflect values that extend beyond utility-maximization, rights-protection, and the prevention of
unjust enrichment. See, e.g., Richard Wright, “Justice and Reasonable Care in Negligence Law,” 47 Am. J. Juris.
143 (2002); Richard Wright, “Negligence in the Courts: Introduction and Commentary,” 77 Chi.-Kent L. Rev.
425 (2002); Richard Wright, “Hand, Posner, and the Myth of the ‘Hand Formula’,” 4 Theoretical Inquiries in
Law 145 (2003).
8
For classic statements of this claim in the context of the criminal law, see Jerome Hall, “Negligent Behavior
Should be Excluded from Penal Liability,” 63 Colum. L. Rev. 632 (1963); Glanville Williams, Criminal Law: The
General Part, 2d ed. (London: Stevens and Sons, 1961).
390 HEIDI M . HURD

materialize, the law seemingly cannot affect the incentives of those who cause harm
accidentally. It would thus appear that card-carrying utilitarians—most obviously in
this context, those within the law-and-economics tradition in tort theory—must
reserve liability only for those who have made informed choices about their harmful
conduct, on pain of the law missing its only morally legitimate mark.
Similarly, choice would seem pivotal to those who reject the utilitarian ambition to
use liability as a means of social engineering, and who instead take liability to be
justified only when it achieves corrective justice—that is, only in circumstances in
which its imposition is morally deserved. One is typically not thought to deserve a
penalty, or even disapprobation, unless one has perpetrated a wrong (a rights viola-
tion) in a manner that is culpable. On standard theories of responsibility, an actor’s
relative culpability is a function of her mental state: she acts culpably if, in her mind,
she is doing an act that is, as an objective matter, wrongful. She need not draw the
conclusion that it is wrongful, but she does have to have a description of the act in her
mind that embodies the characteristics of the act that make it objectively wrongful.
Of course, both utilitarian and desert-oriented tort theorists do advocate penalties
for those who cause harm accidently. The question is how they justify the imposition
of liability on those who have not chosen their harmful actions under any description
that might trigger their appreciation that their actions might be harmful. For example,
when Richard Posner insists that the imposition of tort liability should “bring about
the level of accidents and safety that the market would bring about if transactions were
feasible,”9 he is surely presupposing that somehow the threat of tort liability will
penetrate the thoughts of actors who are, ex hypothesi, without thoughts concerning
the prospect of causing accidents so as to induce them to do what a market transaction
would otherwise induce them to do were bargaining feasible. How does he square this
theoretical circle? And Holmes famously insisted that tort law would have no reason to
adopt the voluntary act requirement10 if it did not take choice to be morally relevant.
“The requirement of an act is the requirement that the defendant should have made a
choice.”11 As he argued, “the only possible purpose of introducing this moral element
is to make the power of avoiding the evil complained of a condition of liability.”12 And
“[a] choice which entails a concealed consequence is as to that consequence no
choice.”13 Inasmuch as a negligent actor ex hypothesi acts in the face of a concealed
consequence (the consequence of harming another), it would appear that Holmes
would have to admit that negligence lacks the touchstone requirement for tort liability.
But he does nothing of the sort. By what witchery does he escape the horns of the
dilemma that he himself appears to have crafted?

9
Richard Posner, “A Theory of Neligence,” 1 J. Legal Stud. 29 (1972), 37.
10
The voluntary act requirement demands that the plaintiff ’s harm have been the product of willed bodily
movements on the part of the defendant.
11 12
Holmes, The Common Law (note 1) at 95. Holmes, The Common Law (note 1) at 95.
13
Holmes, The Common Law (note 1) at 94.
 FINDING NO FAULT WITH NEGLIGENCE 391

In what follows, I will outline five accounts of how theorists might defend (and in
some cases, have defended) negligence law’s indifference to the absence of choice on
the part of defendants who have caused harms. To summarize in advance, apologists
of the law might argue that we are entitled to blame, and therefore to transfer the costs
of accidental injuries, to negligent actors because: (1) they deliberately chose to violate
per se precautionary rules that exist to safeguard others from inadvertent injury, and in
so doing they culpably engaged in objectively faulty conduct, even if they did not
subjectively appreciate its risks; or, (2) they deliberately chose to do acts that they knew
would make their later inadvertence to risks unjustifiably likely, and we can properly
blame them for those prior culpable choices, even if we cannot blame them for the
inadvertence that later attended their injurious actions; or (3) while they made no
relevant choices with regard to the accidental injuries they caused, they possessed
unexercised capacities, which if exercised, would have prompted them to choose a
risk-free course of conduct; or, (4) they possessed defective motor skills, cognitive
capacities, or volitional resources which were themselves morally blameworthy or
which, if not themselves blameworthy, were causally significant to the accidental
injury in ways that could have been prevented; or (5) they possessed defects of
character which were themselves blameworthy, and which caused their failure to
advert to the risks of their behavior, thus making such inadvertence blameworthy.
As I shall argue, none of these arguments succeed in grounding the claim that
negligence is morally blameworthy. None thus provide a basis for thinking that
negligence liability is fault-based. Absent any further account of how inadvertent
injuries are culpably caused, tort law and tort theorists should confess that negligence
liability is just another species of strict liability.

II. Finding Fault in the Action, Not the Actor

The first move that might be (and has been) tempting for jurists and theorists alike, is to
reject the presupposition that the fault-standard in tort law has anything to do with the
moral fault, or culpability, of the defendant. The argument is that the fault in question lies in
the action, not the actor, and so recites numerous torts textbooks.14 According to this
argument, certain actions are objectively or inherently wrongful—e.g., vigorously bouncing

14
Negligence as risky conduct. In modern law, the term negligence in its primary meaning merely
describes conduct. In that sense it has come to mean conduct that is unreasonably risky, such as
driving a car at high speed.
Dan B. Dobbs, The Law of Torts, Vol. 1 (Practitioner Treatise Series) (St. Paul, MN: West Group,
2001), 269.
Negligence Contrasted with Liability Without Fault. The fact that negligence as here defined is
conduct which falls below the standard of behavior established by law for the protection of others
carries with it the idea of social fault.
Robert E. Keeton, Lewis D. Sargentich, and Gregory C. Keating, Cases and Materials on Tort and Accident Law
(American Casebook Series) (St. Paul, MN: West Group, 1998), 173.
392 HEIDI M . HURD

a baby—and those who choose to engage in them are properly blamed, whether at the time
of action they advert to the risks of those actions or not. When those wrongful actions result
in injury to others, the liability that follows is properly thought to be “fault-based” and
properly thought to satisfy the demands of corrective justice.
According to this view, tort law—at least in theory—could mimic the criminal law.
It could follow Holmes’ famous advice to fellow judges in Baltimore and Ohio R.R. v
Goodman,15 and articulate an ever-expanding list of per se rules that capture the moral
injunctions violated by those who perform such objectively wrongful actions—e.g.,
“[s]top, look, and listen before crossing railroad tracks!” But given the burdensome
nature of such a task, and the fact that the law would likely become less, rather than
more, knowable by laypersons who need easy legal heuristics, tort law instead captures
the essential nature of such wrongful deeds by simply declaring them to be of a sort
that a reasonable person would not commit. As this argument goes, the law’s use of the
notion of “reasonableness” in its short-hand summary of objectively negligent (or
faulty) actions, however, should not be mistaken for the adoption of a culpability
requirement. It does not reflect any notion that the actor must have been at fault for his
action; it reflects only the notion that the act that in-fact and proximately caused injury
must be of an inherently or objectively faulty sort.
Such a view allows for a notion of moral fault that, on its face, distinguishes
negligent actions from pursuits that, if actionable, result in strict liability. However,
it simultaneously resists equating fault with subjective culpability. Actions that might
invite strict liability—like blasting tunnels or transporting toxic substances—are not
inherently wrongful on such a view, and thus, were tort law to make a list of actions
that are per se prohibited, it would not include prohibitions against such activities.
But—as the argument continues—it would include such prohibitions as: look both
ways before crossing the street; always wear a seatbelt in a moving vehicle; never point
a gun at another person, even if it is unloaded; never shake an infant; don’t dress
children in flammable pajamas; don’t keep unlabeled medicines in the medicine
cabinet, etc. Because all such actions inevitably involve risks of a greater discounted
value than the benefits they bestow, all are inherently faulty or “unreasonable.” And
thus all persons who choose to perform such actions are (objectively) at fault.
Yet objective fault would appear to be as self-contradictory as heavy metal music. For
objective fault without a morsel of subjective fault would appear to imply no fault at all. To
find fault in an agent’s actions but not in the agent is not to find fault in the relevant sense
in which corrective justice theorists standardly hope that negligence is fault-based—
namely, in the sense that allows them to say that negligence is a basis for desert, a form of
culpability, a foundation for moral blame. To hold people liable for actions absent any
finding that those actions were accompanied by culpability is to hold them strictly liable.
So let me give those who are keen to advance this argument a better version of it.
Let me suppose that what they mean is that when defendants choose to violate known

15
Baltimore & Ohio R.R. v Goodman 275 US 66 (1927).
 FINDING NO FAULT WITH NEGLIGENCE 393

per se prohibitions—when they deliberately opt to do deeds that are enjoined by

precautionary “mini-maxims”—they can properly be held liable, regardless of whether
they have adverted to, or have appreciated, the risks of such per se violations. Yet, on
this argument, such mini-maxims cannot be thought to be mere rules of thumb—mere
heuristics or epistemic proxies—for were they but efficient epistemic constructs, there
would be circumstances in which it would be fully appropriate to second-guess them
and to set them aside as unhelpful. Inasmuch as there is no moral wrong, as such, in
violating a mere rule of thumb, this account must construe the rules that make certain
acts wrongful to have categorical, or deontological, force. Only then will their violation,
itself, be wrongful, so as to generate objective fault despite the absence of subjective risk-
advertence.
There are, however, at least four problems with equating negligence with the
violation of deontological mini-maxims.16 First, the touchstone of an absolute deonto-
logical prohibition is that it cannot be violated in the name of good consequences (e.g.,
an increase in private utility or social welfare). If the per se rules that are violated in
cases of negligence are genuinely of this order, then they cannot be violated even when
the cost-benefit calculus would attach more benefits than costs to so doing. But surely
one can drive a car at high speed in order to escape from a murderous pursuer. One
can point an unloaded gun at an innocent person if it encourages a bank robber to turn
his back, and so permits one to press a silent alarm. And one can refuse to fully stop,
diligently look, and cautiously listen before entering onto train tracks if one’s purpose
is to get to the scene of an accident in time to deliver life-saving aid.
Second, any account of negligence that seeks to equate it with a chosen violation of
one or more categorical mini-maxims is doomed to gross under-inclusiveness. However
long the list or great its detail, it is sure to leave off actions that in their circumstances
are unreasonably risky. A student of Holmes might reply that as we learn more about
the perils of modern life, we can expand the list so that it continues to reflect our
cutting edge understanding of morality’s demands. But such a reply will not do, for the
expansion of the list of prohibitions that collectively comprise the content of the rule
against negligence can only proceed by means of consequential calculations. But this is
to say that they are not deontologically wrongful at all, for deontology ceases to be
deontology when its maxims are simply ever-expanding means of achieving good
consequences. Moreover, if prohibitions against letting children play beside unfenced
railroad tracks and pointing guns at others are justified solely by their consequential
pay-offs, then courts should always inquire into the discounted value of the costs and
benefits of their violation. And this is to say that precautionary rules can never be more
than rules of thumb, and as such, their violation cannot, in itself, be wrongful.

16
For a longer list of problems with construing negligence as a violation of one or more deontological
maxims, see Heidi M. Hurd, “What in the World is Wrong?” 5 J. Contemp. Legal Issues 157 (1994), 196–8;
Heidi M. Hurd, “The Deontology of Negligence,” 76 B.U. L. Rev. 249 (1996), 266–8; Heidi M. Hurd and
Michael S. Moore, “Punishing the Awkward, the Stupid, the Weak and the Selfish: The Culpability of
Negligence,” 5 Crim. L. Phil. (2011), 147–98.
394 HEIDI M . HURD

The third problem with trying to locate the fault of negligence in the chosen
violation of a rule of conduct is that it converts all instances of negligence into instances
of higher-order culpability and so solves the problem by changing the subject. Put
differently, we do not account for the blameworthiness of negligence by converting it
into recklessness. We only account for the blameworthiness of negligence if we can
explain why harms that are never anticipated by those who cause them nevertheless
invite moral condemnation. Proponents of this first “fault-objectification strategy”
might respond by insisting that accidents are never thought to be blameworthy unless
and until they occur after precautionary mini-maxims have evolved that, if followed,
would prevent such accidents, and only if such maxims are deliberately ignored. But
this is just false. There is no well-worn rule that, if followed, would ensure that one
would never inadvertently leave a bottle of prescription pills within reach of a suicidal
visitor. And yet, when a teenager with known suicidal tendencies overdoses on easily
available pills, many will feel fully entitled to demand recompense from the home-
owner whose inadvertence so predictably in-fact and proximately contributed to the
tragedy, and any effort to retro-engineer a precautionary rule is just disingenuous
Monday morning quarterbacking.
Fourth and finally, even if every case of negligence involved chosen violations of
known precautionary rules, those who would press the strategy of substituting a
defendant’s earlier rule violation for the risk-advertence that is missing at the time
of the accident invite the charge that they are engaging in a kind of moral gerryman-
dering akin to that which is used to defend such blatantly unfair rules as the “felony-
murder rule” in criminal law. By substituting a less serious mental state (the defend-
ant’s contemplation and dismissal of a precautionary rule) for a more serious mental
state (conscious awareness of a risk that conduct will cause harm to another), this first
effort to dissolve the contradiction between the conditions of blame and its attachment
to negligence in fact licenses strict liability. For this strategy permits liability grossly in
excess of the only moral shortcoming that is thought to be its predicate.

III. Finding Fault in Prior Risky Actions

The first strategy for dissolving the contradiction between insisting on choice as a
predicate of blame and blaming people for negligence sought to find a choice in every
case of negligence. The second strategy does the same thing. But rather than pinning
liability on a prior choice to ignore a per se precautionary rule, this strategy seeks a
prior choice to engage in conduct which, at the time of choice, was known to derive its
riskiness from the likelihood that it would later lead to inadvertent injury. Such is the
strategy in what I shall call “Decina-like cases”—cases similar to that of People v
Decina,17 in which the defendant, knowing that he was subject to epileptic attacks that

17
People v Decina 138 NE 2d 799 (NY Ct of Apps, 1956).
 FINDING NO FAULT WITH NEGLIGENCE 395

would cause him to lose consciousness for considerable periods of time, nevertheless
chose to drive his Buick on a public highway, suffered just such an attack, lost control
of his vehicle, and killed four pedestrians.18 The court in Decina had no trouble finding
the defendant guilty of manslaughter, for all of the elements of the prima facie case
were present and in compliance with the simultaneity principle: namely, the defend-
ant’s voluntary act of setting the Buick in motion was simultaneously accompanied by
the culpability of recklessness—conscious awareness of a substantial and unjustifiable
risk that such a deed would result in others’ deaths—and such deaths were the in-fact
and proximate results of such recklessness.
There is nothing conceptually confused or morally illicit about this strategy. If one
can go back in time and find an earlier choice that was itself culpable, and that in-fact
and proximately caused accidental injury to another, then one can blame the injurer
for that prior choice, even if one cannot blame her for any inadvertence at the time the
accident occurs. If all cases of negligence were Decina-like, we could surely find fault in
negligence. The fault found would be located at time t1, when a choice was made to
engage in conduct that bore the known risk of in-fact and proximately causing later
inadvertent injury, rather than at time t2, when the predictable inadvertence resulted
in the predictable injury. But apart from the fact that Decina-like cases form a tiny
fraction of accident cases, Decina-like cases are not cases of negligence at all, so it is no
surprise that we can so easily find fault in them. One who knows that his conduct bears
an unjustifiable risk of later making him an inadvertent injurer is reckless with regard
to the later injury he in-fact and proximately causes. The gravamen of our complaint
is that he recklessly rendered himself incapable of advertence; it is not that he was
inadvertent at the time of injury. Precisely because Decina-like cases are not cases of
negligence, they do not constitute examples of cases in which negligence is manifestly
blameworthy so as to provide counter-examples to my claim that one can find no fault
in inadvertent wrongdoing.

IV. Finding Fault in the Failure to Exercise

Capacities of Advertence
Consider now the third account of how negligence can be thought to be blameworthy.
Herbert Hart famously understood that any satisfactory account of the blameworthi-
ness of negligence had to divorce itself from choice-based accounts of culpability.19

18
Others call the strategy in Decina the “tracing strategy,” for one traces a defendant’s legal responsibility
for a harm from some earlier time (when the actor risked harm to others while simultaneously satisfying all of
the conditions for blame) to some later time when the risks actually materialized. See Matt King, “The Problem
with Negligence,” 35 Soc. Theory and Prac. 577 (2009); Manuel Vargas, “The Trouble with Tracing,” 25
Midwest Studies in Philosophy 269 (2005).
19
This discussion of Hart’s theory of unexercised capacities draws heavily from Michael Moore’s analysis in
Hurd and Moore, “Punishing the Awkward, the Stupid, the Weak and the Selfish: The Culpability of
Negligence,” (note 16) at 162–5; Heidi M. Hurd and Michael S. Moore, “The Culpability of Negligence,” in
396 HEIDI M . HURD

He appreciated that it is not possible to identify culpable choices in all (or even many)
cases of negligence, and were it otherwise, he would still need to explain why the moral
outrage we experience in response to a case of negligence very typically takes as its
object the injurer’s inadvertence to the riskiness of her conduct, rather than any of her
prior choices that might be causally connected to the moment of another’s injury. Hart
thus advanced the claim that there are, in fact, two sorts of culpability: the culpability
of choice and the culpability of unexercised capacity.20 As Hart reasoned, if a person
had the capacity to do better than she actually did—as measured by her own capacities
but not others’—she can be held morally and legally responsible for her actions. It follows
that a person can properly be held liable for a failure to advert to the unjustifiable risks
that her conduct imposed on others if, but only if, she had the capacity to advert to those
risks before they materialized.
Hart’s notion that persons have, and can be blamed for, unexercised capacities is
very tempting. Indeed, many of our daily moral judgments about others appear to
invoke this notion. We blame our children for failing to remove roller-skates from
sidewalks; we blame distracted drivers for failing to watch where they are going; we
blame ourselves for failing to remember important appointments. And in every case,
we clearly presuppose that we and others could have done otherwise, and as a result,
should have done otherwise.
Some are convinced that these daily moral practices are just confused, and so is Hart
for seeking to validate them. Larry Alexander and Kimberly Ferzan, for example, insist
that no one can advert to risks not actually adverted to, and thus, that talk of having
unexercised capacities of advertence makes no sense.21 Their argument proceeds from
two necessity claims: first, we can advert to a risk only if we can choose to advert to the
risk; and second, no one can choose to advert to a risk to which they have not already
adverted, for plausibly enough, one cannot will one’s beliefs.22 But Alexander and
Ferzan beg the question against Hart by substituting the capacity to choose to advert
for the capacity to advert. Seeking an alternative basis for culpability judgments, Hart
introduced the notion of an unexercised capacity as distinct from choice. Reinserting
choice as the object of the capacity one has to have in order to have the capacity that
Hart was talking about guarantees only a pyrrhic victory over Hart.

Rowan Cruft, Matthew H. Kramer, and Mark R. Reiff (eds.), Crime, Punishment, and Responsibility (Oxford:
Oxford University Press, 2011), 321–31.
20
H.L.A. Hart, “Negligence, Mens Rea and Criminal Responsibility,” in H.L.A. Hart, Punishment and
Responsibility: Essays in the Philosophy of Law (Oxford: Oxford University Press, 1968), 136–57. Others
followed Hart in finding a culpability of unexercised capacity in cases of negligence. See, e.g., Jeremy Horder,
“Criminal Culpability: The Possibility of a General Theory,” 12 Law and Philosophy 193 (1993).
21
Larry Alexander and Kimberly Kessler Ferzan, Crime and Culpability (Cambridge: Cambridge University
Press, 2009) 83–5.
22
To give Alexander and Ferzan company here, consider the court’s similar claim in Breunig v American
Family Insurance Co., that “it is unjust to hold a man responsible for his conduct which he is incapable of
avoiding and which incapability was unknown to him prior to the accident.” 173 NW 2d 619 (Wis Sup Ct,
1970), 624.
 FINDING NO FAULT WITH NEGLIGENCE 397

While Hart’s conviction that we can rest the culpability of negligence on the claimed
culpability of unexercised capacity cannot be dismissed in the way that Alexander and
Ferzan seek to dismiss it, there is no question that it must sustain some fairly rigorous
poking and prodding. What does it mean to say that a defendant had a capacity to
advert to the risk of another’s injury, and so had the capacity to prevent that injury in a
manner that makes her culpable for a harm she never, in fact, contemplated? Even if
we can make sense of such a capacity, does the failure to exercise it justify tort liability,
on pain of a failure to exact corrective justice?
On the surface, our claims about others’ capacities appear categorical. Parents’
(maddeningly frequent) lectures to their teenagers inevitably embody such statements
as: “You could have gotten a better grade on your chemistry test,” “You could have
come home from band practice when we told you to do so,” “You could have watched
where you were going,” etc. But as J.L. Austin quaintly suggested, such “coulds” appear
to be “constitutionally iffy.”23 Consider the statement, “You could have gotten a better
grade on your chemistry test.” Presumably if a teenager had the capacity to do better
on the chemistry test, this means that he would have done better on the test if one
or more conditions had obtained. It would make absolutely no sense for a parent to
say to a teenager, “I know that all the circumstances necessary and sufficient for you
to do the best that you could on the chemistry test were present. Still, you could
have done better.” Capacity statements at least require that certain counterfactual
conditions (about what would have happened if certain conditions obtained) be true.
And thus, capacity judgments, however categorically expressed, are ultimately innately
conditionalized.24
Now here is the rub. It turns out to be extremely difficult to specify the appropriate
conditions that make capacity judgments true. If the expression “She would have
A-ed” requires the truth of “She would have A-ed if C,” then we need some sense of
how to give content to C.
Consider one morally-relevant context that might shed light on the problem—that
of separating responsible actors from excused actors. According to the so-called
“principle of alternative possibilities” a responsible actor “is one who would (or
could) have done otherwise if he had chosen to do otherwise.” This principle tests
the power of the will to effect changes in the world. However, while this might be
highly relevant to sorting out those who are properly eligible for intentional tort
liability, it is not the capacity that is specifically relevant to negligence liability. Rather,
the capacity in question in a negligence case is the capacity of the defendant to have

23
J.L. Austin, “Ifs and Cans,” 42 Proceedings of the British Academy 109 (1961).
24
One might resist this claim by insisting that when we attribute general capacities to people, we are not
making conditional claims. One might think that the claim “He can bench-press 340 pounds” is uncondi-
tionally true of someone who has certain physical equipment. But conditions are clearly present, if suppressed,
in such claims. A weight-lifter with a general capacity to bench-press 340 pounds is not likely to do so on every
occasion when the opportunity presents itself. Rather, he can do so if he tries hard, if he is not injured, if he is
not distracted, if he is not dehydrated, if he is not cold, if he is in the mood, etc., and any speaker attributing
such a capacity to him means to include such conditions, even if not explicitly.
398 HEIDI M . HURD

adverted to something that would have given him a reason to will one action over
another. The defendant’s capacity to effectuate what he wills is not in question in such
cases, for the typical negligent actor could have done other than he did, if he had
adverted to the need to do so.
What is needed is something like choice (in its generality and moral relevance) that
is not choice (because it must ground culpability judgments in the absence of choice).
To zero in on such a condition or set of conditions, one does well to think about
candidates that have to do with morally relevant features of the actor. There are four
basic attributes of rational agents that make them, simultaneously, moral agents: what
they desire, what they believe is true, what they intend to do in light of what they desire
and believe, and how they execute their more general intentions via the bodily
movements that allow them to navigate through this world.

