CONTEMPL AT ING ART

This page intentionally left blank

Contemplating
Art
Essays in Aesthetics

J ER ROLD LEVIN SON

CLARENDON PRESS OXFORD

 1
Great Clarendon Street, Oxford
Oxford University Press is a department of the University of Oxford.
It furthers the Universitys objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trademark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
Jerrold Levinson 2006
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 2006
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, 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 book in any other binding or cover
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Levinson, Jerrold.
Contemplating art / Jerrold Levinson.
p. cm.
Includes index.
ISBN-13: 9780199206179 (alk. paper)
ISBN-10: 0199206171 (alk. paper)
ISBN-13: 9780199206186 (alk. paper)
ISBN-10: 019920618X (alk. paper)
1. Aesthetics. 2. ArtPhilosophy 3. Arts. I. Title.
BH39.L492 2006
701 .17dc22
2006016276
Typeset by Laserwords Private Limited, Chennai, India
Printed in Great Britain
on acid-free paper by
Biddles Ltd., Kings Lynn, Norfolk

ISBN 019920618X 9780199206186

ISBN 0199206171 (Pbk.) 9780199206179 (Pbk.)

1 3 5 7 9 10 8 6 4 2
 Contents

Acknowledgements vii

Introduction 1

I. ART
1. The Irreducible Historicality of the Concept of Art 13
2. Artworks as Artifacts 27
3. Emotion in Response to Art 38
4. Elster on Artistic Creativity 56

II. MUSIC
5. Sound, Gesture, Space, and the Expression of Emotion in Music 77
6. Musical Expressiveness as Hearability-as-Expression 91
7. Nonexistent Artforms and the Case of Visual Music 109
8. Music as Narrative and Music as Drama 129
9. Film Music and Narrative Agency 143
10. Evaluating Music 184
11. Musical Thinking 209
12. Musical Chills 220

III. PICTURES
13. Wollheim on Pictorial Representation 239
14. What is Erotic Art? 252
15. Erotic Art and Pornographic Pictures 259
vi Contents

IV. INTERPRETAT ION

16. Two Notions of Interpretation 275
17. Whos Afraid of a Paraphrase? 288
18. Hypothetical Intentionalism: Statement, Objections, and
Replies 302

V. AE ST HE T IC PROPE RT IE S
19. Aesthetic Properties, Evaluative Force, and Differences of
Sensibility 315
20. What Are Aesthetic Properties? 336

VI. HISTORY
21. Schopenhauers Aesthetics 355
22. Humes Standard of Taste: The Real Problem 366

VII. OT HER MAT TERS

23. The Concept of Humor 389
24. Intrinsic Value and the Notion of a Life 400

Index 419
 Acknowledgements

The Irreducible Historicality of the Concept of Art, British Journal of Aesthetics

42 (2002): 36779.
Artworks as Artifacts, in E. Margolis and S. Laurence (eds.), Creations of the
Mind (Oxford: Oxford University Press, 2006).
Emotion in Response to Art, (as Emotion in Response to Art: A Survey of the
Terrain), in M. Hjort and S. Laver (eds.), Emotion and the Arts (Oxford: Oxford
University Press, 1997), 2034.
Elster on Artistic Creativity, in B. Gaut and P. Livingston (eds.) The Creation of
Art (Cambridge: Cambridge University Press, 2003), 23556.
Sound, Gesture, Space, and the Expression of Emotion in Music (as Sound,
Gesture, Spatial Imagination, and the Expression of Emotion in Music), Euro-
pean Review of Philosophy 5 (2002): 13750.
Musical Expressiveness as Hearability-as-Expression, in M. Kieran (ed.), Con-
temporary Debates in Aesthetics and the Philosophy of Art (Oxford: Blackwell,
2005), 192206.
Nonexistent Artforms and the Case of Visual Music, in A. Haapala, J. Levinson,
and V. Rantala (eds.), The End of Art and Beyond: Essays after Danto (Atlantic
Highlands, NJ: Humanities Press, 1997), 12239.
Music as Narrative and Music as Drama, Mind and Language 19 (2004):
42841.
Film Music and Narrative Agency, in D. Bordwell and N. Carroll (eds.), Post-
Theory: Reconstructing Film Studies (Madison: University of Wisconsin Press,
1996), 25488.
Evaluating Music, in P. Alperson (ed.), Musical Worlds: New Directions in the
Philosophy of Music (University Park, PA: Pennsylvania State University Press,
1998), 93108. [An earlier version of this essay appeared in Revue Internationale
de Philosophie 198 (1996): 593614.]
Musical Thinking, Midwest Studies 27 (2003): 5968.
Musical Chills. [Earlier versions of this essay were Musical Frissons, Revue
Francaise dEtudes Americaines 86 (2000): 6476; and Musical Chills and Oth-
er Delights of Music, in J. Davidson (ed.), The Music Practitioner (Aldershot:
Ashgate Publishing, 2004), 33551.]
Wollheim on Pictorial Representation, Journal of Aesthetics and Art Criticism 56
(1998): 22733.
What is Erotic Art, (as Erotic Art), in E. Craig (ed.), The Routledge Encyclope-
dia of Philosophy (London: Routledge, 1998), 4069.
viii Acknowledgements

Erotic Art and Pornographic Pictures, Philosophy and Literature 29 (2005):

22840.
Two Notions of Interpretation, in A. Haapala and O. Naukkarinen (eds.), Inter-
pretation and its Boundaries (Helsinki: Helsinki University Press, 1998), 221.
Whos Afraid of a Paraphrase?, Theoria 67 (2001): 723.
Hypothetical Intentionalism: Statement, Objections, and Replies, in M. Krausz
(ed.), On the Single Right Interpretation (University Park, PA: Pennsylvania State
University Press, 2002), 30918.
Aesthetic Properties, Evaluative Force, and Differences of Sensibility, in E.
Brady and J. Levinson (eds.), Aesthetic Concepts: Essays after Sibley (Oxford:
Oxford University Press, 2001), 6180.
What Are Aesthetic Properties?, Proceedings of the Aristotelian Society. Supple-
ment 78 (2005): 21127.
Schopenhauers Aesthetics, (as Schopenhauer, Arthur), in M. Kelly (ed.), The
Encyclopedia of Aesthetics (New York: Oxford University Press, 1998),
24550.
Humes Standard of Taste: The Real Problem, Journal of Aesthetics and Art
Criticism 60 (2002): 22738.
The Concept of Humour, (as Humour), in E. Craig (ed.), The Routledge Encyc-
lopedia of Philosophy (London: Routledge, 1998), 56267.
Intrinsic Value and the Notion of a Life, Journal of Aesthetics and Art Criticism
62 (2004): 31929.
 Introduction

Contemplating Art is the third of my essay collections in philosophy of art,

following clearly in the line of Music, Art, and Metaphysics (1990) and The
Pleasures of Aesthetics (1996). All three volumes are situated in what may
be called mainstream analytic aesthetics, or aesthetics in the tradition of
analytic philosophy. The present volume brings together the bulk of my
work in this vein in the past ten years, and contains twenty-four essays,
making it considerably larger than its predecessors. That it covers a decade
of work accounts in part for its size, but also relevant is the inclusion of one
essay, Film Music and Narrative Agency, that is almost a small book in
itself.
I have grouped the essays into seven parts, on roughly thematic grounds.
Part I contains four essays on art in general, raising issues in art theory not
closely tied to a particular artform. Part II, the longest in the book, contains
essays dealing with philosophical problems specific to music, the art that has
always been my principal occupation as an aesthetician. Part III brings togeth-
er three essays that concern pictorial art, while Part IV brings together three
essays that concern interpretation, and more particularly, the interpretation of
literature and literary language. Part V consists of two essays on the nature
of aesthetic properties, the sort of properties exhibited prominently, if not
exclusively, by works of art, while Part VI consists of two essays that address
issues in historical aesthetics. Finally, Part VII contains essays on two topics,
humor and intrinsic value, falling somewhat outside the scope of aesthetics as
usually conceived, though their relevance to central issues in aesthetics should
nevertheless be apparent.

For an overview of that mainstream the reader is invited to consult J. Levinson (ed.), The
Oxford Handbook of Aesthetics (Oxford: Oxford University Press, 2003), though work outside the
mainstream also receives extensive coverage.
It was in fact published as such in France, under the title La musique de film: fiction et narration
(Pau: Presses Universitaires de Pau, 1999).
2 Introduction

In the opening essay, The Irreducible Historicality of the Concept of Art,

I return once more to the intentional-historical theory of arthood I have
championed since my first paper on the topic in 1979. After succinctly
restating the account, according to which, reduced to essentials, arthood is
a matter of being projected for regard or treatment as some earlier artworks,
or what are taken to be such, are or were correctly regarded or treated, I
underline in particular the historical element in that account, which captures
an inescapable aspect of the modern concept of art, and which dooms to
inadequacy any purely formal or functional account of arthood. Most of the
essay is devoted to responding to recent reservations about the intentional-
historical theory, though since of the making of objections there is no
end, I harbor no illusion of having responded to all the reservations in its
regard to be found in the current literature. The second essay, Artworks
as Artifacts, is concerned with that same account of arthood, but here it is
the nature of the artifactuality of artworks presupposed by the account that
is the focus of attention. I develop my ideas on the artifactuality of artworks
in counterpoint with recent contributions on the subject by Paul Bloom
and Amie Thomasson. Against Bloom, who seeks to extend the intentional-
historical account to all artifacts, I defend the claim that artworks remain a
distinctive sort of artifact in possessing, perhaps alone of all artifact kinds,
only intentional-historical necessary conditions. Against Thomasson, who
maintains that artifact-making necessarily involves a substantive conception
of what is being made, I defend the claim that the conception of artwork
necessarily involved in art-making, although not without content, is about
as insubstantive as an object concept can be.
Emotion in Response to Art is a survey essay on the range of philosoph-
ical problems that can be encompassed under that rubric. It details five such
problems, according most of its attention to the first two of those, namely,
the nature of the emotional responses had to art, and the puzzle of emotional
responses to fictional entities known to be fictional (what is often labeled the

Defining Art Historically, British Journal of Aesthetics 19 (1979): 23250, reprinted in Music,
Art, and Metaphysics (Ithaca: Cornell University Press, 1990). This was followed by two further
essays expounding and defending the theory: Refining Art Historically (1989), reprinted in Music,
Art, and Metaphysics, and Extending Art Historically (1993), reprinted in The Pleasures of Aesthetics
(Ithaca: Cornell University Press, 1996).
Two substantial critiques that appeared after the essay was published, to which I thus do not
there respond, are Nigel Warburton, The Art Question (London: Routledge, 2003), ch. 4, and
Victor Yelverton Haines, Recursive Chaos in Defining Art Recursively, British Journal of Aesthetics
44 (2004): 7383.
 Introduction 3

paradox of fiction). But attention is also given to the puzzle of how people
derive satisfaction from art expressive or evocative of negative emotion (what
is often labeled the paradox of tragedy), and to the question of how abstract
works of art manage to express or evoke emotions at all. Elster on Artistic
Creativity is a study of what of general import might be said about the pro-
cesses or principles of creativity in art, conducted through an examination of
a thought-provoking discussion of artistic creativity by the social theorist Jon
Elster. I take issue with Elsters account of creativity in art as simply a mat-
ter of optimizing choice within constraints following an earlier stage of choice
of constraints, and also take issue with some of the evaluative consequences,
both general and specific, that Elster draws from his account.
All the essays in Part II concern principally the art of music, and most of
them bear connections to earlier writings of mine. The first two essays are
both concerned with the problem of musical expressiveness, how it is to be
analyzed and what it is to perceive or experience it. Musical Expressiveness as
Hearability-as-Expression is continuous with an earlier essay entitled simply
Musical Expressiveness, and defends the analysis of that phenomenon
reflected in its title, according to which music is expressive of an emotion or
other mental state insofar as it induces us to hear it as the personal or personlike
expression of that mental state. Along the way various competing theories
of musical expressiveness, notably those of Malcolm Budd, Stephen Davies,
Robert Stecker, and Roger Scruton, are submitted to critical examination.
Sound, Gesture, Space, and the Expression of Emotion in Music, which
in addition to drawing on Musical Expressiveness also reworks material
from an even earlier essay, Authentic Performance and Performance Means,
emphasizes first the role of grasp of musical gesture in the grasp of musical
expressiveness, and second, the role of spatial imagination in the grasp of
musical gesture.
Nonexistent Artforms and the Case of Visual Music is the oldest of the
essays reprinted here, in terms of its date of composition, having been written
for a conference on the future of art held in Lahti, Finland in 1990. The first,
somewhat fanciful, half of the essay is not specifically concerned with music,
but attempts rather to sketch a general framework for thinking about nonex-
istent yet possible artforms, issuing in a number of schematic formulas for
generating such artforms in the abstract. The second, more concrete, half of
the essay takes as a case study the relative nonexistence of visual music, despite

In The Pleasures of Aesthetics. In Music, Art, and Metaphysics.

4 Introduction

numerous attempts in that direction over the years, and proposes an explana-
tion of visual musics stubborn failure to establish itself as a viable artform.
The next two essays deal, from different angles, with the relationship of
music to narrative. Music as Narrative and Music as Drama pointedly poses
the question of whether music, especially as regards its succession of expressive
properties or states, is fruitfully thought of as a narrative of some sort. The
answer returned is guardedly negative, and the attractions of an alternate
model, one owing to the musicologists Anthony Newcombe and Fred Maus,
of expressive music as dramatized rather than narrated emotion, are touted
instead. Film Music and Narrative Agency, which, as already noted, is the
longest essay in this collection, is as much concerned with film as it is with
music. It seeks to illuminate, on the basis of an account of making fictional
along lines laid down by Kendall Walton, and through an extensive survey
of examples, the ways and means by which extrinsic film music inflects the
fictional content of a film, identifying two distinct modes in which that can
occur, one in which such music is ascribed to the films cinematic narrator,
and one in which such music is ascribed, less commonly, to the films implied
director.
The next essay, Evaluating Music, is an attempt to identify plausible mid-
level principles by reference to which one might conceivably justify an evalu-
ation of some music as good, where by mid-level principles I mean principles
whose specificity lies between the extremes of, on the one hand, musics being
good if it affords appropriate listeners worthwhile experiences, and on the
other hand, musics being good if it displays this or that set of technical fea-
tures held to be productive of musical worth, such as monothematic structure
or coherent harmony. The perspective of Evaluating Music derives from that
developed in an earlier essay, What Is Aesthetic Pleasure?, where I propose
that the distinctive mark of aesthetic satisfaction in art is that it is satisfaction
deriving from attention that focuses, above all, on the relation of content to
form and form to content in the given work of art. The mid-level principles of
musical evaluation arrived at in light of that perspective on aesthetic satisfac-
tion are then illustrated in connection with one of Schuberts piano sonatas,
the Sonata in A major, D. 959.
The last two essays in Part II, Musical Thinking and Musical Chills,
like Nonexistent Artforms and the Case of Visual Music, tackle questions
in musical aesthetics that have not been the focus of much, if any, discussion.

In The Pleasures of Aesthetics.

 Introduction 5

Musical Thinking, which begins with a commentary on Wittgensteins scat-

tered remarks on musical understanding, poses the question of whether there
is a distinctive, non-verbal form of thinking that music, or alternatively, the
composing or performing of music, might be said to exemplify. A positive
answer is returned, and three candidates for such distinctively musical think-
ing are sketched; these are illustrated with a number of musical examples,
most notably, Beethovens Tempest Sonata and Stan Getzs rendition of
The Girl from Ipanema. Musical Chills, of all the essays in this collec-
tion, is the one that has undergone the most evolution since it was first drafted
around 1998, having already been published twice, under different titles, and
in truth my thinking on the subject remains in flux, despite my committing
it to print once again. It is also the only essay I have written to date whose
principal spur was an empirical study, one concerned with a musical phe-
nomenon that has always fascinated me, namely, the distinctive and usually
pleasurable chills, shivers, or frissons that listening to certain passages of music
produces in many listeners. At any rate, after describing the phenomenon and
situating it in the field of musical pleasures as a whole, and after considering
and finding wanting the explanations of the phenomenon and its value that
have been offered by cognitive psychologists, I try to construct a more sat-
isfactory explanation, one illustrated most fully in connection with a piano
piece of Scriabin, his Etude in C# minor, op. 42, no. 5.
Part III initiates a shift of focus to the visual arts. Wollheim on Pictorial
Representation was written as a contribution to a symposium in honor of the
distinguished aesthetician Richard Wollheim, and begins with a sympathetic
summary of his highly influential account of depiction in terms of the
successfully realized intention that viewers have a certain sort of seeing-in
experience faced with a picture depicting a given subject. While agreeing
with the basic thrust of Wollheims account, which makes a certain sort of
visual experience in appropriate viewers criteria of achieved depiction, I differ
with Wollheim as to whether that experience is invariably one of seeing-
in, given the twofold attention to subject and surface that that notion, as
Wollheim conceives it, necessarily involves. I sketch an alternative account,
Wollheimian in spirit, but closer than most recent proposals to the classic
Gombrichian view of depiction as involving something akin to illusion. What
I propose specifically is that a picture that depicts a subject is one fashioned so

Conducted by the neuropsychologist Jaak Panksepp.

6 Introduction

as to yield an experience of as-if seeing of its subject, but not an experience that
engenders the false beliefs typical of illusion.
As is evident, the next two essays in Part III have a common theme, namely,
the erotic in art. What Is Erotic Art?, an expanded version of an encyclopedia
article published in 1998, and my first foray into this domain, straightfor-
wardly addresses the question of its title. The answer offered is not calculated
to astonish: erotic art is, first, art, and second, erotic. In less sphinx-like terms,
erotic art is art that aims to engage viewers sexually through explicit sexual
content, and that succeeds at least to some extent in doing so. This answer is
held up for confirmation to a range of examples of the category, some uncon-
troversial and some less so, and a number of useful subcategories of erotic art
are identified. Erotic Art and Pornographic Pictures, which like its prede-
cessor confines its attention to the visual, was written in response to a 2001
essay of Matthew Kieran, itself prompted in part by remarks on the distinc-
tion between the erotic and the pornographic offered at the end of What Is
Erotic Art?. Whereas Kieran holds that there is no incompatibility, and even
precious little tension, between somethings being pornography and some-
things being erotic art, I hold, and endeavor to demonstrate, that there is
indeed such tension, and that the two statuses are in fact incompatible. That
said, nothing is entailed as to whether pornography, though it is not art, may
or may not be, for various reasons, of value.
Another shift of focus is effected in Part IV, whose three essays concern
for the most part literature and literary language. The first essay, however,
is of somewhat more general scope. Two Notions of Interpretation brings
into relief a distinction among semantic interpretations, or among activities
of semantic interpreting, that cuts across verbal and non-verbal phenomena.
The distinction is between interpretations that aim to answer the question
What does such and such mean? and those that aim to answer the question
What could such and such mean?, the former exemplifying the determinative
mode, and the latter the exploratory mode, of interpreting. In the rest of the
essay I investigate, through a range of examples literary and non-literary, the
relationship between determinative and exploratory interpretation in a given
inquiry, and the varying, sometimes interlocking, motivations with which
determinative and exploratory interpretations are undertaken.
In Whos Afraid of a Paraphrase? I turn specifically to the interpretation
of metaphors. My principal claim, in opposition to the well-known stance of
Donald Davidson, is that metaphors, however much their force or imagery
 Introduction 7

outstrips their semantic content, in fact usually possess relatively definite

meanings, meanings which deserve the label of metaphorical, and which
paraphrases can to a large extent express. The key to the stance on metaphors
adopted is the conception of them as utterances in specific linguistic contexts,
which acquire meanings in such contexts despite there being no rules of a
semantic sort for the projection of such meanings. Examples of metaphors
from both literary and non-literary contexts come in for examination. The
conception of literary meaning as centrally a species of utterance meaning
is the foundation stone of the view of literary interpretation dubbed
hypothetical intentionalism that I have argued for in two earlier essays. In
Hypothetical Intentionalism: Statement, Objections, and Replies I briefly
restate the view, which locates the meaning of a literary text not in what its
author intended it to mean (what one can call utterers meaning), nor in
what the text might be said to mean as a piece of language in the abstract
(what one can call textual meaning), but roughly in what an appropriate
audience would most reasonably hypothesize the contextually situated author
to have meant by composing precisely the text that he or she did. I then
proceed to consider a fair number of objections to the view in the literature
and attempt to respond to them. But as this is currently a very active area of
research I am, as with my replies to objections to the intentional-historical
account of arthood, under no misapprehension that these will constitute the
last words on the subject.
The concerns of the two essays in Part V, which are continuous with those
in my earlier Aesthetic Supervenience, are as much metaphysical as aes-
thetic. The central issue is the nature and objectivity of aesthetic properties,
especially those belonging to works of art. In both essays I defend aesthetic
realism, by which is meant the claim that aesthetic properties exist, that they
are bona fide properties, and that their possession constitutes the truth con-
dition of true aesthetic attributions. In the first part of Aesthetic Properties,
Evaluative Force, and Differences of Sensibility, written for a 1997 confer-
ence in honor of the influential British aesthetician Frank Sibley, I sketch a
largely Sibleyan view of aesthetic attributions, though a more metaphysically

Intention and Interpretation in Literature and Messages in Art, both in The Pleasures of
Aesthetics.
An important recent set of words on the subject, for instance, of which no account is taken
here, is Paisley Livingstons Art and Intention (Oxford: Oxford University Press, 2005).
In Music, Art, and Metaphysics.
8 Introduction

committed one than Sibley was inclined to hold, underlining the extent to
which a descriptive core can be located in almost all such attributions, what-
ever evaluative force they may carry, and however implicitly relativized they
may be to perceivers of certain sorts. I then proceed to formulate and respond
to a number of concerns one might have about this brand of aesthetic real-
ism. In What Are Aesthetic Properties? I extend my defense of aesthetic
realism, devoting most of my attention to the issue of how we should con-
ceive of properties in general and of aesthetic properties in particular. What
I propose is that at least paradigm cases of the latter are to be understood
as higher-order perceptual ways of appearing. In the course of developing this
proposal I address the vexed issue of whether or not aesthetic properties are
response-dependent, or such that they cannot be conceived or analyzed except
in terms of kinds of responses in relevant perceivers, and conclude by sug-
gesting that there is a spectrum, among properties usually thought of as aes-
thetic, from ones that are clearly response-dependent to ones that are clearly
non-response-dependent, with many gradations in-between.
The two essays in Part VI take up themes from the history of aesthetics.
My aim in Schopenhauers Aesthetics, written as an encyclopedia article,
is largely expository. I begin with Schopenhauers relationship to Kant, and
the extent to which the great pessimists aesthetic philosophy relies on Kants
metaphysics even more than it does on Kants aesthetics, and then go on to
highlight the breadth of Schopenhauers vision of the role of art and of the lib-
erating aesthetic experiences it makes possible. At the end I address the puzzle
of how the art of music, which according to Schopenhauer presents us with
blind, ceaseless, and hateful willing in its most unvarnished form, can yet
provide aesthetic experience of the highest order, justifying Schopenhauers
according to music the foremost position among the arts. My aim in Humes
Standard of Taste: The Real Problem, on the other hand, is more polemical
than scholarly. I there formulate a persisting problem about the authority of
art criticism, one that should concern anyone for whom the arts occupy an
important place in life, and situate this problem in relation to Humes search
for the standard of taste in his famous essay of that name. I then sketch a
complex solution to this problem, somewhat provocatively labeled the real
problem left us by Hume, a solution whose complexity is justified by the
thorniness of the problem in question.

As seems to be the case, say, for properties like nauseatingness or disgustingness.

 Introduction 9

The Concept of Humor, also written as an encyclopedia article, surveys

the main theories of humor in the philosophical tradition, and then proposes
a novel account of the essence of humorousness, often regarded as an aesthetic
property. This essence is held to lie not in perceived incongruity, nor in per-
ceived superiority, nor in the power to trigger experiences of relief, but in
the disposition to produce affect of a sort tied identifyingly to laughter. This
account is dubbed the affective theory of humor, and some recent objections
to it are discussed and defused. So far as the causes or mechanisms through
which humorousness is achieved are concerned, I discuss the pros and cons of
the leading theory in that vein, the so-called incongruity theory of humor,
and side in conclusion with those who hold that resolution of incongruity is
perhaps closer to the heart of the matter than incongruity per se.
Finally, in Intrinsic Value and the Notion of a Life, I address a problem in
the general theory of value that goes beyond the concerns of aesthetics as such,
though aesthetic issues at one point serve to bring into relief the nature of
the thesis about intrinsic value ultimately arrived at. That thesis concerns the
form that defensible judgments of intrinsic valueroughly, what is of value in
itself or for its own sakemust take, or equivalently, the sort of thing that may
defensibly be claimed to have intrinsic value. The thesis defended, which tries
to mediate between object-based and experience-based conceptions of the
intrinsically valuable, is that a richly sentient life being a certain way is the only
possible subject of a defensible judgment of intrinsic value. One consequence
of this thesis is a disagreement with G. E. Moore regarding the intrinsic value
of a beautiful world devoid of sentience, a famous thought experiment from
his Principia Ethica. But a more important consequence is the suggestion, if
I am right, of an intimate connection between the notion of a richly sentient
life and the very idea of intrinsic value.

Thanks are owed to all the following for helpful comments on the essays
collected here at various stages in their evolution: Lars-Olaf Ahlberg, Jose
Bermudez, Paul Boghossian, Malcolm Budd, Noel Carroll, David Chalmers,
Ted Cohen, Jean-Pierre Cometti, Jack Copeland, Gregory Currie, David
Davies, Stephen Davies, Rafael De Clercq, Sabine Doring, John Doris,
Hubert Eiholzer, John Fisher, Berys Gaut, Alessandro Giovannelli, Stan God-
lovitch, Mitchell Green, Arto Haapala, Garry Hagberg, Robert Hatten, Peter
Lamarque, Keith Lehrer, Paisley Livingston, Mike Martin, Derek Matravers,
Fred Maus, Aaron Meskin, Daniel Nathan, Alex Neill, David Novitz,
10 Introduction

Elisabeth Pacherie, Derk Pereboom, Paul Pietroski, Diane Proudfoot, Aaron

Ridley, Mark Rollins, Anthony Savile, Martin Seel, Roger Shiner, Elliott
Sober, Robert Stecker, Joseph Tolliver, Saam Trivedi, Kendall Walton,
Alicyn Warren, Susan Wolf, and Nick Zangwill.
Finally, thanks to Peter Momtchiloff for his aid and encouragement, and
to Ludmilla Kolokolova for her love and support, throughout the process of
bringing this book to fruition.
PART I
ART
This page intentionally left blank
 1
The Irreducible Historicality of the Concept
of Art

I INT RODUCT ION

I claim that our present concept of art is minimally historical in the following
sense: whether something is art now depends, and ineliminably, on what has
been art in the past. I claim, in other words, that the concrete history of art is
logically implicated in the way the concept of art operates, and that some part
of that history is involved, either opaquely or transparently, in the claim to
arthood made by any work of art. By contrast, the concepts
square,
red,

pig,
mountain, and so on are not obviously historical in this sense: whether
something falls under them does not seem to depend in the same way on what
specifically fell under them in the past, and to operate those concepts correctly
you do not need to invoke the concrete history of their correct application.
The gist of the intentional-historical conception of art that I advocate is
this: something is art iff it is or was intended or projected for overall regard
as some prior art is or was correctly regarded. As is evident, such a conception
attributes to art, and centrally, the property of minimal historicality sketched
above. In this brief essay I will forgo defense of the sort of complete defini-
tion of art I am inclined to favor, and that I have tried to articulate in three

This chapter was first published in British Journal of Aesthetics 42 (2002): 36779.
It has even been argued recently that this may be true of all artifact concepts, artistic and
nonartistic alike. See Paul Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996):
129. I comment briefly on Blooms intriguing suggestion at the end of this essay, and again in the
following essay.
Jean-Pierre Comettis essay, Misereou grandeurde lhistoricisme?, in Jean-Pierre Cometti
(ed.), Definitions de lart (Brussels: La lettre volee, 2002) has helped me to see the importance of
dissociating the minimal historicism of art claimed by my theory from more robust historicisms of
a Hegelian or Dantoesque sort, such as ascribe to the development of art an inherent goal, or view
the development of art as governed by inherent laws of stylistic evolution. In that light, it might
have been better to denominate my theory of art a retrospectivist or auto-referentialist one, rather
than a historicist one, in order to avoid such unwanted associations.
14 Art

previous essays. I will also largely ignore questions regarding the sufficiency
of an intentional-historical condition for arthood, and questions as to the
necessity of the intentional component of such a condition, in order to focus
on the necessity, in some guise or other, of the historical component.
My ambition in the present outing is thus modest. I aim to do only two
things. One is to underline the necessity of a historical dimension in any
acceptable account of arthood. Two is to sketch answers to certain objections
that have been recently raised for an intentional-historical account of art, most
of which offer a challenge to its insistence on an ineliminable historical element
in any such account. In addition, in the course of underlining the historical
character of the concept of art I hope to show that certain non-historical consid-
erations appealed to by some theorists, for instance, institutional or functional
ones, which appear to weigh importantly in some cases of arthood, in fact have
an underlying or reinforcing rationale of a history-involving sort.

II OBJECTIONS AND REPLIES

I now consider a number of objections that have been lodged against the
intentional-historical theory of art, and offer replies to them.

The Objection from the Implausibility of a Recursive Definition

of Art6
Some writers have objected to the intentional-historical definition of art on the
grounds that it is a recursive definition, or else entails that art can be defined

See Defining Art Historically, British Journal of Aesthetics 19 (1979): 23250, and Refining
Art Historically, Journal of Aesthetics and Art Criticism 47 (1989): 2133, reprinted in Music,
Art, and Metaphysics, 2nd edn. (Oxford: Oxford University Press, 2006), and Extending Art
Historically, Journal of Aesthetics and Art Criticism 51 (1993): 2133, reprinted in The Pleasures of
Aesthetics (Ithaca: Cornell University Press, 1996). I have also replied to criticisms of my theory in
some shorter pieces: A Refiners Fire: Reply to Sartwell and Kolak, Journal of Aesthetics and Art
Criticism 48 (1990): 2315; Further Fire: Reply to Haines, Journal of Aesthetics and Art Criticism
48 (1991): 767; and Art Historically Defined: Reply to Oppy, British Journal of Aesthetics
33 (1993): 3805. See also Robert Stecker, Artworks: Definition, Meaning, Value (University
Park, PA: Pennsylvania State University Press, 1997), 8898, for a critical reconstruction of the
intentional-historical theory of art.
These were queried vigorously by Noel Carroll in his Identifying Art, in Robert Yanal (ed.),
Institutions of Art (University Park, PA: Pennsylvania State University Press, 1994), 339.
See Graham Oppy, On Defining Art Historically, British Journal of Aesthetics 32 (1992):
15361. Some of Oppys criticisms are anticipated in Stephen Davies, Definitions of Art (Ithaca,
NY: Cornell University Press, 1991).
See, for example, Tom Leddy, The Socratic Quest in Art and Philosophy, Journal of Aesthetics
and Art Criticism 51 (1993): 399410
 The Irreducible Historicality of the Concept of Art 15

recursively, neither of which strikes those writers as a happy result. But strictly
speaking, the charge is mistaken. My basic definition of art is a one-step affair,
as is evident even in Steckers reconstruction of it. What I have proposed is that
the full extension of art in a given tradition might be displayed by a recursive
definition, but not that our present concept of art is to be explicated by such a
recursion. Again, its true that my definition implies that the totality of art in a
given tradition has a recursive structure, but that is not tantamount to my having
defined art recursively. In underlining that the intentional-historical definition
of art is not as such a recursive one I am thus denying that the notions of first art
and ur-art, with which such recursions can be thought to begin, are elements
in our concept of art, and that what we mean by an artwork is something that
either is or stands in the appropriate relation to instances of first art or ur-art.
This is, of course, all to the good, since it would be implausible to maintain
that such notions are a part of the ordinary grasp of what arthood is.

The Objection from Unwanted Descendants of the Ur-Arts7

Ancestors of art activities, such as ritual cave paintings, may also turn out to be
ancestors of present-day activities that are clearly non-art, such as deer hunt-
ing with high-caliber rifles. But then my definition, it seems, will wrongly
count these latter activities as art.
My reply to this is as follows. Though that sort of misfiring of the definition
is conceivable, it is arguable that in presumed cases of this sort the links from
remote to present-day activities are not precisely of the right sort, that is, of
the backward-looking-intentional-invocation-of-regard sort. In other words,
the generating principle of these other sequences, ones that begin with some
ur-art and issue in clearly non-art activities, is likely not precisely of the sort
involved in the generation of artistic chains. One would have to examine
closely a putative concrete aberrant chain, leading from unequivocal ur-art
to unequivocal non-art, to assess fairly the strength of this objection. But it is
unclear that any such chains survive scrutiny.

The Objection from the Obsolescence of Art-Regards

Here is a forceful statement of this objection, as put by Noel Carroll:
Levinson supposes that something might be art now just in case it supports any type
of regard, treatment, or mode of appreciation that was appropriate to at least some

See Noel Carroll, Historical Narratives and the Philosophy of Art, Journal of Aesthetics and
Art Criticism 51 (1993): 31326
16 Art

works of art in the past. The problem here is that not every mode of appreciation
that was lavished on artworks in the past is eternally available. Some modes may have
become historically obsolete . . .
Carroll then goes on to give a concrete counterexample to illustrate his
abstract charge. This involves a benighted chicken farmer named Jones, one
steeped in old-time religion, who sets out to make an artwork. Specifically,
he shoots a mass of chickens in record time in order to propitiate the gods.
Moreover, he presents the massacre as an artwork: onlookers are invited to
appreciate it, assess it, or regard it in terms of its effectiveness as a means
of propitiating the gods. According to Carroll, since propitiating the gods
was an aim of some past artworks, and since such artworks were correctly
appreciated in terms of such an aim, my theory must acknowledge
for
appreciation as an instance of god-propitiation as an art-making regard, and
thus Joness action as an artwork, which, given Jones cannot be understood as
either a Conceptual or a Performance artist, seems wrong. [Joness] intention
is simply to make something that is to be regarded . . . as a vehicle for
propitiating the gods, where propitiating the gods was once an acknowledged
purpose of art.
But Carrolls conclusion is unwarranted. The problem with his objection is
that it mistakes a single, isolated regard appropriate to some past artworks for
a complete, integral ensemble of regards appropriate to some past artworks. It
is only in being intentionally projected for the latter, not the former, that an
object acquires the status of artwork on my theory. Although some ancient
artworkssay, certain tragedies or templeswere intended, let us assume,
for appreciation as instances of god-propitiation, it is certainly not the case
that they were intended solely for appreciation in that respect. Surely they were
also intended for other regards, involving attention to those works emotional,
formal, and symbolic aspects. Hence Joness chicken-slaughtering act, being
intended only for appreciation as an instance of god-propitiation, is not an
act intended for regard in the way any past artworks were as a whole correctly
regarded. So the charge that the definition misfires when it invokes obsolete

See Carroll, Identifying Art, 334.

Ibid. 345. Ibid. 35.
I in fact dealt with this objection preemptively in my Refining Art Historically. I there
underlined that only relatively complete or total ways of regarding are to be allowed as substitution
instances of the artwork formula thing which has been seriously intended for regarding-as-a-
work-of-art, i.e. regard in any way preexisting or prior artworks are or were correctly regarded
(p. 24).
But, it may be asked, what if Jones does intend that those other sorts of regards be taken to
his act as well, those for which those ancient temples and tragedies were also intended? In that
 The Irreducible Historicality of the Concept of Art 17

ways of regarding art itself misfires. Integral ensembles of regards appropriate

to past artworks in fact never become obsolete, so far as their potential to
enfranchise possible future artworks is concerned, though of course they may
fall out of favor or fashion.

The Objection from the Putative Ascension of Attractive

Nonartworks to the Ranks of Artworks
There are admittedly many cases of attractive purely utilitarian objects
subsequently treated as artworks by some individual or individuals, counter
to or in the absence of any artistic intentions on the part of their creators.
But it is an error, I suggest, to think this makes such objects into artworks:
audiences, appreciators, consumers cannot make things art merely by treating
them as such. Of course, as far as aesthetically appealing objects of real-world
human cultures are concerned, e.g. pots, knives, masks, curtains, rugs, the
idea that any of them were ever conceived or projected in a purely utilitarian
way is implausible. Thus, when we exhibit such objects in art museums we
dont need to be thought of as transforming or altering their status, but as
simply acknowledging the quasi-artistic status they already have, at least in
part, as created. On the other hand, it is possible that some such objects, e.g.
magic-ritual ones aimed only at invoking the spirits or manipulating natural
forces, really dont belong in art museums, given their original constitutive
projection by their makers, however aesthetically interesting or artistically
advanced they might seem.
In any event, the least that can be said is that such examples of putative
artworks lacking the appropriate sort of intentional projection by their
makers are too much in dispute to serve as decisive counterexamples to an
intentional-historical conception of arthood.

The Objection from the Puzzling Status of First Art13

What secures for first art its status as art such that it is art from the outset and
thus capable of anchoring the chain of artworks that, according to historical
theories of art, reaches from first art to the present? Stephen Davies puts the
difficulty as follows: First art must be art already at the time second wave

case, it seems my theory might have to count Joness act as art; but then it would also be less
counterintuitive to so count it.
See Stephen Davies, First Art and Arts Definition, Southern Journal of Philosophy 35 (1997):
1934.
18 Art

pieces become art, otherwise second wave pieces could not be art as a result of
standing to first art in the art-defining relation. The intentional-historical
theory of art seems to have no acceptable answer to how first art becomes art,
since first art is, by definition, not related to any prior art by the art-making
relation the theory proposes as crucial.
I have elsewhere attempted to supply answers to that question, some of
which were, admittedly, not acceptable. The answer I favor now is that first
art is indeed art at the time of making, but that it is art in a somewhat different
sense, or for a slightly different reason, than all subsequent art. Obviously,
first art in a given tradition bears to ur-artthe ultimate non-art progenit-
ors of artworks in that traditionsomething like the relation that second art
bears to first art, and that all subsequent art bears to art preceding it. The rela-
tion is roughly that of being projected for treatment or regard in ways that
earlier objects were appropriately regarded or treated. But since ur-art objects
are not art, this cannot be described univocally as projection for regard in
the way earlier art was appropriately regarded. There remains an irreducible
difference between ur-art and first art apart from mere temporal precedence,
which consists in the fact that, for both ur-art and first art objects, while there
are certain regards or treatments that are appropriate to them, only the lat-
ter are projected for regard or treatment in the way the former appropriately are
regarded or treated. Though both first art and ur-art objects are artifacts, whose
identities are governed by intentions, the intentionality that makes first art art
ineliminably refers to earlier things and activities, but not so the intentionality
that makes ur-art ur-art.
Anyway, it looks like an expanded definition of art is needed if both first
art and later art are to be comprised, and that this will need to be disjunctive
in form. It would be that something is art iff either (a) it satisfies the basic
definition or (b) it is an instance of first artthat is, one of those things from
which all other art, that satisfying the basic definition, springs.

Ibid. 21.
For example, the suggestion that first art acquires its art status retroactively, after the art
tradition which it stands at the beginning of gets going. (See my Art Historically Defined: Reply to
Oppy.) As Davies rightly notes, an only-retroactively-conferred art status for first art would vitiate
the recursive chain of artmaking from the outset. It is only retrospectively that we can determine,
given adequate archeological research, the identity of the ur-arts for a given tradition of art, but
it remains true that they always were that. It is not a matter of their retroactively becoming those
ur-arts.
See Extending Art Historically.
 The Irreducible Historicality of the Concept of Art 19

The Objection from Anthropocentricity17

It is surely the case that ancient Martians, active in the remote past, could
have made art even though their artifacts would not have been intentionally
related to any preceding works of human art, as the intentional-historical
theory seems to require. Suppose we respond, then, by liberalizing the theory
so as to allow for an objects being art in virtue of being intentionally linked in
the right way to succeeding works of human art. Matters would not, however,
thereby be much improved. For granted that liberalization of the theory it
would still be entailed that Martians could not even have known that their
artifacts were art until humans came along, thousands of years later, which
seems unintuitive. The intentional-historical theory thus fails to capture the
concept of art in its full generality, the objection concludes, because it takes
what is contingent . . . [namely, arts concrete historical realization] . . . for
something essential in art.
Examples of this sort, which highlight for our attention possible
artworks and art practices that predate human history altogether, do appear
to call for some modification of the intentional-historical definition as
originally proposed. For clearly these would be artworks that failed to
possess intentional connections, even opaque-to-their-creators intentional
connections, to earlier works of human art, there being by hypothesis no
such earlier works. The liberalization I favor to meet that problem is not,
however, that ventured above, that the works in question are art through
having the right intentional relations to later works of human art. Rather,
what I propose is that we might with justice consider such works as ancient
Martians may have produced to be art in virtue either of (a) their having
come about in a reflexive practice of making-and-intending-for-regard-as-
earlier-such-objects-were-regarded similar to our own practice of art, or
(b) their having been made and intended for particular regards acknowledged
in our own contingently evolved art history, or (c) the conjunction of

See Gregory Currie, Aliens, Too, Analysis 53 (1993): 11618. For another reply to Currie,
see Robert Stecker, Alien Objections to Historical Definitions of Art, British Journal of Aesthetics
36 (1996): 3058, and reprised with small changes in ArtWorks, 1078.
Currie, Aliens, Too, 118. See also Oppy, On Defining Art Historically.
Currie correctly notes that the intentional connections between later and earlier artifacts may
very well be of this kind, whereby the intention governing later artifact B connects it to earlier
artifact A even though the maker of B is oblivious of A. See Aliens, Too, 117.
20 Art

(a) and (b). Call the intentional-historical theory liberalized in some such
fashion the extended intentional-historical theory.
But what now about the argument from knowledge, which applies to
extended intentional-historical theory as much as it applies to the theory as
originally proposed? There are, it appears, two questions to address. First,
could such artistic Martians have known that they were, by the theorys lights,
making art, and second, to the extent that they could not, is that indeed
unacceptably counterintuitive?
The answer to the first question is complicated. Here are some things that
those Martians, in the course of making art, could have known: (a) that they
were making things for a certain sort of regard or treatment; (b) that they were
making things to be regarded or treated as certain earlier things made by them
were properly regarded or treated. Here, on the other hand, are some things
that they, in the course of making art, clearly could not have known: (c) that
they were making things for a sort of regard or treatment featuring in some
future earth art tradition; (d) that they were making things to be regarded
or treated as certain earlier things made by them were properly regarded or
treated, in just the same manner as that in which art on earth typically relates
to preceding art on earth. But it is (c) or (d), it seems, that they would have
had to have known in order to have known that they were making art on
the extended intentional-historical theory of art. So, indeed, they could not
have known they were making art, by our present concept of art, even if the
extended intentional-historical account of that is correct. That is, they could
not have known that their stuff was art in the full historically-reflexive sense I
claim is now ours, though they could of course have known that what they
produced was *art*, where *art* is some non-historicist predecessor of our
current concept of art, one that was roughly adequate to artistic production in
our culture prior to the early twentieth century.
Now for the second question: how counterintuitive is that? Properly
viewed, not very. What we have just seen is that those Martians could not
have known that their stuff was art as we now understand that, that is, art

I earlier ventured such a suggestion in Extending Art Historically, 4223. Compare a similar
and later suggestion by Stecker: One can say that although Martian modern art does not stand in
an appropriate relation to previous human art, it does stand in such a relation to previous Martian
art . . . So Martian modern art is art in virtue of its relation to previous Martian art, and ultimately
to Martian ur-art (ArtWorks, 108).
As Stecker puts it, But why should [the assumed Martians] complain that they cannot apply
our concept of art to their art? They would be no more able to do this than would ancient Egyptians
or Greeks be able to apply our concept to their art (ArtWorks, 1078).
 The Irreducible Historicality of the Concept of Art 21

in the specific actual-history-reflective sense I claim is now ours. But as we also

saw, they might very well, in virtue of knowing (a), have known that what
they practiced was *art*. In other words, they could have known that they
were making in art in a stripped-down, form-and-function-based sense not
equivalent to the sense we operate with at present.
Another, more recent, version of the anthropocentricity objection against
historical theories of art goes as follows: their account of an art-historical
relation is insufficiently projectable: there could be art objects which are recog-
nizable as such, but which stand in no art-historically significant relation at all
to any of our art. But it is not that a tradition of object-making must already
stand in an art-historically significant relation to something in our tradition of
art in order to be an art-making tradition, but rather that, insofar as anything
outside our art tradition is properly said to fall under our concept of art, it is
because we can appropriately relate it to our tradition of art, and in particular
to the normative regards that have, as a contingent matter of fact, emerged in
that tradition.
Consider finally the hyperbolic question: could there have been art a mil-
lion years ago, on a planet of the star Betelgeuse, if human history had never
occurred at all? Well, yes and no. The answer is no, I think, if one means art in
the specific actual-history-conditioned sense it now has, in the early twenty-
first century. The answer is yes, I think, if one just means objects made in
certain ways, for certain kinds of reception, all intrinsically defined. But that,
though it may once have served as our concept of art, and perhaps as recently
as a century ago, no longer does.

The Objection from the Multiplicity of Art Traditions23

This objection, originally voiced by Stephen Davies, has been concisely for-
mulated by Robert Stecker:
Davies admits that historical definitions of art explain how something is an artwork
by relating it to a given tradition. However, he claims that such definitions will
be incomplete until a basis is provided for distinguishing art traditions from other
historically continuous cultural processes or practices. If there are different art tradi-
tions, and if something is an artwork only in relation to some of these traditions, then

Berys Gaut, Art As A Cluster Concept, in N. Carroll (ed.), Theories of Art Today (Madison:
University of Wisconsin Press, 2000), 36.
Davies, Non-Western Art and Arts Definition, in N. Carroll (ed.), Theories of Art Today.
22 Art

the explanation of why something is an artwork will not be complete without some
account of what makes something an art tradition.
And here is the point in Daviess own words: A definition that character-
izes art making as artworld-relative and that also concedes the existence of
autonomous artworlds must explain how artworlds are of a single type. An
account is required of what makes the various artworlds artworlds.
As it turns out, Davies is sympathetic to the extension of historical definitions
of art to accommodate non-Western art, provided one can say what makes
a given practice of object-projection characterized by backward intentional
invocation of predecessors an art practice, rather than something else. Davies
suggests that this is to be done by recognizing the essentiality of aesthetic interests
and regards, at least in initial stages, to any practice that might be accounted
artistic, however far it may have diverged from its original aesthetic roots.
This is a plausible proposal as to what, in surveying the known art traditions
of the world, makes them all art traditions rather than internally historical
traditions of some other sort of making. But whereas Davies views the neces-
sary aesthetic origins of any tradition that can be recognized as artistic as a
ground-floor fact, arrived at after reflection on both the ubiquity of art in
human culture and the universality of aesthetic interests at the beginning of
art traditions, I would be inclined to historicize further, and so render more
contingent, the role of the aesthetic in the characterization of artistic practices.
What I want to say is, yes, possibly anything we would recognize as an
artworld or an artistic practice would display aesthetic concerns, at least at
its origins, but that is because aesthetic concerns emerged centrally and persisted
centrally for thousands of years in the Western tradition of artmaking. In other
words, it is neither a strange accident nor a conceptual truth that anything
we recognize as an art practice will have involved cultivation of and attention
to humanly bestowed aesthetic features of things; it is just that those are the
concerns that were original to and for ages almost uncontestedly dominant in
what we unreflectively know as our practice of art.
The Objection from the Putative Two-Dimensional Semantics
of the Term Art26
It has been claimed by Gregory Currie that if art is to be a historical concept
in the relevant sense, that is, one in which the very identity of our concept
ArtWorks, 108. Davies, Non-Western Art and Arts Definition, 212.
Gregory Currie, A Note on Art and Historical Concepts, British Journal of Aesthetics 40
(2000): 18690.
 The Irreducible Historicality of the Concept of Art 23

of art depends on history, then it will have to be a concept (a) that is our
actual concept, (b) where its being our concept depends on some contin-
gent matter of history, (c) where some other concept or concepts might, if
history had gone differently, been our concept of art, and (d) where there is
some overarching concept that binds them all together, showing that they
are all, indeed, concepts of art. Currie further claims that this is the conceptu-
al structure exhibited by natural kind concepts such as
water, and thus that

art, if it is to be historical, must exhibit the same general feature as natur-
al kind concepts, even though it is not thereby claimed that art is a natural
kind.
The general feature in question, as the Putnamian analysis of water sug-
gests, is a two-dimensional semantics in which the meaning of a term depends
on two things, a qualitative, purely observable notion like that of waterish
stuff (i.e. stuff resembling water) in the case of water, and certain contin-
gent facts about what in a given world constitutes the stuff in question, such
facts being, in the case of water and the actual world, that the waterish stuff
of our acquaintance is H2 O, or hydrogen oxide.
Thus, as applied to art, this analysis claims that if art is to be simil-
arly historical there must be a qualitative, purely observable covering concept
like artish thing, and then certain contingent facts about what things actu-
ally fall under that concept in a given world, in order for the meaning of
art to be fixed in that world. But if so, then there is no avoiding the observ-
able concept artishthe concept is presupposed in making out the putatively
two-dimensional character of the concept of art and, the objection con-
cludes, there is no reason to think the concept of art we actually use is other
than that of artish thing, thus undermining the rationale for a historical ana-
lysis of arthood.
At bottom, I think this critique of Curries mistakes its target, taking the
intentional-historical theory of art to be a claim about the dependence of the
concept of art on art histories in other possible worlds, whereas the claim is
rather that the concept of art is such that what can be art at a given time in a
world logically depends on what is already art at that time in that world, since
artmaking minimally involves an agents intentionally relating a proposed
object to the body of already existing art. In other words, what is maintained
is the dependence of the possible extension of art at t on the actual extension
of art prior to t, but not the dependence of our present concept of art on the

Ibid. 187. Ibid.

24 Art

actual path of art history. So far as I can see the intentional-historical theory
of arthood is not committed to the claim that our present, minimalist and
intentionalist, concept of art would have been different if actual art history had
been different. I want to say that the concept of art with which we presently
operate, if historical in the way I have in mind, is not different in different
possible worlds. It is rather, in any world, the concept of an object-identifying
practice where what can count as art at t depends on what, contingently,
already counts as art prior to t.
The variation in the possible extension of art from world to world occa-
sioned by the contingencies of what has already fallen under the term art at a
given time, to which the intentional-historical theory is committed, does not
seem to me modellable on the two-dimensional meaning account apt for nat-
ural kind terms, which involves an implicitly referenced underlying nature.
The concept of art, I suggest, is structurally unlike that of water.
Let us look more closely at the disanalogy between water and art. In the
case of water, assuming Putnams account of natural kinds, when we consider
other possible worlds where persons have, let us say, the same basic idea of
water as we do (i.e. watery stuff), we do not count as water everything they
count as water; we only count something in that possible world as water if it
has the same underlying structure or material composition as our water.
In the case of art, assuming my account of arthood, it is true that in
other possible worlds persons will be counting different things as art than we
do, given the contingent development of their art historymost obviously,
there may be objects in that world that do not exist in ours. But when we
consider those possible worlds from this one, I claim, we have every reason
to count or acknowledge as art what they count or acknowledge as art, that
is, things intended for regard the way earlier presumed art in that world was
correctly regarded. It is perhaps true that we can only identify what is their art
practiceas opposed to other practices they might have that hold togeth-
er through reflexive-retrospective intentionsby noting that it is the one
where the normative regards invoked in the reflexive-retrospective intentions
are those which contingently emerged as such in our art practicee.g. aes-
thetic, formal, expressive, communicative ones. This is the key way in which
art as we now deploy it is anchored in the contingency of what is our own
art history. Still, it seems true that we will count as art in that world what they
count as art in that world.

In particular, what I above labeled extended intentional-historical theory.

 The Irreducible Historicality of the Concept of Art 25

Thus the contrast with how water works remains. Given what we mean
by water, where underlying constitution is essential, what persons on Twin-
Earthwhere XYZ, not H2 O, fills the water rolethink of as water,
nevertheless isnt water. But given what I maintain we now mean by art,
and have meant by art for at least half a century, then if persons in another
possible world think of something as artthat is, they label things that way
within a practice we can recognize as an art practice because of its structural
(backward-looking reflexive intentions) and substantive (involving the sort
of normative art regards, ones it is correct to bring to such objects) parallels
with our contingently developed art practicethen those things are art, albeit
art in that world, even by our lights. Thus art, even if historical in the
way I propose, is not plausibly regarded as a two-dimensional concept a la
water.
Shorn of technicalities, the gist of Curries critique is that a concept
operative in a given world and dependent on the contingent history of that
world regarding what falls under it cannot be shown to be a concept of art
unless a qualitative notion like artish is assumed to operate in conjunction
with the first concept. More concretely, the implication seems to be that the
backward-looking act of reference posited by the theory, which by linking
present objects to past artworks makes of the former artworks as well, must
involve a content on the order of preexisting artish things, thus undermining
the theorys pretension to outline a purely historical, non-qualitative concept
of art.
But that is not so. Intentional reference of the art-making sort need not be
secured by appeal to an independent, qualitative notion of the artish. Rather,
it can be secured purely demonstratively, as by a speech act or thought of form
as those things are properly regarded, where the things demonstrated are in
fact artworks, or else via paradigms, as by a speech act or thought of form
as Beethovens 5th Symphony and the like are properly regarded. In other
words, on the theory of arthood I espouse the facticity of arthood, including
not only what things are art but what ways of regarding them are normative
for them, goes all the way down. Contra Currie, then, the intentional-
historical theory of art need not fall back self-defeatingly on a quasi-observable
notion such as that of the artish. On my theory, a qualitative notion like artish
plays no role in how, in any given world, the future extension of art necessar-
ily relates to its past extension.

Or at the least, all the way to the ur-arts, whatever those turn out to be.
26 Art

III ARE ALL ARTIFACT CONCEPTS ESSENTIALLY

HISTORICAL?

The psychologist Paul Bloom has proposed extending the intentional-

historical theory of artworks to cover artifacts of all kinds. Bloom proposes,
roughly, that for any artifact kind K, to be a K is to be successfully intended
to be a K, where what it is to be a K in the latter context is given, inevitably,
by past instances of K. Bloom is persuasive in pointing up the superiority of
his proposal to existing competitors, those which analyze artifact concepts in
terms of necessary and sufficient conditions, family resemblances, character-
istic functions, or prototypes.
But if Bloom is right, then what remains of the special historicality of the
concept of art as opposed, say, to those of chair, pencil, or house? Two things,
it seems.
First, it should be observed that on Blooms analysis something is a K in
virtue of being intentionally related in the right way to preceding Ks generally.
But something can be an artwork, on my analysis, through being intention-
ally connected in the right fashion to some particular past artwork or art-
worksthe history of art thus enters more concretely into what art is and
can be at any point than does the history of a given artifact kind into what is or
can be an instance of that kind at any point. To know that something was art
thus might require tracing relations to a particular episode or domain of arts
history, but nothing comparable would seem to be required to establish that a
candidate chair, pencil, or house was indeed a bona fide instance thereof.
Second, it is arguable that artifact concepts, in contrast to that of artwork,
usually retain at least rough necessary conditions as regards either form or
function; for example, a chair must exhibit shape within a given broadly
circumscribed rangecertain shapes, e.g. that of a javelin, would seem to be
excludedand must answer to or have been designed to answer to a certain
purpose or useto wit, that of being sat upon. But that is not the case
with the current concept of artwork, which, if I am right, and in contrast to
perhaps every other artifact concept, retains only certain purely historical and
intentional necessary conditions.

See P. Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996): 129.
 2
Artworks as Artifacts

I T HE INTENTIONAL-HISTORICAL CONCEPTION
OF ART

What kind of artifact is an artwork? The answer to that question depends,

clearly enough, on the conception of art that one is inclined to adopt. Past
conceptions of art, according to which art was essentially a mode of represent-
ation, or a vehicle of emotional expression, or a display of skill in fashioning,
or an exploration of form as such, or the pursuit of the beautiful, no longer
seem remotely adequate to the nature and range of what have been accoun-
ted artworks in the past hundred years or so. The abstract canvases of Kupka
and Kandinsky are almost a century old; John Cages aleatoric music of the
1960s seems devoid of emotional expression; Robert Rauschenbergs Erased
De Kooning Drawing displayed no notable skill; Tolstoys The Death of Ivan
Ilych does not strike one for its exploration of form; and Francis Bacons tor-
tured portraits are anything but beautiful. Accommodating arts development
since the nineteenth century seems to call for a more circumspect approach,
one that is noncommittal as regards medium, style, form, content, and artistic
objectives.
I have defended a conception of art of this sort, one along intentional-
historical lines, according to which something is art in virtue of being gov-
erned by certain intentions with an essential historical, or backward-looking,
content. More specifically, what I claim is that an artwork, in the current
understanding of the term, is something that has been intended by someone
for regard or treatment in some overall way that some earlier or preexisting

First published in E. Margolis and S. Laurence (eds.), Creations of the Mind (Oxford: Oxford
University Press, 2006).
28 Art

artwork or artworks are or were correctly regarded or treated. The artmak-

ing intention involved may be either of an opaque sort, having roughly the
content just expressed, one that simply references prior art as such, or of a
transparent sort, invoking specific ways of regarding or treating objects that,
as a matter of fact, and whether known to the agent or not, figure in the set
of correct ways of regard or treatment for earlier or preexisting artworks. In
either mode of artmaking, the concrete history of artmaking up to a given
time is thus ineliminably implicated, in whole or in part, in any artmaking
undertaken at that time.
This conception of arthood has obvious points in common with the
art-theoretical and social-institutional conceptions of arthood elaborated
earlier by the philosophers Arthur Danto and George Dickie. Like those
conceptions, it looks for a relational, situational, or contextual defining
feature of art, rather than a formal, intrinsic, or perceivable one. All three
conceptions have their roots in the enforced revision of traditional ideas about
art that was effected by certain revolutionary ventures in artmaking in the
early and middle twentieth century, notably those of the Dadaists, Marcel
Duchamp, Andy Warhol, Jasper Johns, Robert Rauschenberg, and others. By
appropriating, reframing, reconfiguring, and reprojecting as art any number
of things theretofore assumed to lie outside of the ambit of art, artists such as
the preceding managed to establishsince their ventures must be regarded,
at least from our present vantage point, as undeniably successfulthat more
or less any object could be made into or could become a work of art, if suitably
repositioned, reconceived, or, in Dantos famous term, transfigured. Among
the objects that were thus transfigured into art in those years, with little or
no physical alteration or manipulation, were the following: a urinal, a snow
shovel, a bottle rack, a beer can, a coffee cup, a disordered bed, and a postcard
reproduction of the Mona Lisa. It no longer seemed necessary, then, that an
artwork be fashioned by its maker with technical skill, that it make use of
traditional materials in its construction, that it display form of any notable
complexity, that it have any obvious aesthetic appeal, or that it inevitably
reflect in its handling the individual personality of the artist. It is difficult to
deny that the concept of art that emerged in the wake of those developments,
now almost a century old, was an altered and notably broadened one, covering

See Defining Art Historically, British Journal of Aesthetics 19 (1979): 23250, and Refining
Art Historically, Journal of Aesthetics and Art Criticism 47 (1989): 2133, reprinted in Music, Art,
and Metaphysics (Ithaca: Cornell University Press, 1990).
 Artworks as Artifacts 29

all that had previously been recognized as art, to be sure, but much that would
not have been recognized as art under the traditional concept that had held
sway, with only minor modifications, since at least the Renaissance.
The intentional-historical conception of art differs from the art-theoretical
and social-institutional ones, though, in positing as the crucial contextual
condition of arthood not a relation to some prevailing artistic theory, nor a
relation to a surrounding social institution, but a relation to the concrete his-
tory of artmaking and art-projection into which the candidate object hopes to
enter. The intentional-historical conception differs also from its contextualist
predecessors in taking its most direct inspiration not so much from the ready-
made and appropriational modes of artmaking that had been established by
Duchamp and others, but from the subsequent, more radical activities of
Conceptual artistssuch as Robert Barry, Robert Morris, John Baldessari,
Joseph Kosuth, Sol LeWitt, Vito Acconciwhich seemed to establish that
art per se had no need even of any concrete object, whether appropriated,
readymade, or fashioned from scratch, but could apparently consist merely in
concepts, words, statements, gestures, thoughts, and the like, with the appar-
ent consequence that anything, or at least anything thinkable, demonstrable,
or designatable, of whatever metaphysical or logical sort, could be, or at least
could become, a work of art.
Not surprisingly, the intentional-historical conception of art has elicited a
certain number of critiques, turning on such issues as the apparent circularity
of such a conception, the status of first art on such a conception, the extend-
ability of the conception to cultures or histories other than our own, and the
problematic recursiveness of the procedure for identifying objects as art that
the conception appears to entail. I have addressed those critiques elsewhere, if
perhaps not to the satisfaction of all, so will not address them again here. I
will instead simply assume that the intentional-historical conception is more

The ontology of Conceptual Art is not as simple as it seems. In particular, the identity of a
work of Conceptual Art cannot be equated with that of the concepts it invokes or deploys. There
is arguably always something concrete involved in the making of a work of Conceptual Art, and in
which its identity as that artwork, of that artist, is anchored. This applies even to one of the most
emblematic of Conceptual artworks, Robert Barrys All the things I know but of which I am not at
the moment thinking: 1:36 p.m., 15 June 1969, New York. For this work was created by a particular
individual at a particular time and place, through a concrete act of articulation, one embodied in
some particular physical inscription, to which its identity as an artwork would seem to be tied.
See Extending Art Historically, Journal of Aesthetics and Art Criticism 51 (1988): 41123,
reprinted in The Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996), and The Irreducible
Historicality of the Concept of Art, British Journal of Aesthetics 42 (2002): 36779 (reprinted in
this volume).
30 Art

or less adequate to what it now is to be an artwork, in the most comprehensive

sense, in order to ask what that implies for the status of artworks as artifacts,
and for the extent to which the artifactuality of artworks differs, if at all, from
the artifactuality of artifacts in general.

II ART WORKS VERSUS OT HER ARTEFACTS

The psychologist Paul Bloom has attempted to extend the intentional-

historical theory of artworks so as to cover artifacts of all kinds. Bloom
proposes that for any artifact kind X, to be an X is to be an object successfully
created with the intent that it be an X, where what it is to be an X at a given time
is informed inescapably by past instances of X. Blooms insight is thus that
all artifact concepts, and not just that of artwork, have an essential historical
component, so that the past deployment of such concepts ineluctably enters
into their present and future deployment, through the backward-directed
intentions that the makers of such artifacts must of necessity possess. Blooms
explicit statement of his proposal is as follows: We construe the extension of
an artifact kind X to be those entities that have been successfully created with
the intention that they belong to the same kind as current and previous Xs.
Bloom is persuasive in pointing up the superiority of his proposal to existing
competitors, those which analyze artifact concepts in terms of necessary-
and-sufficient conditions, family resemblances, characteristic functions, or
prototypes. But it remains to be seen whether Blooms own, original and
sweeping, proposal is entirely acceptable.
Suppose for the moment that Bloom is right, and that an analysis of the
sort that captures what it is to be an artwork also captures what it is to be an
artifact of any sort. What, if anything, would remain of the special historic-
ality of the concept of artwork, as opposed to those of chair, pencil, house,
or other standard artifacts? Two things, it seems. First, on Blooms analys-
is something is an X in virtue of being intentionally related in the right way
to preceding Xs generally. But on the intentional-historical analysis of art-
hood, something can be an artwork through being intentionally connected
in the right manner to a particular past artwork or artworks, whether or not
intentionally connected to past art invoked generally. For example, someone
could make an artwork of a sculptural sort by assembling pieces of wood

See Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996): 129.
Ibid. 10.
 Artworks as Artifacts 31

and plastic with the intention of the assemblage being regarded in an over-
all manner appropriate to Henry Moores Reclining Nude, but without any
intent explicitly invoking the category of art or even the subcategory of sculp-
ture. The history of art, it appears, enters more concretely into what can be
art at a given point in time than the history of a given artifact kind enters
into what can be an instance of that kind at that time. So to establish that
something was an artwork might require tracing intentional relations to a par-
ticular item or episode in the history of art, but nothing comparable would
seem to be required to establish that a candidate chair, say, was an instance of
that kind.
Second, it is arguable that standard artifact concepts, in contrast to that of
artwork, retain at least some necessary conditions as regards form or function,
whatever the historical dimension of their correct deployment. For instance,
a chair must exhibit shape within a given broadly circumscribed range, with
certain shapes being excluded in advance. And a chair must answer to a cer-
tain purpose in the case of chairs, that of being sat upon with some degree of
comfort, or at the very least, be aimed at answering to such purpose.
But I maintain that that is not the case with artworks as such, which in
contrast to perhaps every other sort of artifact, retain only certain purely
intentional-historical necessary conditions. In other words, nothing can be
declared a failed artwork, in the sense of not succeeding in being an artwork
at all, through failing to display a certain broadly specified form or a partic-
ular sort of functionality. But something can be declared a failed chair, in
the sense of not even being a chair, if shaped like a javelin or if incapable
of being sat upon at all. Thus even were an intentional-historical account
of artifacts in general to be accepted, artwork would remain distinctively
historical, in contrast with other artifact concepts, in respect of the cre-
ation involved requiring only the satisfaction of certain intentional-historical
conditions.
It is difficult to say whether these differences between artworks and
other artifacts, which amount to the latter being less purely intentionally-
historically determined, would be contested by Bloom. And that is because
of the specific way he formulates his intentional-historical theory of artifacts,
in which crucial appeal is made to the action of successfully creating an X.
What is it, though, to successfully create an X? Does successfully creating an
X differ from simply creating an X? If so, it should be possible to create an
X, but unsuccessfully, which is not, I think, what Bloom is after. So successfully
creating an X, it would appear, is just creating an X. What the adverb serves
32 Art

to call attention to, however, is a minimal success condition that Bloom

apparently regards as constraining intentional-historical artifact making in
general. This is evident in his illustration of how chair-making, for example,
might fail, even though the intentional-historical condition of such making
was in place: If someone intends to create a chair, but it falls to pieces as
soon as it is finished, the person would not view this creation as successfully
fulfilling his or her intent, and thus has not created a chair.
Thus for Bloom, something intended to be a chair but that was merely a
heap of materials incapable of being sat upon would not be a chair regardless
of how firm the intention involved that it belong to the category of chairs.
That seems right, but Blooms justification of this judgment is rather peculiar.
He implies that such an object would fail to be a chair not because it could
not fulfill the basic function of affording single seating, but rather because
its creator would not recognize it as the successful product of an intention to
create a chair, that is, something effectively affording single seating. This is
peculiar because it seems that whether or not something counts as a chair,
though it may depend crucially on the intentions of its maker, should not
ultimately depend on whether from the point of view of its maker those inten-
tions are fulfilled, but rather on whether, from some objective point of view,
those intentions really are fulfilled. For after all, a would-be chair maker may
be deluded or confused, thinking that a pile of nails or a coiled length of rope
for which he is responsible conforms well enough to past chairs to count as a
successfully created new one.
It seems that what is relevant to satisfaction of the minimal success condi-
tion is not the makers conception of a chair based on past acquaintance with
them, but rather the conception of a chair endorsed by competent users of
the term chair in general, one that imports at least some minimal features of
form or function. Given satisfaction of that condition, the identity of a can-
didate thing as a chair may indeed be entirely determined, as Blooms account
would have it, by an appropriate chair-history-invoking intention in its mak-
ing. But the insistence on a minimal success condition, which rules out piles
of nails, lengths of rope, decks of cards, javelin-shaped rods, and so on, as
chairs, shows that some non-purely-historical conception of chairhood is in
play in circumscribing the boundaries of the category. As regards artworks,
however, it is far from clear that any such non-purely-historical conception of

See Bloom, Intention, History, and Artifact Concepts, Cognition 60 (1996): 129.
 Artworks as Artifacts 33

arthood is in play, or that there is any minimal success condition of a substant-

ive sort on the making involved.
Why is that so? It helps to recall what, in the Post-Duchampian era, are two
salient features of artmaking in contrast to standard artifact-making. First,
one can more or less simply declare something a work of art, and it becomes
such. Or at least one can, in certain contexts, or with a certain standing, do
so. Second, anything, whatever its material constitution, cultural category, or
ontological status, can become, or can be incorporated into, a work of art.
These features are arguably enough to distinguish the concept of artwork
from that of other artifacts, even if such artifacts, if Bloom is right, share with
artworks the primary determining of their categorial status by a historical, or
past-invoking, intention in their making. What is special about the artifact
concept artwork, one might say, is that it is a wholly relational one; it is more
like those of observed thing or beloved object or prize winner than it is like
those of standard artifacts, such as chair or cup or cabin, for which there are
at least minimal conditions of form as regards finished shape, of constitution
as regards material, of making as regards the activity of the maker, or of
functional success as regards usability of the final product.

III ARTMAKING AND SUBSTANTIVE CONCEPTIONS

If the above is correct, then however sound the inspiration of Blooms

intentional-historical theory of artifacts, he errs in blurring the difference
between artworks and other artifacts, failing to appreciate that though
minimal success conditions, rooted in some not-purely-relational conception
of the kind of artifact in question, are ineliminably involved in the making of
the latter, that is not the case with the former.
Amie Thomasson, in a careful essay highlighting the insufficiently ac-
knowledged role that background conceptions of artifacts play in their cre-
ation, holds Blooms analysis at fault precisely for not sufficiently
acknowledging that role, and for thinking that artifact creation can proceed in
a more conceptually thin or purely historical way than it in fact can. Accord-
ing to Thomasson, even the intention to create something of a given artifact
kind K cannot consist merely in intending the object to belong to the same

See Thomasson, Artifacts and Other Creations, in E. Margolis and S. Laurence (eds.),
Creations of the Mind (Oxford: Oxford University Press, 2006).
34 Art

kind as existing examples of K to which one can point or refer. Thomasson

instead maintains, as a completely general principle, that a substantive concep-
tion of an artifact kind K must be involved in the intentional production of an
artifact of that kind: the relevant sort of intention to make a thing of artifac-
tual kind K must thus involve a substantive, and substantially correct, concept
of what a K is, including an understanding of what sorts of properties are K-
relevant and an intention to realize many of them in the object created.
She adds that, for an artifact of that kind to be created, the intention in
question must be largely successfully realized, which is the minimal success
condition implicit also in Blooms account of artifactuality.
Assuming that Thomassons principle is true for the making of garden-
variety artifacts, is it also true for the making of artworks, that most
elusive species of artifact? Can one create an artwork without a substantive
conception of what artworks are? More specifically, does artmaking on
an intentional-historical account of it require that an artmaker have a
substantive conception of what he or she is making?
The answer depends, in part, on how substantive is substantive. Can one
make an artwork merely by intending something for the sort of regard or
treatment appropriate to artworks, but without knowing what artworks are,
in any qualitative sense, but only that there are such things, and that they are
some sort of artifact, and without knowing what sorts of regards or treatments
are appropriate to them, but only that there are such? I claim one can, and if
so, one neednt have a substantive concept of what an artwork is, one implic-
ating characteristic properties or functions. Does one need to possess in any
measure a theory of art, a la Danto, in order to make art, or need one only
know that there are such things as artworks and that there are ways it is correct
to approach them? I claim not, and if so, once again an artmaker need not pos-
sess a substantive concept of what an artwork isthough of course virtually
all artmakers will possess such, which concept will vary from artist to artist,
and from artform to artform.
Elsewhere in her discussion Thomasson offers an argument that could be
seen as directly aimed at undermining the possibility just affirmed, one that
according to the intentional-historical theory is sometimes realized in the
making of art, that an artifact of kind K might be made merely by intending
an object to stand in certain relations to existing instances of K:

See Thomasson, Artifacts and Other Creations, in E. Margolis and S. Laurence (eds.),
Creations of the Mind (Oxford: Oxford University Press, 2006).
 Artworks as Artifacts 35

If a later artisan does know of the existence of extant Ks, then he may have the inten-
tion to create one of those, with an implicit reference back to earlier K. Yet even so
. . . his intention cannot be a mere transparent intention to create one of those . . .
without any substantive concept of what those are, of what features are relevant to
being of artifactual kind K. In order for his making to be controlled and directed,
his intention to make a K must be filled out with intentions regarding what features
are to be imposed on the object of his creation in order to succeed at realizing his
intention to make a K.
But I think it clear that the stricture Thomasson here invokes, of a mak-
ing filled out with feature-conscious or feature-directed subsidiary intentions,
whose satisfaction is necessary for the making to succeed, though applicable
to the making of standard non-art artifacts, such as chairs, as well as tradi-
tional art artifacts in established media, such as paintings, is inapplicable to
artmaking in an appropriational or conceptual mode. Arguably nothing more
is needed for successful artmaking in that mode than the belief that there is
a practice of art, that various things are exemplars of it, and that there are
correct ways of regarding, treating, or interacting with those things. So far
as I can see, this necessarily involves the maker in some conceptions about
art, to be sure, but not in substantive conceptions, in the sense Thomasson
seems to have in mind, about the nature of artworks and their characteristic
properties.

IV TRADITIONAL ARTMAKING

It would be remiss to end this short essay on the nature of artworks as artifacts
without some remarks on the special character of artmaking of a traditional
sort, that is, all artmaking before Duchamp, Warhol, and Conceptual
Art, and most artmaking after them as well. In artmaking as traditionally
conceivedand for simplicity I confine my attention to artmaking in the
visual artsthere are distinctive raw materials, e.g. paint, clay, charcoal;
there are distinctive techniques, e.g. carving, etching, impasto; and there
are distinctive aims, such as visual beauty, representational verisimilitude,
and emotional expression. But the making of chairs and pencils also
involves distinctive materials, distinctive techniques, and distinctive aims,
albeit utilitarian ones. So even if artmaking in the comprehensive, post-
Duchampian sense distinguishes itself from other sorts of artifact-making

Ibid.
36 Art

by its presumed purely intentional-historical character, is artmaking of the

traditional sort, though issuing in a physical object whose interest is primarily
aesthetic rather than utilitarian, fundamentally different from the making of
physical artifacts generally, including the making of craft objects such as rugs
or pots? To a degree.
Dewey and Collingwood were two philosophers of art who had insight-
ful and consonant things to say on the distinctive character of the making
involved in traditional art, especially as in contrast with the making involved
in the overtly similar activity of craft. What both thinkers stressed is that the
making of an artwork is an open-ended, indefinitely extended creative-critical
process, with alternating phases of making and assessing, or doing and under-
going, but one not governed by any fixed goal or preconceived idea of what
the artwork must be, or how it must turn out. An artist making a sculpture,
for example, in contrast with a craftsman making a rug or a pot, need not
envisage what its dimensions will be, what it will look like, or what form it
will have. This is unsurprising if one recalls that making a traditional artwork
is, as much as anything, an expressive activity, but one in which, as Colling-
wood underlined, the artist does not know precisely what he has expressed
until the process is completed. The maker of a craft object, though, must first
and foremost assure the creation of a usable object of the craft in question,
some of whose features, such as flatness or water-holding capacity, are accord-
ingly non-negotiable, thus enjoining a preconception of some specificity on
the craftsmans part of the object to be created.
Granted the above, the upshot for our discussion is this. If, as Thomasson
urges, the making of standard artifacts is always governed by a substantive
conception of the artifact in question, one that sets clear terms for success and
failure in such makings, and if traditional artworks are accounted standard
enough in that respect, then what is most noteworthy about the making of
such artifacts is that the substantive conceptions involved in their creation
are relatively insubstantial, that is, not such as to constrain notably them in
formal, material, or functional ways. A sculpture, say, needs to be physical,
perceivable, and perhaps smaller than the planet, but apart from that, it can be
of any size, any composition, any shape, any color, and any subject. The rel-
ative insubstantiality of the conceptions governing the making of traditional

See John Dewey, Art as Experience (New York: G. P. Putnam, 1934), and R. G. Collingwood,
The Principles of Art (Oxford: Oxford University Press, 1938).
 Artworks as Artifacts 37

artworks fits well, of course, with the innovative and exploratory aim often
ascribed to art, both traditional and non-traditional.

V CONCLUSION

So what sort of artifact is an artwork? In the past, and thinking primarily of

the visual arts, one might have answered: a physical object, fashioned with
skill, involving a recognized medium, designed to be of aesthetic interest, and
whose making is governed by a fairly substantive conception of the genre of
artwork in question. And such an answer would still be largely adequate to at
least traditional artmaking today. But at present, and just confining ourselves
to the activities of visual artists, such an answer is no longer even remotely
adequate. That is because of alternate modes of artmaking that have become
entrenched in the past hundred years, whereby artworks need not be fash-
ioned by their creators, need not involve recognized artistic media, need not
be aimed at satisfying aesthetic interests, and whose making need not be gov-
erned by any very substantial conception of a genre in which the artist is
working. Those modes of artmaking have revolutionized the concept of art,
making it the case that the concept-of-art-2005 is something fundamentally,
and not just marginally, different from the concept-of-art-1905. Artworks
are necessarily artifacts, since they are things intentionally brought into being
through human agency. That much remains true. But if I am right, to be an
artwork today is simply to be something governed by an intention relating it
in a certain way to what have been accounted artworks in the past. By con-
trast, more is required to be an artifact of a standard sort, such as a chair, even
if the intentional-historical connection sufficient for being an artwork plays a
crucial role there as well, in the manner that Bloom has underlined.
 3
Emotion in Response to Art

Responding emotionally to artworks is a familiar enough occurrence, and

hardly seems puzzling, recalled at that level of generality. Why should not
works of art, in company with people, animals, natural objects and political
events, produce emotions in us? Philosophers have, however, raised questions
about emotional responses to art in particular contexts, or when viewed from
certain angles. These questions suggest that there is indeed something puzz-
ling about such responses.
One such context is that of response to fictions, whether literary, dramatic,
or cinematic ones, where emotions appear to be had, not only for the work
or representation itself, but for the fictional characters or situations represen-
ted therein, even though these are perfectly well understood not to exist. A
second such context is that of abstract or non-representational art, with music
the example par excellence, where it is unclear both what could elicit such a
response and what its object could be. A third context is that in which art-
works expressive of negative emotion, for example tragedies, requiems, and
tales of horror, engender parallel responses in perceivers without evoking
avoidance or disapproval. And a fourth context in which emotional response
to art has struck philosophers as problematic is where the proper appreciation
of art is at issue, since such appreciation may be thought to be incompatible
with experiencing life emotions of a familiar sort that art seems capable of
raising in us.
We might formulate the main philosophical questions concerning emo-
tion in response to art as follows. (1) What kind of emotions are had in
response to works of art? (2) How can we coherently have emotions for fic-
tional persons or situations, given that we do not believe in their existence?
(This query relates to what is known as the paradox of fiction.) (3) How

First published as Emotion in Response to Art: A Survey of the Terrain, in M. Hjort and S. Laver
(eds.), Emotion and the Arts (Oxford: Oxford University Press, 1997), 2034.
 Emotion in Response to Art 39

do abstract works of art, especially musical ones, generate emotions in audi-

ences, and toward what do audiences then have these emotions? (4) How can
we make sense of the interest that many appreciators have in experiencing
empathetically art that is expressive of negative emotions? (A particular form
of this query is the paradox of tragedy.) (5) Is there a conflict between
responding emotionally to art and what the aesthetic appreciation of art
demands?
Answers to these questions depend, to some extent, on the conception of
emotion adopted. I thus begin this essay by sketching a conception of the
emotions that mediates between the sensationalist model of early twentieth-
century psychology and the cognitivist model widely favored in current philo-
sophy. With that as background, various answers to the above questions will
be critically reviewed.