V. Finding Fault in the Possession of Defective

Physical, Cognitive, or Conational Attributes
The question is thus whether tort law can justify attaching liability to defendants on
the basis of defects in such agency-defining attributes. If a defendant can be blamed for
(1) physical, (2) cognitive, (3) conational, or (4) motivational lapses, then we can frame
relevant counterfactuals that test whether he would have adverted to the risks of his
actions if he had not suffered from such defects. And we may then be entitled to infer
that he could have adverted to the risks in question in the morally relevant sense of that
term. But notice that if we find any such condition blameworthy, they will themselves
be the desert-determiners in cases of negligence. They will ground the culpability of
negligence directly: the judgment that a defendant had “the capacity to advert to risks
if C” will be a mere implication of the independent moral significance of C.25
I do not propose to give space here to the consideration of the first two of these
failings—that is, to the questions of whether we can morally condemn a person (1) for
being “born hasty and awkward” so as to be congenitally disposed towards “having
accidents and hurting himself or his neighbors,”26or (2) for being “not of the highest
order of intelligence.”27 So long as these failings do not stem from or collapse into the
other two failings above—conational failings (weakness of will) or motivational lapses
(defects of character)—it would seem that those who suffer from physical or cognitive
impairments are poor candidates for moral disapprobation. Of course, those who
possess such impairments are obligated to mitigate their harmful effects by taking

25
Hart did not appreciate this. He thought that the relationship between unexercised capacity and C was
the reverse: namely, it is because we have the capacity to choose otherwise that choice is a source of culpability
for chosen acts. Yet this is backwards: it is because choice makes one culpable that one picks it out as the
implicit conditional in cases in which one claims that another had the capacity to have done other than he did
(for he would have done other than he did if he had chosen to do so).
26 27
Holmes, The Common Law (note 1). Vaughn v Menlove (note 3).
 FINDING NO FAULT WITH NEGLIGENCE 399

available precautions against their own lack of fine motor skills, or their own ignor-
ance, poor memory, short attention span, compromised perceptual abilities, biased
inference-drawing tendencies, wishful thinking, susceptibility to distraction, breaks
with reality, visions and delusions, etc. We can unproblematically blame the blind man
for choosing to drive and Farmer Menlove for choosing to ignore his wise neighbor
who warned of damp hay’s propensity for spontaneous combustion, but this simply
reflects the principle that we can blame people for their choices; it does not mean that
we can blame them for the failings whose effects they ought to choose to mitigate.28
Things get considerably less clear, however, when one turns to the blameworthiness
of weakness of will—that is, to those who demonstrate conational difficulties in acting
on the conclusions that are drawn from the beliefs and desires that form the premises
of their practical syllogisms. Weakness of will is quite commonly a subject of self-
condemnation and an object of blame by others. When St. Paul complained that he
willed the good, but did the bad, he was passing moral judgment on his perceived
weakness.29
Solving the question of how best to understand akrasia (if, as some contest, it is to be
thought a psychological phenomenon distinct from ignorance, indecision, or cognitive
dissonance), is far beyond the scope of our project. And without engaging the
extensive literature on the topic, one can hardly pretend to resolve its morality. But
notice how conceptually thin a phenomenon it must be if it is to be distinguished from
emotional states (like fear) that result in physical inability to act on one’s intentions
(e.g., muscular paralysis); and if it is to be distinguished from ignorance, indecision, or
cognitive dissonance, all of which can result in one’s doing other than one (in some
sense) intended or (in some sense) really wanted to do. What is crucial is to recognize
that if weakness of will is both a distinct phenomenon and morally blameworthy, then
it may provide an eligible basis for asking whether a defendant who inflicted accidental
harm on another would have done otherwise if he had acted as he (truly) desired,
believed, and intended.
I suspect that our response to many cases of weakness of will is a reflection of our
condemnation of the character of those who succumb to temptations. We use their

28
Tort law, of course, does find liability in many cases in which the physical and cognitive impediments of
injurers have played a significant role in their harming others. Only when a deficiency meets the criteria of “the
distinct defect test,” classically articulated in Holmes, The Common Law (note 1) at 109, will it justify the
application of a standard of reasonableness relativized to that condition. But to meet this test, the impairment
must (1) provably impair the defendant’s ability to take precautions of the sort that the iconic “reasonable
person” would take; (2) not be the defendant’s fault; (3) be measurable (in the sense that courts can gather facts
to assess its presence and its deleterious effects on cognition); and (4) be distinct (in the sense that the average
person will not be lulled into believing that she might get the benefit of its application). While the blind man “is
not required to see at his peril,” all sorts of physical and mental impediments that are not moral failings flunk
these conditions, and thus leave those they plague at risk of legal liability. In such circumstances, it is hard to
say that the liability is anything other than strict.
29
“For that which I do I know not: for what I would, that do I not; but what I hate, that do I.” King James
2000 Bible (2003). “For what I am doing, I do not understand; for I am not practicing what I would like to do,
but I am doing the very thing I hate.” New American Standard Bible (1995).
400 HEIDI M . HURD

akrasia as evidence of their possession of character traits on which we pass independ-

ent judgment—sloth, gluttony, greed, cowardice, etc. But this brings us to the question
of whether we are entitled to blame people for bad character, or whether we ought to
think of such traits as being as immune from moral condemnation as color-blindness,
paraplegia, hypertension, autism, mental retardation, and senility.

VI. Finding Fault in Attributes of Poor Character

Can we rightly blame people for the unfortunate character traits that they possess—for
the vices that exert themselves in the form of evil or unworthy desires, base or
gluttonous passions, or ugly prejudices and distorting biases? If unfortunate attributes
of character are eligible bases for assessing fault, then perhaps we can derivatively
blame people for any instances of inadvertence such character flaws induce.
There seems little doubt that the moral evaluation of human character traits is the
preoccupation of our daily gossip with family, friends, and colleagues, and that
it has provided a bottomless well of inspiration and insight for centuries of
great novelists, playwrights, poets, librettists, filmmakers, biographers, theologians,
journalists, and historians. We move very quickly from an evaluation of others’
(often small) deeds to (often sweeping) evaluations of their character. So whatever
the answer is to the question of whether we are entitled to blame people for
their character, the fact of the matter is that we do, and we do so every day, in
every way.30
I want to suppose, for the sake of argument, that we can vindicate the morality of
passing moral judgment on the personalities of one another—on the desires, passions,
and dispositional beliefs that make us “who we are” despite our never having chosen
them, and despite their resistance to our choices to affect their revision. If a person’s
character invites blame and praise independently of the blame and praise that attach to
a person’s choices, we can ask whether tort law might rest the blameworthiness of a
defendant’s negligence on the blameworthiness of her character, at least when an
unworthy disposition was the cause of the inadvertence that resulted in her accidental
injury of another? After all, if she could have (and would have) adverted to the
prospect of injuring another had she been of good character, can we not blame her
for the things she caused in its absence if we can blame her for the absence of good
character?
Were this a discussion of whether we could justifiably criminally punish persons for
their bad character, it would be easy to trot out the arguments that I and others have

30
We judge persons according to their choices, on the assumption that they are responsible for the
motivations which drive those choices. If we were not responsible for our motivations, individual
responsibility would be impossible.
Samuel Pillsbury, “Crimes of Indifference,” 49 Rutgers L. Rev. 150 (1996).
 FINDING NO FAULT WITH NEGLIGENCE 401

made for why such a proposal would smack of a background commitment to an

indefensible theory of governmental authority.31 But I will spare readers this exercise
for I take the question of whether character might be a basis for liability in tort law to
be a very different question from whether it might be a basis for punishment in
criminal law. And the fact that this is so reveals something fundamental about the
difference between tort law and criminal law.
When a wrongdoer is punished, the state is authoring a new harm. It is creating
suffering that would not otherwise exist. It is not surprising that when the state creates
fresh pain, as it does when it uses the sanctions of the criminal law to deprive people of
life and liberty, it must justify so doing by reference to conditions of desert that give us
no pause, and by reference to a political theory that specifies when the state can extend
the long arm of the law so as to take away yet another liberty. Tort law, on the other
hand, need not justify the deliberate invention of fresh losses. Rather, it simply seeks
tipping principles that justify reallocations of already-generated accidental losses. Its
goal, at least according to those (like me) who are not utilitarians, is to achieve
corrective justice—the restoration of the ex ante status quo, or its moral equivalent.
It should require no consultation with political theory to exercise its mandate, for
assuming the ex ante status quo was just (and that is not, itself, a question for tort law),
tort law’s role is solely to restore it. The narrow question that tort law must answer in
any given case is whether there is an applicable moral tipping principle that justifies
transferring a loss to a defendant who failed to foresee that his actions would cause
injury to a plaintiff.
Corrective justice theorists who are eager to find some form of fault at the founda-
tion of negligence law might think, then, that tort law can rely on the culpability of bad
character in a way that the criminal law cannot. They might argue that tort law can
(and perhaps does) use bad character as a reason to tip the balance of damages to a
defendant who has inadvertently caused harm, when such bad character was the
reason for his inadvertence. For as between two parties, let he who caused a harm as
a result of faulty character bear the inevitable loss.
Yet as tempting as this tipping principle is, it is ultimately to be rejected on both
legal and moral grounds. As a legal matter, such a principle fails both descriptively and
prescriptively. It is a poor fit with existing doctrines of negligence law, and the
substantial doctrinal changes that negligence law would have to undergo in order to
realize the goal of calibrating liability to character judgments would be practically
unworkable. And as a moral matter, such a principle should not be mistaken for a
fault-based principle. For it does not follow from the fact that poor character is
blameworthy that the inadvertence caused by such character is thereby blameworthy.
Let me take these arguments in turn.

31
For lengthy discussions of how a character theory of punishment runs afoul of a liberal political theory,
see Heidi M. Hurd, “Why Liberals Should Hate ‘Hate Crime Legislation’,” 20 Law and Philosophy 215 (2001);
Heidi M. Hurd and Michael S. Moore, “Punishing Hatred and Prejudice,” 56 Stan. L. Rev. 1081 (2004).
402 HEIDI M . HURD

First, if the fault within negligence is a function of the blameworthy aspects of the
defendant’s character, how is it that tort law calibrates damages to the magnitude of
the harm caused to the plaintiff, rather than to the magnitude of the moral turpitude
that caused the defendant’s inadvertence? It might be argued, of course, that whatever
tort law’s account of the fault of negligence, it has never calibrated the damage award
owing the plaintiff to the relative fault of the defendant. It has instead employed the
defendant’s negligence (however slight) as a basis for transferring to him the entirety
of the plaintiff ’s loss (so long as the plaintiff was not, herself, contributorily negligent).
It is none the worse for a character-based account of negligence than for any other
account of negligence that this doctrinal regime appears to be wrong-headed, if in fact
it does.
Perhaps this is so. But what I suspect is this: The blameworthiness of the defendant’s
poor character might give those who are fearful of a strict liability standard just enough
moral comfort to paper over their closet intuition that “he who causes harm should
pay for it.” In fact, the heavy lifting that is needed to justify the law’s willingness to
transfer the whole loss when it transfers any at all is done by purely causal intuitions.
After all, in the quest for tipping principles, one has practically reached one’s destin-
ation simply by invoking the role of causation: as between two innocents, let the causer
pay! The finding of bad character conveniently masks this unspoken but powerful
tipping principle so as to relieve defenders of the significant theoretical problems
posed by any effort to base liability solely upon (in-fact and proximate) causation.
Those who want to resist this charge have to take seriously that a character-based
account of negligence does poor work to explain why we transfer all of an innocent
plaintiff ’s loss to a defendant when we transfer any of it.
Second, what should we make of out-of-character negligence? Presumably, if the
only fault that can be assigned to a defendant’s inadvertent causation of injury derives
from the fault that we can ascribe to his character, then when inadvertent injuries are
out of character, they cannot be thought to be blameworthy. And is it not the case that
such injuries occur with a fair bit of frequency? Is it not the case that careful people do
careless things, and that good people sometimes exhibit “admirable immorality,” as
Michael Slote has termed it, by pursuing a virtue in a way that proves unfortunate.32
When such out-of-character moments result in accidental injuries to others, is it really
the case that tort law does not, or should not, impose liability on the surprised injurer?
Inasmuch as most people are quick to demand recompense from others for injuries
that result from inadvertence without regard to whether that inadvertence was motiv-
ationally aberrational, it would seem that negligence law would be out of step with
common intuitions were it to restrict its understanding of fault and its imposition of
liability to in-character negligence.
Those seeking to preserve a fault-based theory of negligence may conclude, how-
ever, that a prescriptive theory is better than no theory at all. They thus might argue

32
Michael Slote, Goods and Virtues, (Oxford: Clarendon Press, 1983), 78–80.
 FINDING NO FAULT WITH NEGLIGENCE 403

that the law should be more discriminating in its pronouncements of negligence,

because it should inquire into whether a defendant’s inadvertence was a function of
poor character, as opposed to a momentary lapse that prompted a good person to do a
bad act. Tort law should ask in a case like The T.J. Hooper,33 whether the defendant
tugboat owner’s failure to equip his carrier with a radio receiving set so that its captain
would be able to receive naval broadcast storm warnings was a result of a miserly
desire to protect his profits, or an uncharacteristic failure keep up with the technology
of the times.
But not only will such a claim invite defeating moral problems, it will also invite
further legal concerns. The least interesting, but probably most damning, of these is
that any such endeavor would invite extraordinary evidentiary difficulties. How in the
world could judges and jurors adjudicate whether a given moment of inadvertence was
a causal product of miserliness or melancholy? Whether it originated in a person’s
tendency towards slothfulness or was caused by confusion, distraction, or poor
memory? And how could adjudicators compare the nature of the bad character that
motivated the defendant’s inadvertence with the nature of the bad character that
motivated the plaintiff ’s inadvertence so as to apportion damages in proportion to
the relative blameworthiness of the traits that lead them both into harm? Are greed,
jealousy, self-absorption, impetuousness, hastiness, indifference, arrogance, narcis-
sism, foolhardiness, cowardice, and gluttony susceptible to the sort of cardinal or
ordinal ranking that would permit principled comparisons of the vices of parties in
cases of comparative negligence?
However confounding the legal obstacles to piggybacking the culpability of negli-
gence onto the culpability of character, I take the moral obstacles to be even more
serious. For in the end, I think it is clear that while character flaws may be blame-
worthy, the fact that they may cause injurious inadvertence does not make that
inadvertence blameworthy. It is one thing to find character blameworthy, and quite
another to conclude that the blame that we apply to character is transitive so as to
infect whatever deeds may owe their causal genesis to blameworthy character traits.
But is it not intuitive that if we can blame a person for being selfish or impetuous or
cowardly or gluttonous, we can then blame him for whatever carelessness towards
others these vices inspire? What entitles me to think that the blame we attach to
unworthy dispositions is not transitive? The answer requires us to examine the nature
of the blame that we direct at those who possess vices. Such blame and praise is aretaic,
not deontic. It is of a sort unique to character judgments; it does not cross over in
application to judgments concerning actions and the culpability with which those
actions are pursued.
While aretaic duties to cultivate virtuous character traits may de facto require
supererogatory (or “heroic”) actions, they do not (typically) require one to super-
erogate on any given occasion towards any given person. For one may, for example, be

33
The T.J. Hooper 60 F 2d 737 (US Ct of Apps (2nd Cir), 1932).
404 HEIDI M . HURD

a generous person without at any specific moment in time doing a generous deed. And
aretaic duties to suppress vicious character traits may de facto require that one not
suberogate (that one not abuse one’s rights), but they do not require that one refuse to
exercise a right at every turn or on any specific occasion, for one may not be an
intemperate, selfish, or gluttonous person just because one occasionally over-indulges.
Put differently, aretaic obligations compel the cultivation of virtuous character traits or
dispositions which one may possess despite doing specific actions that do not realize
those traits; and aretaic prohibitions compel the suppression of vicious character traits,
but may be satisfied even as one periodically does actions that would be vicious if done
too often.
Crucially, then, the “oughts” that are embodied in our praise of heroic deeds and our
condemnation of abuses of rights derive from aretaic duties, and are thus “oughts” of
obligation. But inasmuch as these “oughts” do not compel or prohibit particular
actions in particular circumstances, they are not deontically obligatory. They are
thus non-obligatory in the sense in which we typically use the notion of obligation,
for that notion is typically reserved for references to deontic obligations. When one
supererogates one expresses a virtue, but the praise that such an action invites is
aretaic, not deontic, for no one has a right to one’s heroism. And when one suberogates
one expresses a vice, but the blame that such an action invites is aretaic, not deontic,
for no one has a right that one forgo the exercise of one’s right. Such suberogation thus
reflects the failure to fulfill an aretaic ought; not the failure to fulfill a deontic ought.
Now, perhaps surprisingly, we are philosophically home. When an actor’s poor
character exhibits itself in a moment of inadvertence that ultimately results in injury to
another, the deontic rights of that injured party have indeed been violated. A deontic
wrong has been done. But the culpability with which that rights violation has occurred
is of an aretaic sort, not a deontic sort. In the defendant’s mind was not a depiction of
the harmful act, complete with its wrong-making characteristics. Rather, in his mind
(perhaps not even fully consciously) were dispositional desires, emotions, or beliefs
that inclined him towards his action and that blinded him, or distracted him, or
otherwise diverted him from carefully contemplating the implications of that action
for others. Such dispositions are the stuff of aretaic culpability. But we cannot marry
such aretaic culpability with the deontic wrong that was done to the plaintiff in order
to declare the defendant a culpable wrongdoer, for by so doing we mix moral apples
and oranges. The defendant did a wrong—a deontic wrong. He did it culpably—but
only aretaically so. He did not possess a mental state that can be described as
blameworthy in the sense of that word that is relevant to law—a sense standardly
reserved for mental states that take as their objects the deontic wrongs that that ground
tort liability and comprise the actus reus conditions of crimes.
To say otherwise would be akin to violating the simultaneity principle in Decina-like
cases. One would be trying to trace the deontic wrong that was done to the plaintiff to a
prior voluntary act by the defendant that was, itself, culpable. But the act that in-fact
and proximately caused the plaintiff ’s injury was not coupled with anything other
than bad character. While bad character might itself be aretaically blameworthy, the
 FINDING NO FAULT WITH NEGLIGENCE 405

prohibitions against it are themselves only aretaic: They speak not at all to actions, and
thus not at all to the particular act that inadvertently brought about the accident. There
is thus no match between the wrong done to the plaintiff and the kind of culpability
that would be required to make the defendant blameworthy for that wrong.
There is, then, a significant gap between finding that character can be blameworthy—
by which we must mean aretaically blameworthy—and concluding that the inadver-
tence caused by such bad character is deontically blameworthy. That a defendant’s bad
character can cause inadvertence that then causes a deontic wrong does not entitle us
to infer that the defendant possessed the kind of culpability required to make such
negligence a matter of fault. For that his aretaic failing caused a deontic failing does not
convert his aretaic culpability into deontic culpability. His negligence remained
blameless in the sense of that term that is morally relevant to tort law.
So even if tort theorists were willing to press upon accident law a characterological
understanding of negligence, their theory would fail to match the wrongs done in cases
of accidental injury to the requisite mental states that would allow us to blame
defendants for those wrongs. We might, in such cases, be transferring liability to
persons of bad character; but we would not be transferring liability to persons who
were culpable for the wrongs they caused.
We are thus without a basis for thinking that tort law’s conception of negligence is,
or can be made, credibly fault-based. We must admit that despite decades of verbiage
devoted to superficial moralizing about the superiority of a fault-based standard to a
standard of strict liability, tort law is and has been attaching liability to persons
without a coherent account of why their inadvertent injuries are credibly things for
which they are at fault.
Of course, in many cases, inadvertence is blameworthy because it can be traced to
blameworthy choices—choices to violate known mini-maxims the spirit of which is to
protect against the risks that the actor’s ensuing conduct inadvertently created; choices
to engage in conduct that bore the known risk of rendering the actor inadvertent to
later risks; choices to avoid precautions that would have insulated the actor from later
inadvertence caused by motor skill failures, cognitive impediments, volitional short-
comings, or characterological failings; and so forth. But when inadvertence owes its
pedigree to culpable choices, we need no theory of negligence, for we can legitimately
rest legal liability on the blameworthiness of the choices that lead to the actor’s
inadvertence. Were tort law only to attach liability for injuries that could be traced
to recklessness on the part of defendants, there could be little doubt that tort law could
defend its regime of accident law as genuinely fault-based. Yet tort law’s net of liability
is thrown as readily over those whose inadvertence is morally innocent as over those
whose inadvertence is a product of culpable choice. If tort law is going to persist in
assigning liability to those who are blameless, then it ought to confess as much. Instead
of pretending that there is a credible conception of fault upon which negligence law
rests, apologists for the present legal regime ought to get on with the business of
vindicating strict liability, for in many cases, negligence liability amounts to as much.
 19
Confused Culpability, Contrived
Causation, and the Collapse
of Tort Theory
Larry Alexander and Kimberly Kessler Ferzan

What justifies tort law? Once we identify a domain that is central to, if not co-extensive
with, “torts,” we will find that it consists of a motley collection of doctrines that are
impossible to justify under any recognizable and attractive normative principles.