I T HE NATURE OF EMOTIONS

In order to fruitfully assess the varieties of emotional response to art it is

obviously of use to have some account of what exactly emotionsin the
occurrent, as opposed to dispositional, senseare. Philosophical debate on
the nature of emotions, informed to greater or lesser degree by available
work in psychology, has in the past thirty years or so revolved around an
opposition between feeling (or sensation) based, and thought (or cognition)
based, approaches. The former holds that at the core of an emotion is an
internal feeling or set of sensations, while the latter holds that at its core an
emotion is a particular kind of thought, judgment, or evaluation. While the
feeling approach has trouble accommodating the intentionality (or object-
directedness) and amenability to reason of many emotions, the thought
approach has trouble with the experiential aspect of emotions, that is, with
what it is to feel them, as opposed to merely having the beliefs or entertaining
the thoughts that may be associated with them, with the evident inertia and
passivity of many emotional conditions, as well as with states of desire, whose

See, for example, among major studies in a cognitivist vein, Robert Solomon, The Passions
(Garden City, NY: Anchor Books, 1978); William Lyons, Emotion (Cambridge: Cambridge
University Press, 1980); Robert Gordon, The Structure of Emotions (Cambridge: Cambridge
University Press, 1987); Ronald De Sousa, The Rationality of Emotion (Cambridge, MA: MIT Press,
1987); and Patricia Greenspan, Emotions and Reasons (London: Routledge, 1989). A useful review
of this literature, to which I am indebted, is John Deigh, Cognitivism in the Theory of Emotions,
Ethics 104 (1994): 82454. An even more extreme cognitivism about emotions is presented in
Martha Nussbaums recent Upheavals of Thought (Cambridge: Cambridge University Press, 2001).
40 Art

connection with many emotions seems more than contingent. Still, while the
feeling approach can be faulted for too mindless a picture of emotions, it
is right to insist on bodily response and inner affect of some sort as a sine
qua non of emotion, and while the thought approach can be faulted for too
mindful a picture of emotions, it is right to emphasize that many emotions
include cognitive elements essentially, e.g. thoughts with specific contents,
which contents are, in many cases, socially shaped.
At present there appears to be some consensus that, in perhaps the
majority of cases, an emotion is best thought of as a bodily response with a
distinctive physiological, phenomenological, and expressive profile, one that
serves to focus attention in a given direction, and which involves cognition
to varying degrees and at various levels. The level of cognitive involvement
runs from mere registering of presence, to ways of seeing or regarding that
which is registered, to propositional conceptions of the object responded
to, to articulate beliefs about or attitudes toward the object of response.
Alternatively put, an experienced emotion can be said to have as its core
a bodily reactioncomprising physiological sensations, feelings of comfort
and discomfort, and orientings of attentionwhich reaction is often caused
or modified by, and is sometimes necessarily bound up with, cognitions of
various sorts and strengths, depending on the type of emotion involved. Note
that on such a view of emotion, in which cognitive representations on the
order of beliefs (or for that matter, desires) are seen as characteristic of, but
not essential to, experienced emotion, the intentionality or directedness of
emotions (as opposed, say, to moods) is preserved by the root feature of
orientation of attention to or focusing of concern on that which the subject
registers as significant.
On the other hand, there is also a growing acknowledgement that the pre-
theoretically recognized emotions constitute an irreducibly heterogeneous
class, i.e. that they do not form a natural kind. It seems reasonable to
recognize a spectrum of emotional states experienced by humans, from
the startle reaction, involving minimal cognition, at one end, to pride,
envy, shame, jealousy, grief, remorse, embarrassment and the like, involving
complex and often morally conditioned cognitions, at the other end, with
hunger, surprise, lust, fear, anger, joy, sorrow, and so on filling in the vast

See, most relevantly, Jenefer Robinson, Startle, Journal of Philosophy 92 (1995): 5374.
See on this, in addition to Robinson, ibid., Paul Griffiths, What Emotions Really Are: From
Evolution to Social Construction (Chicago: University of Chicago Press, 1996).
 Emotion in Response to Art 41

middle ground. The emotional responses typical of engagement with art,

though, tend to be of the moderately, or highly, cognitively involved sorta
fact relevant to some recent attempts to dissolve too quickly the paradox of
fiction by appeal to what need not be true of all cases of emotion.
Although it is convenient to speak of emotions having elements or com-
ponents of various sorts, e.g. thoughts, sensations, desires, feelings, pleasures,
pains, shifts of attention, these should not be thought of as merely bundled
together, and the emotion as a mere conglomeration. The truth is rather that
an emotion is an ordered complex or structure of the elements it is taken
to comprehend, with causal relations prominent among those in which this
order consists. My anger at my daughter for having carelessly misplaced my
keys, for example, is a bodily response, rooted in physiology and reflected in
countenance, involving a focusing of attention on her, and feelings of agit-
ation and displeasure, which feelings result jointly from my thought of her
action and my desire that she not have so acted, while fueling, perhaps, my
desire that she in some way pay for having so acted.

II EMOTIONAL RESPONSE TO REPRESENTAT IONAL

ART: THE PARADOX OF FICTION

The much-discussed paradox of fiction can be formulated as a set of three pro-

positions, to each of which we seem to have strong allegiance, but which are
jointly inconsistent, and thus impossible to maintain coherently as a set. Solu-
tions to the paradox, then, typically take the form of rejecting one or more of
the propositions, with a reasoned justification for doing so. The propositions
are these. (a) We often have emotions for fictional characters and situations
known to be purely fictional; (b) Emotions for objects logically presuppose
beliefs in the existence and features of the objects in question; (c) We do not
harbor beliefs in the existence and features of objects known to be fictional.
In the extensive discussion of this conundrum in the literature, almost every
possible solution to it has been essayed. The following comprise most of the
solutions that have found adherents.
(1) The Non-Intentionalist solution: emotional responses to fictions are
not, despite appearances, instances of emotions as such, but rather of less
complex states, such as moods (e.g. cheerfulness) or reflex reactions (e.g.
shock), which lack the full intentionality and cognitivity of emotions per se.
As is evident, this solution involves the denial of (a). But the diagnosis it
42 Art

offers seems to comfortably apply to only a small portion of the full range
of developed responses to fictions.
(2) The Suspension-of-Disbelief solution: while caught up in fictions,
consumers thereof temporarily allow themselves to believe in the nonexistent
characters and situations of the fiction, and thus to have bona fide emotions
for them, reverting to standing beliefs in their nonexistence once the fiction
no longer actively engages them. Such a solution, turning on a denial of (c),
though popular in the nineteenth century, unacceptably depicts consumers of
fiction as having both a rather tenuous grip on reality and an amazing ability
to manipulate their beliefs at will.
(3) The Surrogate-Object solution: emotional responses to fictions take as
their real objects not known-to-be-nonexistent persons and events in fictions,
but other, existent and believed-to-be-existent, objects. This solution, in one
way or another, thus calls (a) into question.
On one version of this solution, the object of response is simply the fictional
work or artistic representation itself, or parts thereof. On another version,
the objects of response are rather the descriptions, images, propositions,
or thought contents afforded by the fiction or representation. And on
a third version, different enough from the preceding two to deserve a
separate labelthe Shadow-Object proposalthe objects of response are
real individuals or phenomena from the subjects life experience, ones
resembling the persons or events of the fiction, and of which the fiction puts
the subject covertly or indirectly in mind.
The Surrogate-Object solution in its first two guises distorts the logic and
phenomenology of emotional response to fictions. Whatever the nature or
status of our response to fictional characters or situations, it is an emotional
response to them, not to something else. Our responses, however ultimately
analyzed, have those characters and situations as their evident objects, and
not the vehicles that bring them to us or the thoughts through which
they are delineated. Much the same complaint can be brought against the
Shadow-Object proposal, though here it is clear that the sort of response
to which the proposal draws attention does indeed often accompany and

For discussion, see Eva Schaper, Fiction and the Suspension of Disbelief , British Journal of
Aesthetics 18 (1978): 3144, and Noel Carroll, The Philosophy of Horror (London: Routledge, 1990).
See, for example, Peter Lamarque, How Can We Fear and Pity Fictions?, British Journal of
Aesthetics 21 (1981): 291304, and Carroll, The Philosophy of Horror, ch. 2.
See, for example, William Charlton, Feeling for the Fictitious, British Journal of Aesthetics 24
(1984): 20616.
 Emotion in Response to Art 43

underlie the emotional response to fictional matters per se. Still, despising
a fictional character, say, is not simply reducible to despising people of
that sort generally, or to despising some actual similar individual of ones
acquaintance.
(4) The Anti-Judgmentalist solution: emotional responses to objects do not
logically require beliefs concerning the existence or features of such objects,
but only weaker sorts of cognitions, e.g. seeing a certain way, or conceiving
in a certain manner, or regarding as if such and such; thus, there is no good
reason not to categorize as standard emotions the emotional responses had
toward fictions, since they satisfy the demands of a more relaxed cognitivism
about emotions. This approach to the paradox, which directly challenges (b),
has a growing number of proponents, and merits extended discussion.
The instances of emotional response that challenge judgmentalismthe
view that the cognitive element involved in all emotions is a judgment or
beliefare mostly of two types. The first type is where there is insufficient time
for substantial cognition, so that no real representation of the object responded
to is formed, there being only a virtually instantaneous reaction, instinctive or
reflexive in nature, unmediated by conscious thought (examples: apprehension
at a suddenly looming shape, disgust at an accidentally felt slug). A second
type is where, though cognition is involved in generating the response, the
representation thus formed is either not propositional in nature, or else does
not have the status of a judgment, or both (examples: phobic fear of garter
snakes, unfounded resentment of female superiors).
As noted earlier, the emotions involved in responding to fictions, ones such
as pity, sorrow, love, admiration, anger, hate, hope, lie in the main in the
middle and upper ranges of cognitive complexity for emotions. It thus seems
undeniable that, whether or not they involve beliefs, such emotions are cent-
rally mediated by representations of various sorts, such as views, conceptions,
or evaluations, which serve to characterize the object of response.
But now even if emotions at this cognitive level do not necessarily involve
beliefs of a characterizing sort about their objects, such emotions, it seems,
must still involve existential beliefs in regard to those objects, or something
very close to that, i.e. attitudes or stances on the order of taking to exist or
regarding as existent. Otherwise, the state attributed becomes unintelligible,

See, for example, John Morreall, Fear Without Belief , Journal of Philosophy 90 (1983):
35966; for a response, see Alex Neill, Fear and Belief , Philosophy and Literature 19 (1995):
94101.
44 Art

whether as an emotion or anything else. How can one be said to pity,

fear, admire, or hate something that one does not, concurrently with ones
emotion, at least take or regard as existing, now or at some other time? If
indeed that cannot be said, then the problem resurfaces, despite what is right
in the critique of judgmentalism: since sane consumers of fiction do not take,
regard, or view fictional characters as existing, even when fully engaged with
them appreciatively, they cannot really be in the full-fledged emotional states
they are casually said to inhabit. The paradox of fiction is proof against anti-
judgmentalist dissolution, even if we grant that emotions can occur without
characterizing beliefs.
The sticking point of the paradox of fiction is the dimension of existence
and nonexistence, as this connects to the cognitive characterization that emo-
tions of the sort in question minimally require. When we view or conceive an
object as having such and such properties, whether or not we strictly believe
that it does, we must, on pain of incoherence, be taking said object to exist or
regarding it as existent. For nothing can coherently be viewed or conceived as
having properties without at the same time being treated as existent. A case of
genuine emotion of a cognitively mediated sort, unlike a corresponding emo-
tional response to a fictional character, involves at least viewing or conceiving
an object as having such and such features, which thus in turn presupposes
regarding it as existent or taking it to exist.
But I do not, when reading Dostoyevskys The Brothers Karamazov, take
Smerdyakov to exist, and so cannot strictly be viewing or conceiving him as
having properties, such as being base or being a murderer. And though my
evaluative Smerdyakov-thoughts, generated as I read, may largely be what
causes my hateful response, directed ostensibly at him, for that response
strictly to have him as its object, and so count clearly as an instance of hatred
of Smerdyakov, requires, once more, that I take him to existwhich I clearly
do not.

It is a mistake, in particular, to try to assimilate cases of fictional fear to cases of phobic fear,
even apart from their evident divergence in behavioral consequences. With phobic fear we can say
that, though the subject doesnt believe the animal in question is dangerous, the subject at least
conceives or views the animal as dangerous, all the while clearly believing that the animal exists, that
there is something the subject is so conceiving or viewing. With fictional fear, however, it is not
open to us to say that the subject even conceives or views some fictional individual as dangerous or
threatening, since the subject does not, if in his right mind, believe that such an individual exists;
he does not believe there is any such thing to be so viewed or conceived. Full-blown fear of X has
an irreducible cognitive component, one part of which is a viewing or perceiving X as dangerous, but
another part of which is a taking or regarding X to exist.
 Emotion in Response to Art 45

It may, however, be the case that I imagine, or make-believe, that someone

of that sort exists, and that as a result I imaginarily, or make-believedly, exper-
ience for him an emotion of hate. Elaborating, my response to Dostoyevskys
character can be interpreted, not as truly one of hatred for Smerdyakov, but
as various sensations, feelings, and focusings of attention caused in me by my
Smerdyakov-thoughts in the course of making-believe that he and his world
exist, which may as a result amount to my make-believedly, or imaginarily,
hating Smerdyakov. (For more on this, see section (7) below.)
(5) The Surrogate-Belief solution: certain emotional responses to fictions,
e.g. that of pity, require only beliefs that, in the fiction, the character exists and
is or does such and such, and those beliefs are indeed widely held by rational
consumers of fiction. This solution thus rejects (c), though not in the manner
of Suspension-of-Disbelief theorists.
However, the beliefs this proposed solution highlights, ones about what is
fictionally the case, can only ground the truth of ones fictionally, or imagin-
arily, pitying a character, not of ones literally doing so. Furthermore, that
such beliefs play a role in generating emotional responses to fictions does not
touch the heart of the paradox, which is that intelligible emotions for objects
of the sort typical of engagement with fiction conceptually require beliefs in
the existence of such objects, or at a minimum, an existential stance toward
them. Beliefs about how things are fictionally can cause emotional reactions of
some sort, to be sure, but they cannot logically ground intelligible emotions for
entities whose existence is denied. Even where the emotion in question is such
as to constitutively require beliefs, they are the wrong sort of beliefs to partly
constitute such emotions. The beliefs I have in connection with Anna Karen-
ina, say, cannot coherently make her the proper object of any pitiful reaction
I might have. Pity involves concern for the welfare of and distress at the suf-
fering of some creature. If one doesnt believe that such welfare or suffering is
actual, what can one be concerned for or distressed about? Pity may likewise
involve wishes or desires, with respect to the thing pitied, but absent a belief in
the thing, or more loosely, an existential stance toward it, there cannot coher-
ently be any such wishes or desires.
(6) The Irrationalist solution: while caught up in fictions, consumers of fic-
tion become irrational, responding emotionally to objects that they know do

See, for example, Alex Neill, Fiction and the Emotions, American Philosophical Quarterly
30 (1993): 113; and Robert Yanal, The Paradox of Emotion and Fiction, Pacific Philosophical
Quarterly 75 (1994): 5475.
46 Art

not exist and thus do not have the features they are represented as having.
Irrationalists either implicitly deny (c), proposing that we in some manner do
endorse the existence of fictional characters and eventswhile apparently at
the same time disavowing themwhich qualifies as irrational, in the sense of
inconsistent, or else implicitly deny (b), holding that we can have emotions
for such as fictional characters and events, toward which we lack the usual
beliefs, but qualifying such emotions consequently as irrational, in the sense
of unwarranted.
On the first construal the Irrationalist solution approaches closely that
of Suspension-of-Disbelief, with the difference, perhaps, that no attempt
is made to mitigate the clash of existential stances involved by suggesting
that they are not simultaneously in full force. On the second construal the
Irrationalist solution holds appreciators of fiction at fault, not for believing
what they already believe the negation of, namely, that fictional characters
and events exist, but for emotionally responding to such characters and events
in ways contraindicated by their beliefs.
It might seem that the Irrationalist solution, on this second construal, is
saved from being a non-starter by the rejection of judgmentalism, since oth-
erwise it could be held to be simply impossible, rather than just possibly irra-
tional, to experience full-fledged emotions in the absence of certain beliefs.
But as suggested earlier, if the critique of judgmentalism, applied to emotions
of the sort that fiction typically elicits, shows only that characterizing beliefs,
as opposed to existential ones, may be absent in such cases, the logical space
this construal hopes to occupy may not be available.
In any event, in the judgment of most commentators, portraying the nor-
mal consumer of fiction as fundamentally enmeshed in irrationality, however
this be understood, is too high a price to pay for this to be an acceptable solu-
tion to the paradox.
Before completing our survey of responses to the paradox of fiction, it is
worth revisiting an issue we may have too quickly settled in setting up the
terms of the problem. The issue is this: Is what underlies the paradoxicality
of emotional responses to fiction a matter of the conceptual impossibility
of a response being an emotion if it does not include, or is not premised
on, certain beliefsas embodied in (b) above in our formulation of the

See Colin Radford, How Can We Be Moved By The Fate of Anna Karenina?, Proceedings
of the Aristotelian Society Suppl. 49 (1975): 6793, and Harley Slater, The Incoherence of the
Aesthetic Response, British Journal of Aesthetics 33 (1993): 16872.
 Emotion in Response to Art 47

paradoxor is it, rather, a matter of the irrationality of responding with

emotion to things believed not to exist or not to be as they are described?
We are now in a position to see that this may depend, to some extent, on
whether one focuses on existence-beliefs or characterizing-beliefs regarding
creatures of fiction. If the formerfor instance, that Anna Karenina exists
or existedthe problem is perhaps best understood as one of rationality: it is
simply irrational to have an emotional response to something whose existence
you dont credit in any degree. If the latterfor instance, that Anna Karenina
suffers or sufferedthen although the issue might again be construed as one
of rationality, since pitying someone you dont believe suffers seems to qualify
as irrational, it is perhaps better understood as concerned with the logic of the
emotional concept in question. For the emotion pity for X might reasonably
be held to strictly presuppose the belief X suffers, so that whatever one feels
toward an item X, it cant be pity unless one believes X to be, as it were, a
logically fit object of pity.
(7) The Make-Believe, or Imaginary, solution: emotional responses to fic-
tions cannot, despite appearances, be instances of the ordinary emotions with
whose names we tend to label them, but are instead instances of imaginary,
or make-believe, emotions. For first, the standard emotions of life arguably
have belief or belief-like presuppositions, notably existential ones, that are not
fulfilled in normal engagement with fictions, and second, such emotions have
motivational or behavioral consequences that are not in evidence in the course
of such engagement.
The proposal is that in our interactions with works of fiction we experience
make-believe emotions, or make-believedly experience emotions, for fictional
characters and situations; it is thus (a) that is rejected on this solution.
Make-believedly experiencing fear, say, is enough like really experiencing fear,
especially internally, that it is easily confused with it, and yet make-believedly
experiencing fear can be reconciled, while really experiencing fear cannot,
with the absence of existential endorsement and motivational upshot vis-a-vis
the fictions that are feared. In this way is the paradox finally resolved.
In considering this solution, it is important to distinguish the claim that
what we feel for fictional characters is some kind of emotion, or constitutes
emotional response in the broad sense, from the claim, here disputed, that
what we feel for fictional characters and describe with some ordinary emotion

See Kendall Walton, Mimesis as Make-Believe (Cambridge, MA: Harvard University Press,
1990) and Gregory Currie, The Nature of Fiction (Cambridge: Cambridge University Press, 1990).
48 Art

word is literally an example of such emotion. We are indeed moved, this

solution affirms, but not strictly to the standard emotions whose names
come to our lips. The issue is not whether making-believe can cause various
emotional reactions, but whether those reactions, given certain cognitive
conditions are not satisfied, qualify as full-fledged emotions of the ordinary
sort. Note also that to classify our emotions for fictions as imaginary is to
say that they are ones we imagine ourselves to be having, on the basis of
experiences, contributory to emotion, that we are actually having, but does
not imply that such emotions are illusory or unreal.
What makes some reluctant to accept that our emotional relations to fic-
tional objects might be of a different stripe from our emotional relations to
objects we take as existent, as the Make-Believe theory insists, is the sense that,
on the inside, they seem very much the samethey feel the same, we might
say. But as has been observed, there is more to emotional conditions than
feelings. Cognitive and conative commitments play a role in the identity of
many, though not all, emotions; thus, if those commitments vary, so may the
emotion which is present.
But suppose it is replied that consumers of fiction do take or regard the
characters encountered in fictions as existing, precisely insofar as they imagine
them to exist as they engage with them? Very well. This can only mean that,
imaginarily, they take them to exist, or that, in the fiction, they take them to
existnot that they take them to exist, period.
A dilemma presents itself, in short, for those who would resist the Make-
Believe solution to the paradox of fiction. Either such taking to exist of their
objects as these emotions must be understood to involve amounts to belief,
in which case the subject, who denies that fictional entities exist, is mired in
inconsistency, or else such taking to exist amounts to making-believe to exist,
in which case any emotion both based on that stance and directed on its object
will be make-believe emotion, of the appropriate sort.
Though the Make-Believe proposal thus probably provides the best res-
olution to the paradox of fiction as such, a full account of our emotional
responses when engaged with fictionsas opposed to our emotions for fic-
tional characters per se will want to acknowledge what is called to our atten-
tion by the Non-Intentionalist and Surrogate-Object proposals as well. And
even the Irrationalist proposal, on the first construal, may contain a grain of

For discussion, see Jerrold Levinson, Making Believe, in The Pleasures of Aesthetics (Ithaca:
Cornell University Press, 1996).
 Emotion in Response to Art 49

truth, for perhaps we are, at least at moments of maximum involvement, in

the incoherent states of mind it postulates as ours throughout.

III EMOTIONAL RESPONSE TO ABSTRACT ART:

MUSIC AND FEELING

Emotional response to abstract art is puzzling, principally, because the strategies

that provide obvious explanations of both why we respond emotionally, and
what we are responding to, in the case of representational art, here seem not
to be available. A novel, film, or Impressionist landscape gives me the image
of a human world, elements of which I can empathize or identify with, react
to sympathetically or antipathetically, or even mirror unthinkingly, by a sort
of natural contagion. But with a symphony, sonata, minimalist sculpture, or
Abstract Expressionist painting such explanations appear to have no purchase.
Human beings and their predicaments are notably absent, at least as far as rep-
resentation is concerned. So why, or how, does perception of such artworks
raise emotion in us, and on what is such emotion directed?
Concentrating for brevitys sake on the art of music, rough answers are as
follows.
Insofar as music is capable of eliciting emotions in listeners, this appears
to work through two different routes or mechanisms, typically operating in
tandem. The first we may label the sensory, or cognitively unmediated route,
and the second the perceptual-imaginative, or cognitively mediated route.
It seems undeniable that music has a certain power to induce sensations,
feelings, and even moods in virtue of its basic musical properties, virtually
without any interpretation or construal on the listeners part. Particular
timbres, rhythms, intervals, dynamics, and tempi exemplify this power most

See Malcolm Budd, Music and the Emotions (London: Routledge, 1985) and Values of Art
(London: Penguin, 1995); Peter Kivy, Sound Sentiment (Philadelphia: Temple University Press,
1989) and Music Alone (Ithaca: Cornell University Press, 1990); Colin Radford, Emotions and
Music: A Reply to the Cognitivists, Journal of Aesthetics and Art Criticism 47 (1989): 6976;
Jerrold Levinson, Music and Negative Emotion and Hope in The Hebrides , in Music, Art, and
Metaphysics (Ithaca: Cornell University Press, 1990), and Musical Expressiveness, in The Pleasures
of Aesthetics; Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press,
1994); Jenefer Robinson, The Expression and Arousal of Emotion in Music, Journal of Aesthetics
and Art Criticism 52 (1994): 1322; Francis Sparshott, Music and Feeling, Journal of Aesthetics
and Art Criticism 52 (1994): 2335; Alan Goldman, Emotions in Music (A Postscript), Journal of
Aesthetics and Art Criticism 53 (1995): 5969; Aaron Ridley, Music, Value, and the Passions (Ithaca:
Cornell University Press, 1995); and Roger Scruton, The Aesthetics of Music (Oxford: Oxford
University Press, 1997).
50 Art

clearly. Such properties need only be registered, as it were, to have their effect,
at least for one acclimatized to a given musical culture. The rise in heartbeat
caused by rapid tempo, the discomfort occasioned by dissonant intervals, the
kinetic impulses induced by dancing rhythms, the excitement produced by
quick alternations of soft and loud, the relaxation engendered by a certain
tone color or manner of articulation, are all familiar phenomena.
But if the capacity of music to elicit emotion were exhausted by the direct
effects of sensing basic musical features, it would be a poor thing, falling far
short of the evocation of emotions proper, or even the semblance of such. The
gap is filled by the second, or cognitively mediated, route to such evocation.
In addition to presenting an array of sonic features, simultaneously and suc-
cessively, much music offers the appearance of human emotion, or of persons
outwardly manifesting emotional states; arguably, that is what the expressive-
ness of music largely consists in. In other words, music is often heard as, or
heard as if, or imagined to be, the expression of emotion by an unspecified
individual, whom we may call the musics persona. The degree of resemblance
between the shape of music and the behaviors through which emotions are
commonly expressed in life will have something, though not everything, to
do with our being disposed to hear music in such ways. In any event, once
this occurs, the mechanisms mentioned above and familiar from appreciation
of representational artmirroring, identification, empathy, sympathy, anti-
pathycan come into play, resulting in the arousal in the auditor of those
same emotions, or else the feelings characteristic of them, or else those emo-
tions on an imaginary plane. The sensory aspect of music alone indeed seems
capable of inducing in us at least a number of simple states of arousal typic-
ally identified as constituent elements of one or another emotion. But it is the
perceptual-imaginative aspect, manifested in our disposition to hear emotion
or emotional expression in music, that is surely primarily responsible for the
complex, more robustly emotional responses to music, whether mirroring or
reactive, that so many listeners report.
It remains to add that these mechanisms do not operate in total isolation
from each other. The emotion I hear a passage as expressing may soften or
accentuate the particular psychological effect some basic musical feature pro-
duces on me, while the effect induced in me, largely unthinkingly, by some
basic musical feature may influence and constrain the emotion I am disposed
to hear an image of in the music.
But if emotions are often produced in listeners in virtue of auditing emo-
tionally expressive music, toward what are such emotions directed? Music
 Emotion in Response to Art 51

neither supplies any objects, nor appears itself to be an appropriate object, for
at least the vast majority of such emotions as are putatively aroused. In addi-
tion, music does not seem to provide anything that would justify the beliefs or
attitudes toward objects that many objects can be held to require. Among the
ways of responding to this difficulty are the following.
It can be held that music produces in listeners only moods, which
intrinsically lack intentionality, e.g. anxiety or elation, or else objectless
emotions, ones characteristically taking objects but somehow lacking them
when aroused by music, e.g. sadness or joy directed on nothing, or nothing in
particular.
Alternatively, it can be held that music produces in listeners just the feeling
component of an emotion, together with the sense of focus or directedness
inherent in the bodily response at the emotions core, but not the cognitions
which characteristically accompany or even partly constitute the emotion.
Finally, it could be maintained that what music occasions in many listen-
ers are states of imaginary emotion. The idea is that listeners readily erect,
upon a basis of feelings produced in them by music whose expressiveness
they empathetically grasp, imagined emotions of a corresponding sort, and
that they do this through imagining, usually tacitly, objects and thoughts
suitable to the emotions in question. The object of musical emotion, then,
is not missing, but merely posited indefinitely in imagination, or perhaps
appropriated, as it were, from the emotion imaginarily ascribed to the musics
persona.

IV EMOTIONAL RESPONSE TO NEGATIVELY

EMOTIONAL ART: THE PARADOX OF T RAGEDY

The paradox of negative emotion in artof which the paradox of tragedy is a

classical illustrationis this. Art that is negatively emotional, i.e. art that rep-
resents, expresses, or otherwise deals with emotions such as shame, grief, hor-
ror, sorrow, anger, remorse, despair and the like, seems to have a propensity
to elicit parallel responses in appreciators. But if that is so, one would expect
appreciators to avoid, or at any rate judge as inferior, art of this nature. Yet not
only do they not do so, but often they hold such art to be the highest or most
rewarding art of all.
A number of possible explanations have been given for how it is that per-
sons rationally desire or value the empathic experience of negatively emotional
52 Art

art, given the ostensibly negative character of that experience. Here is a rough
categorization of such explanations.
(i) Compensatory explanations: negative emotion aroused by negatively
emotional art is, as such, unpleasant, but undergoing it offers other rewards
that compensate for this. (ii) Conversionary explanations: negative emotion,
which is initially or ordinarily a disagreeable response, is transformed, in the
context of artistic appreciation, into something that is in fact agreeable, or
at any rate, capable of being enjoyed. (iii) Organicist explanations: negative
emotion aroused by negatively emotional art is an essential element in a
total experience, an organic whole that is desired or valued. (iv) Revisionary
explanations: neither negative emotions, nor the feelings they include, are
intrinsically unpleasant or undesirable, and thus there is nothing odd about
appreciating art that induces such emotions or feelings. (v) Deflationary
explanations: despite appearances, neither negative emotions, nor the feelings
they include, are really aroused in us by negatively emotional art.
Compensatory explanations include Aristotles doctrine of catharsis, under-
stood as a purging or purification of excess or unruly emotions of pity and
fear through engagement with tragic drama, which justifies the raising of such
emotions in the course of that engagement. Another such explanation appeals
to the value of knowledge of important truths of human existence that emo-
tional engagement with negative art is said to afford. A third explanation
endorses such engagement, not for the knowledge of life it may afford, but
rather for the knowledge of the artwork it facilitates, emotional engagement
with a work being seen as a necessary cost, in many cases, of fully understand-
ing it. A fourth such explanation invokes the moral exercise that is provided,
or the moral deepening that results, as a benefit of engagement with negatively
emotional art. And a fifth explanation appeals to purely aesthetic pleasures in
the beauty, lifelikeness, virtuosity, or cognitive interest of the representation
or expression itself, positing these as enough to outweigh whatever negative
emotion is undergone in their appreciation.
Conversionary explanations include Humes explanation of the appreci-
ation of tragedy; like that just noted, Humes explanation highlights the pleas-
ure in artistic representation and expression as such, but premises that this

See, for instance, Mark Packer, Dissolving the Paradox of Tragedy, Journal of Aesthetics and
Art Criticism 47 (1989): 21219.
See, notably, Nelson Goodman, Languages of Art (Indianapolis: Bobbs-Merrill, 1968), ch. 6.
Noel Carrolls resolution of the paradox of horror, a first cousin of the paradox of tragedy, is
largely of this sort. See his The Philosophy of Horror.
 Emotion in Response to Art 53

pleasure, being greater than the pain of the negative emotions concomitantly
raised, does not simply offset that pain, but rather overwhelms and absorbs it,
leaving an experience of uniformly positive character. A rather different con-
versionary explanation proposes that since the negative emotions raised by a
work of art have no life implications for spectators, calling for no actions and
betokening no real harms, such emotions must evidently be so altered by the
artistic conditions under which they issue that, though still recognizable as
this or that negative emotion, disagreeable affects intact, they are yet capable
of being relished or enjoyed for experiences sake.
An example of an organicist explanation would be one invoking a satisfac-
tion in some negative emotions being raised in one by a work of art, perhaps
because the emotion strikes one as appropriately raised in such circumstances,
and oneself as admirably human for being thus susceptible. Such a satisfaction
would obviously be inseparable from the negative emotion raised, in the fact
of which satisfaction is taken. Another such explanation would appeal to
the value of working through negative emotions in connection with a work of
art, via immersion in its formal, narrative, or dramatic structure, the emotions
raised thus being an essential element in the experience valued as a whole.
Revisionary explanations go something like this. The experience of neg-
ative emotions is not intrinsically unpleasant; the affects, that is, sensations
and feelings, involved are not in themselves disagreeable, and can be unprob-
lematically savored as such, in appropriate contexts. What is negative about
negative emotions is only the evaluation of their objects that is central to such
emotions. Thus, there is no special difficulty about people seeking these emo-
tions from art.
Deflationary explanations come in at least three varieties. One hypothes-
izes artistic analogues of the life emotions, distinct from them in hedonic tone,
conative connectedness, and behavioral implication, and proposes that only
these are raised in us by engagement with emotional art, and not the life emo-
tions themselves. Another deflationary explanation simply flatly denies that

See Marcia Eaton, A Strange Kind of Sadness, Journal of Aesthetics and Art Criticism 41
(1982): 5163, and Levinson, Music and Negative Emotion.
See in this vein Susan Feagin, The Pleasures of Tragedy, American Philosophical Quarterly 20
(1983): 95104.
See John Morreall, Enjoying Negative Emotions in Fiction, Philosophy and Literature 9
(1985): 95103; Jenefer Robinson, Leducation sentimentale, Australasian Journal of Philosophy
73 (1995): 21226; and Levinson, Music and Negative Emotion.
A suggestion of this sort can be found in Kendall Walton, Mimesis as Make-Believe (Cambridge,
MA: Harvard University Press, 1990), ch. 7, and also in Berys Gaut, The Paradox of Horror,
British Journal of Aesthetics 33 (1993): 33345.
54 Art

anything like the garden-variety emotions are evoked in subjects in the course
of engaging with emotional art, and suggests that the subjects response, inso-
far as it is emotional, is exhausted by properly appreciative reactions, such as
being moved by a works beauty of expression. A third deflationary explan-
ation maintains that spectators are always only make-believedly in states of
negative emotion in virtue of engaging with a work of art, and that on the
assumption that make-believe emotions of the negative sort are not inherently
displeasing, there is no special problem about people tolerating, or even act-
ively pursuing, such experiences.
Detailed assessment of these proposals awaits another occasion, but in my
view there is more merit in compensatory and organicist explanations, and
in the second of the conversionary explanations sketched above, than in revi-
sionary or deflationary ones.

V E MOTION AND THE APPRECIATION OF ART

Are there emotions unique to the appreciation of art, or aesthetic emotions per
se, had when and only when a work is apprehended aesthetically? Past theorists,
notably Clive Bell, have posited something of this sort, but such a posit has
not lately found favor, nor does it appear to answer to any pressing theoretical
problem about art. There may, on the other hand, be an interesting category
of positive emotions that, if not had uniquely for art, are both distinctive of
the appreciation of art and not of the sort that typically figure in the content
of art. Candidates for membership in this category would be emotions such
as admiration for a works skill, fascination with a works form, delight in
a works beauty, or awe at a works depth of expression. What might also
figure here are experiences, remarked by many, of momentary will-lessness or
self-transcendence occasioned by intense absorption in a work of art.
The question may also be raised as to the appropriateness of emotional
responses to art of the ordinary sort. One form of this question concerns an
apparent tension between the familiar picture of an emotion as a disturbing
temporary derangement of the psyche and the image of aesthetic appreciation
as a state of calm and unclouded attention to a work of art.

A suggestion of this sort, applied to music, can be found in Kivy, Music Alone, ch. 8.
Suggestions of this sort are also to be found in Walton, Mimesis as Make-Believe.
But see the debates on this, pro and con, in Monroe Beardsley, The Aesthetic Point of View
(Ithaca: Cornell University Press, 1983), and George Dickie, Art and the Aesthetic (Ithaca: Cornell
University Press, 1974).
 Emotion in Response to Art 55

The traditional notion of the aesthetic attitude, whose roots are in Kant,
Schopenhauer, and the eighteenth-century theorists of taste, depicts a frame
of mind characterized by disinterestedness, detachment, and disengagement
from the practical. Charitably construed, such a notion demands only that
ones personal situation or condition not be what primarily drives or directs
ones response to a work, but instead, the humanly significant material that
the work presents. In other words, such a notion need not call for suppression
of emotional receptivity generally. So long as ones emotional response is a
way of connecting to a work, of tracing its expressive outline, or grasping
its dramatic import, rather than a means of being distracted from it, a
springboard to simply wallowing in ones private concerns, then there is no
conflict between responding to a work with a range of ordinary emotions, on
the basis of ones life experience and individual sensibility, and appreciating
a work in an aesthetically appropriate manner, as the specific embodiment of
human content it is. By contrast, disinterestedness or detachment understood
not as a principle for maintaining focus on a work rather than ones own
circumstances, but as a desired end-state of impassivity or imperturbability,
is nothing that an account of artistic appreciation need embrace.
Finally, accounts of the value of emotional response to art can be divided
roughly into those that exploit the value of emotional experience generally, for
example in contributing to a full life, and those that seek instead to identify a
particular value of emotional experience in the context of art and its appreci-
ation, for example as a mode of understanding a work more fully or a means of
reaping more efficiently the benefits a work has to offer.

For further discussion, see Jerrold Levinson, What Is Aesthetic Pleasure?, in The Pleasures of
Aesthetics.
See, for example, the discussion as applied to music in Davies, Musical Meaning and Expres-
sion, ch. 6.
 4
Elster on Artistic Creativity

1. The French novelist Georges Perec wrote a novel, La Disparition, in which the
letter e was nowhere used . . . This is an extreme example of the more general idea
that artists may impose constraints on themselves in order to create better works
of art. In Perecs case the constraints were entirely idiosyncratic. In other and more
frequent cases the constraints take the form of conventions that define a particular
genre. Although freely chosen, in the sense that it is up to the artist whether to submit
to the laws of the genre, they are not invented by the artist. In still other cases the
constraints are imposed from outside . . . (1756)
So begins Jon Elsters discussion of creativity and constraints in art in the
third part of his recent study, Ulysses Unbound: Studies in Rationality, Pre-
Commitment and Constraint. I return later to La Disparition, indicating my
differences with Elster on the virtues of that singular work, but I begin by
exploring how far we might take the tripartite division of artistic constraints
that Elster here suggests, into chosen, invented, and imposed ones.
Chosen constraints are ones adopted by the artist from a pool of preexisting
styles, genres, or forms. Invented constraints are ones devised by the artist,
and to which the artist subsequently adheres. And imposed constraints are
ones rooted in external conditions that cannot be altered or evaded by the
artist.
Some examples of chosen constraints are: sonata form; sonnet; haiku;
iambic pentameter; still-life painting; charcoal drawing; two-person play,
comedy of manners; Greek temple; Roman arena. Some examples of invented
constraints are: twelve-tone composition; musique concrete; prepared piano
pieces a la John Cage; collage painting; kinetic sculpture; readymades a la

First published in B. Gaut and P. Livingston (eds.), The Creation of Art (Cambridge: Cambridge
University Press, 2003), 23556.
Jon Elster, Ulysses Unbound: Studies in Rationality, Pre-Commitment and Constraint (Cambridge:
Cambridge University Press, 2000). Page references to this book in the text are given in parentheses.
 Elster on Artistic Creativity 57

Duchamp; rayograms a la Man Ray; and absurdist theatre a la Beckett or

Ionesco.
Some examples of imposed constraints, finally, are these: the maximum
length of breath possible in singing; the maximum loudness attainable in solo
piano music; the maximum height of jump in dance; the rectangularity of
frames or shots in film; the irreversibility of shaping in carved sculpture; the
prohibition of live firearms in installation art; the prohibition of certain sub-
ject matters in PG-rated film.
With both chosen and invented constraints there is evidently an element
of voluntary participation by the artist, whereas with imposed constraints
this appears to be lacking. But it is worth observing that even with imposed
constraints, the artists agency is not entirely annulled. In the first place,
the artist is still faced with the choice between various attitudes that can
be taken to the imposed constraint. For example, cheerful acceptance, stoic
resignation, angry defiance, or blithe obliviousness. In the second place, the
artist is always free to decline the gambit entirely, electing not to create subject
to the constraint in question, and simply passing on or forgoing the artform or
endeavor in which that constraint is unavoidable.
If an artist chooses a constraint, says Elster, he presumably believes he will
benefit artistically from having a smaller sphere of choice. This is a case of
standard pre-commitment, and the constraint is thus an essential one, i.e.
consciously chosen for the expected benefits of it. In other cases a constraint
is adopted or accepted, but not with an eye to reaping a benefit; it is then an
incidental constraint. An incidental constraint may end up being of benefit,
but such benefit is not the reason for its adoption or acceptance. And an
incidental constraint may also evolve into an essential constraint, if endorsed
by the artist for its expected benefits after initial adoption.
2. One of Elsters main themes is the contrast between choice of con-
straints, at the outset of the creative process, and choice within constraints,
occurring throughout the creative process:

A nice example of a fruitful attitude one might adopt toward imposed constraints is provided
by the following anecdote, from liner notes by Harry Halbreich to Arthur Honeggers Le Roi David
(Erato 2292-45800-2). Honegger, having been commissioned to write music for the theatre piece in
question, but chafing under the rather peculiar musical conditions the commission involved, asked
his fellow composer Stravinsky for advice. This was the advice offered: Its quite simple . . . Act
as if you had wanted exactly that ensemble, and compose for a hundred singers and seventeen
instrumentalists. Honegger comments on this advice that it constituted for him an excellent
composition lesson: never consider the givens of a commission as something imposed, but on the
contrary as a personally chosen stricture, or an interior necessity.
58 Art

The creation of a work of art can in fact be envisaged as a two-step process: choice
of constraints followed by choice within constraints. The interplay and back-and-forth
between these two choices is a central feature of artistic creation, in the sense that
choices made within the constraints may induce an artist to go back and revise the
constraints themselves. (176)
Elsters description has the ring of truth: there are undoubtedly phases in the
process of creating works of art, and the phases may typically be related in
the way that Elster proposes. But still we must ask, if the artist is free to go
back at will, at any point, and revise the constraints under which he had been
operating, is it really useful to think of them as constraints, rather than as, say,
provisional guidelines, or working assumptions? Elsters idea of constraints
as a constant and defining feature of the creative process seems somewhat
too . . . constraining.
To be fair, Elster acknowledges that there is an interaction between choice
of constraints and choice within constraints in the making of art. An example
he gives is this:A writer may initially plan to develop an idea in a full-length
novel, and then, finding that it will not sustain that format, turn it into a short
story of thirty pages. (201). Even so, Elster arguably underestimates the purely
heuristic value for artistic creativity of mere opening moves or posits, ones
not thought of as binding, and thus not reasonably construed as constraints.
The self-binding of an artist, in other words, is rather unlike the self-binding
which served as a paradigm in Elsters earlier studies of the rationality of pre-
commitment, namely that undertaken by Ulysses in having himself bound to
the mast of his ship as a prophylactic against the sirens. For Ulyssess self-
binding was entered into (a) deliberately and explicitly, (b) categorically and
unrevisably, and (c) with a definite objective in view. But the self-binding of
an artist, if we consent to call it that, is typically (a) inexplicit or only partly
explicit, (b) eminently revisable, and (c) often entered into with no very definite
objective in view. In short, Elsters model of precommitment is unnecessarily
inflexible, and sorts ill with the way artists actually go about creating works.
3. Elster is explicitly committed to an account of artistic creation that sees
it as a project of knowingly engaged-in maximization:

See, in particular, his Ulysses and the Sirens: Studies in Rationality and Irrationality (Cambridge:
Cambridge University Press, 1990).
Michael Bratman, by contrast, offers a more flexible model of pre-commitment, one that is
perhaps a better fit for the artistic case. Bratman stresses the need for rational agents to frame
future-directed intentions that have a certain inertia or weight, yet not be bound to pour good
money after bad. See his Faces of Intention: Selected Essays (Cambridge: Cambridge University
Press, 1999).
 Elster on Artistic Creativity 59

. . . both choice of constraints and choice within constraints can be represented as a

form of maximization. Specifically, artists try to maximize artistic value. (178)
The process of artistic creation is guided by the aim of maximizing aesthetic value
under constraints . . . Creativity is the ability to succeed in this endeavor. (200)
The idea that artists seek to maximize aesthetic or artistic value certainly
has some plausibility. If agents are generally, to a rough approximation,
utility-maximizers according to their own, often idiosyncratic, utility
functions, then it is not too big a stretch to suggest that artists are, to some
extent and in some contexts, maximizers of the utility that corresponds to
aesthetic or artistic value.
But is not Elster guilty of making the creation of art out to be a too pre-
dominantly rational affair? First, it is not clear that the maximizing aim stated
above can plausibly be attributed to most successful artists, even at a less than
fully conscious level. Whether on the basis of express avowals, observed beha-
vior, or the character of completed works, the evidence does not make such
an attribution inescapable. Second, even those most committed to rational
explanation of human behavior usually allow that there is some share of the
irrational, or at least nonrational, in creative activity, as in related activities
of a sexual or religious kind, with their not wholly transparent bases. Now
Elster might try to accommodate that by suggesting that acceding to a limited
irrationality in the service of creative ends can be seen as a strategy that is over-
all rational, and perhaps even maximizing of aesthetic value, but surely this
is to ascribe too calculating an attitude to artists who open themselves to the
irrational in the name of art. In other words, though it is true that irrational
methods are sometimes rationally chosen by artiststhink of the automat-
ic writing experiments of the Surrealistssurely not all irrationality issuing
in something of artistic value has been deliberately opted for by the artist on
rational grounds.
Third, and most importantly, there are positive aims we have good grounds
to ascribe to artists as motivating them that do not reduce to that of max-
imizing aesthetic or artistic value. Some artists are driven to discover new
relationships between elements of their chosen media, or new ways of express-
ing states of mind, or new techniques for achieving realistic representation,
where their commitment to those goals as such overrides any direct concern
they might have to maximize aesthetic value within given constraints. In even
simpler terms, artists are most often driven to make works the works they want

Elster is not too careful of this distinction, a matter to which I return below.
60 Art

them to be, whether or not they are seen to maximize aesthetic value, even by
the artists own lights.
4. Elster, like some other recent writers, takes a fairly firm stand
against originality: I distinguish creativity (working within constraints) from
originality (changing the constraints), arguing that the latter has no intrinsic
relation to aesthetic value (180). Elster sustains that judgment even in
light of this later, ostensibly approving gloss of what originality involves:
Originality . . . can be defined as a durable break with existing conventions
rather than as a momentary departure from them. The emergence of free
verse, nonfigurative art, and atonal music are obvious examples from the
past century (224). Elster recognizes that originality in a work of art can
coexist with greatness, and that the search for originality may even enhance
the pursuit of creativity, but he remains opposed to the idea that originality
can be part of the value of a work of art.
Elsters judgment against originality per se, however, fails to recognize
artmaking as an activity that, like most activities, involves the possibility
of accomplishment or achievement, and the artwork that results as the
embodiment of such accomplishment or achievement, which makes the
work logically and appreciatively inseparable from the activity that generates
it. Part of the achievement of some works of art is precisely the striking
originality, whether of means, of ends, or of ends-in-relation-to-means, that
they manifest. Once the achievement in question is solidified and secured,
usually through the production of a series of related works, then later works,
however similar, do not partake of that achievement and do not exhibit that

This observation, which I owe to Richard Wollheim, abstracts from purely extrinsic or strategic
considerations, such as might induce an artist to make a work a certain way to please a patron, to
secure a commission, or to annoy a rival. In a sense, an artist in such cases is making a work the
way he wants it to be, but not for its own sake; in other words, he is not making quite the work he
wants, qua artist, to make.
See, for example, Frank Sibley, Originality and Value, British Journal of Aesthetics 25 (1985):
16984; and Bruce Vermazen, The Aesthetic Value of Originality, Midwest Studies in Philosophy
16 (1991): 26679; and more evenhandedly, Paul Crowther, Creativity and Originality in Art,
British Journal of Aesthetics 31 (1991): 3019. Sibleys argument against originality as such goes like
this: original is very commonly used of style, manner, technique, medium; Cubism, Abstract
Expressionism, Pointillisme, the twelve-tone system exemplify large innovations . . . In themselves,
these were extremely original but aesthetically quite neutral . . . The value of such innovations is
instrumental; it lies in what new aesthetic characters and values they render possible . . . So the
enthusiastic praise often lavished on works employing innovatory techniques [or] . . . novelties of
form or medium is misplaced unless these bring, or have the potential to bring, some new aesthetic
character of value (1745).
There is some irony, though, in Elsters opposition to artistic originality, given the reputed
creed of Elsters musical hero, Lester Young: Youve got to be original, man.
 Elster on Artistic Creativity 61

originality, and hence are not as admirable artistically. The locus of originality
in art is the historically rooted artwork itself, not the works maker as an
individual, and not the works abstract form or detachable content.
Arguments that originality is not an artistic value, moreover, tend to equate
artistic value, or value attaching to art as art, with narrowly aesthetic value,
that is, value derived from sensuous-perceptual-imaginative experience of an
objects formal and aesthetic qualities. But the two are not equivalent.
Artistic value is a broader notion than aesthetic value, though it includes
aesthetic value as its core. For clearly the experiential rewards afforded an
audience by an artwork, and by design, are a large part of its value as art. But
some of the value of an artwork is not anchored in its character as an object
for perceptual or imaginative experience, which it may share with portions of
nature, but is instead rooted in its being a human artifact, in virtue of which
it, say, displays originality, expresses singular attitudes, communicates moral
insight, or influences in a positive manner artworks to come. These contribute
to its artistic value, though not to its aesthetic value in the narrow sense.
5. Aristotle famously offered a criterion of organic unity, the essence of
which is that a work possessed of such unity can only be altered for the worse.
Elster, it appears, is inclined to accept Aristotles dictum: One piece of evid-
ence for the view that artists are engaged in maximizing is the widespread
belief that in a good work of art nothing can be added and nothing be subtrac-
ted (Aristotle, Poetics.) . . . Another piece of evidence is the widespread prac-
tice of artists of experimenting with small variations until they get it right
(202).
But Aristotles dictum should not in fact be so uncritically accepted. From
the fact that in a good artwork every element plays some role in the function-
ing of the work or the emergence of its artistic content, one cannot validly
infer that only those elements could have served that function or have resul-
ted in that artistic content. More generally, one cannot conclude that only
the specific ensemble of elements that constitutes the finished work can be
thought to optimize the value such a work might possess. This is because there
may be several, equally effective ways of getting it right, that is, bringing a
work to successful completion once it has reached a certain point. The prac-
tice of trying out small variations need not be seen as a search for a unique

For more on artistic value versus aesthetic value see my Art, Value, and Philosophy, Mind
105 (1996): 66782. On the related though not equivalent distinction that can be made between
the artistic and the aesthetic properties of works, see my Artworks and the Future, in Music, Art,
and Metaphysics (Ithaca: Cornell University Press, 1990).
62 Art

local maximum. In fact, it need not even be viewed as a search for any loc-
al maximum, but might instead be seen as governed by a principle of good
enough, that is, as conformable to a satisficing, rather than optimizing, model
of human behavior. Even if getting it right generally motivates the trying out
of small variations of a work in progress, that may not require making it best,
but only, say, really quite good.
The error of unqualified adherence to the doctrine of organic unity, which
maintains that successful works of art are so finely tuned that no alteration in
them can fail to have an adverse effect, is related to that involved in unres-
tricted allegiance to the principle of aesthetic uniqueness, which holds that any
two perceptually different works of art belonging to the same genre neces-
sarily differ in some aesthetic respect or other. But though close to true, that
principle falls short of universal validity. Though perceptually different art-
works in a given genre necessarily afford different experiences and perhaps
even experiences that differ aesthetically, if the perceptual basis of an aesthetic
property always comes within the scope of its appreciationthat does not
yet imply that the works themselves necessarily differ aesthetically. Think of
two abstract, calligraphic paintingssay in the vein of Mark Tobeywhere
one is a mirror image, or else a slight internal rearrangement, of the other. Is
it clear that these will differ at all in aesthetic impact, despite their being eas-
ily perceptually discriminable? Both such paintings might work, so to speak,
and in exactly the same way, and without inclining one to think that no other
arrangement of similar elements could have worked as well.
6. Elsters discussion of the dimensions of value in temporal artforms,
and their particular manifestation in the sphere of classic jazz, is to my mind
the most satisfying and stimulating part of his inquiry into artistic creativity in
Ulysses Unbound. I thus take the liberty of quoting from it at some length.
I distinguish between two sources of emotional satisfaction through the arts. On
the one hand, many works of art can generate non-aesthetic emotionsjoy, grief,
and the like. On the other hand, all works of art, if they have any artistic value at
all, induce specifically aesthetic emotions by means of rhythm, echoes, symmetries,
contrasts, repetitions, proportion, and similar devices . . . The specifically aesthetic
emotions include wonder, amazement, surprise, humor, relief, and release. (206)
In literature and music, sublime artistic effects are created when these two emo-
tional effects go together and reinforce each other. As if by magic, the pleasure of a

See my Aesthetic Uniqueness, in Music, Art, and Metaphysics.

See my What Is Aesthetic Pleasure?, in The Pleasures of Aesthetics (Ithaca: Cornell University
Press, 1996).
 Elster on Artistic Creativity 63

rhyme that falls into place adds to the poignancy of the words. Although the com-
ponents are analytically separable, they are not experienced separately. (206)
In the case of jazz, the non-aesthetic emotions in the listener are produced by
music that has what I shall call emotional depth. The aesthetic emotions can arise
at two levels. At the simplest level, they are produced by music that possesses what I
shall call taste. At a more advanced level, the aesthetic emotions are produced by the
story told by the music . . . (247)
Tastethe sense of order, balance, proportion, timingis an essential prerequis-
ite for the production of the specifically aesthetic emotions . . . Emotional depth refers
to the capacity to generate strong non-aesthetic emotions in the listener. (248)
Taste and emotional depth are only two of the relevant dimensions of quality in
jazz. The thirdand the only one in which jazz differs radically from other musical
performancesis inventiveness in storytelling. (252)
The ability to tell a story through melodic innovation is related to taste, but goes
well beyond it. (253)
There is much with which to agree in the above extracts, and the discussion
from which they are drawn is unquestionably an insightful contribution to
the aesthetics of jazz. But I have, nevertheless, certain qualms.
My first qualm concerns Elsters claim that the storytelling dimension of a
jazz solo, its ability to suggest, through its melodic, harmonic, and dynamic
evolution, an abstract narrative of some sort, enters into the generation
in the listener of aesthetic emotions, but not non-aesthetic ones. This
seems unwarranted. For the narratives that music can suggest are often
themselves emotional ones, and it stands to reason that our response to an
emotional sequence might be a further non-aesthetic or life emotion that
would be appropriate to the sympathetic contemplation of such a sequence.
For example, a solo conjuring up a narrative of trouble, followed by hope
for that troubles disappearance, leading only to the dashing of that hope,
would naturally conduce to a feeling of sorrow or distress in a listener, apart
from any purely aesthetic reaction on the listeners part to the rightness of the
emotional narrative or its musical underpinning.
My second qualm relates to the degree to which taste, characterized
as a formal or configurational matter, and emotional depth, characterized

What precise distinction Elster has in mind by aesthetic vs. non-aesthetic emotions is not
entirely clear. Some emotions which he denominates aesthetic, for instance wonder, can certainly be
had in connection with objects that are not works of art. However, I will accede in the distinction as
drawn, taking it to be roughly this: aesthetic emotions are ones characteristically had for or directed
on the aesthetic aspects, including formal ones, of objects, most notably works of art.
For an emotional scenario of that sort, and its role in the generation of musical frissons, see
the discussion of a Scriabin etude in my Musical Chills (Ch. 12, this volume).
64 Art

as a material or substantive one, are entirely independent and summable

dimensions of musical value, as is suggested by Elsters amusing diagram
(251) locating various jazz artists in a two-dimensional space whose y-axis
is taste and whose x-axis is emotional depth. I think these dimensions
are rather interlocking ones, with a musicians formal sense of timing,
proportion, and flow impacting on the non-aesthetic emotions his discourse
will be able to elicit. For example, the concision, clarity, and drive of John
Coltranes Giant Steps, presumably the result of the great saxophonists
taste, have quite a lot to do with the bold, life-affirming confidence the
music expresses. And the relative structural freedom and relaxation of
Coltranes A Love Supreme is surely not separable from the serenely
rapturous sense of the goodness of things that that music projects, especially
in its first two sections.
In addition, we may note that the formal and emotional dimensions of
musical value in jazz or any other music are arguably subordinate to another,
overarching and non-summative, dimension of value we can label emotion-
in-relation-to-form, or content-in-relation-to-configuration. That is, our
ultimate judgment of the value of a piece of music arguably does not rest only
on our separate estimations of the quality of its content and the excellence of
its form, but also, and most particularly, on the specific relation or fitting-
ness of the former to the latter.
7. Elster has interesting things to say on the topic of pre-commitment to
chance or randomness in the making of art:
. . . each choice made in the creation of a work of art serves as a constraint on later
choices. An artist may decide, however, to generate the constraints randomly rather
than intentionally. (242)

Although I agree with most of the critical judgments embodied in Elsters diagram, and concur
in particular with the placement of Lester Young, Billie Holiday, Johnny Hodges, and Django
Reinhardt in its favored upper right corner, I would explain the relative undervaluation of Art
Tatum, Sarah Vaughan, and Frank Sinatra, who though accorded high y-values are accorded only
middling x-values, in terms of Elsters simply undervaluing certain realms of expression, namely
the lithe/blithe/carefree, relative to others, namely the heavy/doleful/careworn. Only if emotional
depth is narrowly equated with the latter does Elsters downgrading of Tatum, Vaughan, and
Sinatra seem justified. I should also note that Elster in fact recognizes that his two dimensions
of musical value are not entirely independent: Taste and emotional depth do not . . . vary entirely
independently of each other. Total lack of taste is incompatible with great emotional force (250).
But I submit that he still overestimates the degree of their independence, as I try to show in the text.
See my Evaluating Music, in Philip Alperson (ed.), Musical Worlds (University Park, PA:
Pennsylvania State University Press, 1998), and Ch. 10 in this volume.
Understood here to include also its degree of inventiveness in storytelling, which is Elsters
third proposed dimension of musical value.
 Elster on Artistic Creativity 65

Thus Francis Bacon began his pictures by throwing paint at the canvas, so that the
resulting blots would serve as constraints on the further work, limiting his freedom
and presumably enhancing his creativity. Jackson Pollocks poured paintings may
be seen as a variant of this idea. (243)
But as regards the first of these cases, Bacons accidentally produced blots
need not be taken as constraints on, rather than prompters of, further creative
decisions, since it is in the nature of the medium of painting that any such
blots can be covered over or scraped off, if need be. And to a lesser extent that
would be true in the case of Pollocks paintings as well.
Elster next turns his attention to that enfant terrible of modern music, John
Cage:
To the extent that it relies on randomizing or on producing periods of silence, Cages
work is entirely unserious. The most charitable interpretation is that it is a gigantic
and successful put-on. (245)
. . . the use of objective or epistemic randomness to select within constraints has
no aesthetic justification . . . by removing choice rather than restricting it one destroys
creativity rather than enhances it. (246)
This strikes me as a fairly uncomprehending reaction to Cages work as an
artist, however one rates that work in the last analysis. Cages creativity mani-
fests itself, in part, in the astonishing variety of evocative means he devised to
circumvent the operation of human choice at the level of sound selection, in
the name of an individual quasi-Buddhist philosophy of music and life. The
irony of Cages work, though, is that his very distinctive personality shows
itself, at a second level, through the specific range of meansdice, star charts,
paper imperfections, the I Ching he devises for the elimination of person-
ality at the first level, and even through the very undertaking of the project of
eliminating subjectivity or preference from music to begin with.
Richard Wollheim, in a famous essay on art of the kind that subsequently
became known as minimal, sketched a conception of negative work in
artmaking, that is, the sort of work appropriate to the production of objects
of excessive simplicity and relative absence of articulation. Such negative
work, suggests Wollheim, is largely a matter of decisions about what a piece
will be like, ones involving negative acts such as renouncing, eschewing,

Personality is a flimsy thing on which to build an art, Cage once wrote. Yet he himself
seems to make his most direct and appealing impact through his own personality (John Rockwell,
All American Music (New York: Vintage Books, 1984), 58).
Minimal Art, Arts Magazine, January 1965; reprinted in Gregory Battcock (ed.), Minimal
Art: A Critical Anthology (New York: Dutton, 1968).
66 Art

withholding, or restraining, in relation to media, techniques, styles, or aims

available in the artists milieu. If it is possible for an artwork whose genesis
predominantly involves negative work to exhibit creativity, then surely the
creativity involved in some of Cages more radical musical self-removals is of
that sort.
As it happens, Elsters skepticism about the value of avant-garde modes of
art is not restricted to music: [literary] modernism fails because it rests on an
erroneous criterion of adequacy of a text to its object. Just as a flawed world
is not better represented by a flawed text, a fluid or chaotic world is not better
represented by a fluid or chaotic text (190).
But this blanket condemnation ignores, among other things, the exempli-
fying and expressing functions of a literary text, as opposed to its represent-
ing function per se. As Goodman and others have stressed, artworks mean in
ways other than by representing. Thus, a work of art may at least need to
approach, if not wholly embrace, chaos and fluidity in order to exemplify or
express, rather than just represent, those properties.
8. I return now to Elsters discussion of the experimental novel La Dispar-
ition, invoked at the beginning of this essay:
Preexisting or preset constraints enhance and stimulate the creative process . . . Yet
not all constraints will do equally well. The constraint of writing a novel without
the letter e may not make the task of the writer more difficult than writing in the
demanding form of terza rima. Yet the latter constraint, unlike the former, can
contribute directly to aesthetic value, over and above the indirect contribution that
follows from the focus-enhancing effect. Because rhythm and meter generate an
organized form, they have intrinsic aesthetic potential; the absence of a given letter
in the alphabet does not. (209)

But one can argue that the systematic withholding of a certain letter, espe-
cially one as central to the French language as the letter e, gives by its absence
a kind of form to Perecs text, though obviously not form in the sense of meter
or rhythmic scheme. And that form can even be expressed in a positive man-
ner, as the requirement to craft sentences using only the vowels a, i, o, and u.
For most writers, having to write around the letter e would be an irritating distraction
rather than a stimulation, an obstacle to be overcome rather than a challenge to be
met. (210)

See his Languages of Art (Indianapolis: Hackett, 1976) and How Buildings Mean, in
Reconceptions in Philosophy and Other Arts and Sciences (Indianapolis: Hackett, 1988).
 Elster on Artistic Creativity 67

That is no doubt true. But even so, it only means that willingly avoiding the
letter e in writing would not, for most writers, be a spur to creativity. This says
nothing, however, as to whether it was in fact a spur to creativity in a given
case, namely that of Perec.
Recalling Elsters useful trichotomy of chosen, invented, and imposed
constraints, we can ask whether the express avoidance of e in La Disparition
counts as a chosen or an invented one. The question, in effect, is whether
that sort of restriction preexisted Perecs writing of his novel, or whether in
so writing it, and in writing it so, he instituted that sort of restriction on
literary production. Though Elster seems inclined to treat La Disparition as
a case of invention, there is actually reason to see it as a choice to operate in an
existing category, that of the lipogram, but in a more thoroughgoing way than
had been previously attempted. A lipogram is precisely a text produced under
the constraint of eschewing a given letter or letters of the alphabet, and was a
form much favored by the group of avant-garde French writers to which Perec
belonged, the Oulipo or Workshop of Potential Literature.
9. Constraints like that adhered to by Perec in writing La Disparition are
sometimes derided as purely arbitrary and unmotivated, and thus as issuing
in nothing of any artistic worth. But if artistic creativity is, as Elster pro-
poses, centrally a matter of operating innovatively within constraints, then
how does one know which constraints are likely to give rise to artistic value
and which not? Can there be a reasoned prospective criterion of goodness of
constraints, rather than the obvious retrospective one of judging them by their
fruits?
These are difficult questions, which I am unsure how to answer. I want
instead just to consider whether the success of Perec in crafting La Dispar-
ition should indeed be regarded as without artistic point, a pure exercise in
self-binding for its own sake. I think not. For one, there is no reason to
regard the skill, cleverness, and ingenuity of Perec in meeting the constraint
he set himself as somehow beyond the pale of artistic appreciation. Surely we
are practicing artistic appreciation when we admire the interest that Barnett
Newman gave his paintings of very peculiar vertical format, roughly six feet
high and four inches wide, or the elaborately constructed montage in the

The name is a contraction of Ouvroir de Litterature Potentielle.

This is a good point to mention that there has been, astonishingly enough, a translationor
really, recreationof La Disparition in English. It is by the novelist Gilbert Adair, and is cleverly
titled A Void (San Francisco: Harper, 1995). Adairs work presumably has some of the virtues of
Perecs, but not all, and arguably has some that Perecs does not.
68 Art

Odessa Steps sequence from Eisensteins Battleship Potemkin, or the

resourcefulness with which Mozart manages the voices in the sextet from The
Marriage of Figaro, or the magnificence of Racines alexandrines, unfailingly
ten syllables to a line and a rhyme every time. Virtuosity, which broadly con-
strued covers those cases and that of Perec, is unquestionably an artistic virtue,
if not the highest of them, and is something appreciable for its own sake.
The interest of La Disparition, in other words, is surely not merely that which
Samuel Johnson infamously allowed that the phenomenon of female sermon-
izing might be said to have.
But the interest of La Disparition may in fact go beyond the attention its
virtuosic surmounting of challenges properly merits. This is due to special fea-
tures of both the writer and the context of writing. Born in 1936, Perec was
the child of Polish-Jewish parents who died during the Second World War,
one in the Resistance and the other in Auschwitz; he was subsequently raised
in a Catholic boarding school in the Isere during the Occupation. Is it too
much of a stretch, then, to see La Disparition, published in 1969, as referring
not only to the victim in the murder mystery it ostensibly relates, and not only
to the letter e mysteriously banned from its pages, but to the disappearance
from Europe not long before of millions of individuals as arbitrarily declared
persona non grata as that unlucky vowel? Might not La Disparition be seen as
a meditation on how easy it was to get rid of, or do without, six million Jews,
as much as an exercise in getting rid of, or doing without, tens of thousands of
es? We know also that every one of Perecs published works is an exercise in a
different style. Does this not fit all too well with one who had to dissimulate
and reinvent himself from a very early age?
In any event, if such further dimensions are attributable to what might
seem the purely formal stylistic exercises of a Perec, they give those exercises
added value, in virtue of the fittingness between content and form invoked
earlier as a primary locus of aesthetic value.

See Thomas Mark, On Works of Virtuosity, Journal of Philosophy 77 (1980): 2845.

Sir, a woman preaching is like a dog walking on hind legs. It is not done well, but you are
surprised to find it done at all.
The roman nouveau of Alain Robbe-Grillet, Nathalie Sarraute, and others might be mentioned
here as another example of an artform subject to self-imposed constraintsones as Draconian in
their way as those under which Perec workedwhere the resulting stylistic spareness and sobriety of
detail has expressive value. Robbe-Grillets novel La Jalousie, for example, arguably expresses a world
of individuals locked into their separate consciousnesses, their phenomenological identification
with the objects of their experience suggesting a kind of ultimate isolation and alienation. But
such expression is clearly a product of the severely detached and methodically descriptive mode of
narration adopted by the author.
 Elster on Artistic Creativity 69

10. Shifting to cinema for another case of chosen constraints of an

ostensibly limiting sort, consider those with which Sidney Lumet operated in
making his powerful film of jury deliberation, 12 Angry Men: a single, sparsely
furnished room; a fixed set of characters; a continuous segment of time; and
black and white cinematography. Yet 12 Angry Men is as transfixing and
rewarding a film as cinema has to offer. Clearly in this case Lumets chosen
constraintshis form of provisional self-bindingwere a spur to creativity
rather than a shackle on it.
Some of Lumets specific techniques for sustaining cinematic and
dramatic interest given those constraints deserve remark: (a) focusing
on different characters, and different character pairs, at different points;
(b) temporarily subtracting characters through the device of exits to the
restroom; (c) rhythmic but not mechanical varying of long shots, medium
shots, and close-ups; (d) alternating of evidence-sorting and reason-weighing
discussions, on the one hand, with character-revealing and tension-provoking
personal encounters, on the other hand, so far as the pattern of scenes is
concerned.
So Lumets film fits nicely enough Elsters picture of artistic creativity as
successful maneuvering within constraints. But it is worth observing that
even if all artworks involve choices made within constraints of some sort or
otherwhether self-imposed or imposed from without, whether revisably
or unrevisably adopted, whether conventional or idiosyncraticthere is a
useful distinction to be made between artworks which foreground the fact of
their operating under constraints, and those which do not do so, in some
cases even managing to make us unaware of the constraints that governed
their creation. In the first instance works are partly about their constraints,
i.e. about virtuosity or the surmounting of challenges, whereas in the second
instance they are not. In the first category belong, obviously, La Disparition;
Hitchcocks Rope, shot virtually in one continuous take; and Machauts
rondeau Ma fin est mon commencement, a musical palindrome. In the
second category, more likely, belong 12 Angry Men, Coltranes Giant Steps,
and the sonnets of Keats or Millayworks produced within formal confines
but which do not especially draw attention to them.
Twelve-tone music, the basic principle of which is that a given pitch
cannot be repeated until all eleven other pitches of the chromatic scale
have been soundeda method which effectively destroys the feeling of
tonality or keyis sometimes offered as a mode of musical composition
that is too confining, one involving constraints that fetter rather than abet
70 Art

creativity. But it seems as if there are simply different cases here, which
make any such generalization suspect. For Schoenberg the method was a
way to inject new local interest into old global forms associated with Brahms
and his predecessors in the Viennese Classical tradition, an effort in which
Schoenberg succeeded notably in the Third and Fourth String Quartets, the
String Trio, the Piano Concerto, and the Serenade Op. 24. For Berg the
method was from the outset something to be intuitively modified and relaxed
in the interests of dramatic expression, not a code requiring strict adherence,
and yet it remains essential to the rigor and power of his twelve-tone
masterwork, the opera Wozzeck. For Stravinsky, who came to twelve-tone
music late and after much resistance, the method offered itself as something to
selectively adapt and fuse with his preexisting neo-classical style, with perhaps
greatest success in the ballet Agon; in other cases, e.g. the Movements for Piano
and Orchestra, the results were crabbed and less appealing. Finally, for some
composers, like the American George Rochberg, twelve-tone musicor
rather, its subsequent generalization, serial musicbecame an indirect prod
to creativity by serving as an orthodoxy against which to rebel, leading to that
musics total abandonment in Rochbergs largely tonal and pointedly neo-
romantic Third String Quartet.
11. One of Elsters examples of an externally imposed, yet ultimately
fruitful, artmaking constraint concerns the nonfeasibility, prior to 1940, of
recording popular music over three minutes in length, due to the limits
of the old 78 rpm technology. In relation to this he comments that jazz
improvisation at a high level of quality is so hard to sustain that the 78 record
with three minutes playing time was just about optimal (194).
To my ears, this sounds like a too convenient retrospective Leibnizian
explanation. For my impression of even the best solos of the great jazz players
in the 1930s big bands, such as Coleman Hawkins, Ben Webster, Johnny
Hodges, or Lester Young, is that those solos would in fact usually have
benefited from being somewhat longer, that they sometimes hardly get going
before they are over. Certainly, there are many cases of mesmerizingly good
improvised jazz solos extending over large stretches of time, for example, those
by Coltrane and McCoy Tyner in My Favorite Things, which last upwards
of four minutes each. Is it not rather plausible to think we might have had
even more sublime solos from Hawkins and Young had they been able to
dilate, on occasion, for two minutes or even three?
12. Let us turn to two other examples, ones which, at least at first
blush, comport well with Elsters thesis. The first is Michelangelos celebrated
 Elster on Artistic Creativity 71

Laurentian Library vestibule. Michelangelo, who had no formal training as

an architect, accepted as a constraint to employ the vocabulary of High
Renaissance architecturecolumns, corbels, pilasters, balusters, pediments,
niches, and architravesbut not to employ them in the traditional ways.
Rather, he set out to employ them, it seems, in pointedly illogical, that
is, non-structurally justified ways, with the evident aim of more effectively
stimulating the free play of the imagination in their regard. Michelangelo,
untrammeled by habits he might have acquired had he been of the architects
guild, was instead freed to create expressive form for its own sake: this
anteroom, with its unusual somber white and grey color scheme, its high
and narrow spatiality, its flowing and spreading staircase, and its constellation
of familiar yet oddly deployed constructional elements, is one of the most
imposing architectural ensembles I know. Its success indeed seems to be
a function of adhering to certain constraints derived from the prevailing
culture, while simultaneously explicitly flouting others. As Vasari apparently
said of Michelangelos architecture, it broke the bonds and chains of
common usage.
A second example is Picassos creation of the Portrait of Kahnweiler
as a response to both a highly general, tradition-rooted Charge and a
more specific, circumstances-involving Brief, in the terms of the convincing
analysis of Michael Baxandall. Picassos Charge, suggests Baxandall, was
the standard painterly one of producing an object with intentional visual
interest, while Picassos Brief, largely self-assumed, was the three-pronged
one of reconciling the claims of color and form, reconciling the claims of two-
dimensional patterning and three-dimensional representation, and being true
to painting as both cumulative process and completed image. According to
Baxandall this posited Brief, together with the assumed background Charge,
explain much about how Picassos portrait actually turned out.
Still, with both Michelangelos vestibule and Picassos portrait, there is a
real worry as to whether our hypothetical reconstructions of the constraints
under which and in response to which those works emerged are not merely
just-so stories. Certain constraints are perhaps undeniable and document-
able, particularly material and physical ones. But hypotheses about constraints
at the discretion of the artist, however, seem more liable to be products of

Quoted in H. H. Janson, History of Art (New York: Harry Abrams, 1962).

See Patterns of Intention: On the Historical Explanation of Pictures (New Haven: Yale University
Press, 1985).
72 Art

unfettered backward thinking in light of what actually results. The fact that
hypotheses are usually underdetermined by evidence urges caution here, at
least, as to the reliability of such speculative reconstructions.
Consider now a case of maximizing artistic value by violating a freely chosen
constraint. This is illustrated by an anecdote told of Arthur Rubinstein during
one of his first concerts in Paris in the 1920s. The first piece on his program,
the then recently composed Valses Nobles et Sentimentales of Ravel, had not
been well received, so following a second half consisting of pieces of Chop-
in that was rather better received, Rubinstein answered the clamor for an
encore by playing the Valses Nobles et Sentimentales once again. Now the art
in question here is, of course, a performing one, that of recital-giving, and the
constraint in question, which comes with the territory, is that encores be both
shorter than, and other than, the pieces making up the main program. But
arguably Rubinstein gave a finer recital on that occasion precisely by flouting
one of the defining norms of the performing art he was engaged in.
13. Is there a necessary connection between creativity and problem-
solving? Assume for the sake of argument that there is a necessary connection
between creativity and constraints, that creativity is roughly, as Elster
suggests, a matter of value-maximizing choice within constraints. The
question then becomes that of the relation between problem-solving and
constraints. Are all cases of constraints in operation cases of problems calling
for solution?
One reason to deny this is that a problem, unlike a mere set of constraints,
has something like a principle of unity, being often a matter of resolving
some conflict, or meeting some need, or finding the answer to a definite ques-
tion. Another reason is that a problem has at least a prima facie claim on
ones attentionthat it is something that, at least on the face of it, merits
addressing. Thus it is a problem how to compose comprehensible and enga-
ging music outside the bounds of tonality, or how to maintain the integrity
of the picture plane without giving up blatant imageryproblems that were
solved respectively by Schoenberg and Berg, on the one hand, and Johns and
Dubuffet, on the other. But writing a poem of fifteen lines that employs the
word dog in every line, or making a film that last fifteen minutes and costs
less than $150,000, would seem to be cases of operating under constraints,
but not really cases of solving problems. The two notions, then, should not be
thought interchangeable.
14. By way of conclusion, let me move from the sphere of fine art to that
of applied art, and briefly contemplate constraints on the design of chairs,
 Elster on Artistic Creativity 73

those varied receptacles for our backs and bottoms. The case is well chosen,
I think, to illustrate the reciprocity between constraints and creative activity
conducted under such constraintsa reciprocity to which Elsters analysis
of creativity, however illuminating, seems insufficiently sensitive. For highly
innovative chairs challenge, question, and refashion the very idea of a chair,
subjecting its presumed necessary featurese.g. possessing legs, being sturdy,
having a seat, being articulatedto intense conceptual pressure.
Consider, in this connection, the intentional-historical theory of artifact
concepts proposed by the psychologist Paul Bloom. Blooms theory takes
off from and generalizes upon my own intentional-historical theory of the
concept of artwork, whose basic idea is that an artwork is something that is
intentionally connected by its maker to a preceding practice or tradition of
artmaking, possibly identified only through exemplars rather than a covering
description, and whose art status is taken for granted. On Blooms related
theory a chair, say, is not something of a certain circumscribed form or some-
thing fulfilling an unambiguously understood function, but roughly some-
thing intended to belong to the category of chairs as antecedently exemplified,
that is, something intended for regard or treatment as preexisting chairs were
regarded or treated.
The immense profusion of chairs displayed in the astounding book 1000
Chairs, which presents the most comprehensive survey of chair design from
1808 to the present, not only makes Blooms theory attractive, but makes a
theory of that sort almost unavoidable, rendering doubtful any conception
of chairs in terms of essential shape or structure or purpose. The assault
on chairhood represented by the many outre examples of chairs in 1000
Chairs might profitably be compared with the challenge to what a building is
provided by the most envelope-pushing architectural offerings of postmodern
architects such as Frank Gehry, Peter Eisenmann, and Rem Koolhaus.
Now Elster does, it is true, recognize a distinction between what he calls
rebellion and revolution in art, which he glosses as follows: Rebellions violate
existing conventions, whereas revolutions abolish them and create new ones
(p. 223). What I am gesturing at in the case of chairs and buildings, how-
ever, is something in-between, neither the simple abrogation of existing rules

Intention, History, and Artifact Concepts, Cognition 60 (1996): 129.

See Defining Art Historically and Refining Art Historically, in Music, Art, and Metaphysics,
and Extending Art Historically, in The Pleasures of Aesthetics.
Charlotte and Peter Fiell, 1000 Chairs (Cologne: Taschen, 1997). The quote is from the back
cover.
74 Art

for such things nor the institution of new rules in their place, but the forced
reevaulation, by an ostensible chair or building, of what the rules for such
things really were or are. Creativity, in short, is sometimes a matter or recon-
ceiving or reinterpreting or reconstruing given constraints, and not always a
matter of either remaining inventively within them or entirely abandoning
them.
PART I I
MUSIC
This page intentionally left blank
 5
Sound, Gesture, Space, and the Expression
of Emotion in Music

I INT RODUCT ION

The main concern of this essay is musical expressiveness, and the role that
the imaginative construal of sound in gestural ways plays in grounding that
expressiveness. But I will come to that main concern obliquely, beginning
instead with some reflections on the nature of sounds and the sense of hearing.
What I will be arguing, in the central sections of this essay, is that to hear
the expressiveness of music is to hear it as personal expression; that to hear
it as personal expression is to hear a sort of gesture in the music, or the
music as gesturing in a certain manner; that to hear such musical gesture is
to deploy a capacity to imagine in spatial terms, most obviously because that is
required for apprehension of the behavioral gestures and performing gestures
that underlie musical gesture. Thus, grasp of expressiveness is music is not
detachable from a comprehension of the range of expressive human gesture
and a comprehension of the means by which music is actually produced, each
of which implicates possession by the hearer of a robust spatial imagination.

First published as Sound, Gesture, Spatial Imagination, and the Expression of Emotion in Music,
European Review of Philosophy 5 (2002): 13750.
This essay, originally presented at a colloquium entitled Objets et espaces sonores held at
the Ecole Polytechnique in Paris in March 2001, draws significantly on two earlier essays of mine,
Authentic Performance and Performance Means, in Music, Art, and Metaphysics (Ithaca: Cornell
University Press, 1990), 393408, and Musical Expressiveness, in The Pleasures of Aesthetics (Ithaca:
Cornell University Press, 1996), 90125. For a broader perspective on problems of emotion in
relation to art see my Emotion in Response to Art: A Survey of the Terrain, in Mette Hjort and
Sue Laver (eds.), Emotion and the Arts (Oxford: Oxford University Press), 2034 (and reprinted as
Ch. 3 in this volume).
78 Music

II THE NATURE OF SOUNDS

At the heart of their book, La Philosophie du son, Roberto Casati and Jerome
Dokic propose what they call a theorie evenementielle of sounds. Accord-
ing to Casati and Dokic, sounds are neither auditory sensations occurring in
our heads, nor sound waves arriving at our ears when we hear, nor secondary
qualities of any sort, but rather something happening in the resonant object
itself. They summarize their theory in the following formulation: sounds are
vibratory events involving objects.
I believe Casati and Dokic are right in that contention. Sounds are indeed
events, and they are located just where the objects that produce them or in
which they occur are located. But it seems to me that the proposal they
advance is in need of some qualification, since not every vibratory event that
concerns an object can reasonably be accounted a sound. For example, the
atomic and subatomic vibrations that characterize every parcel of matter, the
slow oscillation of an aspic de legumes, the gentle trembling of the hand of
a sufferer from tremor, are all vibratory events in objects, but not thereby
sounds involving those objects.
On the other hand, as Casati and Dokic rightly insist, we should not require
of sounds that they be humanly audible, otherwise we would unfairly exclude
ultrasounds and infrasounds, which we have reason to consider sounds from a
philosophic point of view even though they do not lend themselves to human
hearing. We might thus add to the formulation offered that sounds are vibrat-
ory events in objects that are either humanly audible or of a physical order
comparable to those that are humanly audible, and so at least a possible object
of audition to other creatures we might credit with a sense of hearing.
One might also remark that, without modification, Casatis and Dokics
proposal, identifying sounds with vibratory events in objects, has the con-
sequence that the music of the spheres, of which the ancients spoke, is not
after all a metaphor, but something to be taken literally. At least that is so
if one supposes, as seems reasonable, that the movements of revolution and
rotation of the planets are a sort of vibration involving them. However, I do

Roberto Casati and Jerome Dokic, La Philosophie du son (Nmes: Editions Jacqueline Cham-
bon, 1994).
Les sons sont des evenements vibratoires interessant un objet (La Philosophie du son, 49).
The same thesis, one may note, is defended in Robert Pasnau, What is Sound?, Philosophical
Quarterly 49 (1999): 30923, though with less precision.
 The Expression of Emotion in Music 79

not regard it as a plus for a theory of sounds to count among sounds the sup-
posed music of the spheres.

III THE APPRECIATION OF MUSIC AND THE SENSE

OF HEARING

One of the main claims in La Philosophie du son is that the sense of hearing,
properly speaking, involves the formation of beliefs regarding the spatiality
of sounds understood as vibratory events in objects, and thus presupposes an
image of the space in which such events are located. Casati and Dokic go on to
relate this conception of hearing to the phenomenon of music, in the follow-
ing striking observation:
If what counts in the appreciation of music is being able to have auditory sensations,
rather than being able to form beliefs regarding the location of sounds in the envir-
onment, it follows that to appreciate music there is no need, strictly speaking, of a
sense of hearing. One could in effect conceive of a perfectly solipsistic musical sub-
ject, one without any beliefs concerning external sonic events, but who, in prey to
hallucinations, would hear whole symphonies.
In my opinion this observation is partly just and partly not. I agree with
Casati and Dokic that the sense of hearing, robustly understood, goes bey-
ond the capacity to have auditory sensations, and involves the ability to form
beliefs about the sources of sounds heard. But these beliefs, we may note, are
of two kinds: one concerns the spatial location of the source of the sound rel-
ative to the hearer, while the other concerns the spatial nature of the source
of the sound, apprehended in terms of size, shape, movement, and orienta-
tion. There are thus at least two separable abilities involved. Let us label the
first of these abilities specific spatial imagination and the second generic spatial
imagination.