I. The Target Field

If one looks inside a standard torts casebook, one finds a vast and quite variegated
collection of cases. Nevertheless, all torts casebooks do contain two groups of cases
with the following four characteristics. (1) In both groups, the plaintiff has claimed
that his person has been (or will be) injured, his property damaged, or his livelihood
adversely affected by defendant’s conduct, conduct that does not involve a breach of
contract or of trust. (2) In one group, the defendant has claimed that he did not mean
to injure, damage, or adversely affect plaintiff ’s interests, nor was he aware of an
unjustifiable risk that his conduct would do so. (3) In the other group, defendant was
either aware of an unjustifiable risk that his conduct would adversely affect plaintiff ’s
interests or acted with the purpose of adversely affecting those interests. (4) In both
groups, the plaintiff is seeking monetary compensation for that injury or damage.
Can cases in those two groups be explained and justified by attractive normative
principles? To answer, one ought to take into account other mechanisms for dealing
with loss and culpability. Tort law functions in a rather ad hoc manner, only repairing
losses caused by other humans and not losses caused by Mother Nature, and imposing
liability on those who cause harm and not those who merely risk it.1 We think that an

1
See Ronen Avraham and Issa Kohler-Hausmann, “Accident Law for Egalitarians,” 12 Legal Theory 181
(2006).
 THE COLLAPSE OF TORT THEORY 407

insurance scheme, coupled with administrative regulations and the criminal law, may
be a better mechanism.

II. Is There a Duty in the House?

We agree with Gregory C. Keating that tort law cannot be explained and justified
without locating some duty that defendant has breached that accounts for why he and
not others—or the plaintiff himself—must compensate plaintiff for plaintiff ’s losses in
those cases in which the plaintiff prevails. What is it about defendant that justifies
plaintiff collecting from him? Providing recourse to an injured plaintiff does not
explain and justify giving plaintiff recourse against defendant. Nor does invoking
“corrective justice” tell us why it is “just” in any sense of just for defendant to be
held liable.2
But what would account for plaintiff ’s claim against defendant for compensation?
For Keating and for us, the most obvious candidate would be that defendant breached
a duty toward plaintiff that resulted in plaintiff ’s loss.
Coming up with a duty that defendants have breached in these cases is no easy task,
however. Remember, in one group of cases the defendant acted without intending to
harm plaintiff ’s interests and without adverting to an unjustifiable risk to those
interests that his conduct was creating. Let us look at such cases to see if we can locate
a breach of duty.

A. A duty not to cause harm?

If we had a duty never to cause harm to others’ interests, we would have made some
progress in locating the duty we need. But such a duty can neither explain nor justify
the case results.
To begin with, if there were such a duty, we would be hard pressed to explain all
those cases in which the defendant was held to have caused plaintiff ’s losses but in
which defendant prevailed. There are many such cases and countless that are never
litigated because the plaintiff knows he will not prevail.
It seems doubtful those cases are all incorrect. One reason it should be doubtful is
the Coasean point that plaintiff ’s injuries have also been caused by his conduct.3
Indeed, if plaintiff recovers, defendant’s resulting loss has been caused by plaintiff ’s
acts of, say, using defendant’s product, or driving on the same road as defendant, plus
his act of suing defendant.

2
On these points, see Gregory C. Keating, “The Priority of Respect Over Repair,” 18 Legal Theory 293
(2012). The law and economics, cheapest cost-avoider, approach is not duty-based but rather is an ex ante
insurance, not an ex post damage, approach.
3
Ronald H. Coase, “The Problem of Social Cost,” 3 J. Law and Econ. 1 (1960), 2.
408 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

There are other reasons aside from the Coasean such “strict liability” cannot be
justified. Richard Epstein’s argument in favor of strict liability on the ground that it
followed logically from plaintiff ’s entitlement to his life, bodily integrity, and property
was question-begging.4 For the question can be reposed: why do plaintiff ’s entitle-
ments extend to entitlements to be compensated for accidental damage, and why is
defendant not entitled to secure his property against accidental damage claims by
plaintiffs? There is no good answer to these questions.

B. A duty not to cause harm through “faulty” conduct?

In a large number of cases in the torts casebooks, plaintiff recovers his losses from
defendant because the trier finds that defendant caused plaintiff ’s losses through faulty
conduct. For many, the fact that defendant’s conduct was faulty justifies making
defendant, and not others, pay for plaintiff ’s losses.
But what kind of fault is exhibited by defendant’s conduct in these cases? It cannot
be this: (1) that had defendant not acted exactly as he did, plaintiff would not have
suffered the losses he suffered; and (2) that the social benefits (however calculated) of
defendant’s acting exactly as he did were not sufficient to justify plaintiff ’s losses. Even
leaving aside the problematic counterfactual in (1)—a point we shall return to later in
this chapter—(1) and (2) are present even in those cases in which plaintiff lost because
the court did not find defendant to have been at fault. In other words, leaving aside
cases in which defendant’s act was a justifiable one on the condition that he pay for any
damage it causes—cases that can be considered a species of private eminent domain—
in the bulk of cases meeting conditions (1) and (2), the plaintiff will lose.
What condition, in addition to (1) and (2), is required for “fault” to exist? Remember,
we are not at this point dealing with a defendant who is conscious of an unjustifiable
risk to plaintiff ’s person or property, or who intends to injure plaintiff ’s person or
property. We are dealing with defendants who have not adverted to those conditions
that were present when they acted and the presence of which in conjunction with
defendant’s act brought about plaintiff ’s injuries or property damages.

1. What is the fault in not adverting?

The orthodox doctrinal response is that defendant’s “fault” in these cases was his
failure to alter his conduct because he was not adverting to those harm-causing
conditions that were present when he acted as he did. But what kind of fault is not
adverting?
First, it is worth noting that in some pockets of tort law, even the most ardent
proponents of the reasonable person test for fault must admit that their test amounts
to a form of strict liability. This is because when the defendant is held to an objective
standard that he utterly lacks the capacity to meet, he is not demonstrating any fault on

4
Richard A. Epstein, “A Theory of Strict Liability,” 2 J. Legal Stud. 151 (1973).
 THE COLLAPSE OF TORT THEORY 409

his part.5 When being the best that you can be is simply not good enough for tort law,
there’s a name for that—strict liability.6
Second, we see the problem as going beyond these handful of examples. Consider
two cases:
Hard Worker
Kimberly has an important job and has been asked to write a legal memorandum by the
end of the week. She works through lunch and leaves the office to drive home at 7:30 p.m.
Because she is still deep in thought, because she has been looking at a computer screen all
day, and because her blood sugar is low from not eating, she mistakenly interprets the light
as a green light for her. She hits the accelerator and collides with another car.
Excellent Driver
Mike reads a scientific article that discusses the strength of peripheral vision and
becomes convinced that reading the newspaper while driving is not problematic. Mike
reads the paper while driving and rear-ends a car that he failed to see had braked.

Kimberly and Mike each get the Learned Hand Carroll Towing7 formula wrong.
Kimberly fails to run the formula at all because she never sees the risk. Mike, we
shall assume, is falsely confident in his abilities and therefore discounts the probability
of harm to a point at which his running the risk is justified under the Hand formula.
Indeed, this is why the Hand formula is not a test for inadvertent negligence. Rather,
that test posits a defendant who weighs the costs and benefits of an act, finds that the
costs outweigh the benefits, and then despite this chooses to engage in that act. Such a
defendant would indeed be culpable, but he would be culpable because he is con-
sciously disregarding an unjustifiable risk. He is reckless, not inadvertently negligent.
(The inadvertently negligent actor is not adverting to the Hand formula’s costs and
benefits; and the costs and benefits of uncovering the true costs and benefits do not
appear to him to be Hand-formula cost-justified.)8
Tort law would find that both of these individuals are “at fault,” but wherein does
the fault lie? In our book on criminal law theory, we devoted a chapter to establishing
the proposition that merely not adverting to dangerous aspects of one’s conduct is not
culpable.9 In the criminal law context, theorists have struggled in their attempts to
drive a wedge between Kimberly, whom many theorists think should not be held
criminally responsible, and Mike, whom many think should be responsible, allegedly
because he manifests an indifference to others that Kimberly does not.10 We reject that
distinction, but no matter; for tort law makes Kimberly pay, too. But why?

5
See, e.g., Vaughan v Menlove [1837] 132 ER 490.
6
See also Avihay Dorfman, “Reasonable Care: Equality as Objectivity,” 41 Law & Phil. 369 (2012).
7
United States v Carroll Towing Co. 159 F 2d 169 (US Ct of Apps (2d Cir), 1947).
8
A more complete discussion of this problem of non-actuarial risks is found in Larry A. Alexander,
“Causation and Corrective Justice: Does Tort Law Make Sense?,” 6 Law & Phil. 1 (1987), 18–23.
9
See Larry Alexander and Kimberly Kessler Ferzan, Crime and Culpability: A Theory of Criminal Law
(Cambridge: Cambridge University Press, 2009), Chapter 3.
10
See, e.g., Peter Westen, “Individualizing the Reasonable Person in Criminal Law,” 2 Crim. L. & Phil. 137
(2008).
410 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

As we noted in our book, “an actor may fail to form a belief (or a correct belief) if he
(1) lacks the requisite background beliefs, (2) lacks the intellectual ability, or (3) lacks
the motivation to form the belief.”11 Michael Moore and Heidi Hurd have identified
those capacity defects as motor control defects, cognitive defects, conational defects,
and motivational deficiencies.12 As they clearly elucidate, anytime we wish to say that
someone could have adverted, what we are saying is that the actor would have adverted
if some condition C. The moral work, then, is done by articulating why the lack of C is
blameworthy. (We assume that tort law does not mean “fault” as simply in some way
faulty or deficient, but means to carry some suggestion of moral blame, and thus must
show why the failing in question is blameworthy.)
These failings themselves, however, may not be “faulty.” Even if it is true that
Kimberly could have adverted to the red light had some condition not been present
that was present, wherein lies the fault in her perceptual failure? Even if it is true that
Mike could have correctly appraised the chance of an accident had he had a lower
opinion of his abilities, wherein lies the fault in having a high opinion of his driving
skill and perceptual abilities? Unless tort law aims to be an ad hoc redistributive system
whereby when one’s perceptual or intellectual mechanisms falter and cause harm,
one’s goods are simply redistributed to another human being, tort law must explain
wherein the fault in having the faulty mechanism lies.
Moreover, when there is fault, it is because we can point to a prior point in time
when the actor made a deliberate choice. That is, we can trace the present defect to an
earlier act of recklessness—a conscious decision to take an unjustified risk. But tort law
never directly focuses on these upstream measures, nor does it ever assess whether
those upstream measures would have been reasonable to take. What would Kimberly
have had to forgo (1) to write a note to remember to eat, and (2) to take that lunch
break? She clearly would have done less work, and important work at that. What
should Mike have done to assess the validity of the scientific study he read or to
evaluate further his perceptual abilities? And why would Kimberly and Mike believe
they needed to do these things? After all, if you believe the risk created by your act is X,
and X is low enough to make your act justifiable, what reason do you have to desist and
investigate further? Desisting and investigating are costly in time and resources. It is
true, of course, that you may discover that the risk of your act is Y, which is higher than
X and sufficiently high to render your act unjustifiable. But it is also true that you may
discover the act’s risk is Z, which is lower than X, rendering the act even more
justifiable than you thought. And there is no probability that you can attribute to
discovering that either Y or Z is the case. If you believe that the risk presented by your
act is X, then you must predict that upon further investigation, the risk will still turn
out to be X, and you will have wasted time and resources on the investigation. If the act

11
Alexander and Ferzan, Crime and Culpability (note 9) at 83.
12
Michael S. Moore and Heidi M. Hurd, “Punishing the Awkward, the Stupid, the Weak, and the Selfish:
The Culpability of Negligence,” 5 Crim. L. & Phil. 147 (2011).
 THE COLLAPSE OF TORT THEORY 411

turns out to cause damage that could have been averted at a justifiable cost—if its
actual risk of damage was 1, thus higher than X (and Y)—then it is true in one sense
that defendant’s act, which he thought justifiable, was not justifiable, and he should not
have taken that act. That is the “should” of 20–20 hindsight. And that is really the only
“should” that makes sense in claiming that defendant should have adverted to the
higher risk. But the same “should” is available in the strict liability cases and is not a
“should” that suggests any culpability-reflecting “fault.”
Merely claiming that a “reasonable” and implicitly nonculpable person would have
adverted is not itself an argument that confronts the points that Moore, Hurd, and we
have made, points that can be encapsulated in the observation that we cannot, at the
time of action, directly control to what we are adverting. (We would have to be
adverting to that to which we are not adverting in order to do so.)
Are we inconsistent in deeming actors culpable for their misperceptions of reasons
but not for their misperceptions of risks? If the latter misperceptions are outside of the
actor’s control, are not the former misperceptions as well?
There are deep waters here that we can only dip a toe into in this chapter. A quick
response, which will have to suffice, is that giving the interests of others too little
weight—insufficient concern—is the basis of culpability. Why the culpable actor gives
others’ interests insufficient concern would move the discussion into the heart of the
free will/determinism/moral responsibility debates as well as into the debate over
whether psychopaths, who give others’ interests no weight, can be deemed morally
responsible and culpable. We admit that these issues must be resolved to have a
complete account of culpable acts. We admit also that there are difficult issues raised
when actors give others’ interests insufficient concern because they subscribe to moral
norms—norms which activate in them the reactive emotions of blame, indignation,
and guilt—that require insufficient concern for some others. (Honor killings may be an
example.) Here, we assert only that non-psychopaths who, in acting, give others’
interests insufficient weight are culpable and distinguishable from those who misassess
risks.
Finally, some will agree with us that inadvertent negligence is not culpable. None-
theless, they will respond that we are wrong to assume that “fault” in tort law is
equivalent to culpability, the concern of criminal law. However, because tort theorists
distinguish between fault-based liability and strict liability, for this criticism to have
any traction, there must be a notion of fault that is neither culpability nor strict
liability. In the following section, we deny that possibility.

2. Why negligence is indistinguishable from strict liability

Not only is inadvertent negligence not culpable—or so we contend—but inadvertent
negligence is indistinguishable from strict liability. We have elsewhere argued that the
distinction between negligence and strict liability cannot be maintained because there
is no non-arbitrary way to construct the notion of the “reasonable person” who would
412 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

have adverted to the risk created by his proposed conduct.13 Here, we want to argue
against the distinction between negligence and strict liability on different, but ultim-
ately related, grounds. We want to argue that there is no non-arbitrary way to assess
the riskiness of conduct ex post that will produce a negligence-strict liability distinction
among those defendants who failed to advert to the risk.14 There is no non-arbitrary
way because, ex post, the risk of harm created by conduct is either one or zero, and
neither risk will produce the needed distinction.
To illustrate our point, we offer some cases:
(1) Nellie, looking for some spices in a cabinet, takes out some rat poison and places
it on a kitchen counter. She is so preoccupied with her cooking that she forgets
she has left the poison there. Later, her three-year-old brother Oscar comes into
the kitchen, thinks the poison is sugar, and eats it, becoming very ill.
(2) Same case as (1), except Oscar slips while he is reaching for the rat poison, hurts
himself, and begins crying. Nellie rushes into the kitchen, where she sees that
Oscar had almost eaten the poison. With a great sense of relief, she puts the
poison back in the cabinet.
(3) The Wrinkled Prune Company (“Wrinkled”) sells pitted prunes. Its prunes are
pitted by a state-of-the-art technological process. There is no cost-justified
technology or process design that will guarantee that no pits will remain.
Wrinkled puts a notice on every box of its prunes that warns consumers that
some pits may remain. Paula, a six-year-old, loves to eat prunes. She bites down
on one of Wrinkled’s prunes and breaks a tooth on a pit that had not been
removed. A worker at Wrinkled could have inspected this particular box of
prunes for less cost than Paula’s dentist bill.
(4) This case is the same as (3), only Paula discovers the pit before biting down and
avoids injury.
Traditionally, these four cases would be analyzed along the following lines. In case (1),
Nellie was negligent, and her negligence resulted in injury to Oscar. In case (2), she was
again negligent, but her negligence led to no harm. In case (3), in many jurisdictions,
Wrinkled would be strictly liable to Paula because its product, pitted prunes, was
defective and caused Paula’s injury. Wrinkled is not negligent, however, because its
process for pitting was cost-justified, even given the risks of an occasional incompletely
pitted prune. In case (4), there is neither negligence nor strict liability.
Now what were the risks of harm in these cases? Because we are operating ex post,
the answers are easy. In cases (1) and (3), the risk of harm was one. In cases (2) and (4),

13
See Alexander and Ferzan, Crime and Culpability (note 9), Chapter 3. See also Larry Alexander,
“Reconsidering the Relationship Among Voluntary Acts, Strict Liability, and Negligence in Criminal Law,”
7 Soc. Phil. & Pol’y 84 (1990).
14
This argument draws from Larry Alexander, “Foreword: Coleman and Corrective Justice,” 15 Harv.
J.L. & Pub. Pol’y 621 (1992), 631–6.
 THE COLLAPSE OF TORT THEORY 413

it was zero. The actual risks fail to distinguish negligence and strict liability. They
distinguish only the cases of harm from the cases of no harm.
What if we add the condition that the risks of harm must be unjustifiable if conduct
is to be deemed negligent? That still does not produce the outcomes of the traditional
analysis. In case (3), given that the actual risk of harm to Paula was one, Wrinkled was
“unjustified” in not having a worker inspect the box bound for Paula’s house. Of
course, it did not know that particular box contained an incompletely pitted prune.
But neither did Nellie know that she had left the rat poison out, at least in the sense
that she was not adverting to that fact. Moreover, in case (2), the risk is zero, which can
hardly be deemed an unjustifiable risk.
To make the traditional analysis work, we must delete the information about injury
that we possess ex post in such a way as to produce risks in cases (1) and (2) that are
between one and zero, equal to each other, and higher than the risks in cases (3) and
(4) (which are also between one and zero and equal to each other). The problem is how
to select which information about the actual cases to delete.
This is not a problem in cases of recklessness. There, we let the actor’s subjective
estimate of the risk ex ante determine our characterization of his conduct. If he
estimates the risk as high, so high that we would deem taking such a risk unjustifiable,
then he is reckless whether the risk was one or zero (that is, whether the harm risked
came to pass).
In cases (1)–(4), however, we cannot avail ourselves of the actor’s ex ante subjective
estimate of risk if we want to produce the traditional outcomes. Nellie undoubtedly
believes her activity, which she would describe as “cooking,” is only minimally risky,
because she is unaware that it includes her leaving rat poison in a place accessible to
Oscar. Wrinkled has an ex ante estimate of the risks of injury to Paula which equals the
risks to any consumer of its pitted prunes chosen at random. By hypothesis, that risk is
low and reasonable to impose. The risk is ascertained by abstracting from the details of
all reported cases of incompletely pitted prunes information that can serve as the basis
for efficient actuarial categories. Thus, the percentage of prunes that are incompletely
pitted and the total damage caused by incompletely pitted prunes will probably be the
only information that is gathered, the more specific details being too expensive to
gather or incapable of affecting Wrinkled’s conduct even if gathered. (An example of
the latter information would be information that consumers are more likely to bite
down on a pit at breakfast, when they are less alert, than at lunch; here, unless the
differential risk were sufficient to warrant a special “breakfast warning,” Wrinkled
must lump the risks together.)
The cases of Nellie and Wrinkled resist being distinguished upon close analysis
despite the traditional view that they are different. In all four cases, the actors have in a
sense misgauged the risks. In cases (1) and (3), Nellie and Wrinkled have underesti-
mated the risk—Nellie, because she views her act as “cooking” rather than as “cooking
while leaving rat poison where Oscar will eat it”; Wrinkled, because it is not concerned
with the risks in any individual case but only with the average risks in the aggregate. In
both cases the actual risk of one could have been reduced to zero by cost-justified
414 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

actions, but neither actor was aware of the factors that would justify those actions.
Nellie was unaware that she had forgotten to put the rat poison back in the cabinet.
Wrinkled was not aware that the box bound for Paula contained an incompletely
pitted prune. So although Nellie and Wrinkled believed that their conduct was risky,
but justifiably so under the description “cooking” and “selling prunes pitted by a
particular process,” both were unaware that their conduct caused a risk of one under
the descriptions “cooking (or selling prunes) under the circumstances that actually
exist in these cases.” Likewise, in cases (2) and (4), where no harm occurs, Nellie and
Wrinkled have overestimated the risk in the particular circumstances.
To summarize the analysis thus far, the risk in every case is actually either one or
zero. In each case there are general features on which the actor will focus to predict a
risk of harm that lies somewhere between one and zero. The actor who predicts a risk
that is unjustifiably high and then proceeds to act is reckless even if the harm risked
does not eventuate (the actual risk is zero). The actor who predicts a low risk—a risk
that it is justifiable to take—is not reckless if he acts. If his prediction is actuarially
sound, yet harm eventuates (the actual risk is one), then any liability he faces is
paradigmatic strict liability (for example, our case (3)). This is so even though in the
particular case, acting was not cost-justified. (Wrinkled was not cost-justified in selling
the particular box of prunes to Paula in case (3), or was not cost-justified in failing to
inspect that particular box.)
What if the actor incorrectly underestimates the actuarial risk?15 This may occur
when an act has a general feature or features that lend themselves to determining the
actuarial risk, but the actor is either unaware of those features or is aware of the
features but not of the actuarial risk associated with them. These are the cases of
inadvertent negligence, such as Nellie’s leaving the rat poison within Oscar’s reach
(cases (1) and (2)). In these cases we must ask why the actor is ignorant of the
important features or the risks associated with them. And in doing so we repeat the
preceding analysis, only this time we replace risk of harm with risk of ignorance (of
features or risks).16 The risk of ignorance may be either reckless, or it may be cost-
justified. (For example, Nellie may be employing an efficient level of advertence that in
cases (1) and (2) just happened to result in inadvertence to risk.) Or the actor could be
ignorant of the risk of ignorance, in which case the analysis repeats again. Ultimately,
we end up with either a reckless actor or a strictly liable one. Inadvertent negligence
breaks down into one of those two categories; it is not a third category.
True risks—one or zero—do not distinguish within the class of human-caused
accidents, because the risk is always one in those cases. The focus must therefore be
on ex ante subjective estimates of the risk. The subjective estimates may show the
action to be ex ante cost-justified (strict liability) because the action is a particular

15
Here we continue a discussion begun in Alexander, “Causation and Corrective Justice: Does Tort Law
Make Sense?” (note 8) at 18–23.
16
See also Mark Grady, “Why Are People Negligent? Technology, Nondurable Precautions, and the
Medical Malpractice Explosion,” 82 Nw. U. L. Rev. 293 (1988).
 THE COLLAPSE OF TORT THEORY 415

instance of a general class whose actuarial risk of harm is low relative to its benefits and
the costs of refining the actuarial estimates. The subjective estimates may show the
action to be reckless. Or the subjective estimates may, were they correct, show
the action to be ex ante cost-justified, but they may be incorrect subjective estimates.
In the latter case our focus must switch to the subjective estimate of risks associated
with features of acts, the subjective estimates regarding the degree of confidence
warranted in the subjective estimates of risks associated with features of acts, the
subjective estimates of the benefits of using different actuarial categories, or the subject-
ive estimates of the likelihood of not adverting to a feature whose associated risk is
known. The subjective estimates of those subjective estimates may either reveal
reasonable behavior, and hence strict liability, or reveal recklessness.
In short, every case of inadvertent negligence—Nellie in cases (1) and (2)—is at
bottom either a case of recklessness or a case of strict liability. And every case of strict
liability—Wrinkled in case (3)—can be viewed as a case of inadvertent negligence
if we ignore subjective estimates of risk and focus on the particulars that make the
true risk one.
To the extent that tort law is concerned with deterring culpable behavior and
requiring culpable actors to pay for the damage caused by their acts, both inadvertent
negligence and strict liability are outside its purview. To the extent that tort law is
concerned with placing the costs of interactions on those—injurers or injured, third-
party insurers or first-party insurers—who possess actuarial information about risks,
providing (cost-justified) incentives to obtain actuarial information, or effecting dis-
tributional patterns and reducing transaction costs (including litigation costs), cases of
both inadvertent negligence and strict liability are within its purview, although the
distinction between them is wholly chimerical and irrelevant. This is why, from an
economist’s view, when strategic concerns relative to the costs and likelihood of
proving negligence in court are put to one side, the choice between a negligence rule
and a strict liability rule is inconsequential: They would lead to exactly the same
conduct.
Ultimately, everything comes down to what to do about non-actuarial risks. When
we lack information we cannot act on the information we lack, nor can we assess the
value of obtaining that information. (We must have the information in order to assess
the value of obtaining it.) Ignorance cannot be assigned an actuarial value. Except for
reckless, knowing, or intentional harm causing, all harm-causing results from ignor-
ance of true risk. An actor who acts in the face of his own estimate of unjustifiably high
risks has acted unjustifiably and can be dealt with in the same manner as are knowing
and intentional injurers. But an actor whose ignorance leads him to assess a risk as
sufficiently low to make taking it justifiable is an actor whose liability, if any, is strict.
And when the strict liability attaches to ignorance that, at the level it occurs, cannot be
actuarialized, the case for liability and internalizing costs rather than socializing them
cannot rest on attaching the proper incentives to rational calculation.
We conclude that, limiting ourselves to cases of inadvertent negligence and strict
liability, whatever “fault” is present in the former is also present in the latter, and that
416 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

form of fault is not a form of culpability. Any alleged duty to avoid faulty conduct in
this sense really boils down to a duty not to cause harm, and we have already argued
that such a duty cannot be justified and, moreover, cannot explain torts cases.