Si ce qui compte, pour apprecier de la musique, cest davoir des sensations sonoresplutot
que detre capable de former certaines croyances quant a la localisation des sources sonores presentes
dans lenvironnementalors pour apprecier de la musique, point nest besoin de loue. On peut en
effet concevoir un sujet musical solipsiste, qui naurait aucune croyance sur des evenements sonores,
mais qui, en proie a des hallucinations, entendrait des symphonies (La Philosophie du son, 28).
An illustration of the distinction is the following. In perceiving a particular sound I might have
the thought that it is situated down the street, and moving in my direction, and also the thought
that it is the sound of a car engine, understood as an object of a certain shape consisting of parts
moving relative to one another in a certain way. The former thought is the product of specific
spatial imagination, the latter of generic spatial imagination.
80 Music

Now it is true that to appreciate music of the standard sort we may perhaps
dispense with beliefs of the first kind, and so with the first kind of spatial ima-
gination. But it is not at all clear that we can dispense with beliefs, and so
spatial imagination, of the second kind, and thus at least a part of the sense
of hearing robustly understood. The reason is a fact about the appreciation
of music that I will be foregrounding in this essay, namely that to appreciate
traditional instrumental music adequately we have need at least of generic spa-
tial imagination, enabling us to imagine possible sources of heard sounds, and
possible agents and actions that could be responsible for those sounds. Thus,
although Casati and Dokic may be right that appreciation of such music does
not necessarily call upon the full sense of hearing, they are wrong to think that
the mere capacity to experience auditory sensations could be adequate to such
appreciation. A satisfactory account of a listeners grasp of musical expressive-
ness entails that the listener has the capacity to imagine in spatial terms the
possible sources of sounds heard.
More broadly, we can take the question to be whether the appreciation
of music is detachable from the normal operation of the senses with which
human beings have been endowed. Assuming a capacity to imagine actions
in a real space as possible sources of sounds heard is appreciatively relevant to
most music, and that that capacity could only be firmly in place if at least one
sensehearing, vision, or touchcapable of providing spatial information
and of priming the spatial imagination was at the disposal of the subject, the
answer must be no. Contra the solipsistic supposition entertained by Casati
and Dokic, then, the appreciation of music is not entirely separable from the
senses in their normal functioning.
A key premise in the argument just sketched is that the imagination of the
sources of sounds, in terms of generating actions and objects, is relevant to the
appreciation of music of the ordinary sort. I now devote attention to establish-
ing that.

IV PE RFORMING GESTURE, MUSICAL GESTURE,

AND SPATIAL IMAGINATION

What I hope to show in this section is that a part of the expressive character of
a passage of traditional instrumental music as heard derives from the impres-
sion one has of the manner in which it has been produced and the correlation
of that impression with the sound of the passage narrowly construed. The
 The Expression of Emotion in Music 81

sound of the sound, so to speak, does not suffice in itself to fix the expressive
character of the passage, but only that sound in conjunction with a presumed
manner of production. The same sounds present different appearances, and
affect us differently, according to the notions we entertain at the same time
regarding the actions or processes that have engendered them.
Mozarts Serenade in E-flat, K. 375, begins with a unison statement of the
full cohort of windsoboes, clarinets, horns, and bassoonsa five-note dot-
ted phrase with marked sforzandi. That beginning has an assertive quality,
and suggests a call to attention. But from what, exactly, does that assertive,
attention-getting character derive? It derives, in part, from the fact that that
stretch of music comes across as an instance of honking, as passages involving
double reed instruments played forte often do. Given the manner in which
the sounds have actually been produced, that is, by actions of wind players
similar to those of waterfowl engaged in honking, that passage could in fact be
considered to be a honking, broadly understood.
Hearing the sounds of which that stretch of music is composed as hav-
ing been produced by the rapid passage of air through narrow openings in
wooden tubes, that is, as a honking, rather than by some entirely other pro-
cess, makes for an experiential difference. Were we convinced that the sounds
we were hearing did not result from the playing of wind instruments in the
normal manner, but issued instead from a perfect synthesizer, that stretch of
music would make a somewhat different impression on us. The reason is that
a synthesizer, however powerful or accurate, cannot truly honk. That is to say,
it cannot do what oboes, klaxons, and geese all naturally can. And this mod-
ifies, in a subtle manner, how the sounds the synthesizer emits are received,
since those sounds are gauged against the background of a presumed manner
of production.
I have been arguing that knowledge of the actual sources of sounds affects
how those sounds strike us. I have also been suggesting that such knowledge
properly enters into aesthetic assessment of passages of music, which are pre-
sumably to be taken as what they are, say the product of windplaying, and
not what they are not but only appear to be, as when a synthesizer mim-
ics such windplaying. But in fact that is a stronger thesis than is needed for
present purposes, so let me also articulate a weaker, less controversial one. All
that is needed to show the indispensability of generic spatial imagination to a
grasp of the expressive character of music is that a sense of the possible sources
of sounds, not necessarily knowledge of their actual sources, is required if the
expressiveness of sequences of sound is to be registered.
82 Music

Clearly the assertiveness of the opening of Mozarts serenade only emerges

if that opening is heard as a woodwind proclamation, leaving aside whether
on a given sounding occasion it actually is, and whether it is known to be.
This point is easily generalized. Descriptions of musical passages in terms
of actions are extremely widespread. Passages can be described as sighing,
whispering, chirping, purring, squawking, roaring, sawing, hammering,
pounding, slashing, caressing, swooping, and so on. Now obviously one
cannot recognize the applicability of such descriptions without possessing an
image of the action literally denoted. For example, one cant hear a passage as
sighing unless one has a conception of what sighing is. But the applicability
of such descriptions is also predicated on a sense of related performing actions
from which the sounds in question are assumed to result. Absent a sense of
those background generative actions such passages would not seem to merit
those descriptions, or at any rate, would not do so unequivocally. Ones
sense of a passage as a slashing one is not completely detachable from the
impression that it has been produced by a violin played staccato, and ones
impression of a passage as a purring one is similarly not entirely independent
of the impression that it has been produced by an oboe played legato. A
slashing or purring sound not thought of as the product of those sorts of
performing actions does not slash, does not purr, in quite the way it does
when believed to issue from violin or oboe.
Now just as generic spatial imagination is required to form a conception of
the actions literally denoted in the action-based descriptions recalled above, so
is it required to envisage the background performing actions behind musical
passages which contribute to the applicability of those descriptions. For both
sorts of actions, e.g. the literal roaring we may hear in a passage of music, and
the rising declarations in the brass that underlie it, necessarily manifest them-
selves in space. And these action-based descriptions, in turn, have an obvious
bearing on the expressive character musical passages are perceived to have, giv-
en the aptness of many such actions to the expression of emotion.
Let me approach the conclusion that spatial imagination is necessarily
involved in a grasp of musical expression in another way. The expressiveness
of music is grounded on the factwhich I will shortly explore in more
detailthat the actions or gestures that one hears in a passage of music recall
the actions or gestures that serve as behavioral expressions of emotions, which
allows us to hear the former as the latter, and so the passage as expressive of
 The Expression of Emotion in Music 83

those emotions. But the gestures rightly heard in music are only heard in all
their specificity if the apparent performing gestures behind the sequences of
sounds per se are taken into account. For the gestures we are right to hear in
musical sequences are those we hear in them when we are cognizant of the
instrumental actions understood to generate such sequences.
It is important to be clear on a crucial step in the preceding argument. It is
not that the gestures heard in a passage of musicwhat in effect we hear the
music as doing are identical with the performing gestures that we imagine as
responsible for the sounds we hear. It is rather that those gestures, which for
clarity I label musical gestures, are in part a function of the performing gestures
we are right to imagine in auditing a given passage. In other words, the ges-
tures we are correct to hear in a passage of music, and on which depends our
estimation of its emotional expressiveness, are partly determined by what we
take performers of the passage to literally be doing in producing it.
Some additional examples will serve to support this thesis of the intimate
relationship between performing and musical gestures, and thus ultimately,
of the necessary involvement of spatial imagination in the grasp of music-
al expression. Consider the rapid upward keyboard glissando. This familiar
musical formula conveys most often an impression of gaiety or insouciance.
The best-known instances of it are probably Chico Marxs one-finger antics
in various Marx Brothers movies, but a number of examples enliven as well
the scores of Ravel, Prokofiev, and Gershwin. I claim that the characterist-
ic impression produced by the upward keyboard glissando has its source, in
part, in an image we form of the flicking or sweeping gesture behind the rising
tones that we hear, a gesture that occupies a certain place for us in the field of
expressive behavior as a whole. The imagined executive action melds with the
rising tonal movement so as to create a musical gesture in which we readily
hear a gay insouciance.
One of the most vivid portraits in all music is that found in the sixth section
of Job: A Masque for Dancing by Ralph Vaughan Williams. There the composer

A word about the gestures of conductors during performance is in order. Such gestures,
probably the ones most on display in the musical arena, do not count as examples of what I mean by
performing gestures. That is, such gestures do not figure among those a listener must have an image
of in order to grasp the expressiveness of the music being heard, that is, in order to grasp its musical
gesture. However, a conductors gestures might very well symbolize the musical gestures contained
in the music being conducted, and reflect the conductors understanding of those musical gestures,
but it is only rarely that there is anything approaching identity between musical and conductorial
gesture at a given point.
84 Music

depicts the hypocrites gathered around the suffering Job. The slimy nature
of these false comforters and the flavor of their viscous consolations could
not be more clearly conveyed. The musical phrases charged with achieving
thisslow, descending seconds and thirds, alternately major and minor, in a
mellow timbreare perfectly apropos, but they only make their optimal effect
if they are taken precisely as the saxophone gestures they are. It is only in the
light of that image that the music presents itself unequivocally as a stylization
of the whining, honey-tongued vocal behavior typical of hypocrites.
Take finally the standard expressive contribution of percussion in orches-
tral music. Notes on the timpani carry a powerful association of pounding or
batteringits a matter of mallet strokes, after alland in passages where
timpani are prominent this effect can be overwhelming. However, it is not
timpani sound per se that is efficacious, but the conjunction of such sound and
the action of which that sound is the sign. That a phrase sounds a certain way,
in the narrow sense, is not the only thing that counts so far as the musical ges-
ture we will be disposed to hear in the phrase is concerned; equally important
is our sense that the phrase has been sounded in a certain way. In the scherzo
of Beethovens Ninth Symphony the timpanis statements are aggressive and
interruptive, while in the first movement of Nielsens Fifth Symphony those
of the snare drum are warlike and menacing, in virtue of their maddening
repetitiveness. But in both cases, these effects are precisely what they are only
if the utterances in question are heard as strikings and hammerings, and not as
pure disembodied sequences of sound.
This is a good place to offer a further observation on musical gesture, which
as we have seen depends on performing gesture though without reducing to it.
In many cases in which musical gesture is perceived one is able to characterize
the musical gesture in familiar terms, metaphorically employed; for example,
one might say of a passage of music that it was dancing, or striding, or thrust-
ing, or hectoring, or meditating, or recoiling, or questioning, etc. But this will
not be true in all cases in which one nonetheless registers gesture in music;
in such cases one might perceive music to be gesturing in a certain manner,
and as a result hear the music as, say, heroic, but without being inclined to
identify, or even being able to identify, the manner of gesturing in terms of
some prior category of action.

Of course it will always be possible to characterize retrospectively the manner of gesturing as

heroic, but that would not be to invoke an antecedently familiar action concept so as to characterize
the manner of gesturing heard in question.
 The Expression of Emotion in Music 85

V M USICAL EXPRESSIVENESS AND

HEARABILIT Y-AS-PERSONAL-EXPRESSION

At several points in the preceding discussion I have leaned on a theory of

musical expressiveness, but without elaborating it. It is now time to lay out
that theory more clearly. I cannot here undertake a full defense of the the-
ory, which I have argued for at length elsewhere, but I will try to make its
attractions evident.
The central idea is that musical expression of mental statesand for reas-
ons of simplicity I stick to the most important case, that of emotionsmust
model itself on the primary expression of emotions by human beings, that
is, expression of emotions through behavior or other outward manifestation,
including countenance, posture, bearing, demeanor, actions, gestures, and
modifications of voice. That is not to say that music is expressive in precisely
the same sense as is behavior, for several reasons. One, music doesnt actu-
ally exhibit behavior as such; two, musical gesturing is not literal but rather
metaphorical; three, the character of expressive music is neither the immedi-
ate result of emotion experienced by the music, music not being sentient, nor
always the upshot of emotion experienced by the composer; and so on. But
despite these manifest differences between musical expression of emotion and
human expression of emotion, it seems we cannot consider a piece of music
to be strictly expressive of an emotional state Srather than, more loosely,
simply possessing a correlative emotional quality Qunless we regard it as
analogous to a being endowed with sentiments capable of announcing them-
selves in an external manner. In short, music expresses an emotion only to
the extent that we are disposed to hear it as the expression of an emotion, but
through different means, by a person or personlike entity.
More formally, what I have proposed as an analysis of expression in music
goes like this: a passage of music P is expressive of an emotion E iff P, in
full context, is readily heard, by a listener appropriately backgrounded in the
musical genre in question, as the expression of E in a sui generis, purely music-
al manner, by an indefinite agent, what we can call the musics persona. The
concept of musical expressiveness is thus a complex one, in which the ideas of

See Hope in The Hebrides , in Music, Art, and Metaphysics (Ithaca: Cornell University
Press, 1990), 33675, and Musical Expressiveness, in The Pleasures of Aesthetics (Ithaca: Cornell
University Press, 1996).
86 Music

personal expression, human gesture, musical gesture, imaginary agency, and

hearing-as each plays a role.
What is it to hear a stretch of music as something else? For present purposes
it suffices to locate hearing-as among perceptual acts that partake freely of or
substantially enlist the imagination. The agents one hears in music when one
hears it as an expression of emotion are thus inescapably imaginary ones, ones
displaying the indefiniteness characteristic of all imaginary objects. To hear
music as such and such is, perhaps, to imagine that the music is such and such,
and, more specifically to imagine of the music, as you are hearing it, that it is
such and such.
In earlier discussions of the problem of musical expressiveness I
allowed myself to alternate the locution hear X as Y with the more
counterfactuallyflavored locution hear X as if it were Y, suggesting that in
connection with music these come to more or less the same thing. I am now
less sure of that, so here confine myself to the first locution. The notion of
hearing X as if it were Y may not be an entirely happy one. I suspect that
is because the modifier as if it were is more appropriate to acts of treating,
regarding, or behaving towards than it is to acts of perceiving as such. For
example, though one clearly understands what it would be for X to treat Y as if
she were the only person in the world, it is less clear that one understands what
it would be for X to see Y as if she were the only person in the world.
What about the evocation of emotion by music? Is not music that is express-
ive of an emotion precisely music that evokes in us this same emotion, or that
at any rate has a disposition or tendency to evoke in us that emotion? This is
not the place to demonstrate that this popular theory of musical expressive-
ness is erroneous. However, neither is it necessary. Because we can willingly
admit that the emotion, and also the feelings and sensations, that a passage of
music has a tendency to evoke in a listener influences what emotion the listen-
er will be disposed to imagine the music to be the expression of, without it
being necessary to identify the emotion expressed with that emotion, or those

For further ref lections on the concept of musical expressiveness, and doubts whether the
expression that music is heard as when its expressiveness is being perceived should be held to be a
sui generis mode of expression, rather than just expression tout court, see my Musical Expressiveness
as Hearability-as-Expression (Ch. 6, this volume).
Those cited in n. 9.
But see Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press,
1994), and Peter Kivy, New Essays on Musical Understanding (Oxford: Oxford University Press,
2001). Still, the evocation theory of musical expression continues to have advocates, of whom Derek
Matravers, Art and Emotion, Oxford: Oxford University Press, 1998), is the most sophisticated.
 The Expression of Emotion in Music 87

feelings and sensations. What one has a tendency to imagine, and more par-
ticularly to hear-as, in relation to a passage of music one is auditing, remains
the key to its expressiveness, and not whatever power it may possess to pro-
voke in us a corresponding emotion.
Such is the theory of musical expressiveness to which I subscribe, and which
figured in the background of my earlier remarks. What follows, as already
intimated, is that generic spatial imagination is indispensable to the appreci-
ation of traditional instrumental music. The argument to that conclusion, in
summary form, is this: appreciation of traditional instrumental music requires
grasp of musical expression; musical expression of emotion presupposes the
notion of personal expression of emotion and rests on the emergence of
musical gestures; personal expression of emotion, on the one hand, is fun-
damentally behavioral and so necessarily manifested in space, while musical
gesture, on the other hand, is a function of both performing gestures under-
stood as the sources of musical sequences heard and perceived resemblances
between musical sequences and behavioral expressions; hence grasp of per-
forming gestures and of behavioral expressions, and thus ultimately of musical
gestures, requires generic spatial imagination; hence grasp of musical expres-
sion requires generic spatial imagination; hence spatial imagination is neces-
sary to the appreciation of traditional instrumental music.

VI MUSICAL EXPRESSIVENESS AND MENTAL

SIMUL ATION

A question that we can now pose is this: does the imagining of indetermin-
ate personlike agents in music, who seem to be expressing their own emo-
tionswhich I take to be the core of an experience of musical expressive-
nessinvolve us in acts of mental simulation, and if so, of what are these acts
the simulation?
Though I am unsure how to answer this question, I will permit myself a
measure of speculation in connection with it. Faced with a passage of music
that strikes us as behaving or gesturing in such and such fashion in virtue of
its musical movement, its underlying performing actions, and other aspects
of its sonic appearance, perhaps we frame for ourselves, without being fully

For further illumination on this point, see my Musical Expressiveness and Jenefer Robinson,
The Expression and Arousal of Emotion in Music, in Philip Alperson (ed.), Musical Worlds
(University Park, PA: Pennsylvania State University Press, 1998), 1322.
88 Music

conscious of it, the hypothetical query, If someone was comporting himself in

that way, what emotion is it likely he was feeling? However, perhaps it is the
case that faced with such a passage we do not theorize in this manner, even
subconsciously, in the aim of inferring what the musics expressiveness must
be, but that we rather try to imagine ourselves in the place of the music, that
we assume as our own the musical gestures we hear the passage to be suffused
with, as a consequence of which we find ourselves feeling, in imagination,
such and such emotion, and so in that way come to know what the music
expresses. It remains to add only that such postulated acts as putting oneself
in the place of the music and assuming as ones own the musics perceived gestures
might well be realized by some sort of simulation procedure.
More precisely, in order to effect such a simulation one would have to enter
into the presumed mental simulator in off-line mode thoughts or beliefs such
as [I am the music], [I am behaving in the manner ], [I am making the
musical gesture ], the content of the blanks being supplied by acts of
mental ostension or demonstration, with the expectation of finding oneself
consequently in such and such mental state, a state one would then be in a
position to say was the state expressed by the music. I am not prepared to con-
jecture that we in fact do something of this sort, but such a scenario seems
to me at least possible. At any rate, the theory of musical expressiveness here
defended is certainly amenable to interpretation in a simulationist manner, if
not such as to require that.

VII EXPRESSION IN ELECTROACOUSTIC MUSIC

The necessity of a robust sense of hearing, including the ability to localize

sounds in space, to the overall appreciation of electroacoustic musicmusic
written not for instruments, but composed on computer, tape, or synthesizer
for direct transmission to loudspeakerscan be taken for granted, given that
such music uses space as a primary structural material and derives probably its
most striking effects from spatial manipulations. Appreciating such music
draws on not only generic but specific spatial imagination in the most obvious

The process of imaginative projection by which we endeavor to seize the expressiveness

of music is convincingly described in Roger Scruton, The Aesthetics of Music (Oxford: Oxford
University Press, 1997), though in terms of Verstehen rather than of simulation.
I should make clear that I am here thinking of a music of multiple channels, typically six
or eight, capable of generating an ample acoustic space. Another term for electroacoustic music,
especially in France, is acousmatic music, though the latter has a slightly narrower meaning.
 The Expression of Emotion in Music 89

way. What is at issue at present, rather, is whether generic spatial imagination

is necessary to a grasp specifically of the expressiveness of such music.
With instrumental music the need for generic spatial imagination in
gauging expressiveness is a consequence of the need to imagine the
performing gestures behind the sounds, and behind those performing
gestures, the gestures and actions that serve as the behavioral expressions of
emotions. It is because those gestures and actions take place in space that a
capacity to imagine them goes hand in hand with a capacity for at least generic
spatial imagination. How, then, do matters stand with electroacoustic music?
Since such music is not performed, grasp of its expressiveness cannot require
us to imagine performing gestures from which it is presumed to issue.
Still, though the expressiveness of electroacoustic music thus necessarily
differs somewhat in its basis from that of traditional music, an argument can
be made in its regard connecting a grasp of expressiveness to a spatial imagina-
tion of sources.
As we have seen, the capacity to imagine the presumed sources of sounds,
in terms of instruments, agents, and actions, plays a crucial role in estimating
the expressiveness of traditional music. With electroacoustic music, by con-
trast, we generally have only a vague idea as to the sources of the sounds heard,
because the sounds in question are usually unfamiliar and difficult to categor-
ize, and information about their provenance is rarely indicated in ancillary
material. Nevertheless, it seems that our habits of hearing-as are likely to be
transferred, to some degree, from traditional music to electroacoustic music,
without our being conscious of any categorization of sounds according to
their likely sources. For instance, a sound resembling that of a chainsaw but
not consciously classified by us as such would still probably be subconsciously
assimilated by us in those terms, inducing us to take it as emanating from
something like a chainsaw, with all that that implies expressively. Thus, even
with electroacoustic music, where it is not incumbent upon listeners to hear
stretches of sound in terms of their presumed sources, me may conclude that
the capacity to imagine in a spatial manner possible sources of sounds heard
will play a role in the expressiveness such music wears for us.
On the other hand, insofar as the expressiveness of music rests in part on the
sources one imagines sounds to have had, one can also see there an explana-
tion of why the expressiveness of electroacoustic music, which would be hard
to deny, remains yet more elusive than that of traditional instrumental music.
It would be because in this music, as just noted, we find it difficult to clas-
sify the majority of sounds encountered as to their ostensible sources, and
90 Music

probably even our subconscious classifiers of sounds, such as they are, often
rest undecided before a good portion of the sounds that escape conscious cat-
egorization. If so, then by consequence we will also find it difficult to grasp
with any definiteness the musical gestures embodied in a piece of electroacous-
tic music, in contrast with those embodied in a piece of music composed for
and sounded on familiar instruments. However, and there is perhaps appro-
priate compensation in this, to the degree that the expressiveness of electroa-
coustic music remains elusive to us we are in a certain sense liberated, and
so free to deploy in an even more creative manner than usual the auditory
imagination that is ours.
 6
Musical Expressiveness as
Hearability-as-Expression

I E XPRESSION AND EXPRESSIVENESS

As a number of philosophers have rightly underlined, expression is essentially

a matter of something outward giving evidence of something inward. In
other words, expression is essentially the manifesting or externalizing of
mind or psychology. The scope of expression is thus not, pace Goodman
and others, properties in general, or properties metaphorically possessed,
but rather psychological properties, those pertaining to the mental states of
sentient creatures. For only those can be intelligibly expressed, whoever or
whatever is doing the expressing. And this holds as well for expressiveness,
which we can initially understand as the sort of expression that some objects,
and perhaps most notably musical works, manage to achieve, despite their
not literally having inner lives. Exactly what sort of expression expressiveness
amounts to will emerge as the argument proceeds. My discussion focuses on
the case of musical expressiveness, but at a later point I offer some reflections
on artistic expressiveness in general.
II MUSICAL EXPRESSIVENESS

I here defend an account of musical expressiveness elaborated in earlier

papers, replying to certain objections it has elicited, and underlining its

First published in M. Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art
(Oxford: Blackwell, 2006), 192206.
For instance, Alan Tormey, The Concept of Expression (Princeton: Princeton University Press,
1971); Bruce Vermazen, Expression as Expression, Pacific Philosophical Quarterly 67 (1986):
196224; and Aaron Ridley, Expression in Art, in Jerrold Levinson (ed.), Oxford Handbook of
Aesthetics (Oxford: Oxford University Press, 2003), 21127.
Hope in The Hebrides , in Music, Art, and Metaphysics (Ithaca: Cornell University Press,
1990) and Musical Expressiveness in The Pleasures of Aesthetics (Ithaca: Cornell University Press,
1996).
92 Music

superiority to the most plausible competing accounts on offer. Crucial to my

account is the idea that the expressiveness of music resides in the invitation
that music extends to the listener to hear it as expression in the primary
sensethat is, expression, by persons, of inner states through outer signs.
Thus, to hear music as expressive is to hear it as an instance of personal
expression. Expressive music is music we are disposed to hear as expressing;
and that is what its expressiveness fundamentally consists in.
Since the primary vehicle of the expressing of states of mindas opposed
to articulate thoughtsis gesture, broadly understood, and since music is
naturally assimilated to that kind of expressing, that means that we hear
expressive music as gesturing of some sort, that we hear a sort of gesture in
expressive music. Call this musical gesture. It is important to stress that though
musical gesture is related to both the ordinary behavioral gestures connected
to expression of states of mind and the specific performing gestures involved
in the sounding of music, it is not equivalent to either of those. It is a matter,
at base, of what we hear the music to be doing, in virtue, most importantly, of
the movement we hear in music.
Summon up in your mind the opening of Brahmss First Symphony. You
are inescapably presented, in listening, with the image of someone in the
throes of emotion, which emotion is being manifested to you, through what
one might call musical gestures. You may perhaps not know, or be able to
articulate, what emotion the agent heard in the music is in the grip of, but
in the grip of it he is, and this is something you directly hear. It is as if an
emotion is being expressed, in the most literal sense, though it is somehow
happening through music.
The musical expression of states of mindincluding emotions, feelings,
attitudes, desires, and beliefsmust thus be modeled on the primary
expression of such states by persons or human beings, whereby such states are
revealed or evinced through behavior or other outward manifestation. In the
case of emotions, the usual focus of musical expression, such manifestations
include countenance, posture, bearing, demeanor, actions, gestures, and
modifications of voice.
Naturally there are differences between human expression and musical
expression. For one thing, music is not literally behavior, and musical gesture

See Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994)
and Roger Scruton, The Aesthetics of Music (Oxford: Oxford University Press, 1997) on musical
movement, and Scruton, ibid., and Jerrold Levinson Sound, Gesture, Space and the Expression of
Emotion in Music (Ch. 5, this volume) on musical gesture.
 Musical Expressiveness as Hearability-as-Expression 93

is not literal gesture. For another, expression in music is not the result
of emotion experienced by the music, since music is not sentient, nor is
it invariably the upshot of emotion experienced by the composer, since
lived emotion and musically imagined emotion can readily diverge. But
despite those evident differences between musical expression of emotion and
human expression of emotion, we should not consider a piece of music to
be strictly expressive of an emotionrather than standing in some other,
weaker, relation to it, such as possessing a perceptual quality associated with
the emotionunless we regard it as analogous to a being endowed with
sentiments capable of announcing themselves in an external manner. In short,
music expresses an emotion only to the extent that we are disposed to hear
it as the expression of an emotion, although in a non-standard manner, by a
person or personlike entity.
More formally, what I have proposed as an analysis of expression in music
goes something like this: a passage of music P is expressive of an emotion E
if and only if P, in context, is readily heard, by a listener experienced in the
genre in question, as an expression of E. Since expressing requires an express-
er, this means that in so hearing the music the listener is in effect committed
to hearing an agent in the musicwhat we can call the musics personaor
to at least imagining such an agent in a backgrounded manner. But this agent
or persona, it must be stressed, is almost entirely indefinite, a sort of minimal
person, characterized only by the emotion we hear it to be expressing and the
musical gesture through which it does so. It is important to keep that in mind
when entertaining skepticism as to whether understanding listeners normally
hear or imagine personae when they apprehend expressiveness in music.
My basic analysis of musical expressiveness, again, is that music expressive
of E is music heard as, or as if, someone expressing E. Of course one could
fairly expand this as music heard as, or as if, someone experiencing, and as a
result expressing, E, since A expresses E presupposes A experiences E. But
it is presumably the expressing part alone that enters into what the music can
intelligibly be heard as, since we can, it seems, have no idea what it would be
to hear music or musical process, as if it, the music or musical process, were an
experiencing of something.
It has been suggested, incidentally, that not all expressions are such that
we think of their possessors or bearers as engaged in acts of expressing, in
the sense of intentionally trying to communicate a state of mind. Since I

Vermazen, Expression as Expression, 1989.

94 Music

analyze expressiveness in terms of as-if expression, not as-if-expressing, then

even if that suggestion is valid, it is immaterial for my account. The sugges-
tion can, in any case, be challenged. For it can be argued, pace Vermazen,
that all expressions ascribable to an agent in fact are cases of expressings by
the agent. It is just that not all such expressions are cases of intentional or
self-conscious expressing.
Earlier formulations of my proposal appealed to the idea of a sui generis
mode of expression of emotion, suggesting that music, when heard as express-
ing, was such a mode of expression. But that idea may be an unfortunate one,
and has attracted its share of criticism. Some commentators, for instance, have
charged that there is something incoherent in the suggestion that one hears a
passage of music as a sui generis expression of some emotion, on the grounds
that we can form no conception of a mode of expression declared to be sui gen-
eris; others have charged that the appeal to a sui generis mode of expression in
connection with music implies that we experience music as a novel corporeal
means of sound production or else as an odd creature that somehow behaves
musically.
In light of these charges, it is probably a mistake to insist that the expression
that music is heard as when its expressiveness is being perceived is sui generis
expression, rather than just expression simpliciter. Little appears to be gained,
and only confusion sowed, by such insistence. And yet the notion of a sui
generis mode of expression, which the singularity of our experience of music as
expressive suggests to us, may still have a role to play in the full explanation of
musical expressiveness.
What must be avoided, it seems, is making the idea of a sui generis mode of
expression part of the content of the hearing-as experience involved in regis-
tering musics expressiveness. (After all, thinking in Latin can hardly be a
prerequisite for the grasping of expression in music!) The sui generis aspect
of musical expression, such as it is, will have to be reflected in the analysis in a
different manner.

Derek Matravers, Art and Emotion (Oxford: Oxford University Press, 1998), 131.
Kendall Walton, Listening with Imagination: Is Music Representational?, Journal of Aesthetics
and Art Criticism 52 (1994): 4761 (p. 56); Malcolm Budd, Values of Art (London: Penguin, 1995)
132.
Kendall Walton, Projectivism, Empathy, and Musical Tension, Philosophical Topics 26 (1999):
40740 (p. 435).
 Musical Expressiveness as Hearability-as-Expression 95

Compare hearing the sound of a trains wheels as a baby whining (but in an

unusual, because regular and rhythmic, way), and hearing a stretch of music
as an expression of sadness (but of an unusual, because non-corporeal and
extraordinarily fluid, sort). Now, do the parentheticals in those cases give
part of the content of the hearing-as experiences in question? Perhaps yes,
if we view such contents as highly backgrounded ones. But we might with
more justice answer no, viewing those parentheticals not as giving part of the
contents of those experiences, but rather part of the contents of subsequent
reflections on those experiences. Thus perhaps the qualification sui generis that
attaches to the expressing that we hear music as doing when we register its
expressiveness does not enter into the hearing-as experience itself, but only
into characteristic further thoughts about that experience. The modifier, in
sui generis fashion, plausibly belongs to the content of a subsidiary thought on
the expressing we hear in expressive music, not part of the content of the core
experience of hearing the music as, or as if it were, an expression of some sort.
The notion of hearing-as, it will have been noted, has been relied on rather
heavily in the preceding discussion. So what is it to hear a stretch of music as
something elseor alternatively, to hear that something else in that stretch
of music? This remains a difficult matter, but for present purposes it suffices
to locate hearing-as and hearing-in among perceptual acts that partake freely
of, or that substantially enlist, the imagination. The agents one hears in music
when one hears it as an expression of emotion, sui generis or not, are thus ines-
capably imaginary ones, ones displaying the indefiniteness characteristic of all
imaginary objects. To hear music as such and such is, perhaps, to imagine that
the music is such and such, and more specifically, to imagine of the music, as
you are hearing it, that it is such and such.
The worry is sometimes voiced, concerning the appeal to imagination in
the analysis of musical expressiveness, that imagination is too unconstrained
to secure the degree of objectivity that musical expressiveness appears to enjoy.
But this worry can be put to rest by recalling that the appeal is not to what a
passage might perhaps be imagined to be the expression of, but rather, to what
a passage is most readily and spontaneously imagined to be the expression of, in
which case the exercise often has a fairly unequivocal outcome.

On the effective equivalence of hearing-in and hearing-as in regard to music, see my Musical
Expressiveness, in Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996).
96 Music

III RESEMBL ANCE-BASED VIEWS OF MUSICAL

EXPRESSIVENESS

Malcolm Budd, in a trenchant discussion of our topic, distinguishes a minim-

al, or basic, concept of musical expressiveness, and then identifies three accre-
tions to that minimal concept, resulting in three more elaborate conceptions
of musical expressiveness, the third of which is more or less the conception
defended here. The minimal conception goes like this: a stretch of music is
expressive of E if one hears the music as sounding like the way E feels, or per-
ceives a likeness between the music and the experience of E, and it is correct to
do so.
My main objection to Budds minimal conception, which is a resemblance-
based one, is that it is simply too minimal. What it defines isnt yet musical
expressiveness, but at most a rough precondition or typical upshot of music-
al expressiveness. Perceiving a likeness between two things A and Band
especially where, as in the present instance of musical passages and emotion-
al states, it is a matter of cross-categorial perceptionis not sufficient for
hearing A as B. The latter is a distinct occurrence, which neither entails nor
presupposes the former. Resemblance in various respects between the sound
and shape of a passage of music and the inner experience or outer expression
of an emotion is undoubtedly one of the chief grounds of musical expressive-
ness, but neither the resemblance as such, nor the capacity to make listeners
aware of that resemblance, constitutes the expressiveness in question.
I can perceive that or acquire the perceptual belief that a leafy tree resembles
a bushy beard, for instance, and not have the experience of seeing the tree as a
beard. I can notice the likeness between the two and yet not see the one in the
other. But surely we cannot speak of a musical passage being expressive of an
emotion unless listeners are induced to hear the emotion, or more precisely,
an expressing of the emotion, in the passage, whatever degree of resemblance
they might note, in whatever respects, between the passage and the emotion.
Note further that even if a resemblance-based account of musical
expressiveness could deliver the right verdicts in individual casesthat is,
even if the degree of resemblance that an ensemble of musical features
needs to have to some emotion for a passage possessing such an ensemble
to be expressive of the emotion in question could be specified in some

Values of Art, ch. 4.

 Musical Expressiveness as Hearability-as-Expression 97

general mannerthat would not constitute an acceptable analysis of musical

expressiveness. For it would not elucidate what musical expressiveness was,
but only what ensembles of musical features were coextensive with and
underlay such expressiveness.
Stephen Davies, the most prominent defender of a resemblance-based view
of musical expressiveness, maintains that emotion words used to describe
appearances, whether in persons, natural objects, or works of art, are parasitic
on the use of such words to refer to felt emotions; they thus represent a
secondary, though literal, use of such words. In this secondary use, says
Davies, emotion words describe emotion-characteristics-in-appearance, and
it is in these that musical expressiveness lies. According to Davies, the
expressiveness of music consists in its presenting emotion characteristics
in its appearance. . . . These expressive appearances . . . are not occurrent
emotions at all. They are emergent properties of the things to which they
are attributed. Davies goes on to explain that such musical expressiveness
depends mainly on a resemblance we perceive between the dynamic character
of music and human movement, gait, bearing, or carriage. In conclusion,
Davies affirms that emotions are heard in music as belonging to it, just as
appearances of emotion are present in the bearing, gait, or deportment of our
fellow humans and other creatures.
Daviess view might, I think, be stated as follows: P is expressive of E iff P
exhibits an emotion-characteristic-in-sound associated with E, that is, exhib-
its a sound-appearance analogous to the human emotion-characteristic-in-
appearance of E.
Though I am not unsympathetic to the basic thrust of this account of
musical expressiveness, I have major qualms about the central notion in
terms of which it is framed, namely that of musical emotion-characteristics-
in-appearance. The problem is that the appearance of a passage of music is
not precisely that of a person, or a persons face or body in any condition,
or a persons behavior at any moment. It is instead a matter, when a passage
displays an emotion-characteristic in its medium of sound sequences, of an
appearance similar to that presented by a person in some state. But since
everything is similar to everything else to some degree, the issue then becomes
one of how similar such an appearance must be to one presented by human
behavior in order to constitute an emotion-characteristic-in-sound of the

See his Musical Meaning and Expression. Ibid. 228. Ibid. 229.
Ibid. 239.
98 Music

emotion in question, or else, as Davies sometimes puts it, of how similar the
experience of musical movement and of expressive behavior must be, in order
for the appearance generated by such movement to constitute an emotion-
characteristic-in-sound of the emotion in question.
I think it is plain that there is no answer to this question except by appeal
to our disposition to hear that emotionrather than another, or none at
allin the music, that is, by appeal to our disposition to aurally construe
the music as an instance of personal expression, perceiving the human appear-
ances in the musical ones, in effect animating the sounds in a certain manner,
to use a phrase given currency by Peter Kivy. Only if this occurs does the
music have the expressiveness in question, regardless of the degree of similar-
ity between the musics appearances and the human appearances by relation
to which it ends up being expressive, or alternatively, the degree of similarity
between the experiences of those appearances.
There is simply no independent conception of and no access to what Davies
calls musical emotion-characteristics-in-appearance apart from satisfaction
of the hearability-in-the-music-of-an-expressing-of-emotion condition vis-a-
vis attuned listeners; the latter is what gives content, ultimately, to the
former, however familiar the appearances in question might be. What
the musical emotion-characteristic-in-appearance of sadness is, in general,
cannot be derived from the behavioral emotion-characteristic-in-appearance
of sadness in persons. There is no translation rule from behavioral
appearance-characteristics to musical appearance-characteristics; only the act
of perceiving in music the outlines of the former gives rise, so to speak, to the
latter.
Though Davies does not want to be committed to the view that musical
expressiveness consists in analogy or resemblance to literally expressive
behavior, his invocation of emotion-characteristics-in-sound as something
founded in and emerging out of such analogy or resemblance in any event
suggests that such characteristics are specifiable independently of experiences
of hearing emotional expression in music, like the behavioral emotion-
characteristics-in-appearance associated directly with felt emotion. But that
is to overlook the real differences between human emotion-characteristics-in-
appearance and the supposed musical emotion-characteristics-in-appearance
being appealed to; the former can to some extent be catalogued independently

See Kivy, Sound Sentiment (Philadelphia: Temple University Press, 1989).

 Musical Expressiveness as Hearability-as-Expression 99

of individual judgments of expressive import on the part of perceivers,

whereas the latter cannot. Let me elaborate on this.
We can give content to sad human appearance by glossing it as the
appearance or kind of appearance sad humans typically display. But we cant
analogously give content to sad musical appearance. There is no such thing as
the appearance or kind of appearance that sad music typically displays. There
is no extractable profile of sad musical appearance, as there is of sad human
face or sad human posture; sad musical appearance, unlike sad human
face and sad human posture, is not, as it were, paraphrasable. The only way
to anchor sad musical appearance, I submit, is in terms of our disposition
to hear such music as sad. The analysis of musics expressiveness must thus
foreground that perceptual-imaginative experience, and not the resemblances
that, no doubt, underlie it. Musical emotion-characteristics-in-appearance,
if they are supposed to be something identifiable apart from experiences of
hearing such and such emotion or expression of emotion in music, and as
parallel to human emotion-characteristics-in-appearance, are an illusion.
Take sadness. Sadness is an emotion, that is, a mental condition with
various cognitive, conative, affective, evaluative, behavioral, and possibly
physiological aspects. Next we have sad look, which is more or less the
lookin face and bodythat sad people typically wear, or wear when they
are not trying to conceal or suppress their sadness. And then we have sad
sound, which is more or less the soundvocal, for the most partthat
sad people typically make, at least when not trying to conceal or suppress
their sadness. So characterized, sad look and sad sound are human emotion-
characteristics-in-appearance, as is their conjunction, which we might label
sad appearance.
But now we come to the alleged corresponding musical emotion-
characteristic-in appearance, or sad musical sound. How is that to be cashed
out, in light of the characterizations of sad look and sad sound just given? One
possibility would be, as the musical sound that sad people typically make.
But that cant be right, since sad people dont typically make musical sounds
of any sort, and the sounds they do makeweeping, sighingare clearly
distinguishable from music of an ordinary sort. A second possibility would be,
as the sound that sad pieces of music have in common. But that is an even worse
proposal, since it presupposes the prior identification of pieces of music as sad.
A third possibility would be to appeal to purely technical or structural features
of music, such as those of melody, harmony, tempo and texture. But that
would be of no use, since even if there is a complex disjunction of technical
100 Music

or structural features coextensive with sadness in music, such a disjunction

would not serve to explicate the concept of sad musical sound, nor would such
a disjunction appear to play a role in our identifying passages as sad.
The only proposal with any chance of success, then, must be that sad
musical sound is sound resembling sad sound (the standard aural appearance
of sadness) or, cross-modally, sad look (the standard visual appearance of
sadness), or both. But to what degree? Theres the rub, for as we all
know, everything resembles everything else, yet the degree of resemblance
to an emotion required to make a musical appearance a musical emotion
characteristic in appearance of that emotion cannot be specified in terms of
some fixable degree of resemblance between the two. It can only be specified,
it seems, as whatever resemblance is sufficient to induce appropriately
backgrounded listeners to hear the music as sad, or as expressing sadness. But
that, surely, is to give up the idea that there is a recognizable musical emotion-
characteristic-in-appearance of sadness, somehow analogous to the human
emotion-characteristic-in-appearance of sadness, on which the analysis of
musical expressiveness can rest.
In sum, if there really are musical emotion-characteristics-in-appearance,
to which the explication of musical expressiveness must advert, we should be
able to identify them other than simply as being appearances in which the cor-
responding emotion can be heard. Thus, for sad musical sound, there should
be some possible specification or profile, however schematic, of what sort of
sound that is, other than sound that invites hearing-as-sad. But there isnt.

IV INFERENCE-BASED VIEWS OF MUSICAL

EXPRESSIVENESS

Bruce Vermazen, Jenefer Robinson, and Robert Stecker are philosophers who
subscribe to inference-based views of musical expressiveness. In the view of
each of them, the expressiveness a passage of music possesses is something like
the conclusion of an inference to the best explanation. In Vermazens case it
is an ascription of a state of mind to the imagined utterer of the passage that
best explains its distinctive features; in Robinsons case it is an ascription of a

See Vermazen, Expression as Expression; Jenefer Robinson, The Expression and Arousal of
Emotion in Music, in Philip Alperson (ed.), Musical Worlds (University Park, PA: Pennsylvania
State University Press, 1998), 1322; and Robert Stecker, Expressiveness and Expression in Music
and Poetry, Journal of Aesthetics and Art Criticism 59 (2001): 8596.
 Musical Expressiveness as Hearability-as-Expression 101

state of mind to the imaginary protagonist of the passage that figures in the
best interpretation of the musical work taken as a whole; in Steckers case it is
a best hypothesis on the part of an ideal listener as to what state of mind the
composer of the music intended such a listener to hear in the passage.
Though I share with Vermazen and Robinson the commitment to personae
in the analysis of musical expressiveness, and with Stecker the notion that
certain aspects of artistic meaning are amenable to analysis in hypotheticalist
terms, I disagree with all three that basic musical expressivenessthat is,
the expressiveness of individual passagescan be constitutively tied to the
sustaining of inferences about the music. That is due to my belief that basic
musical expressivenessthough perhaps not all sorts of expressiveness, such
as that more typical of literature, involving articulate states of mind, nor that
perhaps attaching to works of music as wholesis something directly heard,
not inferred, by attuned or properly backgrounded listeners. Otherwise put,
inferentialist views of expressiveness fail to capture the immediacy with which
we register basic musical expressiveness.
I thus continue to think, pace Stecker, that immediacy is a proper
desideratum for an account of musical expressiveness, and that my
ready-hearability-as-expression account acknowledges that better than the
hypotheticalist account that Stecker proposes instead. It is true, as Stecker
points out, that hearing emotion in music and judging the expressiveness of
music are not the same things. But there is a third thing, perceiving the musics
expressiveness, and that is not the same as judging its expressiveness to be such
and such. The expressiveness of music, I claim, and not just the emotion
in music, is standardly something directly registered, not just conjectured
about. My account can accommodate this fact, as can also a resemblance-
based account, but an inferentialist account cannot.
Now Stecker, seeking to hoist me with my own petard, argues that my
ready-hearability-as-expression proposal is ultimately as inferentialist as his
own, since in order to determine that a work was expressive of E on the basis
of his hearing an expression of E in it, even a properly backgrounded listen-
er would have to engage in inference, taking the fact of his so responding
as a premise together with the premise that he was in fact a properly back-
grounded listener listening properly. But I anticipated and replied to that
line of criticism in my previous essay on this topic. I there emphasized that

Stecker, Expressiveness and Expression in Music and Poetry, 92.

Musical Expressiveness, 11819.
102 Music

a qualified listener who hears the expression of some emotion in music nor-
mally acquires, without reflection, the conviction that the music is expressive
of that emotion, that is, would be readily hearable as such by other qualified
listeners. And that, I suggest, counts as perceiving the musics expressiveness,
or as much as anything could. Whether the listener in addition thereby knows
that the music is thus so expressive is another matter, one that may indeed
involve further reflection or investigation.
Here is a variant of that reply. Qualified listeners arguably tacitly assume,
while listening to music, that they are qualified listeners, and are listening
appropriately. Thus, on the view I favor, hearing an expression of E in a
stretch of music becomes, for such listeners, tantamount to hearing the musics
expressiveness of E. All qualified listeners need do to hear the expressiveness
of the music is to readily hear expression of emotion in it. Since they are, and
unreflectingly assume they are, qualified listeners, listening appropriately, their
readily hearing such and such an expression in the music directly manifests its
being readily so hearable by such listeners! Note, however, that this line would
not work to secure the immediacy of musical expressiveness on a hypothetical-
intentionalist view of it, even were we to grant a parallel tacit assumption on
the listeners part, to the effect that he was an ideal listener. For arriving at a
best hypothesis of what the historically rooted composer intended for one to
hear in a given passage remains an ineliminably inferential affair.
There is still another reason to resist a hypothetical-intentionalist account
of musical expressiveness. Expressive content in music and expressive content
in poetry certainly seem to be quite different sorts of things, hence it would
not be surprising if they lent themselves to different sorts of analyses. The lat-
ter is largely propositional, and so reasonably assimilated to the kind of mean-
ingbasic literary meaningthat a view like hypothetical intentionalism
is designed to account for, whereas the former is largely non-propositional,
hence not reasonably assimilable to basic literary meaning.
I am inclined to think that a perceivability-as-if-expression account is apt
not only for music, but for non-representational art generally. However,
where we have to do with representational art, and perhaps especially, literary
artincluding narrative painting, epic poetry, theatre, cinema, and the
novelemphasis on the immediacy of expressiveness seems less apt, and the
merit of a more inferentialist account of expressiveness seems correspondingly
greater. The reason may simply be that immediacy of expression is an

See chs. 17 and 18 of the present volume.

 Musical Expressiveness as Hearability-as-Expression 103

appropriate demand for affective states or attitudes, the core expressive

content of non-literary arts such as music or abstract painting, but not for
propositional beliefs or thoughts, the core expressive content of literary arts
such as cinema or the novel.
We can now offer an observation about expressiveness in art applicable
across the arts, whether representational or non-representational, literary
or non-literary. Remember that expression in general can be characterized
roughly as the evidencing of a state of mind through some sort of external
manifestation. In the case of art, the external manifestation is not behavior,
but rather the artwork itself, in all its perceivable particularity. But just
as we grasp literally expressed states in and through the behavior, verbal
and otherwise, that expresses them, so with expressive art we grasp the as-
if expressed states in and through the concrete vehicle of the work, its
metaphorical body. With expressive art, expressiveness is grasped through
perceiving the work in its specific detailwhether of word, paint, sound, or
stoneand the states expressed are ones that perceivers are thus aided to
enter into in imagination precisely in virtue of perceiving the work through
which, either immediately or inferentially, they grasp what states those are.

V SOME OBJECTIONS TO THE

HEARABILIT Y-AS-EXPRESSION VIEW OF MUSICAL
EXPRESSIVENESS

The first and most common objection is this. A number of writers charge
that competent listeners do not in factor at least not all of them, all of
the timehear or imagine personae in music whose expressiveness they are
registering, and thus that such imaginative hearing cannot be constitutive
of hearing music as expressive and the disposition to induce such hearing as
constitutive of the musics being so expressive.
Well, that may be how it sometimes seems, or seems on the surface, but
if expressive music is, as I maintain, music readily heard as, or as if, expres-
sion, and if, in addition, expression requires an expresser, then personae or
agents, however minimal, just are presupposed in the standard experience of
such music. But, of course, one may not always notice or acknowledge what

Stephen Davies, Contra the Hypothetical Persona in Music, in Mette Hjort and Sue
Laver (eds.), Emotion and the Arts (Oxford: Oxford University Press, 1997), 95109; Walton,
Projectivism, Empathy, and Musical Tension; and Stecker, Expressiveness and Expression in
Music and Poetry.
104 Music

is presupposed in ones imaginative hearing of music. The claim is not that

listeners are always explicitly aware that personae are involved in their hearing
music as expressive. For people are often not entirely aware of what is implic-
ated in a perception or experience they are having.
A point I have made before is worth recalling here. It may indeed be true
that listeners who recognize the expressiveness of a passage of music do not
invariably hear it, then and there, as the expression of an emotion. But all
that the theory requires is that they recognize the music as readily hearable
as the expression of an emotion, even if, for one reason or another, they do
not themselves give in to that inducement on a given occasion. Moreover,
quite possibly all such cases are ones in which listeners are recognizing pas-
sages they have previously heard as the expression of given emotions, or else
as ones highly similar to such passages, thus presupposing occasions on which
the expressiveness in question was in fact grasped through an experience of
hearing-as-the-expression-of.
Furthermore, some of the discrepancy with the sincere avowals of listeners
on this subject, I suggest, is that a passage of music may more loosely have an
emotional quality, in virtue of suggesting an emotion through its appearance,
without being strictly speaking emotionally expressive, understood as being
such as to induce hearing-as-expression of that emotion. The finale of
Beethovens Fifth Symphony, to take a stock example, is expressive of
something like triumphant joy, and I think it is hard not to hear that finale
as if there is someone, or some agent, there who is expressing his, her, or its
triumphant joy in those familiar musical gestures, the character of which is
rendered especially vivid in virtue of the movement in which they occur being
the successor and culmination of the three that precede it. By contrast, the
opening Prelude of Bachs Well-Tempered Clavier has an emotional quality
one might describe as contentment or equanimity, and yet perhaps one is not
induced to hear it as, or as if, the expression of that state of mind. But then it is
probably also right to deny that it is expressive of contentment or equanimity,
in addition to just possessing the corresponding emotional quality.
The second objection is this. It has been suggested that the appeal to
apt hearability-as-expression as a benchmark of real, as opposed to merely
apparent, expressiveness, a feature of my original formula, will either not do
the work that is required of it or else is called upon to do too much work.

Stecker, Expressiveness and Expression in Music and Poetry, 914.

 Musical Expressiveness as Hearability-as-Expression 105

I am at this point inclined to agree. The qualification of apt in my

original formula was an evasion. What I now think is that the burden of
securing the objectivity of expressiveness, or equivalently, the normativity of
judgments of expressiveness, must simply rest on the properly backgrounded
listener and his or her hearing of a passage in its proper intrawork and
extrawork context. Where there is ready hearability-as-expression under such
conditionsevidenced most clearly by convergence in experiences among
such listenersthen objectivity and normativity are present. Where not,
then not. Exactly how large is the domain of objective expressiveness in music,
which depends on such convergence, thus remains an open question.
The third objection is centered on a related worry about my original for-
mula voiced by Roger Scruton, to the effect that appeal to a reference class of
qualified listeners whose ready hearing of a passage as expression of E serves
as the mark of the passages being truly expressive of E is doomed to vacuity,
because the reference class can only be characterized as the class of listeners
who in fact perceive the passages expressiveness.
But that is not so. It is like suggesting that the only way to characterize the
class of appropriate, objectivity-anchoring perceivers for the color of a given
patch of greenish paint is as perceivers who correctly perceive that the patch in
question is greenish. The reference class of listeners anchoring the objectivity
of expressiveness in a given musical genrewhat I mean by properly back-
grounded listenersis roughly that of listeners demonstrably competent at
understanding such music, such competence being manifested through vari-
ous recognitional, continuational, and descriptive abilities, and whose other
judgments of expressiveness are in line with established ones in uncontrover-
sial cases. There is perhaps a certain amount of bootstrapping involved in this
picture of the qualified listener for a given musical genre vis-a-vis the express-
ive and other qualities of works in that genre of whose objectivity the qualified
listener is to serve as a benchmark, but there is nothing, I think, fatally circular
in it. At any rate, it seems the sort of difficulty that affects all attempts to ana-
lyze perceptual properties in terms of appearances or dispositions to appear
relative to a class of perceivers of a certain sort.
Paul Boghossian has recently seconded Scrutons worry in a more general
form. Boghossian charges that the appeal to qualified listeners in the analysis

Scruton, The Aesthetics of Music, 353.

Boghossian, Musical Experience and Musical Meaning, in Kathleen Stock (ed.), Philosophers
on Music (Oxford: Oxford University Press, 2006).
106 Music

is vacuous because there is no way to characterize what makes for a qualified

listener without presupposing an understanding of musical expressiveness
that does not invoke such listeners or the ways such listeners are disposed to
hear or otherwise respond to music.
Again, qualified listeners for a given piece of music are naturally not to be
characterized, unhelpfully, as ones who correctly hear the expressiveness of
the given piece. How, then, are they to be characterized? As suggested earlier,
perhaps as musically competent listeners whose judgments of expressiveness
in the tradition in question accord with accepted ones in paradigm cases. But
what are paradigm cases of musical expressiveness? Well, one conception of
them would be as pieces in a given tradition on whose expressiveness almost all
at least minimally musically competent listeners agree.
Boghossian retorts that even on this suggestion, paradigm cases
having the expressiveness they do remains unanalyzed, thus presupposing
some ultimately non-experiential, non-response-dependent notion of such
expressiveness. But I am not so sure. For the expressiveness of paradigm cases,
we may suggest, comes to precisely the same thing as it does for non-paradigm
cases, namely, the musics being most readily hearable as the expression of
such and such emotion by qualified listeners, the only difference being that,
since these are paradigm cases of expressiveness, they will be so heard by
virtually all listeners who are at least minimally musically competent.
The fourth objection is another difficulty Scruton has articulated for my
proposal that is worth addressing here. It is a particular elaboration of the
skepticism acknowledged earlier as to whether we can form an idea of the
singular way of expressing emotions that expressive music on my proposal is
made out to be, at least in imagination, and whether or not in so doing we
characterize it to ourselves as sui generis. I quote the objection in full:
When we hear expression in music, Levinson suggests, this is like hearing another
person express his feelings. But in what way like? We have no prior conception of
what it would be to express feelings in music: if we can think of someone doing this,
it is because we have an idea of the expressive character of music, and therefore can
imagine someone choosing just this piece of music, to convey just this state of mind.
Our ability to imagine a subject expressing his feelings in just this way is predicated
upon our ability to recognize the expressive content of music. Only if we can inde-
pendently recognize the emotional content of music, therefore, can we embark on
the thought-experiment required by Levinsons definition.

The Aesthetics of Music, 352.

 Musical Expressiveness as Hearability-as-Expression 107

Scruton claims that my account is committed to listeners being able to con-

ceive what it would be to express an emotion, in the literal sense, in or through
music. He then suggests that their only way of doing so would be by ima-
gining someone choosing suitable music to convey the emotion in question,
which obviously presupposes an antecedent grasp of the musics expressive-
ness, thus rendering the putative account of such expressiveness otiose.
But both points in this objection are misplaced. First, my account does not
imply that listeners who register expressiveness in music possess a concrete
conception of what literally expressing emotions through music instead of
behavior would amount to, or how such expressing would work. The account
requires only that listeners are able to imagine music to be such a literal
expression. Second, listeners who do that are not constrained to think of
the musics persona as somehow choosing from available musical items ones
suitable to convey his changing moods, like a sound editor selecting tracks to
go with the successive scenes of a film.
Scruton has simply misunderstood the nature of the thought-experiment
that, if my account of musical expressiveness is correct, one is effectively called
upon to perform in order to grasp the expressiveness of music. It is not ima-
gine someone choosing from among preexisting music to convey a given state
of mind; now, what state of mind would that be if he chose the music you
are hearing? It is, rather, imagine the music you are hearing to be the liter-
al expressing of a state of mind; now, what state of mind does that appear to
be?, or perhaps, imagine that the musical gesture you hear in the music was
your own; now, what state of mind do you appear to be in? That thought-
experiment, which my proposal is committed to, grounds identification of the
musics expressiveness, but does not, like the one Scruton saddles me with,
presuppose such identification. If I am right, one grasps what a musical pas-
sage expresses precisely in virtue of imagining, or being disposed to imagine,
a mental state it is as if the literal expression of. That there is no algorithm or
procedure for this thought-experiment, unlike the one Scruton would substi-
tute for it, does not entail that it cannot be carried out. We carry it out, in
fact, every time we attend to some musics expressive dimension and attempt
to articulate for ourselves what it is.

VI PE RSONAE IN MUSIC

Many writers, including those sympathetic to imagined expressions and

their personae, have voiced concerns regarding the indeterminacy of
108 Music

musical personae and the indefiniteness of the principles governing their

postulation. When is the persona of one passage the same as the persona
of another passage? When is there continuity and when discontinuity of
persona, as a work progresses from beginning to end? Might there be
multiple personae present in a single passage? Might personae heard in
different passages be related to one another through recognition, solidarity,
or opposition?
I cannot address these questions here, which go well beyond the scope of
this essay. What I wish to underline, in closing, is just that the sustainability
of the thesis of a minimal persona we are induced to hear in expressive music,
and typically do hear in it when listening attentively, is not affected by worries
of this sort, even if they are ultimately unresolvable. The persona implic-
ated in the ready-hearability-as-expression account of musical expressiveness
is merely the agent of the expression we hear in expressive music, or the owner
of the musical gesture that is the vehicle of that expression. Whether that per-
sona persists as the music proceeds, whether a given persona is accompanied
by others, whether personae enter into relation with one another, and so on,
are matters on which the account of basic musical expressiveness here defen-
ded can remain agnostic.
See Walton, Listening with Imagination; Davies, Contra the Hypothetical Persona in
Music; Fred Maus, Music as Drama, Music Theory Spectrum 10 (1998): 5673; and Gregory
Karl and Jenefer Robinson, Shostakovichs Tenth Symphony and the Musical Expression of Cog-
nitively Complex Emotions, in J. Robinson (ed.), Music and Meaning (Ithaca: Cornell University
Press, 1997), 15478.
 7
Nonexistent Artforms and the Case
of Visual Music

There is a difference between the realm of art as a whole, however we charac-

terize that, and the class of all artforms that exist, or have existed, at a given
point in time. That difference, seen from one angle, is the class of nonexistent
but possible artforms. This chapter will be devoted at the outset to exploring
this latter class. My tools, however, will be merely those of a philosopher, not
a soothsayer.
The problem concerning nonexistent artforms of perhaps greatest interest
is why certain theoretically possible ones have not been essayed, or else have
not proven successful even if essayed to some extent. Examples of nonexistent
artforms, by at least the second of these tests, are kinetic painting, word-
less song, danced poetry, olfactory art, and visual music. One reason this
problem is interesting is that insofar as art as a whole has often evolved expli-
citly through the emergence of new artforms, the problem bears directly on
the future of art; if there are fewer untried artforms out there, or if those
which remain unpracticed even though tried are indeed inherently problem-
atic, then the future of art is, in this respect, somewhat more dim than it
would otherwise be.
This essay has two parts. In the first part I propose some ways of thinking
about the field of nonexistent arts as wholein effect making some advance
inroads into itand in the second part I try to cast light on why certain art-
forms that seem eminently possible in fact fail to exist, through a case study of
one such would-be art, that of visual music, i.e. a structured organization of
colored presentations in time, such as might be provided through the medium
of color film. Why doesnt visual music, despite experiments in that direction,

First published in A. Haapala, J. Levinson, and V. Rantala (eds.), The End of Art and Beyond: Essays
after Danto (Atlantic Highlands, NJ: Humanities Press, 1997), 12239.
110 Music

exist as an entrenched artistic activity, as a medium with promise and a loyal

audience? One may be forgiven for suspecting, in a pessimistic vein, that the
potential for a significant artistic visual analog to music is just not there. I will
investigate, first, why this should be so, and second, how such an art might
best proceed anywayagainst the odds, as it were.

I T HE FIELD OF NONEXISTENT ARTS

A central question is how the domain of nonexistent but possible artforms

is to be circumscribed or delineated. How can we get a handle on, or sketch
some of the features of, the domain we are interested in? One way to construe
this question is as a request for acceptable principles of generating potential
artforms. Taken that way, it seems there are two approaches we could adopt,
which we might characterize, respectively, as the combinatorial and the extra-
polational.
The first, combinatorial, approach would be to set up an abstract matrix
of artistic properties (materials, modalities)e.g. spatial, temporal, repres-
entational, sonic, monochromatic, figurative, narrative, one-stage, stone-
using, wood-using, human-body-employing, etc.in which every existing
and also nonexisting artform would be represented as a particular conjunction
of such properties (or their negations). Thus, classical Greek sculpture would
be approximated by the conjunction: spatial/static/figurative/stone-using/
manually-worked, while a presently nonexistent artform would be represen-
ted by the conjunction: stone-using/temporal/dramatic/multi-colored.
The other, extrapolational, approach would be to go beyond existing art-
forms by the use of various intelligible rules of projection, such as juxtapos-
ition, reduction, fusion, opposition, or mutual accommodation. Thus, we
might have an artform which involved the simple addition of dance to paint-
ing, or which resulted from the fusion of poetry, music, and calligraphy.
On either approach, of course, the issue will remain of whether what is
formally generated by combination of analytic elements or extrapolation from
actual artforms is conceptually coherent, practically feasible, and artistically
promising. That issue will be engaged, at least in one instance, in my consider-
ation of visual music in the second part of this essay.
One problem with the combinatorial approach to identifying nonexistent
potential artforms is that we have not said where the primitive termsthe
properties whose combination is to represent any given artformare to come
 Nonexistent Artforms and the Case of Visual Music 111

from. What justifies the particular set of elements out of which the matrix
would be constructed? Why are stone-using, temporal, and narrative reason-
ably included in this set but not, say, water-soluble, high-risk, expensive, or
innate? The answer must be that we only arrive at a plausible set of artform-
constituting properties by working from the artforms we are acquainted
with, and by relying on our background understanding of the nature of art
in general, its typical objectives and achievements, and its overall place in
human affairs. That is to say, we start with familiar, understood arts, then
decompose them into their salient features, ones that both differentiate them
from other arts and are important in criticism and appreciation. Having ar-
rived at such elements, we can then mentally experiment with recompos-
ing them in different waysgrouping them in different packagesand
reflecting on the outcome. It must be observed, though, that we will not,
by this method, really have constructed all possible arts from the ground
upthat is, from a purely abstract field of possibilities. For the terms of
our matrix will inevitably come from the arts we know and love, as they are,
or have been, constituted. The combinatorial method, despite initial appear-
ances, could not be deployed ahistorically, at least not with any hope of
interesting results.
There is something unappealingly inefficient, as well, in the combinatorial
method of generating possible artforms as I have sketched it. For if we are
going to have to think in terms of existing artforms anyway, why not skip the
step of decomposition into elements and just consider combinations of, or
confrontations between, existing artforms themselves?
The combinatorial method seems inefficient, further, in that given a cer-
tain constellation of defining features, for a given position on the matrix,
the actual shape of the artform or artforms so conforming to that position
or set of properties will usually be excessively indeterminate, hard to form an
image of. Third, the pure combinatorial approach seems quite likely to yield
a numberand perhaps a majorityof combinations which are either con-
ceptually problematic or practically infeasible.
All told, then, I suspect we are perhaps better off pursuing the extrapol-
ational methodcombining or modifying existing artforms directlyas a
way of thinking our way toward the field of nonexistent artforms. Such a
procedure has also the advantage of paralleling, to a much greater degree, the
For example, we know that the arts are activities, that they typically aim at pleasure or
satisfaction of some sort, that they usually involve skills, that they are modes of meaning and
articulating the world, and so on.
112 Music

evolutionary and synthetic process by which new artforms (e.g. film, kinetic
sculpture, video art) have actually been generated in the past.
Advances in art on the level of artforms have regularly come about through
the impact of disparate arts on one another, or through the modification of
existing artforms in the direction of others, but not through positing artforms
de novo to correspond to abstract congeries of artistic properties not previously
bundled together.
In an earlier essay, Hybrid Artforms, I explored the issue of existing art-
forms that have arisen through the combination of two (or more) preceding
oneslabelling these hybrid artformsand suggested a threefold categor-
ization of such hybrids, into juxtapositional (additive), synthetic (fusional),
and transformational ones. The distinction between them is roughly as fol-
lows. In a juxtapositional hybrid the arts are combined in a largely discrete,
non-interpenetrating manner, so that the contributions of each artform enter-
ing into the result are independently identifiable. In an object of the hybrid
art, objects of the contributing arts are still discernible as simply parts, spatial
or temporal, of the hybrid art object. In a synthetic hybrid, by contrast, the
contributing artforms are more or less fused in the result, so that an object
of the hybrid fails to belong clearly to either of the contributing artforms, in
their standard conceptions, and fails also to exhibit parts which can be simply
assigned to one artform or the other. Finally, a transformational hybrid is
intermediate between the previous two cases, and occurs when there is some
interpenetration, rather than mere addition, of two artforms, as evidenced
in the nature of the resulting art object, but where one artform clearly holds
sway in the end and can still claim the product as one of its own, if singularly
modified in the direction of the other art. Examples offered of these three
categories were the following: (a) symphony with light show, dance with calli-
graphic drawings [juxtapositional]; (b) music drama, concrete poetry, shaped
canvas [synthetic]; (c) kinetic sculpture [transformational].
Visual music, in the form of abstract color film, could probably be con-
sidered a transformational caseas color film pushed in the direction of pure
music, adopting both its means and the sort of experience and expressive-
ness characteristically aimed at. This would be transformational, rather than
See Jerrold Levinson, Hybrid Artforms, in Music, Art, and Metaphysics (Ithaca: Cornell
University Press, 1990). This essay was originally published in 1984.
I also count as transformational hybrids cases of new artforms in which paradigm objects of the
art are altered relative to those of , the parent art, in a manner producing tension with the implicit
structural norms of , even if not in the direction of some identifiable second art, . (These are in
effect cases of self-induced hybridization, or differentiation.)
 Nonexistent Artforms and the Case of Visual Music 113

synthetic, because the result would remain unequivocally a film, though one
whose obvious self-restriction (e.g. non-photographicness, non-representa-
tionality) would appear bizarre except for the intended parallel with music.
More obviously, such visual music would be transformational, rather than
juxtapositional, because music there serves as a model or inspiration, and not
as a literal component in the result.
I start, then, with these three categories of hybrid, in terms of the sorts
of extrapolation they suggest for artforms which do not yet exist, and then
proceed beyond them to consider further, or finer-grained, ways of project-
ing from existing artforms. So, first, we can imagine all those artforms that
would involve merely juxtaposing objects or activities of preexisting arts that
have not till now been placed in conjunction. Thus, we might have musidraw-
ing, the art of drawings exhibited for prearranged lengths of time simultan-
eously with selected or newly composed stretches of music, or pulpture, the art
of pairs of paintings and sculptures made to be presented and experienced
together, in definite spatial relation to one another. An important variable
even for artforms such as these, arrived at merely through unprecedented jux-
tapositions, would be the degree of interrelatedness envisaged between the
components brought into conjunction. This could range anywhere from self-
contained, oblivious standing-alongside-of, to delicately nuanced mutual
sensitivity and adjustment, of the components vis-a-vis one another.
The synthetic paradigm of hybridization, in which the identities of con-
tributing arts are partly dissolved, is a more complicated one, and we can
derive from it at least two distinct rules or procedures for generating not-
yet-existent arts. One is this. Given two artforms, and , posit an artform
whose objects are structurally or formally halfway between those of and those
of . The genres of shaped canvases (e.g. those of Frank Stella) and canvases-
with-attachments (e.g. those of Jasper Johns), are ones that have come about
through that sort of cross-fertilization. Both cases mediate between painting
and sculpture, in that objects of these genres are neither as flat and rectangu-
lar as the paradigms of the former, nor as three-dimensionally rounded and
conceived as paradigms of the latter. But consider now the art of danceotry,
intermediate between dance and poetry. A typical work of danceotry would
involve a performer who goes through a series of movements during her recit-
ation, but ones of more restricted scope and ambit than are generally found

Sound film in the familiar sense, though, could be considered as roughly a juxtapositional
hybrid of silent film, music, and dialogue.
114 Music

in dance, and whose recitation, in turn, would be less elaborate and self-
contained, more sparse, than that normal for poetry offered on its own. Or
consider archisculpture, whose objects are house-like constructions which both
stand erect and enclose an interior space, but which otherwise flout many
of the exigencies of the architecture of dwellingse.g. intact roofs, usable
entrances, sound construction materialsin favor of the formal freedom of
sculpture. Or, third, consider sonopoetry, which involves the production of
measured sequences not of words but only non-verbal vocal sounds, e.g. hic-
cups, snorts, gasps, whistles, whispers. Danceotry, archisculpture, and sono-
poetry would all be applications of the rule of projecting an artform that is
structurally intermediate between two given ones.
A close cousin of this rule, though still distinct from it, would be one
which projected artforms which in some way joined or compounded salient
structural features of the objects of two arts, rather than, so to speak, splitting
the difference between them. Of course in some cases mediationthe pro-
cedure involved in the previous ruleis the only manner of combination
that makes sense; for example, if it is a question of a two-dimensional art
(painting) and a three-dimensional one (sculpture), a synthetic hybrid can
be more readily conceived whose objects are somewhere between two- and
three-dimensional, loosely speakinge.g. Johns paintings with protuber-
ancesthan one whose objects would be clearly both two-dimensional and
three-dimensional. But in other cases, combination of arts could conceiv-
ably involve amalgamation, or compounding, of their distinctive traits or
potentials. Consider aquarelloetry, a given specimen of which displays both
the ordered sequence of carefully chosen words definitive of poetry and,
superimposed upon it, an abstract pattern of water colors designed to com-
plementto enhance and be enhanced bythe sounds and sense of the
words on the page. Or novellasong, which would involve short stories com-
posed explicitly to be chanted, or intoned, by an unaccompanied performer;
one can imagine this being done in such a way that the result seemed roughly
equally a musical object and a literary narrative.
An object that may be an advance member of this artform was on display in the National
Gallery of Art East, in Washington, D.C., during the summer of 1990. It consisted roughly of
a cube about 10 feet on each side, whose walls were mirrors inside and out, and whose rather
cramped interior one could enter through a single opening. Perhaps fun houses and chambers of
horror in certain amusement parks would also come close to qualifying.
This genre was actually devised and practiced by Dadaist pioneer Hugo Ball early in the
twentieth century.
Some entries under the category of Performance Art perhaps already exemplify this descrip-
tion. Note also that the description is intended to exclude lieder, obviously already an existing,
 Nonexistent Artforms and the Case of Visual Music 115

I turn now to the transformational model of hybridization for additional

suggested rules for generating new artforms. The distinction between trans-
formational hybrids and synthetic ones, recall, is that in the former case the
new artform is most naturally seen as a variant of just one of its parents, the
other, if any can be identified, serving as more of a godfather.
One principle of generation with wide application might be this. Given an
existing artform, whose objects exhibit certain standard (or defining) features,
and certain variable (or elective) features, and for which certain features are
contra-standard (or disqualifying), posit another artform that is just like the
first except that some standard feature of the objects of the first artform is
now allowed to become a variable, or contra-standard, feature, or else some
variable or contra-standard feature allowed to become a standard feature.
Thus, if we consider traditional sculpture and imagine that one of its standard
(or category-defining) features, immobility and fixed spatial relationship of
its parts, is allowed to become a contra-standard (or category-disqualifying)
one, everything else remaining the same, then we end up roughly with kinet-
ic sculpturesculpture turned in a danceward direction. Or take traditional
oil painting, for which a variable feature is the use of reds, and convert this
to a standard feature, thus generating a monochrome subgenre of painting in
which reds are de rigueur, and the use of other colors a flagrant violation of
an implicit norm. Or take plays, for which number of characters is variable,
and standardize that at two, thus generating a two-person subgenre of theatre,
in which the appearance of a third party, or a lengthy absence of the first or
second parties, is strongly contraindicated. The general idea behind this prin-
ciple is one of modifying existing artforms by either imposing a restriction
on a dimension where freedom had previously reigned, or else opening up
a dimension which serves as a defining constraint, either positively or negat-
ively, on the artform as normally constituted.
Another transformational rule, perhaps a special case of the above, is famil-
iar because it has often been invoked in this century. Take an existing artform
and envisage a reductive or minimalist version of it, where this involves signi-
ficantly limiting the structures, subjects, or contents of the objects produced,
relative to normal procedure in the art.

well-established art, whose allegiance to the category of music clearly dominates its relationship to
literary storytelling.
The notions of standard, variable, and contra-standard properties relative to a category of art
(e.g. artform, genre, medium, style) are taken from Kendall Walton, Categories of Art, Philosophical
Review 79 (1970): 33467.
116 Music

The minimal sculpture of artists such as Donald Judd or Carl Andre can
be seen to have developed by taking constructivist or assemblage (as opposed
to carved or cast) sculpture as it then existed and restricting it to a small diet
of geometrical forms, a unity of material and color, and a severe set of pos-
sible spatial interrelationships. The early minimalist music of Steve Reich or
Philip Glass defined itself as a genre by explicitly eschewing the melodic and
harmonic complexity, the reliance on key movement, and the driving, devel-
opmental, and goal-oriented aspect of traditional tonal music, leaving manip-
ulation of pulse, rhythm, and phase as almost its sole sphere of operation. The
field of abstract painting generally, obviously enough, can be thought of as
having arisen as an application of this rule, by deliberate exclusion of rep-
resentational matter from the possible content of a painting. An important
variety of minimalist cinema (as exemplified by some films of Michael Snow)
can be understood to have come about through taking the filmmakers usual
freedom of shot and camera movement and jettisoning those in favor of a
self-imposed, spartan limitation to very long takes and fixed camera positions.
The appropriational art of the photographer Sherrie Levine, finally, can be
construed as an instance of the rule, where photography in the robust sense is
transformed into a peculiar minimalist subgenre governed by the restriction
that the only possible subject for a photograph will be a photograph of some
earlier, usually famous, photographer.
Imagining further applications of the rule of minimalist or reductive
versions of existing arts is not difficult. Take kinetic sculpture, already a clear
product of redrawing the lines around existing arts, and posit a subgenre
which allows motion only in a single plane, where that plane is perpendicular
to the line of sight of the viewer; the objects of such a restricted kinetic
sculpture would tend to establish a kinship with painting, equal to any
they retained with traditional sculpture in the round. Take the genre of the
short, descriptive poem (e.g. haiku) and impose on it the restriction that all
words except nouns be banned, thus confining the poet to getting his effects,
creating images, conveying ideas, etc. merely through carefully judged and
paced lists of things. Take modern dance and imagine a form of it in which
all movements were to be carried out extremely slowly, almost at the limit
of detection of movement. Now consider another principle for projecting a

The writer Georges Perec initiated a genre of this sort, by producing a viable instance himself:
a novel written without using the letter e, entitled La Disparition. For additional discussion, see
Elster on Artistic Creativity, Ch. 4 in this volume.
 Nonexistent Artforms and the Case of Visual Music 117

transformation of a given art. Instead of imagining a reshuffling of defining

and optional features of objects in an artform, or imagining a reductive
version of a given artform, we might try to imagine how some artforms could
remain themselves and yet be transformed so as to achieve something of the
distinctive effects or characteristic experience of another artform.
Kinetic sculpture has already been offered as an art viewable as the upshot
of the rule of transforming a standard feature of an artformimmobil-
ityinto a variable or contra-standard one, but we could alternatively see
it as originating in the notion of making sculptures that would achieve some
of the distinctive effects of dance, and not initially in an envisaged structur-
al reconception of the art. Now reversing this last thought we could ask, in
accord with our latest principle, how dance, while in some sense remaining
dance, might yet answer to some of the aims of sculpture. Possibly through
the medium of choreotableaux, or static dance. A work of choreotableaux
would consist of an unmoving arrangement of individuals on a stage, with
all aspects of posture and gesture fixed, which would be assumed by dan-
cers and held, for contemplation, during a period of a few minutes. Visual
music, in the form of abstract color film, could be laid down to this same
projective principle: take some visual art and modify it in such a way that
it achieves some of the characteristic effects or experiences provided by pure
music. Obviously this kind of thought projection will be successful in some
cases but not in others; certain artforms are structurally and thematically so
different that we get no purchase on the idea of the one realizing somehow, in
its own medium, the qualities of the other. Thus, if we ask whether the novel
might be modified so as to achieve the characteristic objectives and results of
sculpture, we would have, so far as I can see, no foothold at all.
I would suggest just one further transformational principle for generating,
in the abstract way we have been pursuing, artforms that do not currently
existone that, once more, yields visual music as an output. It is this:
posit an artform, as far as possible like an existent artform, in which the
sense modality centrally appealed to in the latter is replaced by some other
sense modality. The idea is to see whether unprecedented arts, modeled on
established ones, can be built around senses other than those already in play.
This idea is perhaps particularly apt in connection with those unexploited, so-
called lower senses, e.g. touch, taste, smell, proprioception, which are not the
primary basis of any established art at present. In the first case, if we start with
ordinary sculpturewhich is, by and large, for seeing, i.e. ambulatory visual
118 Music

beholdingand replace seeing by touching, we arrive at tactile sculpture.

In tactile sculpture objects are fashioned for palpation and caressing with no
concern for, or even explicit disregard of, how they look. What is only just
tolerated, or even proscribed, for ordinary sculpture, i.e. tactile exploration,
would be here the invited and exclusively prescribed mode of address. Nor
would such tactile sculpture be hard to arrange, in practical terms: blindfolds,
or enclosures with small openings for the hands, could be made good use of. I
leave it to your imaginations what olfactory music, gustatory painting, and
the like, might involve.
I have not mentioned, in the foregoing, principles for projecting nonex-
istent artforms that could someday exist by appeal to radically new means
and media that technological advances will make available. Most of the new
artforms of the future, it is fair to say, will have such advances as a partial
cause. The only problem, from present point of view, is that there is no way to
anticipate what these might beas noted earlier, my toolkit as essayist con-
tains no crystal balland so nothing contentful on this wavelength to add
to my sketch of the domain of possible but unrealized arts. It is hard not to
appreciate the essential technological contributions to, or at least precondi-
tions of, the emergence of photography, etching, cinema, computer music,
musique concrete, video art, holography, earthworks, and jazz; these did not
come about merely because some overlooked combination or transformation
of existing arts was finally hit upon, some underutilization of an existing sense
capacity finally noticed, or some approach of one art to the aims of another
finally worked out.
Still, in many of these cases, one might locate the spur to develop tech-
nological means enabling these arts to exist, or at least a rationale for their
emergence after the fact, in projections of the kind I have been pursuing. On
this way of thinking, photography arises in part because a visual artform cap-
able of the detail and precision of painting or engraving, but executable with
the spontaneity and speed of, say, a charcoal sketch, was really asking to be
brought into existence. Holography comes to fill the space opened up logic-
ally by positing a cross, in the biological sense, between photography and
sculpture; a hologram is a synthetic hybrid that no longer comfortably resides
I am aware that artforms along these lines have already been pursued among and for the blind,
and that the great sculptor Constantin Brancusi created some pieces in this vein, but that does not
alter the fact that tactile sculpture does not exist as a flourishing artform in mainstream artistic
culture.
I am thinking here of the crucial role of the saxophone, a late nineteenth-century Belgian
invention, in the evolution of jazz sound.
 Nonexistent Artforms and the Case of Visual Music 119

under either banner, photography or sculpture, but exhibits features of both:

fashioning by light, and three-dimensional form. Musique concrete, or col-
lage music, while obviously unrealizable without the magnetophone or tape
recorder, just as obviously gets off the ground conceptually by attempting to
transpose the collage and constructivist procedures of modern painting and
assemblage to the realm of sounds in time.
In another vein, it could also be objected to the domain-exploring proced-
ure I have adopted that it mainly looks for new artforms to generate them-
selves out of formal impulses, rather than expressive, moral, social, or
political ones. That is largely so, but I think the form-based procedure adop-
ted provides the only effective means to sketch any significant portion of the
uncharted terrain. In any event, certainly the motivation toward new artforms
can be either formal or material. In searching for new formal combinations,
we may discover that there are things we can and want to express that wer-
ent possible before, or ways of embodying moral attitudes or advancing social
claims that hadnt been open to us. On the other hand, in searching for out-
lets for new expressive impulses, or in striving to put forward novel moral,
social, or political perspectives, we may equally well induce new formal com-
binations in response. There is, in short, a healthy dialectic between form-
al/structural and material/expressive impulses toward, and justifications of,
previously nonexistent artforms. New vessels may call forth new contents, and
new contents may call forth new vessels capable of holding them.
In summary, we have seen that there is a rich domain of nonexistent art-
forms that can be at least abstractly mapped by reference to the field of exist-
ing artforms. If and are existing arts, then there are those whose objects
involve mere juxtaposition of s and s, ones whose objects are structurally
intermediary between s and of s, ones whose objects are structural com-
pounds of s and s, ones derived from or by revaluation of what prop-
erties are standard, variable, or contra-standard for objects of the art, ones that
represent minimal or reductive versions of a given artform, ones that aim at
the characteristic effects and experience of a given artform through other means,
and ones that derive from a given artform by substitution of the sense modality
involved.

II THE CASE OF VISUAL MUSIC

I turn now to the question of why nonexistent yet possible artforms remain
nonexistent, or more exactly, why any number of them that seem appealing
120 Music

in prospect, conceptually unproblematic, and technologically feasible are yet

not pursued with any vigor. I think it is clear that there is no informative gen-
eral answer to be given. Naturally we can always say that either the potential
for fruitful development is simply inadequate, or that for historical, cultural,
or even political reasons such potential as is present is simply not exploited,
but such blanket, almost dormatively virtuous, explanations are singularly
unilluminating. Case-by-case investigations thus seem to be what are called
for instead. So what I do next is examine the case of visual musicto me the
most intriguing of the envisagable yet stubbornly absent artsto see what
insight we can get into at least its nonexistence. We may, in the course of
this investigation, glean some hints as to what distinguishes those arts, con-
ceivable in the abstract but remaining stubbornly unrealized and unpursued,
from those that have a chance of emerging from the shadows of the merely
possible.
Why, then, are there no successful analogs to pure music in other sensory
realms, no enthralling temporal patterns of colors, smells, feels, tastes? Why,
in particular, isnt the art of abstract color filmat least in any manifestation
known to mecomparable to music in interest and value? Why dont pure
patterns of colors in timetemporal successions of huesgrab us as do tem-
poral patterns of pitched sound? Why dont they transfigure us the way great
music can, or at least captivate us, as music that is merely good does?
A short answer would be that there is surely some insufficiency in the struc-
tures of which abstract color film is capable, or some insufficiency in the
human visual system relative to the auditory, or else some mismatch between
those structures and the capacities of that visual system. But let us look to
more specific explanatory factors, in the hope of giving more body to the
schematic diagnosis just tendered.
First, there is the fact that musictraditional Western music, at any rate
makes use of tones at fixed pitch levels, not of sounds drawn at will from
just anywhere in the pitch continuum. For abstract color film to mirror this
feature would require at least a prior systematic decision to employ a certain
subset of hues, and no others, at precise points along the color spectrum.

The hero of J. K. Huysmanss novel A Rebours, Des Esseintes, has a mouth organ constructed
that squirts a sequence of liqueurs on his palate in analogy with the peal of a carillon. That fictional
example is about as close as anything known to me gets to what one might call taste music.
For the purposes of this essay when I speak of music I have predominantly Western tonal
music in mind.
 Nonexistent Artforms and the Case of Visual Music 121

Second, there is the crucial role in tonal music of scales, chords, octaves, or
more generally, preferred or privileged relations between tones, many of which
are grounded in the harmonic series of a vibrating string. These relations,
which are the precondition of so many of the fundamental qualities and effects
of such music, e.g. tension and release, consonance and dissonance, cadence
and closure, tendency and resolution, do not seem to have any obvious par-
allel in the color realm. The only possible candidate, perhaps, would be rela-
tions of complementarity among hues, but that is not nearly as rich as the web
of relationships governing how the different notes of a musical scale stand to
one another.
Third, there is the notion of key (and key feeling), and the consequent pos-
sibility of key change (or modulation), which is the basis for so much of the
large-scale dialectic of a piece of music, and for a good deal of its emotional
power. It is hard to see what might do duty for this in the color realm, unless
it would be something like the contrast between the warm colors (red, yel-
low, orange) and the cool ones (blue, green, violet), and the potential force of
shifts from one set to the other.
Fourth, phenomenologically speaking the series of pitches employed in ton-
al music form a space-like dimension, in which pitches seem to exist at definite
distances from one another, and in which an experience of motionrather
than just one of successionis thus enabled in connection with change of
pitch over time. We hear movement in the course of a melody, and not just
one note replacing another; we hear the music, as it were, go up and down,
and do so with various predicates of pace, strength, intensity, and effort.
The series of hues in the spectrum, by contrast, does not appear to form a
space-like dimension in the same way; there is no clear impression of direc-
tionality in the succession of hues from red to violet, nor is there the sense
of an underlying variable property which binds together the different hues,
and in respect of which alone they differ from one another. By contrast,

Several of the factors I will cite as militating against the potential of the color realm for
musicality are discussed in chapters VII and VIII of Edmund Gurneys magnum opus, The Power
of Sound ([1880], New York: Basic Books, 1966). (See my Gurney, Edmund, in Edward Craig
(ed.), Routledge Encyclopedia of Philosophy (London: Routledge, 1998).)
One development of this idea is Gurneys notion of Ideal Motion, the distinctive characteristic
of music as heard. An essay that deals with the essential spatiality and motion of music as experienced
is Roger Scruton, Understanding Music, in The Aesthetic Understanding (London: Methuen, 1983),
77100. See also his monumental Aesthetics of Music (Oxford: Oxford University Press, 1997).
To spell this out a bit more, I am suggesting that there is no inherent necessity in the order of
the colors in the spectrum, though of course that order is physically fixed and grounded, and that
different orders, within limits, could be allowed to be as acceptable progressionsnot violating
122 Music

when one proceeds from C to C there is an unavoidable impression of rising,

and the strong sense of a binding propertyhigh/low-ness of pitchthat is
being progressively modulated. For these reasons, it seems, succession of hues
is not inevitably, and certainly not readily, perceived as a kind of motion or
movement, but only as change simpliciter.
If this is so, then there are more weighty consequences as well. Without a
sense of dimensionality and directionality in the realm of pitches, i.e. without
there being a space of pitches, there could be no experiences of movement
through music, or even of melodic shape or contour. But without the possib-
ility of hearing shape and movement in music, we would be barred, a fortiori,
from hearing such shape and movement as gesture, and in turn, such gesture as
the imagined expression of emotion. Thus, music without its spatial/kinetic
aspect would be incapable of embodying expressiveness in the chief way that
it seems to do. So if color sequences are unable, through measured succes-
sion of colors, to provide the experience of shape and motion in the color
realmand not just of rhythmic patternthen their potential for emotion-
al expression even faintly approaching that of music seems dim. In general,
the prospects of color sequences ending up meaning anything much, even if
they managed to be perceptually absorbing, seem less than bright if we end up
being unable to perceive such sequences in a spatially dynamic manner, which
further resonance with worldly and human matters would seem to require.
If in light of the above we accept that there can be nothing like the spatial
and motional possibilities we are familiar with in the auditory realm available
to us in the chromatic one, the artistic burden carried by rhythm per sethe
temporal patterning of color presentationswill be correspondingly greater.
Unfortunately, though, there is reason to believe that our sensitivity to visual
temporal rhythm is considerably poorer than our sensitivity to rhythm in the
auditory sphere. Small alterations in auditory rhythm are much more readily
perceivable by us, it appears, than comparable alterations in visual rhythms.

any phenomenological impression of underlying directionas that which nature has ordained. For
instance, the order of hues might have been given as GBIVROY, rather than ROYGBIV, without
affronting any indelible sensory intuition.
Also adding to the hearability of gesture in musical progression itself is the fact that in
traditional performed music a listener has a sense of the instruments and physical motions involved
in the rendering of the sounds. It is clear that this, as well as the spatiality of the pitch dimension,
funds the hearing of gesture in music on the abstract plane we are here concerned with. For
additional discussion, see Sound, Gesture, Space, and the Expression of Emotion in Music, Ch. 5
in this volume.
For claritys sake I acknowledge that there is a use of visual rhythm in connection with
static visual presentations, for instance, the progression of colors across a canvas, or the repetition
 Nonexistent Artforms and the Case of Visual Music 123

Our finely hewn sensitivity to speech, as a patterning of sound over time, in

all its inflections and nuances, undoubtedly carries over to hearing and dis-
criminating temporal relationships in music. There is nothing, however, in
the natural repertoire of the developing human that calls upon fine sensitivity
to timed sequences of light.
This bears directly on the critical issue of recognizable basic units in our
proposed visual music. Can the realm of visual music sustain something
analogous to the reidentifiable melodic-rhythmic unitsmotives, phrases,
figureswithout which musical experience in the ordinary sense would
be unthinkable? To what degree could such visual motives or phrases be
remembered as such and recognized, if only subconsciously, upon re-
occurrence? How fine may our discriminations for such temporal visual
motives be? How easily would we tell apart two color motives which are
the same except for a small difference in internal rhythm, e.g. a ratio of
two-to-one versus three-to-onea difference, however, which is salient to
the ear? I suspect that our discriminations of such motives would be much
coarser, and our retention of them less reliable, than of motives in music.
Curiously enough, this might even have something to do with the fact
that musical phrases are something we have the capacity to give back or
reproduce by singing, whistling, hummingwhereas we obviously have no
natural capacity to generate or mimic measured sequences of color or light;
the capacity to reproduce may facilitate, or even be part and parcel of, our
grasp and retention of temporal sonic form.
There is, next, the question of transposition of such visual motives or phrases
to other hue levels. Motives and phrases in music can be identified when
transposed to another pitch level or harmonic setting, and this possibility
is clearly crucial for most kinds of musical development (most obviously,
for melodic sequencing). It is not clear, however, whether a temporal visual
motivea sequence of colors set to a rhythmcould have that kind of
reidentifiability. To compound matters, the color spectrum is not extensive
enough to allow for much moving around of a color phrase in hopes of
reidentification at a lower (redder) or higher (violeter) position; if we think
of it, charitably, as having two octaves, red-to-green and green-to-violet,

of ornamental figures on an architectural facade, but when I speak of visual rhythms here I mean
temporal visual rhythms, that is, patterns of durations of color presentations in time.
The fact that we are well adapted to picking out the motions of persisting objects, such as
hands or birds, through space does not automatically argue for a capacity to grasp the temporal
patterns of changing visual presentations at fixed positions in space.
124 Music

then this allows at a maximum twenty or so hue levels at which such a

phrase might be perceived, as opposed to around a hundred for the standard
compass of a piano, string quartet, or symphony orchestra. But it is quixotic
to think that such a phrase would, in virtue of the relational invariants it
embodied, be reidentified at any other hue level; we just dont seem to be
sensitive to, and abstractive of, rhythmic color sequences in the right way.
The limited extent of the color spectrum, in terms of what seem significant
differences in hue across its span, would also, obviously, impact negatively on
the number of theoretically possible color phrases or motives.
I mention just three more factors which seem to militate further against
the possibility of a successful visual music. First, the experience of duration,
or sustainedness, of merely a held tone, is more vivid than that we receive
from a held visual presentationthe former is perceived as a happening or
event, even while in some sense unchanging, while the latter strikes us as a
non-happeningless of an event than a fact or condition. Thus, even where
there is no progression in a straightforward sense, a steady musical state or
situation will always seem more alive to us than a corresponding visual one.
Second, the ease and naturalness of combining separate lines individual pro-
gressions of tonesin the auditory realm, yielding polyphony or counter-
point, is in strong contrast with the artificiality of an analogous procedure
in (say) an abstract color film. Two or more tones or progressions of tones
can be heard simultaneously and distinctly in the same region of pitch space
(or register), whereas two or more colors cant occupy the same region in
two-dimensional space, or in ones visual field, at the same time. Thus, to
realize a multiplicity of chromatic lines in an abstract color film would neces-
sarily require their presentation side-by-side, or one atop the other, or in some
other spatial configuration, it being arbitrary which was elected, and it being
far from clear that such independent and spatially separate streams could be
integrated by the eye into anything like the blending and interpenetration of
lines we experience in the sonic domain.
Third and finally, counting once more against the possibility of a
comparably gripping visual music, is the close connection of real, audible
music with primitive emotional utterance. The prominent place of sound,
especially vocal sound (yells, roars, laughs, grunts, squeals), in the expression
of basic emotions and the communication of simple feelings and attitudes, is
something that music, though a highly elaborate and rarefied use of sound,

Differences of timbre obviously have a lot to do with enabling this.

 Nonexistent Artforms and the Case of Visual Music 125

undoubtedly taps into. There seems nothing comparably visceral in human

behavior for temporal patterns of color to resonate with. This might partly
explain why color sequences, however diverting, seem relatively inert, without
life impact, by comparison with sound sequences.
The strikes against visual music just reviewed, uncovered in the course of
our abstract examination of the question, seem a plausible explanation of why
visual music is not a thriving enterprise. But it would be premature to fore-
close on visual music completely, simply because of the possible handicaps
we can expect it to labor under. For only empirical experimentation, after
all, carried out with artistic imagination, will really show conclusively what
can, or cannot, be achieved in this direction. Since I am not equipped to
conduct such experimentation, I will just conclude with a few suggestions
guesses reallyas to how ventures in abstract color film might best proceed,
with the aim of achieving something of musical effect.
First, it seems clear that a limited set of positions along the spectrum of
hues should be selected and adhered to, as a background matrix, analogous
to the domain of fixed pitches out of which tonal music is made. Second, the
kinds of rhythms employed should be fairly simple, and the differentiation
between different rhythmic figures fairly blatant. Third, moderate tempos are
probably advisable, if there is to be much hope of fixing on isolatable tem-
poral color patterns as distinctive and memorable, in partial analogy with
motives and melodies. Fourth, efforts should be made to exploit the rela-
tions of complementarity among colors, as the only likely analog of those
special relations of kinship and remoteness, attraction and repulsion, among
pitchesthe fifth, the third, the leading tone, the octavewhich have so
great a role in giving tonal music its teleological aspect, its tendency and dir-
ectedness. Fifth, the potential of split-screen, in the attempt to generate some-
thing like polyphonic and interactive effects between two color streams, or
even just in emulation of the effect of tune-and-accompaniment, should not
be ignored, though the problem of the inherent artificiality of such a device
should be somehow faced head-on. Sixth, incremental use could certainly be
made of variation in parameters which I have assumed held fixed until now,
namely those of brightness and saturation, as opposed to exact hue, of each
color presentation or visual note. If variation in hue serves, in color music,
as the only possible analog of movement in pitch, then variation in brightness
and variation in saturation might, suitably managed, come to seem analog-
ous to dynamic (loudness) and timbral (tone color) variation in the musical
realm.
126 Music

Finally, if all choices in the dimensions just reviewed were to prove in

vain, so far as generating a sequence of color presentations that would
have the interest, appeal, and expressiveness of even the simplest music was
concerned, one might then open the door to manipulations of size, shape, and
two-dimensional patterning of color presentations as wellexpanding and
contracting color fields, swirling and intermingling chromatic constellations,
many-hued patches dancing from left to right, and corner to corner, etc.
But note that such visual music would no longer be a strict analog of
monophonic or simple polyphonic instrumental music, whose basic material
is just sequences of pitches-with-durations. The sort of visual music we have
just envisaged, which is in effect a kind of abstract painting in time, might
well be expected to have more power and impact than the strict, spartan
music of colors whose prospects have been my proper target in this section;
for one, it would partake of all sorts of representational resonances which
forms in space, and especially changing ones, possess. Thus, whatever greater
success such expanded visual music may achieve does not thereby vindicate
visual music in our original conception. Moreover, it is far from clear, from
extant examples of the genre, whether spatially unrestricted sorts of abstract
color film, even if aesthetically more engaging than what can be achieved by
pure chromatic sequences per se, are more engaging in virtue of affording an
experience closer to that which musicliteral musicaffords than is open
to us through visual music narrowly conceived.

III AFTERTHOUGHTS

Not long after completing this essay I discovered some reflections of Ernst
Gombrich on the subject of visual or color music. He notes that attempts
to build a color piano may go back to the sixteenth-century painter Arcim-
boldo, that the idea was revived in the eighteenth century in the form of

The films of the Canadian filmmaker Norman McLaren, which might be thought to be an
actual counterexample to my pessimistic diagnosis of the possibility of pure color music in this
section, are, so far as I can tell, of this impure type. That is to say, they make use of resources,
particularly spatial ones, which go far beyond the idea of a rhythmic succession of simple hue
presentations, or even two or three such successions simultaneously. In addition, a number of them
even have soundtracks, which work in counterpoint with the complex color images constituting the
purely visual component of these films.
See Gombrich, The Sense of Order. (Ithaca: Cornell University Press, 1979), epilogue.
 Nonexistent Artforms and the Case of Visual Music 127

a keyboard instrument that produced colored ribbons as well as notes with

every touch, and that Sir David Brewster, the inventor of the kaleidoscope,
predicted in 1819 a great future for an art wherein combinations of forms
and colors may be made to succeed each other in such a manner as to excite
sentiments and ideas with as much vivacity as those which are excited by
musical composition. Gombrich wonders, as I have done, whether color
music can be viable in the absence of something which, in chromatic terms,
would be a functional equivalent of the tonal system:
Could a visual music be developed which would not only offer a succession of moods
of pleasing kaleidoscopic patterns, but set up a field of force which could lead to
that interplay of expectation and fulfilment, of tension and resolution, which is the
stuff of music? He would be a bold man who would dare to predict that such an
experiment could never succeed.

I am not that bold man, but on the other hand, neither would I venture the
opposite prediction, that such experiments must inevitably succeed.
Another writer who has treated this topic, I belatedly discovered, is Peter
Kivy. Speaking of his sense of wonder at the very existence of music alone,
Kivy suggests that it begins with what seems to me to be the genuine, if insol-
uble, mystery of why we have pure music at all, and why, since we do, we
dont have music for our other sense modalities. Nevertheless, Kivy pro-
poses an evolutionary explanation for why humans have not developed visual
music, in particular. It is, according to Kivy, because we have evolved hard-
wired to see defensively, and so to unstoppably place a representational (or
realistic) interpretation on visual perceptions, thus foreclosing on a tempor-
ally extended art of purely visual phenomena.
This explanation, even on its own terms, seems to me to fail. For in fact
we do not, faced with abstract color sequences, have a noticeable inclination
to perceive them representationally or realistically. They remain for us what
they at first seem to be, namely, temporally evolving chromatic patterns. The
explanation for their relative non-involvingness has to be sought elsewhere
than in their putatively putting us frustratingly at cross-purposes with what
evolution has disposed us to do with visual data.

Ibid. 287. Ibid. 305.

See Kivy, Music Alone. (Ithaca: Cornell University Press, 1990), ch. 1.
Ibid. 12. Ibid. 4.
128 Music

More importantly, even were this explanation on the mark, it would remain
too external. Surely we need to appeal to something about the intrinsic nature
of sounds as opposed to colors, or at least to salient features of familiar musical
systems that have been constituted on the basis of the inherent potentialities
of the sound medium. It is that sort of structural, and thus largely intern-
al, explanation that has been offered here for the persisting absence of visual
music.
 8
Music as Narrative and Music as Drama

I INT RODUCT ION

To what extent can instrumental music be viewed as narrative in character, or

understood as involving some sort of narrative? That is the central question of
this essay, but a number of subsidiary questions will also guide my reflections.
I will be interested in the varying potential for narrative construal of different
forms of music, some of which may invite, and some of which may resist, such
construal. I will be interested in what musical narrative, when present, might
be a narrative of. I will be interested in whether musical narratives can pos-
sess certain of the features of standard narratives, such as literary or cinematic
ones. But I will finally also explore, at some length, the appeal of an alternative
construal of music: as dramatic, rather than narrative, in nature.
It is salutary to ask ourselves, at the outset of an inquiry like this, exactly
how often pure instrumental music impresses itself upon us as needing to
be construed narratively in order to be understood. The answer, it seems to
me, is not very often. This is of course not to deny that in appreciating such
music we are made to focus on sequence and progression, from note to note,
phrase to phrase, and section to section. For that is what following music by
ear largely consists in. But it is a large step from that to the claim that music is,
in its sequence and progression, narrating a story of some kind as it unfolds.
Still, the idea that instrumental music, and especially the extended musical
essays of composers of the eighteenth, nineteenth, and twentieth centuries,
might be understood as narratives is a staple of humanist music criticism,
as exemplified by Donald Tovey, George Bernard Shaw, Leonard Bernstein,
Antony Hopkins, Charles Rosen, Andrew Porter, Alex Ross, and others. It is,
after all, not surprising that music, as an intentionally arranged, temporally

First published in Mind & Language 19 (2004): 42841.

130 Music

extended sequence of sounds, one that often displays a character of utterance,

is readily thought of as recounting something or other, and likely something
that is itself temporally extended, such as a sequence of actions, events, or
mental states.

II THE CONCEPT OF NARRATIVE

How should narrative be characterized? For our purposes, the more simply
the better. One recent writer suggests that a genuine narrative requires the
representation of a minimum of two events and some indication of the order-
ing in time of the events depicted. If so, then there are three crucial elements
to narrative: representation, events, and temporal relations. Another recent
writer has proposed that a narrative must, in addition, indicate causal rela-
tions obtaining among represented events: the basis of the narrative connec-
tion is that earlier events and/or states of affairs are at least causally necessary
conditions, or contributions thereto, for the occurrence of later events in the
relevant stories.
Applying this to music, then, if music is to be narrative (a) it must repres-
ent; (b) it must represent events or states of affairs; and (c) it must repres-
ent temporal and/or causal relations among those events or states of affairs.
Accepting these as the minimal features of narrative, the task would then be to
assess whether any pure instrumental musicthat is, music without program
or textin fact displays them. The prospects do not seem bright. The third
condition, in particular, seems especially hard to meet, as it looks as if it would
require the sort of temporal and singular referential devices that language, but
not music, possesses. Yet that is where the distinctive feature of narration, as
opposed to non-narrative representation, would seem to lie.

George Wilson, Narrative, in Jerrold Levinson (ed.), Oxford Handbook of Aesthetics (Oxford:
Oxford University Press, 2003), 392407, at p. 393.
Noel Carroll, On the Narrative Connection, in Beyond Aesthetics (Cambridge: Cambridge
University Press, 2001), 11833 at p. 133.
An alternative tack would be to locate the specificity of narrative representation in there being
a discernible narrator internal to the representation, who tells or recounts the events and relations
in question. But since a number of philosophers have argued for the possibility of narratorless
narratives (see Kendall Walton, Mimesis as Make-Believe, (Cambridge, MA: Harvard University
Press, 1990); Gregory Currie, The Nature of Fiction (Cambridge: Cambridge University Press,
1990); and George Wilson, Le Grand Imagier Steps Out: The Primitive Basis of Film Narration,
Philosophical Topics 25 (1998): 295318), it is not a tack to be taken lightly.
 Music as Narrative and Music as Drama 131

III OBJECTS AND CONTENTS OF MUSICAL

NARRATIVE

If musical narratives narrate events, a question arises immediately as to wheth-

er those events are musical or non-musical ones. In other words, one possibil-
ity is that music somehow tells a tale of musical events, such as the inversion
of a motive, or the arrival of a cadence, or a modulation from B-flat to E-flat.
Another is that music somehow tells a tale of non-musical events. Presumably
the narratives of interest in music are of the second kind, for those of the first
kind, to the extent they could be made out, would seem redundant. Since the
sequence of musical events is directly present and immediately heard, what
would be the point of its being narrated as well?
In order to locate more plausible objects of musical narration, it will be
helpful to offer a brief sketch of the musical apprehension of the ordinary,
though musically sensitive and practiced, listener, at least as other than a strict
formalist would conceive it. The experience of apprehension can be seen as
comprising different stages or levels, though it is a mistake to suppose that
these are either entirely separable or clearly sequential. In any case, first there is
the level at which the most elemental properties of tones, such as pitches, dur-
ations, and timbres, are perceived. Then, following immediately upon that,
is the level at which rhythms, motives, phrases, melodies, and harmonies,
the fundamental building blocks of music, are perceived; note that that level
involves the hearing of motion in music, rather than mere succession or altera-
tion. Those two, perhaps not fully distinguishable, stages might be labeled the
configurational level of musical apprehension. Next there is the level of gesture
or action, that at which one hears the music doing something, something it is
not literally doing. Note that the sense of the real gestures of the performers
who are performing the music enters into the gesturecall that musical ges-
ture that one hears the music to be engaged in, but the musical gesture and
the conjectured or imagined performing gesture are not the same. We can
label that the gestural (or actional) level of musical apprehension. Next there
is the level of states of mind heard behind the gestures or actions perceived at
the preceding level of which those gestures or actions are the expression; that
is the expressive level of musical apprehension.
Assuming this rough picture of musical experience, what are then possible
narrative objects of music? They would seem, at a minimum, to be these:
(a) gestures, (b) actions, (c) expressions, (d) mental states. As for the content
132 Music

of a musical narrative, it would presumably be some sort of sequence of the

preceding. Suppose for the sake of argument we focus on option (d), and
construe the narrative content of music to be a sequence of mental states. A
typical specimen would then go something like this: First S1, then S2, then
S3 . . . , where the Sns are mental states, plus whatever relations among the
Sns might be implied, such as S1 evolves into S2, S2 results from S1, S2
is a reaction to S1, and the like. So, does expressively varied music indeed
relate such a narrative? Do we, at any rate, hear such a narrative in such music?
Agnosticism about that seems highly warranted, to say the least.
We should observe, in addition, that certain features central to standard
narratives of a literary or cinematic sort, even if not accounted essential to
narrative, seem virtually impossible to locate in instrumental music. These
include the capacity to predicate of a subject, the capacity for reflective self-
commentary, and the capacity to clearly signal pastness or futurity. To those
three features difficult to imagine pure music exhibiting, we can add others,
such as narrative voice, narrative point of view, and narrative true-to-lifeness.

IV MUSICAL REPRESENTAT ION

We need to examine more closely the idea of representation presupposed in

the foregoing discussion. Taking representation to be a sine qua non of narra-
tion, if instrumental music is to count as narrative we will have to establish,
first, that such music does indeed represent, second, that it represents non-
musical events, and third, that such music represents in a narrative mode, that
is, by somehow conveying or recounting temporal or causal relations among
those non-musical events.

See Fred Maus, Narratology, Narrativity, in Stanley Sadie (ed.), New Grove Dictionary of
Music (Oxford: Oxford University Press, 2003). Indication of pastness may not, for all that, be
beyond the powers of music to achieve. The most promising case is probably the emergence of
nostalgia in music, whereby a passage of music appears to relate regretfully to an earlier one, or to
the sentiment or action with which the earlier passage was associated.
This is not to deny that point of view, one attributable to the implied composer, plays a
role in music involving humor, parody, allusion, and the like. Examples include Mozarts Musical
Joke, Bartoks Sixth Quartet, Haydns Joke Quartet, and parts of Ivess symphonies. Such cases
typically involve implicit commentary on something musically referenced then and there. Despite
such examples, it is hard to see how there could be scope for the operation in music of full-fledged
narrative point of view. As for true-to-lifeness in music, for an attempt to theorize that in regard
to facts concerning the realm of emotions, see my Truth in Music (1982), in Music, Art, and
Metaphysics (Ithaca: Cornell University Press, 1990). Though I do not entirely repudiate it, I am
now less sanguine about the approach taken there.
 Music as Narrative and Music as Drama 133

But with the very first task we confront an obstacle. Even if expressive
music, according to the sketch of musical experience offered above, read-
ily induces appropriately backgrounded listeners to hear gesture and expres-
sion in it, is that quite enough to say that the music represents those things?
Unfortunately not. For artistic representation, as exemplified most clearly by
pictorial representation, is a strongly intentional notion. Thus, for a picture
to represent an oak tree an oak tree must be seeable in it, to be sure, but the
maker of the picture must also actually have intended such seeing-in. Artistic
expressiveness, on the other hand, is not a strongly intentional notion, nor is
that of gestural content; a passage of music might be expressive of an emotion
or might embody musical gestures, without its composer having intended the
hearing of those gestures or that emotion in it, and in some cases, without the
composer even having foreseen or anticipated such hearing.
In light of that, it may be that the relation between music and the gestures
and expressions it induces us to hear in it is more properly one of suggestion,
rather than representation. Be that as it may, I will put aside this reservation
for present purposes, and continue to speak of expressive music as represent-
ing the gestures, actions, and expressions that it perhaps strictly speaking only
suggests. That will allow the question of musics narrativity to remain open.

V M USIC AS EXTERNALLY NARRATIVE VS. MUSIC

AS INTERNALLY NARRATIVE

A distinction regarding musics narrativity that will prove to be of some use

is that between music as externally narrativeas something that is being told,
by the composerand music as internally narrativeas something that is
telling of something else. In the first case, the sequence of musical events is
the story; in the second case, the story, if any, is what the musical events
are about. Of course, where music can be made out to be both externally
and internally narrative, the responsibility for the internal narrative ultimately
rests with the composer as well, the composer being the teller, in the first
instance, of the sequence of musical events, and in the second instance,
though sometimes unknowingly, of the sequence of non-musical events the
musical sequence represents.
In connection with musics external narrativitythat is, the idea that the
events of which the music consists are themselves being narratedthere is
also the role of the performer to consider. The performer might be thought
134 Music

of as occurrently narrating the events that the composer has only narrated
in a standing fashion, or alternatively, as the only proper narrator of musical
events on any occasion of performance, the composer then not being a narrat-
or of any sort, but only the designator in the abstract of events for narration.
Having for the moment gone as far as I can with the idea of music as narrat-
ive, I want now to consider an alternative idea, that of music as drama.

VI MAUS AND NEWCOMB ON DRAMA IN MUSIC

In this section I summarize two analyses of musical compositions by

philosophically informed musicologists that can be seen as recommending
a dramatic, as opposed to narrative, model of the events musical compositions
appear to image forth. The first is due to Fred Maus, and concerns the opening
movement of Beethovens Quartet in F minor, Op. 95; the second is due to
Anthony Newcomb, and concerns the scherzo of Mahlers Ninth Symphony.
Mauss analysis, which focuses rather minutely on the arresting first
eighteen measures of Beethovens quartet, identifying in it individual gestures
standing in certain relations to one another, issues in the following general
conclusion:
It would be natural to call the quartet a conspicuously dramatic composition. The
analysis makes the sense of drama concrete by narrating a succession of dramatic
actions: an abrupt, inconclusive outburst; a second outburst in response, abrupt and
coarse in its attempt to compensate for the first; then a response to the first two
actions, calmer and more careful, in many ways more satisfactory.
Maus notes that his analytical description of the passage explains events
by regarding them as actions, and by venturing motivations for those actions.
But to whom, Maus asks, are those actions and motivations being ascribed?
Neither the composer nor the performers, it seems, for though they are the
authors of certain music-related actions, namely, composing and performing,
they are not comfortably thought of as the authors of the actions heard in the
music, actions such as asserting, objecting, responding. Rather, such actions,
and associated motivations, are to be ascribed to a persona, or personlike

Maus, Music as Drama, in Jenefer Robinson (ed.), Music and Meaning (Ithaca: Cornell
University Press, 1997), 10530, at pp. 11819 (originally in Music Theory Spectrum (1988)).
Note that although Maus invokes narrating in this quotation, the narrating is being done by the
analysis, not by the music being analyzed.
 Music as Narrative and Music as Drama 135

agent, who is imagined as present, performing those actions in real time. In

listening to a piece, Maus observes, it is as though one follows a series of
actions that are performed now, before ones ears, not as though one merely
learns of what someone did years ago.
Maus proposes that a piece of instrumental music such as the quartet move-
ment he analyzes be assimilated to a play, and thus seen to have an essen-
tially dramatic character, rather than assimilated to a narrative form like the
novel. Plays have a number of salient features which music can share. They
present a series of actions, which are the actions of fictional characters or per-
sonae, which actions are experienced as occurring as they are perceived, and
which form a plot, or at least, make some kind of sense as a whole. Of course
the actions heard in music are not as concrete or detailed as the actions of a
stage play, which more nearly approximate those of life. The agents, objects,
and motivations of musically embodied actions remain much more indeterm-
inate. Music is, as Maus suggests, a kind of drama that lacks determinate
characters. This difference in degree of indeterminacy of the characters or
personae involved in music and in theater should not be thought to under-
mine the validity of the analogy. For, after all, playwrights such as Strindberg
and Beckett have created stage plays, such as Endgame and The Ghost Son-
ata, whose personae are almost as indeterminate as those to be heard in the
expressive instrumental music with which we are concerned.
Anthony Newcomb, for his part, maintains that music is heard is a reenact-
ment of a complex pattern of intentional human action, and his analysis of
Mahlers movement is designed to illustrate that. According to Newcomb, the
imagination of agency in music, in schematic form, goes as follows:
[first] the selection of [or focusing on] musical attributes . . . the interpretation of
these musical attributes as attributes of human character or behavior . . . the com-
bination of these human attributes in various configurations as possible or plausible
human agencies . . . [and finally] the understanding of . . . these fictional agencies as
relevant in the unfolding of a plausible chain of human actions and events.

Maus, Music as Drama, 121. Ibid. 128.

Anthony Newcomb, Action and Agency in Mahlers Ninth Symphony, Second Movement,
in Jenefer Robinson (ed.), Music and Meaning (Ithaca: Cornell University Press, 1997), 13153,
p. 131.
Newcomb hastens to add this qualification: It is important to realize that in music, as in other
arts . . . aspects of agency are not continuously displayed . . . Even the most expressive music . . .
at times simply swirls or dreams or chugs along in its decorative function (ibid. 133).
136 Music

Newcomb locates in Mahlers scherzo three dances of different character,

which are introduced individually, and then interwoven as the movement
proceeds. The first is a medium Landler (A), the second a fast Waltz (B), and
the third a slow Landler (C). When the second dance first appears, it suggests
an agency distinct from that of the first dance, but there are two reasonable
possibilities for what that agency is: either a second, entirely distinct persona,
or else an element within the personality of the first persona. The cultural-
historical context of early twentieth-century Vienna, awash in the ideas of
Schopenhauer and Freud, and sympathetic to the notion of hidden sides of
the human psyche, favors the latter way of reading the contrast between the
agency embodied in dance A and that embodied in dance B. Overall, New-
comb suggests, the movement offers a picture of a clumsy and coarse rustic
personality swept away by a sophisticated and confident urban one, followed
in due course by a sober and reflective personality, associated with dance C,
who serves to rein in the second and perhaps restore, in some measure, the
honor of the first.
In the remainder of this essay I attempt to bring into relief further aspects
of the dramatic and narrative models of musical content, and to weigh the
respective merits of those models in regard to music of different kinds.

VII MUSIC AS DRAMATIC VS. MUSIC AS NARRATIVE

Music is expressive, I maintain, when it prompts us to hear the music as anim-

ated by agency of a certain sort, more specifically, when it induces us to hear
the music as the expressing of a mental state, or perhaps equivalently, when it
induces us to imagine a persona expressing a mental state through the vehicle
of music. I call this the hearability-as-expression view of musical expressive-
ness. But is a sequence of passages that are expressive of a sequence of states
of mind thereby an emotional narrative? Only, it seems, if we have the sense
that the first sequence involves acts of relating or telling, ones attributable to
an agent who stands apart from the imagined agent or agents who are the sub-
jects of the mental states and acts of expression that constitute the musics
expressive substance. Otherwise, as the analyses of Maus and Newcomb

See Musical Expressiveness as Hearability-as-Expression, in Matthew Kieran (ed.), Contem-

porary Debates in Aesthetics and Philosophy of Art (Oxford: Blackwell, 2005), and Ch. 6 in this
volume.
That is, in the terms introduced earlier, if the music appears internally, and not just externally,
narrative.
 Music as Narrative and Music as Drama 137

suggest, we would seem more justified in construing the music according to

a dramatic model, on which personae and their expressive actions appear dir-
ectly on the musical stage, than according to a narrative model, on which
those personae and expressive actions are instead being recounted to us by a
narrative agent.
The locus classicus of the distinction between the narrative and the dramat-
ic is Aristotles Poetics, the first serious attempt in Western thought to offer
a taxonomy of the arts. Aristotle invokes the distinction to explain how epic
and tragedy are distinct. He notes that though they have the same objects of
imitation, namely, noble human beings in serious situations, and roughly the
same means of imitation, namely, persons uttering words, they differ non-
etheless in the manner of imitation: whereas epic speaks of the events with
which the epic is concerned, tragedy presents those events directly. Alternat-
ively, the epic reciter describes the events of the story, while tragic actors enact
those very events.
Again, on the view of musical expressiveness I hold, we hear expressive
music precisely as if it were an expression of mental states. More generally,
we hear such music, and much not specifically expressive music, as imbued
with action. That is, we standardly hear music as acting in ways it is not liter-
ally acting, or doing things it is not literally doing, or gesturing in ways it is
not literally gesturing. In the case of expressive music, the actions are, natur-
ally, expressive actions, but in other cases they are actions of other sorts. From
such a perspective, clearly, the most natural way to view instrumental music
is as dramatic, that is, as offering a sequence of actions to be directly ima-
gined, rather than as narrative, that is, as offering a sequential relating of such
actions.

VIII CONDITIONS FAVORING NARRATIVE

CONSTRUAL OF M USIC

When are we inclined to regard the actions that we hear in music as directly
present to us, enacted by personae as we listen, and when as matters that are
not directly present, but instead represented in a narrative conveyed to us by a
narrating agency, whether the composer, the performer, or a narrator internal
to the music? A most difficult question. Let me, then, pose a simpler version of
it. When do we have the sense that unadorned instrumental music is relating a
story to us, that such music is, in terms invoked earlier, narrating, and not just
being narrated?
138 Music

I suggest, first, that the music must have a marked character of utterance, of
seeming to speak, if that sense is to emerge. And not all music displays that.
But second, there must also be a character more specifically of storytelling.
What I mean by that is a measured, deliberate, reflective character, such as
is conjured by phrases like once upon a time or long ago and far away.
Music of storytelling character recruits features traditionally associated with
the storytelling mode of discourse, one we think of perhaps above all as unhur-
ried. Thus slowish tempo, relaxed rhythm, and restrained dynamic are among
the musical features conducive to the emergence of a storytelling character.
An example of music with something like this storytelling character would
be the opening of Smetanas Ma Vlast. This is, admittedly, an explicitly pro-
grammatic work, but non-programmatic examples may also be found. The
openings of Bruckners Seventh Symphony, Francks Symphony in D minor,
Mendelssohns Reformation Symphony, or Dvoraks New World Sym-
phony will perhaps serve.
Another generalization we might try on for size is this: The more music
strikes us as a direct utterance from the composer, the more we are likely to
construe it as narrative, as in effect the composer testifying to or recounting
something in sound; and the more music strikes us as constituting a world of
its own, in which events occur independently of a guiding force, the more we
are likely to construe it as dramatic, that is, as involving agencies that appear,
interact, and depart before our ears. Or again: the more the agents imagined
in connection with expressive music seem autonomous or self-directed, the
more apt is a dramatic construal of the music; and conversely, the more the
agents imagined seem framed or subject to outside control, the more a narrat-
ive construal of the music recommends itself.
In any event, plainly not all instrumental music displays narrative character
or lends itself to narrative interpretation. The extent to which it does
seems to depend to a fair degree on the genre of music involved. With
minuets, scherzos, toccatas, etudes, canons, variations, and the like, the
form of such pieces dictates certain structural repetitions or certain kinds
of musical filler which tend to block narrative suggestion or undercut
narrative momentum, whereby such pieces come close to fitting Peter Kivys
characterization of instrumental music as sonic wallpaper. With sonata
movements, by contrast, internal narrativity has more purchase, sonata form

Peter Kivy, The Fine Art of Repetition, in The Fine Art of Repetition (Cambridge: Cambridge
University Press, 1993), 359.
 Music as Narrative and Music as Drama 139

being inherently progressive and developmental, and so more susceptible to

narrative construal.
Still, the most natural construal of a sonata movement is as dramatic rather
than narrative. Sonata movements generally come across like a stage on which
actors appear, directly express their emotions, and move on. (Alternatively, a
single actor may be thought present, who passes through a sequence of states
somewhat in the manner of a soliloquy.) In the opening movement of Charles-
Valentin Alkans little-heard Symphony for Piano, for example, one readily
hears a protagonist, embodied most clearly in a particular five-note motive,
express its peculiar mixture of yearning and suspicion. But one does not seem
to hear a voice that tells of a third party. One does not have the sense of a
mediator between oneself and the expressive gestures that one perceives in the
music.
Perhaps it would be best to admit that music can generally be heard as either
strictly narrative or strictly dramatic, even if most music lends itself more
readily to the latter. When music is regarded as the utterance or voice of the
composer, then construing it as a narrative, one whose content consists of the
gestures, actions, or emotions of the composers alter ego, is natural enough.
On that model, the composer is analogous to a lyric poet, and the performer to
a reciter or rhapsode. When music is regarded rather as the organized product
of creative activity offered for our engagement, then construing it as a drama,
one whose content consists in the gestures, actions, or emotions of various
shifting personae, is arguably more natural. On that model, the composer is
analogous to a playwright, and the performer to a director or producer.
In the last analysis, construing music either narratively or dramatically,
and not simply expressively, might best be regarded as an appreciative option,
not something correct appreciation absolutely enjoins. Even when the option
is exercised it is often irresolvable whether the music in question is better
construed one way or the other. One likely source of this irresolvability as
between narrative and dramatic construals of music is the fact that there are
so many different candidates in the domain of articulate actions for what a
given passage of music might be heard as fictively engaged in: monologue,

A similar claim could be made for fugue, which also avoids strict repetition, and which one
could hear as relating a story consisting in the vicissitudes of its theme, and for jazz improvising
on standard tunes, given that such improvising is often understood as at least in part a kind of
commentary on the tune or its associated chord changes. (For interesting reflections on the narrative
dimension of jazz improvisation, see the third part of Jon Elster, Ulysses Unbound (Cambridge:
Cambridge University Press, 2000).)
140 Music

dialogue, recitation, aside, soliloquy, meditation, diatribe, ballad, chronicle,

report, confession, exhortation, peroration, elegy, admonition, and so on,
some of which are dramatic actions, and some of which are narrative actions.

IX NARROW AND BROAD SENSES OF NARRATIVE

IN MUSIC

Experienced from within, so to speak, a sonata movement strikes us as a

drama, with actors corresponding to themes or motives. This is the way son-
atas are usually experienced, as illustrated by the Alkan example above. Yet
experienced from without, as the deliberate product of a creator, a sonata
movement may take on the aspect of a narrative being related by the com-
poser through the musical drama more immediately apprehended in listening.
Thus, though not narrative in the narrow sense in which it contrasts with
dramatic, the music of a sonata can be considered narrative in a broad sense,
one in which narration is effected by means of drama.
Consider an army commander narrating a past battle, using actors or pup-
pets with recognizable identities or powers. This looks to be analogous to the
stage-managing a composer engages in regarding his musical material and the
personal agents such material gives rise to for imaginative hearing. What this
amounts to, then, is the conveying of a narrative by the creating of a drama.
Of course in this broad sense of narrative playwrights, who directly fashion
dramas for enactment, are also engaged in narrative, even more clearly than
are composers. Macbeth, for instance, certainly conveys a story, and there
seems no objection to thinking of Shakespeare as, in some sense, the conveyer
of that story. But neither in theatre nor in music does this amount to narrating
in the stricter sense that applies to, say, literary fiction, in which a narrator can
be identified in the fiction, in which there is a discernible narrative voice, in
which narrative point of view is robustly present, and in which there is a clear
distinction between narrative discourse and narrated story.

X M USICAL PE RSONAE

What disposes us to think of a musical persona heard in one stretch of music

or portion of the musical fabric as the same as the musical persona heard in
another such stretch or portion? It is very hard to generalize here, but perhaps
at least a contiguous series of musical gestures of similar character will be
interpreted, ceteris paribus, as the gestures of a single, continuing persona.
 Music as Narrative and Music as Drama 141

Yet there is very likely ineliminable indeterminacy in such matters. Speak-

ing of the dramatic personae hearable in the opening of Beethovens Op. 95,
Fred Maus observes that:
. . . nowhere is it clear whether the response to the first outburst is made by the
same agent or agents. If the continuity in the performing forces suggests continuity
of the musical agency, the registral discontinuity and utterly different treatment of
the ensemble may suggest that a distinct agent or agents have entered to respond . . . .
The actions that a listener follows in listening to the Beethoven passage do not belong
to determinately distinct agents. More precisely, as the listener discerns actions and
explains them by psychological states, various discriminations of agents will seem
appropriate, but never with a determinacy that rules out other interpretations.
Nevertheless, music for a single performer is more often than not heard as
communicating a slice of the psychological life of a persisting persona. Even
where there is a clear melody and accompaniment structure, or a contrapunt-
al texture with several distinct lines, one generally has the impression of a
single expressive agency, at the service of which are the various distinguish-
able musical components. By contrast, in certain genres, such as that of the
concerto, with its manifest oppositional structure, playing off soloist against
orchestra, the presumption of single expressive agency is just as clearly over-
ridden. And duo sonatas and trios may also often invite the kind of hearing
normative for concertos, in which different instruments or performing forces
are heard as the vehicles of different personae.

XI A SONATA OF SCHUBERT

I end with a musical example, the first movement of Schuberts Piano Son-
ata in a minor, D. 845. It is instructive to trace the experience of gesture,
expression, and action in this stretch of music, one that, like most movements
in sonata form, sustains hearing in dramatic mode rather better than it does
hearing in narrative mode. A pragmatic difficulty looms, though, in the effort
to convince an audience of this through description of what appears to be
going on in the music. For in doing so one is inevitably involved in narrat-
ing the succession of agents and actions that one hears. Still, the fact that
in making a brief for the dramatic content of the music one perforce narrates
ones experience of that content does not turn the music itself into a narrative
rather than a kind of drama.
Maus, Music as Drama, 123. See n. 6 for the same caution.
142 Music

Three contrasting themes or theme groupscall them A, B, and Ccan

be discerned in this movement, with which three personae or personality
aspects can be readily associated. The opening motive, A, is of a grim, inward
character, while the next, B, is of restless character overall, consisting of three
sub-motives of respectively rocking, flowing, and declamatory nature. The
next motive, C, has a martial and strutting character, which is then succeeded
by a fusion of C and B, which, though still vigorous, assumes a more lyrical
character than anything that has come before, and which suggests in its fusion
of motives a kind of rapprochement between them. Eventually C drops out
and the flowing, arpeggiated sub-motive of B predominates, until the return
of A in a key a third above the tonic. From there until the end of the
exposition, another twenty or so measures, there is an alternation between A
and C which has the air of a dialogue, one concluded gently by C.
The beginning of the development section sees an alternation of A with
itself, distinct contrasts of register and dynamic demarcating the two guises
of A, giving the passage the sense of a soliloquy in which different sides of a
question are being weighed. This is followed, after two held chords serving as
a transition, by a section of more pronounced fantasy character: here the per-
sona of A, though shadowed by the rocking, syncopated sub-motive of B in
the bass, takes wing, ventures into unknown regions, and beginning around
measure 120, wrestles with doubts, the upshot of which is by no means clear.
Dark thoughts continue to accumulate, and the sense of crisis is accentu-
ated by the fragmentation of A, reduced to its last four descending notes,
which plunge somewhat desperately, again and again. This eventually sub-
sides, allowing A to reassert itself once more in full, though now uncertain,
spent of energy, and winding in on itself in a series of remote and inhospitable
keys. There I shall leave our motives and their associated personae, with the
return to the initial order, in the recapitulation, still some sixty measures off.
In the foregoing narrative of a listeners experienceor at any rate, this
listeners experienceof this music, I have mixed technical, expressive, and
agential vocabulary. But the technical bits in the narrative serve only to
pinpoint what is going on in the music on the other two levels, those of
expression and agency. And what is going on, I would hope your experience
confirms, is not so much doings that occurred in another place and time, of
which one is receiving a report, but a drama of events happening here and
now, with indefinite personae which are the shifting loci of the emotions and
actions encountered throughout.
 9
Film Music and Narrative Agency

In this essay I address certain issues about paradigmatic film music, that is,
the music that is often heard in the course of a fiction film but that does not
originate in or issue from the fictional world revealed on screen. What most
interests me is the question, which confronts every filmgoer at some level, and
to which he or she must, explicitly or implicitly, accord an answer, of who or
what is responsible for such music. That is to say, to what agency is film music
assigned by a comprehending viewer, and what is this music understood to be
doing, in relation either to the films internal narrative, the viewers experience
of that narrative, or the film as an aesthetic whole? Furthermore, by what prin-
ciple does a viewer assign, however tacitly, responsibility for the music he or
she hears?
It will turn out that different answers to this question of agency are in order
from one film to another, and even from one cue to another within a given
film. The upshot is a basic division within the realm of film music, one I have
not seen marked elsewhere, but which is probably more fundamental than
others that are regularly noted.

II

I begin with some preliminaries. First, the music I am concerned with is usu-
ally designated nondiegetic film music, that is, music whose source is not the
story (or diegesis) being conveyed by the films sequence of images. It is some-
times also designated soundtrack as opposed to source music, and sometimes

First published in D. Bordwell and N. Carroll (eds.), Post-Theory: Reconstructing Film Studies
(Madison: University of Wisconsin Press, 1996), 25488.
144 Music

extrinsic, as opposed to intrinsic, music. Second, the films I am concerned

with are all narrative fiction films, both of the classical (or Hollywood) sort,
and the modernist (or art film) sort, though not any of the more extreme
examples of the latter, in which the bounds of fictionality or narrative coher-
ence are stretched to their limits. Third, I consider film music here only
as an integral component of a complete film, and not as a genre of music
which, in the form of suites or soundtracks, might be enjoyed and evaluated
on its own.
Certain kinds of answers to our opening queries can be put aside immedi-
ately as not to the point. For instance, the source of nondiegetic film music
might in one sense be said to be the composer who composes it, or the pro-
ducer who commissions it, or the sound editor who integrates it into the
finished film, but this does not address the question of where, in relation to
the fictional world projected, the music is situated or positioned in compre-
hending the film. Similarly, the function of nondiegetic film music might be
said to be, somewhat vaguely, the aesthetic enhancement of the film, or more
specifically, the emotional manipulation of the film viewer, or more crassly,
the augmentation of the films marketability and secondary profits, but none
of these answers addresses the question of how such music is understood to
function in relation to the central narrative of sight and sound, and thus to
contribute ultimately to a films meaning.
It should be noted straight off that there are two basic sorts of musical score
regularly encountered in the domain of the sound film: the first, more tradi-
tional, sort consists of music composed specifically for the film in question,
and generally tailored by the composer to the rough cut, scene by scene; the
second sort consists of pre-existent music chosen by the filmmaker, often in
conjunction with a musical consultant, and applied or affixed to scenes or
parts thereof. Call the former sort a composed score, and the latter an appro-
priated score.
Directly we can make at least two observations about these two types of
score. First, with appropriated scores the issue of specific imported associ-
ations, deriving from the original context of composition or performance
or distribution, rather than just general associations carried by musical style

For more on these categories, see David Bordwell and Kristin Thompson, Fundamental
Aesthetics of Sound in Cinema, in Film Art: An Introduction, 4th edn. (New York: Knopf, 1993).
For instance, Michael Snows Wavelength or Alain Resnaiss LAnnee Derniere a Marienbad.
 Film Music and Narrative Agency 145

or conventions, is likely to arise. Second, with appropriated as opposed to

composed scores, ironically there will generally be more attention drawn to
the music, both because it is often recognized as such and located by the
viewer in cultural space, and because the impression it gives of chosenness,
on the part of the implied filmmaker, is greater. To these two observations I
add a third, more contentious one, that later discussion will support: music
composed for a film (e.g. the soundtracks of Vertigo or The Heiress or On the
Waterfront or La Strada), is more likely to be purely narrative in function
than pre-existing music appropriated by a filmmaker (e.g. the soundtracks of
A Clockwork Orange or Barry Lyndon or Love and Death or Death in Venice).

III

There are some theoretical claims prevalent in the recent literature on film
with which I will be disagreeing, and it is best I signal what they are at the out-
set. One is that nondiegetic film music is standardly inaudible, i.e. is not,
and is not meant to be, consciously heard, attended to, or noticed. This
seems to be clearly false, or at any rate, false for a wide range of films in
which soundtrack music calls attention to itself unmistakably, or requires the
viewer to attend to it explicitly if he or she is not to miss something of narrat-
ive importance. The inaudibility claim seems most true for what is called
underscoring, music at a low volume that serves as a sort of aural cushion
for dialogue that remains the main order of business, or for melodically and
rhythmically unmarked music helping to effect transitions between scenes of
notably different character. Even here, when the music hovers in the penum-
bra of consciousness, it is rarely very far from being consciously focused, as is
perhaps reflected in the fact of being immediately noticed if stopped. If non-
diegetic film music were generally unheard, or not consciously noted by the
viewer, then there would not be much of an interpretive issue for the viewer
of how to construe such music in relation to the rest of what is going on in
the film. But there manifestly is an issue of some significance, with respect to
many films. Finally, even if it were the case that casual viewing of films with
significant music tracks often goes on without a viewers explicit awareness of

This is a central thesis in Claudia Gorbmans Unheard Melodies: Narrative Film Music
(Bloomington: Indiana University Press, 1987), and is echoed by other recent psychoanalytically
oriented writers on film.
146 Music

that music, it hardly follows that an aesthetically justified or optimal viewing

of such films remains similarly oblivious.
Another idea with some currency is the disavowal of what might be called
narration properthe conveying of a story by an intelligent agentas actually
characterizing the standard fiction film. One variant of this has it that such
films are not really narrated by anyone or anything within the film world, but
instead narrate themselves. A second variant insists that such films are consti-
tuted as narratives only by the viewer, and contain no narration apart from
that. A third variant maintains that such films are not only constituted as nar-
ratives by viewers, but are in fact narrated by viewers to themselves as well, in
the course of viewing.
I reject the first sort of disavowal on grounds of incoherence; if narration
means anything, it is the conveying or imparting of a story by means that are
distinct both from the story being conveyed and from that which is doing the
conveying; if the film, or its processes, are the means of narration, then it, or
they, cannot also be conceived to be the agent or source of narration. I reject
the second and third sorts of disavowal because they seem based on conflat-
ing the viewers actual task of comprehending a films story and significance
by actively reconstructing or piecing together the narrative on offer, with the
viewers literal creation of that narrative, which would thus not exist apart
from the viewer. But this is unnecessarily fanciful; our responsibility as film-
goers is to grasp what the narrative is, so as to reflect further on what it might
signify, rather than to create that narrative for ourselves. Furthermore, were
we really to create the narrative for ourselves, its significance would not, at any
rate, be that of the film we were putatively attempting to understand.
So I am going to assume, following Seymour Chatman, that if there is nar-
ration in a fiction film, if a comprehensible story is being conveyed to us, then
there is an agency or intelligence we are entitled, and in fact need, to imagine

Though he argues vigorously against the third form of disavowal in his attack on enonciation
theorists, there remains something of the first and second in the constructivism about film meaning
defended by David Bordwell; see his Narration in the Fiction Film (Madison: University of Wisconsin
Press, 1985) and Making Meaning (Cambridge, MA: Harvard University Press, 1989). For a critique
of this aspect of Bordwells otherwise eminently sane approach to film, see Berys Gaut, Making
Sense of Films: Neoformalism and Its Limits, Forum for Modern Language Studies 31 (1995): 823.
Bordwells rejection of narrative agents in film as such is also criticized by Seymour Chatman,
Coming to Terms (Ithaca: Cornell University Press, 1990), ch. 8.
See his Coming to Terms, especially chs. 5, 7, and 8. Another writer who seems to accept the
necessity of positing narrative agency in narrative film, though he verges on abstracting this to the
point of abandonment, is Edward Branigan, Point of View in the Cinema (The Hague: Mouton,
1984).
 Film Music and Narrative Agency 147

is responsible for this, i.e. doing the narrating. That is, a narrator, though not
necessarily an ordinary human being.
There are, of course, alternatives to this assumption. As noted above, there
are those who propose that films or filmic processes are themselves the per-
formers or executants of narration, there not being of necessity any narrator
within the films world, on the same plane as the events being displayed. But
in addition to the fundamental incoherence remarked above, this proposal, to
the extent it can be made out, is simply less interpretively useful than that of a
narrator, however minimally characterized, for every successful narration. My
response to yet another alternative, that in many cases of filmic narration, we
imagine we are presented directly with the events of the story, without ima-
gining there is any agent presenting them to us, is much the same: the postu-
late of narrative agency in cinema does a better job of accounting for how we,
admittedly largely implicitly, make sense of films as conveyors of stories.
For those who yet balk at this postulate, I would offer this. What I want
to say about assigning nondiegetic music to narrative agents as opposed to
implied filmmakers can, I believe, be translated so as to require instead only
the assumption of narrative processes or mere appearances-of-being-narrated.
So even if one does not regard the positing of internal narrators or presenters
in film as inevitable, the issue will still remain of whether soundtrack music
is to be thought of as an element in the narrative process or appearance of
narrative presentation, as opposed to an element in the construction of the
film by a filmmaker, standing outside both the story and its narration. It is
that issue I hope to illuminate here.

IV

I now review Chatmans brief for both cinematic narrators and cinematic
implied authors (that is, implied filmmakers). Chatman begins with an
appeal to ordinary language, one it is hard to gainsay:
It stands to reason that if shown stories are to be considered narratives, they must
be narrated . . . I would argue that every narrative is by definition narratedthat is,

A position taken, for example, by Gregory Currie, in Visual Fictions, Philosophical Quarterly
41 (1991): 12943. I respond to Currie in Seeing, Imaginarily, at the Movies, Philosophical
Quarterly 43 (1993): 708.
All text references in this section are to Coming to Terms.
148 Music

narratively presentedand that narration . . . entails an agent even when the agent
bears no signs of human personality. (115)
If narrative films are then necessarily narrated by a narrator, what kind of
narrator is this? Not, of course, the essentially linguistic narrator of a standard
literary fiction:
Film often has nothing like a narrative voice, no tell-er. Even the cinematic voice-
over narrator is usually at the service of a larger narrative agent, the cinematic show-
er. But that shower can reasonably be called a presenter . . . (113)
Films, in my view, are always presentedmostly and often exclusively shown, but
sometimes partially toldby a narrator or narrators. The overall agent that does
the showing I would call the cinematic narrator . . . The cinematic narrator is not be
identified with the voice-over narrator . . . . (1334)
Chatman also proposes that a cinematic narrator, operating mainly through
the affording of sights and sounds, is closely analogous to the mute presenter
of dialogue in a purely dialogic short story. So though in film a teller,
whose standard format is that of the voice-over, is usually absent or
secondary, a shower or better, because the term covers more comfortably
aural information, a presenter can be taken to be invariably in place, and
the primary agent of narration. The presenter in a film presents, or gives
perceptual access to, the storys sights and sounds; the presenter in a film is
thus, in part, a sort of perceptual enabler. Such perceptual enabling is what we
must implicitly posit to explain how it is we are, even imaginarily, perceiving
what we are perceiving of the story, in the manner and order in which we are
perceiving it. The notion of a presenter, whose main charge is the providing of
perceptual access on the fictional world, is simply the best default assumption
available for how we make sense of narrative fiction film.

This formulation of Chatmans is actually somewhat off the mark: its not the film that
is presented by the narrator, but various perceptual contents, various sights and sounds, i.e. what one
is enabled to see and hear, courtesy of the presumed powers of such a narrator. The film as such is
rather presented by the filmmaker or, interpretively, the implied filmmaker.
This is not to deny that it is sometimes in the purview of the cinematic narrator to present the
mental contents of some character, e.g. memories, fantasies, dreams, visualizations. But two points
about this should be noted. One, it may be unclear in such cases whether it is the cinematic narrator,
acting on the characters behalf, who shows the characters mentation, or rather the character, acting
as his own narrator, who is doing so. Two, the possibility of this sort of presenting requires a
background of presentings of perceptual reality at a more basic story level.
Problems of terminology loom here which a preemptive strike of clarification might dispel.
Of the three ideas, cinematic narrator, filmic presenter, and perceptual enabler, the first is perhaps the
broadest and the third the narrowest. Certainly there are actions of the cinematic narrator which
go beyond those of perceptual enabling or filmic presenting, most notably in this context, narrative
pointing through nondiegetic music. Whether there is a distinction worth making between filmic
 Film Music and Narrative Agency 149

While I thus accept Chatmans postulate of narrative agency wherever there

is narration, I do not endorse certain of his claims about the separation of nar-
rators from the story worlds they are narratively presenting. Chatman says,
for instance, that the [literary] narrator, by definition, does not see things in
the story world; only characters can do that, because only they occupy that
world.(120), and that the narrator cannot impinge on story space but must
stay within the bounds of discourse space (123). It is, however, incoherent to
postulate a narrator who offers us a window on or reportage concerning the
doings of a set of individuals the narrator takes and presents as real, and yet
insist the narrator is on a different plane, fictionally speaking, from those indi-
viduals, and in principle incapable of perceptual awareness of them. Chatman
is confusing a narrators fictional level, which must standardly be the same as
that of the other characters whose doings he/she/it is purporting to convey,
and the narrators degree of story involvement, which is variable, often rather
small, and in the limit, nil.
A narrator and the events narrated by the narrator must be on the same
fictional plane, otherwise cognitive relations posited between narrator and
events would not make sense. The cinematic narrators logical status vis-a-vis
the film world is to be distinguished from the narrators degree of involve-
mentcausal, emotional, experientialin the story, i.e. what literary theor-
ists mark as the narrators being either homodiegetic or heterodiegetic. Being
heterodiegetic, or an outsider to the events being related, does not remove
a filmic narrator ontologically from the characters he/she/it serves to offer us
perceptual access to. Chatman fails to see that the narrator must per-
force share the fictional plane of the characters, since they are apparently real
and reportable to that narrator, and this is true whether the narrator is homo-
diegetic, i.e. involved in the story events, or heterodiegetic, i.e. uninvolved

presenter and perceptual enabler is less clear; if so, the former would include the latter but comprise
in addition resources such as character voice-overs or mind-overs, affording access to the fictional
world in a wider-than-perceptual vein.
Except when the narrators relationship to the story being presented is clearly signaled, in the
novel or film, as one of relating a fiction as such, e.g. as through a disclaimer like this is only a
story, it never happened. But this is quite rare in fiction film; it is even rarer in literary fiction,
Thackerays Vanity Fair standing as a classic example, and John Fowless French Lieutenants Woman
as a recent, though more ambiguous, one. Kendall Walton, in Mimesis as Make-Believe (Cambridge,
MA: Harvard University Press, 1990), marks this distinction as one between reporting narrators
and storytelling narrators (36872). The overwhelming majority of narrators in narrative fiction are
reporting narrators, and as Walton points out, in such cases narrator and events narrated necessarily
belong to the same world.
As in the event of a wholly effaced and omniscient third-person narrator, the norm for
cinematic fictions.
150 Music

in them, standing to those events in merely a witnessing and transmitting

capacity.
I turn now to the notion of implied author in film. The need for this
concept is clear from the fact that films, like novels
present phenomena that cannot otherwise be accounted for, such as the discrepancies
between what the cinematic narrator presents and what the film as a whole implies . . .
unreliable narration presents the clearest but not the only case for the implied author
[in film] (1301)
. . . in cinema as in literature, the implied author is the agent intrinsic to the
story whose responsibility is the overall designincluding the decision to commu-
nicate it through one or more narrators. Cinematic narrators are transmitting agents
of narratives, not their creators. (132)
In short, for films as for novels, we would do well to distinguish between a presenter
of the story, the narrator (who is a component of the discourse), and the inventor of
both the story and the discourse (including the narrator): that is, the implied author
. . . (133)
So in film we must generally distinguish between, on the one hand, the nar-
rator or presenter of the story and, on the other hand, the ostensible inventor
of (all of) the narrator, the story narrated, the narrative structure, and the
cinematic entirety in which these are embodiedto wit, the implied film-
maker. The implied filmmaker is the agent who appears to have invented,
arranged, and integrated the various narrative agents and aspects of narra-
tion involved in the film, as well as everything else required to constitute the
film as a complete object of appreciation. The implied filmmaker, in short,
is the image we construct of the films makerbeliefs, aims, attitudes, val-
ues, and personalityon the basis of the film viewed in its full context of
creation.
A films narrator presents the events of the films world from within it,
whereas the implied author of a film, if he or she can be said to present any-
thing other than the film itself, presents the world of the film, at one doxastic
remove, from a position external to it. For the implied filmmaker, as for the
viewer, but in contrast to the films narrator, the films world is a fictional one,
acknowledged as fictional throughout. The implied filmmaker cant be in the
position of directly affording us, as with a silent gesture of behold!, the vis-
ion and audition of something that is only fictional with respect to himself,
namely, the characters and their circumstances; that remains the prerogative
of the films narrator or presenter, who is, in a fundamental sense, and pace
Chatman, one of them.
 Film Music and Narrative Agency 151

Before proceeding to my main concerns I will address some worries about

the general postulate of a filmic narrator formulated by George Wilson in his
penetrating study, Narration in Light. These worries are the most substantial
and explicit of which I am aware, so if they can be allayed, the ground for such
a postulate will be that much clearer.
A reason Wilson offers initially for being wary of a standing postulate of
cinematic narrator is that such a narrator is often conceived as an agent who is
necessarily observing events, the image track being thus identified as the visu-
al experience of that fictional observer. But this, as Wilson quickly notes, is
unnecessary. The essential function of such a narrator is to show us what is to
be seenor more broadly, to present to us what is to be seen and heardin
other words, to enable perception, albeit fictional perception, of those events.
The agent who shows, or permits us to see, need not be thought of as seeing
as well. This is apparent in Wilsons own useful sketch of what a cinematic
narrator would have to be: a fictional or fictionalized being, presupposed in
any viewing of the film narrative, who continuously provides to the audience,
from within the general framework of the fiction, the successive views that
open onto the action of the film.
The alternative, then, is to conceive the cinematic narrator as a kind of per-
ceptual pilot through the film world, rather than as an observer of it whom we
opportunistically inhabit.
Considered in this fashion, the narrator is a fictional figure who, at each moment of
the film, asserts the existence of certain fictional states of affairs by showing them to
the audience demonstratively; that is, by ostending them within and by means of the
boundaries of the screen. It is certainly part of our experience in film viewing that we
feel, usually subliminally, a constant guidance and outside direction of our percep-
tion toward the range of predetermined fictional facts which we are meant to see.
Having so well formulated this alternative, what problem does Wilson find
with it? Just this: that the entity described is analogous not to the narrator of
a novel, but rather to its implied author. His reason for this reluctant conclu-
sion seems to be that although the exact style and manner of this guidance
the fine-grained articulation of the processes of showingmanifests traits of

George Wilson, Narration in Light (Baltimore: Johns Hopkins University Press, 1986). See
also my review in Journal of Aesthetics and Art Criticism 47 (1989): 2902.
Narration in Light, 132. Ibid. 1334.
152 Music

sensibility, intelligence, and character, these traits are ones we will naturally
take to define the persona of the filmmaker as expressed through the film, or
equivalently, the personality of the films implied author, thus leaving no room
for a filmic narrator as such.
This seems to me too quick. First, Wilson gives no reason why such traits
should not be assigned, in some cases, to a filmic narrator, and in others, to
the implied filmmaker. Second, Wilson fails to consider that the assignment
of traits to the implied filmmaker might very well interpretively depend on
the assignment of traits, the same or different ones, to such a narrator, much
as our image of the implied author of a novel is necessarily based, in large part,
on our image of the narrator and how the narrator is managed or positioned
by the author.
But third, and most importantly, Wilson overlooks the fact that the implied
filmmaker just cannot occupy the role of perceptual guide to the films occur-
rences, and so, a fortiori, his particular mode of doing that cannot be what
cues us to some of his traits. And that is because the implied filmmaker cannot
logically be the presenter and ostender of events that are fictional with respect
to him. That is to say, if we imagine anyone giving us access to those events,
it cannot coherently be the filmmaker, in any guise. To be sure, the filmmaker
can present representations of those events, i.e. the shots or images the totality
of which constitute the film, but he cannot offer us the vision and audition
of those events themselves. If he be allowed a surrogate, however, a narrat-
ing agent presupposed by the process of narration, and fictionally on the same
level as its subject matter, then this difficulty disappears.
A fourth and related reason why the filmmaker cannot do duty for the
films narrator in this connection is this. Often we want not only to attrib-
ute traits of character, sensibility, or intellect to some agent connected with
the film, but more specifically, attitudes or views concerning the story that is
unfolding. But the filmmaker is not in the right cognitive position for this;
that is to say, he or she will not actually have attitudes or views towards the
fictional personages or occurrences involved in the story, knowing they are
merely fictional.
I suspect that Wilson is unable to find a place for, and thus underestim-
ates the rationale for, the cinematic narrator, because he subtly conflates the

Of course he can and does offer us the vision and audition of various events which took place
during the filming, namely, the enacting of various roles by various actorsbut that is another
matter.
See Walton, Mimesis as Make-Believe, 366.
 Film Music and Narrative Agency 153

devices or powers at the service of a director as crafter of a representation, and

those in the command of an imagined perceptual guide to the world such a
representation makes fictional; but these are different, and the ways they
function to control our experience as viewers differ too. The filmic narrator
allows us to perceive first this person, then that, at such and such an apparent
range, and for so long, clearly or not so clearly, etc., all of which manner of
showing may give us a certain impression of the showers attitudes or motiv-
ations; the filmmaker chooses or stages the profilmic events that are to be
filmed, decides on the camera distances and movements required for a shot,
determines the lighting and length of shots, orders those shots in a certain
fashion, etc., thus ultimately composing a narrative of a particular sort, with
a particular sort of implied narrator, all of which manner of making may give
us a certain impression of the makers personality or outlook. But the view
we form of the narrating agency or intelligence, from the way it carries out its
main charge as perceptual facilitator, need not coincide with the view we form
of the human maker of the film, from the way he or she fulfills the demands of
filmmaking.
Curiously, Wilson allows that we can indeed imagine a film where we
would have
grounds for a distinction between a voyeuristic filmic narrator and a satirizing im-
plied film maker. In such a case, there would be enough of a personification of the
manner in which the action is shown and enough of a contrast between the per-
sonification and what is implied about the filmmakers views of this to motivate the
identification of [the former] as a narrator.
But to my mind, the very possibility of this kind of divergence is predicated
on and presupposes the logical distinctness of the roles of filmic narrator and

Nor are supporters of the notion of a narrator internal to film immune to this confusion.
Consider the formulation of Nick Browne, quoted by both Branigan and Chatman: the authority
which can be taken to rationalize the presentation of shots (The Rhetoric of Filmic Narration (Ann
Arbor: UMI Research Press, 1982), 1). But since shots are constructional elements of films as made
objects, that authority can only be the implied filmmaker; the authority, or agency, that Browne is
really after is that which appears to rationalize the presentation of views, or sights and sounds.
There are other sources of Wilsons reluctance to embrace filmic narrators on a standing basis.
He suggests at one point that a filmic narrator distinct from the implied filmmaker would, by
analogy to literature, have to be a character that the text depicts directly or indirectly (Narration
in Light, 136). But this is only half right; the narrator is indeed a kind of fictional character, but
not one that need be depicted, as I understand that term. If purely dialogic short stories, or purely
epistolary novels, have narrators, then such narrators are not depicted, either directly or indirectly.
So the fact that cinematic narrators are standardly not depicted, i.e. nothing in films shows or
announces them as such, is not a principled impediment to acknowledging them.
Ibid. 1367.
154 Music

implied filmmaker, even if in most films, unlike the one Wilson conjures
up, the personal traits ascribable to the occupants of the two roles tend to
coincide.
As we have seen, Wilson questions whether every standard film narration
must be understood to entail an implicit narrator distinct from the filmmaker.
He maintains that in films governed by the classical paradigm of transparency,
which covers almost all narrative film, we simply see the fictional events for
ourselves, defeasibly taking the facts about them to be what we see them to be.
But to my mind this sidesteps the question of how it is we are seeing what we
are seeing, however reliable or unreliable it turns out to be. If this question,
however, is not simply set aside then the only satisfying answer to it is that we
are being shown such and such, by some agent, in some perhaps unspecifiable
manner. That is to say, the posit, however unvoiced, of an agency that is offer-
ing us sights and sightsan agency with certain powers, motivations, and
limitationsseems inescapable if we are to justify our taking anything to be
fictional in the film world, on the basis of the moving images that are the only
thing we are literally confronted with. It is not enough to just say that, with
fiction film, the films world is made visible to us, perhaps adding that there is
a convention to that effect. Reasonalbeit reason operating in service of the
imaginative understanding of fictiondemands an answer to how it is that a
world is being made visible to us, and that demand, it appears, is only satisfied
by the assumption of an agency responsible for that.
One might still seek to avoid this conclusion by adopting the following
stance: it is, indeed, as if we are being shown such and such, from a giv-
en perspective, by an agent within the films world, with certain powers to
make views of that world available to us, but we need not assume that there
is such an agent. Here, though, we arrive at a distinction virtually without a
difference. If it seems to us, at some level, as if we are being shown such and

My discussion so far may give the impression that I regard the role of the implied presenter
of a narrative film as confined to that of providing views on events understood as fully constituted
independently of the showers activity. But while I think that that is indeed the dominant role
of a films presenter, it need not be the exclusive one. The cinematic narrator might in part be
thought of as a fashioner or shaper of events that are only then presented, more straightforwardly,
in certain ways (e.g. in certain lights, or in a certain order). This fashioning or shaping of fictional
events thought of as existing, on a more basic level, prior to narrative attention, can be seen as a
more subtle way of presenting events taken to belong to the underlying event structure with which
narration is concerned. If so, then the narrative structuring of films is a two-stage affair, and that
which is effected through camerawork and editing is subsequent to that understood to be achieved
in the staging of action and the manipulation of setting. (For more on this dimension of cinematic
narration, and the rationale of its recognition, see Bordwell, Narration in the Fiction Film.)
 Film Music and Narrative Agency 155

such, are we not in effect imagining that we are being shown such and such,
and thus, finally, that there is, on the imaginative plane, something doing the
showing? So I would claim.

VI

That nondiegetic music standardly serves to advance a films narrative is some-

thing on which theorists of film appear to agree:
Narrative is not constructed by visual means alone. By this I mean that music works
as part of the process that transmits narrative information to the spectator . . .
Voice-over is just one of many elements, including musical scoring, sound effects,
editing, lighting, and so on, through which the cinematic text is narrated.
The moment we recognize to what degree film music shapes our perception of a
narrative, we can no longer consider it incidental . . .
Another point widely agreed upon is that even if the primary purpose of
nondiegetic film music is the advancing of the narrative, there may very well
be others. Here is a typical admonition concerning film musics multiplicity
of ends:
There is not one and only one function that music can perform in relation to movies.
Aaron Copland suggested five broad functions: creating atmosphere, underlining the
psychological states of characters, providing background filler, building a sense of
continuity, sustaining tension and then rounding it off with a sense of closure. These
do not seem to be necessarily exclusive categories, nor do they exhaust the range of
functions that music can perform in movies.
Not surprisingly, I am happy to join this double consensus: film music
often serves narrative in some way, but there is a range of other functions that
such music sometimes performs. What I am concerned to demonstrate, how-
ever, goes beyond those two pieces of received wisdom. It is that the most
fundamental division in the realm of film music concerns the viewers assign-
ment of responsibility for such music, i.e. the agency the viewer posits, usually
implicitly, as responsible for the music being heard. It will turn out that there
is a rough coincidence between film music to which we intuitively accord

Kathryn Kalinak, Settling the Score: Music and the Classical Hollywood Film (Madison:
University of Wisconsin Press, 1992), 30.
Sarah Kozloff, Invisible Storytellers: Voice-over in American Fiction Film (Berkeley: University
of California Press, 1988), 434.
Gorbman, Unheard Melodies, 11.
Noel Carroll, Mystifying Movies (New York: Columbia University Press, 1988), 216.
156 Music

narrative significance and film music for which we implicitly hold an internal
cinematic narrator accountable, and between film music to which we do not
accord narrative significance and film music that we implicitly assign directly
to the implied filmmaker.
When, though, can film music be said to have narrative significance? When
does nondiegetic music function narratively? In order to answer this question
we must have a plausible criterion of narrativity, or of actions within the pur-
view of a narrator. In trying to arrive at one, it will be helpful to have before
us a survey of the various functions that critics or theorists have observed film
music to perform.
These functions include: (1) the indicating or revealing of something about
a characters psychological condition, including emotional states, personality
traits, or specific cognitions, as when the music informs you that the heroine
is happy, or that the hero has just realized who the murderer was; (2) the
modifying or qualifying of some psychological attribution to a character inde-
pendently grounded by other elements of the film, as when the music tells you
that a characters grief over a loss is intense; (3) the underlining or corroborat-
ing of some psychological attribution to a character independently grounded
by other elements of the film, as when music emphasizes something about a
situation on screen which is already fully evident; (4) the signifying of some
fact or state of affairs in the film world other than the psychological condi-
tion of some character, e.g. that a certain evil deed has occurred, offscreen;
(5) The foreshadowing of a dramatic development in a situation being depic-
ted on screen; (6) the projecting of a story-appropriate mood, attributable to
a scene as a whole; (7) the imparting to the viewer of a sense that the happen-
ings in the film are more important than those of ordinary life, the emotions
magnified, the stakes higher, the significances deeper; (8) the suggesting to the
viewer of how the presenter of the story regards or feels about some aspect of
the story, e.g. sympathetically; (9) the suggesting to the viewer of how he or
she is to regard or feel about some aspect of the story, e.g. compassionately;
(10) the imparting of certain formal properties, such as coherence, cogency,
continuity, closure, to the film or parts thereof; (11) the direct inducing in
viewers of tension, fear, wariness, relaxation, cheerfulness, or other similar
cognitive or affective state; (12) the lulling or mesmerizing of the viewer, so as
to facilitate emotional involvement in the fictional world to which the viewer
would otherwise prove resistant; (13) the distracting of the viewers attention
from the technical features of the film as a constructed artifact, concern with
 Film Music and Narrative Agency 157

which would prevent immersion in the filmic narrative; (14) the expressing
by the filmmaker of an attitude toward, or view on, the fictional story or
aspect thereof; (15) the embellishing or enriching of the film as an object of
appreciation.
Without deciding, for each of these functions, which are properly con-
sidered narrative and which not, it would appear that some unequivocally are,
and some unequivocally are not. What I will do at this point is explore a num-
ber of suggestions as to what the criterion of narrativity might be in regard to
nondiegetic film music, assessing them against the background of this array
of observed functions, some of which, at any rate, would have to come out
counting as narrative, some clearly not, and some having a status that might
only be settled, clarifyingly, once a given suggestion were adopted.
One possible criterion is this: (C1) does the music seem to issue from, be in
service of, the agency one imagines to be bringing one the sights and sounds
of the films world? If so, then it can be reckoned part of the narration proper,
and assignable to the cinematic narrator. Perhaps an equivalent formulation
would be: (C2) does the intelligence one thinks of as bringing one the music
seem to be the same as that charged with conveying the storyas opposed to
that charged with constructing the film? If so, then the music can be reckoned
part of the narration, and assigned to the cinematic narrator.
Though I think these criteria point in the right direction, there is an evident
problem with them, insofar as we hope to look to them for guidance, espe-
cially in difficult cases. And that is that they are uncomfortably close to what
they purport to analyze or elucidate, namely, whether a use of nondiegetic
music is narrative or not. So if we are unsure whether a given cue is function-
ing narratively, we are likely to be almost equally unsure whether it feels as if it
derives from the films narrative agent. Thus, it would seem desirable to have
some other mark, could we discover one, whose conceptual distinctness from
the idea of narrative functioning was greater than that of C1 or C2.
Such a mark might be that of making a difference in the narrative. Instead
of appealing directly to an intuition of a connection of the music to a films
internal narrator, we can appeal instead to the notion of making-fictional, or
generating fictional truths, in a film. A criterion of nondiegetic music having

This would apply, note, even when such music is unforegrounded: if it appears to respond to
the demands of storytelling, broadly understood, then it can be construed as something like musical
musing, sotto voce, on the cinematic narrators part.
158 Music

a narrative function, and thus being attributable to a narrative agent, could

be this: (C3) the music makes something fictionally truetrue in the story
being conveyedthat would not otherwise be true, or not to the same degree
or with the same definiteness. A counterfactual form of the suggestion is
perhaps more transparent: (C4) would deleting the music in a scene change
its represented content, i.e. what is fictional in it, or only how the scene affects
viewers? If the former, then the music is an aspect of narration; if the latter,
then not.
We must briefly discuss what it means to make something fictional in a work
of fiction such as a narrative film. Something is fictional in a film, according to
a well-developed recent account, if it is to be imagined to be the case by viewers
concerned to experience the film properly. What thus makes somethinga
proposition about the films worldsomething that is to be imagined in the
course of viewing is perceivable features of the film, a public object, taken as
a prop for guided imaginings.
When we make-believe in accord both with the features of artistic props
and the usually tacitly grasped principles for imagining that are in effect in a
given artform, we are engaged in tracing out imaginary worlds, ones in which
things are make-believedly, or fictionally, so. The fictional world of a repres-
entational art work, unlike that of a daydream or fantasy, is as it is because
features of the associated proptext, canvas, filmproperly construed, are
the way they are; not all is up to the imaginer. Props, through their exist-
ence and nature, generate fictional truths independently of what individual
perceivers might choose to imagine.
What does it mean for a proposition to be fictional, or true in a fictional
world, in respect of a given work of art? Simply that there is a prescription to
imagine it, a prescription encoded in the particulars of the artifact that serves
as a prop for making-believe, and whose force derives from underlying con-
ventions of construing works of the sort in question. Being fictional thus has
an ineliminable normative dimension: it is what is to be imagined in a given
context, rather than merely what may be imagined.
For example, in Citizen Kane, Orson Welless image on screen being that
of a large man makes it fictional that Charles Foster Kane is a large man;
the opening shotsa series of lap dissolveshaving a certain visual content

See Walton, Mimesis as Make-Believe. For an entree into this important work, see my critical
notice, Making Believe, Dialogue 32 (1993): 35974 (reprinted in The Pleasures of Aesthetics).
 Film Music and Narrative Agency 159

makes it fictional that at the beginning of the story one is shown Kanes estate,
Xanadu, from a distance and shrouded in mist, and then at progressively
closer range; Ray Collinss voice saying certain things on the soundtrack in the
scene at Susan Alexanders apartment makes it fictional that Collinss char-
acter, Jim Gettys, has threatened Kane; the way the shot of Kane expiring is
sequenced in relation to others which are understood as a flashback to Kanes
childhood, makes it fictional that Kanes dying word, rosebud, refers to his
beloved old sled, etc. Of course, much of this generation of fictional truths
will be indirect, dependent on various conventions of the medium in effect
and on other things taken provisionally as fictional, and accordingly, much
of our knowledge of such fictional truths will be inferential. And sometimes,
what is made fictional by a films narration is orthogonal to, or even the oppo-
site of, what first appears to be the case, that is, what it initially seems we are
to imagine is the case; unreliable, uninformed, or unforthcoming narrators,
though not as common in film as in literature, are still a significant possibility.
Applying this suggestion to the issue of narrativity in film music, then, the
question becomes, of a given cue, whether it generates, contributes to gener-
ating, or at a minimum, more firmly grounds, a fictional truth in the scene
which it accompanies. Thus, film music which, when interpreted in the light
of prevailing conventions of the medium and the surrounding narrative con-
text, indicated that a character was afraid or was remembering a past incident,
or that a man had been executed or an agreement reached, or that a situation
was fraught with danger or else full of hope, where these things would not
be established, or not so definitely, without the music, would clearly count as
narrative.
We should note that nondiegetic music may, indeed, generate fictional
truths even if only attended to with half a mind, or not consciously remarked
at all while present. It will do this by causing a viewer, say, to perceive a scene
as fraught with danger, even if the viewer is not aware of what is making her
have that perception. Nevertheless, if such an imaginative perception is reli-
ably produced in attuned viewers, and not undermined by subsequent aspects
of the narration, then it may well be fictional that the scene is fraught with
danger, even though the rest of the narrative indicators are insufficient to

Film music . . . often contributes subtly but effectively to the generation of fictional
truthshelping to establish, for example, that fictionally a character is nervous or cocky or
ecstatic (Mimesis as Make-Believe, 172).
160 Music

establish that and the viewer never realizes that it is the background music
that in fact makes it so.