C. A duty not to cause harm through culpable conduct

There is surely a general duty not to engage in culpable conduct. For us, culpable
conduct, rightly conceived, is conduct that the actor believes imposes risks to others’
legally protected interests that are beyond the actor’s ability to reduce once he engages
in the conduct in question, and that are not justified by the reasons that favor the
conduct that the actor believes exist.17 Therefore, on our conception of culpable
conduct, the actor must be adverting to the relevant risks and reasons that determine
his culpability. The question of whether tort law is an appropriate means for dealing
with culpably caused harms then leads to our second worry—causation.

III. Causation
For criminal culpability, what happens once the actor engages in culpable conduct—
whether his conduct does or does not cause harm to others’ legally protected
interests—should be immaterial, or so we have argued elsewhere.18 In the realm of
tort law, however, if culpable conduct does not produce harm to the plaintiff ’s legally
protected interests, then no tort has occurred. So, for purposes of tort law, when we
focus on truly culpable conduct, we focus solely on culpable conduct by the defendant
that has resulted in damage to legally protected interests of the plaintiff.
The “resulted in” notion is typically cashed out in terms of causation. Defendant’s
culpable conduct must have “caused” plaintiff ’s injuries. But this causation relation,
without which there would be no recognizable tort law, has proven to be quite
troublesome. If it cannot be adequately theorized in the context of truly culpable
conduct, then tort law in that context cannot be fully theoretically justified. And if it
cannot be theoretically justified when defendant has acted culpably, then it surely
cannot be theoretically justified when defendant’s conduct has displayed only a
nonculpable “fault.”

A. Problem one: The mismatch between degree of culpability and

degree of harm, or does corrective justice trump retributive justice?
Suppose Ed commits a minimally culpable act, one for which he would deserve only a
smallish penalty—say, a $50 fine. Unfortunately, Ed’s minimally culpable act results in

17
Or might exist, discounted by their improbability as the actor perceives it.
18
See Alexander and Ferzan, Crime and Culpability (note 9), Chapter 5.
 THE COLLAPSE OF TORT THEORY 417

Edith’s suffering catastrophic losses. Her two children—prodigies with heretofore

fabulous earning capacities—are severely crippled. Her Westminster dog show first-
prize winning Papillon is killed. Her fourteen Picassos are destroyed. Compensating
Edith for her losses will bankrupt Ed, who has been quite prosperous due to having
provided many socially valuable products and services.
Should Ed have to compensate Edith for her losses beyond $50 and his fair share of
the balance? (We leave unstated what that fair share might be.) Tort law says “yes.” If
Ed had intentionally lightly kicked Edith’s shin, a minimally culpable act that Ed
expected to cause a slight momentary pain but no injury, but Edith had the proverbial
“eggshell skull” (or shin) and suffered massive injury or even death, tort law has it that
Ed would be on the hook for all the damage that Edith (or her heirs) suffered.
Is it just that Ed should be impoverished by Edith’s tort suit? Why should not first-
party insurance take over after Ed has paid Edith $50? It would surely be prudent for
Edith, if she has an eggshell skull, or if she is transporting prodigies, prize-winning
Papillons, and Picassos, to have first-party insurance. After all, if she doesn’t, then she
is extremely fortunate if Ed is capable of giving her full compensation.
Or if first-party insurance seems like an unfair burden to place on Edith, how about
social insurance to cover the balance? Once Ed pays the $50, he has suffered to the
extent he deserves for his culpable act, and he is effectively as innocent as Edith and
everyone else.
It is possible perhaps that Ed could purchase liability insurance that covers his
culpable conduct so that he can socialize the risk just as easily as can Edith or the rest
of society. So we might get at the basic issue more easily if we completely eliminate
insurance from the picture. Suppose Ed and Edith are the only two people who exist.
The costs of their activities cannot be spread to others. If Ed’s slightly culpable conduct
results in catastrophic losses for Edith, should Ed have to pay full compensation, or
should Ed and Edith have to split the damages after Ed pays $50? Tort law says Ed
must pay full compensation. Perhaps the logic is that if you engage in culpable
conduct, you should bear the full risk of whatever losses result. Some culpable
actors will pay nothing, for the risks they foresee will never materialize. Others will
pay amounts that are proportionate to their culpability. And some, like Ed, will pay
amounts greatly in excess of the risks they foresee. But if you engage in culpable
conduct, no matter how minimal your culpability, you are gambling that the losses
that result are no greater than what you deserve to pay, and that is a gamble that you
may lose big time.
We are not going to contest tort doctrine on this point. For we are dealing by
hypothesis with truly culpable conduct. Yet tort law extends its “eggshell skull” and
other doctrines of cataclysmic damages to conduct that is “faulty” but not culpable.
And even if it is fair to impose the risk of cataclysmic damage awards on the minimally
but truly culpable, it seems ludicrous to assert that it is also fair to impose that risk on
the nonculpable—the stupid, the ignorant, the forgetful, and so on. For however
“faulty” their conduct is in some sense, it is not truly culpable. Nor is it distinguishable
from the harm-producing conduct involved in cases of strict liability.
418 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

So we conclude that tort doctrine gives the wrong verdict in cases of catastrophic
injury when the defendant’s conduct is only “faulty” but not culpable. And we have
left unresolved the question whether tort doctrine has the correct verdict when it
is culpable conduct that results in catastrophic injury—though we are doubtful that
it does.
We leave the latter question unresolved because even in cases of truly culpable
conduct, we need to establish that defendant’s conduct resulted in plaintiff ’s injuries.
In other words, we need to link plaintiff ’s injuries to defendant’s conduct, whether or
not that conduct is culpable. Causation is supposed to be the linkage that we need. But
causation as the link between defendant’s conduct and plaintiff ’s injuries has proved
to be quite problematic.

B. Proximate causation
Before turning directly to factual causation, we should start with our significant
reservations about proximate causation. Determinations about proximate causation,
whether harms are “foreseeable” or “within the risk” or otherwise attributable to the
defendant, are certainly not metaphysical causation judgments. They are policy judg-
ments about the limits of liability.19 We will therefore put these questions to the side as
we address causation question. Policy limits are policy limits. (But even policy requires
principled arguments that are not masked as metaphysics.)
Moreover, recently, Michael Moore’s extensive study of the metaphysics of caus-
ation has sought to replace proximate causation with a more nuanced understanding
of causation itself.20 He parses what we think of as causation into counterfactual
dependence and true causation. He argues that both of these concepts are scalar. He
dispenses with foreseeability (as it crucially depends on the description of the harm,
and there is no principled method for selecting that description)21 and harm within
the risk (as all harms that do or might occur are within the risk, a point to which we
will return below). Moore then recasts the only remaining usage for proximate
causation—for when truly culpable actors (intentional, knowing, or reckless actors)
cause harm in a manner that they did not foresee—as questions of culpability, not
causation. Notably, this final arbiter of responsibility (a view we call below the
“mechanism within the plan”) is unavailable in cases of negligence and strict liability,
and therefore outside the arsenal of tort theory, because in those cases there is no plan.
Before leaving the policy decision masked as causation, it is worth making one final
observation about a rather odd discontinuity in tort law. If John rear-ends Jerry’s car,
and Jerry is an “eggshell skull” plaintiff, then as discussed in the previous section,

19
See Kimberly D. Kessler, “The Role of Luck in the Criminal Law,” 142 U. Pa. L. Rev. 2183 (1994), 2187–8.
20
Michael S. Moore, Causation and Responsibility: An Essay in Law, Morals, and Metaphysics (Oxford:
Oxford University Press, 2009), 371.
21
See Michael S. Moore, Placing Blame: A General Theory of the Criminal Law (Oxford: Oxford University
Press, 1997), Chapter 8.
 THE COLLAPSE OF TORT THEORY 419

despite the lack of foreseeability and so forth, John is on the hook for the full scope of
Jerry’s damages. But if John drives negligently, but the way that Jerry’s harms occur is
as unforeseeable (whatever that may mean) as Jerry’s eggshell status, then John is not
responsible for those harms. We fail to see a principled argument for distinguishing
the two; and importantly for our purposes here, if there is such an argument, it cannot
be grounded in the metaphysics of causation. It has to come from elsewhere.

C. Counterfactual dependence
The standard test for whether defendant’s conduct caused plaintiff ’s injury is the so-
called “but for” test: If plaintiff would not have suffered injury “but for” defendant’s
conduct, defendant’s conduct is the “cause” of plaintiff ’s injuries.22 This test, which
Michael Moore labels counterfactual dependence (CD),23 seems quite intuitive as what
we ordinarily mean when we claim that X “caused” Y’s injuries. Moore doesn’t believe
that CD is real causation, but he does believe that CD is sufficient to make defendant
responsible for plaintiff ’s injuries.24 (Moore needs CD to supplement “real” causation
as a basis for responsibility because, for example, “real” causation cannot account for
harm resulting from omissions or preventions.)25
CD as the linkage between defendant’s conduct (or omission, prevention, etc.) and
plaintiff ’s injuries is, however, beset by several difficulties. We shall leave aside cases of
overdetermination—cases in which there are two or more independently sufficient
causes of plaintiff ’s injuries, as when two independently set fires join to burn down
plaintiff ’s house, and each would have done so to the same extent at the same time had
it been the only fire. There are, in the literature, ingenious amendments to CD to
handle overdetermination cases (though one might justifiably regard these as ad hoc
attempts to save CD from a devastating counterexample, attempts that in essence
concede the inadequacy of CD.)26
We shall also put aside those cases in which defendant’s conduct results in plain-
tiff ’s death but also results in prolonging his life. (Defendant doctor maliciously gives
plaintiff a dose of radiation that will cause him to die of cancer in twenty years;
however, though it does kill plaintiff in twenty years, it also cures an undetected cancer
that would have killed plaintiff in one year. Or prospector B fills prospector C’s
canteen with sand, resulting in C’s dying of thirst. Unbeknownst to prospector B,
prospector A had filled C’s canteen with poison, which would have killed C much
sooner.) The CD test is met: But for defendant’s conduct, plaintiff would not have died

22
Restatement (Third) of Torts: Phys. & Emot. Harm } 26 (2010).
23
Moore, Causation and Responsibility (note 20) at 371.
24
Moore, Causation and Responsibility (note 20) at 391–425.
25
Moore, Causation and Responsibility (note 20) at 435–69.
26
Moore, Causation and Responsibility (note 20) at 87. See also Richard Wright, “Causation in Tort Law,”
73 Calif. L. Rev. 1775 (1985).
420 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

when and how he did. But it is also the case that but for defendant’s conduct, plaintiff
would have died differently but sooner.
One big problem for the CD test is illustrated by cases such as Berry v Village of
Sugar Notch.27 In Berry, recall, the question was whether driving the trolley over the
posted speed limit caused the injury that resulted from a dead tree falling on the
plaintiff. The claim was that, “but for” the speeding, the plaintiff would not have been
under the tree at the moment it fell. The court gave this claim a swift back of the hand.
But why?
It might be supposed that one reason for rejecting the claim is that had the motor-
man’s conduct been even worse than it was—had he been speeding even more—the
tree would not have fallen on the plaintiff. It would have fallen behind him, just as, had
there been no speeding, it would have fallen in front of him. The problem with this
response, however, is that it is often the case that if defendant’s conduct had been
worse than it was, the plaintiff would not have suffered harm. Had defendant speeded
down the residential street at 60 mph and not 50 mph, the infant whom he struck
when the infant crawled into the road from behind a parked car would have been
behind rather than in front of defendant’s car. Such examples are legion.
The other possible defense of Berry is to claim that the speed of the trolley does not
affect the risk of being hit by a falling dead tree. One is just as likely to be struck when
the trolley is traveling at or below the speed limit as when it is traveling above the limit.
But this reply, too, is incorrect in any case, such as Berry, in which plaintiff is hit by a
falling tree.
This latter “harm within the risk” (HWR) approach to limiting CD is misconceived
for various reasons. First, risk is an epistemic notion. It is someone’s estimate from
some perspective of the probability of some event’s occurrence given a certain act.
“Actual risks” are either one or zero, depending on what happens. In tort law, the
actual risk is always one: for plaintiff has been injured as a CD result of defendant’s act.
Second, the risks that are relevant to culpable actors are the basket of all the various
harms with their various probabilities that the actor ex ante would estimate. If one
takes each discrete risk separately and is sufficiently fine-grained in describing it, each
will be sufficiently improbable so as to render the defendant’s contemplated act
justifiable were it the only risk. The risk that defendant’s act will cause “bodily injury”
or “property damage” may be so high that defendant would be culpable in imposing
that risk. However, the risk that his act will break plaintiff ’s arm exactly three inches
above the wrist and cause plaintiff ’s car to suffer a precise amount of damage to the left
side passenger door may be tiny from an ex ante perspective (even though those risks
are one ex post).28 All potential harms go into the risk basket that determines whether
defendant was culpable for acting. If defendant has a duty to pen livestock on board his
ship in order to prevent the spread of disease, and if his failure to pen the livestock
increases, however slightly, the chance that the livestock will be swept overboard in a

27
43 A 240 (Pa S Ct, 1899).
28
Moore, Causation and Responsibility (note 20) at 194–5.
 THE COLLAPSE OF TORT THEORY 421

storm, the latter risk is part of the basket of risks that make defendant culpable. And
the same can be said for negligence; as Moore observes, “once one appreciates that all
risks created by a defendant count in an assessment of negligence, it would appear that
any and all harms that materialize from the defendant’s conduct are within the
category of risks that makes such conduct negligent.”29
Suppose one were to respond to the preceding argument by pointing out that if one
compares the reference class “speeding trolleys” with the reference class “non-speed-
ing trolleys,” the per mile incidence of injuries caused by falling rotting trees is exactly
the same for both reference classes. On the other hand, if one compares those reference
classes in terms other types of injuries and losses, one will find that speeding trolleys
result in more of them per mile than do non-speeding trolleys. Likewise, no matter
how improbable, the incidence of livestock washed overboard will turn out to be
higher if they are unpenned than if they are penned. Thus, on this reference class
interpretation of HWR, Berry is different from the other examples we gave, and the
harm there was not within the risk.
The problem with the reference class approach to HWR is that the selection of the
reference classes is completely arbitrary. If the reference classes are not “speeding
trolleys” and “non-speeding trolleys” but instead are “trolleys travelling at exactly the
same speed as Berry’s at exactly the same place and time, etc.,” and “all other trolley
trips,” one would find the incidence of loss in the former reference class—100 per
cent—is much higher than it is in the latter. And the use of these reference classes is no
more arbitrary than the use of “speeding trolleys” and “non-speeding trolleys.” Remem-
ber, we are already assuming the actor is culpable based on the risks he foresaw. And we
are not using reference classes for insurance purposes, which require, for actuarial
data, reference classes defined rather generally. HWR is neither a culpability measure
nor an insurance program.
A second big problem with CD as the linkage between defendant’s conduct and
plaintiff ’s injuries is CD’s unlimited scope, its “for want of a nail . . . ” aspect. If
defendant acts culpably, the culpable act’s consequences can radiate over space and
time without limits. If defendant culpably threatens plaintiff with a gun, and plaintiff,
fearful and anxious to get out of town, buys a ticket on an airplane that then crashes,
plaintiff ’s death is linked to defendant’s conduct by CD. The same is true if defendant
culpably drops a brick on plaintiff ’s toe, causing plaintiff to go to the medical clinic,
where he is stabbed by a deranged patient. The stabbing is linked by CD to defendant’s
act. And the linkage can be much more remote than that involved in these examples.
One possible rejoinder here is to claim that CD can be scalar.30 Defendant’s conduct
can be linked by CD to plaintiff ’s injuries in a “more of less” fashion.
It is difficult to see how CD can be scalar, however. Something either is or is not a
“but for” cause. If there are to be limitations on CD, they will have to come from
something other than scalarity.

29
Moore, Causation and Responsibility (note 20) at 185.
30
Moore argues that CD can be scalar. See Moore, Causation and Responsibility (note 20) at 468.
422 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

Moore argues that CD’s scalarity is a function of how close are the possible worlds
in which “but for” the defendant’s act, the plaintiff escapes or does not escape
injury.31 If there are only very nearby possible worlds where defendant acts as he
actually did and plaintiff is injured, then defendant is only a little CD responsible for
plaintiff ’s injuries. On the other hand, if there are more remote possible worlds in
which defendant acts as he actually did and plaintiff is injured, defendant’s CD
responsibility for the injuries is large.
As we said in an earlier article, we find the possible worlds account of CD scalarity
curious.32 Moreover, in order to save CD from affirming recovery in Berry, he has to
swallow an equally unappetizing result in cases of omission liability. For Moore,
because he rejects the possibility of omissions “causing” harm, omitters can only be
responsible for harms through CD. But if CD can be scalar, and that is why the
speeding in Berry is not responsible for the harm, then parents who omit to feed their
child are only a wee bit responsible for the child’s death through starvation when Aunt
Emma would have arrived with food had she not been riding on the Sugar Notch
trolley and knocked unconscious by the falling tree.
If CD is nonscalar, as we believe it must be, then, because it will be present in all
cases of Moore’s true causation—other than the very rare cases of overdetermination
of exactly the same harm at exactly the same time—we will be left with counterintuitive
assignments of responsibility for all harms resulting from culpable acts no matter how
remote. The death of the person who flees the defendant’s beatings by buying a ticket
on a plane that crashes will be the defendant’s responsibility, for CD is present.

D. True causation
If liability for harms resulting from culpable acts cannot be limited in cases of CD, can
it be limited in cases of true metaphysical (non-CD) causation?

31
See Michael S. Moore, “Causation Revisited,” 42 Rutgers L.J. 451 (2011), 454.
32
We noted, moreover, that there are an indefinite number of possible worlds in which the tree injures the
plaintiff in Berry:
Moreover, why are the nearest possible worlds those in which we hold everything constant except the
motorman’s speed? Change the color of the motorman’s shirt and he still gets hit by the tree. Change the
number of birds singing and he still gets hit. Change his speed a tiny bit, his time of departure a tiny bit,
and he still gets hit. Indeed, one can change an infinite number of things without altering the result. If
there are an infinite number of nearby possible worlds in which the same result occurs, then even if there
are an infinite number in which it does not, infinity over infinity does not lead to a “small CD”
conclusion.
Larry Alexander and Kimberly Kessler Ferzan, “ ‘Moore or Less’ Causation and Responsibility,” 6 Crim. L. &
Phil. 81 (2012), 85–6.
Indeed, we can have an infinite number of possible worlds that are quite remote from the actual world
but in which the same result occurs. All we need are the time of the falling tree, the time of departure of
the trolley, the distance from the departure point to the tree, and the speed.
 THE COLLAPSE OF TORT THEORY 423

Moore claims not only that CD is scalar but also that “real” causation is scalar as
well. He analogizes “real” causal chains to the ripples in a pond caused by tossing a
pebble into it. The ripples gradually peter out, and so too does causation.
The idea that the causal linkage between a culpable act and its injurious effects is one
of gradual diminishment seems to work for physical causation. It works less well, if at
all, for culpable acts that cause harm by giving the victims or intermediaries reasons
for acting. If the defendant lures his victim into harm’s way, solicits a hit man to break
the victim’s legs, or gives a letter bomb to the mail carrier with instructions to deliver it
to the victim, it is hard to see how scalarity gets into the picture.
Take a case of luring. Suppose Defendant wants to kill Victim and knows that the
Bridge of San Luis Rey will collapse under Victim’s weight. Victim is planning to walk
to town. One route takes him across the Bridge. It is longer but more scenic. The other
route is shorter and less scenic. It is a close choice, but ultimately Victim decides to
take the shorter route. At this point, Defendant calls Victim and urges him to take the
longer route over the Bridge, offering, as an inducement, companionship on the walk
once Victim crosses the Bridge. That tips the balance of reasons for Victim in favor of
the longer route, and he subsequently plummets to his death.
Just how much, on a scalar view of causation, is Defendant causally responsible for
Victim’s death? He offered Victim a reason to cross the Bridge, a reason which perhaps
was not terribly weighty but was just weighty enough to shift the balance of reasons in
favor of crossing the Bridge. Is Defendant, therefore, only a small cause of Victim’s
death and only minimally causally responsible for it? How does causation’s scalarity
work here?
Or take the letter bomb case. If Unabomber 1, who hatches the plot and builds the
bomb, gives it to his confederate, Unabomber 2, who delivers it to Victim, it would
seem to us that Unabomber 1 is as responsible for Victim’s death or injury as is
Unabomber 2. Moore says that accomplices are less responsible causally than are
principals, which would imply that Unabomber 1 would be less responsible than
Unabomber 2.33 And if there were a third confederate, Unabomber 3, to whom
Unabomber 2 delivers the better bomb, Unabomber 1 would be even less responsible
for the harms suffered by Victim.
All of this is wildly counterintuitive. To see why, assume Unabomber 1 is the only
culpable actor. He merely gives his letter bomb to an unsuspecting mail carrier, who in
turn hands it over to equally unsuspecting personnel at the local post office, who in
turn hand it over to other unsuspecting postal workers, one of whom ultimately gives
it to the Victim. There may be many innocent actors in the causal chain. We would
assume that in such a case—the real one that occurred—Unabomber 1 would be fully
responsible for Victim’s death or injuries. Yet the causal chain from Unabomber 1’s
culpable act to Victim’s injuries is just as attenuated (or not so) as it is when the
intermediaries are in on the plot. Scalarity does not appear to be in play here.