VII

It is time to look at a range of illustrative examples of film music. I begin

with examples whose narrative functioning is obvious, and which conform,
expectedly, to the making-fictional criterion proposed above. I then explore
another range of examples, ones that exhibit a different sort of narrativity,
and show how, on a more encompassing construal of the making-fictional
criterion, these can be accommodated as well. Eventually, though, I turn to
films containing nondiegetic music that is not, by that criterion or any other,
reasonably construed as narrative. The music in such films instead serves other
sorts of artistic function, ones attributable directly, I will argue, to implied
filmmakers.
One of the least ambiguous narrative uses of soundtrack music in main-
stream film occurs in Steven Spielbergs 1975 blockbuster, Jaws. I have in
mind the shark motto devised by the composer, John Williams. This consists
of an ostinato alternation of low staccato notes at the interval of a seconda
kind of aural sawing. The motto has an unarguable informational mission,
namely, the signaling of the presence of the shark. It is true that there is anoth-
er, visual, indicator of the sharks presence when unseen, namely, shots from
an offshore point of view, at the water line or slightly below it. But that indic-
ator is not invariant in meaning, since it is sometimes employed when there is
no shark about. The musical shark motto is the only reliable signifier of the
shark, and so has an ineliminable fact-conveying function. Correspondingly,
it is clear that it is the presence of that motto on the soundtrack at a given
point that makes it fictional that the shark, though as yet unseen, is in the
vicinity of what is shown.
David Raksins haunting score for Otto Premingers Laura provides some
further instances of straightforward narrative use of film music. The Laura
theme, first encountered diegetically on a record player in the apartment of
the ostensibly (but not actually) murdered heroine, pervades critic Waldos
(Clifton Webb) represented recollections of the early days of his relation-
ship with Laura (Gene Tierney), and unmistakably signifies his delight in
her companionship. Subsequently we are treated to apprehensive versions of
the Laura theme as detective McPherson (Dana Andrews), alone in Lauras
 Film Music and Narrative Agency 161

apartment, studies the portrait of Laura over the fireplace; this cue then cli-
maxes unsettlingly, revealing or underlining McPhersons frustration with his
investigation at this point. The most striking cue, one much noted in the film
music literature, is a weird version of the theme produced by playing it on
a piano but only recording the overtones of each note struck. This is heard
as McPherson views Lauras portrait on a second occasion, before then drink-
ing too much and falling asleep, and suggests the ghostly influence Laura is
beginning to exert over his poor detectives mind. In each of the foregoing
cases, the music is plausibly viewed as making, or contributing to making,
something fictional in the story: that Waldo delighted in Laura inordinately,
that McPherson is (earlier) almost terminally frustrated with Lauras case, that
McPherson is (later) succumbing to bewitchment by Lauras spirit.
Another film rich in narrative pointing of a theoretically unproblematic
sort is Martin Scorseses Taxi Driver. Regarding a scene in which Travis
(Robert De Niro), the films semi-psychotic protagonist, is induced to move
his cab away from the Manhattan workplace of a girl he is infatuated with
and back into the grime and disorder of the city, one writer affirms that
the music . . . here reveals that Traviss thoughts are not with the street but
with Betsy. And of the bluesy, sensual saxophone tune itself, which stands
for Betsy (Cybil Sheppard) in Traviss mind, the same writer has this to
say: Traviss vision of idealized womanhood, the music implies, is strongly
erotic. Thus, Bernard Herrmanns music does not serve merely to inform
us about Traviss mental life, or to second redundantly what other elements of
the film establish about his mentality, but rather enters into making it fictional
in the film that Traviss mental life is a certain way at a certain time.
Commenting on the blade-game fight scene in Nicholas Rays Rebel
Without a Cause, Noel Carroll offers the following: The uneasy, unstable
quality of the music [by Leonard Rosenman] serves to characterize the
psychological turmoilthe play of repression and explosive releasewith
which the scene, and the movie, is concerned. If Carroll is right, the
music of this scene, which intuitively seems an aspect of its narration, serves
to underwrite as desired a fictional truth about the specific, highly volatile,
character of the turmoil afflicting the young protagonists. Another instructive
example from Rebel Without a Cause occurs later in the film, and consists of

See Kalinaks informative discussion in Settling the Score, 178.

Graham Bruce, Bernard Herrmann: Film Music and Narrative (Ann Arbor: UMI Research
Press, 1985), 68.
Mystifying Movies, 217.
162 Music

a montage of two-way phone calls among various adults concerned with the
whereabouts of the three main youngsters. This montage is covered by tense
nondiegetic music, displacing the dialogue that would ordinarily be heard, the
music thus signifying that the conversations, whatever their specific contents,
are anxious ones.
The opening of Elia Kazans On the Waterfront affords another illuminat-
ing example. An establishing shot of city docks, ocean liner in the distance,
gives way to a street scene in which longshoreman Terry (Marlon Brando)
becomes the focus of attention. Leonard Bernsteins jazz-inflected score at this
point involves a persistent drum tattoo overlaid with saxophone insinuations.
Terry, in the darkening street, yells up to friend Joeys window, persuading
him to go to the roof to recover one of his pet pigeons, where unbeknownst
to Joey, two men are waiting for him. After Terry releases the pigeon he has
been holding, and promises to join Joey in a moment, the score becomes
loud, aggressive, and insistent, its rhythms more syncopated. The music tele-
graphs us that something bad is in store, that the men glimpsed on the roof
are trouble; the music can be said to prefigure Joeys fall, pushed off the roof
by thugs of the corrupt union boss, though without defining precisely what
is about to happen. The cue is clearly narrative, and just as clearly, makes it
fictional that Joey is in danger, even before he leaves the window for his fatal
visit to the roof.
Later on, after the boss tells right-hand man Charley to straighten out his
brother Terry or else, Charley leaves union headquarters to do something, we
know not what. Bernsteins music at this point is very dramatic and tense: a
series of rising notes in the brass, leading to a rhythmic explosion, the whole
heard twice. The cue arguably conveys Charleys complex state of mind, faced
with the necessity of keeping his errant brother, who is threatening to do
the right thing, in line: a mixture of anger, shame, and angst. If it does not
singlehandedly make it fictional that that is Charleys state of mind, the cue
contributes ineliminably to making it so. A dissolve leads directly to the fam-
ous conversation between the brothers in the rear of a taxi.
Consider, lastly, the final sequence in Fellinis La Strada. Having five years
ago abandoned his erstwhile assistant, the childlike Gelsomina, after she be-
comes too withdrawn and depressed to work, Zampano the strongman
discovers, by accident, what became of her. That evening he does his act per-
functorily, gets drunk, starts brawling, then goes down to the beach, which
reminds us of where he first acquired Gelsomina from her impoverished fam-
ily. He walks into the water, goes back out, looks up at the sky apprehensively,
 Film Music and Narrative Agency 163

then starts to bawl and grasp at the sand, on which he has flung himself in
despair. At this point the La Strada theme on the soundtrack removes all
doubt as to what it is Zampano is bemoaningnamely, the loss of Gelsom-
ina and her innocent love.

VIII

Clearly, making something fictional in a film is a sufficient condition of music-

al narrativity. Is it, however, a necessary one? Though providing the basic
fictional truths of a story may be the central activity of a narrator, there are
others that are almost equally paradigmatic of narration. One is the evincing
of attitudes or feelings on the narrators part toward the story presented, in
virtue of how the story is presented; another is the inviting of the viewer to
adopt certain attitudes or feelings toward the story presented. In other words,
in addition to giving access, in a particular manner, to the fictional states of
affairs that constitute a story, a narrator generally manifests attitudes regard-
ing the states of affairs to which access is afforded, and thereby suggests to the
narratee attitudes to be adopted. In literature, for example, the narrator stand-
ardly tells us what happened, after his or her fashion, reveals, knowingly or
unknowingly, his or her view of these happenings, and also suggests, explicitly
or implicitly, how we should view what we are told happened.
Now it seems plain that such narrational effects are often achieved by
appropriate nondiegetic music: the music tells you how the presenter of the
story regards the events being presented, or else how he would like you to
regard them. But on the surface, this does not appear to be a matter of estab-
lishing, nuancing, or even confirming a fictional state of affairs in the story.
So in light of that, can making-fictional be sustained as the effective mark of
musical narrativity?
I believe so. We need to make a distinction between what is fictional in a
films story and what is fictional in the world of a film. The latter is a broader
notion than the former. What is fictional in the films world comprises, in
addition to the facts of the story, the facts of its narration by the special, often
almost effaced, fictional agent known as the narrator. All that is still within
the sphere of the fictional, that is, of propositions to be imagined by a viewer

In some literary fictions, for example, Hemingways The Killers or Robbe-Grillets La Jalousie,
this latter function may seem to have lapsed. But I would argue that even in such fictions there
are attitudes the narrator implicitly invites the reader to adopt, precisely in virtue of so pointedly
eschewing normal commentary.
164 Music

in comprehending the film. The films story consists of what is fictional about
the characters who figure in the action; the films world includes, as well, what
is fictional about the narrator, in relation to either the story narrated or the
implied audience of that narration.
Returning to film music, a plausible construal of some nondiegetic cue will
often have the implication, not that it makes something fictional in the story,
but that it makes it fictional either that the cinematic narrator has a certain
attitude or feeling toward some event being presented, or that the narrator
encourages viewers to have such an attitude or feeling toward it. In either
case, musical narrativity will still correlate with musics making something
fictional, only here it is a making-fictional in the films world, as opposed to
a making-fictional in the embedded story. Some examples will serve to cla-
rify this more encompassing interpretation of musical narration in terms of
making-fictional.
Music functions narratively, by any intuitive assessment, in Hitchcocks
Shadow of a Doubt, particularly at junctures when a scrap of Lehars Merry
Widow waltz intrudes itself, suggesting the Merry Widow Murders that are
central to the plot. Several characters are heard singing or humming the tune
in the course of the film, these occurrences being of course diegetic, but the
tune is heard, in an altered form, as early as Dmitri Tiomkins title music,
which accompanies a stylized shot of waltzing couples. Two notable nondie-
getic occurrences after that are these. First, a few bars of the waltz theme in
the cue that accompanies the familys greeting of Uncle Charlie at the train
station, as they walk off to their car to take him home: a tracking shot of
the group, heading toward the camera, is eventually reframed so that only
Uncle Charlie is in view, and that is when the scrap of tune is heard. Second,
a more prominent statement of the theme when Uncle Charlie gives young
Charlie an emerald ring, and she notices it is already engraved inside with an
unknown someones initials. In both cases, the music arguably serves to com-
municate something to the viewer about Uncle Charlies identity, connecting
him in some as yet unexplained way to the waltzing image presented at the
beginning.
But does the music make, or even contribute to making, something true
in the films story as such, something that would not otherwise be the case?
It is not clear that it does. To consider just the most obvious candidates,

Of course, when a narrator in a film is also a character in the action, as with a homodiegetic
voice-over narrator, then certain facts about such a narrator are also facts of the story.
 Film Music and Narrative Agency 165

neither cue makes it trueeven viewed in retrospect, when a connection to

Lehars tune is understood to import as well a connection to the Merry Wid-
ow murdersthat Uncle Charlie is the murderer, nor does the second make
it true, say, that young Charlie suspects that he is. The reason is that those fic-
tional truths are firmly established, and independently, by other elements in
the film.
What, then, might they be doing? I suggest that the first cue makes it fic-
tional that the narrator is obliquely hinting to viewers with regard to Uncle
Charlies identity, and the second makes it fictional that the narrator is, even
more directly, connecting Uncle Charlie to something sinister in his past,
though at that point viewers have no notion of what it might be. The second
cue may, in addition, function as the narrators proposing of a deep psychic
link between Uncle Charlie and young Charlie, one that her subsequent mor-
al corner-cutting, in dealing with an uncle she then knows to be an unhinged
killer, partially bears out. In any event, the status of these cues as narrative
can be recovered in the guise of what is made fictional, not in the story as such,
but in the narrators attitudes or actions with respect to viewers.
But what of the curious musical image of waltzing couples first encountered
in the title sequence, which recurs nondiegetically and unchangingly at three
crucial points in the story? In each case the image is superimposed over the
action already on view, which continues underneath. The first occurrence is
after the interaction between Uncle Charlie and young Charlie over the emer-
ald ring, as young Charlie goes off to clear the supper dishes, leaving only
Uncle Charlie on screen. The second occurrence is at night in the town lib-
rary, when young Charlie, after reading the newspaper account of the Merry
Widow murders, gets up, almost reeling, as the camera tracks upward and
away from her. The third and last is just as Uncle Charlie falls to his death
beneath the wheels of a hurtling locomotive.

For example, that Uncle Charlie is the murderer is underwritten by his unexplained money in
the opening hotel room scene, by his evident concern to keep an item in the daily newspaper unread,
by his unreasonable aversion to being photographed, by his maniacal utterance at the dinner table
about fat, wheezing, useless widows, by the already inscribed ring itself, etc. That young Charlie
suspects him does not become true until she is informed about the manhunt by one of the two
detectives who have been trailing Uncle Charliethough of course there have been signs, intended
for and readable by the viewer, well before that.
Their psychic kinship is adumbrated earlier in the film, in the parallelism of our first views
of them both, reclining on beds with their hands behind their heads, in the worried, almost
cynical remarks about family values that young Charlie makes when we first hear her speak, and
in the coincidence of young Charlie deciding to send her uncle a telegram just hours after he has,
unbeknownst to her, sent one in her direction.
166 Music

The first and second of these might be interpreted as the narrators display
of the mental contents of the character then in frame, in the one case signify-
ing Uncle Charlies meditation on his hidden identity, in the other, his nieces
realization of that identity. But in addition to being implausible because it
does not reflect the very different emotional tones with which uncle and niece
would have contemplated this identity, this sort of interpretation seems
unavailable for the last occurrence, where ascription to the terrified and soon-
to-be-obliterated Uncle Charlie of a contemplative thought about his past
strains credulity to the breaking point.
This suggests that the recurrent waltzing image should be construed as a
form of narrators commentary: it is employed by the cinematic storyteller at
crucial moments to underline in an intentionally jarring mannerbecause
achieved through the elegance and innocence of a waltzUncle Charlies
horrific identity. Thus, what is made fictional by these musical cues is not that
Uncle Charlie is the murderer, but that the narrator is adverting to that fact,
almost sardonically, both before and after it is narratively established.
Nicholas Rays Rebel Without a Cause provides another example whose ana-
lysis helps us to see our way here. The opening scene unfolds at a police
station, where three juveniles whose lives will soon importantly intersect find
themselves separately in trouble. At one point Jim Stark (James Dean), who
has been talking with a sympathetic counselor, bangs and kicks a desk in
frustration, at the counselors explicit invitation. As his outburst concludes,
dissonant music surges up briefly on the soundtrack. This undoubtedly adds
tension to the scene, but does it contribute to defining the fictional world in
any way? That Jim is wildly and angrily frustrated is fully established by what
the perceptual enabler of the film has allowed us to see and hear of his out-
burst. What, then, is the music, which certainly seems to have narrative force,
doing there in narrative terms?
Perhaps this: it serves to get across the phenomenology of Jims feelings,
giving viewers access to the quality of his outburst from the inside, supple-
menting the access afforded from the outside by the ordinary perceptual data
of the scene. Suppose that is so. Then on the one hand, this could be con-
strued as a subtle sort of making-fictional in the story, namely, making it
fictional that the quality of feeling in Jims outburst was precisely such and
suchthe quality the musical cue in question is expressive of. On the other
hand, this could equally well be construed as a making-fictional concern-
ing not Jim, whose emotional condition is perhaps overdetermined by other
indicators in the scene, but instead the narrators stance toward the audience.
 Film Music and Narrative Agency 167

That is, perhaps the cues cash value is that the narrator is inviting viewers
to share in rather than merely observe what Jim was feeling, and as a con-
sequence, encouraging viewers to adopt a sympathetic attitude to him. The
cues narrativity, in other words, may be a matter of its definition of the fic-
tional world of the film, comprising both narrator and story narrated, rather
than that of the story per se.
Consider now the common use of background music to create atmosphere
in a scene, but without attributing mental states to any character therein. Is
there anything that can thus be said to be made fictional in the film world?
In cases where an appropriate atmosphere is created, i.e. one that seems con-
sonant with the way the story is otherwise told, what is made fictional might
be that the narrator wants the viewer to assume a particular mood or frame
of mind as certain events are presented for perception. In cases, though, where
the atmosphere created does not gibe with the style or tone of narration
already established, then even indirect fictional generation of that sort may
be absent. The musical creation of mood may then have to be understood
not as a narrative action, but rather as one which aims to affect the viewer
immediately in a way that has no fictional upshot. Where nondiegetic music
adds atmosphere to a scene without plausibly making anything fictional in the
films world, simply producing a mood in viewers, it seems that responsibility
for it, as for other nonnarrative, purely compositional elements of a film, must
rest directly with the implied filmmaker.
Exploring the interpretive option just broachedof assigning musical cues
to the implied filmmaker rather than the films narrative agentwill be the
focus of the remainder of this essay. But before turning to that I conclude
this section with a brief look at narrative uses of nondiegetic music in Hitch-
cocks Vertigo. Vertigo boasts perhaps the greatest of classical film scores, and
its greatness as a film is due, in no small measure, to that score and its mas-
terful integration into the film in almost every respect. The intrinsic interest
and sophistication of Bernard Herrmanns score has been much discussed,
but what is most striking about it in the context of the film is how significant a
burden it bears for limning the mental states and traits of characters, by com-
parison with most other films. Vertigo abounds in occasions where not only
are viewers fictionally informed about the inner lives of the characters through
soundtrack music, but the music is what in large part makes it fictional that
their inner lives are to be so characterized.
When Scottie (James Stewart) first sees Madeleine/Judy (Kim Novak) at
the rear of a restaurant in San Francisco, the music serves significantly to
168 Music

characterize her for us and for him: if the camera movement toward
Madeleine lets us experience the physical nature of Scotties immediate attrac-
tion to her, it is the music that most fully conveys the sensual mystery of
the woman. This scene is instructive in other ways as well. Madeleine gets
up to leave, comes toward Scottie, pauses momentarily, and is very notice-
ably framed and lit in profileshown, in effect, to best advantage. But who
is doing that? The cinematic narrator, in order to indicate something about
Madeleine and the overwhelming psychic effect she has on Scottie on first
encounter. The filmmaker, Hitchcock, cannot do thatthough he can do
certain parallel things to Kim Novak and the set in order to bring about,
on a fictional plane, the narrative result. The cinematic narrator is the one
who, fictionally, showcases Madeleine, for our benefit as trackers of the story,
and then underscores this showcasing through the musical resources under its
control, e.g. by crescendoing at the point of held close-up.
After the crisis of the first part of the film, Scottie spends some time in
a sanitarium, sunk deep in depression and aimless longing. Soon after his
release, we are given a high pan over the front of Madeleines apartment build-
ing, as the love motifa four-note Tristan-like descending figureis soun-
ded romantically by French horns. This foreshadows Scotties appearance in
frame at the end of the camera movement, with Madeleine obviously in mind:
he approaches a blonde woman in front of the building, about to get into
what was Madeleines car, only to discover that it isnt her. The exact content
of his hope and then disappointment is supplied by the musical cue.
Scotties vertigo first occurs in the films opening scene, while he is hanging
from a rain gutter, high above the city, having slipped in the course of pur-
suing a fleeing felon. This is importantly recalled in the plots pivotal event,
occurring halfway through the film, which takes place at the Mission of San
Juan Battista, from whose tower the real Madeleine, unwanted wife of Gav-
in Elster, will appear to have leapt to her death. As Madeleine rushes into the
church, and Scottie begins to follow, Herrmanns music foretells the recur-
rence of Scotties vertigo: milder variants of the clash of tonalities which were
heard in the [opening] rooftop sequence hint at the probable effect climb-
ing the tower will have upon Scottie. The musical cue, it seems, generates

Bruce, Bernard Herrmann, 143.

This scene illustrates nicely a narrative possibility mentioned above (see n. 21), whereby a
cinematic narrator might be thought of as presenting story events, conceived of as already existing
fictionally at a basic level, in a certain way, through a partial shaping of the event being viewed.
Ibid. 173.
 Film Music and Narrative Agency 169

the fictional truth, at the point it sounds, that Scottie is going to experience
vertigo when he climbs, though he is not experiencing it now. In other words,
that Scotties vertigo is coming becomes something that is to be imagined by
viewers at that point in the film. Alternatively, perhaps the truth is gener-
ated that Scottie knows it is coming, or is concerned that it might. In the
films final scene, also set in this tower, the tremolo trills which are prominent
during this, Scotties second ascent, suddenly cease, suggesting he has at that
point overcome his vertigo and will be able to complete his trip to the top.
At the start of the letter scene, the moto perpetuo string figures prominent
in the opening rooftop scene recur, in an overwrought vein, accompanying
Judys detailed recollection of the tower incident and her role in the
deception perpetrated there. This underscores sonically how emotively
charged the incident remains for her, and helps us understand why she is
ultimately unable to carry through the writing of the letter of confession.
In the famous nightmare sequence, the habanera music associated with
Carlottaa dead woman with a tragic past with whom Madeleine appears
to identifybecomes more discordant, almost parodic, through the addition
of stereotypical castanets and tambourine, conveying unmistakably the
intensity of Scotties oppression by Carlotta/Madeleine. But more specific
psychological pointings yet have been laid at the door of the scoring in this
film, with some plausibility. According to one writer, the rather banal music
that accompanies a walk taken by Scottie and Judy in the park adjacent to the
Palace of Fine Arts, soon after he meets her and senses a kinship with the lost
Madeleine, suggests Scotties feeling of dissatisfaction with this working-class
version of the elegant, sophisticated woman of his memory.
On a more general plane, Herrmanns music helps forge a connection
between Scotties vertigo in the literal sense, that is, dizziness caused by heights,
and vertigo in a metaphorical sense, that is, emotional and metaphysical disor-
ientation, which in Scotties case results from his obsession with someone who
in effect does not exist. While the dual nature of vertigoits involving both
attraction and repulsionis realized visually through an unprecedented com-
bined forward zoom and reverse tracking shot, musically it is realized through
rolling arpeggiated seventh chords punctuated by harshly bitonal ones. In some
subtle but undeniable way it is due to Herrmanns vertiginous musical cues

Even more conservatively, perhaps the only fictional truth generated is that the narrator is
reminding us of the possibility of Scotties imminent vertigo, without it yet being fictional either
that it is imminent, or that Scottie believes that it is.
Bruce, Bernard Herrmann, 163.
170 Music

that it is fictional, in Vertigo, that the physiological and the psychic aspects of
Scotties affliction are but two sides of the same coin.
I have tried to show, through the varied examples in this section, the viabil-
ity of a making-fictional criterion of narrativity for nondiegetic film music.
There is, I submit, an intuitive match between the concepts: any nondie-
getic music we would regard as narrative in status is music that can be seen
as contributing to making something fictional in the world of the film and
vice versa.

IX

Narration, though, however broadly construed and however subtly carried

off, is not always the basic charge of nondiegetic film music, and serving a
narrative function not always the best explanation of its presence. I want now
to consider films where nondiegetic music is featured that appears not to be of
a narrative sortwhere thus, in my terms, the music does not make anything
fictional in the world of the film and is not reasonably assignable to the films
internal narrator. Instead, the music seems best understood as directly at the
service of the implied filmmaker. I begin with some films that are in different
ways intermediate or borderline in regard to the contrast I want eventually
to draw.
In Fellinis semi-autobiographical 8 1/2, Guido, a famous but floundering
director, has gone to a fashionable spa to try to recover his mental equanimity
and decide on a direction for his new film. We find him in a spacious bath-
room, as Wagners Ride of the Valkyries begins on the soundtrack. There is
a cut to masses of people taking the waters at the spa, walking in rows and
carrying parasols, among whom Guido eventually takes his place and receives
his allotted glass. We see a conductor conducting, though with no orchestra
in sight, and later see that he is leading a small salon groupone that could
not be the source of the music we hear in the form we hear it. That cue ends
and Rossinis overture to The Barber of Seville immediately starts up, but with
a robustness, once again, that surpasses the resources of the musicians visually
established as present. The effect of both cues, it seems, is one of gentle mock-
ery of the behavior and attitudes of the spas clientele.
The musical soundtrack during this sequence is what one might call quasi-
diegetic. That is to say, the music can be thought to be audible in the world of
 Film Music and Narrative Agency 171

the story, because it is fictionally grounded in an observable source, and even

confirmed later as something heard by a character (as by Guidos subsequent
whistling of snatches of the Rossini)but not in the precise form heard by
the viewer, in respect of volume, instrumentation, or performance quality.
The same quasi-diegetic status attaches to the music in the final scene, the
press conference-cum-party, designed to launch Guidos supposed film, at the
extravagantly erected Spaceship site. We hear Nino Rotas excited music,
which begins with a variant on Khatchaturians Saber Dance, and eventu-
ally brings in almost all the other motives heard earlier in the film, as Guido is
mobbed by impatient questioners and alternately shielded or prodded by his
handlers, all captured in swooping, restless camera movement. Once again,
we are shown a small band set up on a platform, and can even observe at one
point the synchronization of the soundtrack with the rhythm, visually appar-
ent, of the bands drummer, but there is still a discrepancy between what we
can hear of Rotas marvelous score and our sense of what sort of sound the
band visually in evidence could have produced.
So, does such quasi-diegetic music serve a narrative function? To the extent
the music is considered nondiegetic, its function seems to be, in the first scene,
satirical commentary, and in the second, mood enhancement, both arguably
from a point of view internal to the film. So despite their peculiar status, these
cues, insofar as they are nondiegetic, are plausibly ascribable to the cinemat-
ic narrator. They make things fictional: in the first instance, that the narrator
views the spa goings-on satirically, and in the second, that the narrator wants
to induce a certain mood in viewers with regard to the final episode. The
soundtrack musics equivocal status as diegesis does not thus seem to yield
anything correspondingly intermediate as regards narrative assignability.
Another example of intermediate status occurs in the rogue auto scene in
Hitchcocks North by Northwest, in which the villain Van Dams henchmen
attempt to kill the hero, Thornhill, by forcing him to drive down a danger-
ous cliff road while completely inebriated. I would claim that the music of this
scene not only generates tension and underlines the drivers state of drunk-
enness, but at the same time signals, through its jokey style and lighthearted
character, the absence of any real danger for Thornhill. Is this then a commu-
nication from the cinematic narrator, or from the implied filmmaker? That
is to say, is it fictional that Thornhill is not truly in peril, or at least that the
narrator knows he is not? Or is it rather that Hitchcock is telling us, on the sly,
172 Music

that he does not intend to do away with his main character at this point? It is
hard to say which, but in a film whose borderline self-conscious or modernist
character has often been remarked, this is perhaps not surprising.
Most of the music in Peter Weirs Witness, composed by Maurice Jarre,
functions in the by now familiar mood-setting, character-delineating, attitude-
evincing, or thought-specifying way, and is unproblematically categorizable
as narrative. It begins with floating, gently pulsating synthesized chords, as
images of Amish farmers looking up out of fields, and buggies traveling down
roads, occupy the screen. What is conveyed is a sense of harmony and awe,
a sense of the homogeneous spirituality of the world inhabited by the Amish,
especially as compared with the vulgar and violent world of Englishers (the
Amish term for their secular neighbors). During a sequence in which an Amish
boy in Philadelphias 30th St. Station gazes high above him at an erotic statue
of two mythic figures in some sort of embrace, the pulsating music, in voice-
like chords, comes back, suggesting his bewilderment and wonder at the statue
and what it depicts. A variation of this gently pulsing music is prominent dur-
ing detective John Books night of healingwith Rachel, a beautiful Amish
widow, at his bedsideat whose farmstead he has ended up with a gunshot
wound. The music serves to suggest the growing intimacy and spiritual bond
between them. After the violent climax, in which Book manages to dispose
of his corrupt pursuerswith the help of some Amish corn, providentially
stored in a silothe pulsing music underscores the long, silent glances of
farewell between the two protagonists, reaffirming the essential goodness of
their interaction, which stops poignantly short of actual love-making. In all the
foregoing, the music is naturally construed either as establishing something
about the characters or else as evincing the attitudes of the cinematic narrator
towards themattitudes we are clearly invited to share.
However, there is the virtuoso Barn Raising scene, located roughly in the
center of the film, to consider. This provides the occasion of the films main
musical cue: an extended piece, lasting about four minutes, on the order of
Pachelbels Canon (that is to say, variations on a ground bass). The image
track shows us wagons laden with supplies, coming together, people on foot
congregating, getting ready to work, and then, in stages, the raising of the

See Wilsons discussion in Narration in Light, ch. 4.

The particular visual look of these images is due to Weirs trademark use of idealizing telephoto
shots.
 Film Music and Narrative Agency 173

barn, beginning with walls assembled at an earlier time, and finishing with the
whole superstructure in place. The music, by means of its unity, solid flow,
and arching sureness of direction, admirably symbolizes the strength of the
Amish and the spirit of life-affirming communitarianism exemplified in the
activity of cooperatively building a newlywed couple the barn they will need
to sustain themselves.
What, then, gives any pause in regarding this cue as wholly narrative? Only
this: the meter and rhythms of the music in this scene are largely and signi-
ficantly, though not slavishly or mechanically, synchronized with the actions
visually depicted. The pace and pattern of the visual editing seem to respond,
not so much to any internal narrative demand, but rather, to the steady pro-
gression of the music. The cue is not so much designed to flesh out the scene
as the scene seems designed to illustrate the cue. All told, this suggests assign-
ment of the cues music to the implied filmmaker, as opposed to the internal
narrator, since the artful synchronization noted is most naturally taken as an
aspect of the aesthetic construction of the film as the conjunction of an image
track and a sound track, rather than an aspect of how the narrator is present-
ing, through resources available to him, the story. It seems plausible to regard
the music of Barn Raising as attributable, at least in part, directly to the
implied filmmaker.
The main cue in Hugh Hudsons Chariots of Fire occurs near the begin-
ning of the film, accompanying a scene of athletes in training: two dozen or
so men running along the ocean in gym whites, represented as the fifty-year-
old memory of one of the runners. Vangeliss synthesized music, a tune of
simple nobility over a throbbing bass with snare drum-like accents, is heard
throughout, as the credits roll. The cue lasts a few minutes, and the scene ends
visually with the group of men cutting inland and returning to the grounds of
a building in Kent, where they have gone to train in preparation for the 1928
Olympics.
Now this cue may contribute in part to narrationunderstood as making-
fictionalby making more precise the state of the runners as exhilaration,
as opposed to mere determinedness, or by evincing a narratorial attitude, e.g.
one of confident control, or by indicating a mood the narrator would like
to impose on the viewer, e.g. one of heroism. But there still seems to be a
certain surplus value, as it were, to the cue. Those narrative ends do not
appear to exhaust the functioning of the cue; its scale and expressiveness seem
more than is called for with respect to those ends, imparting to the activity
174 Music

of jogging on the beach an almost godly aspect, without it becoming fictional

in the story that such activity really has such status, or even that the narrat-
or believes that it does. Instead, it seems tempting to regard it as attributable,
at least in part, to the implied filmmaker directly: it appears to testify to the
almost religious regard in which he holds the athletic efforts of those young
Britishers of yesteryear. The emotive surplus value of this cue, as far as plaus-
ible narrative functioning is concerned, is what points, it seems, to the implied
filmmaker as a locus of attribution.

Having uncovered some cases of film music with equivocal or partial narrative
status, we are now ready to contemplate cases of substantially, perhaps wholly,
nonnarrative film music. My claim is that such music, which I characterize as
additive (or juxtapositional) film music, is attributable directly to the implied
maker of the film. Such music alters, often powerfully, the artistic content or
effect of the complete film, but it does not do so by nuancing narration, i.e. by
making or helping to make things fictional in the films world.
As a first example, consider Robert Bressons Mouchette. There is only one
significant musical cue in the film, a segment of Monteverdis Magnificat. It is
heard very near the opening, during which the titles are projected, and again
at the end, when Mouchette, an abused country girl of 13 or so, commits sui-
cide by rolling in a sheet into a pond and drowning. Lindley Hanlon gives a
sensitive reading of the music in this film that supports, I think, a largely non-
narrative understanding of it, an understanding that connects it rather more
closely with the filmmaker than with the films internal storyteller:
From Mouchette on, Bresson uses music only at the beginning or the end of a film
unless the source of the music can emanate from the space and situation of the film
narrative . . . It is a more subtle, less intrusive means on Bressons part of authori-
al commentary on the action of the film . . . Recurring after Mouchettes death, the
Monteverdi music seems to function as Bressons requiem for the girl, who has
wrapped herself in shroudlike vestments . . . The words of the Magnificat affirm the
possibility of another life after death and sanctify Mouchettes decision to escape
from the despair of her own life.

Sound in Bressons Mouchette, in Elisabeth Weis and John Belton (eds.), Film Sound (New
York: Columbia University Press, 1985), 32930.
 Film Music and Narrative Agency 175

The music here is most plausibly assigned to the implied filmmakeras affirm-
ing the general possibility of grace as exemplified in the tale of Mouchette
rather than to the films relatively effaced internal presenter, especially as it
seems to frame the fictional narrative from without, like a pair of musical
bookends, as opposed to shaping it from within.
Terrence Malicks extraordinary film, Badlands, provides an outstanding
example of an appropriated score, consisting mainly of extracts from Carl
Orffs Musica Poetica and Erik Saties Trois Morceaux en Forme de Poire.
This score also serves as one of my key examples of nondiegetic film music
that is not, in the main, usefully construed as narrative.
Badlands, based loosely on the Charles Starkweather shooting spree of 1958,
contains a partially unreliable narration, since two components of it, the image
track and the voice-over narration by one of the main characters, Holly (Sissy
Spacek), are at odds with one another (in some respects, only at certain points,
and in other respects, throughout). Here, as is customary, the visual repres-
entation is taken to be the more truthful, on the convention that seeing is
believing, and so when what is shown conflicts with what is told, we are
inclined to credit the former.
Orffs and Saties music, I maintain, is characteristically employed in Bad-
lands in a mode of distanced and reflective juxtaposition to the story narrated,
by an intelligence standing just outside that narration. It is not, in general,
attributable to the films narrating agent, but only to the implied filmmaker.
To make this point I examine at length one particular cue. Fairly early on, we
are shown Kit (Martin Sheen), the films other main character, working cattle
in a feedlot, after having been fired from his job as a garbage collector. On the
soundtrack is a striking, far from inaudible, portion of Orffs score, consist-
ing of sharply rhythmic xylophone or marimba music, built on an exotic scale,
having no obvious connection with, or fittingness to, gritty scenes of cows
being force fed, almost expiring in the heat. That is to say, there is nothing
in the character of the states of affairs depicted that the music could plausibly
be thought to second, nor anything indeterminate about those states of affairs
that the music might plausibly be thought to specify.
Could it be narrative in the sense of expressing the cinematic narrators
view of the situation depicted? This seems unlikely, if only because it is rather

In identifying this theme as of grace I of course rely on a knowledge of Bressons oeuvre as a

whole, and of the artist implicit in that oeuvre, one with a deeply Catholic vision of the world.
Chatman, Coming to Terms, 136.
176 Music

unclear what sort of attitude could be signaled by such music in relation to the
events shown. In addition, the cinematic narrator, who often visually corrects
or gainsays Hollys romantic and simplistic notions of what has transpired in
her time with Kit, comes across as an agency too sober and straightforward,
almost nonhuman in its detachmentconsider the odd montages of nature
shots that occur occasionally during the film, giving the impression of an
iguana-eyes point of viewto be credited with a sentiment as quirky and
mischievous as that expressed by this musical cue.
Might the music be narrative in virtue of acting to characterize Hollys
recollective impression of Kits job at the feedlot? Such a hypothesis is mul-
tiply problematic. First, we havent been given any reason to think the non-
diegetic music is in the service of the voice-over narrator, but at most, the
cinematic narrator operating from the point of view of or on behalf of some
character; that is to say, there must be rather special indications, not here
present, before we will think of nondiegetic music as a resource belonging
to, rather than applied in elucidation of, a character in the story. Second,
since there is reason not to regard the image track as an accurate version of
Hollys memoriesit regularly outstrips, and occasionally contradicts, her
verbal narration of what happenedthe ground for thinking of the sound-
track music as signifying Hollys impression of those sights seems lacking.
There is little reason to think, in particular, that she ever visited Kit at the
feedlot or witnessed the kinds of scenes we see on screen. Third, whatever
attitude we found such music to connote, it seems not to be one we would
ascribe to hazy-minded Holly while the thought of Kit at the feedlot was
before her mind.
This leaves as the only interpretively live possibility the assignment of the
music to the implied filmmaker who, from a point outside both the story and
its narration, has apparently added this music as a kind of counterpoint to the
fictional drama. To what end? It is hard to say, especially without an interpret-
ation of the film as a whole, but possibly one of aesthetic embellishment, or
derangement of the viewers moral compass, or refraction of the storys con-
tent in a distorting mirror, or external meditation on the films happenings.

The sequences that make it clearest that the image track is not to be thought of as a reliable
representation of Hollys occurrent memories are one in which we see Kit shoot a football and then
hear Holly, a minute later, recount this event, and another in which we see Kit, trying to outrun his
police pursuers, suddenly stop his car, get out, shoot its left front tire flat, and then blatantly await
capture, while Holly alludes to the incident, never observed by her, in a mode of speculation rather
than reportage: Many times Ive wondered about why Kit didnt get away. He said he had a flat,
but from the way he kept coming back to that, I doubt it.
 Film Music and Narrative Agency 177

Now the music of Orff and Satie is characterized in general by an inten-

tional simplicity, a primitiveness of musical materials, and a studied directness
of effect, and that employed in this film is no exception. Thus perhaps the
function of this music in the filmon a global plane, rather than scene-by-
scenecould be said to be a reflection of the basic childlikeness and oblivi-
ousness to social reality of the two principals, and especially that of Holly, the
verbal narrator. I think that is so, but for the reasons given above this music,
and that aspect of the films content, is best laid at the door of the implied
filmmaker, rather than any agent internal to the narrative.
My next examples come from Woody Allens Love and Death, whose
appropriated score is derived entirely from the suites to Lt. Kije and Alexander
Nevsky by Prokofiev. The sleigh-like music from Lt. Kije starts up after
Natasha (Diane Keaton) announces her engagement to a herring merchant,
and extends through her riding off in a carriage and subsequent shots of
Russian troops in training, marshaled to protect Russia against Napoleon.
This music has a satirical effect, more properly attributed to Allen as auteur,
than to Boris, Allens character, as narrator, or even the cinematic narrator
conceived as encompassing Boris verbal narration. A farcical battle scene
between Russian and French troops, shortly thereafter, is accompanied by the
grim and heavy music for the Battle on the Ice from Alexander Nevsky; the
mismatch is palpable, and the implied equation of the two battles laughable.
Both the satirical intent inherent in this juxtaposition, and the frame of
cultural reference with which it operates, seem to put it beyond ascription
to either Boris or the cinematic narrator.
The last example I discuss of a film much of whose musical soundtrack
is best seen as additive or juxtapositional, rather than narrative, is Stanley
Kubricks A Clockwork Orange. The opening creditthe words A Clock-
work Orange on a garish orange fieldis followed by a close-up of Alex
and his pals (droogs) in a bar that dispenses drugged milk (moloko), dispos-
ing its consumers toward acts of ultraviolence. Soon Alexs voice-over is
heard, which establishes what we will soon see as Alexs recollections of his
recent past. Walter Carloss synthesized music here is a slow-moving, quasi-
Handelian progression, with a hint of Dies Irae. It functions narratively in
setting an appropriate mood, in suggesting something of the effect of moloko-
drinking, and in perhaps foreshadowing some of the grim doings the narrator,
acting on Alexs behalf, has in store to present to us, in due course. But the
appropriated music employed in the film, notably that of Rossini and Beeth-
oven, functions rather differently.
178 Music

Rossinis La Gazza Ladra Overture begins on the soundtrack as an old man

is being beaten by Alex and his droogs, continues over a cut to another gang
of youths assaulting a naked girl on a stage, leading to a fight between the two
gangs, and covers the escape of Alex and his droogs from the scene by car,
fading out only as they approach a house in the country whose occupants they
are going to terrorize.
There is no obvious narrative appropriateness to the music: it seems neither
to convey information about the events shown, nor to suggest the narrators
perspective on those events, nor to suggest an attitude that viewers should
plausibly adopt toward them. I take it that the first and third points will be
granted without dissent; the second, though, might be supported further, as
follows. If the claim of narrative function is to be sustained on that ground,
it seems we would have to posit either a perversely inhuman cinematic nar-
rator, whose lighthearted view of the proceedings is reflected in the music, or
else a psychologically more normal one who merely signals to us, through the
music, Alexs perversely comic perspective on the violence he is perpetrating
on others. The first possibility strikes me as unmotivated, while the second,
though more promising, faces the problem that it casts Alexs reactions on
perhaps too high a level of sophistication.
Thus we arrive, once again, at the assignment of this music directly to the
implied filmmaker as interpretively the most reasonable option. As such, how
does it function? Pasted onto the scenes of violence presented by the films
internal narrator, it invites us, at least initially, to see them as a joke, thus mak-
ing us complicit in the mindless pleasure of Alex and his pals in inflicting pain,
in the expectation, presumably, of getting us to be even more horrified when
we realize what weve been duped into. Kubrick, and not the cinematic nar-
rator, is addressing us directly through this odd and unsettling juxtaposition
of music and story.
A similar scene takes place in Alexs room at home, with two girls he has
picked up in record shop. It is filmed in extremely fast motion, to the accom-
paniment of Rossinis William Tell Overture. Here both the fast motion film-
ing and the superimposed frenetic music seem to reflect the activity of the
implied filmmaker, as opposed to that of the films perceptual enabler or inter-
nal commentator.
A related, though distinct, use of music occurs in a scene also set in Alexs
room, to which he has repaired after the first nights round of ultraviolence.
He deposits things in his booty drawer, checks on his pet python, and puts the
scherzo of Beethovens Ninth Symphony on his sound system. The music,
 Film Music and Narrative Agency 179

here diegetic, is synched to a montage of close-ups of statue parts, as Alex

imagines acts of sex and violence, recounted in voice-over. But this intrastory
perversion of Beethoven by the protagonist echoes and parallels the implied
filmmakers superficially warped overlaying of Alexs recollections of occa-
sions of torture and fornication with Rossinis diverting scores.
Near the very end of the film, Alex is being questioned by a few intellec-
tuals, including the writer he crippled earlier in the film, about behavioral
conditioning via background music. It is not too much to suggest that this
scene obliquely raises within the film the issue of film musics legitimacy and
role, and of its possible subversive effects, e.g. the undermining of autonomy
or the blunting of rationality. This self-consciousness in the film about what
we may call nonaesthetic or incidental uses of music reinforces the assignment
of additive, as opposed to narrative, status to the Rossini overtures appropri-
ated by Kubrick for A Clockwork Orange.

XI

Though in many cases where nondiegetic film music is more reasonably

assigned to the implied filmmaker rather than the films narrative agent, we
find that such music is being used ironically or satirically, e.g. as with Love and
Death and A Clockwork Orange, it is important to remember that that is not
the only possibility. The examples of Mouchette and Badlands, and in a par-
tial vein, Witness and Chariots of Fire, illustrate as much. And we may also
observe, at this point, that nondiegetic music is not the only music in a film
responsibility for which may redound, without intermediary, to the implied

A contrasting, rather more cynical, view of the mode of film-scoring of which A Clockwork
Orange was perhaps the pioneer is provided in this recent commentary: Faced with the task
of differentiating their scenes of brutality and mayhem from all the other scenes of brutality
and mayhem, filmmakers are using music to distance the viewer from violenceor to comment
ironically on it. As the images get more explicit, the accompanying tunes seem to get more frothy.
Everything from Bach to hook-laden pop-rock songs provides background for images of fist fights,
shootings, stabbings and torture (Kenneth Chanko, Its Got a Nice Beat, You Can Torture to It,
New York Times, Feb. 20, 1994).
The scene may in fact be what Wilson calls a rhetorical figure of narrative instruction,
something offered by the filmmaker to the viewer as a key to interpreting the films narration
generally. See Narration in Light, 4950.
Another intriguing case is Slava Tsukermans Liquid Sky (1983), whose soundtrack employs
an overmodulated synthesized harpsichord version of eighteenth-century composer Marin Maraiss
hypnotically repetitive Sonnerie de Ste. Genevieve du Mont. It is not clear to meon the basis of
a single viewing, many years agowhether that music belongs in the satirical or the non-satirical
subcategory of additive film music.
180 Music

filmmaker, and which may be read by us as a direct reflection of authorial

stance or personality. Jane Campions film, The Piano, offers a case of film
music commissioned and composed for diegetic insertion in a filmMichael
Nymans music for mute protagonist Adas pianismrather than nondie-
getic accompaniment. The musics characterization of its fictional originat-
orAdais a function that can only be assigned, it seems, to the implied
filmmaker, as the agent who has chosen the characters, their actions, and
their traits, in constructing and arranging the elements of the filmic object
as she has.
How does the distinction we have been exploring, of film music as additive
versus film music as narrative, relate to another standard classification, namely
that of film music as commentative? The answer is: not simply. The equation
of narrative and commentative will not do, for two reasons. First, some music
of clearly narrative function is not reasonably thought of as commentative,
unless all information-conveying counts as commentary. Second, some addit-
ive music seems to supply a commentary, if oblique, on matters with which
a film is concerned. In light of this, we might distinguish between extern-
ally commentative music, assignable to the implied filmmaker, and internally
commentative music, assignable to the cinematic narrator.
Still, it is important to stress that musical commentary on the events of a
fictional story as such, or the characters figuring in those events, remains a
possibility only for the cinematic narrator internal to the fiction. Additive
music, assignable to an implied filmmaker, might generate, as noted, a
kind of commentary as well, but it could not be on the fictional events
themselves, from a perspective internal to the fictional world, but at most
on the representation of those events or on the significance of events of that
type. The implied filmmaker of a fiction film is not on the same plane as the
events of the films worldwhich are for him, as for us, fictionaland so
his direct commentary on those events is not a coherent option. For instance,
if the Magnificat cue at the end of Mouchette expresses Bressons attitude of
consolation toward Mouchettes suicide, this has to be understood not as an

One critic has remarked on the music for this film as follows: Both the orchestral and solo
keyboard music suggest a modern minimalist gloss of Chopin and Liszt but spun off plain, abrupt
folk tunes . . . the pianism suggests someone doggedly trying to speak through the keyboard . . . As
distinctive as it is, the music is strangely cramped and emotionally arid . . . the solo piano passages
sound too much like elementary practice exercises to soar into the stratosphere (Stephen Holden,
New York Times, Jan. 30, 1994).
 Film Music and Narrative Agency 181

attitude literally directed on the suicide of Mouchettean event in which

Bresson presumably does not believebut instead as an attitude bound up
with the films representation of that event, or directed toward events of the
sort represented by the film.
A standard function of nondiegetic film music, we have observed, is to
reveal, confirm, or make precise a characters feelings or attitudes toward
something or other in the story. Such a function makes most sense in
connection with a narrator, rather than an implied filmmaker, since it pre-
supposes an agent on the same plane, fictionally, as the characters, whose
existence the narrator credits, and whose lives the narrator selectively presents
to us. The deliverances of narrative film music seem to come from one who
shares a world with the characters, rather than one who has invented them,
and everything else in the fictional world, from whole cloth.
On the other hand, another standard function of nondiegetic film music
is to bathe the incidents of a film in a common atmosphere. The thematic,
instrumental, and stylistic continuities typical of film scores help to create a
consistency of tone or feeling across the span of a film, especially where the
events presented are not very tightly connected in a dramatic sense. Thus this,
rather than any narrative task, seems to be the main function of Rotas score
for Fellinis Amarcord. When nondiegetic film music has this function, it is
more naturally ascribed to an implied filmmaker than to an internal cinematic
narrator. Nondiegetic film music bridging scenes of different character, say,
or smoothing over large lapses of time, is of this sort. Such music, like the
presentational, voice-over, and mind-over narrators in a film, is understood
primarily as constructed or arranged by the implied filmmaker in putting
together the aesthetic object that is the total film, rather than as something
used or employed by the cinematic narrator in its different narrative
capacities.
Returning to the five functions of film music recognized by Copland, I
would suggest that only twounderlining characters psychological states
and sustaining and releasing tensionare clearly assignable to the cinematic
narrator. The othersensuring continuity, providing background filler, and
creating atmospherecan with equal, or more, justice be regarded as activ-
ities of the implied filmmaker, in that they seem aimed directly at the viewer

A function highlighted by Noel Carroll in Mystifying Movies, 21623; Carroll labels film
music of this familiar type modifying music.
182 Music

as an aesthetic subject, at causing his or her experience to be a certain way,

rather than at defining or delineating the films fictional world. If we con-
sider, similarly, the list of functions drawn up by Gorbman in her study of
the operation of classical film music, I would suggest that twothe signify-
ing of emotion, and the referential and connotative cueing of narrativeare
assignable to the cinematic narrator, while the remaining twothe provision
of continuity and the achievement of unitymake most sense as assignable
to the implied filmmaker.
What of my own list of fifteen functions of film music, drawn up earlier,
in Section VI? By present lights, I think they sort out as follows: functions
(1), (2), (3), (4), (5), (6), (7), (8), and (9) are arguably narrative, in that they
involve making something fictional in the film, and so music functioning in
such ways is assignable to the cinematic narrator. Functions (10), (11), (12),
(13), (14), and (15) are arguably nonnarrative, and are often achieved through
music of additive status, assignable only to the implied filmmaker. There is
not, however, a perfect correspondence between the division of functions as
either narrative or nonnarrative, and the categorization of cues as either nar-
rative or additive, because a cue can have significant functions of both sorts.
What is true is roughly this: if a cue has significant narrative function, wheth-
er or not it functions in addition nonnarratively, then it is a narrative cue,
whereas if a cue has no significant narrative function, then it is an additive cue.
The question I have been exploring in the latter part of this essay can be put
as follows: when is nondiegetic film music primary a compositional element
in a film, at the command of the implied filmmaker, and when is it instead,
or in addition, an instrument we imagine as at the service of the cinematic
narrator, generating truths in the world of the film, either about the story as
such or about the act of its narration? But perhaps the same question poses
itself, on close examination, for a number of other filmic elements viewed ini-
tially just as compositional, e.g. lighting or camera angle. When is the dim
or filtered quality of light in a sceneas in Vertigo, when Judy reemerges
into Scotties presence as Madeleinemerely a directorial choice and when
a manifestation, as well, of narrative activity on the part of the films intern-
al presenter, showing things in a light they would not otherwise appear in?
When is an off-kilter view of a man running across a squareas in The Third

Though in regard to the last of these, creating atmosphere, it was suggested earlier how this
can in many cases be understood as having narrative status, if the atmosphere involved is one the
films narrative agent can be plausibly thought of projecting.
Unheard Melodies, 73.
 Film Music and Narrative Agency 183

Manjust a matter of the directors tilt of the camera in relation to the actor
being filmed, and when is it to be regarded as well as connoting an interven-
tion of the cinematic narrator, as showing us the character from an oblique
perspective, with whatever that suggests about either the character or the nar-
rators view of him? The issues addressed here concerning the interpretation
of nondiegetic film music resonate, I suspect, across the whole spectrum of
meaning-making elements in film.
 10
Evaluating Music

Wagners music is better than it sounds.

(Mark Twain)
If it sounds good, it is good.
(Duke Ellington)

The above opposed epigrams about musical worth neatly serve to introduce
the sorts of issues I want to explore. Is Twain right, in terms of what he implies
about the grounds for evaluating music, or is Ellington? If they are both right,
how can we reconcile the apparent conflict between the principles suggested
by their respective observations?
One possibility for reconciliation is that Ellingtons epigram is to be taken
straight, while Twains is to be understood as tongue-in-cheek. Another is
that Twain is alluding to secondary or sophisticated aspects of musical worth,
while Ellington is focused on primary or elemental ones. A third is that their
epigrams simply apply to different, and disjoint, spheres of music, Twains to
the classical tradition, and Ellingtons to that of jazz. A fourth is that Twains
observation reflects the fact that, for some music, evaluations evolve over time,

First published in P. Alperson (ed.), Musical Worlds: New Directions in the Philosophy of Music
(University Park, PA: Pennsylvania State University Press, 1998), 93108. An earlier version of this
essay appeared in Revue Internationale de Philosophie 198 (1996): 593614.
These epigrams are not, of course, strictly inconsistent with one another. For Twains epigram
does not even entail that Wagners music is good. But suppose, for the sake of argument, we take
Twain s remark to have the force, that there is music that is good, but does not sound good.
Even so, there is no conflict per se with Ellingtons dictum, but only with its converse, to wit, if
music is good, it sounds good. However, if we construe Ellington as implicitly committed to the
biconditional, which seems reasonable, then a contradiction emerges. Otherwise put, if we elaborate
the principle behind Ellingtons remark to be music is good just insofar as, and to the degree that, it
sounds good, and that behind Twains remark, taken straight, to be music is not always good just
insofar as, and to the degree that, it sounds good, then the logical opposition of these principles is
plain.
 Evaluating Music 185

so that impressions formed on first exposure are replaced, on deeper acquaint-

ance, by opposite ones, while Ellingtons observation underlines that, for much
music, the ears first impression is a pretty reliable guide to musical worth.
I will forbear trying to decide among these possibilities for reconciliation.
I will, however, assume there is some truth to Twains remark underneath its
display of wit. We shall see a reflection of this shortly in our discussion of
whether all the value of music as music can be encompassed under the rubric
of how good it sounds, or even that of how rewarding it is to experience.

When confronting the issue of musical value, two questions must be

distinguished at the outset. One question is that of the value of music
generally. Why is any music valuable, and how does music, of any sort, add
to or enrich human life? A particular form of this first question, though
one that takes it perhaps closer to the second question, asks what makes
music distinctively valuable, as opposed to other arts; that is, what does
musicmore or less any musicoffer that other arts or activities do not,
or at least, not to the same degree?
A second question is that of the value of particular pieces, genres, or styles of
music, and is more obviously inherently comparative in nature. What makes
a given piece (genre, style) of music valuable, or alternatively, what makes this
piece (genre, style) of music more valuable than that one?
In even pithier guise, the two questions are, in effect, (a) what makes music
a good, or contributory to human good, and (b) what makes this music good,
or better than other music?
Once these questions are roughly distinguished as above, an immediate
concern is what is the relationship, if any, between acceptable answers to
them? Does knowing the answer to (a) help with the answer to (b), or vice
versa? Is there a constraint in either direction, so that, for example, what
makes music valuable generally is also essentially what makes particular pieces
valuable, or that what makes particular pieces of music valuable is, writ large
or in some way summed, what makes music valuable generally? One possible
relationship might be that a work that displayed in high degree the properties

See Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994),
2756.
186 Music

that make music valuable generally, or one that fulfilled well the function ful-
filling of which makes music of value generally, was therefore a relatively valu-
able piece of music. But is that so? It seems unlikely that things are as simple
as that.
In any event, a clear case of divergence between the value of music as
a wholeof the phenomenon or practice of musicand the value of an
individual piece of music would be the opportunity music affords to bring
people together in a social setting for shared experience and interaction. Call
that the social value of music. This appears to be a value of music generally
without being a differentiating value of any individual piece of music. Other
candidates for values of music generally that are yet not such as to add much,
if anything, to the comparative value of an individual work of music, would
be these: serving as a vehicle of relaxation; serving as a distraction from
practical concerns; serving as an accompaniment to and facilitator of activities
involving bodily movement, such as dancing, marching, exercise, or physical
labor.

II

In an important study devoted to the second of our questions, namely, that of

the artistic value of individual works of art, Malcolm Budd offers a completely
general conception of such value, meant to cover all artworks in all artforms.
Budd aims to identify an artworks artistic value in such a way as to differen-
tiate it from other values it may possess, e.g. as social record, religious artifact,
financial investment, or totem of prestige. Budds straightforward proposal is
that the artistic value of a work of art, its value as art, is determined by or is a
function of the intrinsic value of the experience the work offers.
By the experience the work offers Budd means an experience in which the
work is fully and correctly understood, its individual nature grasped for what
it is. By the intrinsic value of the experience Budd means the value of having
such an experience for its own sake, rather than for the sake of any effects or
consequences the experience may engender.
What is excluded from the artistic value of an artwork is thus (a) anything
about the work not directly reflected in the right sort of experience of it, and
(b) any value of such experience that is not purely intrinsic, i.e. is in some way
instrumental.

Values of Art: Pictures, Poetry, and Music (London: Penguin, 1995).

 Evaluating Music 187

Budds proposal, applied to music, is thus in line with, if obviously a refine-

ment of, Ellingtons dictum that musics value lies, above all, in how it sounds
when you listen to it. There is little point in gainsaying the healthy intuition
behind both Ellingtons dictum and Budds proposal. Still, though Budds
proposal seems roughly acceptable, there are at least three grounds on which it
can be challenged.
One concerns the presupposition of a viable division between the effects or
consequences of an experience, and the parts or elements of an experience, a divi-
sion that may be difficult to sustain. The second concerns the restriction to
the intrinsic value of the experience a work offers as the sole gauge of a works
value as art, a restriction that appears inadequately justified. And the third
concerns the confinement of artistic value to that which is manifested in or
through experience of a work, a confinement that seems at odds with certain
firmly grounded judgments of artistic value. I elaborate on these in turn.
The distinction between an experience and its effects, though unproblem-
atic on its face, has some tendency to dissolve under scrutiny. Experiences
often have no unequivocal beginning and ending points. They characterist-
ically do not start up with the sharpness of a pistol crack, nor do they charac-
teristically close with a full stop. Many experiences have indeterminate
beginnings, and take shape slowly. Often, rather than ceasing abruptly, they
simply fail to continue developing or ramifying, though exactly where and
when may remain elusive. This blurriness-around-the-edges is evident enough
with traumatic experiences, such as losing a loved one, but attaches, if less
blatantly, to many more ordinary experiences, appreciative ones among them.
Turning to the case at hand, the endpoint of the experience of a musical work
in audition, for instance, is fairly fuzzy, with no clear cut-off between the
experience itself and what might be called, fittingly enough, its echoes and
reverberations.
A second, more important, challenge to the formula Budd offers us is this.
Even if we assume the artistic value of a work of music to lie wholly in the
value of the experience, suitably demarcated, that the work offers, it may not
be defensible to restrict that value to the intrinsic, as opposed to instrumental,
value of the experience, that is, the value of the experience itself as opposed
to its effects. Granting for arguments sake the workability of the division
in questionbetween effects of, and elements in, a given experiencethere
It may be that the broader value these instrumental goods contribute tothat a life is a
certain way or possesses certain featuresis intrinsic, even if they remain instrumental goods of the
experience. (For discussion see Ch. 24 of this volume.)
188 Music

is insufficient reason to hold that only the intrinsic value of the experience a
work offers is a measure of its artistic value, rather than what accrues in virtue
of the experiences effects.
Suppose a given work of music is valued in part because it gives insight,
when properly experienced, into the character of romantic love or the
inevitability of suffering. If so, would this count as an intrinsic or an
instrumental value of experiencing the music? The best answer would seem
to be that it is both. For it is intrinsically valuable to have such insights
while listening, or in the course of subsequent reflection on the music, but
also instrumentally valuable, because the insights are thus, after all, acquired,
something one can summon up for further use or benefit. If something counts
as an insight, and if music affords it, then it is then an enduring asset,
something whose value goes beyond the confines of the experience in which
it is acquired. The same could be said about music that gave one access to a
point of view that had not previously been available to one; the value of this
would seem to transcend the value of the experience of achieving such access
through the music, and thus be partly instrumental.
It is also natural to consider here the possible moral effects of music. If
there are any such, and if they could be shown to issue with some regularity
from the comprehending experience of certain music, then they would consti-
tute an instrumental value of the experience of such music that would seem to
be at least a candidate for inclusion in the musics artistic value.
But is there not a Catch-22 of sorts lurking in regard to the putative
improving tendency of certain music? It seems that either you already have
a refined nature or developed moral sensibility allowing you to appreciate
such music, in which case you will not improved by it, or else you do not,
in which case you will not be able to properly appreciate the music, and so
it will fail to have its proper effect. In other words, that you appreciate such
music presupposes that you are a person of some moral capacity; thus, such
music would seem to be powerless to transform you into that.
A partial answer to this quandary is as follows. Even if you have to have cer-
tain minimal moral capacities in order to adequately appreciate great music,

Two essays on this topic are Colin Radford, How Can Music Be Moral?, Midwest Studies 16
(1991): 42138, and Donald Walhout, Music and Moral Goodness, Journal of Aesthetic Education
29 (1995): 516. The topic is also explored at length in Kathleen Higgins, The Music of Our Lives
(Philadelphia: Temple University Press, 1991). Higgins makes a good case for recognizing a moral
dimension to music, though her focus is on plausible ethical effects of whole genres and practices,
as opposed to individual musical works.
 Evaluating Music 189

calling as it does on emotional repertoire and practical insight as much as on

perceptual ability, exposing yourself to such music plausibly helps to devel-
op or reinforce such capacities, through providing a controlled arena in which
such capacities to respond are exercised in a specific way, serving as a kind
of touchstone of what it is to be human. In other words, it is true that you
might already have to be disposed, in some measure, to a moral form of life,
in order for great music to be of benefit to you, but immersing yourself in it
might still function, ceteris paribus, to make you more human than you were
beforewithout, of course, guaranteeing any such result.
The example of Hannibal Lecter from Jonathan Demmes Silence of the
Lambs is usefully recalled here, for that unparalleled film villain notoriously
displays both the most malevolent cannibalism and the most cultivated
appreciation of J. S. Bach. If one is so unkind as to observe that Hannibal
Lecter is a fiction, and a meretricious one at that, it remains true that the
coexistence of moral turpitude and aesthetic refinement in a given person is
more than merely imaginable: Richard Wagner approached this, and many of
the Nazis who later gloried in his music exemplified it rather fully.
But such cases, fictional or actual, are largely a red herring in regard to
claims of a moral dimension to music and the possible relevance of such to
some musics artistic value. No one would claim that great music can, entirely
on its own, make appreciators of it better people, nor that great music, how-
ever supplemented and seconded, is likely to make all appreciators of it moral.
The most anyone can sensibly propose is that such music, properly grasped,
exerts, through the attitudes or states of mind the music projects or the com-
plexes of feeling it evokes, a humanizing and moralizing force, though one
easily enough overridden or neutralized, and thus that, all things being equal,
people exposed to such music tend to be morally better, more humane, than
they would otherwise be. Of course even this remains unproven, but it is not
to be dismissed out of hand.
At the least, good music may help to remove barriers to moral education
by increasing an individuals awareness of the subjectivity of others, which is
The examples of Hannibal Lecter and Nazi doctors enjoying their chamber music in the
evening strike us so forcefully, I suggest, precisely because they are, in fact, exceptional. They violate
an empirically grounded regularity of artistic taste comporting with some degree of moral awareness.
While it seems almost natural for purveyors of gangsta rap, say, to be engaged in shady or illegal
activities from time to time, given the thrust of the music and the implausibility of attributing
it all to a persona unrelated to the rapper, it is more surprising when professional pianists and
violinists are found to be so engaged. The exceptionof immoral behavior from devotees of fine
musiccalls attention to the presumption that there is likely some degree of correlation between
aesthetic refinement and moral awareness.
190 Music

clearly a prerequisite to treating others as ends in themselves and taking their

interests into account in deciding how to behave. Appreciation of good music
may plausibly lead to a more vivid imaginative grasp of the mental life of oth-
ers, a necessary condition for regarding those others in a morally appropriate
manner.
But does great music simply inform us of morally relevant data or acquaint
us with morally relevant perspectives, thus supplying a necessary condition
of acting morally, or does it in addition motivate us, to some extent, to act
morally? If it does not, then it will not be, even all things being equal, a mor-
ally improving force, but only something that lays the groundwork for acting
in a moral way. However, I think there is reason to believe that some music,
at any rate, is not only of epistemic value, but also motivating in relation to
moral life.
In sum, then, it is reasonable to take a musical work to be greater, and great-
er as art, on the assumption that it has moral effects or tendencies of the sort
postulated among those who experience it fully and correctly, leaving aside
whether it can be shown to have them. And this constitutes a ground of value,
or reason for valuing, distinct from that of the specific musical merits in virtue
of which the work possesses such moral force as it does.
This is not, of course, to license every demonstrable benefit of the experi-
ence of a work as contributory to its artistic value. Being instrumentally bene-
ficial in the waysmoral, cognitive, and emotionalI have been discussing
is arguably a part of arts proper purpose or missionunlike, say, the capa-
city of experience of a work to alleviate mental illness, induce sleep in the
weary, or promote a sense of self-satisfaction. The instrumental benefits to
which I draw attention are consonant with historically prevalent intentions
of composers for their works, as well as being implicitly involved in received
judgments about greatness in art. Such benefits rightly enter, it seems, into the
assessment of a musical works goodness as art.
There are surely other instrumental benefits of the experience of music
that could reasonably be reckoned relevant to its worth as art. Music can be
instrumentally valuable in virtue of providing, through its sounding form, a
paragon or practicum of how to move or to behow, in effect, to go on.
Good music, adequately experienced, can serve as a highly abstract, though
suggestive, design for living.

This relates to Monroe Beardsleys suggestion that music might reasonably be held to symbolize
or exemplify general patterns of continuation, growth, or development. See his Understanding
 Evaluating Music 191

Some music, correctly appreciated, may be instrumentally valuable in vir-

tue of the influence exerted on ones general outlook, enabling one to think or
feel the world differently, thus enlarging ones life as a result. Such music does
so by embodying a process of thought or a frame of mind, one that through
attentive and sympathetic listening one is allowed to enter, and that might not
be otherwise communicable, or communicable in as vivid and effective a way.
Some music, finally, may be instrumentally valuable through contribut-
ing to the sense of self and the formation of individual personality. Music-
al works can arguably help to constitute and define the self that attends to
them, internalizes them, and identifies with them. Some music may even
have the disposition to produce such effects on personality as to count as
transfigurative.
It is important to note that the objective value of a piece of music in this
respect is distinct from its personal value in that respect to a given individual.
Both are comparative, but only the latter is relative in a strong sense. The
former concerns a pieces power or potential, among pieces of music gener-
ally, to contribute to self-definition and the like in virtue of its formal and
expressive qualities; the latter concerns the actual historical contribution of a
given piece of music to shaping some persons identity. The latter may be due,
in part, to the former, but there are usually idiosyncratic factors at work as
well. That some piece of music had a profound effect on Sam, and became
a touchstone of his emotional and intellectual life thereafter, means that it
is, unquestionably, a piece of music with personal value for Sam. But it may
not, for all that, have much objective value in this regard, that is, it may lack
significant potential to so affect prepared listeners generally.

III

I turn now to my third challenge to Budds proposal about artistic value.

Even if the artistic value of music is acknowledged to be centrally a matter
of the value of experience of it, whether intrinsic or instrumental, that does
not exhaust what such value comprehends. Why? Because certain things enter

Music, in Kingsley Price (ed.), On Criticizing Music (Baltimore: Johns Hopkins University
Press, 1981).
See Leonard Meyer, Some Remarks on Value and Greatness in Music, in Music, the Arts,
and Ideas (Chicago: University of Chicago Press, 1967), 2241; Anthony Savile, Kantian Aesthetics
Pursued (Edinburgh: Edinburgh University Press, 1993), ch. 6; and Higgins, The Music of Our
Lives, ch. 6.
192 Music

into an artworks artistic value that are not reducible to the value of exper-
iencing the work in the prescribed manner. The artistic value of a work of
music, in other words, may quite reasonably outstrip its experiential value,
even broadly understood.
An important component of musical artistic value is what may be
called influence-value, the impact a musical work has for the better on the
future course of music. Examples of works in whose artistic value there
is a significant component of influence-value include Beethovens Eroica
Symphony, Debussys Prelude to the Afternoon of a Faun, Schoenbergs
Piano Pieces, Op. 23, and Stravinskys Rite of Spring. This component of
the artistic value of music goes beyond the value, intrinsic or instrumental,
of the experience it offers. It is reflected in the works of value that
the aforementioned works have spawned, the new avenues of musical
composition they have opened up, or the unforeseen sorts of musical
experiences they prepare the ground for, but do not themselves afford. When
we identify musical works as seminal, revolutionary, or ground-breaking,
and praise them as such, we are in the realm of artistic value I am labeling
influence-value.
It may be suggested, though, that actual influence-value is one thing, and
potential influence-value another. The former would consist in actual positive
effects on the future of music, whereas the latter, by contrast, would consist in
being such as to give rise to such effects, conditions of reception being favor-
able. In other words, a works potential influence-value would amount to its
having the capacity to beneficially influence the future of music through its
directly appreciable artistic features, to wit, its form, expression, or technique.
Having distinguished actual from potential influence-value in this manner,
one might then go on to suggest that only potential influence-value, which
flows directly from a works having the right stuff, is relevant to claims of
artistic value.
Yet it seems to me that, despite the validity of the distinction between
them, actual influence-value, and not just potential influence-value, is prop-
erly accounted part of the artistic value of works of music, especially when
such works are seen not in isolation, but as part of an ongoing tradition of
music-making and musical thinking. No doubt having the right stuff, and

I am ignoring, for simplicity of discussion, positive effects on other arts or spheres of culture. But
in principle there is no reason to exclude these as irrelevant to an assessment of artistic value. Also,
in speaking of influence-value I shall, unless otherwise indicated, have in mind positive influence
value, that of influencing the future of art for the better.
 Evaluating Music 193

at the right time, figures in a works artistic value, but so, it seems, does what
becomes of that historically. Since actual influence on the history of music
depends, as we know, not only on the nature of the work and the relations
it bears to its antecedents, but on a contingent degree of receptivity to and
uptake of what it offers, we may need to recognize a measure of artistic luck in
how much artistic value accrues to a work. For that is what actual, as opposed
to merely potential, seminality, revolutionariness, and so on, require.
But what of influence that is freaky, undeserved, or even a result of
repugnanceas when an inferior work, in an unpromising mode, prompts
an artist to angrily create something superior in a wholly other vein? The
sort of influence that would seem germane to artistic value is where earlier
art prompts emulation, adaptation, or further exploration in the same or
related directionswhere it serves as an example and inspiration to later art,
rather than merely a negative spur to it. Perhaps, then, what is defensible
is only that some actual influence-value contributes to a works value as art;
perhaps only when such actual influence-value is coupled with, or rests on, a
works potential influence-value does it do so. Still, in those cases, contribute
it does.
But suppose that, contrary to what I have just been arguing, it is really only
potential influence-value that can be held germane to artistic value, a works
particular fortunes in the subsequent history of music being discounted as
irrelevant. The existence of such influence-value would still constitute a chal-
lenge to Budds proposal. For a works potential influence-valueits power
or propensity to alter the stream of musical culture for the betterremains
something over and above the features of the work in which that power or
propensity inheres, and is distinct from and not reducible to the value, wheth-
er intrinsic or instrumental, of experiencing the work correctly.
Having said this much, influence-value may now rightly be qualified as
a secondary sort of artistic value, though a real one nevertheless. The reas-
on is as follows. Such value is clearly parasitic on primary, experience-based
artistic value, in the sense that it is value that accrues to a work in virtue
of its issuing in or paving the way for other works that have a value beyond
influence-valuepresumably, experiential value. To have led to the creation
of other works, but ones without notable experiential value, or ones with

Note also that influence-value, of either sort, might very well reside naturally in a set of
pieces, rather than any particular one of the set, though those individual pieces would then share
or participate in the influence-value of the whole. A good example would be the last ten or so of
Mozarts piano concertos.
194 Music

only influence value, would seem not to amount to artistic value of any kind.
Influence-value is thus like a promissory note that needs to be redeemed by
the exhibition of subsequent induced works of positive experiential value. But
such notes can be redeemed, and have been, time and again.
Finally, are there other varieties of artistic value possessed by individual
works of music, apart from influence-value, ones that go beyond the value,
intrinsic or instrumental, of experiencing the work comprehendingly? Here
are some candidates:
(a) Problem-solving value: part of the artistic value of a musical work
might reside in the problems it solves subject to various constraints, formal
or expressive. It is true that the way a work answers to problems set by its
predecessors might enter into how it strikes an informed ear, with awareness
of such solutions being intrinsically rewarding to sustain, but the fact of
constituting a solution to a preexisting artistic problem of some importance
would seem to be a ground of value in itself, not reducible to the value of a
listeners awareness of such a solution having been arrived at.
(b) Originality-value: part of the artistic value of a musical work might lie
in its originality or innovativeness relative to its tradition. Now although this
might again be encompassed in the experience the work offers to an informed
ear, in that originality, a backward-looking characteristic, comes across when
one holds a work up against its context of emergence, its prototypes and pre-
decessors, while listening. However, the originality of a work vis-a-vis its back-
ground, a complex relational property, is of value in a way that goes beyond
the value of appreciating such originality.
(c) Performance-value: part of the artistic value of a musical work might
be as a source of pleasure for performers in negotiating its difficulties, or as
a vehicle for a performers display of emotionality or taste. Clearly, music
might be especially enjoyable to play, or pose challenges to execution that
are exhilarating to overcome, or provide unusual opportunities for self-
expression, without being particularly rewarding to listen to. Furthermore,
some music might be said to be more valuable in virtue of allowing, more

See Stephen Davies, Musical Understanding and Musical Kinds, Journal of Aesthetics and Art
Criticism 52 (1994): 6981; Anthony Savile, The Test of Time (Oxford: Oxford University Press,
1982).
A similar brief might be lodged for constructional features of a musical work that are not
appreciatively accessible, yet are partly causally responsible for features that are. One would have to
make the case that there was value here that outstripped both that of the features made accessible
to the ear and that derived from appreciative reflection on the role played by such constructional
features.
 Evaluating Music 195

than most, for differently revealing interpretations from performance to

performanceperformances that work out differently the relation between
the musics form and content.
So it seems that there certainly are other non-appreciative-experience-based
sources of artistic value, apart from influence-value. I will not, however, try to
settle here exactly how many, or how important, those sources are.

IV

My aim in the remainder of this essay is to explore the primary, that is to

say, intrinsic-experiential, value of a piece of music for a listener. I want to
try to pinpoint what that fundamentally consists in, and what it might, most
generally, be gauged by. As the preceding discussion has made clear, I hold
that a significant part of the artistic value of a piece of music, that is, its value
as art, may be non-experiential, consisting in such things as originality, or
influentiality, or being a solution to a standing problem in a musical tradition.
In addition, I hold that the experiential value of a piece of music as art may
in principle go beyond what is intrinsically valuable in such experience, and
properly reflect extrinsic benefits of certain kinds. My target here, however,
is precisely that central dimension of musics value for listeners, namely, that
which is intrinsic to the listening experience itself, wherein such experience is
deemed worth having for its own sake.
Even if we agree that the artistic value of a work of music is primarily giv-
en by the intrinsic value of experiencing the work with understanding, our
criteria for assigning musical value are naturally apt to be considerably more
concrete than that. Asked to defend the judgment that some piece of music
was good as music, a listener will rarely, and certainly not only, submit that
experience of it is of high intrinsic value. There seems surely to be a place
for, or a role played by, various low-level criteria, such as attractive melody,
interesting rhythm, intense expression, pleasing timbres, inventive harmony,

Another sort of value that might seem to figure in artistic value may be denominated composer-
value. The composer-value of a musical work would reside in its functioning as a model for other
composers of forms, techniques, or procedures for realizing valuable artistic ends. However, such
value would appear to be pretty clearly parasitic on the value of works for listeners and performers,
the intended recipients of music. That is to say, a musical work will be valuable in virtue of guiding
and motivating other composers to further composition only insofar as it itself embodies value of
some sort for listeners or performers; otherwise, that it prompts and enables further composition
along the same lines would be, if anything, to its discredit.
Note that this might rightly be taken to include, in addition to apprehension of music as it
sounds, retrospective reflection on music after audition is complete.
196 Music

intelligible overall form, and so on. Such criteria are, of course, invoked more
frequently on the ground than is the inherent rewardingness of an under-
standing experience of music.
Yet clearly, such low-level musical merits lack generality, and fail to be
even presumptively good-making, taken by themselves. Music can be good,
and easily so, without attractive melody (for example, Stravinskys Rite of
Spring or Weberns Five Movements for String Quartet), without significant
harmonic invention (for example, much of Palestrina or Handel), without
substantial rhythmic interest (witness Saties Gymnopedie No. 1), without
pleasing timbres (witness Bachs keyboard music), without evident emotional
expressiveness (as with Conlon Nancarrows studies for player piano), and
so on.
What is interesting is whether there are any defensible intermediate prin-
ciples, so to speak, ones less abstract than that of the intrinsic reward-
ingness of the experience of listening, but not so concrete as those just
recalled. Between the most abstract condition that might be taken to ana-
lyze the core of musics artistic valuethat experience of it be intrinsically
worthwhileand the most concrete experientially satisfying merit features
appealed to in casual justification of verdicts of goodness of musice.g.
attractive melody, varied orchestrationmay lie midlevel principles that cap-
ture the ground on which such verdicts and justifications rest, while provid-
ing a specification of general validity of what experiential musical value
centrally consists in.
One approach popular in philosophical aesthetics for almost fifty years
though it has somewhat wilted of late under attacks from particularists about
evaluationis to identify general properties of works of art that invariably
conduce to or underlie artistic goodness across the artsand presumably do
so because such properties are inherently satisfying to experience, alone or in
combination, under appreciative conditions. The modern source of this tra-
dition of theorizing is Monroe Beardsley, who proposed that there were three
primary canons of criticism: unity, intensity, and complexity. The claim is
that insofar as a work exhibits one of these features it is ipso facto better as art
than it would otherwise be, and that any of its features that appear to contrib-
ute to its value as art can be shown to be forms of, or to rest on, those three
primary criteria: unity, intensity, and complexity.

These were first enunciated in his Aesthetics: Problems in the Philosophy of Criticism (New York:
Harcourt, Brace & World, 1958).
 Evaluating Music 197

But there has been significant skepticism over whether those three are always,
in all circumstances, positive-tending features; whether there arent sometimes
interactive effects among them, ones that prevent an increase in one such
dimension from always being value-enhancing; whether those three are all the
independent positive-tending features with regard to artistic goodness; wheth-
er there is any effective way to sum such criteria, so as to justify comparative
rankings of works exhibiting the three primary criteria to different degrees; and
whether the presence of such features, even in high degree, is ever sufficient by
itself to support a judgment of high comparative artistic value.
So rather than attempt rehabilitation of that approach, I want to pursue
another tack, one that tries to bring into relief the distinctive grounds of value
for individual musical works, as opposed to works of art generally, while at the
same time hewing more closely to the intuition that artistic value is centrally
a function of the worthwhileness of the experience it offers. What I have in
mind are proposals that might naturally be regarded as specifications of that
intuition as it applies in the musical case.

Consider these two approving responses to music: (a) I like how it sounds;
and (b) I like how it goes.
While neither of these responses, which strictly speaking are simply subject-
ive judgments of approval, constitutes an adequate basis for judgments of the
artistic value of a piece of music, (b) both says more than (a), and also comes
rather closer to being such a basis than does (a). And that is because (b) cap-
tures a feature fundamental to musics being good, at a level less abstract than
that of experience of it being intrinsically rewarding.
Someone offering response (a) commits himself only to approving, first,
the mere sonic appearance of the music, and second, an aspect of the music
manifestable in the smallest perceivable doses. Someone offering response (b),

Some writers, notably George Dickie, have expressed this skepticism by doubting whether
there are any strong, as opposed to weak, principles of criticism involving such criteria. Weak
principles get one only to judgments to the effect that a work has some artistic valuewhich is, of
course, compatible with its being a bad work of art. See his Evaluating Art (Philadelphia: Temple
University Press, 1988).
But see Frank Sibley, General Criteria and Reasons in Aesthetics, in J. Fisher (ed.), Essays on
Aesthetics (Philadelphia: Temple University Press, 1983); and John Bender, General but Defeasible
Reasons in Aesthetic Evaluation: The Particularist/Generalist Dispute, Journal of Aesthetics and Art
Criticism 53 (1995): 37992.
198 Music

on the other hand, has gotten closer to the musical heart of the matter, since
such a response reflects enjoyment taken in musical progression, in how music
evolves from point to point, whose most obvious manifestation is perhaps
rhythm, and the way musics parts, and especially those of small span, are
joined together from moment to moment. I like how it goes is a more telling
specification of I find experience of it intrinsically rewarding than I like
how it sounds, because the latter, strictly speaking, indicates nothing of the
essentially kinetic nature of musics basic appeal. I like how it sounds, one
may observe, would apply equally well to an imaginary art of short-duration
sound bursts, carefully synthesizer-designed for aesthetic delectation. The
appeal of the products of such an art overlaps only slightly, I would claim,
with the appeal of music, which is indicative of why how it sounds is less apt
as an epitome of what makes music good than how it goes. I like how it goes
connects with the absolutely crucial notion of following music. There is in fact
a rough equivalence between I like how it goes and I find the experience
of following it intrinsically rewarding. To take satisfaction in some music
is, above all, to enjoy following it, and its value as music is plausibly quite
centrally its enabling an experience of following its evolution over time that is
intrinsically rewarding.
Leonard Meyers central idea on the topic of musical worth is very much
in this vein. It is that music is valuable insofar as it sets up expectations in
the listener for how it will proceed, and then subsequently fulfills or frus-
trates those expectations in various ways, to varying degrees, and with more or
less delay. The most valuable music, for Meyer, provides a structure in which
there is delayed, but not indefinitely postponed, fulfillment of expectations,
the gratification that results ultimately being the more valuable for the delay
involved in attaining it.
Clearly this is a further concretization of the idea that music is valued inso-
far as one relishes how it goes, or takes satisfaction in following it, one that
gives it more specific content still. Unfortunately, and especially as it seems to
privilege the unlikely or deviant continuation, it may be too specific, too lim-
iting, to cover all ways that music may go, i.e. unfold or develop over time,
that people find intrinsically rewarding. Sometimes a continuation that just
seems right, beautiful, or cogent, may not be particularly unexpectedmay
not, in terms Meyer favors in a later formulation, impart a large amount of
information relative to alternatives.