33
See Moore, Causation and Responsibility (note 20) at 319–23.
424 LARRY ALEXANDER AND KIMBERLY KESSLER FERZAN

E. Recasting proximate causation as culpability

If intervening human acts do not break causal chains—and we agree with Moore that
they do not;34 if scalarity does not work as a limitation on responsibility; and,
moreover, if nonscalar CD is omnipresent, then there are no obvious limitations on
culpable acts that reduce or eliminate responsibility for resulting harms. In this final
section, we turn to Moore’s last method for cabining causation by viewing it as
culpability.
At the outset, it is worth noting that Moore specifically denies that this approach
works for negligence.35 There is no safe harbor for tort theorists. They will need to
come up with a way to limit causation that neither depends upon metaphysics nor
fault.
Back to Moore. We return to an example we have used elsewhere of two Pierres.36
One Pierre intends to kill Monique with a bullet but misses, with the result that a
frightened Monique runs outside and is killed by lightning. The other Pierre intends to
miss Monique with a bullet and frighten her into running outside in a dangerous
thunderstorm, a plan that succeeds. At one point Moore acknowledges that both
Pierres are equally causes of Monique’s death, but are only “small” causes.
Treating the second Pierre, whose plan succeeded precisely as he intended, as only a
small cause of Monique’s death, and therefore only minimally responsible, seems
perverse. And indeed Moore introduces a new concept—the “mechanism within the
plan” (MWP)—to distinguish the Pierres in terms of their causal responsibility.37 The
second Pierre’s causal chain satisfies MWP, whereas the first Pierre’s does not.
(Of course, there could be a third Pierre, one who, like the first Pierre, was trying to
kill Monique with the bullet, but who also, like the second Pierre, contemplated the
possibility that the bullet might miss but frighten Monique into running outside and
getting struck by lightning. Would MWP be satisfied with this third Pierre? If so,
should this slight difference between the first Pierre and the third in terms of what they
contemplated result in a greater causal responsibility for the latter?)
The problems with MWP as a limitation on either causal or CD responsibility are
fatal, however. First, it is difficult to understand why the content of defendant’s plan,
which has already been factored into his culpability, should then reappear as a limit on
causal responsibility. Why does the first Pierre, who at first seems quite culpable
because he intends to kill, somehow become less culpable when he says, “I intended to
kill her and she died, but it didn’t happen the way I planned.” What is culpability
diminishing about that? In addition, Moore admits, as he should, that defendant’s

34
Moore, Causation and Responsibility (note 20) at 278–9.
35
Moore, Causation and Responsibility (note 20) at 102 n. 69.
36
Alexander and Ferzan, Crime and Culpability (note 9) at 184–5.
37
MWP is our term, not Moore’s. What Moore is asking is how closely does the actual causal (or CD) path
from defendant’s culpable act to the victim’s injuries match the path that defendant envisioned when he acted.
(In all cases of true culpability, defendant will be aware that his conduct risks harm.)
 THE COLLAPSE OF TORT THEORY 425

mental states themselves do not produce effects in the world beyond their effects on
how he acts. There is no telekinesis, says Moore, and we agree. But if defendant’s
mental states themselves are not causes, why are they limitations on causal
responsibility?
Second, for those torts that are based on recklessness, negligence, and strict liability,
defendant’s plan is immaterial. For inadvertent negligence and strict liability, defend-
ant’s mental state at the time he acts is not material. Recklessness does turn on
defendant’s mental state, but the reckless defendant merely adverts to the basket of
risks his act creates and does not involve any “plan” or specific conception of the many
ways those risks might be realized. Recklessness culpability is awareness-based but not
plan-based.
As we said elsewhere regarding a MWP limitation on recklessness:
Consider People v Acosta.38 Acosta, driving fast to escape pursuing police cars and police
helicopters, was aware that he was acting in an unjustifiably risky manner and thus
recklessly. Moore would ask whether Acosta was reckless as to the actual harm he caused
(the helicopter crash). Suppose Acosta considered each risk he was creating in a fine-
grained manner, including the risk of the helicopter crash. And suppose none of those
risks taken individually would render Acosta reckless. Only those risks in the aggregate
would. It would be counterintuitive to say that Acosta was not reckless with regard to the
helicopter crash (or any other harm) even though that perceived risk, taken in isolation,
would not have made him reckless. And the analysis should not change if Acosta did not
individuate the risks in this fine-grained way but considered his conduct “very risky” in
light of the aggregated various dangers of pursuing helicopters or of the risk they might
crash any more than it would matter that he would have been unaware of the particular
pedestrian whom he might have hit while recklessly speeding.39

We conclude that the quest for principled limitations on causal or CD responsibility

has come up empty. It is not for lack of trying by many able minds. Rather, we think
the quest is doomed to fail. We can assess the defendant’s culpability. And perhaps we
can assess whether defendant’s culpable act was a CD “cause”—though we believe that
the CD inquiry may rest on shaky metaphysical grounds. But if culpability plus CD
seems too capacious an ambit of tort liability, we see no tenable way to limit it.

IV. Conclusion
If there is no fault in not adverting, if there is no principled distinction between
negligence and strict liability, and if there is no principled limit on causation, then
nothing cabins the reach of tort law. An ad hoc and unprincipled system is simply not
the way to run a railroad. It isn’t even the way to run a trolley system in the Borough of
Sugar Notch.

38
284 Cal Rptr 117 (Cal Ct of Apps, 1991).
39
Larry Alexander and Kimberly Kessler Ferzan, “Ferzander’s Surrebuttal,” 6 Crim. L. & Phil. 463 (2012).
 BIBLIOGRAPHY

Books and Articles

Abraham, Kenneth, The Liability Century: Insurance and Tort Law from the Progressive Era to
9/11 (Cambridge, MA: Harvard University Press, 2008)
Acton, H.B. (ed.), The Philosophy of Punishment: A Collection of Papers (London: Macmillan,
1969)
Adams, Robert Merrihew “Involuntary Sins,” 94 Philosophical Review 3 (1985)
Aristotle, Nichomachean Ethics, trans. Roger Crisp (.) (Cambridge: Cambridge University
Press, 2000)
Aristotle, Nicomachean Ethics, trans. T. Irwin (2d ed.) (Indianapolis: Hackett, 1999)
Alexander, Larry, “Causation and Corrective Justice: Does Tort Law Make Sense?” 6 Law and
Philosophy 1 (1987)
Alexander, Larry, “Reconsidering the Relationship Among Voluntary Acts, Strict Liability, and
Negligence in Criminal Law,” 7 Social Philosophy and Policy 84 (1990)
Alexander, Larry, “Foreword: Coleman and Corrective Justice,” 15 Harvard Journal of Law and
Public Policy 621 (1992)
Alexander, Larry, “The Moral Magic of Consent (II),” 2 Legal Theory 165 (1996)
Alexander, Larry and Kimberly Kessler Ferzan, Crime and Culpability: A Theory of Criminal
Law (Cambridge: Cambridge University Press, 2009)
Alexander, Larry and Kimberly Kessler Ferzan, “Ferzander’s Surrebuttal,” 6 Criminal Law and
Philosophy 463 (2012)
Alexander, Larry and Kimberly Kessler Ferzan, “‘Moore or Less’ Causation and Responsibility,”
6 Criminal Law and Philosophy 81 (2012)
Anscombe, G.E.M., “Who is Wronged?” 5 Oxford Review 16 (1967)
Ashworth, Andrew, Sentencing and Criminal Justice (5th ed.) (Oxford: Oxford University Press,
2010)
Austin, J.L., “Ifs and Cans,” 42 Proceedings of the British Academy 109 (1961)
Austin, Lisa and Dennis Klimchuk (eds.), The Rule of Law and Private Law (Oxford: Oxford
University Press) (forthcoming)
Avraham, Ronen and Issa Kohler-Hausmann, “Accident Law for Egalitarians,” 12 Legal Theory
181 (2006)
Bacon, Francis, “Of Revenge,” in Brian Vickers (ed.), Francis Bacon: The Major Works (Oxford:
Oxford University Press, 2002)
Baier, Kurt, “Types of Responsibility,” in Peter A. French (ed.), The Spectrum of Responsibility
(New York: St. Martin’s Press, 1972)
Barber, Nicholas, The Constitutional State (Oxford: Oxford University Press, 2010)
Barker, Vanessa, The Politics of Imprisonment: How the Democratic Process Shapes the Way
America Punishes Offenders (New York: Oxford University Press, 2009)
Barry, Brian, Justice as Impartiality (New York: Oxford University Press, 1996)
Bass, Gary, “Jus Post Bellum,” 32 Philosophy and Public Affairs 4 (2004)
Beever, Allan, Rediscovering the Law of Negligence (Oxford: Hart, 2007)
 BIBLIOGRAPHY 427

Beever, Allan, “The Form of Liability in the Torts of Trespass,” 40 Common Law World Review
378 (2011)
Bennett, Christopher, The Apology Ritual (Cambridge: Cambridge University Press, 2008)
Benson, Peter, “The Basis of Corrective Justice and its Relationship to Distributive Justice,” 77
Iowa Law Review 515 (1992)
Bishop, William, “The Rational Strength of the Private Law Model,” 40 University of Toronto
Law Journal 663 (1990)
Blackstone, William, Commentaries on the Laws of England, A.W. Brian Simpson (ed.)
(Chicago: University of Chicago Press, 1979)
Blom-Cooper, L., G. Drewry, and B. Dickinson (eds.), The Judicial House of Lords 1876–2009
(Oxford: Oxford University Press, 2009)
Bohlen, Francis, “Incomplete Privilege to Inflict Intentional Invasions of Interests of Property
and Personality,” 39 Harvard Law Review 307 (1926)
Boxill, Bernard, “A Lockean Argument for Black Reparations,” 7 Journal of Ethics 63 (2003)
Braithwaite, John, “Repentance Rituals and Restorative Justice,” Journal of Political Philosophy
8.1 (2000)
Bridgeman, Curtis, “Note, Corrective Justice in Contract Law: Is There a Case for Punitive
Damages?” 56 Vanderbilt Law Review 237 (2003)
Brudner, Alan, The Unity of the Common Law: Studies in Hegelian Jurisprudence (Berkeley:
University of California Press, 1995)
Calabresi, Guido, “Some Thoughts on Risk Distribution in the Law of Torts,” 70 Yale Law
Journal 499 (1961)
Calabresi, Guido, The Costs of Accidents (New Haven: Yale University Press, 1970)
Calabresi, Guido, “Concerning Cause and the Law of Torts” 43 University of Chicago Law
Review 69 (1975)
Calabresi, Guido, “Torts—The Law of the Mixed Society,” 56 Texas Law Review 519 (1978)
Calabresi, Guido, “Toward A Unified Theory of Torts,” 1 Journal of Tort Law 3 (2007)
Calabresi, Guido and A. Douglas Melamed, “Property Rules, Liability Rules, and Inalienability:
One View of the Cathedral,” 85 Harvard Law Review 1089 (1972)
Calhoun, Cheshire (ed.), Setting the Moral Compass: Essays by Women Philosophers (Oxford:
Oxford University Press, 2003)
Campbell, Joseph Keim, Michael O’Rourke, and David Shier (eds.), Law and Social Justice
(Cambridge, MA: MIT Press, 2005)
Cane, Peter, Tort Law and Economic Interests (2d ed.) (New York: Clarendon, 1996)
Cane, Peter, “Damages in Public Law,” 9 Otago Law Review 489 (1999)
Cane, Peter, “Distributive Justice in Tort Law,” 4 New Zealand Law Review 401 (2001)
Cane, Peter, Responsibility in Law and Morality (Oxford: Hart, 2002)
Cane, Peter, Administrative Law (5th ed.) (Oxford: Oxford University Press, 2011)
Cane, Peter and John Gardner (eds.), Relating to Responsibility: Essays in Honor of Tony Honoré
on his 80th Birthday (Oxford: Hart, 2001)
Claudia Card, The Atrocity Paradigm (New York: Oxford University Press, 2002)
Chamallas, Martha and Jennifer B. Wriggins, The Measure of Injury: Race, Gender, and Tort
Law (New York: New York University Press, 2010)
Chamberlain, Erika, “Misfeasance in a Public Office: A Justifiable Anomaly within the Rights-
Based Approach?” in Donal Nolan and Andrew Robertson (eds.), Rights and Private Law
(Oxford: Hart, 2012)
428 BIBLIOGRAPHY

Christie, Nils, “Conflicts as Property,” 17 British Journal of Criminology 1 (1977)

Claeys, Eric, “Property 101: Is Property a Thing or a Bundle?” 32 Seattle University Law Review
617 (2009)
Claeys, Eric, “Jefferson Meets Coase: Land-Use Torts, Law and Economics, and Natural
Property Rights,” 85 Notre Dame Law Review 1379 (2010)
Claeys, Eric, “Exclusion and Exclusivity in Gridlock,” 53 Arizona Law Review 9 (2011)
Claeys, Eric, “Private Law Theory and Corrective Justice in Trade Secrecy,” 4 Journal of Tort
Law 2 (2011)
Claeys, Eric, “Exclusion and Private Law Theory: A Comment on Property as the Law of
Things,” 125 Harvard Law Review Forum 133 (2012)
Claeys, Eric, “Locke Unlocked: Productive Use in Trespass, Adverse Possession, and Labor
Theory” (unpublished) (http://ssrn.com/abstract=1759551)
Claeys, Eric, “On the Use and Abuse of Overflight Columns,” 2 Brigham-Kanner Property
Rights Conference Journal (forthcoming)
Claeys, Eric, “Productive Use in Acquisition, Accession, and Labor Theory,” in James Penner
and Henry Henry Smith (eds.), Philosophical Foundations of Property Law (Oxford: Oxford
University Press) (forthcoming)
Coase, Ronald, “The Problem of Social Cost,” 4 Journal of Law and Economics 1 (1960)
Cohen, David, “Suing the State,” 40 University of Toronto Law Journal 630 (1990)
Cohen, David and J.C. Smith, “Entitlement and the Body Politic: Rethinking Negligence in
Public Law,” 64 Canadian Bar Review 1 (1986)
Cohen, G.A., “On the Currency of Egalitarian Justice,” 99 Ethics 906 (1989)
Cohen, G.A., “Where the Action Is: On the Site of Distributive Justice,” 26 Philosophy and
Public Affairs 3 (1997)
Cohen, Gerard, Rescuing Justice and Equality (Cambridge, MA: Harvard University Press, 2008)
Cohen, Morris, “Property and Sovereignty,” 13 Cornell Law Quarterly 8 (1924)
Colby, Thomas, “Beyond the Multiple Punishment Problem: Punitive Damages as Punishment
for Individual, Private Wrongs,” 87 Minnesota Law Review 583 (2003)
Coleman, Jules, “The Structure of Tort Law,” 97 Yale Law Journal 1233 (1988)
Coleman, Jules, “Adding Institutional Insult to Personal Injury,” 8 Yale Journal on Regulation
223 (1991)
Coleman, Jules, Risks and Wrongs (Cambridge: Cambridge University Press, 1992)
Coleman, Jules, “The Practice of Corrective Justice,” in David Owen (ed.), Philosophical
Foundations of Tort Law (Oxford: Oxford University Press, 1995)
Coleman, Jules, The Practice of Principle: In Defence of a Pragmatist Approach to Legal Theory
(Oxford: Oxford University Press, 2001)
Coleman, Jules, “Facts, Fictions, and the Grounds of Law,” in Joseph Keim Campbell, Michael
O’Rourke, and David Shier (eds.), Law and Social Justice (Cambridge, MA: MIT Press, 2005)
Coleman, Jules, “Epilogue to Risks and Wrongs,” (unpublished manuscript) (http://ssrn.com/
abstract=1679554)
Coleman, Jules and Jody Kraus, “Rethinking the Legal Theory of Rights,” 95 Yale Law Journal
1335 (1985)
Coleman, Jules and Gabriel Mendlow, “Theories of Tort Law,” in Edward N. Zalta (ed.), The
Stanford Encyclopedia of Philosophy (Fall 2010 edition) (http://plato.stanford.edu/archives/
fall2010/entries/tort-theories/)
 BIBLIOGRAPHY 429

Coleman, Jules and Arthur Ripstein, “Mischief and Misfortune,” 41 McGill Law Journal 91
(1995)
Coleman, Jules and Scott Shapiro (eds.), The Oxford Handbook of Jurisprudence and Philosophy
of Law (Oxford: Oxford University Press, 2004)
Cooley, Thomas, A Treatise on the Law of Torts or the Wrongs Which Arise Independent of
Contract (Littleton: Fred B. Rothman and Co., 1993)
Cooter, Robert, “Prices and Sanctions,” 84 Columbia Law Review 1523 (1984)
Cornford, Tom, Towards a Public Law of Tort (Aldershot: Ashgate, 2008)
Craswell, Richard, “Deterrence and Damages: The Multiplier Principle and its Alternatives,” 97
Michigan Law Review 2185 (1999)
Cruft, Rowan, Matthew H. Kramer, and Mark R. Reiff (eds.), Crime, Punishment, and Respon-
sibility (Oxford: Oxford University Press, 2011)
Dagan, Hanoch, “The Distributive Foundations of Corrective Justice,” 98 Michigan Law Review
138 (1999)
Dagan, Hanoch, The Law and Ethics of Restitution (Cambridge: Cambridge University Press,
2004)
Dan-Cohen, Meir, “Decision Rules and Conduct Rules: On Acoustic Separation in Criminal
Law,” 97 Harvard Law Review 625 (1984)
Darwall, Stephen and Julian Darwall, “Civil Recourse as Mutual Accountability” 39 Florida
State University Law Review (2012)
Dau-Schmidt, Kenneth, “An Economic Analysis of the Criminal Law as a Preference-Shaping
Policy,” 1990 Duke Law Journal 1 (1990)
Dobbs, Dan (ed.), Prosser and Keeton on the Law of Torts (5th ed.) (St. Paul: West, 1984)
Dobbs, Dan, The Law of Torts, Vol. 1 (Practitioner Treatise Series) (St. Paul: West, 2001)
Dobbs, Dan, P. Hayden, and E. Bublick, The Law of Torts (Eagan: West, 2011)
Dodge, William, “The Case for Punitive Damages in Contracts,” 48 Duke Law Journal 629
(1999)
Dolinko, David, “Three Mistakes of Retributivism,” 39 University of California Los Angeles Law
Review 1623 (1992)
Donohue, John and Justin Wolfers, “Uses and Abuses of Empirical Evidence in the Death
Penalty Debate,” 58 Stanford Law Review 791 (2005)
Donohue, John and Justin Wolfers, “Estimating the Impact of the Death Penalty on Murder,”
11 American Law and Econonimcs Review 249 (2009)
Dorfman, Avihay, “The Normativity of the Private Ownership Form,” 75 Modern Law Review
981 (2012)
Dorfman, Avihay, “Reasonable Care: Equality as Objectivity,” 41 Law and Philosophy 369 (2012)
du Bois, Francois, “Human Rights and the Tort Liability of Public Authorities,” 127 Law
Quarterly Review 589 (2011)
du Bois, Francois, “Social Purposes, Fundamental Rights and the Judicial Development of
Private Law,” in Donal Nolan and Andrew Robertson (eds.), Rights and Private Law (Oxford:
Hart, 2012)
R.A. Duff, Trials and Punishments (New York: Cambridge University Press, 1991)
R.A. Duff, Punishment, Communication and Community (New York: Oxford University Press,
2001)
R.A. Duff, Answering for Crime (Oxford: Hart, 2007)
R.A. Duff, “Towards a Modest Legal Moralism,” 8 Criminal Law and Philosophy (forthcoming)
430 BIBLIOGRAPHY

R.A. Duff, Lindsay Farmer, S.E. Marshall, Massimo Renzo, and Victor Tadros (eds.), The
Constitution of Criminal Law (Oxford: Oxford University Press, 2012)
Dworkin, Ronald, Taking Rights Seriously (Cambridge, MA: Harvard University Press, 1977)
Dworkin, Ronald, “What is Equality? Part 2:00 Equality of Resources,” 10 Philosophy and Public
Affairs 283 (1981)
Dworkin, Ronald, Law’s Empire (Cambridge, MA: Harvard University Press, 1986)
Dworkin, Ronald, Sovereign Virtue: The Theory and Practice of Equality (Cambridge, MA:
Harvard University Press, 2000)
Eisenberg, Theodore, M. Heise, and M. Wells, “Variability in Punitive Damages: An Empirical
Assessment of the U.S. Supreme Court’s Decision in Exxon Shipping Co. v. Baker,” 166
Journal of Institutional and Theoretical Economics 5 (2010)
Ellis, Dorsey, Jr., “Fairness and Efficiency in the Law of Punitive Damages,” 56 Southern
California Law Review 1 (1982)
Enoch, David, “Luck between Morality, Law, and Justice,” 9 Theoretical Inquiries in Law 23 (2008)
Enoch, David, “Giving Practical Reasons,” 11 The Philosopher’s Imprint 4 (March 2011)
Enoch, David, “Being Responsible, Taking Responsibility, and Penumbral Agency,” in Ulrike
Heuer and Gerlad Lang (eds.), Luck, Value, and Commitment: Themes from the Ethics of
Bernard Williams (Oxford: Oxford University Press, 2012)
Enoch, David and Andrei Marmor, “The Case against Moral Luck,” 26 Law and Philosophy 405
(2007)
Epstein, Richard, “A Theory of Strict Liability,” 2 Journal of Legal Studies 151 (1972)
Epstein, Richard, “Defenses and Subsequent Pleas in a System of Strict Liability,” 3 Journal of
Legal Studies 165 (1974)
Epstein, Richard, “Nuisance Law: Corrective Justice and Its Utilitarian Constraints,” 8 Journal
of Legal Studies 49 (1978)
Epstein, Richard, “Possession As the Root of Title,” 13 Georgia Law Review 1221 (1979)
Epstein, Richard, Torts (New York: Aspen, 1999)
Epstein, Richard, “Intel v. Hamidi: The Role of Self-Help in Cyberspace?” 1 Journal of Law,
Economics, and Policy 147 (2005)
Epstein, Richard, “Toward a General Theory of Tort Law: Strict Liability in Context,” 3 Journal
of Tort Law 1 (2010)
Esper, Dilan and Gregory Keating, “Abusing ‘Duty’” 79 Southern California Law Review 265 (2006)
Esper, Dilan and Gregory Keating, “Putting ‘Duty’ in Its Place: A Reply to Professors Goldberg
and Zipursky,” 41 Loyola of Los Angeles Law Review 1225 (2008)
Farmer, Lindsay, “Codifying Crime, Finding Government: Reconstructing the English Codification
Debate: The Criminal Law Commissioners 1833–45,” 18 Law and History Review 397 (2000)
Feinberg, Joel, Doing and Deserving (Princeton, NJ: Princeton University Press, 1970)
Feinberg, Joel, “The Nature and Value of Rights,” 4 Journal of Value Inquiry 243 (1970)
Feinberg, Joel, Harm to Others (Oxford: Oxford University Press, 1986)
Feldthusen, Bruce, “Failure to Confer Discretionary Public Benefits: The Case for Complete
Negligence Immunity,” 5 Tort Law Review 17 (1997)
Finkelstein, Claire, “Is Risk a Harm?” 151 University of Pennsylvania Law Review 963 (2003)
Finnis, John, Natural Law and Natural Rights (Oxford: Clarendon, 1980)
Finnis, John, “Natural Law: The Classical Tradition,” in Jules Coleman and Scott Shapiro (eds.),
The Oxford Handbook of Jurisprudence and Philosophy of Law (Oxford: Oxford University
Press, 2004)
 BIBLIOGRAPHY 431

Fischer, John Martin and Mark Ravizza, “Précis of Responsibility and Control: A Theory of
Moral Responsibility,” 61 Philosophy and Phenomenological Research 441 (2000)
Flaspohler, Paul, J. Elfstrom, K. Vanderzee, H. Sink, and Z. Birchmeier, “Stand by Me: The
Effects of Peer and Teacher Support in Mitigating the Impact of Bullying on Quality of Life,”
46 Psychology in the Schools 7 (2009)
Fletcher, George, “Fairness and Utility in Tort Theory,” 85 Harvard Law Review 537 (1972)
French, Peter A. (ed.), The Spectrum of Responsibility (New York: St. Martin’s Press, 1972)
Fried, Barbara, “The Limits of a Nonconsequentialist Approach to Torts,” 18 Legal Theory
231 (2012)
Friedmann, Daniel and Daphne Barak-Erez (eds.), Human Rights in Private Law (Oxford: Hart,
2001)
Galanter, Marc and David Luban, “Poetic Justice: Punitive Damages and Legal Pluralism,” 42
American University Law Review 1393 (1993)
Gardner, John, “The Gist of Excuses,” 1 Buffalo Law Review 575 (1998)
Gardner, John, “Obligations and Outcomes in the Law of Torts,” in Peter Cane and John
Gardner (eds.), Relating to Responsibility: Essays in Honor of Tony Honoré on His 80th
Birthday (Oxford: Hart 2001)
Gardner, John, “Backward and Forward with Tort Law,” in Joseph Keim Campbell, Michael
O’Rourke, and David Shier (eds.), Law and Social Justice (Cambridge, MA: MIT Press, 2005)
Gardner, John, “Crime: In Proportion and in Perspective,” in John Gardner, Offences and
Defences: Selected Essays in the Philosophy of Criminal Law (Oxford: Oxford University
Press, 2007)
Gardner, John, “Justifications and Reasons” in Offences and Defences: Selected Essays in the
Philosophy of Criminal Law (Oxford: Oxford University Press, 2007)
Gardner, John, “Torts and Other Wrongs,” 39 Florida State University Law Review 43 (2011)
Gardner, John, “What Is Tort Law For? Part I—The Place of Corrective Justice,” 30 Law and
Philosophy 1 (2011)
Gardner, John, “Criminals in Uniform,” in R.A. Duff, Lindsay Farmer, S.E. Marshall, Massimo
Renzo, and Victor Tadros (eds.), The Constitution of Criminal Law (Oxford: Oxford Uni-
versity Press, 2012)
Gardner, John, Law as a Leap of Faith (Oxford: Oxford University Press, 2012)
Gardner, John, “Some Rule-of-Law Anxieties about Strict Liability in Private Law,” in Lisa
Austin and Dennis Klimchuk (eds.), The Rule of Law and Private Law (Oxford: Oxford
University Press) (forthcoming)
Gavison, Ruth (ed.), Issues in Contemporary Legal Philosophy (Oxford: Clarendon, 1987)
Geistfeld, Mark, “Placing a Price on Pain and Suffering: A Method for Helping Juries Determine
Tort Damages for Nonmonetary Injuries,” 83 California Law Review 773 (1995)
Geistfeld, Mark, “Economics, Moral Philosophy, and the Positive Analysis of Tort Law,” in
Gerald J. Postema (ed.), Philosophy and the Law of Torts (New York: Cambridge University
Press, 2001)
Geistfeld, Mark, “Reconciling Cost-Benefit Analysis with the Principle that Safety Matters More
than Money,” 76 New York University Law Review 114 (2001)
Geistfeld, Mark, “Punitive Damages, Retribution, and Due Process,” 81 Southern California
Law Review 263 (2008)
Geistfeld, Mark, Tort Law: The Essentials (New York: Walters Kluwer, 2008)
432 BIBLIOGRAPHY

Geistfeld, Mark, “Efficiency, Fairness, and the Economic Analysis of Tort Law,” in Mark D.
White (ed.), Theoretical Foundations of Law and Economics (New York: Cambridge Univer-
sity Press, 2009)
Geistfeld, Mark, “The Principle of Misalignment: Duty, Damages, and the Nature of Tort
Liability,” 121 Yale Law Journal 142 (2011)
Geistfeld, Mark, Principles of Products Liability (New York: Foundation Press, 2011)
Geistfeld, Mark, “Tort Law and the Inherent Limitations of Monetary Exchange: Property
Rules, Liability Rules, and the Negligence Rule,” 4 Journal of Tort Law 1 (2011)
Geistfeld, Mark, “The Coherence of Compensation-Deterrence Theory in Tort Law,” 61 DePaul
Law Review 383 (2012)
Ghirardi, James and John Kircher, Punitive Damages Law and Practice (West, 1997)
Gold, Andrew, “The Taxonomy of Civil Recourse,” 39 Florida State University Law Review 65
(2011)
Goldberg, John, “Twentieth-Century Tort Theory,” 91 Georgetown Law Journal 513 (2002)
Goldberg, John, “Two Conceptions of Tort Damages: Fair v. Full Compensation,” 55 DePaul
Law Review 435 (2006)
Goldberg, John and Benjamin Zipursky, “The Moral of MacPherson,” 146 University of
Pennsylvania Law Review 1733 (1998)
Goldberg, John and Benjamin Zipursky, “Shielding Duty,” 79 Southern California Law Review
329 (2006)
Goldberg, John and Benjamin Zipursky, “Intervening Wrongdoing in Tort: The Restatement
(Third)’s Unfortunate Embrace of Negligent Enabling,” 44 Wake Forest Law Review 1211
(2009)
Goldberg, John and Benjamin Zipursky, The Oxford Introductions to Law: Torts (New York:
Oxford University Press, 2010)
Goldberg, John and Benjamin Zipursky, Torts (New York: Oxford University Press, 2010)
Goldberg, John and Benjamin Zipursky, “Torts as Wrongs,” 88 Texas Law Review 917 (2010)
Goldberg, John and Benjamin Zipursky, “Civil Recourse Revisited,” 39 Florida State University
Law Review 243 (2011)
Goldberg, John and Benjamin Zipursky, “Rights and Responsibility in the Law of Torts,” in
Donal Nolan and Andrew Robertson (eds.), Rights and Private Law (Oxford: Hart 2012)
Goldberg, John and Benjamin Zipursky, “Civil Recourse Theory Defended: A Reply to Posner,
Calabresi, Rustad, Chamallas and Robinette,” 88 Indiana Law Journal (forthcoming)
Goldberg, John, Anthony Sebok, and Benjamin Zipursky, Tort Law: Responsibilities and Redress
(3d ed.) (New York: Wolters Kluwer, 2012)
Gotanda, John, “Charting Developments Concerning Punitive Damages: Is the Tide Chan-
ging?” 45 Columbia Journal Transnational Law 507 (2007)
Goudkamp, John, Tort Law Defences (Oxford: Hart, 2013)
Grady, Mark, “Why Are People Negligent? Technology, Nondurable Precautions, and the
Medical Malpractice Explosion,” 82 North Western University Law Review 293 (1988)
Gregory, Charles, “Trespass to Negligence to Absolute Liability,” 37 Virginia Law Review 359
(1951)
Grey, Thomas, “Accidental Torts,” 54 Vanderbilt Law Review 1225 (2001)
Griswold, Charles, Forgiveness: A Philosophical Exploration (New York: Cambridge University
Press, 2007)
 BIBLIOGRAPHY 433

Grosskopf, Ofer, “Horizontal Equality and the Law of Torts,” in Daniel Friedmann and Daphne
Barak-Erez (eds.), Human Rights in Private Law (Oxford: Hart, 2001)
Grotius, Hugo, De Jure Belli Ac Pacis Libri Tres, trans. Francis Kelsey (Oxford: Clarendon, 1925)
Hall, Jerome, “Negligent Behavior Should be Excluded from Penal Liability,” 63 Columbia Law
Review 632 (1963)
Hanser, Matthew, “The Metaphysics of Harm,” 77 Philosophy and Phenomenological Research
432 (2008)
Hanser, Matthew, “Still More on the Metaphysics of Harm,” 82 Philosophy and Phenomeno-
logical Research 459 (2011)
Harlow, Carol, State Liability: Tort Law and Beyond (Oxford: Oxford University Press, 2004)
Harlow, Carol, “A Punitive Role for Tort Law?” in Linda Pearson, Carol Harlow and Michael
Taggart (eds.), Administrative Law in a Changing State: Essays in Honour of Mark Aronson
(Oxford: Hart, 2008)
Hart, H.L.A., The Concept of Law (Oxford: Clarendon, 1961)
Hart, H.L.A., “Negligence, Mens Rea and Criminal Responsibility,” in H.L.A. Hart, Punishment
and Responsibility: Essays in the Philosophy of Law (Oxford: Oxford University Press, 1968)
Hart, H.L.A., Punishment and Responsibility (2d ed.) (Oxford: Oxford University Press, 2008)
Hart, H.L.A. and Tony Honoré, Causation in the Law (2d ed.) (Oxford: Clarendon, 1985)
Harvey, Jean, “The Emerging Practice of Institutional Apologies,” 9 International Journal of
Applied Philosophy 2 (1995)
Hershovitz, Scott, “Two Models of Tort (and Takings),” 92 Virginia Law Review 1147 (2006)
Hershovitz, Scott, “Harry Potter and the Trouble with Tort Theory,” 63 Stanford Law Review 1
(2010)
Hershovitz, Scott, “Corrective Justice for Civil Recourse Theorists,” 39 Florida State University
Law Review 1 (2011)
Herstein, Ori, “Responsibility in Negligence: Why the Duty of Care is not a Duty ‘To Try’,” 23
Canadian Journal of Law and Jurisprudence 403 (2010)
Heuer, Ulrike and Gerlad Lang (eds.), Luck, Value, and Commitment: Themes from the Ethics of
Bernard Williams (Oxford: Oxford University Press, 2012)
Hieronymi, Pamela, “Articulating an Uncompromising Forgiveness,” 62(3) Philosophy and
Phenomenological Research 529 (2001)
Hitchcock, Christopher, “Counterfactual Availability and Causal Judgment,” in Christoph Hoerl,
Teresa McCormack, and Sarah Beck (eds.), Understanding Counterfactuals, Understanding
Causation: Issues in Philosophy and Psychology (Oxford: Oxford University Press, 2011)
Hobbes, Thomas, Leviathan, Richard Tuck (ed.), (Cambridge: Cambridge University Press, 1991)
Hoerl, Christoph, Teresa McCormack, and Sarah Beck (eds.), Understanding Counterfactuals,
Understanding Causation: Issues in Philosophy and Psychology (Oxford: Oxford University
Press, 2011)
Hohfeld, Wesley, “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning,” 23
Yale Law Journal 16 (1913)
Holmes, Oliver Wendell, “The Theory of Torts,” 7 American Law Review 652 (1873)
Holmes, Oliver Wendell, “Primitive Notions in Modern Law” (pts. 1 and 2), 10 American Law
Review 422 (1876); 11 American Law Review 641 (1877)
Holmes, Oliver Wendell, The Common Law (Boston: Little, Brown and Co., 1881)
Holmes, Oliver Wendell and Mark Howe (ed.), The Common Law (Cambridge: Belknap Press,
1963)
434 BIBLIOGRAPHY

Honoré, Tony, “Responsibility and Luck,” 104 Law Quarterly Review 530 (1988)
Honoré, Tony, Responsibility and Fault (Oxford: Hart, 1999)
Horder, Jeremy, “Criminal Culpability: The Possibility of a General Theory,” 12 Law and
Philosophy 193 (1993)
Horder, Jeremy (ed.), Oxford Essays in Jurisprudence (4th ed.) (Oxford: Oxford University
Press, 2000)
Huigens, Kyron, “The Dead End Of Deterrence, And Beyond,” 41 William and Mary Law
Review 943 (2000)
Hurd, Heidi, “What in the World is Wrong?” 5 Journal of Contemporary Legal Issues 157 (1994)
Hurd, Heidi, “The Deontology of Negligence,” 76 Boston University Law Review 249 (1996)
Hurd, Heidi, “The Moral Magic of Consent,” 2 Legal Theory 121 (1996)
Hurd, Heidi, “Why Liberals Should Hate ‘Hate Crime Legislation’,” 20 Law and Philosophy
215 (2001)
Hurd, Heidi and Michael Moore, “Negligence in the Air,” 3 Theoretical Inquiries in Law
2 (2002)
Hurd, Heidi and Michael Moore, “Punishing Hatred and Prejudice,” 56 Stanford Law Review
1081 (2004)
Hurd, Heidi and Michael Moore, “The Culpability of Negligence,” in Rowan Cruft, Matthew
H. Kramer, and Mark R. Reiff (eds.), Crime, Punishment, and Responsibility (Oxford: Oxford
University Press, 2011)
Hylton, Keith, “Punitive Damages and the Economic Theory of Penalties,” 87 Georgetown Law
Journal 421 (1998)
Jaffey, Peter, “Duties and Liabilities in Private Law,” 12 Legal Theory 137 (2006)
Jaffey, Peter, “Liabilities in Private Law,” 14 Legal Theory 233 (2008)
Jarecke, George and Nancy Plante, Seeking Civility: Common Courtesy and the Common Law
(Boston: Northeastern University Press, 2003)
Kagan, Shelly, The Limits of Morality (New York: Oxford University Press, 1989)
Kahan, Dan, “New Path of Deterrence,” 95 Michigan Law Review 2477 (1977)
Kahan, Dan, “What Do Alternative Sanctions Mean?” 63 University of Chicago Law Review
591 (1996)
Kahan, Dan, “Between Economics and Sociology: The New Path of Deterrence,” 95 Michigan
Law Review 2477 (1997)
Kaplow, Louis and Steven Shavell, Fairness versus Welfare (Cambridge, MA: Harvard Univer-
sity Press, 2002)
Keating, Gregory, “Reasonableness and Rationality in Negligence Theory,” 48 Stanford Law
Review 311 (1996)
Keating, Gregory, “Distributive and Corrective Justice in the Tort Law of Accidents,” 74
Southern California Law Review 193 (2000)
Keating, Gregory, “A Social Contract Conception of the Tort Law of Accidents,” in Gerald
Postema (ed.), Philosophy and the Law of Torts (New York: Cambridge University Press,
2001)
Keating, Gregory, “Rawlsian Fairness and Regime Choice in the Law of Accidents,” 72 Fordham
Law Review 1857 (2004)
Keating, Gregory, “The Priority of Respect Over Repair,” 18 Legal Theory 293 (2012)
Keeton, Robert, “Conditional Fault in the Law of Torts,” 72 Harvard Law Review 401 (1959)
 BIBLIOGRAPHY 435

Keeton, Robert, L. Sargentich, and G. Keating, Cases and Materials on Tort and Accident Law
(American Casebook Series) (St. Paul: West, 1998)
Kent, James, Commentaries on American Law (New York: Da Capo Press, 1971)
Keren-Paz, Tsachi, “Egalitarianism as Justification: Why and How Should Egalitarian Consid-
erations Reshape the Standard of Care in Negligence Law?” 4 Theoretical Inquiries in Law
275 (2003)
Keren-Paz, Taschi, Torts, Egalitarianism and Distributive Justice (Aldershot: Ashgate, 2007)
Kessler, Kimberly, “The Role of Luck in the Criminal Law,” 142 University of Pennsylvania Law
Review 2183 (1994)
King, J.H., “Causation, Valuation, and Chance in Personal Injury Torts Involving Preexisting
Conditions and Future Consequences,” 90 Yale Law Journal 1353 (1981)
King, Matt, “The Problem with Negligence,” 35 Social Theory and Practice 577 (2009)
Klepper, Howard, “Torts of Necessity: A Moral Theory of Compensation,” 9 Law and Philoso-
phy 223 (1990)
Klimchuk, Dennis, “Unjust Enrichment and Reparations for Slavery,” 84 Boston University Law
Review 1257 (2004)
Klimchuk, Dennis, “Property and Necessity,” in James Penner and Henry Smith (eds.),
Philosophical Foundations of Property Law (Oxford: Oxford University Press) (forthcoming)
Kornhauser, Lewis, “Wealth Maximization,” in Peter Newman (ed.), The New Palgrave Dic-
tionary of Economics and the Law (New York: Stockton, 1998)
Korsgaard, Christine, “The Reasons We Can Share: An Attack on the Distinction Between
Agent-Relative and Agent-Neutral Values,” in Christine Korsgard, Creating the Kingdom of
Ends (New York: Cambridge University Press, 1996)
Kraus, Jody, “Transparency and Determinacy in Common Law Adjudication: A Philosophical
Defense of Explanatory Economic Analysis,” 93 Virginia Law Review 287 (2007)
Kumar, Rahul, “Defending the Moral Moderate: Contractualism and Common Sense,”
28 Philosophy and Public Affairs 275 (1999)
Kumar, Rahul, Consensualism in Principle: On the Foundations of Non-Consequentialist Moral
Reasoning (New York: Routledge, 2001)
Kumar, Rahul and David Silver, “The Legacy of Injustice,” in Lukas Meyer (ed.), Justice in Time
(Baden-Baden: Nomos, 2004)
Kutz, Christopher, “Justice in Reparations: The Cost of Memory and the Value of Talk,”
32 Philosophy and Public Affairs 3 (2004)
Kymlicka, Will, Contemporary Political Philosophy: An Introduction (New York: Oxford
University Press, 1990)
Lamont, Julian and Christi Favor, “Distributive Justice,” in Edward N Zalta (ed.), The Stanford
Encyclopedia of Philosophy (Fall 2008 edition) (http://plato.stanford.edu/archives/fall2008/
entries/justice-distributive/)
Laycock, Douglas, “The Death of the Irreparable Injury Rule,” 103 Harvard Law Review 687
(1990)
Lazar, Seth, “Corrective Justice and the Possibility of Rectification,” 11 Ethical Theory and
Moral Practice 4 (2008)
Lazare, Aaron, On Apology (New York: Oxford University Press, 2005)
Lens, Jill, “Punishing for the Injury: Tort Law’s Influence in Defining the Constitutional
Limitations on Punitive Damage Awards,” 39 Hofstra Law Review 595 (2011)
Lewens, Tim (ed.), Risk: Philosophical Perspectives (New York: Routledge, 2007)
436 BIBLIOGRAPHY

Lewis, Sinclair, The Job (New York: Harper and Brothers, 1917)
Locke, John, An Essay Concerning Human Understanding, Peter Nidditch (ed.), (Oxford:
Clarendon, 1979)
Locke, John, Second Treatise of Civil Government (Buffalo: Prometheus Books, 1986)
Locke, John, “Two Treatises of Government,” Peter Laslett (ed.), Two Treatises of Government
(Cambridge: Cambridge University Press, 1988)
Lomasky, Loren, “Compensation and the Bounds of Rights,” 33 Nomos 13 (1991)
Lucy, William, Philosophy of Private Law (Oxford: Oxford University Press, 2007)
Lyons, David, “On Formal Justice,” 58 Cornell Law Review 833 (1972)
Mandel, D. and D. Hilton (eds.), The Psychology of Counterfactual Thinking (London: Routle-
dge, 2005)
Markel, Dan, “How Should Punitive Damages Work?” 157 University of Pennsylvania Law
Review 1383 (2009)
Markel, Dan, “Retributive Damages: A Theory of Punitive Damages as Intermediate Sanction,”
94 Cornell Law Review 239 (2009)
Marshall, Sandra and R.A. Duff, “Criminalization and Sharing Wrongs,” 11 Canadian Journal
of Law and Jurisprudence 7 (1998)
May, Larry, After War Ends: A Philosophical Perspective (New York: Cambridge University
Press, 2012)
McCarthy, David, “Liability and Risk,” 25 Philosophy and Public Affairs 238 (1996)
McCarthy, David, “Rights, Explanation, and Risks,” 107 Ethics 205 (1997)
McClurg, Andrew, “Dead Sorrow: A Story About Loss and a New Theory of Wrongful Death
Damages,” 85 Boston University Law Review 1 (2005)
McLean, Janet, Searching for the State in British Legal Thought: Competing Conceptions of the
Public Sphere (Cambridge: Cambridge University Press, 2012)
McMahan, Jeff, “Self-Defense and the Problem of the Innocent Attacker,” 104 Ethics 252 (1994)
McMahan, Jeff, The Ethics of Killing: Problems at the Margins of Life (Oxford: Oxford Univer-
sity Press, 2002)
McMahan, Jeff, “Causing People to Exist and Saving People’s Lives,” Journal of Ethics
(forthcoming)
Mendlow, Gabriel, “Is Tort Law a Form of Institutionalized Revenge,” 39 Florida State
University Law Review 129 (2011)
Menzies, Peter, “The Role of Counterfactual Dependence in Causal Judgements,” in Christoph
Hoerl, T. McCormack, and S. Beck (eds.), Understanding Counterfactuals, Understanding
Causation: Issues in Philosophy and Psychology (Oxford: Oxford University Press, 2011)
Merrill, Thomas and Henry Smith, Property, Principles, and Policies (2d ed.) (New York:
Foundation Press, 2012)
Merryman, John, The Civil Law Tradition: An Introduction to the Legal Systems of Europe and
Latin America (3d ed.) (Stanford: Stanford University Press, 2007)
Meurkens, Lotte and Emily Nordin (ed.), The Power of Punitive Damages: Is Europe Missing
Out? (Mortsel: Intersentia, 2012)
Meyer, Lukas (ed.), Justice in Time (Baden-Baden: Nomos, 2004)
Miller, William, Eye for an Eye (New York: Cambridge University Press, 2006)
Minow, Martha, Between Vengeance and Forgiveness: Facing History after Genocide and Mass
Violence (Boston: Beacon Press, 1998)
 BIBLIOGRAPHY 437