See Some Remarks on Value and Greatness in Music.

 Evaluating Music 199

A related suggestion, in this middle terrain between the conceptually unas-

sailable and the empirically false, which concretizes further the idea of enjoy-
ing following music as the key to its central experiential artistic value, was
offered by Edmund Gurney. The idea is that music, or at least music of the
sort Meyer has labeled teleological, that is, giving an appearance of purpos-
iveness or goal-directedness, has experiential value almost entirely in virtue of
the satisfyingness of its individual parts experienced sequentially, i.e. in virtue of
the satisfyingness of its small-scale evolution or progression, from moment to
moment.
Now another such suggestion, on roughly the same level of concreteness,
would bring in, finally, that on which I have so far been silent, namely, the
expressive side of music, and locate the experiential value of music centrally in
the satisfaction of apprehending and responding to musics expressive aspect. If the
previous suggestion aims to capture most comprehensively what music offers
on a formal level, as a process uninterpreted in other than musical terms, then
the present suggestion aims to capture most comprehensively what music
offers on the level of content, or as a process interpreted in terms of human life.
For surely much of the interest of music is wrapped up in what it intimates of
human gesture, feeling, and agency.
But a principle more adequate than either of these may be arrived at by,
as it were, putting them together. More than merely enjoying following the
music in its concrete particularity or finding satisfying precisely how it goes,
and more than enjoying perceiving and responding to the gestural, affective,
and agential qualities that emerge as the music unfolds over time, what one
finds intrinsically rewarding in the experience a good piece of music offers,
and what perhaps most importantly determines its artistic value, is its very
particular wedding of its form and content. That is to say, with a good piece
of music one enjoys how it goes, to be sureits individual, temporally
evolving formand again one enjoys what it conveysthe attitudes,
emotions, qualities, actions, or events it suggestsbut above all, one enjoys

See his major work, The Power of Sound (1880).

I here make use of conveys as a blanket term for musics relation to its content, broadly
speaking, i.e. anything beyond the properties it possesses as merely a sequence of sounds. Conveys
is thus intended to cover expresses, exemplifies, represents, symbolizes, signifies, suggests,
even evokes; that is, it is being used to stand for the whole array of meaning relations music may
exhibit. Much the same, by the way, is true of expressiveness and expressive, as those terms figure
in later formulations; they too should be understood as covering dimensions of musical meaning
beyond that which can strictly be denominated musical expressiveness, as I would analyze that
notion. (See Chs. 5 and 6 of this volume.)
200 Music

and finds intrinsically rewarding the fusion of how it goes and what it conveys,
the precise way in which what it conveys is embodied in and carried by how it
goes.
If we equate the how it goes of a piece of music with its configurational
(or kinetic) form, and the what it conveys with its expressive (or interpret-
ive) content, then the what it conveys in relation to how it goes of such a
piece might reasonably be identified with its form or its content, more com-
prehensively viewedthat is, with its significant form, or immanent content.
Call this picture of things Model 1.

Model 1
(i) How it goes: how note follows note, chord chord, motive motive, phrase
phrase, and passage passage: configurational/kinetic form.
(ii) What it conveys: gesture, action, feeling, mood, emotion: expressive/
interpretive content.
(iii) What it conveys in relation to how it goes: significant form/immanent
content.
Of course, marking off configurational/kinetic form from express-
ive/interpretive content in this manner is not meant to suggest that these
are standardly distinguished as such in concrete listening experience. Usu-
ally one attends to content-infused form, or formally-embodied content, rather
than either form or content per se, though it remains possible, with effort, to
focus abstractly on such form or content as such. The progression or move-
ment of music is usually heard as both configurational and expressive, with
those aspects fused together. Indeed, the boundary between purely intra-
musical relations, motions, and tensions grasped at the level of configuration
(or kinesis) and gestural, affective, or actional contents grasped at the level of
expression (or interpretation), is perhaps essentially blurred. The above mod-
el, then, just makes explicit the two poles, so to speak, of the object on which
musical appreciation properly focuses.

Of course this is a familiar idea in the annals of aesthetics, if not put in precisely these terms.
Certainly Croce, Collingwood, and Dewey defended versions of the intimacy of form and content
in art, and Budd endorses the idea at several points in his book. A congenial development of the
idea can be found in an essay of Richard Eldridge, Form and Content: An Aesthetic Theory of Art,
British Journal of Aesthetics 25 (1985): 30316, though it is there elevated, mistakenly in my view,
into an account of the notion of art itself.
 Evaluating Music 201

What is here proposed as the main focus of musical appreciationthat is,

what music conveys vis-a-vis how it goeshelps explain, I think, why the
prevailing attitude in relation to music one loves is the desire to experience
it repeatedly and fully in actual hearing. For only through such experience,
or the simulacrum of it that mental simulation can provide, can one access
the specific fusion of human content and audible form my working formula
points up. Merely abstractly recalling the form, or even reviewing it concretely
but unrespondingly, does little to satisfy the distinctive hunger for a sorely
missed piece of music. Even less does abstractly recalling the musics content,
referenced in some manner or other, without retrieving and rehearing its
specific sonic embodiment or vehicle.
The above model, and the formula from which it is derived, is perhaps most
adequate to the experience of music on the basic or ground level, as a succes-
sion of events of relatively short duration, each with a significance of its own,
more or less absorbing. But another, more complex, model also recommends
itself to us, in which it is acknowledged that what I have designated the side
of content itself admits of formthat is to say, a how it goesat a high-
er level, which form then generates in turn a further dimension of content,
thus providing ultimately for the relationship between that content and its
underlying form. Call this Model 2.

Model 2
(i) How it goes on an expressive level: how an episode of one expressive
character follows an episode of another such character, and the pattern
of this succession as a whole: expressive form.
(ii) What it conveys in virtue of how it goes on an expressive level: dramatic
content.
(iii) What it conveys dramatically in relation to how it goes expressively:
(global) significant form/immanent content.

For simplicitys sake I am leaving out the fact that high-order, or dramatic, content depends
on and emerges out of not only high-order, or expressive, form but also large-scale configurational
form, that is, a piece of musics architectonic structure. But to acknowledge that in this model
would reduce its transparency even further. (For more on such complications, but a defense of
the primacy, nonetheless, of small-scale configurational form, see my Music in the Moment (Ithaca:
Cornell University Press, 1998).)
Dramatic is not an ideal name for all varieties of global expressive content, but dramatic
here, like conveys and expressive earlier, is intended broadly, covering all sorts of content of a
global sort that music may convey, in virtue of its overall form and its pattern of expressiveness as
a whole.
202 Music

To complete the picture we would need to recognize that relationships

between higher-order forms and contents and those on the ground level
might also enter into the appreciation of music, that is, serve as potential
foci of intrinsically rewarding attention. Thus, if from Model 1 we have
three elements, configurational form, expressive content, and their resultant
low-order significant form/immanent content, and from Model 2 three
elements as well, expressive form, dramatic content, and their resultant
high-order significant form/immanent content, then in theory there are
another nine relationships that become candidates for musical apprehension
or contemplation, and hence, loci of evaluative assessment. Of those, the
relationships between expressive and configurational form, and between
dramatic and expressive content, would seem the most relevant. But at this
stage of reflection we surely begin to lose our grip on what the experiential
value of music mainly consists in. Thus I think we may safely ignore any
further relationships of this sort, and rest with the two models already
articulated, in which the central object of musical appreciation is identified
as the relationship of content to form, on lower and higher levels, with
such form and content themselves functioning, as well, as secondary and
supporting objects of attention.

VI

Confining our sights to Model 1, the simpler and more fundamental of

our two models, it might seem that there is an asymmetry between the
component labeled how it goes (configurational form) and that labeled
what it conveys (expressive content), as potential loci of musical value.
For the formermusical formappears to have an independent claim
on listeners, that is, it is of interest to apprehend on its own, while the
latterextramusical contentmay appear, taken by itself, to lack any such
claim. Budd, for example, adopts such a viewpoint: our experience of the
expressive aspect of music is not separable from our experience of the music,
and we value the expressive aspect not in itself but as realized in the music. In
other words, that a work of music has a configurational form that is absorbing,
expressive meaning aside, is already a ground of musical value, such form
being an appropriate, if limited, object of musical appreciation; but not so,

Values of Art, 152.

 Evaluating Music 203

it seems, that a work has this or that expressive content, identified apart from
the musical form that realizes it.
But is this really so? I think not. What is true is that there is an asym-
metry in degree in that a works expressive meaning abstractly identified is
of less artistic interest, and less contributory to artistic value, than the quality
of its configurational form or the manner in which the expressiveness emerges
from said form, but it is false that such meaning, considered apart from its
vehicle, is of no such interest, or that its nature has no bearing on the artistic
value of a work. Unusual, rare, subtle, deep, profound targets of expression, if
attained, make a musical work better, because such contents are themselves
more rewarding to contemplate, engage with, or respond to. For instance,
a musical work expressive of bittersweet melancholy, or communicating the
gestures of resignation, is arguably artistically more valuable, all things being
equal, than one expressing simple cheerfulness or conveying an ordinary sort
of anger.
Recall that for Budd, the question about a musical works artistic value
is simply this: Is the experience the work offers intrinsically valuable? The
question I have highlighted, by contrast, is rather: What is the nature of the
experience musical works typically offer, so far as the intrinsic value of such
experience is concerned, and what more specifically are the dimensions of
the experience in which such value resides? I have ventured to generalize about
the value-relevant dimensions of the experience of music, and have concluded
that they are irreducibly three: experience of configurational or kinetic form,
experience of expressive or interpretive content, and experience of the embod-
iment or realization of the latter in the former. So if the experience a work
offers possesses intrinsic value, this will be because the work is found inher-
ently rewarding expressively, or configurationally, or in terms of its fusion
of the expressive and the configurational, or in all three ways. Thus we can,
if we like, now identify three prima facie criteria of experiential goodness in
music: one, how rewarding it is to experience how the music goes, that is,
how rewarding it is to follow as tonal process; how rewarding it is to register
or respond to what it conveys; and how rewarding it is to experience what it
conveys in relation to, or as embodied in, how it goes.
Does it then follow that a work that rates highly in all three dimensions of
the experience it affords is thus necessarily better than a work that rates highly
in only two such dimensions? For instance, that a work with a distinctive
representational content that exhibits a musically absorbing form and weds
the two in a successful manner is necessarily better than a work with a more
204 Music

commonplace expressive content exhibiting musically absorbing form suc-

cessfully wed to that content?
The answer, I believe, is that it does not follow. This is because of the
possibility, even likelihood, of interaction or interference between the dimen-
sions of such experience. For example, it may typically be the case that works
of markedly representational character, even when musically absorbing and
exhibiting good integration of the representational and configurational, do
not afford an experience on the whole as inherently rewarding as much music
of nonrepresentational and only middling expressive character; the value of
the appreciative experience of such a work is not the sum of the values of the
three dimensions of appreciation considered in isolation, because in trying
to realize them all in a single appreciative experience one finds that attention
to the representational dimension harmfully competes with attention to the
configurational, more so than when music has extramusical but not represent-
ational content.
Still, though the value of musical experience will not, if this is right, be
simply the sum of the value of the three dimensions of such experience that
I have identified, those dimensions would seem to be the ones in which all
such value resides, and the satisfyingness in those dimensions taken separately
the only prima facie reasons to regard a work as musically good.

VII

How do the formulas and models of experiential value proposed above fare
when held up against a piece of unquestionable and surpassing musical worth?
By way of conclusion I choose, for brief examination in this light, Schuberts
Piano Sonata in A major, op. post., D. 959. I suggest that this sonatas exper-
iential value, that is, its central value as music, fits comfortably under the
umbrella provided by those formulas and models. For, in short, its goodness
as music is a matter of how satisfyingly it goes as a purely musical process, of
how satisfying it is to engage with the content it conveys, and above all, of
how satisfying it is to experience the way in which what it conveys is embed-
ded in, intertwined with, and borne by how it very precisely goes.
But let us try to identify, in regard to this particular musical composition,
some of what, in particular, makes it so good. Now the schematic midlevel
answer to the question Why is Schuberts sonata musically so good? is that
the experience it affords, on a configurational level, on an expressive level, and
on the level of their interrelation, is so inherently satisfying or worthwhile. On
 Evaluating Music 205

the purely kinetic level, it is highly absorbing to follow, in its small-scale and
large-scale movement; from beginning to end the ear is regaled with beautiful,
cogent, and original forms, or ways of sounding in time. On the level of con-
tent, what it expresses or conveys is distinctive, intense, and well worth enga-
ging with. And on the level of fusion, the specific manner in which the work
musically conveys its content seems wholly compelling, exhibiting through-
out an elegance and rightness of means to ends.
Still, it will not be amiss to go into somewhat more detail in regard to the
expressive and dramatic dimensions of Schuberts musical essay. I begin with
the former.
This A major sonata possesses an intensity and variety of emotional expres-
sion quite out of the ordinary, covering almost the full range of human feel-
ing. It displays perhaps the greatest range of moods and affects of any of
Schuberts works, yet manages to tie them all together into such a satisfying
whole that listening to it is like living a human life in microcosm.
What are salient loci of expressive goodness in this sonata? The following
are four instances of distinctive expression, distinctively achieved, in most
cases through music that is, in addition, itself entirely satisfying on a
configurational level:
(1) The uneasy stasis, suggestive of restless anxiety or obsession, at the begin-
ning of the first movements development, achieved through harmonic
oscillation between C major and B majoras opposed to the more usual
nonreversing journey from key to key.
(2) The unique nostalgia of the first movements coda, in which the move-
ments opening rhetorical theme is recollected as if through a hazeits
vital, almost peremptory force drained off, but its essential identity intact.
(3) The unparalleled violence, approaching chaos, of the slow movements
middle section, a remarkable evocation of someone going desperately out
of control.
(4) The charming perkiness of the scherzo movements main theme, which
charm is based in part on the piano writings suggestion of string pizzicati.

The means here would include the configurational form basis of the musics local expressiveness,
and the expressive form basis of the musics dramatic content, or global expressiveness. Some
expressive forms, or patterns of expressiveness, are less satisfying than others, whether as forms per
se, or in virtue of the sort of higher-order content they generate or fail to generate.
The specific image of a man distraught to the point of tearing out his hair is, for me, almost
inescapable.
206 Music

I turn now to dramatic, or global expressive, content. A striking example is

afforded by the intense, soul-searching dialogue in which the musics persona
is engaged at the climax of the second episode of the sonatas rondo finale,
at measures 180-210. In this passage the musics persona seems to pose, then
reluctantly answer, questions it would rather not face, but that can no longer
be avoided. And a sympathetic listener cannot help but feel they are the life
and death questions of his or her own existence as well. As for the source,
formally speaking, of the passages dialogic quality, it is the stark alternation
of treble and bass in the sounding of the passages sharply etched, individually
expressive motifs.
Another example of global expressiveness in this sonata is something almost
impossible to put into words: it is reflected in the impression a listener receives
during its first movement, and perhaps most pointedly, its opening statement,
of being addressed by an adult, and as an adult, of reflectively and unhurriedly
being given the benefit of someones wisdom and maturity. This is one of sev-
eral quite singular attitudes of mind somehow communicated by this sonata
in the large, attitudes that give it the sort of moral force I earlier speculated
may belong to certain pieces of music.
A final example of global expressiveness in this sonata concerns a special sort
of unity that some pieces of music attain. One kind of most valuable music is
that which displays unity in such manner and degree as to stand as a power-
ful emblem of the unification of opposites and the reconciling of the diverse,
not through the subduing or overpowering of one element by another, but
through the evincing of a deeper relatedness despite superficial differences.
Call unity of this sort transcendent unity, in contrast to the merely formal
unity that is, in most cases, its substantial underpinning. Such emblemati-
city of wholeness as I have in mind is not just an abstract relationship that one
may reflect on intellectually, but rather something one feels or registers in the
course of listening.
Schuberts sonata displays that sort of unity, if any piece of music does.
There is a striking thematic, rhythmic, and harmonic unification effected in
subtle ways throughout the sonata, despite the rich surface variety of its four
movements, which helps to generate the pronounced transcendent unity the
piece evinces. Here I will just note some aspects of the sonatas formal unific-
ation, ones that seem to play some part in its achievement of the other sort of
unification.
 Evaluating Music 207

The opening six bars of the first movement are echoed, in loose retrograde,
by the closing eight bars of the fourth and last movement; the first move-
ments development section is derived almost entirely from the opening figure
of the movements second theme; there is a clear reminiscence of the agita-
tion of the slow movements middle section in the interpolated repeated-note
triplet figures in the reprise of the slow movements main theme; the tensions
resident in the first and second movements reappear in the second half of
the scherzo movements main section, employing a variant of the figure from
the first movements development; the opening motif of the trio section of the
scherzo is a loose inversion of the first movements opening motif; rising or
falling semitone motion is an important element in the main themes of the
first, second, and fourth movements, and includes the mysterious alternation
of A major and B-flat arpeggios at the very end of the first movement; finally,
there is a clear locus of harmonic gravitation in the sonata constituted by keys
a third up from the tonic. Of course elements of formal unity in a music-
al composition, even when widespread, do not inevitably issue in the sort of
global expressiveness I have labeled transcendent unity, whose ultimate mark
is a sense of transfiguring oneness in the listener, but I submit that in this case,
they do.
It would be quixotic to think that this incomplete and selective survey
of loci of goodness in one very fine musical composition can conclusively
establish the validity of the schematic midlevel formulas proposed earlier for
that in which musical value residesto wit, configurational, expressive, and
expressive/configurational satisfyingness, or at a higher level, expressive form,
dramatic content, and dramatic content/expressive form satisfyingnessbut
at least it has turned up nothing that fails to fall handily under them.
In any event, it is only through such midlevel formulas, I think, that we
may perspicuously bring together on this subject the view from above, in
which musical value is understood in terms of the intrinsic rewardingness
of experience of a work, and the view from below, in which musical value
is seen in terms of familiar featuresones of sound, melody, rhythm, or
mooddirectly cited by listeners as grounds for approving a work.
This page intentionally left blank
 11
Musical Thinking

If a lion made music, would we get what he played?

(pseudo-Wittgenstein)

1. It has sometimes been remarked that making musicthat is, compos-

ing, performing, or improvising itinvolves thought, or is a form of thought.
If so, what is the nature of the thinking that goes on in making music? And
what of listening to music? Is the experience of the comprehending listener
also a kind of thinking? How does musical thinking differ from the paradigm
of thinking, that is, the formulation and manipulation of thoughts in words?
Can musical sequence itself, rather than the activity of producing or auditing
it, be regarded as a kind of thinking? In short, is music thought?
In the course of trying to shed light on these issues I will take as a spring-
board various remarks of Wittgenstein on music that are to be found here and
there in his writings. I will also yield to the temptation to emulate, in a small
degree, Wittgensteins elliptical, oracular manner, a manner particularly apt
to the exploratory stages of a philosophical investigation, which is certainly
the case here. Whether what results should be considered an homage, a par-
ody, or some mixture of the two, I leave to my readers to decide.
2. It seems clear from a number of Wittgensteins remarks, especially ones
directed to particular composers, that he was indeed inclined to regard music
as thinking. In one place we find the following invocation: The strength of
the thoughts in Brahmss music (CV, p. 23). In another place we are told that
one can point to particular places in a tune by Schubert and say: look, that is
the point of the tune, this is where the thought comes to a head (CV, p. 47).

First published in Midwest Studies 27 (2003): 5968.

The following abbreviations are used for the citations in the text: PI = Philosophical Invest-
igations, 3rd edn. (New York: Macmillan, 1953); BBB = The Blue and Brown Books, 2nd edn.
(Oxford: Blackwell, 1958); CV = Culture and Value (Chicago: University of Chicago Press, 1980).
210 Music

What is most striking about these observations is how natural it seems for
Wittgenstein to think of music as a kind of thinking, how little in need of
defense he appears to take that view to be. What if one invoked, by contrast,
The strength of the thoughts in the cuisine of les freres Troisgros, or The
strength of the thoughts in Michael Jordans basketball playing? Would this
seem as natural? Could we easily speak of a moment in Jordans progress to
the basket, or of a dish in a ten-course meal at Troisgros, where the thought
comes to a head? I suggest not.
3. In the Investigations and elsewhere, Wittgenstein remarks that one
might describe the effect of a passage of music by saying Here it is as if a
conclusion were being drawn (PI, p.182)
There are a couple of things to note about this. First, Wittgenstein does not
say that, in such a passage, a conclusion is being drawn; rather, it is as if a con-
clusion were being drawn. So far, then, we are in the realm of analogy or meta-
phor, or perhaps of the dawning of an aspect. Second, the character of some
passages of music to which Wittgenstein is calling attention is specifically that
of seeming to draw or reach a conclusion, as after a period of reflection; it
is not the idea of merely concluding, in the sense of stopping or terminating.
Compare the endings of Beethovens Piano Sonata op. 110 or Dvoraks Sev-
enth Symphony, which seem to sum up and crystallize what has gone before,
with the endings of, say, minuet movements from symphonies of the Classical
period, even great ones such as Mozarts Fortieth or Forty-first. The former
have this special rhetorical character of concluding, whereas the latter have
only the mundane character of coming to a closehowever satisfyingly.
4. Wittgenstein treats the phenomenon further in another place: If I say,
for instance: here its as though a conclusion were being drawn, here as though
someone were expressing agreement, or as though this were a reply to what
came beforemy understanding of it presupposes my familiarity with con-
clusions, expressions of agreement, replies (CV, p. 52).
What Wittgenstein is underscoring here about the appreciation of music
is this. Music is not understood in a vacuum, as a pure structure of sounds
fallen from the stars, one which we receive via some pure faculty of musical
perception. Music is rather inextricably embedded in our form of life, a form
of life that is, as it happens, essentially linguistic. Thus music is necessarily
apprehended, at least in part, in terms of the language and linguistic practices
that define us and our world.
But by the same token, should we not expect that our understanding
of linguistic phenomena will sometimes be inflected by our musical
 Musical Thinking 211

understanding, especially in light of the fact that our musical capacities are
awakened at least as early as our linguistic ones? For example, we may describe
certain speech as sing-songy, a conversation as not having the right rhythm,
and the papers at a conference as not harmonizing. Furthermore, in tonal
languages, such as Japanese or Indonesian, the distinction between speaking
and singing is to some extent effaced. Though language may be essential to
the human form of lifewhereas music, though universal, arguably is not,
since we can presumably imagine human life without music, but not without
languageonce both are present their interpenetration is assured, and we
cannot help interpreting the one in terms rooted in the other.
5. It is true that the question What are you thinking? most often elicits
a verbal answer, such as Its going to rain, or I need to buy milk soon, or
She is very attractive. But why not, on some occasion, a musical phrase, or
even a particular rendition of a musical phrase? If someone asks me what I was
thinking, can I not sometimes truthfully say the opening of Mendelssohns
Violin Concerto? Could I not, in response, even whistle that opening, and
in a particular way? Note that the former response would not be the same
as saying I was thinking of the opening of Mendelssohns Violin Concerto.
For of course anything might be an object of thought. But that doesnt make
it into an example of thought. No, the Mendelssohn opening is what I was
thinking, not what I was thinking of.
6. We say of some music that thought went into it, or there was
thought behind it, and mean to contrast that with cases we might
rather describe as thoughtless note-spinning. Is the distinction between
thoughtful and thoughtless musicor thought-filled and thought-free
musiccoincident with that between good and bad music? If not, it is
probably not too far removed from it.
When we estimate the quality of music we often refer to the mind that is
revealed in it, the mind one comes into contact with in listening to it, the
mind that is reflected in it, and so on. Granted that there is more to mind
than thinking, can there be less to mind than that? If not, then can we easily
deny the label of thinking to music of any worth, given the mind that stands
before us in sound when such music is played?
7. Wittgenstein remarks in several places that it is common to experience
a musical phrase as a question. It is also not uncommon to experience anoth-
er phrase as an answer. And experiencing music in such ways seems part of
what it is to understand music. Note that this is a matter of phrases striking
us as questions and answers on more than the purely musical plane; that is,
212 Music

we are here speaking of more than the sense in which one phrase can serve as
musically the answer to another, in terms of completing its melodic arch or
balancing its harmonic movement.
Now, if two phrases of music strike us as having a more-than-musical
question-and-answer relation, must there be a content to the question that
the first phrase seems to embody? In other words, must it be possible to say
what exactly the first phrase is asking? If not, then what does the claim that
the phrases have a more-than-musical question-and-answer relation amount
to? Perhaps just that they convey the character or physiognomy of questio-
ning-and-answering, though without constituting a specific question-and-
answer.
8. In The Brown Book Wittgenstein observes:
. . . if repeating a tune to ourselves and letting it make its full impression
on us we say, This tune says something, and it is as though I had to find what
it says. And yet I know that it doesnt say anything such that I might express in
words or pictures what it says (BBB, p. 166).
As we have already noted, very often music makes on us the impression of a
communicative act, and more specifically, a speech act or utterance. There is
nothing more common than the sense that expressive music is speaking to us,
and though the embedded claim is perhaps not to be taken literally, neither is
it merely a weak metaphor whose cash value would be simply that the music
seems meaningful, or that one gets something out of it. As Wittgenstein says,
the impression of speech from music is so strong that we often feel impelled,
however misguidedly, to try to ascertain exactly what is being said. And we are
not satisfied, it seems, unless we can exhibit what is said in other than musical
terms, and preferably verbal ones.
But should we be thus dissatisfied? Is there a communicative medium that
should be privileged above all others which help to constitute the lived world?
One is reminded of the anecdote in which Beethoven, having played for some
visitors his latest piano sonata, was asked, But what does it mean, Herr
Beethoven?, to which his response was just to play the sonata over again.
To require that musical thought, if it is to truly deserve that label, must be
such that it can be rendered articulate or verbally paraphrased, would seem to
smack of a double-standard. It would not impugn the claim to being thought
of a stretch of discourse to note that what it conveyed could not, so far as
we could see, be put into music. Why, then, should it be held to impugn the
claim to being thought of a stretch of music that what it conveys cannot, in
general, be put into words?
 Musical Thinking 213

9. It is instructive to draw up an illustrated catalog of thoughtful actions

that we can hear in musical passages, or that we can hear musical passages as
instantiating: asserting (e.g. the opening of Schuberts Piano Trio, Op. 100
No. 2), questioning (e.g. the opening phrases of Beethovens Piano Sonata
No. 18, Op. 31 No. 3), musing (e.g. Schumanns Des Abends), imploring
(e.g. the flute introduction to Bellinis aria Casta Diva), angrily despairing
(e.g. the opening of Mahlers Second Symphony), menacing (e.g. the opening
of Stravinskys Symphony in Three Movements), defying (e.g. the opening
of Beethovens Fifth Symphony), cajoling (e.g. the sixth part of Vaughan
Williams Job: A Masque for Dancing), comforting (e.g. the first section
in moderate tempo near the beginning of Faures Requiem), disapproving
(e.g. the orchestral interjections in the first part of the finale of Beethovens
Ninth Symphony), and even nose-thumbing (e.g. the opening of the finale
of Beethovens Second Symphony). And there are passages in which can be
heard meditating, applauding, bemoaning, heaven-storming, and so on. Can
a medium capable of summoning up such a range of mindful actions be a
domain in which thought is absent?
10. So one kind of musical thought is this: musical passages wearing an
appearance of thoughtful acts, such as questioning, concluding, searching, and
the like. But another kind, surely, is this: musical passages giving evidence of
thought processes in their creator.
Let us bring the contrast between these two senses of musical thinking into
clearer relief. One is thinking that seems as if it is embodied in musical process,
that is, thinking that the music itself strikes us as being engaged in, or perhaps,
that we are induced to imagine that the music is engaged in. Two is think-
ing in the composer that we take to be implied by musical process, that is,
thinking that the music betokens on the composers part. We might even go
so far as to say, as Wittgenstein would urge us to do, that we directly hear the
composers thought in the musical process. For we are confronted with com-
positional choices at every turn that we cannot but regard as manifestations of
mind.
Some examples of the first kind of musical thinking, of which we had a
number of illustrations earlier, would be where music seems to be embarked
on reflection, or to be lost in wonder, or where one musical phrase seems to
answer the question posed by a preceding one. Examples of the second kind
of musical thinking would be the assessment we infer Bach must have made
in devising a fugue theme combinable with itself in counterpoint, or the judg-
ment we suppose Mozart to have exercised, in composing a piano sonata, in
214 Music

designing a second theme whose character would contrast suitably with that
of the first theme, or the vision we understand Beethoven to have displayed in
opting for a C-sharp rather than a C-natural in the fourth bar of the opening
theme of the Eroica Symphony, setting up a tension exploited significantly
later in the movement.
Again: embodied thinking in music is thinking we ascribe to the music, as
something it appears to be doing, and has no identifiable object, whereas
implied thinking in regard to music is thinking we ascribe to the composer,
and has a quite definite object, namely the evolving composition itself.
11. Yet possibly the most important way in which music is a kind of
thought does not reside either in musics frequent suggestions of thoughtful
actions, in or its implications of thoughtful fashioning on the composers part.
It may reside instead in the mere succession from chord to chord, motive to
motive, or phrase to phrase at every point in any intelligible piece of music,
whether or not there is any suggestion of recognizable extramusical action,
or any implication of specific compositional deliberation. Call such musical
thinking intrinsic musical thinking.
But why call such succession thinking? Obviously this is not enjoined by
any rule of language. Still, musical succession has features that set it apart
from succession in general. It is a purposive-seeming, goal-directed temporal
process, an intelligent form of continuation in time, and one naturally subject
to assessment in cognitive terms, such as coherence or logicality or mak-
ing sense. In addition, it is succession that we know emanated from a human
mind, and that we hear under the influence of that postulate. If one insists
that that is not enough for thinking, is one not just assuming that thinking is
necessarily in words? And why should one assume that? Of course, if music be
admitted to be thinking on the grounds just offered, the door is also open for
dance, mime, and abstract film to be considered thinking as well. But such an
implication is not, I think, to be feared.
12. Let me attempt to trace the process of embodied thought in the first
movement of Beethovens Tempest Sonata, Op. 31 No. 2, one of the most
rhetorical pieces of music in all Beethovens oeuvre.
The movements opening gesture, a four-note rising motif in largo tempo
beginning with an arpeggiated A major chord, has about it a pronounced air
of uncertainty and wonder (ms. 12). It is followed by a descending allegro

The details of my account follow in broad measure the analysis given in Charles Rosen,
Beethovens Piano Sonatas: A Short Companion (New Haven: Yale University Press, 2001), 16870.
 Musical Thinking 215

motif in D minor which anxiously frets, ending in an adagio turn of ques-

tioning character (ms. 26). Next the largo motif returns, to be followed
by a more excited variant of the earlier allegro D minor music (ms.716),
whereupon the music gathers resolve in a passage in octaves (ms. 1718),
before issuing in a full cadence on the tonic and the first episode of pure
affirmation, a declaration in the bass and in allegro tempo of the opening
four-note rising motif, now wearing a minatory cast, but rounded off in the
treble by a new motif, plaintive and supplicating.
Move now to the beginning of the development, where the rhetorical char-
acter of the movement becomes even more pronounced. The opening four-
note largo motif returns, with its initial arpeggiated chord extended, and is
heard three times, each time outlining a different chord, ever more removed
from the tonic (ms. 938). It is hard not to hear this sequence of returns as
a deepening of the uncertainty and wonder expressed by the largo motif on
its first appearance, and the agitated minor key music that succeeds it as an
exacerbation of the minatory proclamations of the exposition (ms. 99118).
Consider, finally, an episode at the end of the development prior to the recapi-
tulation, where the music takes on even more unmistakably than before the
appearance of a mindful agent. After six measures of sustained chords and a
descending passage in bare octaves (ms. 13342), which strike one as clearing
a space for reflection, there follows a recitative with all the earmarks of a solilo-
quy, in part due to the thinning out of the musical texture and the starkness of
the melodic line that remains (ms. 14358). This utterance is at turns medit-
ative, questioning, and anxiousthe last of these due to an eruption halfway
through of the fretting motif from the sonatas beginning, though in muted
form.
As one attends closely to this movement, one cannot fail to be struck by
the mind manifested in its progression, and more specifically, by the series
of communicative acts incarnated in the music itself. Yet whose mind is so
manifested, one may ask? In one sense, it is the mind of the imagined agent
of those acts, what one may call the persona of the music. In another sense,
it is the mind of the composer, who has in effect constructed, or caused to
emerge, the persona to whom the communicative acts heard can be directly
attributed.

This is a good point at which to note that the opening arpeggiated chord of the movement,
which recurs in various guises, is a sixthree chord, the type of chord that typically introduces
recitative in opera, and which thus might be said to adumbrate the recitative that occurs later in the
middle of the movement.
216 Music

But at this point an objection may be raised. Let us grant that the music
of the first movement of the Tempest Sonata exhibits a series of images of
thought-filled actions. Does that show that the movement literally consti-
tutes thought, or is literally a thought process? Here we must add that this
movement of the Tempest Sonata is, of course, not a random or accident-
al concatenation of such images. It is, rather, a meaningful concatenation of
them, one that makes sense to us, and one that induces us to imagine a mind-
ful agent of those acts, of whose mental life the music then appears as the
narrative. However, the objector continues, all that shows is that music, or
at least some music, is the narrative of an imaginary thought process, not a
thought process itself. But how much of a difference is that?
13. Consider now, in a change of gears, how the sorts of thinking
involved in the activities of composing, performing, improvising, and
listening to music saliently differ. One way they differ is this: following is a
key idea in the last of theselisteningbut not the others. To understand
music to which one is listening is, at bottom, to follow it, that is, to experience
its evolution in an involved way, exercising certain perceptual abilities and
emotional sympathies, anticipating and projecting that evolution, responding
appropriately in the moment to each twist and turn. That following
musicas opposed to mere listening, or half-listeningis a form of thinking
is evidenced by the near impossibility of doing any other thinking, of an
unequivocal sort, at the same time. Musical process absorbs and effectively
fills the mind that attends to it with any seriousness.
But following is not, it seems, of the essence of composing, performing,
or improvising. Rather, determining that is, the determining of notes as
constitutive of a workwould seem to be the essential activity of composing;
interpreting that is, the interpreting in concrete sound of notes already
giventhe essential activity of performing; and generating that is, the
creating of music on the spot, subject only to relatively loose constraints, the
essential activity of improvising.
Now, on the one hand, these activities of determining, interpreting, and
generating music might all be classed as productive, whereas that of following
music might by contrast be classed as receptive, though that should not make
one lose sight of the anticipatory and constructive element in the activity of
following music by ear. On the other hand, the activities classed as productive

For further description of the activity of following as a core component of the appreciation of
music, see my Music in the Moment (Ithaca: Cornell University Press, 1998).
 Musical Thinking 217

in a sense also involve actions of following: the composer follows one meas-
ure with another as he composes, the performer follows his reading of one
phrase with his reading of the next, the improviser follows what he has just
played by playing something else. But those sorts of following are manifestly
not the same as that involved in listening. In the one case what is central is the
tracking of what already exists, whereas in the other case what is central is a
bringing into being at each step.
14. As an illustration of the thinking involved in improvising, I turn to
Stan Getzs solo on his famous recording of Antonio Carlos Jobims The Girl
from Ipanema.
Knowing Jobims basic tune, and hearing Getzs treatment of it, we marvel
at the melodic, rhythmic, harmonic, and timbral possibilities that Getz brings
out that we didnt suspect the tune possessed. I would single out for remark
just the high melodic leap Getz takes about midway through the repeat
of the first strainwhere the lyric, significantly, has the words sways so
gentleand the playful one-long three-short rhythm he introduces in the
refrain in place of the original dotted one. If this is not thinking in sound,
then what is it? Surely music that in the span of a mere forty bars manages to
suggest a whole way of beingfor my part, I have often wished to live some
of the time as that solo soundscannot be music in which thought fails to be
present.
Note that with improvised music a distinction invoked earlierbetween
embodied thinking in music, wherein music presents us with images of
thoughtful actions of an imaginary persona, and implied thinking in music,
whereby musical process betokens or signifies thought in the composerhas
almost no purchase. Is it Getz or is it the musics persona who exults in that
high-flying turn mentioned a moment ago? Is it Getz or is it the musics
persona who gets down in that one-long three-short rhythm? And is there
much importance to deciding?
Music, we may note, and improvised music especially, stands as one of the
supreme exemplars of the fusion of inner and outer in mental life that Wit-
tgenstein was at pains to underline, a fusion that in the last analysis invites
us to transcend the opposition between inner and outer, a transcendence that
was perhaps the ultimate goal of Wittgensteins philosophy of mind.
15. If music is in some sense thinking, then, as noted earlier, bad music
should tend to have the character of bad thinking. One of the ways we mark
out bad thinking is by the epithets dumb and stupid. Hence, if music is
thinking, at least some bad music should wear the appearance of dumbness
218 Music

or stupidity. One clear example, I would say, from an early work of a great
composer, Franz Schubert, is the beginning of the finale of his String Quartet
No. 7 in D.
Now why is this music so stupid, and thus bad? The problem, zeroing in
on just the opening six bars, which are even repeated, is that it consists in two
largely unrelated ideas, and more specifically, of a first idea of utter banality
followed by a second idea which is an emphatic closing gesture, one entirely
unjustified by the meager four-bar ditty that precedes it.
Compare this, though, to a little known piano sonata by Beethoven, the
Sonata No. 16, Op. 31 No. 1, in G. This music once also struck me as some-
what stupid, because of its quirky premature use of a closing gesture similar
to that featured in the Schubert, and the similarly unpromising character of
its melodic materials. On longer acquaintance, though, the music seems any-
thing but stupid. The Beethoven is, despite its modest materials, miles beyond
the Schubert in development, flow, and organicity.
Of course, not all bad music is stupid, or bad because it is stupid. Some
music is bad because it is bland, or bombastic, or bathetic, or lacks balance
just to stick with b words. This is to be expected, since even if music is think-
ing, and thus sometimes bad in the way thinking is generally, music is oth-
er things as wellmovement, gesture, pattern, expression, narration, depic-
tionand can thus exhibit failures in those respects, and not just fallings-off
from the cardinal virtues of thought, such as cleverness or cogency.
16. Finally, what is the connection between Wittgensteins views on
understanding musicthat it is manifested by a complex of behaviors, such
as illustrative gestures, apt comparisons, suitable hummings, and appropriate
movings to music, that its criteria are neither inner acts of comprehension
nor articulate paraphrases of musical content but a range of outwardly
demonstrable responses and capacitiesand the claim that Wittgenstein also
endorsed, which has been my focus here, that music, no less than language,
incarnates thought?
Here is one way to articulate the connection or connect the dots: Both
music and language are forms of thought. Understanding music should there-
fore be analogous to understanding language. The former, like the latter, is a
matter of use, that is, of knowing how to operate with the medium in question
in particular communicative games, in particular contexts. But knowing how

I here use the expression communicative games, instead of language games, to avoid
privileging language over other forms of communication or meaning-making.
 Musical Thinking 219

in regard to music, as with knowing how generally, does not consist in pro-
positional knowledge but rather in behavioral and experiential abilities and
dispositions. Hence if music is thought we should naturally come to under-
stand it as we come to understand thought in words; not by learning how to
decode or decipher it, but by learning how to respond to it appropriately and
how to connect it to our lives.
17. Parallel to the question at the heart of this essayIs music
thought?would be the question, Is speech thought? In other words, one
might wonder whether a stretch of intelligible verbal discourse was literally
thinking, or was instead only the expression of literal thinking, that is,
certain occurrences or processes in the mind of the speaker. Wittgenstein, of
course, argued that there is no reason to think of thinking as a purely inner
process, of which our observable behavior, however intelligent, can be no
more than the outer shell, and thus no reason not to recognize as thinking
the normal deployment of language. But for those who balk at the idea even
that intelligible verbal discourse is thought, that is, that thought has outer
as well as inner forms, and who claim that such discourse only manifests
thought, the central claim of this essay can be suitably recast. It becomes this:
intelligible music stands to literal thinking in precisely the same relation as
does intelligible verbal discourse. If that relation is not exemplification but
instead, say, expression, then music and language are, at any rate, in the same,
and quite comfortable, boat.
 12
Musical Chills

1. Although the value of music goes beyond, in a number of respects, the

pleasure it provides when appropriately attended to, the capacity to provide
such pleasure is clearly a significant part of that value. To be sure, there are
values that attach to both individual pieces of music and music as a whole that
do not cash out in terms of pleasures afforded listeners, but such values are not
the focus of my present reflections.
When I speak of pleasures properly afforded a listener by music, a particular
sort of attention to music should be understood to be involved. First, such
attention is close and concerted. Second, such attention is locally focused,
though global-context-sensitive. Third, such attention is aesthetic, or appro-
priate to music as art; in other words, attention carries to the role played by
what is precisely heard in the music in the generation of any pleasure that res-
ults, so that the music serves, at least in part, as the object of such pleasure.
Finally, such attention is stylistically and historically informed, if only on a
tacit level, thus allowing each musical characteristic, whether formal, express-
ive, rhetorical, or representational, to be registered for what it is.
2. I recall here some obvious dimensions of difference among musical pleas-
ures. Musical pleasures differ in how active they are, how intellectual they
are, and how essentially physiological they are. Musical pleasures differ with

Earlier versions of this essay were Musical Frissons, Revue Francaise dEtudes Americaines 86
(2000): 6476; and Musical Chills and Other Delights of Music, in J. Davidson (ed.), The Music
Practitioner (Aldershot: Ashgate Publishing, 2004), 33551.
For both sides of the issue, see my essays Pleasure and the Value of Works of Art, in The
Pleasures of Aesthetics, and Evaluating Music (Ch. 10, this volume).
For a defense, inspired by Edmund Gurney, of the primacy of local focus in the comprehension
of music by ear see my Music in the Moment (Ithaca: Cornell University Press, 1998).
For more on this aspect of aesthetic attention, see my What Is Aesthetic Pleasure?, in The
Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996).
For further admonitions in this vein, see my Musical Literacy, in Pleasures of Aesthetics, and
Stephen Daviess Musical Understanding and Musical Kinds, Journal of Aesthetics and Art Criticism
51 (1993): 6981.
 Musical Chills 221

respect to intensity, e.g. they may be acute or mild; with respect to duration,
e.g. they may be passing or long-lasting; with respect to durability, e.g. they
may be one-shot affairs or eminently repeatable; with respect to communic-
ability, e.g. they may be highly esoteric or widely shared. Musical pleasures
arguably differ in the different values thereof, e.g. elevated or trifling, in the
different moral qualities thereof, e.g. humanizing or dehumanizing, and in
their different social imports, e.g. solidarizing or exclusionary.
Musical pleasures differ, further, in when they are taken, so to speak, relat-
ive to when the music to which they refer is heard. That is to say, though most
musical pleasure, I would claim, arises from the real-time following of music
in its formal and expressive evolution, some musical pleasure is anticipatory,
preceding audition, and some musical pleasure is recollective, occurring after
audition.
One dimension of difference among musical pleasures corresponds to two
contrasting modes of listening, or perhaps two contrasting stances towards the
listening that is going on. On the one hand one may, without losing contact
with the music in its full particularity, let a piece of music enfold one, envelop
one, wash over one, so that one gives oneself over to it in a personal way, as to
a lover, or perhaps a trusted therapist. On the other hand one may undertake
to keep music at a distance, so to speak, observing its lapidary details, its emo-
tional maneuverings, its dramatic gestures as something external to and apart
from the self that listens. Each mode carries with it distinct sorts of pleasure,
ones which, manifestly, are not easily combined on a given occasion.
A question that naturally presents itself is whether there is a one-to-one
correspondence between musical pleasures and pleasurable musical features.
That is to say, is it the case that for every musical pleasure there is a musical
feature such that the pleasure is a pleasure in that feature? To me this seems
unlikely. Of course many instances will conform, and obviously so, to the
proposition under consideration. For example, pleasure in the intricacies of
Bachs counterpoint, or pleasure in the mellifluousness of phrase in Mozarts

On the social import of jokes, which serve at the same time to bind together members of
a given group while putting at a distance those of other groups, see Ted Cohen, Jokes, in Eva
Schaper (ed.), Pleasure, Preference, and Value (Cambridge: Cambridge University Press, 1983); the
lesson applies, with modifications, to musical communities that form around a given work, genre,
composer, or performer. On possible moral aspects of music see my Evaluating Music, and also
Kathleen Higgins, The Music of Our Lives (Philadelphia: Temple University Press, 1991), Colin
Radford, How Can Music Be Moral?, Midwest Studies in Philosophy 16 (1991), and Anthony
Savile, Kantian Aesthetics Pursued (Edinburgh: Edinburgh University Press, 1993), ch. 6.
See my Music in the Moment.
222 Music

late piano concerti, or pleasure in the sheer amplitude of Schuberts C major

symphony.
But what of the pleasure of being simply transporteddazzled, blown
away, knocked off ones feetby music? Naturally one can always say that
it was the music as a whole, in all its concreteness, that transported one. But
that doesnt imply that there is any particular featuresay, formal perfection,
sensual beauty, or expressive depththat has claim to being the object of the
pleasure in question. Alternatively, one could postulate a sort of transportative
virtueevidently possessed by the music, given it has transported oneand
maintain that the pleasure of being transported is a pleasure taken specifically
in that power of the music. But that would convince no one. I conclude that
even if most musical pleasures readily reveal themselves to be pleasures taken
in, or turning on, particular musical features, logically this need not be true of
all musical pleasures.
3. I come now to the primary concern of this essay, namely, a strange and
strangely pleasurable response to music that is, I assume, familiar to most ser-
ious music lovers, and one that has interested this music lover for the longest
time. It is the singular phenomenon of music-induced chills, or as I will
also denominate them, frissons. How do such chills or frissons arise, why
do we take pleasure in them, and what broader significance or value might
they have? My point of departure will be a recent empirical study of the phe-
nomenon conducted by the neuroscientist Jaak Panksepp.
The pleasure associated with musical chills is clearly of a sort we can label
physiologically centered. That is to say, musical pleasure in such cases revolves
around a particular physiological effect, in the present instance, the skin-
suffusing chill in question, where such effect is an integral part of the pleasure
experienced. And one reason the musical chill phenomenon is philosophically
interesting is this. How can a mere physiological tingle or shiver, so to speak,
a mere bodily disturbance, be of appreciative significance? Any number of
philosophers of art, most famously Nelson Goodman, have accustomed us
to view as ridiculous, through the ridicule they have heaped on it, the idea that
sensations might as such have a legitimate role in aesthetic response. What
good is a mere sensation, even an agreeable one, in the context of art? What
does it tell of or testify to? Does it inform us of some matter of artistic fact?

Recall Schumanns apostrophe of the symphonys heavenly lengths.

The Emotional Sources of Chills Induced by Music, Music Perception 13 (1995): 171207.
All page citations to this article are given in parentheses.
See his Languages of Art (Indianapolis: Bobbs-Merrill, 1968).
 Musical Chills 223

Does it illuminate some artistic relation of ideas? If neither, then consign it,
if not to the flames, then at any rate to the dustbin of appreciative theory.
Such is the prevailing Goodmanian wisdom on this score. But it is not one I
completely share, hence my interest in musical chills.
I turn now to Panksepps study, beginning with some of his attempts to
characterize the target phenomenon: the tingly somatosensory feeling that
can be evoked by certain kinds of music . . . (172); the provocative and often
delightful bodily experiences that deeply moving passages of music arouse
in many people (173); a bodily rush commonly described as a spreading
gooseflesh, hair-on-end feeling that is common on the back of the neck and
head and often moves down the spine, at times spreading across much of the
rest of the body (173). Panksepp notes that, despite its intriguing nature, the
prickly skin response usually called shivers, thrills, or chills in English
has not received the experimental attention it deserves. He observes further
that people rarely discuss the experience, and there is no unambiguous refer-
ent for it (173).
Taking up that last point, there is indeed a terminological problem for what
we wish to discuss, in that none of chill, thrill, or shiver seems entirely
apt to denote the phenomenon under investigation, each of those terms car-
rying connotations, whether of coldness (chill), or risk (thrill), or tremor
(shiver), that are in some degree undesired. Possibly the term frisson, a
partly nativized immigrant from French, is the best of the designations avail-
able for the phenomenon. At any rate, I will often speak in what follows of
musical frissons, though in discussion of Panksepps paper I will call them
musical chills in deference to his preferred term for the experience in question.
Experiments were conducted on undergraduates at a small midwestern
university in the United States. Panksepp employed as his test material
popular music of the 1970s and 1980s, items having been proposed by his
subjects themselves as chill-inducing. Thus with few exceptions, the test
selections were songs, mainly of the soft and hard rock variety. Ideally, of
course, one would have preferred that the experimenters had used textless
selections devoid of programwhat Peter Kivy calls music alonebut it
is no surprise that the design of the experiment, in which selections were
elicited from the student population, did not conduce to that. Panksepp
seems oddly unconcerned about the possible collateral effects of song lyrics,
with the articulate ideas and sentiments they contain, on the phenomenon
under study, but naturally it is a possible source of reservations about some of
his results.
224 Music

Here is one specific experimental result, rather typical of the study as a

whole:

The highest rate of reported chills was .5 chills/min/person for the beginning 3-min
segment . . . from Pink Floyds album Final Cut, which, on average, yielded essen-
tially the same number of chills as ones own song [that is, the selection provided by
the subject himself/herself] . . . . . . it was clear that the majority of the chills to this
piece occurred in response to the dramatic crescendo at the beginning of the second
minute. (1789)

As regards the most important general result of the study, it would seem
to be this: Overall, the data support the thesis that sadness or melancholy
is an emotional dimension more significantly related to the production of
chills than is happiness (187). In other words, it is negative, rather than pos-
itive, emotion in music that appears more efficacious in inducing the chill
experience.
Yet clearly some positively toned music in the study was found capable of
inducing chills. This prompts Panksepp to the following speculation, one that
would, if sustained, preserve a role for negative emotion in the generation of
chills in all cases: it will be worth considering whether the chills provoked
during happy music are caused by segments where happiness and sadness are
inextricably entwined in bittersweet feelings (187). I will later return to this
conjecture, which I believe to be on the right track.
4. Panksepps studies targeted a number of different factors plausibly
thought to bear on the incidence of musical chills. These included (a) the
gender of the listener, (b) the degree of familiarity with the music, (c) the
degree of liking for the music, (d) the emotional quality of the music, and
(e) the dynamic and tonal contour of the music. Panksepp observed a strong
correlation between chills and degree of both familiarity and liking, a strong
correlation between chills and both rise in volume and rise in pitch, a fairly
strong correlation, which we have already noted, between chills and music of
sadmelancholynostalgic character, and a weak, but statistically significant,
correlation between chills and being female.

Panksepp expands on this thought later on in the following manner: happiness and sadness
work together, and the most moving music allows the two processes to be blended in such a way
as to magnify our sense of ourselves as deeply feeling creatures who are conscious inheritors of the
tragic view (198).
That is, female subjects were generally more susceptible to musical chills than were male
subjectsor at any rate, they were more likely to report having them.
 Musical Chills 225

In my view there are probable determinants of the chill experience that

Panksepps experiments did not target, though they are ones that would
admittedly be hard to investigate quantitatively. Two worth mentioning are
(1) the musics fineness of expression and (2) the musics temporal expressive
shape. Panksepp seems not to have considered the possibility that fineness
of expression, whether glossed as depth or intensity or exquisiteness of
expression, may be crucial in triggering the chill experience, rather than
the expressing of negative emotion per se. Panksepp seems also not to have
considered the likelihood that a pieces expressive structuring in time, that
is, the pattern of succession of its individually expressive parts, what one
might loosely label its expressive narrative, contributes importantly to its chill
potential, with some sorts of succession, some kinds of narrative, being more
apt to elicit chills than others. I elaborate on this observation later on.
5. Recall now Panksepps suggestion that music with the greatest capacity
for inducing chills may well be of an emotionally hybrid naturethat it is
music in which positive and negative affects are in some manner or other
interwoven or combined. This connects to what I am inclined to propose, on
the basis of my own musical experience, as perhaps the crucial determinant of
chill-inducingness, namely poignancy of expression, or perhaps equivalently,
expression of poignancy.
A profound truth about life is that almost all situations or conditions
encountered are of mixed character. One is cognizant of the bad, if only
peripherally, even when firmly engaged with the good, and one glimpses the
good even when caught up in the bad, intermingled as they are in virtually
anything. Nor is this necessarily regrettable. For the mutual focusing of
positive and negative elements that results arguably ends up enhancing the
appreciation of whatever good is being enjoyed. The essential poignancy of
human life, one may suggest, resides in its mixed nature, in the indissociable
union of its joys and ills, the inescapable commingling of its pluses and
minuses. Thus, were we to assume that the prime determinant of musical
frissons was poignancy of expression/expression of poignancy, it would not be
surprising to discover that the music most reliably able to induce such frissons

See Stephen Davies, Musical Meaning and Expression (Ithaca: Cornell University Press, 1994),
ch. 7, for related reflections. Similar thoughts are to be found in Nietzsche, as Anthony Storr
observes: Nietzsche realizedno one more vividlythat the only life we know is constituted by
opposites. Pleasure is inconceivable without pain; light without darkness; love without hate; good
without evil . . . . This is why the greatest art always includes tragedy (Music and the Mind (New
York: Free Press, 1992), 158).
226 Music

was not that of unmitigated despair, nor that of untroubled gaiety, but that in
which there was some admixture of the two.
6. Here is a somewhat haphazard list of pieces containing passages
conducive to the production of frissons or chills, at least in my experience:
Brahms, String Quintet in G, Op. 111, first movement; Brahms, Piano
Trio in B, last movement; Brahms, Intermezzo, Op. 118 No. 1; Schubert,
Piano Sonata in C, D. 958, last movement; Chopin, Prelude Op. 28 No.
6, Mazurka Op. 17 No. 4 and Etude Op. 25 No. 1; Scriabin, Etude Op.
42 No. 5; Sibelius, Symphony No. 5, first movement coda; Faure, Violin
Sonata in A, first movement; Franck, Violin Sonata in A, last movement;
Schumann, Piano Concerto, first and third movements; Saint-Saens, Piano
Concerto No. 2, first movement; Poulenc, Sonata for Flute and Piano, first
movement; Mahler, Symphony No. 5, fourth movement; Mahler, Symphony
No. 6, first and third movements; Shostakovich, String Quartet No. 8, first
and second movements; Ravel, Gaspard de la Nuit, first movement; Nielsen,
Symphony No. 2, first movement coda.
From that list it might appear that only highly charged music of the
Romantic or early Modern period is capable of inducing frissons in this
listener, but that is not the case. Here are some other pieces, of earlier vintage,
that have this power: Beethoven, Piano Sonata in E, Op. 109, andante;
Mozart, Piano Concerto No. 20 in D minor, first movement; Mozart, Piano
Concerto No. 23 in A, second movement; Vivaldi, Concerto for Four Violins
in A minor, Op. 3, first movement; Vivaldi, Concerto for Two Violins in A
minor, Op. 3, first movement; Albinoni, Adagio for Organ and Strings in
G minor; and somewhat surprisingly, the first movement of Haydns sunny
Piano Sonata No. 60 in C, where certain chromatic bridge passages in the
exposition, recapitulation, and development usually produce the effect in my
hearing of them.
Passages with the capacity to induce chills need not even be heard, strictly
speaking, for that capacity to be realized: it suffices in many cases for them
merely to be run through vividly in aural imagination, courtesy of the mental

There is clearly a resemblance between this idea of the poignant as reflecting the essentially
mixed nature of what life has to offer and the traditional idea of the sublime as the correlate of an
aesthetic response distinct from that of the beautiful and involving a fusion of pleasure and pain. Yet
the experience of the sublime and the frisson-centered experience of poignancy cannot be simply
identified, first, because not all sublime experiences involve frissons, and second, because not all
frisson-centered experiences exhibit the specific cognitive character of the sublime, in particular, the
aspect of awe.
 Musical Chills 227

CD player, for frissons to be produced. And it seems generally true that when
one actively seconds or parallels music one is auditing, by a sort of inner
singing, frissons are more likely to occur. If so, that stands as yet another of the
many rewards of active, rather than passive, involvement with music, though
the reward itself, as has been noted, might perhaps be described as passive in
nature, being a sort of sufferance or submission.
The opening movement of Brahmss String Quintet in G, Op. 111, affords
me one of the most sustained chill experiences of any piece that I know. But
why, exactly, do I particularly relish that chill, as it steals up my spine and
pervades my body, suffusing it with a sort of oxymoronic warmth? It is, per-
haps, inherently pleasurable, but arguably no more so than having ones hair
stroked, or settling into a well-stuffed armchair, or consuming a nice mousse
au chocolat. So why do I value it above, or at any rate, differently from, those
other delights?
What makes the musical chill I receive from this piece particularly welcome
must be more than the mild corporeal pleasure it affords. That something
more, in this case, and as far as I can understand it, seems to be an accompa-
nying feeling of surrendering control, of letting go, of delivering oneself to a
powerful force, a guide to the terror and mystery of existence. More generally,
I suggest, chills of the sort in question announce themselves as the mark of a
confrontation with some fundamental truth of life, bodied forth by the music
that so moves us. Such chills are received not as mere physiological disturb-
ances, but as ones fraught with significance.
The echoes of religious experience here are intended. As has often been
observed, the greatest music seems to provide a passable substitute for the
sacred, for those who find themselves doxastically challenged in regard to the
traditional demands of religion. For many music lovers the listening room is a
kind of chapel, at least when certain items are on the order of service.
Anyway, contrast that movement by Brahms with a roughly contempor-
aneous piece, the Russian Easter Overture of Rimsky-Korsakov. Though col-
orful, imbued with feeling, well put together, and possessing a comparable
degree of forward momentum, the Russian Easter Overture is not, I suspect, a

According to anecdote, in composing this resplendent music Brahms was thinking of time
spent at the Prater, an amusement park on the outskirts of Vienna, and the simple joys of life
available there, including, as he was supposed to have remarked, the pretty girls. But that sounds
perhaps too earthbound a note for the almost superhuman affirmation and exhilaration bodied forth
in the music.
Nietzsche realized that, for many people, the concert hall and the art gallery have replaced
the church as places where the divine can be encountered (Storr, Music and the Mind, 155).
228 Music

piece likely to induce musical chills. Though enjoyable, and even absorbing,
as music it is simply too superficial: its energy does not speak to or tap into
anything profound in human nature. And so it does not summon from the
depths, as it were, those frissons that seem so full of import.
7. I return now to Panksepps study, and specifically, to Panksepps evol-
utionary speculations on the underlying cause and biological significance of
musical chills. My interest here is not so much in whether these speculations
are well-founded but in whether, if they are, this must affect the way in which
those who are susceptible to musical chills need regard them.
It is clear that people are most likely to have chills to music that has moved them in
the past . . . However, since unfamiliar sad music was more likely to provoke chills
than unfamiliar happy music, the evidence suggests that there are more primitive
instinctual neuropsychic components that underlie the phenomenon . . . . I will argue
that the chill ultimately reflects a property of ingrained neural systems of our old
mammalian brain that monitor emotions related to social proximity and separation.
(195)
. . . we presently know a great deal about the neural circuits for separation distress
that lead young animals to cry out when they are lonely and lost . . . . Internal feel-
ing of coldness and chills when parents hear separation calls may provide increased
motivation for social reunion. Thus the separation call may have been designed,
during the evolutionary construction of the brains emotional systems, to acoustic-
ally activate a thermally based need for social contact . . . Sad music may achieve its
beauty and its chilling effect by presenting a symbolic rendition of the separation call
(e.g. a high-pitched crescendo or a solo instrument emerging from the background)
in the emotional context of potential reunion and redemption. (1989)

The issue I want to raise is this. Once we have scientific insight into
the causesneurophysiological, biochemical, evolutionary, or what have
youof musical frissons, what impact should this have on our pleasure in
experiencing them or in being subject to them? Need such knowledge have
a deflationary effect, serving to undermine our satisfaction? Once we realize
that the responses in question are just, supposing Panksepps speculations to
be on the mark, a legacy of our evolutionary past, an artifact of a mammalian
brain still sensitive to the separation calls of errant young, must we rationally
cease to regard them as bearers of significance beyond the biological?
I think not. Supposing the underlying cause of musical frissons to be
an approximation to the separation call of lonely adolescent mammals, this
need not invalidate the other dimensions of such frissons in which their
value for us seems to reside. For first, the pleasure in being so affected
 Musical Chills 229

by the music, whatever its remote causes, is real, and can be rationalized
in terms of the beauty or depth or poignancy of the music to which the
pleasure seems most directly a response. Second, that a response has certain
underlying causes of an evolutionary sort does not preclude its also possessing
for us a certain significance, and possibly reflecting recognition of something
important about life as some music seems to embody it.
But in fact the evolutionary account of our susceptibility to frissons
proposed by Panksepp seems to me somewhat unlikely, at least as applied
to the sort of musical frissons I have primarily in mind. For there are really
two kinds of musical frisson that need to be distinguished. On the one
hand there are those, of relatively short duration, that are essentially timbrally
and/or dynamically induced, that is, produced by sound quality as such, and
typified by the effect on many persons of a clear and strong soprano voice.
On the other hand there are those, of relatively long duration, that are for the
most part melodically/harmonically/rhythmically induced, that is, produced
by sound structure as such. Structural features that appear to conduce to
frissons of this latter, more extended, sort include certain kinds of melodic
sequence, certain kinds of chord progression, chromatic intensifications,
pedal points, suspensions, delayed cadences, sustained tremolos, and melodic
leaps.
Thus, that the former sort of frisson rests on the precipitating musics
resemblance to piercing mammalian calls of separation may have some
plausibility, but that the latter sort of frisson has its roots there as well is
rather less plausible. For the latter sort of frisson exhibits more in the way
of temporal shape, of tension accumulated and discharged, of emotion built
up and released. Frissons of that sort thus seem harder to ascribe merely to a
particular color and volume of sound.
8. One researcher apart from Panksepp who has interested himself in music-
al frissons is the prolific psychologist of music John Sloboda. In one study
Sloboda attempts to identify the structural features associated with various

Of course there is also another sort of frisson instrumental music is capable of inducing, that
which relies on the force of external associations, e.g. of a patriotic or sentimental sort. Such would
be the frisson produced in someone by the hearing of his national anthem or the waltz that was
playing when he first laid eyes on his future wife. But I am not concerned with that sort of frisson
here, since clearly it is not rooted in musical sound or structure themselves.
Of course most such features will not by themselves produce frissons, if only because the
specific musical embedding of such features, and their interaction with other features constitutive
of a given passage, is absolutely crucial.
230 Music

pronounced physical responses to music. The results of his study point to

the following six features as strongly associated with musical frissons or, as
he prefers to call them, shivers: melodic appogiaturas; melodic or harmon-
ic sequences; enharmonic modulation; unprepared harmonic change; sudden
dynamic or textural change; and early arrival of expected events.
Appropriating a central idea of music theorist Leonard Meyer, Sloboda
conjectures that what ties such devices together is that they all involve
frustration of musical expectations. Though undoubtedly containing some
truth, this conjecture seems an over-generalization. Consider just melodic
sequences, which appear on Slobodas list of conducing features, or sustained
tremolos, which appear on mine, neither of which seems to contribute to
shivers primarily by countering expectations.
In a more recent study, Sloboda suggests that it is not unexpected music-
al turns per se that conduce to emotional peaks such as shivers, but rather
both the degree of unexpectedness of such events and the density of unexpected
events in a given stretch of music. That improves the explanatory power of
the conjecture, no doubt, but without making it wholly adequate. The reas-
on, as I earlier hinted in discussing Panksepp, is that the explanation accords
no place to the registering of the expressiveness of the music and the character
of that expressiveness. Slobodas explanation of musical chills, in other words,
though of a cognitive rather than an evolutionary sort, is insufficiently cog-
nitive, or more exactly, insufficiently cognitively complex. No formula that is
couched in terms of structural and expectational variables alone, and thus fails
to acknowledge the role of perceived expressiveness, can be entirely predict-
ive of frissons, especially the type of frissons that are of the most appreciative
importance.
To illustrate this point I venture an explanation of this more cognitively
complex sort of my experience of the Scriabin Etude p. 42 No. 5, mentioned
earlier. The tempo of the piece is fast, and the expressive marking is affan-
nato, or breathless. The overall shape of the piece is roughly AABABA, and
though the piece is passionate and agitated throughout, there is a signific-
ant contrast in emotional tone between the A and B sections, with the A
section being troubled and despairing in character, and the B section project-
ing, when heard in context, a more lyrical and hopeful state of mind.

Music Structure and Emotional Response: Some Empirical Findings, Psychology of Music 19
(1991): 11020.
Musical Performance and Emotion: Issues and Developments, in S. W. Yi (ed.), Music,
Mind, and Science (Seoul: Western Music Research Institute, 1999).
 Musical Chills 231

Now when in this marvelous etude B first succeeds A, my spirits momentar-

ily lift, as I detect a will to cast off the cloud of doom conveyed by A, a striving
toward something more positive, though with an undercurrent of anxiety that
does not depart. But when B again gives way to Aor more exactly, when
B is on the verge of giving way to a version of A even more desperate than
before (ms. 2730)I am struck by a palpable sense of the hopelessness of
the aspiration fleetingly perceived in B. B then appears to me to be overcome
by A, to be recaptured by it, and I realize that all is indeed lost, that there was
never really any hope for this doomed passion, that despair has now uncon-
tested domain. This is the juncture at which the occurrence of chills is for me
almost inevitable; and then again, though less powerfully, at the second, less
psychologically crushing shift, in the second half of the piece, from a more
febrile version of B back to A (ms. 467).
What I want to suggest is that only a quasi-narrative account of this sort,
positing a sequence of half-conscious, semi-articulated thoughts taking the
form of an emotional scenario, can adequately explain the power of music
such as this Scriabin etude to induce frissons in a range of listeners. And
though it would be quixotic of me to expect all listeners strongly moved by
this music to confess to precisely the scenario I have sketched, I believe that
the sense of the music that most such listeners would extract from their own
experiences would conform, at least roughly, to such a scenario.
This is a good point at which to acknowledge the important role that spe-
cific performance of a piece plays in the generation of musical chills, given a
structurally based potential for that in the music itself. Undeniably, certain
performances exploit the chill potential inherent in a piece better than others,
through their specific shaping and pacing of the sequence of musical events in
which chill potential evidently resides. In the case of the Scriabin etude just
discussed, for example, the performance by Ruth Laredo is by far the most
effective I am acquainted with in that regard, the performance of the Brahms
quintet by the Juilliard Quartet is more effective than that by the Guarneri
Quartet, and the performance by Emmanuel Ax of the bridge passage in the
Haydn sonata noted above is decidedly more conducive to chill production
than the equally fine but more hurried performances by Glenn Gould and
Mikhail Pletnev. To investigate further the whys and wherefores of the differ-
ential chill efficacy of different manners of performing a given piece of music
would, however, take us too far afield.

Scriabin: Complete Piano Sonatas, Nonesuch Records.

232 Music

9. What lessons about the pleasurability and value of musical frissons might
we now draw from our consideration of their causes and conditions? It will
help to begin by comparing the musical chills with which we have so far been
concerned, and which are ordinarily experienced with pleasure, with some
other species of bodily chill. Consider the chill that shoots through you when
you suddenly think, rightly or wrongly, that you have deleted a computer file
on which you have been working for hours, or the chill that runs down your
spine when you sense that a burglar has just broken into your house through a
ground-floor window. Chills of this sort are obviously far from pleasurable,
and not ones we take satisfaction in, though of course they have a certain
practical value. What makes for the difference in pleasure or satisfyingness
between such chills and the ones typically had from music?
In broad terms, it appears to be the cognitive construal of the frisson, or
the way the frisson is interpreted, that makes most of the difference. More
narrowly, what makes the difference is the sort of perception underlying the
chill in question and bound up with it. The computer-glitch and burglar-
entry chills are unpleasant because they reflect the perception of something
bad, whereas chills had in response to music are pleasant because they reflect
the perception of something good.
However, that may not account for all the difference in the cases. For it is
not clear that the chills in question are precisely the same, physiologically and
phenomenologically speaking. If we may generalize, those which arise in the
course of engagement with music tend to blossom more slowly, to suffuse the
organism more gently, and to have softer edges, whereas those prompted by
sudden perceptions of danger or loss seem more sudden and piercing. These
latter are more truly chilling, as it were. Thus, even among bodily chills to
which we are subject there may be qualitative differences that suit certain of
them to a role in experiences of positive value, while others are excluded, on
the same grounds, from so taking part.
In any case, whatever their distinctive qualitative nature, the pleasurability
of musical chills seems at least in large part dependent on what a listener takes
such chills to signify. For example, if you take your chill to signify that you

At least they are so for the vast majority of music lovers who are subject to them; that is, they
are experienced as hedonic in tone. I have discovered, however, that there are listeners who in fact
do not find chills from music delectable, but on the contrary, unpleasant. My speculation is that
these are listeners of a certain personality type, who resist letting go or surrendering control in even
the smallest measure. It is unclear whether for such listeners there is any scope for cognitive framing
of received chills rendering them more palatable.
 Musical Chills 233

have recognized the musics poignancy or fineness of expression, then you will
likely take pleasure in it, whereas if you take it to signify, say, that the treble is
turned too high on your stereo, you likely will not.
Well and good. But might the chill not in fact be valuable apart from the
pleasure you take in it when you regard it as significant, in virtue of what it
might actually signify? The question before us, in other words, is that of the
epistemic value of musical chills. Do they really have any such value, or are
they merely physiological responses that we take pleasure in when we regard
them a certain way, namely, as having epistemic value?
Musical chills obviously have one kind of epistemic value, but it is not
of a sort that answers to our purposes. That is, they indicate that the music
being heard is, so to speak, chillogenic, or capable of inducing chills in listeners
under certain conditions. Clearly, what we are interested in under the rubric
of the epistemic value of musical chills is not that, but rather what such chills
might possibly indicate beyond themselves, concerning either the music or
the listener.
There are, it seems, two possibilities for what musical chills might reli-
ably signal. One, they might signal the perception of some notable feature of
the music, such perceptions often occurring without the subject being clearly
conscious of them. Two, they might signal the presence of the notable musical
feature itself. In other words, they might directly testify to something hap-
pening in the subject, namely, perceptions at a more or less conscious level
concerning features of the music; or they might directly testify to the presence
of such musical features, somehow functioning as immediate registrations
of them, in something like the way pains function as indicators of bodily
damage.
I am inclined to think that musical chillsor at any rate, many such chills
had by experienced listenershave epistemic value in at least the first sense.
Such chills are the upshot of perceptions of expressiveness in music, and reli-
ably signal those perceptions to perceivers, who are often only dimly aware
of them and their contents. Such musical chills thus have intentionality of a
complex sort: as corporeal markers of perceptions directed on something dis-
tinctive in the music, they are, in the first instance, about the music-directed
perception that is occurring, yet also, in the second instance, about the feature
of the music on which the perception is directed.

Why not take the line according to which such chills directly represent valuable properties of
the music? For it certainly seems as if such chills register something about the music rather than
234 Music

Earlier I gave an account of the sort of thoughts and perceptions involved in

my hearing of a particular Scriabin etude, and which I maintained lay behind
and were crucial to the chills I experience in listening to it. I now briefly do
the same for the Haydn and Brahms examples offered before, in further illus-
tration of the likely epistemic value of at least some musical chills. As with the
Scriabin, this will not be to say that the thoughts and perceptions adduced
were fully formed on first hearing of those pieces, or even on the first hearing
of them in which chills occurred. It is rather a matter of plausible reconstruc-
tion and articulation of what, prior to that, is more inchoately or obscurely
apprehended. Naturally, however, once reconstructed and articulated in a
form that seems to square with the underlying, more inchoate apprehension,
such thoughts and perceptions are likely to become part of the explicit con-
tent of subsequent experiences of the music.
Anyway, my hearing of the Brahms quintet includes a recognition, at some
level, of the intense vitality embodied in that incomparable opening passage,
and especially in the cello part, which strives heroically to assert its individu-
al nature against the generalized, insistent sawing of the other strings. And
my hearing of the Haydn sonata includes a recognition, at some level, of the
subtlety and suppleness of the shifts in harmonic color in the course of that
modulating passage, which displays the marvelous seamlessness of an organ-
ic process. These are highly distinctive features of the music in question, and
my body, according to the present hypothesis, signals to me my emotionally
fraught perception of them by the chills in which those perceptions eventuate,
ultimately prompting me to articulate those perceptions on a more conscious
level. But the chills, I suggest, are usually the first sign that one has registered
something of depth or significance in the music. In other words, chills of the
sort I have been discussing thus serve as focusers of attention, as direct aids to
appreciation, drawing attention to expressive aspects of musical structure that
might otherwise escape notice.
A few more words are in order on the nature of the chills exemplified in
my experiences of these pieces by Haydn, Brahms, and Scriabin. The chills
in question are plausibly the upshot of perceptions of emotionally significant
patterns in music, which perceptions have themselves an emotional character.

something about me. The problem is coming up with a plausible mechanism as to how chills could
be directly sensitive to such complex properties as fineness or poignancy of expression. A two-stage
model seems more promising, as it does for, say, cold-induced chills, whereby shivering from cold
directly indicates coldness of the body (internal temperature), and then indirectly, coldness of the
environment (external temperature).
 Musical Chills 235

These emotion-imbued perceptions reach a level of intensity sufficient to

issue in a physiological responsethe famous chillwhich we can now
understand as a kind of overflow of the emotional perception developed to
that point. Such a chill is thus not merely an indication of an emotional
perceptionas would be, say, a twitching of ones left earlobe or a blue flash
in ones visual field, if regularly conjoined with such perceptionsbut the
natural culmination of the emotional perception itself.
10. What position have we arrived at, then, on the appreciative value of
musical chills? It would seem to be this. First, musical chills can be sources
of aesthetic satisfaction insofar as we take them to signal perceptions of
something significant about the music to which we are listening; and second,
musical chills can have epistemic value insofar as they actually signal that we
have indeed perceived something significant about the music, often serving to
first alert us that we have done so.
But the Goodmanian qualm may now reassert itself. For it may seem that
it is not, after all, chill sensations as such that have appreciative value, but
only the perceptions to which they are connected in favorable cases. And if so,
one might be tempted to insist that it is really the music-directed perceptions
that count appreciatively, not the chill sensations, thus ultimately acceding
in, rather than diverging from, the Goodmanian position according to which
chills as such are both without value and irrelevant to value. If musical chills
are only appreciatively valuable when functioning as signalers of perceptions
regarding the music, then does not all the value in fact reside in those percep-
tions, whether or not chills accompanying them take place?
Not quite. For what is important, and what registering chills while
listening underscores, is the response to music of, so to speak, the whole
person. Responses to music of that sortones that are cognitive, emotional,
physiological, and behavioral at onceare arguably of greater value than
more limited or restricted responses to music being heard. Thus even if such
chills are valuable only when had in connection with certain perceptions
regarding the musics content or structure on the part of those who have
them, the chills remain essential to the full value of the experience in question.
The chills are essential to the full value of the experience, one might say,
because they represent a bodily confirmation of what is registered at the same
time, usually in an obscure manner, by the mind. Arguably, the marking of
perceptual apprehensions by felt frissons imparts a kind of added value to
musical experience. Such frissons stand as corporeal endorsements of what is
concomitantly grasped in more cognitive terms. A central appreciative value
236 Music

of such frissons, I suggest, lies precisely in the affirmation of wholeness that

they afford, of mind and body resonating together in response to a given
musical utterance.
There is thus something right about the idea that not to feel frissons or
chills at certain junctures in a piece of music is to be missing something appre-
ciatively. For not to register such effects is to have an experience of the music
that is not as responsive, in some sense, as it should be. The underlying
rationale may be that, as embodied creatures, the responses we most admire
are not those in which only mind, or only body, are involved. To be most
human, it appears, is to react to things, and perhaps especially works of art,
with our whole selves.

On the account that has the chill directly signaling the perception of a valuable musical
property, such as the musics embodying of a particular emotional scenario, the force of the idea
that the chill in seconding the perception serves to make ones response to the music more whole
may appear undercut, since on that account the perceptions object is the music, while the chills
object is the perception and not the music, so that perception and chill are not clearly responses
to the same thing. But this appearance can be mitigated. For the chill, though directly signaling
the perception, still indirectly signals the musical property, and thus the idea can be preserved that
there is effectively a combined cognitive and physiological response to the music.
There may be a parallel here in the moral sphere, where a person who can, say, classify acts
into good, bad, and heinous, but does not experience any revulsion in contemplating acts of a
heinous sort, is regarded as somehow morally lacking.
 PART I I I
PI C T U R E S
This page intentionally left blank
 13
Wollheim on Pictorial Representation

In On Pictorial Representation Richard Wollheim offers us an elegant

precis of the account of pictorial representation he has developed over the past
thirty years. In addition, he comments on competing views of the matter,
and responds to criticisms or requests for elaboration that his own view has
elicited. Though I eventually express some reservations on the view myself,
the extent of my accord with Wollheim on this topic is rather large, as I now
indicate.
First, I agree with Wollheim that the concept of pictorial representation,
or depiction, cannot be explicated without appeal to a characteristic sort of
experience, the sort of experience Wollheim has denominated seeing-in. Sus-
taining an appropriate seeing-in experience, that is, a seeing-in experience that
conforms with the artistic intention governing a given picture, is what is cri-
terial of such representation, and not anything else.
Second, I agree with Wollheim, as against Budd, that seeing-in is generally
prior to, and not to be analyzed in terms of, the perceiving of resemblances as
such, whether between objects or experiences.
The fundamental rationale for so insisting is this. Though perception
of resemblance, or more narrowly, structural isomorphism, between object
aspects or visual fields, may be a concomitant, trigger, or consequence of

First published in Journal of Aesthetics and Art Criticism 56 (1998): 22733.

Richard Wollheim, On Pictorial Representation, Journal of Aesthetics and Art Criticism 56
(1998): 21726.
Notable bulletins in that development include Seeing-as, Seeing-in, and Pictorial
Representation, in Art and Its Objects, 2nd edn. (Cambridge: Cambridge University Press, 1980);
Imagination and Pictorial Understanding, Proceedings of the Aristotelian Society, Suppl. Vol. 60
(1986): 4560; Art, Interpretation and Perception, in The Mind and Its Depths (Cambridge, MA:
Harvard University Press, 1993); and above all, Painting as an Art (Princeton: Princeton University
Press, 1987).
240 Pictures

seeing-in, it is not equivalent to seeing-in. Seeing-in can occur without

such perceptions, and vice versa, and so there can be no identifying them.
The experiences of perceiving resemblances and seeing-things-in-other-things
are different, and irreducibly so; the former is inherently relational and
comparative, the latter not. We may observe, in addition, that were seeing-
in to be identified with perception of structural isomorphism, then since the
latter is clearly a notion of degree one would expect the former to be as well.
But seeing-in is not evidently a notion of degree, nor is that of depiction,
which seeing-in underwrites; seeing-in and depiction are closer, if anything,
to being on-off or all-or-nothing affairs.
What is likely true in this matter is that a nonzero degree of structural iso-
morphism between a representation and its subject is required for seeing-in to
take place, that is, that some such isomorphism may be a causal precondition
of seeing the subject in the representation; the mechanisms whereby seeing-
ina kind of seeing, after allis enabled to occur seem to require as much.
But even if that is so, the perception of such isomorphism, as opposed to its
mere existence, remains strictly unnecessary to the occurrence of the distinct
experience of seeing-in.
Third, I agree with Wollheim, as against Walton, that seeing-in is generally
prior to, and not to be analyzed in terms of, imagined seeing.
A reason for that insistence, beyond those hinted at by Wollheim, is as fol-
lows. If seeing-in is equated with imagined seeing of a certain kind, that is, if
every case of the former is made out to be a case of the latter, then we lose a
resource for explaining some of the special character, whether of immediacy,
intimacy, absorbingness, or emotional impact, of some pictures as opposed to
others (or alternatively, of some occasions of experiencing pictures as opposed
to others), by appeal to the idea that although all pictures in being perceived
as such induce seeing-in, only some pictures induce (or only some occasions
of experiencing involve) actually imagining seeing the object that a picture

To elaborate: perception of resemblance between visual field 1 and visual field 2 explicitly
involves relating and comparing those items, while seeing object X in painting Y does not involve
a parallel relating and comparing of those items. The second term in such an experience, that
is, Y, does not enter into the content of the experience involved, though naturally it is involved
in generating the experience and in fixing what experience it is. The content of the experience,
consisting as it does in seeing X, in a manner of speaking, is basically just X . (I say basically, since
it might be held that the content in question is not precisely X but something more like image of X.
The point would remain that such content was non-relational, or at least, not such as to involve a
relation to Y .)
 Wollheim on Pictorial Representation 241

represents. Imagining seeing X in viewing Y implies, as a default, imagining

you are face-to-face with X; but it seems doubtful whether one is standardly
doing that merely in virtue of seeing X in Y, that is, recognizing the look of X
in the design of Y.