Moore, Michael, “The Moral Worth of Retribution,” in Ferdinand Schoeman (ed.), Responsi-
bility, Character, and the Moral Emotions: New Essays in Moral Psychology (New York:
Cambridge University Press, 1987)
Moore, Michael, Placing Blame: A General Theory of the Criminal Law (Oxford: Oxford
University Press, 1997)
Moore, Michael, Causation and Responsibility: An Essay in Law, Morals, and Metaphysics
(Oxford: Oxford University Press, 2009)
Moore, Michael, “Causation Revisited,” 42 Rutgers Law Journal 451 (2011)
Moore, Michael and Heidi Hurd, “Punishing the Awkward, the Stupid, the Weak, and the
Selfish: The Culpability of Negligence,” 5 Criminal Law and Philosophy 147 (2011)
Morse, S.J., “Reason, Results, and Criminal Responsibility,” 24 Illinois Law Review 363 (2004)
Murphy, Colleen, A Moral Theory of Political Reconciliation (New York: Cambridge University
Press, 2010)
Murphy, Jeffrie, “Forgiveness and resentment,” in Jeffrie G. Murphy and Jean Hampton (eds.),
Forgiveness and Mercy (Cambridge: Cambridge University Press, 1990)
Murphy, Jeffrie and Jean Hampton (eds.), Forgiveness and Mercy (Cambridge: Cambridge
University Press, 1990)
Murphy, John, “Misfeasance in a Public Office: A Tort Law Misfit?” 32 Oxford Journal of Legal
Studies 51 (2012)
Nagel, Thomas, “Libertarianism Without Foundations,” 85 Yale Law Journal 136 (1975)
Nagel, Thomas, “War and Massacre,” in Thomas Nagel, Mortal Questions (New York: Cam-
bridge University Press, 1979)
Nagel, Thomas, Equality and Partiality (New York: Oxford University Press, 1991)
Newman, Peter (ed.), The New Palgrave Dictionary of Economics and the Law (New York:
Stockton, 1998)
Nickel, James, “Justice in Compensation,” 18 William and Mary Law Review 379 (1976)
Nozick, Robert, Anarchy, State, and Utopia (New York: Basic Books, 1974)
Nozick, Robert, Philosophical Explanations (Cambridge: Belknap, 1981)
Oberdiek, John, “Lost in Moral Space: On the Infringing/Violating Distinction and its Place in
the Theory of Rights,” 23 Law and Philosophy 325 (2004)
Oberdiek, John, “Specifying Rights Out of Necessity,” 28 Oxford Journal of Legal Studies 1 (2008)
Oberdiek, John, “Towards a Right Against Risking” 12 Law and Philosophy 367 (2009)
Oberdiek, John, “Method and Morality in the New Private Law of Torts,” 125 Harvard Law
Review Forum 189 (2012)
Oberdiek, John, “The Moral Significance of Risking,” 18 Legal Theory 339 (2012)
Oman, Nathan, “Why There Is No Duty to Pay Damages: Powers, Duties, and Private Law,” 39
Florida State University Law Review 137 (2011)
Orend, Brian, “Jus Post Bellum,” 31 Journal of Social Philosophy 1 (2000)
Oshana, Marina, “Ascriptions of Responsibility,” 34 American Philosophical Quarterly 71 (1997)
Owen, David, “A Punitive Damages Overview: Functions, Problems and Reform,” 39 Villanova
Law Review 363 (1994)
Owen, David (ed.), Philosophical Foundations of Tort Law (New York: Oxford University Press,
1995)
Owen, David, “Response, Aggravating Punitive Damages,” 158 University of Pennsylvania Law
Review PENNumbra 181 (2010)
Parfit, Derek, Reasons and Persons (Oxford: Oxford University Press, 1984)
438 BIBLIOGRAPHY

Parfit, Derek, On What Matters (New York: Oxford University Press, 2011)
Parker, Jeffrey, “The Economics of Mens Rea,” 79 Virginia Law Review 741 (1993)
Pearce, Natasha, D. Cross, H. Monks, S. Waters, and S. Falconer, “Current Evidence of Best
Practice in Whole-School Bullying Intervention and its Potential to Inform Cyberbullying
Interventions,” 21 Australian Journal of Guidance and Counselling 1 (2011)
Pearson, Linda, Carol Harlow and Michael Taggart (eds.), Administrative Law in a Changing
State: Essays in Honour of Mark Aronson (Oxford: Hart, 2008)
Penner, James, The Idea of Property in Law (Oxford: Oxford University Press, 1997)
Penner, James and Henry Smith (ed.), Philosophical Foundations of Property Law (Oxford:
Oxford University Press) (forthcoming)
Perry, Stephen, “The Impossibility of General Strict Liability,” 1 Canadian Journal of Law and
Jurisprudence 147 (1988)
Perry, Stephen, “Comment on Coleman: Corrective Justice,” 67 Indiana Law Journal 381 (1992)
Perry, Stephen, “The Moral Foundations of Tort Law,” 77 Iowa Law Review 449 (1992)
Perry, Stephen, “Protected Interests and Undertakings in the Law of Negligence,” 42 University
of Toronto Law Review 247 (1992)
Perry, Stephen, “Risk, Harm, and Responsibility,” in David Owen (ed.), Philosophical Founda-
tions of Tort Law (Oxford: Oxford University Press, 1995)
Perry, Stephen, “Libertarianism, Entitlement, and Responsibility,” 26 Philosophy & Public
Affairs 351 (1997)
Perry, Stephen, “On the Relationship Between Distributive and Corrective Justice,” in Jeremy
Horder (ed.), Oxford Essays in Jurisprudence (4th ed.) (Oxford: Oxford University Press, 2000)
Perry, Stephen, “Responsibility for Outcomes, Risk, and the Law of Torts,” in Gerald Postema
(ed.), Philosophy and the Law of Torts (New York: Cambridge University Press, 2001)
Perry, Stephen, “Harm, History, and Counterfactuals,” 40 San Diego Law Review 1283 (2003)
Perry, Stephen, “Risk, Harm, Interests and Rights,” in T. Lewens (ed.), Risk: Philosophical
Perspectives (New York: Routledge, 2007)
Pillsbury, Samuel, “Crimes of Indifference,” 49 Rutgers Law Review 150 (1996)
Pogge, Thomas, “On the Site of Distributive Justice: Reflections on Cohen and Murphy,”
29 Philosophy and Public Affairs 137 (2000)
Polinsky, Mitchell and Steven Shavell, “Punitive Damages: An Economic Analysis,” 111 Harvard
Law Review 869 (1998)
Pollock, Frederick and James Avery Webb, A Treatise on the Law of Torts (3d ed.) (St. Louis:
F.H. Thomas Law Book Co., 1894)
Pomeroy, John, A Treatise on Equity Jurisprudence as Administered in the United States of
America (San Francisco: A.L. Bancroft and Co., 1883)
Posner, Richard, “A Theory of Negligence,” 1 Journal of Legal Studies 29 (1972)
Posner, Richard, “The Concept of Corrective Justice in Recent Theories of Tort Law,” 10 Journal
of Legal Studies 187 (1981)
Postema, Gerald, “Introduction: Search for an Explanatory Theory of Torts,” in Gerald Postema
(ed.), Philosophy and the Law of Torts (New York: Cambridge University Press, 2001)
Postema, Gerald (ed.), Philosophy and the Law of Torts (New York: Cambridge University
Press, 2001)
Priel, Dan, “That Can’t Be Rights,” 2 Jurisprudence 227 (2011)
Quinn, Warren, “Actions, Intentions, and Consequences: The Doctrine of Doing and Allow-
ing,” in Warren Quinn, Morality and Action (New York: Cambridge University Press, 1993)
 BIBLIOGRAPHY 439

Radin, Margaret, “Compensation and Commensurability,” 43 Duke Law Journal 56 (1993)

Radzik, Linda, Making Amends: Atonement in Morality, Law, and Politics (New York: Oxford
University Press, 2009)
Rawls, John, “Two Concepts of Rules,” 64 Philosophical Review 3 (1955), reprinted in H. B.
Acton (ed.), The Philosophy of Punishment: A Collection of Papers (London: Macmillan,
1969)
Rawls, John, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971)
Rawls, John, Political Liberalism (expanded edition) (New York: Columbia University Press,
2005)
Raz, Joseph, The Morality of Freedom (Oxford: Oxford University Press, 1986)
Raz, Joseph, “Autonomy, Toleration, and the Harm Principle,” in Ruth Gavison (ed.), Issues in
Contemporary Legal Philosophy (Oxford: Clarendon, 1987)
Raz, Joseph, “The Inner Logic of the Law,” in Joseph Raz, Ethics in the Public Domain (revised
edition) (Oxford: Oxford University Press, 1995)
Raz, Joseph, Value, Respect, and Attachment (Cambridge: Cambridge University Press, 2001)
Redish, Martin and Andrew Mathews, “Why Punitive Damages Are Unconstitutional,”
53 Emory Law Journal 1 (2004)
Reich, Charles, “The New Property,” 73 Yale Law Journal 733 (1964)
Rhee, Robert, “A Financial Economic Theory of Punitive Damages,” 111 Michigan Law Review
33 (2012)
Ripstein, Arthur, “Private Law and Private Narratives,” in Peter Cane and John Gardner (eds.),
Relating to Responsibility (Oxford: Hart, 1981)
Ripstein, Arthur, Equality, Responsibility, and the Law (Cambridge: Cambridge University
Press, 2001)
Ripstein, Arthur, “As if it Never Happened,” 48 William and Mary Law Review 1957 (2007)
Ripstein, Arthur, “Tort Law in a Liberal State,” 1 Journal of Tort Law 2 (2007)
Ripstein, Arthur, Force and Freedom: Kant’s Legal and Political Philosophy (Cambridge, MA:
Harvard University Press, 2009)
Ripstein, Arthur and Benjamin Zipursky, “Corrective Justice in an Age of Mass Torts,” in
Gerald J. Postema (ed.), Philosophy and the Law of Torts (New York: Cambridge University
Press, 2001)
Robinson, Paul and John Darley, “The Utility of Desert,” 91 Northwestern University Law
Review 453 (1997)
Rotberg, Robert and Dennis Thompson (eds.), Truth v. Justice: The Morality of Truth Com-
missions (Princeton, NJ: Princeton University Press, 2000)
Rousseau, Jean-Jacques, The Social Contract and the Discourses, trans. G. D. H. Cole (New York:
Dutton, 1979)
Sampsell-Jones, Ted, “The Myth of Ashby v White,” 8 University of St. Thomas Law Journal
40 (2010)
Scanlon, T.M., “Contractualism and Utilitarianism,” in Amartya Sen and Bernard Williams
(eds.), Utilitarianism and Beyond (Cambridge: Cambridge University Press, 1982)
Scanlon, T.M., What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998)
Scanlon, T.M., Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, MA: Harvard
University Press, 2010)
Schlueter, Linda and Kenneth R. Redden, Punitive Damages (4th ed.) (LEXIS, 2000)
440 BIBLIOGRAPHY

Schoeman, Ferdinand (ed.), Responsibility, Character, and the Moral Emotions: New Essays in
Moral Psychology (New York: Cambridge University Press, 1987)
Sebok, Anthony, “What Did Punitive Damages Do?” 78 Chicago-Kent Law Review 163 (2003)
Sebok, Anthony, “Punitive Damages: From Myth to Theory,” 92 Iowa Law Review 957 (2007)
Sebok, Anthony, “What is Wrong About Wrongdoing?” 39 Florida State University Law Review
209 (2011)
Sebok, Anthony, “The U.S. Supreme Court’s Theory of Common Law Punitive Damages: An
Inauspicious Start,” in Lotte Meurkens and Emily Nordin (ed.), The Power of Punitive
Damages: Is Europe Missing Out? (Mortsel: Intersentia, 2012)
Sen, Amartya and Bernard Williams (eds.), Utilitarianism and Beyond (Cambridge: Cambridge
University Press, 1982)
Sharkey, Catherine, “Punitive Damages as Societal Damages,” 113 Yale Law Journal 347 (2003)
Sharkey, Catherine, “The Exxon Valdez Litigation Marathon: A Window on Punitive Damages,”
7 University of St. Thomas Law Journal 25 (2010)
Shavell, Steven, “Criminal Law and the Optimal Use of Nonmonetary Sanctions as a Deterrent,”
85 Columbia Law Review 1232 (1985)
Sheinman, Hanoch, “Tort Law and Corrective Justice,” 22 Law and Philosophy 21 (2003)
Sher, George, “Transgenerational Compensation,” 33 Philosophy and Public Affairs 2 (2005)
Sherwin, Emily, “Compensation and Revenge,” 40 San Diego Law Review 1387 (2003)
Shiffrin, Seana, “Harm and its Moral Significance,” 18 Legal Theory 357 (2012)
Simester, Andrew and Andreas von Hirsch, Crimes, Harms, and Wrongs: On the Principles of
Criminalisation (Oxford: Hart, 2011)
Simons, Kenneth, “Assumption of Risk and Consent in the Law of Torts: A Theory of Full
Preference,” 67 Boston University Law Review 213 (1987)
Simons, Kenneth, “Reflections on Assumption of Risk,” 50 University of California Los Angeles
Law Review 481 (2002)
Simons, Kenneth, “Book Review: The Conceptual Structure of Consent in Criminal Law,”
9 Buffalo Criminal Law Review 577 (2006)
Slote, Michael, Goods and Virtues (Oxford: Clarendon, 1983)
Smiley, Marion, “Collective Responsibility,” in Edward N. Zalta (ed.), The Stanford
Encyclopedia of Philosophy (Fall 2010 edition) (http://plato.stanford.edu/entries/collective-
responsibility/)
Smith, Nick, I Was Wrong: The Meanings of Apologies (New York: Cambridge University Press,
2008)
Smith, Nick, “Against Court-Ordered Apologies,” 16 New Criminal Law Review 1 (2013)
Smith, Stephen, “Duties, Liabilities, and Damages,” 125 Harvard Law Review 1727 (2012)
Sobel, David and Steven Wall (eds.), Reasons for Action (Cambridge: Cambridge University
Press, 2009)
Solomon, Jason, “Civil Recourse as Social Equality,” 39 Florida State University Law Review
243 (2011)
Stapleton, Jane, “The Gist of Negligence, Part 2,” 104 Law Quarterly Review 389 (1988)
Steele, Jenny, Risks and Legal Theory (Oxford: Hart, 2004)
Stevens, Robert, Torts and Rights (Oxford: Oxford University Press, 2007)
Stevens, Robert, “Law of Torts,” in L. Blom-Cooper, G. Drewry, and B. Dickinson (eds.), The
Judicial House of Lords 1876–2009 (Oxford: Oxford University Press, 2009)
 BIBLIOGRAPHY 441

Stith, Kate, “The Arc of the Pendulum: Judges, Prosecutors, and the Exercise of Discretion,”
117 Yale Law Journal 1420 (2008)
Stone, Martin, “The Significance of Doing and Suffering,” in Gerald Postema (ed.), Philosophy
and the Law of Torts (Cambridge: Cambridge University Press, 2001)
Strawson, Peter, “Freedom and Resentment,” 48 Proceedings of the British Academy 1 (1962)
Sunstein, Cass et al., “Predictably Incoherent Judgments,” 54 Stanford Law Review 1153 (2002)
Sunstein, Cass et al., Punitive Damages: How Juries Decide (Chicago: University of Chicago
Press, 2002)
Tadros, Victor, Criminal Responsibility (Oxford: Oxford University Press, 2005)
Tadros, Victor, “Consent to Harm,” 64 Current Legal Problems 23 (2011)
Tadros, Victor, The Ends of Harm (Oxford: Oxford University Press, 2011)
Tadros, Victor, “Harm, Sovereignty, and Prohibition,” 17 Legal Theory 35 (2011)
Thomas, Tracy, “Proportionality and the Supreme Court’s Jurisprudence of Remedies,”
59 Hastings Law Journal 73 (2007)
Thompson, Janna, Responsibility for the Past: Reparation and Historical Injustice (Cambridge:
Polity Press, 2002)
Thomson, Judith Jarvis, “More on the Metaphysics of Harm,” 82 Philosophy and Phenomeno-
logical Research 436 (2011)
Judith Jarvis Thomson, The Realm of Rights (Cambridge, MA: Harvard University Press, 1990)
Tyler, Ransom, A Treatise on the Remedy by Ejectment and the Law of Adverse Enjoyment in the
United States: Embracing in Full the Statutory Policy of the Several States, in Respect to the
Action for the Recovery of Real Property (Albany: William Gould and Son, 1876)
Vargas, Manuel, “The Trouble with Tracing,” 25 Midwest Studies in Philosophy 269 (2005)
Velleman, James, “Persons in Prospect,” 36 Philosophy and Public Affairs 221 (2008)
Waldron, Jeremy, The Right to Private Property (Oxford: Oxford University Press, 1988)
Waldron, Jeremy, “Moments of Carelessness and Massive Loss,” in David Owen (ed.), Philo-
sophical Foundations of Tort Law (Oxford: Oxford University Press, 1995)
Waldron, Jeremy, “Where Money and Markets Don’t Belong,” in New York Review of Books
(August 16, 2012)
Walker, Margaret, Moral Repair: Reconstructing Moral Relations after Wrongdoing (New York:
Cambridge University Press, 2007)
Walker, Margaret, “Truth Telling as Reparations,” 41 Metaphilosophy 4 (2010)
Watson, Gary, “Promises, Reasons, and Normative Powers,” in David Sobel and Steven Wall
(eds.), Reasons for Action (Cambridge: Cambridge University Press, 2009)
Weinrib, Ernest, “Understanding Tort Law,” 23 Valparaiso University Law Review 485 (1989)
Weinrib, Ernest, The Idea of Private Law (Cambridge, MA: Harvard University Press, 1995)
Weinrib, Ernest, “Punishment and Disgorgement as Contract Remedies,” 78 Chicago-Kent Law
Review 55 (2003)
Weinrib, Ernest, “Private Law and Public Right,” 61 University of Toronto Law Journal
191 (2011)
Weinrib, Ernest, Corrective Justice (Oxford: Oxford University Press, 2012)
Weinrib, Laura and Ernest Weinrib, “Constitutional Values and Private Law in Canada,” in
Daniel Friedmann and Daphne Barak-Erez (eds.), Human Rights in Private Law (Oxford:
Hart, 2001)
Wenar, Leif, “Reparations for the Future,” 37 Journal of Social Philosophy 3 (2006)
Wertheimer, Alan, Consent to Sexual Relations (Cambridge: Cambridge University Press, 2004)
442 BIBLIOGRAPHY

Westen, Peter, The Logic of Consent (Burlington, VT: Ashgate, 2004)

Westen, Peter, “Individualizing the Reasonable Person in Criminal Law,” 2 Criminal Law and
Philosophy 137 (2008)
White, Edward, “The Unexpected Persistence of Negligence, 1980 (2000),” 54 Vanderbilt Law
Review 1337 (2001)
White, G. Edward, Tort Law in America: An Intellectual History (New York: Oxford University
Press, 1980)
White, Mark D. (ed.), Theoretical Foundations of Law and Economics (New York: Cambridge
University Press, 2009)
Wilberg, Hanna, “In Defence of the Omissions Rule in Public Authority Negligence Claims,”
19 Torts Law Journal 159 (2011)
Williams, Bernard, “Moral Luck,” in Bernard Williams, Moral Luck (Cambridge: Cambridge
University Press, 1981)
Williams, Glanville, Criminal Law: The General Part (2d ed.) (London: Stevens and Sons, 1961)
Winfield, Percy, The Province of the Law of Tort (New York: MacMillan, 1931)
Wolf, Susan, “The Moral of Moral Luck,” 31 Philosophical Exchange 4 (2000), reprinted in
Cheshire Calhoun (ed.), Setting the Moral Compass: Essays by Women Philosophers (Oxford:
Oxford University Press, 2003)
Woo, Junda, “Blowing Smoke Around Others May Be Battery,” The Wall Street Journal (April
11, 1994)
Woodward, James, Making Things Happen (Oxford: Oxford University Press, 2003)
Wright, Richard, “Causation in Tort Law,” 73 California Law Review 1775 (1985)
Wright, Richard, “Substantive Corrective Justice,” 77 Iowa Law Review 625 (1992)
Wright, Richard, “Justice and Reasonable Care in Negligence Law,” 47 American Journal of
Jurisprudence 143 (2002)
Wright, Richard, “Negligence in the Courts: Introduction and Commentary,” 77 Chicago-Kent
Law Review 425 (2002)
Wright, Richard, “Hand, Posner, and the Myth of the ‘Hand Formula’,” 4 Theoretical Inquiries
in Law 145 (2003)
Zeelenberg, Marcel and Eric van Dijk, “On the Comparative Nature of Regret,” in D. Mandel
and D. Hilton (eds.), The Psychology of Counterfactual Thinking (London: Routledge, 2005)
Zehr, Howard, Changing Lenses: A New Focus for Crime and Justice (Scottsdale: Herald, 1990)
Zimmerman, Michael, “Taking Luck Seriously,” 99 Journal of Philosophy 553 (2002)
Zipursky, Benjamin, “Rights, Wrongs, and Recourse in the Law of Torts,” 51 Vanderbilt Law
Review 1 (1998)
Zipursky, Benjamin, “Pragmatic Conceptualism,” 6 Legal Theory 457 (2000)
Zipursky, Benjamin, “Civil Recourse, Not Corrective Justice,” 91 Georgetown Law Journal
695 (2003)
Zipursky, Benjamin, “A Theory of Punitive Damages,” 84 Texas Law Review 105 (2005)
Zipursky, Benjamin, “Substantive Standing, Civil Recourse, and Corrective Justice,” 39 Florida
State University Law Review 299 (2011)
Zipursky, Benjamin, “Palsgraf, Punitive Damages, and Preemption,” 125 Harvard Law Review
1757 (2012)