II

The basic shape of Wollheims position on pictorial representation is thus

one I find congenial. But I have come to have various qualms about its spe-
cific articulation, qualms that prompt me to a friendly interrogation of some
of its constitutive elements. As a result I am led to venture certain claims
that Wollheim would, I am sure, be reluctant to embrace. Still, the picture
of picturing that I uphold remains, in broad outline, a clearly Wollheimian
one.
The elements of Wollheims position that I will examine are these: the
treatment of trompe-loeil; the status of twofoldness in seeing-in; the recog-
nitional aspect of seeing-in; the scope or range of seeing-in; and the appeal to
the artists fulfilled intention as a standard of representation. At more length,
the questions I want to pursue are as follows. (1) Is trompe-loeil precluded
from being understood as representational because it is designed to forestall
an apprehending experience characterized by twofoldness? (2) Is the experi-
ence of seeing-in in fact necessarily characterized by twofoldness, that is, sim-
ultaneous awareness of medium and of subject, such that seeing-in has always
a configurational as well as a recognitional aspect?; (3) What can be said about
the recognitional awareness that is arguably at the core of seeing-in, especially
if configurational awareness, or awareness of medium, is not always present as
well?; (4) Is seeing-in really the same phenomenon or mental state across all
the sorts of things it is said can be seen in pictures? (5) Is the artists fulfilled
intention to depict such and such an apt criterion of what it is correct to see in
a picture, and so of what it depicts?

On my conception of it imagining is necessarily in some degree active or contributory, though

not necessarily something one is aware of initiating, and not necessarily something under complete
control of ones will. By contrast, seeming to one as if what I propose captures, as well as anything
can, the experience at the heart of pictorial seeingis passive or receptive, not something one brings
about and actively sustains, but something that, in the last analysis, simply occurs. Seeing X in Y is
something that happens to one, even when deliberate mental actions of various kinds, for instance,
framings, thinkings, or suggestings, serve as triggers to such happening.
242 Pictures

III

That trompe-loeil pictures pose a problem for the seeing-in theory of

depiction is, I think, undeniable. If being a depiction requires inviting
and sustaining seeing-in, and if seeing-in is an experience that necessarily
involves twofoldness, and if twofoldness necessarily implicates awareness of
and attention to pictorial surface, at some level, then it seems that trompe-
loeil pictures cannot be depictions. Though Wollheim is content to accept
this consequence of his seeing-in account, it strikes me, as it has others, as
counterintuitive.
Now there is in fact a way to understand trompe-loeil pictures as supporting
appreciative experiences with something like twofold character, and thus as
thereby having clear claim to depictive status, before addressing the question
of whether simple seeing-in is necessarily characterized by strict twofoldness.
It is this. When we see trompe-loeil pictures as pictures, that is, when we
are aware that they are pictorial contrivances, when we are past the point of
being taken in by them, when we recognize them as trompe-loeil while allow-
ing them to continue to fool the eye, then something like twofoldness, or
simultaneous awareness of subject and medium, is present, even though the
medium is, in a way, transparent. In such cases there is a kind of awareness,
perhaps even visual awareness, of the surface, in the sense that visual atten-
tion is carried to it, despite the fact that with a perfect trompe-loeil the surface
remains invisible. Once you grasp that something is a trompe-loeil you can
attend to its surface, and in its visual aspect, even though you cannot by hypo-
thesis see the surface as such. What you can do with a trompe-loeil painting, as
with any painting, is mentally focus on the surface before you at the same time
as you register its pictorial content, notwithstanding the fact that in such cases
the surface does not end up arresting your vision.
But let us put aside that resolution of the difficulty, appealing as it does to
an exceedingly liberal construal of twofoldness, and consider again the prob-
lem generated for the theory of picturing by trompe-loeil. It seems there are
two options open to us. We can either allow that seeing the pictorial content
of a trompe-loeil painting without realizing it is such, and so a fortiori without
any awareness of the paintings surface, is still an instance of seeing-in, and

See, for example, Dominic Lopes, Understanding Pictures (Oxford: Oxford University Press,
1996), ch. 2.
 Wollheim on Pictorial Representation 243

thus that such seeing does not always involve twofoldness (first option), or
else deny that seeing the pictorial content of a trompe-loeil painting without
realizing it is such is an instance of seeing-in, thus retaining twofoldness as a
necessary feature of such seeing (second option).
One might argue in favor of the second option that naively registering the
pictorial content of a trompe-loeil does not involve seeing the picture as a
picture, and for that reason should not be accounted a case of seeing things in
the picture. In addition, since attention to form concurrently with content,
or to content-as-embodied-in-form, is often taken to be the heart of what it is
to carry aesthetic attention to an object, one might further argue, against the
first option, that by its lights seeing-in would not necessarily exhibit aesthetic
character.
However, as is probably apparent, such an argument would be weak, since
the considerations on which it turns seem more convincingly deployed in the
opposite direction. Plausibly not all seeing-in or registering of pictorial con-
tent is aesthetic in character, or even informed by the awareness of pictures
as pictures; for instance, that directed to or had in connection with post-
cards, passport photos, magazine illustrations, comic strips, television shows,
or movies. Thus, any view that builds aesthetic character, or even awareness
of pictures as pictures, directly into seeing-in would seem to have something
amiss. It seems perfectly reasonable to hold that one can be seeing things in
pictures, in virtue of looking at pictures, even when one is not seeing them as
pictures, and a fortiori, without appreciating them aesthetically.
I propose, then, that we embrace the first option, whereby simple seeing-in,
and what we might call pictorial seeing proper, are distinguished, with only
the latter definitionally implying twofoldness. Pictorial seeing, or seeing pic-
tures as pictures, is indeed a sine qua non of aesthetic appreciation of pictures,
but the fact is that there can be seeing-in in connection with pictures that
is not even pictorial seeing, that does not involve any awareness of pictorial
properties or the medium in which they are embedded.
If you see a woman in a picture in virtue of visually processing a pattern of
marks, then of course in some sense you are thereby perceiving the medium

See Richard Eldridge, Form and Content: An Aesthetic Theory of Art, British Journal of
Aesthetics 25 (1985); Malcolm Budd, Values of Art (London: Penguin, 1995); and Jerrold Levinson,
Pleasures of Aesthetics (Ithaca: Cornell University Press, 1996), chs. 1 and 2.
I am not, of course, denying that we may often be cognizant of such pictures as pictures, or
carry aesthetic attention to them, only that we must or even usually do so.
I have been influenced on these points by Lopess critique of seeing-in theory in Understanding
Pictures.
244 Pictures

in which those marks inhere or consist. But it is far from clear that when you
see the woman in the picture you must in some measure be attending to, tak-
ing notice of, or consciously focusing on the pictures surface or patterning as
such. Yet that does appear to be part of the import of twofoldness as Woll-
heim construes it: Looking at a suitably marked surface, we are visually aware
at once of the marked surface and of something in front of or behind some-
thing else. I call this feature of the phenomenology twofoldness. That
twofoldness as Wollheim understands it means that the experience of seeing-
in involves, in its configurational as well as its recognitional aspect, some level
of conscious apprehension and not, say, merely unconscious registering, is
confirmed by this more extensive passage from Painting as an Art:

The twofoldness of seeing-in does not, of course, preclude the one aspect of the com-
plex experience being emphasized at the expense of the other. In seeing a boy in a
stained wall I may very well concentrate on the stains, and how they are formed, and
the materials and colours they consist of . . . and I might in consequence lose all but a
shadowy awareness of the boy. Alternatively, I might concentrate on the boy, and on
the long ears he seems to be sprouting . . . and thus have only the vaguest sense of how
the wall is marked.

A crucial issue, then, would seem to be what, exactly, being visually aware of
a picture surface amounts to. Not, surely, receiving information from the sur-
face, or being sensitive to changes in features of the surface; such construals are
too weak for the purpose, since they are too easily satisfied by mental states,
for example subdoxastic ones, that lie below the level of consciousness. Not,
surely, thinking or reflecting that one is seeing the surface as one sees it; such
a construal would be too strong, collapsing visual awareness per se and self-
conscious visual awareness. Perhaps, then, something like this: attending to the
surface as one views it and is affected by it. But if anything like that construal is
adopted, it is indeed doubtful that the seeing-in involved in grasping pictori-
al content always entails or includes visual awareness of the surface as well.
At any rate, Wollheim has not indicated an intermediate notion of awareness
that might be apt to the needs of the case but that does not import any degree
of attention whatsoever.

On Pictorial Representation, 221.

Painting as an Art, 47 (my emphases). We might, echoing Death of a Salesman, underline that
in order for something to count as awareness attention must be paid, at least in same degree.
Note that if this critique of seeing-in in respect of whether it necessarily displays twofoldness
is correct, we may still rest with Wollheims desired characterization of a depiction as a marked
 Wollheim on Pictorial Representation 245

IV

The task remains, though, of saying what simple seeing-in consists in, giv-
en that it does not inevitably involve visual awareness of medium, that is,
attention in some degree to medium, and yet is not just seeing in the ordin-
ary sense. As has rightly been observed, clarifying what Wollheim calls the
recognitional aspect of seeing-inand what we may now take to be the very
core of seeing-inseems incumbent on a supporter of the seeing-in approach
to pictorial representation.
Here, then, is a stab at what such recognition amounts to. In looking com-
prehendingly at a picture of a woman, say Kees van Dongens engaging and
mildly fauve canvas, La chemise noire, one does not necessarily perceive an
isomorphism between experience of the picture and experience of a woman
(Budd), nor does one invariably imagine seeing a woman (Walton), nor, in all
probability, does it seem to you that a woman is actually before you (Gom-
brich). Rather, I suggest, it seems to you as if you are seeing a woman (altern-
atively, you have an impression of seeing a woman), in virtue of attending
visually to portions of the canvas. The core of seeing-in, in other words, is
a kind of as-if seeing that is both occasioned by visually registering a differen-
tiated surface and inextricably bound up with such registering.
Of course more needs to be said about the tight relation required here
between the registering of visual information and the perception of pictorial

surface intended for seeing-inonly that should now be understood as simple seeing-in, it being
granted that in most cases the surface is also clearly intended for pictorial seeing, with its inherent
twofoldness, as well.
To say it does not involve visual awareness of the medium is not, of course, to say that it does
not involve visual processing of the information embodied in the medium.
See Kendall Walton, Seeing-in and Seeing Fictionally, in James Hopkins and Anthony Savile
(eds.), Psychoanalysis, Mind and Art (Oxford: Blackwell, 1992).
Note that seeming to one as if P, or having the impression that P, are not locutions that
entail believing that P or even thinking it probable that P. For example, It seems to me as if I am
falling unsupported, said in a rapidly descending elevator.
There is a real question whether the experience I have continued to refer to as simple
seeing-in should in fact be so called. Two reasons give pause. The first is that the association
of seeing-inwhich is, after all, a term of art introduced by Wollheimand twofoldness is so
entrenched that an experience of seeing-in sans twofoldness sounds almost oxymoronic. The second
is that conceiving such experiences as the seeing-in label encourages one to do, as a matter of seeing
things in surfaces, does undeniably occasion strain where trompe-loeil pictures are concerned, since
in such cases the surfaces are, by hypothesis, neither seen nor seeable. One might thus concede that
the visual experience of pictures I have been calling simple seeing-in, and that is present even when
twofoldness is not, might in certain cases with more justice be called rather seeing-from.
246 Pictures

content. The relation has to be such as to rule out non-standard causal

routes by which a pictures visual array might lead one to have an impres-
sion of seeing a woman, for example, one where such an array triggered, at
a subperceptual level, a chemical change that in turn issued in a localized
hallucination of a woman just like the woman seeable in the picture. The
impression of seeing, or as-if seeing, at the core of seeing-in is thus one intim-
ately bound up with the registering of the visual data afforded by the picture,
whereby the latter in a sense constitutes or realizes the former.

None of this is to deny that much of the interest and appeal of seeing-in
lies in the possibility of twofoldness in ones experience of a picture, that
is, simultaneous awareness of both picturing pattern and pictured object,
where ones seeing-in thus becomes seeing pictorially, properly speaking.
Yet equally important, I would suggest, is the option, in which one might
at turns indulge, of switching back and forth between awarenesses or
focusings of attention of those two kinds, seeing sometimes only pure
pattern, sometimes only pure object. In fact it would seem reasonable to
include, within the ambit of pictorial seeing, that is, seeing of the kind
normative for pictures understood and appreciated as pictures, both seeing

It is a virtue of Waltons account of depiction, of course, that it secures the desired intimacy in
the most direct fashion, by making the act of perceiving the picture that which the viewer imagines
to be an act of actually seeing the subject of the picture. But it seems to me that what necessarily
happens in such a case of seeing-in is at most that one takes ones apprehending of a surfaces forms
and colors to be a seeing of a woman, in a sense that does not imply that one believes or suspects
that one is seeing a woman, but not that one imagines of such apprehending that it is a seeing of a
woman.
One might worry, finally, that if simple seeing-in is construed so as not to necessarily involve
awareness of a pictures surface, then simple seeing-in and simple seeing will collapse. But this worry
is unfounded. In the case of simple seeing-in you seem to see X, that is, you have an as-if-seeing
experience of X, in virtue of visually registering certain configurations of a surface, rather than in
virtue of being in the visual presence of X. In the case of simply seeing X, it is true as well that you
seem to see X, that is, have an as-if-seeing experience of X, but then there are differences. With
simply seeing X there is, first, the belief, or tendency to believe, that X is before you, and second,
it is X, and not merely a surface configured to afford an impression of X, that indeed is before
you. This is all admittedly rough and ready, not intended as careful analysis. The point is just that
the experiences of simple seeing-in and simple seeing can surely be discriminated, though once
twofoldness is abandoned as a sine qua non of seeing-in such discrimination may not be a wholly
internal matter, but may instead rest on matters such as what is precipitating the experience and
what sort of mechanisms are involved in its doing so.
 Wollheim on Pictorial Representation 247

where there is simultaneous awareness of design and content (i.e. twofoldness),

and seeing in which there was alternation back and forth between phases
of simultaneous awareness of design and content and phases of exclusive
or near-exclusive focus on one or the other. It seems that our knowing
engagement with pictures does in general display an alternation between
phases of simultaneity, often sustained without deliberateness, and ones of
switching, often occasioned by deliberate reflection on what ones experience
is like. Pictorial seeing might thus conveniently be stretched to cover such
activity in all its phases.
It is hard to overestimate the keen interest that viewers of painting naturally
take in bringing simultaneously into relation, or alternating systematically
between, the recognitional and the configurational, or the pictured and the
picturing, in different styles of depiction. This is one of the obvious, but
nevertheless deep, sources of fascination with the differences among Neo-
Classical, Impressionist, Post-Impressionist, Expressionist, Cubist, Surrealist,
and Abstract-Expressionist treatments of what is in some sense the same
subject. We are endlessly amazed with the variety of ways there are to
pictorially construct familiar objects, so that patterns or designs that would
seem to have little in common, compared as such, are revealed to have an
affinity in supporting equally a visual impression of, say, a cow. That a cow
can be made, visually speaking, out of dots, dashes, lines, angles, masses,
smears, or mere chiaroscuro, is something we delight in bringing home to
ourselves through this activity of regularly correlating design and content in
our apprehension of a painting. Each time, after absorption in the represented
world, that we attend primarily to the configurations afforded by a painting,
we derive anew the pleasure of seeing of what the objects of that world have
been made.
But there is yet more than that. Different styles of representational painting
arguably give us access to unique kinds of beings, allowing us to see things
not encountered in the real world at all, rather than merely allowing us to
see familiar things in a new way. What I have in mind are beings such as
these: Ingres-women, Picasso-women, and De Kooning-women; Kirchner-
men, Beckmann-men, and Grosz-men; and finally, Miro-dogs, Klee-dogs,
and Dubuffet-dogs. Paintings of the respective artists familiarize us with
extraordinary creatures of that sort, ones that can enter importantly into
ones imaginative and interpretive repertoire; such paintings do more than
simply show us how those artists, or their implied personae, may be said
248 Pictures

to have viewed ordinary women, men, and dogs. Of course, after making
the acquaintance of Ingres-women, Kirchner-men, or Miro-dogs, one may
then be in a position to spot their instantiations, or near-instantiations,
in the world around one, hors de peinture. That is to say, we achieve
acquaintance with kinds of beings whose exemplars are not all of them, or
not necessarily, fictional.

VI

I have concentrated so far mostly on seeing-in as it applies to objects. But as

Wollheim urges, seeing-in may be held to range over actions and events as
well, and even over individuals-merely-of-a-certain-kind as opposed to partic-
ular individuals. Concern arises, however, as to whether the range of seeing-in
is usefully taken to be as wide as Wollheim proposes. The concern might
alternatively be expressed as one of whether the seeing-in involved in all such
cases is sufficiently of a piece as to merit the single label. Let us look to what
Wollheim says about the outer limits of seeing-in as he descries them, as illus-
trated in the example of the classical landscape with ruins.
Wollheim suggests that a suitably prepared and prompted viewer plausibly
can see, in such a painting, all the following: columns, columns-as-having-
come-from-a-temple, columns-as-having-been-thrown-down, columns-as-
having-been-thrown-down-hundreds-of-years-ago, and columns-as-having-
been-thrown-down-hundreds-of-years-ago-by-barbarians. But such a viewer
cannot, Wollheim submits, see in the painting columns-as-having-been-
thrown-down-hundreds-of-years-ago-by-barbarians-wearing-the-skins-of-
wild-asses.
But why not? Why are all those other qua-objects seeable in the
picturewhich I take to be roughly interconvertible with the seeability
therein of corresponding states of affairsbut not the last? What principle
of cut-off for the qua-objects or states of affairs that can be seen in a picture

A story from the golden era of The New Yorker is relevant here. James Thurber, one of the
early great cartoonists of the magazine, was once the subject of discussion at the weekly art meeting
being presided over by the then editor, Harold Ross. The point at issue was the seal perched on
the headboard of a bed in one of Thurbers most famous cartoons, in which the wife is vocally
skeptical of her husbands claims to have heard a seal bark. Someone at the meeting, noting the
somewhat loosely drawn character of the seal, asked Do seals look like that?. To which Rosss reply
was: Thurbers seals look like that. (From an interview with New Yorker Cartoon Editor Robert
Mankoff, The Washington Post, Dec. 7, 1997.)
On Pictorial Representation, 224.
 Wollheim on Pictorial Representation 249

does Wollheim have in mind? What he offers is an operational test: propose

for seeing-in a candidate state of affairs and note whether it makes a difference
in the suitable spectators experience. Yet in the absence of a clear idea of the
bounds of seeing-in, it is hard to know how one would interpret the results of
such a test. It is not entirely obvious what would rule out seeing the columns
as having been thrown down by barbarians wearing the skins of wild asses.
After all, we no more see the vandals and their destructive acts than we see the
equine pelts they may very well have sported.
Of course we can speculate on what it is that makes a non-manifest state
of affairs or condition a reasonable candidate for seeing-in as Wollheim con-
ceives that. Possibly a condition being such that perceptual inference to it
is highly compelling, or a condition possessing visual traces of a relatively
unequivocal sort, at least for a properly backgrounded viewer, makes such a
condition something that can be seen in a picture. But I am not concerned
to worry further about these or any similar suggestions. The real problem, I
think, is that seeing non-manifest states of affairs in pictures, seeing occurrent
actions in pictures, and seeing objects in pictures may be importantly different
phenomena, whose differences, say as regards spatial localization or permeab-
ility by thought, may be more obscured than illuminated by considering them
together as members of a species. Seeing-in may not be univocal across its
putative instances, especially if, as suggested earlier, twofoldness is not even an
invariant feature of the experience of seeing one thing in another.
It is not clear that the same sort of activity or perception is involved when
going from seeing-in of objects to seeing-in of events to seeing-in of con-
ditions, or from seeing-in of physical events or conditions to seeing-in of
psychological events or conditions. For example, localization, the property of
such-and-suchs being seeable more or less right where the relevant pictorial
design is, may be characteristic of the seeing in paintings of physical objects,
but somewhat less so of the seeing of events, and very much less so of the see-
ing of psychological entities, whether objects or events. And the permeability
of seeing-in to thought, reflection, or conceptualization seems progressively
more pronounced, in general, as one moves from the seeing of objects to the
seeing of events to the seeing of only indirectly evidenced states of affairs in
a painting. And finally one might add, for good measure, that a role for ima-
gination in the robust sense appears considerably more plausible in regard to
seeing-in of this latter sort than to seeing-in of the former two sorts. These
divergences make suspect, at the least, the assumption that seeing-in in all the
cases claimed by Wollheim is of a uniform nature.
250 Pictures

VII

Lastly, there is to my mind a problem about Wollheims appeal to fulfilled

representational intentions as a standard for what a picture represents. Here
is a formulation, though from an earlier paper: Very roughly, P represents
X if X can be correctly seen in P, where the standard of correctness is set
for P by the fulfilled intentions of the artist of P. A difficulty lurks here
that Wollheim and other actual intentionalists about meaning have some
tendency to gloss over. It is this. What it is for the pictorial intentions of
the artist of P to be fulfilled cannot be specified apart from what suitable
viewers are enabled to see in P. Such intentions are fulfilled if viewers are in
fact enabledand enabled without undue thematic prompting or inordinate
mental contortionto see in P what the artist intended be seen there. The
artists fulfilled intention cannot be thought of as an independent condition to
which viewers responses can be held accountable, but can only be understood
in terms of the responses of appropriately primed and backgrounded viewers
being the ones they were intended to be. Another way of making my point
would be to say that the standard of correctness for depiction is not, as
Wollheim sometimes puts it, the fulfilled intentions of the artist, but merely
the intentions simpliciter of the artist for a certain sort of seeing-in, given that
they are capable of being complied with by the pictures intended viewers.
The artists representational intentions only are fulfilled if suitable viewers are
enabled, on reasonable prompting, to see-in the painting in accord with the
artists representational intentions. Thus it arguably makes little sense to say
they comply with the artists fulfilled intentions in this regard, since such do
not, as it were, preexist such compliance.

VIII

Richard Wollheims theory of pictorial representation is the fruit of long

reflection, deep insight, and an intense love of painting. In the course of

Imagination and Pictorial Understanding, 46.

To be fair to Wollheim, the formulation of the intentional condition on representation in the
present essay almost entirely escapes the problem, highlighted here, to which earlier formulations
were subject: Representational meaning, indeed pictorial meaning in general, is, on my view,
dependent, not on intention as such, but on fulfilled intention. And intention is fulfilled when the
picture can cause, in a suitable spectator, an experience that tallies with the intention (On Pictorial
Representation, 226).
 Wollheim on Pictorial Representation 251

this essay I have criticized that theory in a number of respects, notably, its
treatment of trompe-loeil, its conception of seeing-in, and its appeal to the
artists fulfilled intention as a standard. Observe, however, how much remains
of what Wollheim has urged in what I am willing to affirm on this vexed topic:
pictorial representation involves the intentional marking of a flat surface so
as to elicit a distinctive sort of visual experience in appropriate spectators,
an experience we may continue to call seeing-in as long as we understand
that this sometimes amounts only to what may be more transparently labeled
seeing-from, where such experience is indeed elicited from those spectators in
virtue of their attending to the surface as marked.
 14
What is Erotic Art?

Here is an answer to the question of my title: erotic art is art which aims to
engage viewers sexually through explicit sexual content, and that succeeds,
to some extent, in doing so. In addition to developing and defending that
answer, in this short essay I explore the boundaries of the concept of erotic art,
and some of the psychological and social implications of erotic art.
The scope of this essay is restricted in two ways. First, the essay concerns
visual art exclusively, and then, almost entirely, visual art of two dimensions.
But of course, that is not to deny the existence of erotic literature, dance,
cinema, or even, less obviously, music. Second, the examples are drawn only
from Western art, and mainly Western art since the Renaissance. But that is
not to deny the prominence of erotic art in non-Western traditions of art, for
instance, those of India or Japan.
There are arguably strict and less strict senses of the term erotic art. If the
definition offered above captures the strict sense of the term, applicable to
art that is unquestionably or unequivocally erotic, we may also recognize art
that is erotic according to a looser construal of the term, including art that is
only accidentally erotic, art that is only instrumentally erotic, art that is only
covertly erotic, and art that is only erotic, a bit paradoxically, in virtue of being
anti-erotic. These different cases will be addressed in detail in what follows.

I M AIN QUE ST IONS

The chief philosophical questions regarding erotic art would seem to be these.
(1) What is the distinction, within art, between erotic and non-erotic art, and
how sharp is this distinction? (2) What are the normative implications, if any,

First published as Erotic Art, in E. Craig (ed.), The Routledge Encyclopedia of Philosophy (London:
Routledge, 1998), 4069.
 What is Erotic Art? 253

of the different forms and modes of erotic art, and what, in particular, is the
distinction between erotic art and pornography? (3) How can erotic art in fact
be art, that is, something properly eliciting an aesthetic response, traditionally
characterized as disinterested, while also aimed at provoking sexual desire, the
very paradigm of an interested reaction? (4) In what way, if any, do the criteria
for assessing erotic art differ from those appropriate to assessing art of other
sorts, and how does the degree of eroticness of a work of erotic art relate to its
goodness as art? This essay will be devoted almost exclusively to the first and
second of these questions.

II THE CONCEPT OF EROT IC ART

A good proportion of the work of many great visual artistsRubens,

Ingres, Delacroix, Degas, Rodin, Gauguin, Matisse, Magritte, Munch, Klimt,
Schiele, Picasso, Modiglianiis unquestionably erotic. But what is it, pre-
cisely, for art to be erotic? It seems that, at a minimum, it must have sexual
content. Though sexual content may be either overt or covert, let us con-
sider first such art as has overt sexual content. Typically, this takes the form
of depictions of unclothed or semi-clothed human beings, alone or accom-
panied, at rest or performing actions of a sexual nature.
But for art to be accounted erotic, it must do more than represent the naked
human body or otherwise make reference to sexual matters: not all art con-
cerned in some way with sexuality counts as erotic. An anatomical sketch of
private parts by a Durer or a Leonardo, a realistic study of a gynecologists
examining room, a comic strip featuring pneumatic bimbos, are none of them
erotic, despite their inclusion of sexual content.
Rather, erotic art is art that treats its sexual content in a particular way or
that projects a certain attitude toward it. Erotic art is art aimed at arousing
sexual interest, that is, at evoking sexual thoughts, feelings, or desires in view-
ers, in virtue of what it depicts and how it is depicted, and which achieves
some measure of success in that regard. The intent to awaken and reward
sexual interest through what is depicted can be taken as criterial of at least
central cases of erotic art. The erotic work of art does more than merely refer
to or acknowledge human sexuality; rather, it expresses an involved attitude
toward it, whether of fascination, obsession, or delectation, and in addition,
invites the viewers imaginative engagement, along similar lines, with what
is shown.
254 Pictures

Erotic art not only aims at engaging the sexuality of the viewer, but typically
also reflects that of the artist. That is to say, erotic works usually embody a
perspective on what is depicted that suggests sexual interest, and of a particu-
lar sort, on the makers part. Furthermore, the sense of sharing in what at least
appears to have been sexually stimulating to the artist often plays a causal role
in the viewers own stimulation by what is depicted.
It is worth emphasizing that the sexual response occasioned by erotic art
occurs largely on the plane of imagination, consisting primarily of thoughts,
images, and feelings, and rarely goes as far as full physiological arousal; the
upshot of engagement with erotic art is imagined desire as often as it is real
desire. This is not unrelated to the distinction between erotic art and porno-
graphy, which is touched on below.
As suggested above, the term erotic art in its central usage covers art that
aims at, and that at least minimally succeeds at, stimulating sexual thoughts
and feelings in its target audience. But this leads, easily enough, to two looser
usages, according to which, roughly speaking, meeting either the intentional
condition or the success condition independently qualifies a work as erotic.
On the first such looser usage, a work counts as erotic if it is apparently aimed
at stimulating sexual thoughts and feelings even when it is not successful in
doing so, while on the second such usage, a work counts as erotic if it succeeds
in stimulating viewers sexually even when not intended or even apparently
intended to do so. Art of the former sort might be labeled nominally erotic art,
while art of the latter sort might be labeled accidentally erotic art.
Finally, perhaps some artworks reasonably accounted erotic neither aim at
nor achieve viewer arousal as such, that is, sexual thoughts, feelings, or sen-
sations directed towards what is depicted, but instead are erotic in virtue of
facilitating the imagining of erotic states of others, without unequivocal erotic
involvement on the viewers part, that is, without the viewer identifying with
or entering into those states, in either reality or imagination. Such art might
merit the label of obliquely erotic art.

III INSTRUMENTALLY EROTIC ART

AND ANTI-EROT IC EROT IC ART

With some erotic art the evocation of erotic feelings, rather than being the
main order of business, is a secondary aim, and is employed or manipulated
by the artist primarily in order to achieve some further end. For example,
 What is Erotic Art? 255

Tom Wesslemanns caricatures of female pulchritude and Mel Ramoss

exaggeratedly voluptuous pin-ups use erotic images to achieve a kind of
wry humor, Degass monotypes of brothel scenes serve purposes of social
commentary, and the recombinant sexual imagery of Magritte and Dali
is a means to psychological disorientation of the viewer. We might label
such art instrumentally erotic art. As a result of these secondary aims, the
excitatory tendency of such works is generally weakened, and is sometimes
wholly neutralized. In some limiting cases works are in effect about erotic
artthey are commentaries on, or satirical appropriations of the conventions
and mechanisms of, ordinary erotic artbut without being erotic in the
central sense, that is, ultimately aimed at sexually engaging the viewer. Such
art may be accounted erotic in virtue of leading the viewer to question the
presuppositions and consequences, social and otherwise, of erotic responses,
without inviting or even permitting viewers to have such responses.
Some other cases of works representing sexual matters without appearing
clearly erotic will serve to illuminate further the boundaries of the category:
1. Lysippuss sculpture Aphrodite, Botticellis Birth of Venus, and Cranachs
Eve occasion some hesitation if classified as erotic. Probably this is because
we take the primary intent of the artist to have been to embody ideals of
the human form, male and female, and not to prompt imaginative erot-
ic engagement on the part of viewers of either gender. But this may be
ingenuous; at any rate, such a line could not plausibly be extended to
exclude from the erotic Donatellos sensuous, almost coquettish, David.
2. Picassos Demoiselles dAvignon occasions hesitation of a different sort.
Though the painting presents women who are not only nude but in fact
prostitutes, they are depicted in a highly non-realistic mode, which short-
circuits erotic involvement, as well as drawing attention primarily to the
paintings formal and expressive dimension.
3. Judy Chicagos The Dinner Party, an elaborate sculptural installation, uses
female genital imagery, and in a celebratory fashion, but probably not in
an erotic way; its sexual content is of the purely symbolic, rather than sexu-
ally involving, sort.
4. Lucian Freuds paintings of naked subjects, though displaying some of the
hallmarks of erotic art, are not obviously erotic, being more evocative of
the boucherie than the boudoir an observation even truer of the images
of nudes in Francis Bacons paintings. Philip Pearlsteins super-realist
256 Pictures

figure paintings or, in another vein, Dubuffets quasi-paleolithic images

of squashed and splayed humanity, belong here as well.
The works of these latter artistsBacon, Freud, Pearlstein, Dubuffetare
not aptly described merely as non-erotic, as are, say, a Corot landscape or a
Chardin still life, but rather as anti-erotic. But in a broader sense they, unlike
the Chardin or the Corot, are erotic after all, in the sense of being concerned
with sexuality in a way that reflects the sexual interests of the maker and
engages those of the viewer, if not in a positive manner. De Koonings raw
and primitivist images of women come naturally to mind in this connection
as well, though the case can be made that those images project an even more
ambivalent attitude to human sexuality than those of Bacon and Freud, with
their admixture of awe, terror, and admiration.

IV COVERTLY EROTIC ART

It is relatively easy to give plausible examples of erotic art with no explicit

depictions of sexuality or nakedness: Georgia OKeeffes landscapes and still
lifes, with their oblique evocation of female anatomy; Caravaggios paintings
of Bacchus or St John the Baptist, with their coded references to homosexual
experience; or Berninis marble of St Teresa in spiritual ecstasy, a state readily
translated by the viewer into its profane cousin.
What the criterion of covert sexual content is, however, remains unclear.
Depiction of objects recognized as sexually symbolic, such as umbrellas
or fruit, especially when they are juxtaposed with human subjects, may
be a typical indication of such content, but can hardly serve as a general
mark. According to some writers, virtually all art has covert sexual content
in virtue of being the expression of unconscious wishes or fantasies of a
sexual sort. For Richard Wollheim, for instance, Ingress history paintings,
Bellottos landscapes with buildings, and Poussins landscapes with water are
as substantially imbued with sexuality as Goyas Naked Maja or Titians Venus
of Urbino.
Even so, it seems that not all covertly sexual art is usefully considered erotic,
but only that which is plausibly aimed, if unconsciously, at exciting sexu-
al thoughts or feelings in target viewers, and which succeeds in doing so.
In putative cases of covert sexual content, the arousal of the appropriately

Richard Wollheim, Painting as an Art (Princeton: Princeton University Press, 1987).

 What is Erotic Art? 257

backgrounded and oriented viewer may be what in fact signals the presence
of such content, and justifies its ascription.

V T HE REL ATIONALIT Y OF E ROTIC ART

If a painting is erotic, this is in virtue of being aimed at and to some extent eli-
citing an erotic response from a certain class of viewer, the paintings intended
or target audience. Such classes may be delimited not only by requirements
of sensitivity and background knowledge, such as are appropriate to art of
any kind, but also by ones, less acquirable, of physiological makeup or sexual
orientation. Thus, a painting may be erotic in virtue of being designed to pro-
duce, and succeeding in producing, an erotic reaction in heterosexual males,
elderly homosexual males, young heterosexual girls, homosexual women, or
bisexuals of either gender. There is a fact of the matter, if a hazy one, about
whether a given painting is erotic, but it is an inherently relational one, whose
nature is only fully evident when the group targeted for response is identified.
Indeed, according to Linda Nochlin, the very term erotic art is understood
to imply the specification erotic-for-men . Still, once such implicit index-
ing has been made explicit, it may then be cancelled, so as to recognize art that
is erotic relative to other target groups.

VI SOCIAL AND POLITICAL DIMENSIONS

OF EROT IC ART

Recent writers on erotic art stress the way in which entrenched genres
and conventions of representation embody dominant ideas and assumptions
about the nature of men and women and their proper relationship. Paintings
such as Delacroixs Death of Sardanapalus, Geromes Oriental Slave Market,
or the Turkish Bath and Jupiter and Thetis of Ingres lend themselves readily
to such analysis. For example, Linda Nochlin speaks of the power relations
obtaining between men and women inscribed in visual representation as a
focus of her investigations.
Equally frequently noted is the element of voyeurism in erotic art. It is
said that the spectator is a voyeur, at least fictionally, with the artwork often

Linda Nochlin, Woman, Art, and Power and Other Essays (New York: Harper & Row, 1988).
Ibid.
258 Pictures

seconding or echoing this by containing a depicted spectator who, together

with the viewer, regards the erotic object. The implicit or explicit voyeurism
of erotic art is, in addition, sometimes held to reflect the necessary impotence
of the artist in respect of the imaginary, and thus unattainable, individuals
depicted within his art.
Finally, the relationship of erotic art and pornography has been much
queried. The latter may be distinguished from the former in at least two
ways. One, pornography might be said to have, by definition, no significant
artistic aspect. That is to say, pornography in the strict sense makes no
credible appeal to viewers to consider the mode and means of depiction, as
opposed merely to what is depicted; pornography, unlike art of any kind, is
wholly transparent in both aim and effect. Two, pornography might be said
to have, as a central intent and characteristic result, not merely the stimulation
of sexual feelings or fantasies in viewers, but the devaluation or degradation of
its subjects, usually women. By such criteria, although Courbets Sleep, which
depicts two beautiful nude women in the arms of Morpheus and each other,
or Schieles Reclining Woman, which presents its subject provocatively spread-
legged and scarlet-nippled, perhaps court categorization as pornography, on
reflection they remain at some distance from it; though the images in question
are starkly arousing, even exploitative, the technique of their construction, the
style in which they are rendered, the preceding art history they encapsulate,
and the access they afford into their makers psyches, are at least as absorbing
as what those images flatly represent, and conspire to redeem them as art.

For a more nuanced account of the distinction between pornography and erotic art, see my
Erotic Art and Pornographic Pictures, Philosophy and Literature 29 (2005): 22840 (Ch. 15, this
volume).
Some instructive writings on erotic art, apart from those of Wollheim and Nochlin already
noted, are these: John Berger, Ways of Seeing (London: Penguin, 1972); Kenneth Clark, The
Nude: A Study in Ideal Form (Princeton: Princeton University Press, 1956); Edward Lucie-Smith,
Sexuality in Western Art (London: Thames and Hudson, 1972); Linda Mulvey, Visual and Oth-
er Pleasures (Bloomington: Indiana University Press, 1989); Lynda Nead, The Female Nude: Art,
Obscenity, and Sexuality (London: Routledge, 1992); Griselda Pollock, Vision and Difference: Fem-
ininity, Feminism, and the Histories of Art (London: Routledge, 1988); Roger Scruton, Sexual
Desire: A Moral Philosophy of the Erotic (New York: Free Press, 1986); Guy Sircello, Love and
Beauty (Princeton: Princeton University Press, 1990); Leo Steinberg, Picassos Sleepwatchers, in
Other Criteria (London: Oxford University Press, 1972).
 15
Erotic Art and Pornographic Pictures

Only in primitive art, with its urgent need to evoke the sources of fertil-
ity, are the phallus and the vulva emphasized, as it were innocently. By
ancient Greek and Roman times there already existed the special category
of the pornographicgraphic art or writing supposed, like a harlot, or
porne, to sexually stimulate.

As regards philosophical analysis of the opposition between the erotic and

the pornographic, there are a number of reasonable goals one might have: to
preserve as many considered intuitions about the opposition as possible; to
present the opposition in a clearer light than it enjoys when casually invoked;
to propose modest sharpenings to the standard opposition that either account
for our experience in this domain more fully, or allow us to organize our
thinking about the domain more perspicuously. I hope in this essay to make
progress toward some of those goals. Though the scope of the opposition of
erotic and pornographic goes beyond the visual, my focus here will be the
opposition as it exists in the visual sphere, and even more narrowly, in the
sphere of two-dimensional images. In addition to preserving and clarifying a
distinction between erotic art and pornography, I hope also to make an intelli-
gible place for erotica, as something intermediate between the other two.
Here, then, are some intuitions on the erotic and the pornographic:
1. The erotic and the pornographic are both concerned with sexual stimula-
tion or arousal.

First published in Philosophy and Literature 29 (2005): 22840.

John Updike, Can Genitalia Be Beautiful?, review of Egon Schiele exhibition at the Museum
of Modern Art, New York Review of Books, Dec. 4, 1997.
260 Pictures

2. While the term erotic is neutral or even approving, the term porno-
graphic is pejorative or disapproving.
3. While erotic art is a familiar, if somewhat problematic, notion, porno-
graphic art seems an almost oxymoronic one.
4. Whereas pornography has a paramount aim, namely, the sexual satisfac-
tion of the viewer, erotic art, even if it also aims at sexual satisfaction on
some level, includes other aims of significance.
5. Whereas we appreciate (or relish) erotic art, we consume (or use) porno-
graphy. In other words, our interactions with erotic art and pornography
are fundamentally different in character, as reflected in the verbs most
appropriate to the respective engagements.
In what follows I try to accommodate all of those intuitions. As I will need
a distinction, invoked in the first of them, between sexual stimulation and
sexual arousal, let me spell that out before proceeding. By sexual stimulation
I will mean the inducing of sexual thoughts, feelings, imaginings, or desires
that would generally be regarded as pleasant in themselves. By sexual arousal I
mean the physiological state that is prelude and prerequisite to sexual release,
involving, at least in the male, some degree of erection. And by sexual release I
mean something that I take it needs no spelling out.
How to differentiate erotic art from pornography, and from erotica as well,
is of course not the only important philosophical question about erotic art.
Here are two others: (1) How is the erotic aspect of erotic art compatible with
the disinterested or distanced frame of mind that seems required for aesthetic
engagement with or appreciation of a work of visual art? (2) How is the degree
of eroticness of a work of erotic art related to its artistic value? I will briefly
address those two questions toward the end of the essay, but most of my effort
will be devoted to the prior question, that of the differentiation of erotic art,
erotica, and pornography.

II

How, then, to effect that differentiation, in the domain of visual images to

which I have restricted my inquiry? One possibility would be by the specific
kind of response aimed at. Thus erotic art might be said to involve images
intended to stimulate sexually but also to reward artistic interest, erotica
to involve images intended to stimulate sexually but not to reward artistic
interest, and pornography to involve images intended to arouse sexually in
 Erotic Art and Pornographic Pictures 261

the interests of sexual release. Another possibility would be by the character

of the sexual representation involved, so that pornography would involve
sexually explicit images, erotic art sexually inexplicit images, and erotica
images intermediate in their explicitness, irrespective of the ends for which the
images were fashioned. And a third possibility would be by the moral status of
the images involved, with pornography offering morally objectionable sexual
images, erotic art morally unobjectionable sexual images, and erotica sexual
images of borderline moral status.
For reasons that will emerge shortly, the first tack is the only viable one for
making out the threefold distinction we are after. What makes the difference
among the three kinds of sexual image, in other words, is what they are for,
what response they are designed to evoke, what they are meant to do to us or
we with them. The stark contrast is between erotic art, which invites artistic
interest, and pornography, which positively deflects such interest, whatever
degree of artistic interest it might, as it were accidentally, sustain. So for-
mulated the contrast clearly presupposes a satisfactory gloss on what artistic
interest amounts to, but that will be provided shortly.
What we have, then, is a subdivision of the broad category of erotic, or
intentionally sexually interesting, images, into three subcategories: erotic art,
erotica, and pornography. But I maintain neither that the boundaries between
these subcategories are sharp, nor that there are no examples of erotic images
that perhaps fit into none of those subcategories. Though there is little need, I
think, to offer examples of erotic art or pornography, it will be useful to men-
tion an example of what to my mind counts as erotica, namely, classily pro-
vocative lingerie ads of the Victorias Secret variety. Note also that the broad
category of erotic images, at least roughly exhausted by the three subcategor-
ies I have detailed, is not as broad as the category of sexual images generally,
that is, images of any sort depicting sexual phenomena, such as sexual organs
or acts or conditions. Thus, illustrations in textbooks of gynecology or field
guides to baboon mating behavior, though they depict sexual phenomena,
do not count as erotic images, since they are not intended to interest viewers
sexually; nor does security camera footage of a rape or molestation, even if in
fact sexually stimulating or arousing to certain sorts of viewers.

III

Let me return briefly to the other options mentioned earlier for demarcat-
ing the erotic from the pornographic, namely by degree of explicitness of
262 Pictures

representation involved or by moral status of the representations involved, in

order to show their unworkability. The first of these is the more easily dis-
patched. The fact is that, although as a rule erotic art involves representations
of organs or acts that are less explicit than those of pornography, erotic art
also embraces representations that are equally, if not more, explicit, than those
in many instances of pornography. Much of the erotic art of artists such as
Schiele, Klimt, Picasso, or Dali is as explicit as any Penthouse pictorial, and
this is true also of certain traditions of Japanese graphic art. Pass on to the
idea of distinguishing pornography from other varieties of erotic image on
moral grounds. Sometimes the defining mark of pornography is taken to be
its implied attitude toward its subject, an attitude, perhaps, of demeaning
or degrading them. But even if such an implied attitude is typical of porno-
graphy, its not clear it can be held to be an invariable feature of it, unless the
mere fact that such subjects are depicted for the purposes of sexual fantasizing
and arousal is taken to be demeaning or degrading. As for the idea that por-
nography might be conceived more loosely as erotic images that are in some
way or another morally objectionable, there are at least two problems with
that. First, even if a moral case could be made against most pornography, it is
unlikely that it can be made against it all; surely there must be morally accept-
able ways of deploying sexual imagery to aid persons in achieving solitary
sexual release. Second, even if such a case could be made against all porno-
graphy, its morally objectionable qualities would seem to be a consequential
feature of it, rather than a defining one. Compare embezzlement, or the unau-
thorized taking or diverting of funds for private gain in a business context.
That may very well be immoral, but its immorality is not reasonably made
part of the definition of embezzlement. In addition, there are plausibly cases,
such as when embezzlement is undertaken to save ones family from starva-
tion, where embezzlement is not immoral, even though it is as a rule.

IV

Now, what do I mean by the artistic interest or dimension of an erotic image?

Roughly, its form and the relation of that form to its content; the way the con-
tent has been embodied in the form, the way the medium has been employed
to convey the content. We can speak almost equivalently of the artistic intent

For reflections on the morally worrisome dimension of pornography for consumers of

pornography, see my Sexual Perversity, Monist 86 (2003): 3054.
 Erotic Art and Pornographic Pictures 263

of an image as of its artistic interest or dimension, if that is understood in

a hypothetical manner, meaning the intent the image appears to manifest to
have viewers attend to its artistic dimension or artistic interest. I will thus
loosely alternate artistic interest, dimension, and intent in what follows, since
although in another context one might want to tease them apart, for present
purposes there is no need to do so.
So characterized, an image that has an artistic interest, dimension, or intent
is one that is not simply seen through, or seen past, leaving one, at least in
imagination, face to face with its subject. Images with an artistic dimension
are thus to some extent opaque, rather than transparent. In other words, with
artistic images we are invited to dwell on features of the image itself, and not
merely on what the image represents. Both erotica and pornography predom-
inantly aim at sexually affecting the viewer, one with an eye toward stimu-
lation, the other with an eye toward arousal, and accordingly do not seek to
have attention rest on the vehicle of such stimulation or arousal, the medi-
um through which the sexual content is communicated or presented. Erotic
art, though aimed in part at sexually affecting the viewer, at stimulating sexual
thoughts and feelingsthats what makes it erotic art, after allalso aims in
some measure to draw the viewers attention to the vehicle, inviting the view-
er to contemplate the relationship between the stimulation achieved and the
means employed to achieve it, and more broadly, the relationship between the
erotic content of the image and its other contents, such as expressive, dramat-
ic, or religious ones.
Its thus no accident, but highly telling, that photography is the prime
medium for pornography, that which has displaced all other such media in
that connection. For photography is the transparent medium par excellence,
that is, the medium that comes closest to simply presenting the requisite
objecttypically, a woman or a man or combinations thereofdirectly,
as material for sexual fantasy and gratification. Though photographs of
course can be art, and more specifically, erotic art, they also lend themselves
extremely well to non-artistic employment, which makes use of their inherent
transparency, whereby they serve, if we let them, as mere aids to seeing.
As we all know, pornography is essentially a kind of substitute or surrogate
for sex, whether a poor one or not we can leave aside. That is why it is appro-
priate to characterize it narrowly in terms of the facilitation of sexual fantasy

The locus classicus in philosophy on the transparency of photography is Kendall Waltons

Transparent Pictures, Critical Inquiry 11 (1984): 24677.
264 Pictures

in the name of arousal and release. And that is why to fulfill that purpose its
images should be as transparent as possiblethey should present the object
for sexual fantasy vividly, and then, as it were, get out of the way. Nothing
does that better than photography.

I turn now to a provocative essay on our subject by Matthew Kieran, whose

conclusions are diametrically opposed to those I have been trying to estab-
lish. Kieran attempts to make a place for pornographic art, refusing to accept
that the extension of that concept is by definition the null set, that porno-
graphic art is indeed the oxymoron it appears to be. His attempt, however,
simply shows that the works he champions are not pornographic art, or even
erotica, as I use those terms, but instead erotic art of a distinctive kind, in
which pornography or erotica are themselves subjects of the art in question,
or in which pornography or erotica have been turned to artistic ends, and so
transformed into art.
One of Kierans examples is the novel Vox, by Nicholson Baker, which
Kieran argues counts as pornographic art, though of course we are here deal-
ing with literature rather than visual art. But I maintain that Vox is not por-
nographic art, if that means it is both art and pornography, though I grant
that it is, in a sense, pornographic. I will remove the air of paradox from that
assertion a bit further on.
Whats true of Vox is that it mimics pornography, and in particular, phone
sex, appropriating its gestures, tropes, and outer appearance, but does so in
order to produce a work of literature, and thus, art. And one can admit that
it is, at many places, sexually stimulating, even arousingafter all, a simu-
lacrum of something often has many of the same properties and powers. But
the point is that that is not all it is, nor all it is intended to be, nor what it is
ultimately aimed at producing in readers by way of experience. Vox resembles
pornography, to be surethat is obviousbut is not identical to porno-
graphy, because its paramount aim is not that of producing sexual arousal
and release. Yet it is, of course, a mild turn-on, owing to the effective simu-
lation of verbal pornography it presents throughout its length, and couldnt,

See his Pornographic Art, Philosophy and Literature 25 (2001): 3145. Much the same
case was made earlier, though with less philosophical sophistication, by Susan Sontag, in The
Pornographic Imagination, Styles of Radical Will (New York: Farrar, Straus & Giroux, 1969).
 Erotic Art and Pornographic Pictures 265

in fact, achieve its artistic aims, which are variously psychological, parodistic,
and pyrotechnic, if it werent.
Another of Kierans examples is the ensemble of erotic drawings by Gustav
Klimt. His discussion of those drawings is very insightful, showing as it does
how Klimts mastery of artistic techniques helps to focus attention on the
sexual parts, features, and states of the women depicted, making those draw-
ings all the more erotic. But it is one thing to say that certain artistic devices,
masterfully deployed, can enhance the erotic charge of a representation. It is
quite another to say that a viewers focusing on those devices will enhance the
representations erotic charge for the viewer, that is, render it more stimulat-
ing or arousing. There is every reason to think it will not, that it will rather
temper the stimulation or arousal involved, replacing what is thus lost, how-
ever, with a portion of aesthetic pleasure. So what Kierans discussion shows,
to my mind, is that Klimts drawings are not pornography, but rather art,
albeit art that might be mistaken for pornography by inattentive viewers, or
that might be used as pornography by viewers happy to lose sight of its artful
fashioning and just enjoy the erotic upshot thereof.
Making room for pornography that is also art, I suggest, is a bad idea. First,
allowing that something can be pornographyand not just resemble porno-
graphy, or mimic pornography, or have a pornographic flavor, or be quasi-
pornographicand art at the same time, leaves no place for the category of
erotic art as distinct from pornography. Second, the aims of true pornography
and the aims of art, erotic art included, are not compatible, but war against
one another, in the way that has already been sketched. One induces you, in
the name of arousal and release, to ignore the representation so as to get at
what is represented, the other induces you, in the name of aesthetic delight, to
dwell on the representation and to contemplate it in relation to the stimulat-
ingness or arousingness of what is represented.
Now to remove the air of paradox from my assertion above that Vox is art,
perhaps even pornographic art, but not pornography. We can say that there is
pornographic art, if we just mean that there is art that has a pornographic look
or character, but then such art is not yet pornography. That is, it is not both art
and pornography. Analogously, there is art, for example, certain kinds of con-
temporary painting, such as that of Richard Estes or Alex Katz, which might
be described as photographic. This means the art has a certain photographic
look or character, not that it literally is photography. And there is writing,
such as that found in certain kinds of newspaper articles, which can be called
telegraphic, but that doesnt literally make that writing telegraphy.
266 Pictures

There is, in effect, a strong and a weak sense of the term pornograph-
ic in the expression pornographic art. In the strong, or conjunctive, sense,
something is pornographic art if it is both art and pornography; in the weak,
or modifying, sense, something is pornographic art if it is art and has a porno-
graphic character, look, or aspect. It is only in the weak sense of pornograph-
ic, I submit, that there is pornographic art. And in that sense, of course, the
inference from x is pornographic art to x is pornography fails. Recognizing
that the expression pornographic art has a strong and a weak reading helps
explain why many astute and progressive thinkers are as open to the idea as
they are, for the ease of satisfying the weak reading of the expression creates
the illusion that the strong reading is satisfied as well.
Kieran tries to dispel three considerations against pornographic art under-
stood in the strong sense. The first consideration is that pornography is by
definition non-artistic, or without artistic interest, to which Kieran replies
that pornography is just highly explicit erotica, and since all admit that erot-
ica can be art, there is thus no conceptual bar to pornography being art as
well. But we have already seen that degree of explicitness cannot be the distin-
guishing mark of pornography, both because some erotic art is more explicit
than some pornography, and because the essence of pornography arguably
has something to do with what its images are for and not just what they
show. The second consideration is that pornographys central aim, namely, to
arouse sexually through explicit means, militates against its achieving artistic
interest, even if that is not precluded by definition, to which Kieran responds
that such an aim is not in fact incompatible with achieving such interest. Kier-
an is right about that; there is nothing to prevent pornography from having
artistic interest, even though it doesnt aim at having it. What remains to be
shown, though, is that as pornography it can be art, not just that it can have
artistic interest.
The third consideration Kieran addresses comes closest to that on which
my brief against pornographic art is based. It is that there is an appreciat-
ive problem about pornographic art, which makes it impossible to appreci-
ate an object as art and as pornography at the same time, because attention
to its artistic aspect entails inattention to its pornographic aspect, and vice
versa. Kierans response to this, naturally, is to deny the conflict; however, he
supports this denial only by arguing against views, such as those of Martha
Nussbaum and Roger Scruton, according to which pornographic interest is
necessarily objectifying or depersonalizing, and showing that it need not be.
But this seems like a red herring. What needs showing is that attending to
 Erotic Art and Pornographic Pictures 267

an image in order to be sexually aroused by it does not conflict with attend-

ing to an image for its artistic features, not that attending to an image for
its arousal value is not in conflict with regarding the depicted subject as a
person.
Properly characterized, and not simply as highly explicit erotica, porno-
graphys central aim, to facilitate sexual arousal in the name of sexual release,
though it does not, as Kieran notes, preclude artistic interest from being pre-
sent in an image, does, contrary to what Kieran says, militate against a viewers
artistic engagement with the image, because it enjoins treatment of the image
as transparent, as simply presenting its subject for sexual fantasizing, thus
entailing inattention to the form or fashioning of the image. Hence if some-
thing answers to the central aim of pornography it cant at the same time
answer to the aims of art. Thus, at the least, nothing can be coherently pro-
jected as both pornography, in the strict sense, and art.

VI

It might be objected to my claim that the status of art and the status of porno-
graphy are mutually exclusive that pornography, understood as erotic imagery
aimed at facilitating sexual arousal, fantasy, and release, does not preempt
artistic interest on a viewers part, even when it is recognized as pornography,
since some pornography works precisely by engaging the artistic interest of the
viewer. The idea, in other words, is that some images are more arousing for
some viewers when such viewers attend to or concentrate on aspects of the
image as such, such as its form or style or embodied point of view, rather
than merely being affected by them unwittingly in various ways. I see no
harm in granting this, for perhaps there are viewers whose arousal is enhanced
by attending explicitly to aspects of the vehicle of arousal. But even in such
cases, so long as the image is being regarded as pornography, aspects of the
image are not being appreciated for their own sakes, but only as instruments
to more effective arousal, fantasy, and release. If such images are intended to
be so regarded, then they constitute a complex mode of pornography, aimed at
a cognitively atypical viewer, rather than instances of erotic art per se. That

For more on the alleged objectification of depicted subjects by makers or consumers of

pornography, see my Sexual Perversity.
Or at least, cognitively atypical male viewer. For it has been suggested to me that what I here
label atypical is closer to typical for female viewers.
268 Pictures

sort of pornographywhat one might call artful pornography, consisting of

images that invite attention to their artistic aspects precisely so as to enhance
sexual arousal or fantasy involvement on the viewers partis perhaps the
only serious challenge to the claim that art and pornography are necessarily
disjoint. But that challenge fails nevertheless. Though for reasons already giv-
en the strategy of artful pornography is generally self-defeating, even when it
is successful, and arousal is achieved precisely in virtue of the viewers atten-
tion having been drawn to the artistic aspect of the image, if such drawing of
attention is entirely in the service of arousal aimed at, then the image remains
pornography, however artful, and not art.
So there is art that has a pornographic, or intentionally arousing, dimen-
sion. And there is pornography that has an artistic, or aesthetically interest-
ing, dimension. But the former is not thereby pornography, and the latter is
not thereby art. What usefully defines and differentiates pornography and art
are their central aims, and those aims are incompatible. One requires form/
vehicle/fashioning to be transparent, while the other requires them to be, at
least in some measure, opaque.
What is arousing in pornography, generally speaking, is imagining inter-
acting with or doing things to the depicted, usually unclothed, person, not
pausing on the manner or means of depiction. Transparency of medium is
all to the good of arousal, and is thus a virtual sine qua non of pornography.
Opaqueness of medium is all to the good of art, but invariably weakens, and
sometimes even wholly undermines, arousal.
It is instructive in the present connection to consider some famous
examples of European painting which clearly flirt with the status of
pornography. Consider Courbets notorious The Origin of the World, which
graphically displays the midsection of a nude reclining woman, or in a
less flagrant vein, Ingress Turkish Bath or Velasquezs Toilet of Venus or
Bronzinos An Allegory. Were these, perhaps, the pornography of their day,
despite the fact that they now grace the walls of our finest museums of
fine art? Only, I suggest, if we are speaking hyperbolically. The arousal of
male viewers was undoubtedly part of the intention with which they were

In terms of a formulation offered recently by Dominic Lopes in order to capture what it is

to have an aesthetic interest in a photograph, when one consumes a pornographic photograph
one is decidedly not appreciating seeing the photographed object through the photograph (Lopes, The
Aesthetics of Photographic Transparency, Mind 112 (2003): 43348). Rather, one is relishing the
sight of the photographed object simpliciter, the better to be aroused by and to fantasize about it.
The vehiculing of that sight by photographic means is not an object of interest at all.
 Erotic Art and Pornographic Pictures 269

painted; in the case of the Courbet, made on commission for a pashas private
viewing, such arousal was perhaps the overriding intention. But surely the
makers of those paintings intended as well that the attention of viewers be
directed to features of the works themselvesthe handling of paint, the
arrangement of forms, the play of internal perspectivesand the relation
of those features to manifest sexual content. Thus those paintings count as
erotic art, notwithstanding the fact that they can be regarded or employed
pornographically, as perhaps they were by some of their owners.
The work of Egon Schiele, perhaps the greatest of erotic artists, is also prob-
lematic for my thesis, but can be handled similarly. The inconvenient fact is
that Schieles manifest intention for many of his sexually themed drawings
was indeed pornographic, since they were created expressly for male patrons
with precisely that sort of use in mind. There are, from my perspective, two
ways to deal with this fact. The first is simply to accept that, on the con-
ception defended here, those drawings must be accounted pornography, but
pornography that it is uncommonly aesthetically rewarding, and otherwise
justifiable, to treat as erotic art. The second is to posit for those drawings an
implicit artistic intention as robust as the explicit pornographic one, as evid-
enced by the unmistakable aesthetic features of those works, in virtue of which
they can be accounted, though uneasily, erotic art after all.

VII

Viewed in a certain light, it might seem odd that there is even such a thing
as erotic art. The appeal of art is fairly clear, as is that of pornography, but
why should there be something that, so to speak, straddles the two, given their
inherent opposition?
A key to the distinctive appeal of erotic art, I think, lies in this. In normal
circumstances, sexual stimulation leads beyond itself, to sexual arousal and
sexual activity. Sexual stimulation is normally a preliminary, an antechamber,
a way station in relation to what follows. It is thus not something to which one
attends, or on which one concentrates, for itself. But with erotic art, the stim-
ulation definitive of the category is hitched to a concern for the formal and

That said, were one to maintain that the Courbet, in particular, really does fall on the side of
pornography, so prominent is its intention to arouse its target audience, I would not demur.
Uneasily, again, because the pornographic projection of those works is in conflict with their
artistic projection.
270 Pictures

other features of the stimulating representation that anchor it in the broader

category of art, a concern which thus inhibits such stimulation from taking
its usual path towards sexual arousal and activity. We are thus constrained to
dwell in and on our state of stimulation without going beyond it, appreci-
ating instead its relation to and interaction with the other aspectsformal,
expressive, social, politicalof the representation that occasions it. Part of
what we enjoy in a work of erotic art can thus be described as a kind of ten-
sionone between life and art, to put it simplya tension that generates an
edgy pleasure akin to, though not identical with, that of the sublime.

VIII

Before concluding let me return very briefly to the other philosophical ques-
tions concerning erotic art formulated at the beginning of this essay. One
concerns how, with erotic art, stimulation or arousal are compatible with
aesthetic appreciation, where this is understood to entail some degree of disin-
terestedness or capacity for contemplation. A short answer is that stimulation
and arousal have to be held in check, neither suppressed nor given completely
free rein, and attention made to focus on the erotic qualities of the picture in
relation to the formal means that achieve or underlie it.
The other concerns how, with erotic art, the erotic quality of a work and
its artistic value are related. A short answer is that the more erotic a picture is,
while not becoming effectively pornographicthat is, such as to induce full-
blown sexual arousalthe better as art, provided the pictures erotic dimen-
sion is interestingly and intimately related to the other dimensions of the
pictures content, something that it may take sensitive interpretation to estab-
lish. Beyond that it would be unwise to generalize.

IX

The central argument of this essay can be stated as follows:

1. Erotic art consists of images centrally aimed at a certain sort of recep-
tion, R1.
2. Pornography consists of images centrally aimed at a certain sort of recep-
tion, R2.
3. R1 essentially involves attention to form/vehicle/medium/manner, and so
entails treating images as in part opaque.
 Erotic Art and Pornographic Pictures 271

4. R2 essentially excludes attention to form/vehicle/medium/manner, and so

entails treating images as wholly transparent.
5. R1 and R2 are incompatible.
6. Hence, nothing can be both erotic art and pornography; or at least, noth-
ing can be coherently projected as both erotic art and pornography; or at
the very least, nothing can succeed as erotic art and pornography at the
same time.
This page intentionally left blank
 PART I V
I N T E R P R E TATI O N
This page intentionally left blank
 16
Two Notions of Interpretation

I. There are many roads into the contested domain of interpretation. One
road, which I will pursue here, centers on an opposition between two notions
or modes of interpretation, which correspond roughly to two questions that
might be asked about an artifact, event, or object. These questions are What
does it mean? and What could it mean?. I will refer to the first of these modes
as the DM (does mean) mode and to the second of these as the CM (could
mean) mode. The first might also be labeled the determinative mode of inter-
pretation, the second the exploratory mode of interpretation.
An issue concerning interpretation in the CM mode that confronts us at
the outset is how to construe the could in the question what could X mean?
There are several possibilities here. We can construe could epistemically, the
question then becoming what could X mean, given what is known?; or we
can construe could logically, the question then becoming what is it logic-
ally possible for X to mean?; or we can construe could more pragmatically,
yielding a question such as given both what is known and what is logically
possible, what might X reasonably be taken to mean?. I will refrain from
stipulating a construal of the could involved in CM interpretation, but will
assume that the force appropriate in a given case will make itself evident.
Another preliminary matter requires attention. I will here be concerned
only with interpreting as an activity which seeks the meaning, significance,
purpose or role of that on which it is directed, and which issues in an inter-
pretation stating or formulating some such meaning, significance, etc. Call

First published in A. Haapala and O. Naukkarinen (eds.), Interpretation and its Boundaries (Helsinki:
Helsinki University Press, 1998), 221.
There is an obvious affinity between what I am calling exploratory interpretation and
what Arthur Danto has dubbed deep interpretation (Deep Interpretation, in The Philosophical
Disenfranchisement of Art (New York: Columbia University Press, 1986)), and even more between my
notion and what Umberto Eco has dubbed overinterpretation (Interpretation and Overinterpretation
(Cambridge: Cambridge University Press, 1992)). I will forgo, however, any attempt to pin down
the respects in which the three notions differ.
276 Interpretation

this semantic interpreting. Obviously, semantic is here to be understood

fairly broadly.
I am thus putting to the side, most notably, the activity labeled interpretive
in the performing arts which consists in playing, singing, reciting, dancing,
staging, or enacting a pre-existing work in a particular way. Call that per-
formative interpreting. This is not to deny that there are important relations
between performative interpreting and semantic interpreting, nor to deny
that performative interpretation may reflect or contribute to semantic inter-
pretation of something (such as the work being performed). It is just to insist
that performative interpreting is not literally the same activity as semantic
interpreting; that is to say, it does not, as such, seek and propose answers to
the question of what something does or could mean.
I will also put to one side the sort of interpreting that is done, in real-
time, by translators of spoken language for the benefit of those who do not
speak that language. Call that translational interpreting. Though such activ-
ity indeed aims at rendering the meaning of a stretch of discourse, in most
cases it involves little that is either decipherative or conjectural; that is, the
meaning involved is evident to the interpreter, and its reformulation in the
other tongue is a largely, if not wholly, mechanical affair.
II. Before proceeding I recall some received wisdom about semantic inter-
pretation.
1. Interpretation standardly involves the formation and entertaining of
hypotheses, the weighing of possibilities of meaning, significance, purpose, or
role in regard to a given phenomenon or thing.
2. Interpretation standardly involves conscious thought, deliberate
reflection, explicit reasoning, or the like. Not all perception or understanding
or apprehension is properly viewed as interpretive; some such is clearly pre-
interpretive, and serves as that on which interpretation rests, or that from
which it departs.
3. Interpretation standardly presupposes the nonobviousness of what is
being interpreted; if one simply and securely sees that X is F, if there is no
question of choosing or deciding to do so, then remarking that X is F is not a
matter of interpreting it.

See Ludwig Wittgenstein, Philosophical Investigations (London: Macmillan, 1958).

See Richard Shusterman, Beneath Interpretation, in Pragmatist Aesthetics (Oxford:
Blackwell, 1992).
See Annette Barnes, On Interpretation (Oxford: Blackwell, 1988). See also Goran Hermeren,
Expression, Meaning, and Nonverbal Communication, in Jeanette Emt and Goran Hermeren
 Two Notions of Interpretation 277

I will more or less accept these three features as definitive of any activity
worth labeling interpretive: such activity is hypothesis-involving, thought-
involving, and concerned with nonobvious attributions of meaning, signi-
ficance, purpose, or role. I do not claim, however, that these features are
sufficient to mark an activity as interpretive. They may not be. For it seems
that at least many cases of scientific inquiry and philosophical theorizing dis-
play those features without being what we would comfortably categorize as
semantically interpretive activities.
III. The following is a more or less random list of semantically interpreted
things, in the rather broad sense sketched above:
(1) a persons facial expression
(2) a persons gesture or bearing
(3) a persons action or behavior
(4) a rock in the desert
(5) tea leaves at the bottom of ones cup
(6) lines on the palm of ones hand
(7) an ambiguous phrase or sentence
(8) an ambiguous visual figure (e.g. a duck-rabbit)
(9) a situation (e.g. a swinging door or a missing car)
(10) an official act (e.g. the closing of an embassy)
(11) a medical symptom (e.g. a rash or a fever)
(12) an x-ray or radiograph
(13) a Rorschach blot
(14) a work of literature
(15) a philosophical argument
(16) the starry heavens
(17) animal markings
(18) readings and measurements
(19) jokes and witticisms
(20) metaphors and aphorisms
(21) a stain on the sidewalk
(22) an unexpected natural event
(23) a hand shadow figure
(24) movements on the stock exchange
(25) the deliverances of an oracle

(eds.), Understanding the Arts: Contemporary Scandinavian Aesthetics (Lund: Lund University
Press, 1992).
278 Interpretation

(26) a social custom

(27) a graph or diagram
(28) a chess move of ones opponent
(29) a tablet of cuneiforms
(30) a crossword puzzle clue
(31) dreams and daydreams
(32) a passage in the Bible
(33) a book in the Library of Babel
I am inclined to think, in reviewing this list, that the idea can be upheld
that there are basically two different modes of interpreting, and correspond-
ing to them, two different motivations for interpreting. For each of the above
items, either the DM mode or the CM mode of interpreting recommends
itself immediately, or else the two modes strike us as equally appropriate.
Consider some examples. Querying a rash on a childs chest (11), say, seems
straightforwardly a matter of DM interpreting; we want to know what it does
mean, in the sense of what underlying medical condition it betokens. Query-
ing an ambiguous figure (8) is just as clearly a matter of CM interpreting: we
want to know in what ways it could be seen. A puzzling situation in real life
(9) prompts us to discover its actual, and not merely possible, significance,
and thus calls for DM interpreting. Rorschach testing (13), a psychoanalyt-
ic procedure aimed at eliciting construals of any sort, with none counting as
correct or incorrect, is thus a paradigm of CM interpreting.
On the other hand, the starry heavens (16) may well be seen as calling both
for explanation, whether cosmological or theological, and so DM interpret-
ing, and also scanning with an eye for portents, and so CM interpreting. A
hand shadow figure on a wall (23) may be regarded as an occasion for seeing
in a number of waysCM interpretingor else as simply pointing to what
the shadowmaker had in mind to projectDM interpreting. A passage in the

In commenting on this essay in its earlier incarnation Peter Lamarque (Objects of Interpret-
ation, Metaphilosophy 31 (2000): 96124) suggests that to classify Rorschach blots as phenomena
appropriate for CM rather than DM interpreting may be to misclassify them, on the grounds that
Rorschach blots are used diagnostically, as signifying something about the subject who responds to
them. But I remain unconvinced. It may be true that what a person P sees in a Rorschach blot R means
a number of things as regards Ps personality, recurrent fantasies, and unconscious conflicts, at least
if the theory behind the diagnostic use of such blots is valid, but it does not follow that Rorschach
blot R means those things. In other words, that reactions to a given blot are psychologically revealing
doesnt show that there are in fact answers to what a given blot means, nor that we are seeking such
answers in reacting to them. In short, that responses to Rorschach blots mean something doesnt
show that the blots themselves mean something.
 Two Notions of Interpretation 279

Bible (32) may be taken by some, in DM mode, to have just a single mean-
ing, fixed by the intentions of the divine author, while others, operating more
in CM mode, take the passage as a springboard for tracing numerous at least
possible meanings.
But what are the two motivations that correspond to the two modes of
interpretation we have identified, and which different phenomena call forth
in differing degrees? I suggest that behind DM interpreting, in any sphere, lies
a spirit that might be qualified as scientific, practical, and knowledge-seeking.
Part and parcel of this spirit is a desire for understanding, explanation, discov-
ery, or communication. Moved by this spirit we strive to establish or further
our grasp of or contact with the real, whether in the form of nature in gener-
al or some particular human nature, that is, some individual mind or minds.
Behind CM interpreting, by contrast, lies a spirit that might be qualified as
ludic, liberated, and freedom-seeking. Central to this spirit is a desire for cog-
nitive play, much like that which Kant located at the core of the aesthetic,
without a concern for cognitive payoff of a concrete sort, a fascination with
possibilities of understanding, explanation, discovery or communication, but
no care for their actuality. Moved by this spirit we strive to deepen our appre-
ciation of alternatives, of the spaces in which meaning is formed, but without
privileging the actual or the crystallization of meaning that takes place therein;
rather than seeking secure knowledge of the world, natural or man-made, we
prefer to glory in our imaginative apparatus for doing so.
One thing that can be expected to affect the sort of interpretive mode we
adopt in confronting a given phenomenon, of course, is whether we under-
stand it as having been designed for interpreting, or offered for interpretive
activity, rather than as something that simply exists or occurs. But contrary
to what one might at first think, there is no automatic inference from an
assumption of intentional projection to the appropriateness of a DM mode
of interpretation, nor from an assumed absence of intentional projection to
the appropriateness of a CM mode of interpretation.
The relationship between the presence or absence of intention behind a
phenomenon and the mode of interpreting aptly brought to bear on that phe-
nomenon is a complicated one. On the one hand, jokes, radiographs, and
grimaces are products of intention that invite DM interpreting, while meta-
phors and ambiguous figures and moves in chess are products of intention
that invite CM interpreting. On the other hand, tracks, traces, and earth-
quakes invite DM inquiry even though they are not the products of intention,
while certain equally non-intentional phenomena, such as sidewalk blotches,
280 Interpretation

tea leaf patterns, or the constellations of the night sky, strike at least some of
us as an invitation to CM inquiry. Interpretation may be the characteristic
method of the human sciences, and especially in its exploratory mode, but not
all exploratory interpretation aims at the reconstruction or decipherment of
intention.
Now a reason someone might have for refusing to recognize what I call
exploratory interpretation as interpretation at all is this. Interpretation, it
might be said, presupposes the possibility of misinterpretation: misinterpret-
ation could be claimed to be a necessary correlate of interpretation. But if
so, then exploratory interpretation might be disallowed on the grounds that
nothing under that rubric counts as misinterpretation.
Certainly in the case of determinative interpretation misinterpretation has
an obvious place. If you interpret something to mean what it does not, and you
are operating in determinative mode, then you are misinterpreting it. But it is
not clear that misinterpretation can get no toehold in the case of exploratory
interpretation. Admittedly misinterpretation cannot there be glossed as it is
for determinative interpretation, or else virtually all exploratory interpretation
would count as misinterpretation, since it does not supply what a given item
does mean. But misinterpretation in exploratory mode might still have an
application, applying in those cases in which what is offered in response to
the question of what something could mean is something that the given item
could not conceivably mean. For instance, muddy footprints on the doorstep
just could not mean that 3 is the cube root of 27, and a pained expression
just could not mean the Statue of Liberty. Of course, for misinterpretation of
an item to be possible it must in some sense be possible that the item has a
meaning, but exploratory misinterpretation as just glossed requires no more
than that. Thus, in sum, even if interpretation does presuppose the possibility
of misinterpretation, that is no obstacle to counting exploratory interpretation
as interpretation.
IV. I will shortly examine more closely a number of other items on my list
of semantically interpretable things. I will continue to be interested in the
question of whether all varieties of interpretive activity can be illuminatingly
categorized as aiming at answering either the what does it mean? or the what
could it mean? question, but it is fair to say that a positive answer to that
question serves as a working postulate of my investigation.
In addition to that postulate of two basic modes of and motivations
for interpretingwhat might be called Thesis 0 of this essayI offer
 Two Notions of Interpretation 281

for consideration some additional theses, which will find support in the
subsequent discussion.
Thesis 1: DM or determinative interpretive inquiry presupposes, at least
defeasibly, that there is a single answer to the question of an items meaning
or significance, while CM or exploratory interpretive inquiry presupposes, at
least defeasibly, that there are many such answers.
Thesis 2: In some cases CM inquiry is only heuristic or instrumental to
engaging in DM; in other cases CM inquiry is engaged in for its own sake.
When engaged in for its own sake, CM inquiry has no predetermined stop-
ping point, and might in principle continue indefinitely.
Thesis 3: DM inquiry invariably involves a phase or moment of CM inquiry
engaged in heuristically or instrumentally. That is, in the course of determin-
ative inquiry one will invariably consider what an item could mean or signify,
in the interests of turning up reasonable candidates for what the item does
mean or signify.
I turn now to a more detailed look at certain items on my list.
V. An X-ray or radiograph. Consider some typical radiological judgments:
Thats a tibia fracture, This means intercranial bleeding, or That looks to
be an ovarian cyst. What I have in mind are cases one might describe as easy
calls. At first blush, such cases appear to have nothing of the conjectural or
hypothetical about them. That is to say, they are cases in which identification
of what is shown on the X-ray occurs automatically and unreflectively, cases
in which, at least to the radiologist, what is shown is perfectly apparent. Is this
sort of radiology then not interpreting after all? I would argue that it still is.
Why? Here are some possible reasons. (a) Because such cases are continuous
with other cases, that is, most or at least many occasions of radiology, which
are more clearly interpretive. (b) Because such cases are instances of a cer-
tain activity, namely, reading a radiograph, that has an inherently interpretive
status. (c) Because such cases involve applying an expertise, which allows radi-
ologists to discern what is not open to normal vision. Though these reasons
point in the right direction, none of them seems quite right. We will arrive at a
more satisfying answer shortly.
Of course not everything that is perceived is interpreted, as we have already
noted. When I see cups, bowls, or spoons on the kitchen table I dont interpret
those objects as cups, bowls, or spoons, nor arguably do I even see them as cups,
bowls, or spoons. I simply see cups, bowls, or spoons, as the case may be, and
perhaps that they are such. Ordinary (that is, non-artisanal) tableware is not
as such intended to be read or construed, is not designed to carry information,
282 Interpretation

does not call for deciphering; a cup is a cup, but does not mean cup. By contrast,
a given radiograph may show, indicate, indeed mean, such and such a medical
condition, e.g. fracture, tumor, or murmur. Just seeing that something is a
radiograph is not normally the end of cognitive engagement with it.
Return now to the problem of the easy call, such as that provided by a
practiced radiologist on an unproblematic radiograph, and the claim, ventured
above, that this should still be classified as interpretive, by contrast with the
unreflective seeing of spoon or bowl at breakfast. The reason is that though
the radiologist concludes more or less immediately that the X-ray indicates
a fracture, cyst, or whatever, this is still arguably a matter of concluding, as
opposed to simply registering. And part of the reason for so viewing it is that
the radiologist has the responsibility to rule out other possible readings, either
explicitly, through actual consideration of alternatives, or else implicitly, in
virtue of his ability to tell one condition from another in radiographic terms,
an ability acquired via training in which explicit consideration of relevant
alternatives would have figured prominently. In other words, the radiologists
reading of the X-ray is an act in which alternative construals are either manifestly
or latently involved, and is thus not to be assimilated to cases of unmediated
and unproblematic perception. A field of alternatives is part of the activity of
reading a radiograph in a way it is not part of everyday unselfconscious seeing.
This serves, I believe, as an illustration of Thesis 3 above, that DM inter-
preting involves, at least defeasibly, a phase of CM interpreting. For clearly
unproblematic radiography, like radiography in general, is to be understood
on the DM model of interpreting if understood as interpretive at all. But as
we have just seen, it may be right to so classify it precisely because it implies a
space of possibilities acknowledged, in a prior and exploratory moment, even
if this is a space which has been collapsed down to one possibility in the judg-
ment actually issued.
VI. A move in a game of chess. If approached in the DM, or determinative
mode, one attempts to infer from ones opponents last move, in conjunc-
tion with the general position of the board, what ones opponents plan is
for the next stage of the game. Doing this presupposes that there is such a
plan, that ones opponent is not simply moving randomly or instinctively,
and that that plan is singular and determinate. With that assumption in effect,
one attempts to discern what specific succeeding moves ones opponent envis-
ages will ensue. Interpretation in this guise is a matter of ascertaining what the
given move means, in the sense of inducing or inferring the concrete scheme
in which it figures in the mind of ones opponent.
 Two Notions of Interpretation 283

By contrast, if one adopts the CM or exploratory mode of interpretation,

one will seek to ascribe different plans to ones opponent, imagining different
possible justifications for the move in question, but without assuming psycho-
logical reality for any of them. One might interpret a chess move in CM mode
either for the intrinsic interest of doing so or in order to be better prepared
for various eventualities in the working out of the game or the various paths
it might take. In CM mode, which evolution of the pieces on the board is
most plausibly ascribed to ones opponent as the plan he has actually in mind
to carry out is but one focus of interest in the move that ones opponent has
just made. Equally important is the whole field of possibilities a given move
opens up, possibilities that can be viewed as meanings of the move in ques-
tion, in the sense of representing alternative objectives that ones opponent
might have had in view, whether or not he actually did.
VII. A metaphor. In interpreting a metaphor, is one primarily seeking what
it does mean, or what it could mean? That is, does one naturally adopt DM
mode or CM mode? Here is an example I came across recently, a saying
attributable to the novelist Balzac: Fame is the sun of the dead. Now, what
could that mean? It could mean that as the sun sustains the living, fame
sustains, or gives life to, the dead. Or it could mean that as living things nat-
urally turn toward the sun, those contemplating death naturally turn toward
the prospect of posthumous fame, for the consolation it may provide. Or it
could mean that the dead constitute a kind of solar system, one centering on
the property of fame, viewed as some kind of a Platonic entity. Or it could
mean that too close an approach to fame by the dead will cause their wings
to melt.
But what Balzacs metaphorical remark does mean, I take it, is some subset
of the things it could, with no restrictions on appropriateness or plausibility,
mean. Thus, in the present case, what it means is arguably given, at least in
part, by the first two suggestions, but not by the last two. The principle of
such selection, roughly speaking, is something like what meanings it would be
plausible to hypothesize that the utterer of the metaphor intended to convey
to hearers by uttering the metaphor in a given communicative context.
In such cases, one asks initially what it could mean, but with the ultimate
aim, in most cases, of arriving at an answer to what it does mean. Of course

This reflects a hypothetical intentionalist perspective on literary meaning. (See Hypothetical

Intentionalism: Statement, Objections, and Replies, Ch. 18, this volume. See also Whos Afraid
of a Paraphrase?, Ch. 17, this volume, for further discussion of metaphor.)
284 Interpretation

what a metaphor does mean may very well, and usually does, exhibit multipli-
city and ambiguity, but if so, there is a sense in which this simply enters into
the meaning, broadly viewed, that it determinatively has. This is thus another
illustration of Thesis 3.
VIII. A clue in a crossword puzzle. Were one to regard a crossword puzzle
as a sort of found object, as an unpremeditated conundrum posed by a for-
tuitous constellation of lexical elements, a conundrum which perhaps has an
intelligible solution and perhaps does not, nothing would be more natural
than to adopt toward it a CM interpretive stance. Since on that assumption
there would be no question of the solution to the puzzle, the most one could
do would be to hypothesize various possible readings of clues with an eye to
making answers to them fit together with answers to other clues, arrived at
provisionally in the same vein.
Now on the usual and well-founded assumption that an ostensible cross-
word puzzle is an actual crossword puzzle, devised by human hand, we will
also entertain different ways in which its clues might be takenfor example,
as straight, ironic, joking, elliptical, or self-referential. But note that we will
do this only insofar as it is necessary to pin down the sense intended by the
puzzle maker, one guaranteed to lead to an answer that works in solving
the puzzle as a whole. So the ordinary exercise of completing a crossword
puzzle, it seems, is one that involves a determinative stance toward the clues
provided, however exploratory one may wax in the name of such determ-
ination. For our interpretive querying has then a clear target: namely, what
was in the puzzle makers mind in offering us the clues that he did. For only
in ascertaining that will one arrive at the correct solution to the puzzle in
question.
IX. A volume in the Library of Babel. Jorge Luis Borgess marvelous work of
fiction The Library of Babel, the lesser known and less discussed companion
piece to Pierre Menard, Author of the Quixote, elaborates a conceit that is,
in my opinion, at least as intriguing and as fertile as that which figures in
Pierre Menard. As described by Borges, the Library of Babel is an immense,
cyclical, and unending repository of all books possible within certain fixed
parameters. More specifically, the Library contains all possible books of the
following format: 410 pages in length, 40 lines to a page, 80 or so characters
to a line, and employing 25 orthographic symbols, to wit, 22 alphabet letters
plus the comma, period, and space. The number of such books, while not
strictly infinite, is mind-bogglingly large, as carrying out the permutational
calculation reveals: it is greater than 25 to the 1,300,000th power. Moreover,
 Two Notions of Interpretation 285

the length limitation on individual volumes is no real obstacle to the Librarys

containing books of length greater than 410 pages, such as Prousts A la
recherche du temps perdu or Gibbons Decline and Fall of the Roman Empire;
they too exist there, only spread out over several volumes of this sprawling
universe of books. Now of course only a minuscule fraction of this googolplex
of tomes will appear to be in readable English, only a minuscule fraction
in almost-readable French, and only a somewhat larger, though still tiny,
fraction will appear to be written in any known human language. But so long
as there are hands willing to separate the wheat from the chaff, there would
seem to be much worth harvesting.
Thus the first reaction to such a cornucopia, as reported by Borgess narrat-
or, is spontaneous euphoria at the idea of all possible books, in which all con-
ceivable human knowledge, wisdom, insight, and literary achievement would
thus be contained. But this is soon succeeded by the realization that such a
Library must prove entirely worthlessmuch inferior, in fact, to that of the
average junior college. For texts stripped of intentional projection, of context
of generation, of even a specified language in which they are to be understood,
are not works of literature of any kind, and have no determinate content. As
pure texts, that is, as strings of syntactic elements that can be thought of as
deriving from a process of permutation on a fixed set of such elements, there
is nothing to choose among them. What seems, at first glance, to provide all
the meaning in the world in fact provides no meaning at all. The prospect
of an infinity of meanings, which results when interpretation of an item is
entirely unconstrained, turns out to issue in an absence of meaning. Borges
strikingly underlines the uselessnessfrom the point of view of knowledge or
discovery, at leastof even the seemingly most cogent volumes in this most
disillusioning of libraries, by pointing out that every such volume has a multi-
tude of near-doppelgangers that would appear to negate, contradict, or call into
question whatever it was that the given volume might be taken to be saying or
conveying.
The Library of Babel, Borges tells us, is literally a sphere. But it is also, fig-
uratively, a sphere for the purest expression of the CM interpretive impulse.
Pick up any volume, posit a language or scheme of translation in which it is,
at a ground level, to be understood, then posit, in the manner of the renewing

Actually less than a googolplex, but still much more than a googol. (A googol is 10 to the 100th
power, while a googolplex is 10 to the googolth power.)
That would seem a fair description of French without any diacritical marks to individuate
vowels.
286 Interpretation

readings recommended at the end of Pierre Menard, an authorial intention

and historical situation for the resulting text, and then seek to discern what
further meaning or significance might reasonably be ascribed to such a text.
Next, begin again, with new values of the two variables. In this activity, whose
game-like character is evident, there is no truth, no fact, about what the given
volume does mean. But as it turns out, this is all the more liberating, allowing
us the greatest possible freedom to exercise our conjectural and combinatorial
powers concerning what something could mean. From the point of view of
CM interpretation, at any rate, the Library of Babel is a valuable resource, an
inexhaustible springboard to hermeneutical high jinks without limit. It also
provides, incidentally, a confirmation of Thesis 2 above.
X. Though we have only tested its mettle on a fraction of the thirty-three
items on the list in section III, we can now be said to have some reason to
regard the DM/CM opposition as a fruitful lens through which to view what
can seem a bewildering variety of interpretive situations. We can also be said
to have some reason to regard as sustainable as well the three theses enunciated
earlier in connection with that opposition.
Yet it is perhaps time to observe that DM interpreting and CM interpret-
ing are, in a sense, not as separable as my discussion to this point may have
suggested. That DM interpreting invariably involves, if passingly, a phase of
CM interpreting is of course something already enshrined in Thesis 3. But
CM interpreting also exhibits an involvement in the reverse direction, in the
following sense: Asking what a given phenomenon could mean presupposes,
for its intelligibility, at least the idea of what the phenomenon does mean, if
not of course the existence of such meaning. Thus our two modes, though
from one angle contrasting and alternating, can from another angle be seen as

It is a short step from this to the idea of virtual artworks as the product of willful abandonment
of all intentional and contextual constraints on interpretation in regard to a given actual artwork,
holding fixed only manifest or perceivable structure on some construal of that. It can be argued,
however, that virtual artworks, though not without interest, are ultimately much less valuable and
important to us as human beings than are actual artworks. (See Jerry Fodor, Deja vu All Over Again:
How Dantos Aesthetics Recapitulates the Philosophy of Mind, in Mark Rollins (ed.), Danto And
His Critics (Oxford: Basil Blackwell, 1993), for development of the related idea of virtual etiologies
for actual artifacts generally.)
A possible objection to this is as follows: It may be that in most cases it could mean X can be
glossed as it is possible it does mean X. But in some cases, for example, those of Rorschach blots,
it is arguable that it cannot be so glossed. A response: it is true that when inviting interpretations
of Rorschach blots you are interested in what someone just might see in the blot. But if such an
interpretation is conceived as what the blot could mean, as opposed merely to what one might see
in it, then it seems the usual gloss, reflecting the claimed conceptual link between CM and DM
interpreting, would have to be allowed.
 Two Notions of Interpretation 287

mutually implicating. Neither is logically whole, so to speak, without the oth-

er. That interpretation in CM mode in some sense presupposes interpretation
in DM mode might be seen as an illustration of the Aristotelian point that, at
least conceptually, potentiality presupposes actuality.