abnormally dangerous
activities 51, 53, 72n, 76,
77, 83, 84, 294, 295n, 296,
300, 306
accidents 22, 23, 35, 69, 70, 72,
73, 74, 76, 83, 103n, 106,
107, 109, 111, 133, 136, 137,
139, 141–6, 225, 226, 235,
254, 285, 292, 293, 295, 306,
307, 308, 309, 316, 317, 387,
390, 391, 394, 395, 399, 400,
401, 402, 405, 408, 414
accountability 7, 17, 28, 29, 129,
143, 164, 165, 166, 167, 168,
225, 263, 265
adverse possession 125, 131,
132, 138
African-American slavery (see
chattel slavery)
Agency 2, 9, 20, 33, 52n, 107,
212, 221, 238, 254, 256, 257,
258, 261, 265, 266, 267, 268,
296, 310, 370, 398
Alexander, Larry 3, 12, 134n,
277, 396, 397, 406–425
amends 31, 196, 225, 231, 239,
241, 243, 245, 246, 247, 248
apology 94, 96n, 199, 200, 203,
210, 211, 218, 220, 225, 226,
227, 231, 234, 238, 240, 241,
242, 246, 248, 253, 260, 264,
265, 266, 269, 270, 271
Aristotle 89, 233, 307, 308, 349,
354, 355, 356, 359–65, 371
assumption of risk 10, 23, 223,
272–91, 310
authority 27, 116, 117, 149, 151,
152, 153, 156, 161, 205, 206,
207, 209, 216, 227, 241, 244,
330, 331, 340, 401
autonomy 10, 54, 55, 57, 58, 59,
66, 67, 69, 71, 72, 73, 74, 77,
84, 85, 105, 124, 129, 132,
136, 139, 145, 161, 167, 176,
181, 183, 278, 283, 293n,
295, 296, 297, 298, 299, 300,
305, 308, 309, 311
INDEX
bilateral structure 2, 6, 7, 40, 83,
104, 105, 106, 108–12, 115,
117, 119, 120, 155, 156n,
159, 161, 165, 349, 362, 368,
372, 379n, 381, 382
blameworthiness 11, 12, 19n, 21,
22, 28, 30, 34, 35, 52, 54n,
72, 125, 133, 135, 199, 213,
214, 217, 218, 224, 225, 233,
240, 251, 262, 263n, 264,
269, 283n, 284, 287, 299,
314, 320, 325, 326n, 327,
361, 387–405, 406–25
Calabresi, Guido 39, 43n, 83n,
143n, 159, 164, 165, 167,
292n, 295n, 309, 310n, 339,
344, 345
Cane, Peter 3, 7, 18, 122n, 134n,
148–168, 338, 349, 354n,
358, 377n, 379
capacities 9, 19n, 60, 64, 159,
177, 205, 224, 302, 310, 350,
391, 395–8, 408, 410
carelessness (see also negligence;
fault) 11, 22, 23, 24, 43, 235,
380, 389, 403
causation 2, 19n, 281, 416,
418, 425
actual 12, 21, 22, 32, 34, 36,
56, 187, 190, 191, 216, 251,
402, 418, 419–23
proximate 4, 138, 418–19,
424–5
character 248, 365, 391, 398,
399, 400, 401, 402, 403,
404, 405
chattel slavery 193, 194, 195,
196, 197, 199, 204–09,
210, 211
chattels 125, 132, 135–7, 276,
297, 299
choice 26, 64, 67, 69, 70, 72, 73n,
83n, 85, 124, 144, 155, 160,
195n, 274, 277, 282n, 288,
289, 307, 309, 320, 322n,
327, 330, 336, 352, 388, 389,
390, 391, 394, 395, 396, 298,
399, 400, 405, 410
civil law 134, 148n, 219, 354,
356, 359, 362, 363, 365, 366,
371, 372–3, 377, 378, 383
civil recourse 3, 4, 6, 8, 9, 18,
26–35, 36, 106, 116, 119,
122n, 212, 218–30, 247, 249,
300, 304, 308, 313, 328n, 330
civil wrongs 129, 165, 228
Claeys, Eric 3, 6, 122–47
Coleman, Jules 1, 2, 3, 9, 11, 18,
20, 25, 38, 40n, 44, 80, 89n,
104, 105, 122n, 123n, 124n,
129n, 131n, 142n, 143n,
242n, 250n, 251n, 252n,
293–4, 300, 302n, 304, 305,
356n
common law 11, 34, 37, 39, 65,
72, 78, 79, 140, 145, 148n,
150, 152, 156, 163, 165, 168,
294, 295, 306, 312, 313, 314,
315, 316, 317, 319n, 320,
321, 323, 324, 325, 326, 327,
328, 329, 330, 331, 342,
343, 346
comparative fault 10, 22–5, 275,
283, 287n, 290
compensation (see also
damages) 4, 5, 6, 7, 8, 10, 17,
25, 26, 28, 29, 30, 31, 32, 35,
36, 65–85, 88, 96, 97, 104,
108, 149, 155, 159, 160, 163,
164, 165, 166, 167, 171–92,
193–211, 212–30, 231–49,
251n, 253, 302, 303, 306,
315, 316, 317, 318, 319, 320,
321, 322, 324, 327, 328, 329,
371, 406, 407, 408, 417
consequentialism 1, 38, 71n, 103,
104, 105, 112, 113, 114n,
127, 128, 129, 139, 140, 149,
201, 320, 321, 322, 323, 324,
389, 390, 401
contract 27, 29, 30, 74, 76, 77,
129, 130, 138, 140, 152, 159,
164, 167, 280, 324, 325, 327,
444
342, 343, 345, 371, 372, 373,
381, 406
contractualism 2, 6, 103–21
contributory fault (see
contributory negligence)
contributory negligence 23, 144,
213, 272, 275, 349, 376
corrective justice (see also
rights-based accounts of
tort law) 1, 3, 9, 11, 38, 41,
67, 102, 119, 122, 133, 138,
155, 158, 231, 233, 238, 250,
251, 252, 300, 387, 388, 390,
392, 397, 401, 407
and civil recourse 4, 222
and compensation 79–84,
193, 232, 235
and contractualism 6, 112
and distributive justice 11, 84,
156, 157, 335–46, 347, 348,
349, 350, 352, 353, 354–84
and economic accounts of tort
law 5, 17, 104, 105, 116,
251, 327
and property 7, 135
and punishment 313, 321
and reconciliation 232, 234,
235, 237, 239–48, 249
and responsibility 17, 18,
25–6, 270
and retributive justice 416–18
and revenge 6, 89–100
and strict liability 133, 304–08
correlativity (see bilateral
structure)
cost allocation 216, 221, 222,
223, 225, 229–30
counterfactual
account of harm (see harm,
counterfactual account of )
dependence (see causation,
actual)
criminal process (see criminal
law, process of )
criminal law 8, 9, 50, 54n, 60, 99,
100, 124, 125, 134, 135, 137,
159, 164, 165, 166, 190n,
212, 215, 216, 219, 223, 224,
225, 228, 229–30, 232, 248,
270, 319, 320, 322, 323, 324,
325, 330, 383n, 389n, 394,
401, 407, 409
process of 8, 99, 100, 165, 213n,
214, 215, 217, 218, 219, 221,
229, 230
INDEX
and tort law 49, 50n, 55, 99–100,
167, 280, 327, 392, 411
culpability (see blameworthiness)
damages (see also
compensation; harm)
compensatory 6, 10, 25, 26,
29, 30, 35, 65–85, 87, 88, 91,
93, 97, 99, 101, 104, 156,
158, 159, 160, 174, 197, 213,
214, 217, 218, 220, 221, 222,
223, 226, 227, 228, 247, 274,
275, 280, 287n, 296, 335,
347, 348, 349, 351, 353, 401,
402, 403, 417, 419
nominal 28, 138
punitive 11, 28, 33, 78, 86, 87,
88, 96, 97, 99, 133, 134, 138,
167, 215n, 227, 228, 247,
312–31
defences 6, 10, 36, 125, 130–2,
134, 137, 144, 146, 214, 215,
217, 223, 224, 225, 272–91,
308–10, 342, 376–7
deontology (see non-
consequentialism)
deterrence 5, 11, 17, 36, 37, 41,
42, 66, 67, 88, 89, 149, 155,
164, 165, 250, 312–31
distributive justice 2, 3, 5, 11,
106, 108–12, 118, 119, 120,
121, 122n, 211, 253, 308,
335–53, 354–84
and corrective justice 11, 84,
155, 156, 157, 195, 335–53,
354–84
localized 21, 23, 34, 111, 346,
348, 349, 353, 379n
Duff, R. A. 3, 8, 9, 212–230,
231n, 252n, 263n
duty of care 4, 5, 12, 25, 27, 28,
35, 36, 38, 39, 40–64, 66, 78,
80, 81, 82, 115, 116, 118,
119, 143–6, 153, 157, 159,
284, 287n, 289, 290, 293,
300, 305, 363, 407–16
duty of repair (see repair)
Dworkin, Ronald 5, 36, 66, 67,
68, 69, 70, 324n, 338n
economic account of tort law (see
law and economics)
economic loss 42, 62, 75n, 129
egalitarianism 5, 66, 67–70, 71,
84, 85
Enoch, David 3, 9, 250–71
Epstein, Richard 18, 50, 125n,
131n, 135n, 136n, 141n,
294n, 408
equality (see also egalitarianism)
7, 37, 79, 83, 89, 95, 124,
148–68, 354, 360, 361, 362,
363, 364, 366, 370, 374, 375,
378, 383
fairness 21, 29, 32, 38, 104, 107,
108, 111, 112, 119, 124, 164,
166, 197, 207, 217, 233, 244,
278, 289, 296, 303, 305, 306,
307, 308, 316, 330, 380,
394, 417
fault (see also carelessness;
negligence) 10, 11, 12, 17,
20, 21, 22, 23, 27, 29, 30, 32,
34, 36, 51, 52, 70, 71, 72, 83,
103n, 132, 133, 134, 135,
139, 187, 213, 216, 217, 227,
228, 233, 235, 247, 254, 260,
274, 275, 283, 284, 285,
287n, 290, 292–311, 321,
326, 376, 387–405, 408–16,
417, 418, 424, 425
Ferzan, Kimberly Kessler 3, 12,
134n, 396, 397, 406–25
Fletcher, George 1, 2, 3, 6, 11,
104, 107, 108, 109, 121
foreseeability 4, 5, 21, 32, 34, 36,
42n, 43, 51, 60–4, 418, 419
freedom 10, 31, 37, 54, 66, 68, 69,
71, 72, 73, 75, 76, 77, 84, 106,
107, 108, 111, 123, 129, 130,
131, 132, 142, 143, 144, 148,
149, 151, 153, 159, 161, 163,
248, 274, 283, 293n, 298n,
314, 401
Fried, Barbara 56, 80
Gardner, John 3, 11, 36, 49, 50n,
52n, 66n, 81, 82, 89, 90, 92n,
100n, 105n, 123n, 148n,
190n, 224n, 251n, 298n,
302n, 335–53, 356n, 358,
360, 367n, 375, 379, 381n
Geistfeld, Mark 3, 5, 6, 8, 65–84,
105n, 110n, 318n
Goldberg, John 3, 4, 5, 8, 9,
17–37, 39, 40n, 43n, 60–2,
63n, 65n, 66n, 79n, 91n,
115, 116, 119, 122n, 129n,
135n, 150n, 221, 223, 224,

225n, 226, 227n, 228, 229,
247n, 250n, 294, 295, 300,
304n, 324n, 330n, 340n, 356n
Hand formula/test 41, 293, 317,
324, 389, 409
harm (see also damages) 4, 5, 7,
8, 9, 10, 12, 29, 31, 32, 33,
35, 40, 44–60, 61, 62, 64, 65,
66, 69–83, 88, 103, 104, 105,
107, 108, 109, 115, 116, 118,
124, 132, 135–47, 153–9,
165–7, 171–92, 193, 194,
197, 199, 201, 202, 204, 210,
212–30, 232–48, 253, 254,
271, 274, 279, 280–91,
292–311, 316, 321, 326, 347,
350, 358, 363, 373–74, 383,
387–405, 407–25
counterfactual account of 7, 8,
171–92, 201, 202n,
non-comparative account
of 172, 175, 176, 177–81
pre-emption of 172, 181, 184,
185, 186–88, 189
temporal account of 172, 176,
181–5
Hart, H.L.A. 1, 2, 18, 19n, 36, 322,
325, 343, 395, 396, 397, 398n
Hershovitz, Scott 3, 5, 6, 26,
86–102, 122n, 129n, 233n,
246, 247n, 248n
historical injustice 197, 211, 231
Holmes, Oliver Wendell 39, 42,
64, 72n, 303n, 325, 326, 388,
390, 392, 393, 398n, 399n
Honoré, Tony 1, 2, 3, 11, 18, 19n,
20, 34, 36, 350, 351n, 353
Hurd, Heidi 3, 11, 12, 45n, 48n,
50n, 53, 54n, 55n, 61n, 277n,
278n, 387–405, 410, 411
inequality (see equality)
initial entitlements 344, 345
instrumental accounts of tort
law 2, 4, 7, 22, 24, 37, 38,
39, 40, 41, 42, 43n, 52, 60,
61n, 64, 105, 110, 148, 149,
150, 151, 153–68, 250,
251n, 268
insurance 8, 10, 22, 75, 84, 215,
217, 218, 226, 229, 246, 271,
348, 355, 381, 382, 383, 384,
407, 417, 421
INDEX
intentional tort 23, 35, 53, 57,
59, 132, 133, 134, 135, 137,
139, 140, 190, 203, 225,
272–91, 295, 296, 298, 299,
300, 311, 325, 326, 328, 339,
397, 415, 417
interests (see also public
interest) 34, 37, 41, 43, 44,
46, 54–9, 63, 70–7, 83, 84,
104, 107, 108, 113, 116,
123–30, 132, 135, 136, 137,
139, 140, 141, 142, 145, 149,
151, 155, 162, 163, 166, 173,
194, 203, 206, 210, 219, 235,
280, 283, 287, 290, 301, 336,
339, 406, 407, 411, 416
interpersonal justification 6,
103–06, 113–21
juridical inequality 148–68
justice (see corrective justice;
distributive justice;
reparative justice;
retributive justice)
justification (see interpersonal
justification)
Keating, Gregory C. 3, 6, 10,
61n, 63n, 104, 106–08, 109,
110, 111, 112, 119n, 121,
122n, 142n, 292–311,
356n, 391n, 407
Kant, Immanuel 18, 106, 107,
114n, 148n, 149, 236
Kumar, Rahul 3, 8, 114n, 119n,
193–211
labor 125–8, 129, 130, 131, 132,
139, 140, 143, 144, 145,
146, 197
law and economics (see also
corrective justice; and
economic accounts of tort
law) 1, 2, 5, 6, 80, 103,
104, 105, 106, 110, 111, 112,
116, 117, 121, 231, 292, 293,
321, 323, 325, 327, 390
legitimacy 7, 11, 150, 151, 166,
206, 323, 325, 330
liability (see tort liability)
liberty (see freedom)
Locke, John 18, 27,
125–8, 247
loss (see harm)
445
malice 86, 225, 235, 318
Melamed, A. Douglas 83n, 143n,
339, 344, 345
Misfeasance 62, 150, 154, 163
moral responsibility (see
responsibility)
moral theory 3, 18, 39, 49, 50,
112, 117, 143, 149, 236,
313, 321
morality 27, 30, 31, 35, 37, 44,
45, 49, 50, 54, 58, 59, 64,
103, 113n, 114, 117, 125,
126, 127, 128, 130, 140, 143,
145, 146, 149n, 155n, 176,
181, 241, 245–8, 254, 292n,
300, 305–08, 309, 338, 339,
344, 346, 347, 354, 362, 393,
399, 400
Nagel, Thomas 67n, 109n, 114,
115, 119
necessity 125, 126, 127, 128, 131,
132, 139, 140, 141, 214,
283n, 296, 301, 303, 304
negligence 4, 5, 6, 7, 10, 11, 12,
17, 21, 24, 28, 30, 34, 35,
38–64, 66, 70, 72, 73, 74, 75,
76, 77, 78, 79, 82, 84, 107,
108, 116, 123, 125, 131, 132,
134, 135, 141–6, 153, 154,
214, 216, 220, 221, 226, 227,
228, 235, 242, 246, 247, 248,
250, 251, 252, 253, 266, 270,
271, 273–91, 292–311, 324,
325, 326, 328, 342, 363,
387–400, 409, 411, 412, 413,
414, 415, 418, 421, 424, 425
non-consequentialism 1, 2, 37,
38, 54, 55, 106, 112, 320,
322, 329, 365, 377, 393
non-identity problem 179, 197,
198, 199, 204
non-instrumentalism (see
instrumentalism)
Nozick, Robert 89n, 95, 100n,
109n, 111, 173, 174,
175, 341n
Nuisance 125, 141, 142, 295,
296, 300, 303, 311
Oberdiek, John 1–16, 51n, 56,
58, 59, 103–21, 224n
outcome responsibility 3, 20, 21,
23, 29, 30, 32, 33, 35, 72
446
Owen, David G. 1, 3, 12, 318, 320n
Ownership 131, 133, 136
Perry, Stephen 3, 4, 5, 18, 19–22,
23, 25, 29–35, 36, 38–64,
72n, 109n, 110n, 122n,
133n, 173n, 182, 183, 184,
346n, 347n, 348, 356n,
358, 373n
Posner, Richard 2, 103n, 165,
317, 324, 348n, 390
pre-emption (see harm, pre-
emption of )
private law 2, 106, 111, 115, 116,
119, 120, 148n, 149, 150,
153, 154, 155, 159, 161, 166,
167, 313, 318, 319n, 321,
322, 324, 329, 331, 336,
337, 353
probability (see also risk) 24, 46,
48, 56, 279, 282, 284–7, 323,
352, 353, 389, 409, 410, 420
products liability 22–5, 34, 75,
77n, 294, 295, 297, 300, 304,
305, 306
property 6, 7, 10, 24, 26, 43, 54,
57, 61, 68, 71, 72n, 74n, 78,
122–47, 166, 193, 205, 212,
223, 233, 240, 254, 287, 293,
295, 296, 297, 298, 299, 301,
302, 303, 304, 309, 311, 345,
406, 408, 420
public interest 149, 151, 152,
154, 157, 160, 162, 163,
165, 166
public policy 4, 42, 44, 111, 116,
120, 123, 139, 144, 159, 207,
209, 214n, 323n, 418
public reason 206, 207, 208, 209,
210, 211, 330
punishment (see also damages,
punitive) 6, 11, 87, 94, 95,
99, 100, 165, 166, 167, 191,
214, 215, 217, 218, 227, 228,
229, 230, 233, 234, 235, 236,
269, 317–31, 361, 401
punitive damages (see damages,
punitive)
Radzik, Linda 3, 9, 231–49
Rawls, John 2, 3, 6, 103, 106,
107, 108–13, 121, 143n, 145,
146, 206, 321n, 346
Raz, Joseph 55n, 57, 58, 176,
203n, 266, 340n, 342n
INDEX
reconciliation 8, 9, 198n, 231–49
remedies 27, 28, 30, 33, 36, 39, 56,
70, 77, 78, 79, 80, 81, 82, 83n,
85, 99, 105, 112, 116, 124,
142, 143, 156n, 159, 163, 164,
165, 166, 216, 218, 221, 222,
224, 225–30, 236, 247, 275,
280, 281, 296, 320, 322, 324,
325, 345, 358, 372, 373
repair 8, 20, 29, 30, 31, 32, 80,
104, 193–211, 244, 248, 306,
340, 375, 376, 379
reparations (see repair)
reparative justice 195, 371–7
responsibility (see also
contributory negligence;
outcome responsibility; and
tort liability) 17–37, 38, 69,
72, 97, 116, 126, 127, 144,
153, 163, 165, 167, 195, 196,
198, 199, 200, 204, 210, 213,
216, 218, 223, 224n, 226,
238, 240, 246, 250–71, 275,
293, 294, 299, 301, 306, 325,
347, 348, 349, 350, 356, 370,
390, 395n, 400n, 411, 418,
419, 422, 424, 425
restitution (see also
compensation; damages;
repair) 71, 94, 99, 100, 203,
302, 314, 319, 320, 322, 370,
372, 373
retributive justice (see also
retributivism) 99, 100, 319n,
416–18
retributivism 215, 234, 235, 313,
320, 321
revenge (see also punishment;
damages, punitive) 5, 6,
86–102
rights 4, 5, 6, 7, 10, 11, 26n, 27,
28, 29–32, 34, 35, 37, 38–64,
66, 67, 69, 70–85, 111, 115,
116, 117, 118, 122–47,
148–68, 207, 219, 224n,
228, 246, 247, 275, 280, 283,
290, 292–311, 314, 317, 322,
330, 338–45, 351, 352, 353,
372, 377, 379, 389,
390, 404
rights-based accounts of tort
law (see also corrective
justice) 38–41, 44, 50, 53,
60, 61n, 62, 63, 71, 132,
135–48, 295, 297–311, 389
Ripstein, Arthur 18, 25, 40n, 50,
51, 91n, 148n, 150n, 154n,
250n, 252n, 320n, 350n
risk (see also assumption of risk;
foreseeability) 2, 4, 5, 10, 12,
35, 39, 40, 41, 44–64, 69, 70,
71, 72, 73–7, 79, 80, 81, 83,
84, 89, 104, 107, 108, 109,
136, 142, 143, 173n, 212,
213, 214, 217, 218, 244, 251,
253, 270, 271, 296, 302n,
307, 308, 309, 310, 316, 317,
335, 350–4, 387–98, 405,
406–25
Scanlon, T. M. 6, 103n, 106, 112,
113–14, 115, 116 117n, 118n,
119, 121, 149n, 176, 187n
Sebok, Anthony 3, 11, 26n, 88n,
312–31
Sheinman, Hanoch 3, 11, 105n,
354–84
Simons, Kenneth 3, 9, 10, 272–91
strict liability 4, 5, 10, 11, 12, 20,
33, 34, 36, 51, 52, 53, 69, 70,
72, 73n, 74, 76, 77, 81, 82,
83, 84, 108, 133, 141, 142,
282n, 292–311, 325, 326,
328, 387, 388, 391, 392, 394,
402, 405, 408, 409, 411, 412,
413, 414, 415, 417, 418, 425
Tadros, Victor 3, 7, 8, 171–92
Thompson, Judith Jarvis 50,
172n, 177, 178n, 179n,
181n, 183n
trespass 32, 33, 34, 35, 86, 92,
122–47, 224n, 251, 276, 297,
298, 299, 300, 302, 309, 310,
311, 326, 342, 343, 345
ultrahazardous activities (see
abnormally dangerous
activities)
unjust enrichment 197, 371,
372, 389n
utilitarianism (see
consequentialism)
vicarious liability 149n, 254, 304,
306, 307, 311
Waldron, Jeremy 127n, 220n,
254n, 269n, 306n, 317n,
356n, 380

wealth maximization (see law
and economics)
Weinrib, Ernest 1, 2, 3, 10, 11,
18, 20, 38, 40n, 52, 82n, 83,
84, 89n, 92n, 104n, 115,
122n, 133n, 148n, 149n,
150n, 153, 154, 155, 156,
157, 158, 161, 162, 163,
293–94n, 304n, 306n,
INDEX 447
320n, 337, 338, 339, 79n, 80n, 91n, 96n,
344, 346, 347, 348, 353, 105n, 106, 115, 116,
356n, 357 119, 120n, 122n, 124n,
Wright, Richard 71n, 154n, 129n, 135n, 150n, 221,
389n, 419n 223n, 224n, 225n, 226n,
227n, 228n, 247n, 294,
Zipursky, Benjamin 3, 4, 5, 8, 295, 300, 304n, 313n,
9, 17–37, 39, 40n, 43n, 326n, 328n, 330n,
50, 51, 60, 61, 62, 63n, 340n, 356n