It should be admitted, finally, that in many cases of interpretation we seem to be concerned

neither with what something could mean, nor with what it does mean, but rather with what
it is, seeking what Georg von Wright called explicative interpretation. But whether explicative
interpretation falls within the ambit of semantic interpretation, even broadly conceived, is another
question.
 17
Whos Afraid of a Paraphrase?

1. Having never written on the topic of metaphor before, which rightly

or wrongly I had always considered a fairly marginal one as far as aesthet-
ics was concernedso much fuss for one little trope!I decided to begin
where many modern discussions begin, that is, with Donald Davidsons sem-
inal essay on the topic. I attempt to orient myself in this debate by fixing
on certain of Davidsons assertions in that essay and offering my reactions to
them. As will be seen, I come down pretty clearly on the anti-Davidsonian
side of the fence, in that I regard the idea of metaphorical meaning as ulti-
mately defensible, as long as one correctly identifies what has such meaning,
correctly locates how such meaning is acquired, and acknowledges that such
meaning is perhaps not all there is to a metaphor, depending of course on how
broadly or narrowly one chooses to deploy the notion of meaning. Certainly if
metaphorical meaning is restricted to what is in principle capturable by para-
phrases, there is indeed more to metaphor than that. But there is also, and
undeniably, it seems to me, that as well.
There is indeed a tendency, when the object is to challenge the adequacy
of paraphrase for the elucidation of metaphor, to construe the idea of para-
phrase as implicitly importing an ambition of paraphrase without remainder,
thus implying that paraphrases, if sufficiently elaborated, give the whole of
what a metaphor is about and can thus do duty for them. Evidently, insofar as
such an unstated importation is not remarked, the pretension of paraphrase to
a role in the elucidating of metaphor will be unfairly denied, being held to be
necessarily more sweeping than it in fact need be.
The fact that the task of exhibiting in literal language the metaphorical
meaning of a metaphor might not, perhaps, ever be completely dischargedthe

First published in Theoria 67 (2001): 723.

What Metaphors Mean, in Inquiries into Truth and Interpretation (Oxford: Oxford University
Press, 1984). Page references to the edition cited are given in parentheses.
 Whos Afraid of a Paraphrase? 289

fact that it might always be possible to expand or supplement the paraphrases

with which one seeks to cash out such meaningshould not be thought to
license the inference that therefore the task cannot be carried out, and thus
that the paraphrases offered at any given point necessarily fail to articulate any
part of the meaning that a metaphor possesses.
2. But to begin, now, with Davidson.
(1) The concept of metaphor as primarily a vehicle for conveying ideas,
even if unusual ones, seems to me as wrong as the parent idea that a meta-
phor has a special meaning . . . if I am right, a metaphor doesnt say anything
beyond its literal meaning (p. 246).
What this assertion most clearly overlooks is the fact that there are phases in
the comprehension and reception of a metaphor, in the acceptance of it as apt
or just, once proposed. Perhaps oversimplifying, we might posit a first phase
in which the metaphor, typically an evident falsehood if construed literally,
generates cognitive dissonance, shakes up associations, induces seeing-as of an
unfamiliar sort, and so on, all of which is by design, and a second phase in
which, once the dust settles, certain paraphrasable meanings, if the metaphor
is minimally effective, precipitate out of the first, intentionally disorienting
phase.* That the meaning of a metaphor only emerges at this second phase,
while evading us during the first, is no reason, of course, for failing to recog-
nize the existence of such meaning.
(2) The supposed figurative meaning of a simile explains nothing; it is
not a feature of the word that the word has prior to and independent of the
context of use, and it rests upon no linguistic customs except those that gov-
ern ordinary meaning. (p. 255).
This pronouncement fails to recognize that even if a metaphor is based
on nothing but the pre-existing literal meanings of its constituent words,
there may yet be a subsequent, metaphor-specific interpretive custom
or mini-practice that forms around a successful metaphor, so that it
becomes, even before eventual pasturage as a dead metaphorand one
might remark, parenthetically, that all metaphors should be so luckyan
available descriptive resource of the language, one capable of right or wrong
employment.*

Typically but not necessarily, as is shown by examples of twice true metaphors, such as No
man is an island, brought to our attention notably by Ted Cohen. (See his Metaphor and the
Cultivation of Intimacy, Critical Inquiry 5 (1978): 113, and Metaphor, Feeling, and Narrative,
Philosophy and Literature 21 (1997): 22340). Here is another such metaphor: Life is no bowl of
cherries.
290 Interpretation

(3) . . . what we attempt in paraphrasing a metaphor cannot be to give

its meaning, for that lies on the surface; rather we attempt to evoke what the
metaphor brings to our attention (p. 262).
First of all, if it were really true that the meaning of a metaphor lies on
its surfaceand one presumes, in construing this metaphor, that this is
a smooth, unbroken surface, free of nooks, crannies, or potholes in which
things can hidewould anyone ever need to have the meaning of a metaphor
explained? But more importantly, this assertion ignores the fact that what is
evoked by a metaphor often crystallizes in a fairly pronounced mannerthat
is to say, significant intersubjective convergence as to what the metaphor
recommends to our attention manifests itselfthus allowing to paraphrase
the job of articulating, at least partially, the meaning of the metaphor.
Ted Cohen remarks much the same thing in a recent essay on our subject:
The metaphorical content of a metaphorical expression is more or less
specific . . . [while] there is no function that will calculate the metaphorical
content of an expression from its literal meaning . . . there is a content that it is
correct to take from the expression.
Back to Davidson:
(4) . . . in fact there is no limit to what a metaphor calls to our attention,
and much of what we are caused to notice is not propositional in character.
When we try to say what a metaphor means, we soon realize that there is no
end to what we want to mention (p. 263).
This declaration, finally, is blind to the fact that there are many things
which, while they are indeed resemblances or similarities between the terms
of the metaphor, are arguably no part of what the metaphor means, or is
plausibly taken as conveying, as the history of subsequent interpretation of the
metaphor would establish. For example, it is no part of the meaning of the
metaphor No man is an island that no man is made of sand, whereas it is
determinably part of such meaning that no man is inherently isolated from
the society of other men.

Metaphor, Feeling, and Narrative, 227. A similar observation is offered by Anders Engstrom,
in Metaphor and Ambiguity, Danish Yearbook of Philosophy 31 (1996): If there exists an inclination
towards certain interpretations within a community of speakers, it should be clear that a basis for
prescribing these would be possible, and that a notion of metaphorical meaning could be maintained
(pp. 1213).
Of course the determinacy of this meaning is partly a function of the metaphors embedding
in Donnes poetic discourse as a whole, but that does not affect the point at issue.
 Whos Afraid of a Paraphrase? 291

Richard Moran has made the point as follows:

It may be well be true, as Davidson says, that a metaphor makes us attend to some
likeness, but it is not true that attending to just any of the infinite aspects of likeness
between the two things counts as understanding the metaphor . . . the process of
interpretation couldnt even begin without some sense of which are the relevant
dimensions of the comparison.

Understanding metaphor, in other words, involves hitting on and attending

to the right likenesses, which the specific juxtaposition of the terms not cus-
tomarily brought together brings to the fore, in the given context in which
they are juxtaposed.
3. But the central claim in Davidsons brief against metaphorical mean-
ing, and that on which his argument seems almost entirely to turn, is the
claimI spare you a quote confirming that this is indeed the core of the argu-
mentthat in a metaphor the constituent words carry only their original, literal
meanings, and do not acquire new, metaphorical ones. Now with this claim I
am inclined to agree: being used in a metaphor, however successful, does not
affect the meaning of a word as a term in the language, even passingly. The
problem, though, is that this in no way yields the conclusion desired, that
metaphors, that is, metaphorical utterances, lack a meaning, acquired in con-
text, a meaning that one may as well call metaphorical meaning.
That in a metaphor the constituent words do not acquire new, metaphorical
meanings, ones that might figure eventually in the dictionary, in perhaps fifth
or sixth place, and that the metaphor would not work, would not perform as a
metaphor at all, were its constituent words not to retain their ordinary, preex-
isting, meanings, does not entail that the metaphorical sentence, or perhaps
better, the sentence taken as a metaphor, does not acquire in situ a metaphor-
ical meaning, one that paraphrases can be charged with exhibiting. It is hard
to underestimate the rhetorical benefit that Davidson draws from this usually

Moran, Seeing and Believing: Metaphor, Image, and Force, Critical Inquiry 16 (1989): 106.
See also, and earlier, David Novitz, Knowledge, Fiction, and Imagination (Philadelphia: Temple
University Press, 1987), ch. 7.
David Novitz, however, argues against Davidson that in a successful metaphor some constituent
words must in fact acquire new, metaphorical meanings, if the metaphor as a whole is granted to
have a new, metaphorical meaning (Knowledge, Fiction, and Imagination, 1546). But I think that
Novitz is impelled to this conclusion by not distinguishing sharply enough between sentences in a
language, whose meanings are perhaps solely a function of the meanings of their constituent words,
and utterances on an occasion, whose meanings are not.
292 Interpretation

unremarked shift from words or terms to sentences or utterances. The justice

of his position in regard to the former redounds undeservedly, and somewhat
surreptitiously, to the credit of the latter.
4. It is useful to compare the operation of metaphors to that of something
more homely, such as exclamations. Consider the exclamation Fire!.
Exclamations have meanings, in context, which go beyond the meanings
of their constituent wordsor in this case, wordand often in a stable
and persisting manner. This meaning is partly propositional, and thus
paraphrasable, e.g. there is a fire in the vicinity and everyone is urged to
leave, and partly non-propositional, consisting in an expressive, illocutionary
force, though one that can also be described well enough. Would it be
reasonable to argue that since the constituent word of this exclamation retains
its meaning in the exclamation itself, and since the exclamation would not
function as such were that word not to do so, therefore there is no further, as it
were, exclamational meaning to that exclamation, one that outlasts its use on a
given occasion? I think not.
Consider also more closely the issue of dead metaphors. Here are two from
roughly the same sphere, as regards their literal roots: you are the light of my
life and never again darken my door. Obviously such expressions now have
quasi-literal meanings themselves, as not very extended paraphrases of them
would confirm, such as, respectively, you impart meaning and joy to my life
and are central to it and you are not welcome in my home and I hope never
to see you again.
But now try to picture those dead metaphors in their youth, a youth
shrouded in the mists of time and no doubt unrecoverable.* Is it yet plausible
to think that what is captured in the paraphrases offered a moment ago, those
homespun distillates, was no part of the meaning of those metaphors when
newly minted, that those metaphors acquired such meaning, all of a sudden,
only after being laid definitively to rest in the graveyard of spent expressions?*
To take such a line, it seems to me, is to adopt a sort of doctrine of semantic
creation ex nihilo as regards metaphor, the prospects of which seem no brighter
than those of the parent doctrine in theology.
5. It is sometimes remarked that, even granted the differences between
individual words in isolation and complete utterances in context, Davidson
is still right to deny metaphorical meaning to metaphors. And that is because
the supposed meanings of metaphors, unlike the literal meanings of words in
a language or the literal meanings of sentences composed from them in rule-
governed ways, are entirely bound to occasions of use, are simply a matter of
 Whos Afraid of a Paraphrase? 293

what speakers intend to convey on such occasions, and so belong entirely to

pragmatics rather than semantics. The final upshot of this, then, is that that
they do not deserve to be called meanings at all, since they are not possessed
of any generality, stability, or exportability outside the specific circumstances
in which they see light, and thus have no explanatory value as far as semantic
theory is concerned.
But this sort of vindication of Davidson, however guarded, mistakenly
exaggerates the occasion-boundedness of metaphors. Although born on
given occasions, and acquiring concrete, if never completely paraphrasable,
significances in specific contexts of utterance, successful metaphors retain
such significances in enduring fashion, ones they carry with them on future
occasions of use, once successfully constituted. The acquisition of significance
by a metaphor on the occasion of successful use is often a relatively permanent
one, and the more so, one may suppose, the more just or compelling the
metaphor.
The significance of the metaphorical ascription of sunhood to a young
woman by her lover is now, some five hundred years after Shakespeares
Romeo and Juliet, hardly less stable, less exportable, less available for meaning-
making in the sense of speakers meaning, than the significance of a literal
ascription of sunhood to a particular heavenly body. That metaphors are born
in context and acquire their content therein, against a complex background of
shared understandings, assumptions, and dispositions, should not lead us to
think that metaphors necessarily perish once those originating circumstances
have past, and with them the content then acquired. No, the content of
many metaphorsthat is, metaphorical assertions, understood as species
of utterance is as general, stable, and exportable as that of most literal
assertions, and so would seem to count as semantic content of some sort.
If describing some thing as the opiate of the people, or the hobgoblin of
little minds, or a wound in the side of ones country, or a tale told by an
idiot and full of sound and fury, or someone as a snake in the grass, or a
utensil, or an Emma Bovary, or no Jack Kennedyand I have tried here
to avoid metaphors that would be accounted irretrievably deaddoes not

Qualified support of Davidson in this vein, preliminary to advancing a theory of metaphorical

meanings that would not be subject to such objection, can be found in Josef Stern, What Metaphors
Do Not Mean, Midwest Studies in Philosophy 16 (1991): 1352.
I was about to add, or an affair between a predicate with a past and an object that yields
while resisting, but then I had to admit that I couldnt think of a reasonable re-employment of this
wonderful metaphor of Nelson Goodmans, that is, to characterize something other than metaphor
itself.
294 Interpretation

qualify as employing an expression with an available, paraphrasable, occasion-

transcendent meaning, then I begin to be unsure what does. There is some-
thing called cultural literacy, in the sense E. D. Hirsch has made familiar, and
having a feeling for the sense of such metaphors, utilizable outside the site of
their original application, is a part of it if anything is.
6. Another source of skepticism as regards metaphorical meanings has its
roots not in Davidson, but in the later Wittgenstein. Wittgensteins discus-
sion of secondary uses of language, presupposing yet diverging from those
which can be considered primary, is taken by some to show the pointless-
ness of postulating secondary senses borne by such language in its secondary
uses, in addition to the senses such language carries in virtue of its primary
uses, uses that ground the language game in question. But cases of the sort dis-
cussed by Wittgenstein in his remarks on secondary uses of language, designed
to forestall the postulation of parallel secondary senses, seem to me very differ-
ent from those of successful metaphor.
Wittgensteins favored examples, recall, concern the colors of vowels, e.g.
e is yellow, or degrees of stoutness of the days of the week, e.g. Tuesday is
thin. Now such examples of, let us call it, quasi-metaphorical assertion have an
evident whimsicality, idiosyncrasy, its-a-free-country-ism about them that
is quite foreign to cases of genuine metaphor. In genuine metaphor, there is
a rightness, or click, in the disparate things that the metaphor brings into
conjunction, which unleashes a meaning for the sentence that was all along a
potential for specific illumination, if an unnoticed one, of the literal meanings
of the terms employed once conjoined. But even were a certain predispos-
ition to view e as yellow as opposed to other colors, or to view Tuesday as
thin as opposed to fat, to be observable in a given linguistic community, that
would hardly be enough to make yellow or thin a permanent descriptive
resource of the language of that community in regard to vowels or days of the
week. Thus, the fact that secondary uses of language of this sort do nothing
to prompt recognition of corresponding senses of the terms involved does not

As David Hills suggests in a recent insightful essay on our topic, bona fide metaphors possess
some degree of poetic power (Aptness and Truth in Verbal Metaphor, Philosophical Topics
25 (1997): 119). Hills usefully identifies and defends two positions on metaphor that he labels
aestheticism (metaphors are rightly assessed aesthetically, that is, for aptness) and semanticism
(metaphors have semantic content, that is, a distinctive paraphrasable meaning). As should be clear,
these are positions to which I also subscribe.
The idea that Tuesday might generally be regarded as thin, incidentally, has always struck
me as especially misguided, given the conflicting force of the concrete and, one might add, public
association of Tuesday and fatness in the holiday of Mardi Gras.
 Whos Afraid of a Paraphrase? 295

cast much, if any, suspicion on the idea that bona fide metaphorical uses of
language might be said to underpin graspable metaphorical meanings corres-
ponding to them.
7. Elsewhere I have developed a view of literary interpretation, or at least
of the central meaning of a text offered as literature, that I call hypothetical
intentionalism. Insofar as metaphors are utterances, and insofar as they can
be seen in particular as literary utterances, albeit of a small-scale sort, then
a hypothetical-intentionalist view of literary meaning would seem naturally
to recommend itself for getting a handle on metaphorical meaning as well.
What, then, does such a view maintain?
According to hypothetical intentionalism a literary work is to be construed
as an utterance, one produced in a public context by a historically and cultur-
ally situated author, where the central meaning of such a work is thus a form
of utterance meaning, as opposed to either textual meaning, the meaning of the
brute text as a string of words in a language, or utterer meaning, the meaning
the utterer, speaker, or author had in mind and intended to get across. Utter-
ance meaning, in turn, is understood on a loosely Gricean model according
to which what an utterance means is a matter, roughly, of what an appropri-
ate hearer would most appropriately take an utterer to be trying to convey
in employing a given verbal vehicle in the given communicative context. As
applied to literature, and fleshed out in certain ways, what it amounts to is
roughly this: the core meaning of a literary work is given by the best hypothes-
is, from the position of an informed, sympathetic, and discriminating reader,
of authorial intent to convey such and such to an audience through the text in
question.
In the light of that explication, then, we can give a fairly straightforward
hypothetical-intentionalist account of the fact, emphasized by a number of
writers on metaphor, that understanding a metaphor requires identifying
or homing in on the right likenesses or connections between terms in a
given case, putting aside those that are not to the purpose. For correctly

See my Intention and Interpretation in Literature and Messages in Art, both in The Pleasures
of Aesthetics (Ithaca: Cornell University Press, 1996), and Hypothetical Intentionalism: Statement,
Objections, and Replies, in M. Krausz (ed.), Is There A Single Right Interpretation? (University Park,
PA: Penn State University Press, 2002), and Ch. 18, this volume. I am not alone in advocating
such a view: William Tolhurst, Alexander Nehamas, Gregory Currie, and Stephen Davies subscribe
to something similar, and the broadly Gricean outlines of the view will, at any rate, be familiar.
In addition, certain current competing views, such as the actual intentionalisms defended by Noel
Carroll, Robert Stecker, and Paisley Livingston, are perhaps more hypothetically intentionalist than
they appear to be on the surface.
296 Interpretation

understanding a metaphor in context, by hypothetical-intentionalist lights,

involves arriving at the best hypothesis, in epistemic and aesthetic senses, as
to what a speaker would likely have intended to convey or draw attention to,
what similarities a speaker would plausibly have wanted to highlight or bring
into focus, bearing in mind that this will naturally open out into resonances
that a speaker would have endorsed or welcomed in connection with the
metaphor offered, if not ones it is plausible to conjecture were foreseen or
perceived in advance.
Now even if this is accepted, and the identification of metaphorical
meaning, a species of utterance meaning, is understood to run along
hypothetical-intentional lines, it remains true, in my opinion, that whether
or not a given utterance is in fact a metaphorical one is irreducibly an affair
of actual intentions, to be determined by whatever means serve generally
to determine the actual intentions of utterers. That is, in suggesting the
plausibility of a hypothetical-intentionalist view of metaphorical meaning,
as a species of literary meaning, I am not thereby proposing, and in fact I
would explicitly disavow, a hypothetical-intentionalist view of metaphorical
status. For indeed anything of a certain linguistic form, roughly, that of
an assertion, positive or negative, might be being projected as a metaphor
or with a metaphorical intent, and thus the most plausible construction we
might put on an utterance might categorize it incorrectly. On the other hand,
in the absence of knowledge of whether something is in fact a metaphorical
assertion, it would be reasonable to invoke a hypothetical-intentionalist
principle as a justification for revisably ascribing or withholding metaphorical
status to a given utterance.
8. Until now I have been highlighting the character of metaphors as pos-
sessors of relatively stable and graspable meanings, and as reusable on occa-
sions other than those in which they first arise. I want at this point, however,
to underscore an aspect of metaphors that runs in rather the opposite dir-
ection, that is, as against their, as it were, complete exportability. What I
call attention to now is the inseparability of content and form in metaphor,

For discussion, see Intention and Interpretation in Literature.

Thus I can concur with this remark of Max Blacks regarding assignment of metaphorical
status, but only in the sense of defeasible assignment absent knowledge of actual intention: Our
recognition of a metaphorical statement depends essentially upon two things: Our general knowledge
of what it is to be a metaphorical statement, and our specific judgment that a metaphorical reading
of a given statement is here preferable to a literal one (More About Metaphors, in Andrew Ortony
(ed.), Metaphor and Thought, 2nd edn. (Cambridge: Cambridge University Press, 1993), 356).
 Whos Afraid of a Paraphrase? 297

something which, reasonably construed, I take to be valid for all artistic phe-
nomena.
It can be asked whether metaphor should be conceived as primarily a con-
ceptual matter, as primarily a linguistic matter, or rather as one that essentially
straddles the divide. Is the essence of a metaphor a sentence in a language,
that is, a sequence of words, or is it rather a constellation of concepts, which
different strings of words make available? If the latter, how can those concepts
be identified apart from a given linguistic formulation, given the presumed
impossibility of exact translation of almost any term in a natural language
into any other? Does the possibility of neutral designation of the concepts
involved in a given metaphor require us to presuppose something like a lan-
guage of thought, or at least some identity of concepts across users of different
languages, however that comes about?
I am inclined to the view that regards the concrete vehicle of a metaphor
more or less those words, in that language, in that order, with just those pre-
cise rhythms, resonances, and prosodic propertiesas ineliminable, and thus
as not after all the mere vehicle, but the very body and soul of the metaphor.*
That is to say, even once the cognitive content of a metaphor is approxim-
ated through acceptable paraphrases, and its imagistic force, that in virtue of
which we are made to see one thing in terms of another, is identified as well,
there seems to remain a residue that attaches to the specific feel of the words
employed to invoke that paraphrasable meaning and put that imagistic force

I do not assume that verbal metaphors are literally speaking works of art, but only that they
can be appropriately treated in many respects as if they were, e.g. they can be assessed for their
aesthetic merit. See Arnold Isenberg, On Defining Metaphor (Aesthetics and the Theory of Criticism
(Chicago: University of Chicago Press, 1973), 10524), who aptly observes that metaphors are
always strokes, if not always works, of art.
The view that metaphor is essentially a conceptual matterthat what are called conceptual
metaphors are the fundamental ones, and that ordinary verbal metaphors are just the verbal
expression or externalization of such underlying conceptual metaphorsis a view that at present
enjoys considerable currency, owing principally to the work of George Lakoff, Mark Johnson, and
Mark Turner. (See Lakoff and Johnson, Metaphors We Live By (Chicago: University of Chicago
Press, 1980); Johnson, The Body in the Mind (Chicago: University of Chicago Press, 1987); Lakoff
and Turner, More Than Cool Reason: A Field Guide to Poetic Metaphor (Chicago: University of
Chicago Press, 1989); Lakoff and Johnson, Philosophy in the Flesh (New York: Basic Books, 1999).)
I dont wish to deny that there may be, on some psychological level, what these thinkers call
conceptual metaphors. I only wish to insist on the at least equal claim to reality of metaphors as
specific pieces of language, and the prima-facie irreducibility of the latter to the former. (For similar
cautions see Hills, Aptness and Truth in Verbal Metaphor.)
See Morans two-part analysis of metaphor, in Seeing and Believing, according to which
metaphors have, on the one hand, a paraphrasable cognitive content, and on the other, an evidently
non-paraphrasable imagistic force, consisting in a prescribed, asymmetric viewing of one thing
through the lens of another.
298 Interpretation

in play. In short, the verbal substance of a metaphor bids fair to be considered

essential to it, and not just its conceptual structure.
But perhaps we should recognize two notions of metaphor, or two criter-
ia of identity for metaphors, a restrictive and a permissive. On the permissive
criterion, the metaphor is detachable from its linguistic formulation, and is
a sort of thought, or way of thinking, with respect to certain things. On the
restrictive criterion, the metaphor is undetachable from its linguistic formu-
lation, and is thus like a miniature work of poetry, akin to a haiku. Which
of these notions or identity criteria is more apt would seem to depend on the
context of inquiry into metaphor. Thus, it seems advisable to recognize both.
At any rate, with a metaphor considered as something like a work of literary
art, if a tiny one, we must experience the content of the metaphor, including
both its paraphrasable meaning and its imagistic force, through its specif-
ic verbal form, if we are to fully appreciate the metaphor. That outstanding
metaphors can usually be translated from one language to another should not
lead us to forget that the result is a translation, in which some part of the meta-
phor has been lost, in that part of the metaphorical charge of the original is
indissolubly bound up with specific words, specifically deployed.
Un bon mot in one language is not necessarily un bon mot in another, to
which that very phrase bears witness; for a good word neither really means
the same thing as, nor has the same fluid sound as, un bon mot, which sound
contributes to its being the expression it is and to its conveying what it does.
Metaphors are certainly not guaranteed to survive largely intact when trans-
lated, even faithfully, from one language to another. Man is a wolf to man is
one metaphor, in English, but another, Lhomme est un loup pour lhomme,
in French. For notice that the metaphor in French suffers, in comparison to
the metaphor in English, in possessing to a slightly lesser extent, because of
the unavoidability of articles in French, the quasi-palindromic quality which
is a feature of both versions, and which reinforces the idea of reciprocity of
behavior that is at the conceptual core of the metaphor. Assonance, allit-
eration, symmetry, syncopation, and so on are all part of a metaphor as a
verbal entity. Translations of metaphors from one language to another, in
short, might with justice be considered metaphorical cousins of one another,

In Max Blacks terms, I am suggesting that metaphors are paradigmatically emphatic: A

metaphorical utterance is emphatic to the degree that its producer will allow no variation upon or
substitute for the words used (More About Metaphor, 26.)
The Latin original of the metaphor, ascribed to St Francis of Assisi, is homo homini lupus,
which, again, has its own peculiar linguistic flavor.
 Whos Afraid of a Paraphrase? 299

rather than simply the same metaphorical individual in different linguistic

dress.*
9. In approaching the writing of this essay, my first idea, I confess, was
to compose it entirely in metaphor, as a way of demonstrating rhetorically
that effective metaphors, construed in context, have a paraphrasable meaning
operative within a linguistic community, whatever non-paraphrasable force
or charge they may have as well, and can thus communicate a set of thoughts,
advance a position, perhaps even ignite a revolution.
Since that seemed, on reflection, rather too hard to carry off, I decided
instead to cast an interested eye on the novel I had just begun rereading,
David Lodges Small World, combing it for figures with which to test the
validity of the convictions about metaphor I have been airing here. As it turns
out that novel, being written in a casually elegant but not especially lyrical
style, is not overly rich in metaphor, generally achieving its effects by other
means, notably similes, which in comparison are fairly thick on the ground.*
In any event, here is an almost complete listing of metaphors from the first
fifty pages:
1. Dismay had already been plainly written on many faces (p. 3).
2. . . . the narrow beds, whose springs sagged dejectedly in the middle (p. 3).
3. He exhaled rather than pronounced the syllables (p. 9).
4. . . . he bent forward over his roast shoe-leather (p. 11).
5. The rest of the audience was performing the same tableau of petrified bore-
dom as before (p. 16).
6. . . . you have succumbed to the virus of structuralism (p. 27).
7. . . . the individual campus is . . . the heavy industry of the mind (p. 43).
I take it that few of us would have much trouble in discerning what is
conveyed by these figures, in elaborating the connotations the metaphorically

I have, however, employed a larger than usual number of at least somewhat live metaphors
in this essay, which I can now reveal is what the unexplained asterisks after certain sentences the
observant reader will have noted are meant to mark. But, and this was the point of my little conceit,
I take it that none of those metaphors impeded understanding of the cognitive content of this essay,
and may even have facilitated that understanding.
David Lodge, Small World (London: Penguin Books, 1984). Page references are in parentheses.
The qualification almost is required partly because the borderline between metaphorical
senses and second-order or third-order literal senses is irredeemably blurry. So for example, the list
given does not include His face darkened as he added . . . (p. 7), since I regard darkened there as
literal, but the case might be made that darkened in such a context is simply a very tired metaphor.
Full disclosure also prompts me to add that, in those first fifty pages of Lodges novel, two characters
end up discussing a metaphor, that of a woolly fold as it occurs in Keatss poem The Eve of St.
Agnes.
300 Interpretation

invoked predicate carries in relation to the literal subject to which it is, with
a characteristic degree of unusualness, applied. For example, as regards the
last and most ambitious of these metaphors, the individual campus is therein
painted as unwieldy, as on its way to obsolescence, as perhaps even harmful to
the health of its familiars.* Note that the specific novelistic context is not even
necessary for the construal of these metaphors, except in the case of the third,
where it perhaps helps to know that the syllables exhaled form the name of a
girl, Angelica, whom the exhaler has fallen in love with moments after making
her acquaintance.
But rather than devote any more time to metaphors of middling quality,
I look briefly to the opposite ends of the metaphorical spectrum, addressing
first two metaphors which are in my opinion especially fine, as well as being,
as it happens, good examples of humor, and then those that have, as it were,
fallen off the scale, in the sense that it is impossible, or virtually impossible,
to construe them as metaphors at all, despite their partaking of the canonical
form A is B where A is B is manifestly false.*
Here are two candidates for excellence in the category of metaphor, the one
encountered recently, the other an old favorite:
Lotteries are a tax on the mathematically challenged.
A wife is an umbrella; sooner or later one hails a cab.
Ill spare you any paraphrase of these metaphorical assertions, in the first
case because its sense is fairly transparent, in the second case because, though
its sense is not immediately transparent, it is more delicious if deciphered on
ones own.
I turn now to the outcasts of the society of metaphors, those sorry would-
be tropes consisting in manifestly false identity sentences that simply resist
metaphorical redemption.* Note first that it is in fact rather difficult, and
perhaps impossible, to find such entirely metaphor-resistant sentences. It is
instructive to consider Richard Morans throwaway example of a supposedly
completely profitless yoking of two items, the taste of sugar and the discovery
of America, which, so he claims, must fail to strike any metaphorical sparks
because the things brought together have inherently no relation to one
another.* But the statement The taste of sugar was the discovery of
America would actually not be too bad a way of conveying, metaphorically,

Offered by Freud in his Jokes and Their Relations to the Unconscious (New York: W. W.
Norton, 1963).
Seeing and Believing, 106.
 Whos Afraid of a Paraphrase? 301

that the exploitation of cane sugar in the Caribbean islands was a driving force
in the further exploration of the Americas.
Nevertheless, it seems to me that almost metaphor-resistant sentences can
be found, though perhaps not at random. One formula for balking metaphor-
icity would seem to be to equate assertively things from the same family or
category, at roughly the same level of specificity. Thus it appears difficult, if
not impossible, to make anything metaphorically of gold is lead, a fox is a
wolf, or every woman is a man.
On the other hand, given a bit more context and a crucial third element in
the mix, a metaphor equating things on the same level from the same family can
sometimes work remarkably well, as in this curious witticism: Wagner is the
Puccini of music. Now, what does this mean? What is the cognitive content
or imagistic transformation that, once caught sight of, causes us to laugh? It is
admittedly hard to say. But here is a stab at it. Puccini, through being invoked
to position Wagner in relation to music, seems to have been placed outside
of music entirely, a rather grave insult to a composer, one has to admit, while
Wagner, for his part, comes off hardly better, having been figuratively equated
with a composer who was not, so to speak, even a composer at all.
In any case, to revert in closing to what is probably my main theme, note
how this metaphor, once effectively essayed, becomes an available resource,
one deployable, with suitable substitutions, on other occasions, and with what
one sees no good reason, ultimately, not to regard as a metaphorical meaning.
Thus, Rachmaninov and Respighi might also be called, of course unfairly,
Puccinis of music, Queen the Puccini of rock music, George Winston the
Puccini of jazz, Renoir the Puccini of painting, Gore Vidal the Puccini of
American literature, and so on. For all I know, I may well be the Puccini of
metaphor studies. At any rate, Ive made a start toward the title.
As I say, difficult, but not impossible. Thus gold is lead might be employed, in the right
context, to express disapproval of earthly riches, or admiration for golds specific gravity. Work
in linguistics on the permeability and interconnectedness of semantic fields would in fact argue
against there being any cases of sentences whose metaphorical deployment was to be precluded
absolutely. (See Eva Kittay, Metaphor: Its Cognitive Force and Linguistic Structure (Oxford: Oxford
University Press, 1990).) Max Black offers this as an example of a likely metaphor-proof form
of words: a chair is a syllogism (More About Metaphor, 23). Black is silent on why such a
predication resists metaphorical interpretation, but one may speculate that it turns on both the
utter dissimilarity and the categorial remoteness of the terms involved.
What are the musical failings for which Wagner directlyand Puccini indirectly, if more
harshlyare here being reproached? It is hard to say for sure, but I imagine windiness of expres-
sion, cheapness of effect, and dramatic overblownness are among the failings targeted. (I hasten to
add that I am not here endorsing those charges!)
 18
Hypothetical Intentionalism: Statement,
Objections, and Replies

1. In an earlier essay I have defended at some length a view on literary

interpretation which I call hypothetical intentionalism. The view centers on
the idea that a literary work should be seen as an utterance, one produced
in a public context by a historically and culturally situated author, and that
the central meaning of such a work is thus a form of utterance meaning, as
opposed to either textual meaning or utterer meaning. Utterance meaning,
in turn, is understood on a pragmatic model according to which what an
utterance means is a matter, roughly, of what an appropriate hearer would
most reasonably take a speaker to be trying to convey in employing a given
verbal vehicle in the given communicative context. As applied to literature,
and fleshed out in certain ways, what it amounts to is this: the core meaning of
a literary work is given by the best hypothesis, from the position of an appro-
priately informed, sympathetic, and discriminating reader, of authorial intent
to convey such and such to an audience through the text in question. Thus
hypothetical intentionalism is a perspective on literary interpretation which
takes optimal hypotheses about authorial intention, rather than actual authorial
intention, to provide the key to the central meaning of literary works.
Since the key notion here is that of a best hypothesis on the part of
readers, which would seem to entail the consideration by them of a variety
of hypotheses, it might seem that hypothetical intentionalism was committed
to a picture of literary interpretation as a species of what I have elsewhere

First published in M. Krausz (ed.), On the Single Right Interpretation (University Park, PA:
Pennsylvania State University Press, 2002), 30918.
See Intention and Interpretation in Literature, in The Pleasures of Aesthetics (Ithaca, NY:
Cornell University Press, 1996): 175213. The view is further deployed and developed in my
Messages in Art, also in The Pleasures of Aesthetics: 22441.
For an initial statement of the view, see William Tolhurst, On What a Text Is and How It
Means, British Journal of Aesthetics 19 (1979): 314.
 Hypothetical Intentionalism 303

characterized as CM, or could mean, interpretive activity, as opposed to

DM, or does mean, interpretive activity, the latter aimed at identifying
what, if anything, something does mean, with the former concerned rather
with what it is possible, in a manner of speaking, that something might
mean. But this is where a distinction between heuristic (instrumental)
and final (intrinsic) engagement in CM interpretive activity is important.
For hypothetical intentionalism, CM engagement with a work is purely
instrumental, and not undertaken for its own sake. The consideration of
various possibilities of construal serves only to identify what is in fact the best
hypothesisthat is, the most explanatorily plausible and, to a lesser extent,
aesthetically charitable construction we can arrive atregarding a works
intended import. What a work in fact means, however multifaceted that may
be, remains the focus of inquiry.
2. Does a view like hypothetical intentionalism allow for a multiplicity of
distinct, at least nominally incompatible interpretations, when interpretation
is being conducted ultimately in a DM or determinative spirit? The answer
is yes, in virtue of the existence of ties or draws, roughly speaking, among
distinct and competing hypotheses concerning a works import. That is to
say, nothing precludes there being, in a given case, two or more informed
hypotheses frameable as to authorial intent that are explanatorily and aes-
thetically optimal. Even so, there are ways in which to view such multiplicity
so that it becomes, from a certain angle, unitary. I suggest that it is always
possible, in principle, to combine competing reasonable first-order interpret-
ations of a work into a totality that is embraceable from a more encompassing
perspective.
The fact that we have no logical notion handy for representing the ensemble
of acceptable interpretations taken together does not show that the most cor-
rect and comprehensive interpretation of a work of art is anything other than
that ensemble. The logical notions that naturally suggest themselves, conjunc-
tion and disjunction, are both, in different ways, misleading or unsuitable.
We dont wish to say that the overall correct interpretation of W, where R1,
R2, R3 are individually acceptable interpretations of W, is just R1 or R2 or
R3, nor do we wish to say that it is just R1 and R2 and R3. Rather, it is each
and all of R1, R2, and R3, and yet not their simple conjunction or disjunc-
tion. The best, most correct and comprehensive, interpretation of a work of

See Two Notions of Interpretation, in A. Haapala and O. Naukkarinen (eds.), Interpretation

and Its Boundaries (Helsinki: Helsinki University Press, 1999), and Ch. 16, this volume.
304 Interpretation

art subject to multiple individually justifiable readings must be an interpreta-

tion which enfolds all such readings: a kind of global or subsumptive reading,
so to speak, which acknowledges all the individually acceptable readings and
puts them into relation with one another. That is the sort of perspective I am
inclined to adopt toward those examples of multiple acceptable yet incom-
patible readings that have figured so prominently in discussions of artistic
interpretation, prompted by works such as Kafkas Castle, Jamess Turn of the
Screw, Becketts Waiting for Godot, and De Koonings Woman paintings.
To make this a bit more concrete, my global or subsumptive interpretation
of, say, Kafkas The Castle might run roughly as follows. The Castle reason-
ably admits of theological, bureaucratic, psychoanalytical, existential and epi-
stemological readings, and in ways that can be related to one another, mostly
reinforcingly; on the other hand it does not reasonably admit, say, of ento-
mological (castle as teeming beehive), chivalric (castle as prison of beautiful
damsel), or oneiric (castle narrative as dream report) ones, nor do such read-
ings relate reinforcingly to any of those already acknowledged as admissible.
There can be little doubt that if the individual interpretations of The Castle
invoked above, e.g. the theological or bureaucratic ones, have merit, then the
best interpretation, tout court, of Kafkas novel will be an inclusive one having
more or less the form indicated. If that is borne in mind, the narrowly logical
problem of reconciling multiple distinct and individually insightful readings
of such a work, with which philosophical discussion has been excessively pre-
occupied, will seem of reduced importance, or even to have disappeared.
3. But what, more precisely, is the form of global or subsumptive inter-
pretations of the sort I have invoked above? The form of such an interpretation,
I*, would seem to be something like this: Ws meaning is such that it is
partly given by/aptly viewed under interpretation 1, partly given by/aptly
viewed under interpretation 2, . . . and partly given by/aptly viewed under
interpretation n, where those embedded interpretations, I1, I2, and so on,
are understood as first-order sub-interpretations subject to the higher-order

A similar perspective is advanced by Robert Stecker, Art Interpretation, Journal of Aesthetics

and Art Criticism 52 (1994): 193206. Though all true statements are conjoinable, that may
not be the best way to hook up a pair of true interpretations into a more comprehensive true
interpretation . . . If The Turn of the Screw is intentionally ambiguous, it doesnt represent the
governess as battling with hosts and having hallucinations . . . Better, the novella is such that it
can be correctly read either as representing the governess as battling ghosts or as representing the
governess as having hallucinations (p. 201).
 Hypothetical Intentionalism 305

interpretation, I*, which subsumes them, though not simply disjunctively or

conjunctively.
We might further want to distinguish among such global or subsumptive
interpretations those which, like the above, are simply collective (or enumer-
ative), and those which are integrative (or hierarchic), and thus include an
account of the relations of importance or centrality obtaining among the sub-
interpretations brought together in the global interpretation.
Admitting global/subsumptive interpretations of an integrative/hierarchic,
rather than simply collective/enumerative, sort does open the door to
possible ties among competing such interpretations, ones that put different
weights on or differently position the sub-interpretations they acknowledge
in common. For from a hypothetical-intentionalist perspective, two different
ways of organizing in relation to one another the individually attractive sub-
interpretations of a work might be equally plausibly hypothesized, in light of
all the appreciatively relevant data, to be what the contextually understood
author ultimately wanted readers to grasp. Still, given how subtle the
differences are likely to be between two such integrative/hierarchic readings
acknowledging all the same individually acceptable sub-interpretations,
multiplicity at this level, even if not ultimately eliminable, must surely strike
one as not much of a qualification on the idea of there being such a thing,
grosso modo, as the meaning of a literary work.
4. It will be no surprise that the doctrine of hypothetical intentionalism
has failed to win universal acceptance among theorists of interpretation. Tex-
tualists, deconstructionists, and actual intentionalists still abound. What is
more, the unconverted, and particularly actual intentionalists, have not been
shy to voice their criticisms of the doctrine. I thus here outline some brief
responses to objections that have been raised to a hypothetical intentional-
ist account of literary meaning such as I have proposed, whose bare bones
were sketched above. The basic idea, recall, is that on such an account literary
meaning, the object of literary interpretation in a determinative mode, is con-
stitutively bound not to what a historically untethered text might be saying,
nor to an authors actual, psychologically real semantic intentions in compos-
ing the texteven ones that might be said to have been successfully realized
in the textbut to our best hypothesis, as ideally comprehending readers, as
to what the concretely situated and publicly available authors semantic inten-
tions were in composing the text he or she did.
306 Interpretation

Objection 1: Drawing a veil of ignorance across some aspects of a

works actual creative history and not others, as a hypothetical intentionalist
approach to work meaning enjoins, is unacceptably arbitrary.
Response: Erecting a rough cordon around essentially privatewhich is
not to say, epistemically inaccessibleinformation is hardly arbitrary from
a literary point of view. The making of literature is an individual, largely
interior endeavor, but it is also a public, convention-governed one, bound
by mutually understood rules for producing and receiving literary offerings.
These rules might quite naturally specify that facts related to context of origin
beyond what an ideally prepared and backgrounded reader could generally
be expected to know were irrelevant to fixing or constituting the meaning of
the work as an utterance in that context. The artists state of mind is not our
ultimate goal as interpreters of literary works, but rather what meaning can
be ascribed to those works, albeit as the indissociable products of those very
particular communicative agents; thus not all obtainable evidence as to the
artists state of mind is automatically germane to the project of delineating
what a work issuing from that mind and presented in a literary setting argu-
ably means.
Objection 2: Hypothetical intentionalism is committed to a communica-
tion model of the literary domain, but such a model does not sit well with
the appeal to idealized, as opposed to actual, audiences that is a feature of
sophisticated versions of hypothetical intentionalism.
Response: The communicative model arguably presupposed by literary
activity does not commit us to authors projecting their works for specific
and specifically envisaged audiences, ones contemporaneous with the author,
rather than, less restrictively, for whatever audiences, present and future, are
well suited to receive and understand the work in its historical, cultural,
and authorial context. Call the former the narrowly communicative model of
literary activity, if you like, and the latter the broadly communicative model.
The point is that communication with appropriate readerswhoever,
whenever, wherever they might beis still communication, even when such
readers are not narrowly identified or targeted in advance.

See Anthony Savile, Instrumentalism and the Interpretation of Narrative, Mind 105 (1996):
55376; Paisley Livingston, Arguing Over Intentions, Revue Internationale de Philosophie 198
(1996): 61533; and Robert Stecker, Artworks: Definition, Meaning, Value (University Park, PA:
Pennsylvania State University Press, 1997), ch. 10.
See Savile, Instrumentalism and the Interpretation of Narrative.
 Hypothetical Intentionalism 307

Objection 3: Allowing optimally hypothesizable intentions to trump actual

ones, where the basic nature or status of a work is concerned, opens the door
to an unacceptable level of indeterminacy as regards work content.
Response: Hypothetical intentionalism regarding work content or meaning
is rightly coupled with actual intentionalism as regards both the status of
works as literature and their category or genre location within literature.
Note that category or genre specifications can be taken as tantamount to
or generative of very general semantic, or perhaps metasemantic, intentions,
indicating what sorts of meaning, at the least, are to be sought in a given work,
which helps to dissipate the worry sounded by opponents of hypothetical
intentionalism that hypotheticism regarding authorial semantic intentions
will issue too readily in indeterminacy of meaning.
Furthermore, coupling actual and hypothetical intentionalism in this man-
ner strengthens the claim of literature so understood to be communicative,
in almost the narrower sense distinguished a moment ago, and for two reas-
ons. First, readers who attempt to arrive at meaning by hypotheticist lights are
entitled from the outset to know, and so ideally do know, at least what category
of offering they are dealing with. Second, the actual author, in being obliged to
show his opening hand, that is, vouchsafe to readers directly the approximate
nature, if not the precise import, of his work, thus does not remain entirely
behind a veil as far as the constitution of meaning is concerned.
Objection 4: The best hypothesis about authorial intention must, logically
speaking, be that which is correct; thus there can in fact be no divergence
between actual authorial intention and our best hypothesis as to what that
intention is or was.
Response: This is a simple misunderstanding. Obviously, best hypothesis
in the formulation of hypothetical intentionalism cannot be taken to mean
that which in fact happens to be correct, and so best in the sense of true.
Rather, the best hypothesis by hypothetical intentionalist lights is that which
we would have most reason to accept or adopt given the totality of evidence
that is both available and admissible, i.e. given the totality of what is derivable
from the text and its legitimately invoked surrounding context.

Ibid.
On the notion of categorial vs. semantic intentions in relation to a work of literature, see
my Intention and Interpretation in Literature, and in a more critical vein, Paisley Livingston,
Intentionalism in Aesthetics, New Literary History 29 (1998): 83146.
308 Interpretation

Objection 5: Even if there is a distinction, regarding authorial

intention, between a best hypothesis, in the sense invoked by hypothetical
intentionalism, and a true hypothesis, why should we ever favor the former
over the latter, once we have arrived at the latter, by whatever means we have
at our disposal? Surely in science we would not prefer our methodologically
soundest hypothesis regarding some state of affairs over what was in fact the
case, were we to learn what that was.
Response: This objection misunderstands the goal of literary interpretation
as conceived by hypothetical intentionalism, which is not to discover, for its
own sake, the authors intention in writing the text, as if criticism were at
base a matter of detective investigation, but to get at the utterance meaning
of the text, that is, what it not the author is saying, in its author-specific
context. Utterance meaning just is constitutively tied to a most reasonable
projection of utterers intent in the given context, and does not collapse into
utterers meaning. Thus even when the latter is available it does not displace
the former as the object of literary interpretationas opposed to biographical
sleuthing.
Objection 6: Hypothetical intentionalism that acknowledges the necessity
for interpretation to ascertain actual intentions of a categorial or constitutive
sort has already betrayed the vaunted autonomy of the literary workits
independence in a fundamental respect from its creatorthat it claims to
safeguard.
Response: This is not so. A restricted autonomy, to wit, as regards
resultant meaning, is still autonomy; furthermore, it is arguably the only
sort of autonomyas opposed, say, to that requiring detachability from

See Noel Carroll, Interpretation and Intention: The Debate Between Hypothetical and
Actual Intentionalism, Metaphilosophy 31 (2000): 7595. Carroll charges that to proceed in this
way would appear to be fetishizing our method over what the method is designed to secure (p. 83).
Thus the charge that hypothetical intentionalism simply substitutes warranted assertibility for
truth where literary interpretations are concerned (see Carroll, Interpretation and Intention, 84)
is similarly unjustified. Hypothetical intentionalism, at least when advanced with a background
commitment to metaphysical realism, retains that distinction, but relocates it with respect to the
items involved. For hypothetical intentionalism, a true literary interpretation of a work W by an
author A writing in context C is one given by what is, so to speak, optimally warrantedly assertible
about the intention with which A, writing in C, composed Ws text. But if a literary interpretation
is thus true, then it is more than just warrantedly assertible. (For further relevant discussion, see
Gregory Currie, Interpretation and Objectivity, Mind 102 (1993): 41528.)
See Gary Iseminger, Actual Intentionalism vs. Hypothetical Intentionalism, Journal of
Aesthetics and Art Criticism 54 (1996): 31926.
 Hypothetical Intentionalism 309

generative contextthat it seems important to insist on where literature is

concerned.
Objection 7: Hypothetical intentionalism, which identifies the core
meaning of a work with a best projection of authorial intentionwhere
such projection may not in fact coincide with any meaning actually
intendedthus appears in the last analysis to be not really concerned
with either the author or his achievement. Hypothetical intentionalism
unjustifiably severs the work from the agent who has created it.
Response: Again, this is not so. Hypothetical intentionalism accords the
semantic intentions of the actual author a crucial role, only it is a heuristic
rather than a final one. Authorial intention is what truth-seeking interpretive
activity necessarily aims at, the idea being that what one would most
reasonably take to be that intention, on the basis of the text and a full grasp
of its author-specific public context, yields a true interpretation of the literary
work, understood as an artistic utterance, which is embodied in the text.
As to severing a work from its author, hypothetical intentionalism pleads
not guilty; it simply insists that the meaning of a literary work, however
informed its interpretation must be by the authors public identity, and
even by certain of the authors actual intentions, is not constrained to being
just what the author intended it to mean, even where that intention is
fully compatible or consonant with the contextually situated text. And even
though a literary work is inextricably the work of just that author, in that
precise context, the author is not the ultimate arbiter of what his or her work
means, i.e. what it appears to convey or communicate to an appropriately
backgrounded reader. Finally, hypothetical intentionalism doesnt deny
authors their achievements, it just locates those achievements in the utterance
meanings their uttered texts attain, owing for the most part to the ingenuity
with which those texts have been contrived, and not in the utterers meanings,
which those texts also, in favorable cases, subserve.
Objection 8: The defense of hypothetical intentionalism ultimately rests on
the claim that it accords better with current interpretive practices than does
actual intentionalism, alleging that critics in framing interpretive hypotheses
do observe the proposed ban on inherently private information, such as direct

See Iseminger, Actual Intentionalism vs. Hypothetical Intentionalism, and Livingston,

Arguing Over Intentions.
Nor is the author in the best position to discern that, in any event. Authors, because of their
unique perspectives on their own works, are generally very far from being ideal readers of them.
310 Interpretation

but hidden authorial proclamations of what a work is intended to mean, and

so on. But in fact this is not the case.
Response: Hypothetical intentionalism does not ultimately rest on an
empirical claim about actual interpretive practices, taken in their full and
motley variety, but rather on what are arguably norms underlying the most
defensible of such practices, understood as ones that truly answer to our
interests in literature as literature. It is on that elusive and highly contestable
terrain that the dispute about the merits of hypothetical intentionalism must
be conducted, rather than that of statistical conformity or non-conformity
with current practice. Admittedly, a full case for hypothetical intentionalism
on those grounds remains to be made.
5. A literary work is an utterance, of course, but it is a sort of grand
utterance, one governed by different ground rules of interpretation than are
ordinary utterances. Our interests in literature are communicative ones, where
communicative is understood broadly, but they are not, pace certain recent
writers, more narrowly conversational ones. This means, in part, that in
literary contexts, unlike conversational ones, we have a prior and independent
interest in utterance meaning entirely apart from whatever utterer meaning
may stand behind or parallel that utterance meaning as constituted. In
conversation, if we dont understand what someone has said we may quite
properly get him to explain further what he meant, or to qualify or retract
his words. In literature, if we dont understand what a contextually situated
text is saying, we cannot legitimately demand explication from the author, or

See Carroll, Interpretation and Intention.

I note a further objection that is often raised against hypothetical intentionalism, one I admit to
finding more troubling than those reviewed above. The objection is that the distinction presupposed
by hypothetical intentionalism between essentially public and essentially private information regarding
an author, where the former enters into the appreciatively relevant context for the work while the
latter does not, is fundamentally untenable, or fatally blurry. (See Stecker, ArtWorks; Livingston,
Intentionalism in Aesthetics; and Carroll, Interpretation and Intention.) Clearly, however this
distinction is made out, it cannot be equated with that between published and unpublished
information, if only because that would have the implausible consequence that a works meaning,
i.e. what is given by a correct interpretation of it, would change upon publication of certain
appreciatively relevant facts about how a work came to be that it just happened were not known
outside of the authors immediate circle. This is not the place to attempt a full reconstruction
of the needed distinction, but one might begin to refine the concept of a works appreciatively
relevant public context by focusing on what the author appears to have wanted his or her readers
to know about the circumstances of a works creation, beyond what is implicit in the authors
previous work and the authors public identity. At any rate, such a thing will not fluctuate with
the contingencies of actual publication of the information in question. Finally, that the distinction,
however reconstructed, might remain blurry is not fatal to its utility.
See Noel Carroll, Art, Intention, and Conversation, in Gary Iseminger (ed.), Intention and
Interpretation (Philadelphia: Temple University Press, 1992).
 Hypothetical Intentionalism 311

instruct him to modify his offering; at most we can ask him to confirm that
the text he has given us is indeed the text he wants us to have, as we set about
to interpret it as literature.

For further defense of hypothetical intentionalism and criticism of actual intentionalism, see
Saam Trivedi, An Epistemic Dilemma for Actual Intentionalism, British Journal of Aesthetics 41
(2001): 192206, and Gregory Currie, Interpretation in Art, in Jerrold Levinson (ed.), Oxford
Handbook of Aesthetics (Oxford: Oxford University Press, 2003), 291306.
This page intentionally left blank
 PART V
A E S T H E T I C P RO PE RT I E S
This page intentionally left blank
 19
Aesthetic Properties, Evaluative Force,
and Differences of Sensibility

In this essay I revisit a position on aesthetic attributions I have held for some
time, a position rooted in some seminal essays of Frank Sibley, and which can
be labeled aesthetic realism. I reflect on some challenges which have emerged
to the viability of that position, and determine what accommodations, if any,
are called for. First I sketch the position, borrowing with modification from
an earlier short essay of mine. I then formulate a number of worries about the
position which have lately come into view, and try to see where they lead.

Aesthetic attributions to works of art, and the terms used to effect such attri-
butions, are largely descriptive; that is to say, they are based on, and obliquely
testify to the occurrence of, certain looks, impressions, or appearances which
emerge out of lower-order perceptual properties. Insofar as an aesthetic attri-
bution is intended as objective, that is, as the attribution of a property of inter-
subjective import, such looks or impressions or appearances are relativized to

First published in E. Brady and J. Levinson (eds.), Aesthetic Concepts: Essays after Sibley (Oxford:
Oxford University Press, 2001), 6180.
Aesthetic Concepts, Philosophical Review 68 (1959): 42150, of course, and other essays
mentioned in subsequent notes. The view defended also owes something, in broad measure, to
writings of Monroe Beardsley and Kendall Walton.
Being Realistic About Aesthetic Properties, Journal of Aesthetics and Art Criticism 52 (1994):
3514. An earlier and fuller statement of the position, though with different emphasis, can be
found in my Aesthetic Supervenience (1983), reprinted in Music, Art, and Metaphysics (Ithaca:
Cornell University Press, 1990).
See Frank Sibley, Aesthetic and Non-aesthetic, Philosophical Review 74 (1965): 13793;
Frank Sibley, Objectivity and Aesthetics, Proceedings of the Aristotelian Society Supplement 42
(1968): 3154; Monroe Beardsley, The Descriptivist Account of Aesthetic Attributions, Revue
Internationale de Philosophie 28 (1974): 33652. See also Goran Hermeren, The Nature of Aesthetic
Qualities (Lund: Lund University Press, 1988).
316 Aesthetic Properties

a perceiver who views a work correctly, and thus approaches the condition
of what has been called, following Hume, a true critic or ideal judge. That
means, in particular, someone who properly situates a work with respect to its
context of origin, including its place in the artists oeuvre, its relation to the
surrounding culture, and its connections to preceding artistic traditions.
It is true, of course, that some terms used in aesthetic discourse are com-
pletely, or almost completely, evaluative. That is to say, they are terms with no
or almost no descriptive content, meaning they do not imply anything about
the kind, category, or nature of the object to which they are applied. Sibley
called these solely evaluative terms. Such terms either just denote a degree
of aesthetic value or disvalue believed by the speaker to be present, or else
serve to express the speakers attitude, approving or disapproving, toward the
object in question. On either construal such terms have evaluative force, if any
do. Striking, splendid, excellent, mediocre, miserable, execrable, and
so on, are terms of this sort.
Most aesthetic terms, however, clearly have a substantial descriptive con-
tent; they cannot, whatever evaluative force they may have aside, be applied
to just anything. There are, for instance, formal terms, e.g. balanced, chaot-
ic, unified; expressive ones, e.g. melancholy, anguished, cheerful; meta-
phorical but non-psychological ones, e.g. delicate, steely, brittle; and nat-
ively aesthetic ones, e.g. graceful, gaudy, garish.
It is sometimes proposed that the commonest aesthetic terms, such as those
just mentioned, are of mixed character, having both a descriptive and an

For elaborations of objectivity along these lines, see Richard Miller, Three Versions of
Objectivity: Moral, Aesthetic, and Scientific, and Peter Railton, Aesthetic Value, Moral Value,
and the Ambitions of Naturalism, in Jerrold Levinson (ed.), Aesthetics and Ethics: Essays at the
Intersection (Cambridge: Cambridge University Press, 1998).
See Kendall Walton, Categories of Art, Philosophical Review 79 (1970): 33467; Mark Sagoff,
Historical Authenticity, Erkenntnis 12 (1978): 8393; Philip Pettit, The Possibility of Aesthetic
Realism, in Eva Schaper (ed.), Pleasure, Preference, and Value (Cambridge: Cambridge University
Press, 1983), 1738; Gregory Currie, Supervenience, Essentialism, and Aesthetic Properties,
Philosophical Studies 58 (1990): 24357; and Levinson, Aesthetic Supervenience.
See Frank Sibley, Particularity, Art, and Evaluation, Proceedings of the Aristotelian Society
Supplement 48 (1974): 121. Nick Zangwill, in a useful essay, The Beautiful, the Dainty, and the
Dumpy, British Journal of Aesthetics 35 (1995): 31729, calls them verdictive aesthetic terms.
I am not inclined to put beautiful and ugly in this category without comment, as I
think that such terms, in their primary employment in regard to visual objects, imply particular
kinds of phenomenal impression, ones respectively of pleasing harmoniousness and displeasing
disharmoniousness, rather than simply merit or demerit in the abstract. I recognize, though, that
beautiful is also used as a non-specific term of aesthetic appraisal, with the sense merely of
aesthetically excellent.
Zangwill denominates all such terms substantive aesthetic terms. See The Beautiful, the
Dainty, and the Dumpy.
 Aesthetic Properties, Evaluative Force, and Sensibility 317

evaluative component. It would be a rule of the use of such terms that in

applying them one would be indicating both that a thing had a given property
and that one valued or disvalued that property or ascribed value or disvalue to
it. Sibley called these evaluation-added terms. However, as Sibley noted,
there is little reason to think that these common aesthetic terms, ones that
feature prominently in art criticism, are any more inherently evaluative than
those characterizing human conduct in the context of moral assessment, terms
such as courageous, honest, or merciful. Whether someone is courage-
ous, honest, or merciful would seem to be open to straightforward observation,
though observation of a complex sort, and by an observer versed in the human
form of life. Thus, there is at least reason to doubt that parallel terms of every-
day art criticism, denoting generally desired qualities of appearance, are even
evaluation-added terms. They seem rather to be basically descriptive terms,
naming properties which, as it happens, constitute merits or defects in certain
spheres of assessment. Sibley labeled such terms descriptive merit-terms.
We may grant that many common aesthetic terms, for example, gaudy,
appear to entail evaluations on the part of the speaker. Still, most cannot be
held to do so strictly. First, it seems possible to approve a work for its gaud-
iness, say, or despite its gaudiness. This suggests that the essence of gaudiness
is not a judgment of disapprobation on the speakers part but instead a kind
of appearance: a perceptually manifest effect one can register independently
of any evaluative assessment of or attitudinal reaction to that effect. Second,
the evaluative implications, loosely speaking, of terms like gaudywhich
perhaps derive, in part, from past histories of use in connection with partic-
ular canons of criticism or tastecan be explicitly cancelled or disavowed,
without semantic anomaly. Thus, terms of this sort, despite their air of evalu-
ativity, are such that they can nevertheless be ascribed without strictly entail-
ing anything about the speakers evaluative attitudes. Nick Zangwill makes a
helpful observation along the same lines, proposing that for most substantive
aesthetic attributions it is only a conversational implicature, in Grices sense,
that an evaluation is being made.

Particularity, Art and Evaluation, 6.

In moral philosophy such terms are usually said to generate thick descriptions, ones that
evaluate as well as describe. But see Zangwill, The Beautiful, the Dainty, and the Dumpy, for some
disanalogies between the aesthetic and the moral cases.
Particularity, Art and Evaluation, 6.
See on this Hermeren, The Nature of Aesthetic Qualities, ch. 5.
The Beautiful, the Dainty, and the Dumpy, 322. Zangwill later denies, however, that this
line is sustainable for at least one such attribution, that of gracefulness. He claims that although
318 Aesthetic Properties

But suppose for the sake of argument that there is a purely evaluative ele-
ment in certain common aesthetic terms of the substantive and not solely
evaluative sort, ones such as gaudy, or maudlin. Suppose, in other words,
that it is correct to regard such terms as evaluation-added terms, imply-
ing an evaluative stance toward or estimation of an object on the part of a
user of the term. There would, I claim, still remain a purely descriptive, dis-
tinctively aesthetic content in such an attribution, consisting roughly in an
overall impression afforded, an impression that cannot be simply identified
with the structural properties that underpin it. But what evidence is there of
such higher-order perceptual impressions? What reasons are there to accept
their existence?
First, one can often find alternative descriptions, sometimes requiring
several words, of the distinctive experiential contents involved, in which
the evaluation-added element of the original attribution, if any, has been
removed. For instance, one might approximate the descriptive content of
gaudy by bright, non-harmonious, eye-catching color combinations.
Second, one can often get disputing critics to focus on the common
perceptual ground in their aesthetic responses. For instance, a critic might be
brought to admit that he is aware of the look or appearance another critic has
remarked on with evident relish, reserving his right to dislike it, that is, to
exercise his taste in the sense of personal preference with regard to it.
Here is a musical example. The opening of Bachs Concerto for Three
Harpsichords and Strings in D minor, BWV 1063, offers a vivid expression of
grimness; it might even be described as starkly grim. Now, some competent
listeners like grimnesslike that character, like being confronted with
itand others do not. It is easy to imagine those latter folk simply labeling
the opening depressingly dour and having done with it, but it also seems

it is context that makes merits or demerits of most aesthetic properties, and that most are thus
not even pro tanto merits or demerits, gracefulness appears to count only and always to the good:
Can the grace of something ever be a demerit? It is hard to see how it could (p. 324). But I
am inclined to disagree. Grace would seem to be aesthetically contraindicated in an expressionist
painting or sculpture of the mass executions at Babi Yar. If so, then even gracefulness may not be,
tout court, a pro tanto merit in works of art, and the positive evaluative overtones of its attribution
to a work may be only a matter of conversational implication. For further complications regarding
aesthetic attributions, see Iuliana Corina Vaida, The Quest for Objectivity: Secondary Qualities
and Aesthetic Qualities, Journal of Aesthetics and Art Criticism 56 (1998): 28397.
More precisely, by the look characteristic of bright, non-harmonious, eye-catching color
combinations.
 Aesthetic Properties, Evaluative Force, and Sensibility 319

more than likely that they could be brought to acknowledge the aptness of the
characterization starkly grim as well, only adding under their breaths, if you
like that sort of thing.
Thirdly, unless one assumes there are core aesthetic impressions of a qual-
itative sort, distinguishable from reactions of approval or disapproval per se, it
becomes difficult to explain what competent critics with evaluative differences
of opinion could really be talking about. Surely its not just that one approves
a certain arrangement of lines and colors, or pitches and rhythms, or words
and phrases, and the other not. Rather more likely is that each registers the
overall effect of the arrangement in question, that there are descriptions, reas-
onably neutral ones, they could even agree upon to characterize it, but that
one favors it and the other does not, or one thinks it makes the work good and
the other does not. In addition, failing to acknowledge distinctive aesthetic
impressions as the core descriptive content of common aesthetic attributions
makes a mystery out of what the aesthetic experiences of perceivers of any sort
could possibly consist in.
So suppose, for the sake of argument, that many aesthetic terms, for
example gaudy or maudlin, include an evaluative component irreducibly.
This would hardly show that there are no objective aesthetic properties in the
wings when such terms are being correctly applied. Such properties supply
the purely descriptive content of any evaluation-added terms as there are in
the critics repertoire, as well as figuring as the entire content of others, such as
sorrowful, frenetic, serene, passionate, balanced, delicate. Even were
some of the pairs of aesthetic terms of the type highlighted by Sibley, e.g.
delicate/anemic, bold/gaudy, or heartfelt/maudlin, to differ primarily or
only in the implied attitude or reaction, pro or con, of the ascriberon his
or her taste in the preferential sensethe presence of an aesthetic property to
which they advert in common, and not just complexes of non-aesthetic ones,
seems strongly indicated. The fact that we often lack a single word of neutral
cast to refer to the experiential terrain descriptively shared proves little.

In Being Realistic, the same point was illustrated by the bumptious excitement of the
opening of the fourth movement of Tchaikovskys Fourth Symphony and the distraught longing
of the second theme of the first movement of Francks Piano Quintet, to whose aesthetic qualities
evaluative reactions clearly differ. Regarding the latter piece, we may note this recent testimony from
Roger Scruton: The unctuous narcissism of Cesar Francks Piano Quintet is certainly an expressive
feature, but not a virtue in the work that possesses it. Nevertheless, it is part of the power of this
work, that it so successfully conveys this somewhat disreputable state of mind (The Aesthetics of
Music (Oxford: Oxford University Press, 1997), 148).
320 Aesthetic Properties

As we have already remarked, whatever evaluative force is carried by such

terms as we have just been examining, there are clearly descriptive limits on
their application. Not just any visual pattern can be disapproved of by calling
it, say, gaudy, chaotic, or flamboyant. I suggest that such limits are fixed, in
the last analysis, by distinctive phenomenal impressions or appearances asso-
ciated with such terms, not by sets of structural properties that disjunctively
serve as their occurrence conditions. Thus the existence of certain aesthetic
propertiesroughly, the dispositions to afford such impressions or appear-
ancescan hardly be denied, even if such properties are not always all that
is conveyed by standard terms of aesthetic description, if an evaluative com-
ponent in them be admitted. Attitudes toward the impressions afforded by
artistic structures may evolve over time, even fluctuate back and forth, but
surely something unitary of a broadly perceptual sort often remains constant
throughout.
If an aesthetic predicate P is admitted to have, in some sense, an evaluative
vector, perhaps it follows that there is not then a straightforward or purely
descriptive property being P. Still, it seems there will always be, in the
offing, an associated evaluatively neutralized aesthetic property, being Pn ,
a disposition to afford a certain phenomenal impression to adequately
positioned perceivers. More generally, we can say that there is in regard
to a given object almost always a descriptive aesthetic content such that
ideal judgers who would not apply to the object all and only the same
aesthetic predicatesbecause they have, by assumption, different reactions
or attitudes towards that contentcan still agree on what that content is.
Furthermore, we should not rule out that there might be such aesthetic
content even if no suitably neutral terms were readily available to evince the
agreement on it that there might be.

Though for convenience I here adopt the usual dispositional analysis of perceptual properties,
I believe that strictly speaking such properties are not precisely dispositions, but are rather just
strongly supervenient on such dispositions. (Colin McGinn has recently argued this persuasively,
so far as color properties are concerned, in Another Look at Color, Journal of Philosophy 93
(1996): 53753, while Gregory Currie argues in the same vein, as regards aesthetic properties,
in Supervenience, Essentialism, and Aesthetic Properties.) What, then, are perceptual properties
themselves? In my view color properties are roughly manifest ways of appearing visually, related
to but not identifiable with dispositions to appear visually in such and such ways, while aesthetic
properties are roughly manifest ways of appearing phenomenally, where the ways involved are of a
higher order than those involved in the basic sensory properties. Unlike a dispositional conception
of perceptual properties, a way-of-appearing conception is not in conflict with the intuition that, at
least under favorable conditions, one just looks and sees what color property or aesthetic property
something possesses. (For further elaboration, see What Are Aesthetic Properties?, Proceedings of
the Aristotelian Society. Supplement 79 (2005): 21127 (and Ch. 20, this volume).
 Aesthetic Properties, Evaluative Force, and Sensibility 321

II

So what, then, is there to give pause in this picture of aesthetic attributions

and aesthetic properties? What reasons might there be to reconsider some
aspects of this picture? Here are some of the worries that have been raised
in this connection.
One: it is not really possible to separate the descriptive and evluative com-
ponents in an attribution or property. Two: the descriptive component of
such an attribution or property is not given by overall gestalt impressions
but by structural features of a non-aesthetic sort. Three: there just are no
qualia, no phenomenal properties of any sort, so a fortiori there are no aes-
thetic qualia. Four: even if these other worries are laid to rest it is inescapable
that there are irresolvable differences in aesthetic judgments among even ideal
judges, yet it is unclear that a realist perspective on aesthetic properties can
properly accommodate that fact.
In what follows I give each of these worries a hearing.

III

First worry. Despite what was urged above, many will still be inclined to insist
that most aesthetic terms just do have an evaluative component of one sort or
another, at least in practice or in context. Very well: simply focus on the evalu-
atively neutral phenomenal core of such terms and you will arrive at bona fide
aesthetic properties. But as Hamlet famously complained, theres the rub. Per-
haps it is not really possible to identify the purely evaluative component of an
aesthetic term so as to allow its subtraction from the import of the term as a
whole, leaving a purely descriptive component whose cash value is an aesthetic
property; perhaps what is descriptively conveyed by an aesthetic term cannot
be isolated, cannot even legitimately be presumed to exist, apart from what
the term conveys in its concrete use. If that is so, then even modest aesthetic

One philosopher who is clearly skeptical of the conceptual surgery required is Alan Goldman:
A different question is whether we can always analyze evaluative properties into evaluative and
non-evaluative components. Since we have viewed these properties as relations between objective
properties and evaluative responses, it might seem that the answer must be affirmative. But I have
also pointed out that many of the higher-level properties of this sort are unspecific on their objective
sides. Although it should be possible in principle to analyze specific references to such properties
into objective and subjective components, we cannot do so for the properties themselves (Aesthetic
Value (Boulder, CO: Westview Press, 1995), 26).
322 Aesthetic Properties

realism, affirming the existence of evaluatively neutral aesthetic properties at

the root of aesthetic attributions, whether evaluatively charged or not, would
be too optimistic.
What to say? Well, one way it might turn out that the phenomenal impres-
sions associated with aesthetic terms might not be exhibitable denuded of
all evaluative aspect would be if such impressions were themselves inherently
pleasant or unpleasant. The impressions that go with finding something gra-
ceful or harmonious, say, are plausibly of that sort. To the extent that phe-
nomenal impressions are inherently pleasant or unpleasant, they would seem
of necessity to bring in their train corresponding reactions of favor or disfavor.
Thus, to that extent, isolating neutral phenomenal impressions at the core of
aesthetic attributions would be a chimerical pursuit.
But two points bear making in response. First, it is surely not the case that
all aesthetic impressions are inherently hedonically valenced. It rather seems
that most such impressions are, for most perceivers, more or less hedonically
neutral for instance, those of the peacefulness of a landscape, the angularity
of a design, or the urbanity of a passage of violin music. Second, even were all
aesthetic impressions hedonically valenced, it would still be possible rationally
either to approve or disapprove the affording of such an impression at a partic-
ular point in a particular work and to report its occurrence by an appropriate
evaluatively charged term. That is to say, the inherent pleasurability of the
impression at the core of an aesthetic property would still be distinguishable
from the evaluative component of that property, so to speak, when picked out
by an evaluatively charged substantive term.
There are other things to say in regard to the problem of the distinctness
or isolatability of aesthetic impressions, but they will be evident in my dis-
cussion of the third worry noted, that concerning the existence of qualitative
phenomena generally.

IV

Second worry. What exactly does the descriptive component of an aesthetic

property consist in? Is it in fact a unitary impressiona look or appearance
that an object is fitted to afford, as I have argued, or is it rather a plurality of
combinations of non-aesthetic features that an object might possess? Call the
former the phenomenological account of aesthetic properties and the latter the
 Aesthetic Properties, Evaluative Force, and Sensibility 323

structural account. Are there additional reasons that can be given in favor of
the phenomenological account of aesthetic properties as opposed to the
structural one? I am not sure that I have any more to offer than I have already
advanced, here and elsewhere, or that can be found in the writings of Sib-
ley, Beardsley, and Walton. What argues above all for the phenomenologic-
al conception is its comporting better with the evident semantics of aesthetic
attributions: when we ascribe an aesthetic property it seems that what we are
ascribing, at base, is an emergent way of appearing, and not a range of ensembles
of disparate traits that, it so happens, sustain such a way of appearing.
In any event, the phenomenological account presumably gains in convin-
cingness the more light we can throw on exactly what an overall phenomenal
impressionthe suggested core of an aesthetic attributionmight involve.
Though I have until now stressed the perceptual character of such impres-
sions, we need not deny that such impressions might be partly affective as
well. In other words, some aesthetic impressions might be bound up with feel-
ings consequent on apprehending an objects non-aesthetic features. Register-
ing such an impression from an object might, as Derek Matravers has recently
suggested, involve apprehending the objects sensible qualities and forms with
a certain feeling, or having a certain feeling in apprehending them. Further-
more, such impressions, though basically perceptual, might be importantly
mediated or inflected by conceptual activity at some level; for example, the
impression of a pattern as unified might partly consist in a tendency to con-
strue the parts of the pattern under a concept such as that of <fitting or work-
ing together>. Clearly, more work needs to be done on this matteron

One might alternatively label the latter the reductive account.

See my Aesthetic Supervenience.
This is marked by calling them, as is often done, gestalts.
Aesthetic Concepts and Aesthetic Experiences, British Journal of Aesthetics 36 (1996):
26577. Matravers proposes that some aesthetic property terms, such as delicate, have both a
non-aesthetic, condition-governed use and a properly aesthetic, experience-tied use. With the latter,
there is no getting around a phenomenological account of the property being attributed: Consider
the vase again; experiencing it causes an observer to believe it is fluted, blue, made of very thin
glass, about six inches tall, and so on. By itself this may be enough to ground the non-aesthetic
judgement that it is delicate. If, in addition, the observer gets a feeling (perhaps akin to loss, or
anxiety about the objects fragility) this may prompt an aesthetic judgement [that it is delicate]. It
would be senseless to use a different term; what needs to be communicated is the response to just
those properties of the vase that are to do with its being (non-aesthetically) delicate (p. 273).
For suggestive proposals on this score, see Monique Roelofs, Aesthetic Experiences and
Their Place in the Mind (Ph.D. dissertation, University of Maryland, 1997). See also Berys Gaut,
Metaphor and the Understanding of Art, Proceedings of the Aristotelian Society (1997): 22441,
324 Aesthetic Properties

what exactly may be comprised in the phenomenological impression or exper-

iential complex at the core of a given aesthetic property.
We should also not overlook the role that paradigm objects or patterns
might play in pinning down the descriptive content of aesthetic attributions
on a phenomenological view of them. Perhaps, with such a paradigm at
hand, one could indicate the target overall impression or feeling as the
impression or feeling one gets from this, ostending the object or pattern in
question. Of course, there is always the problem of whether people will go
on, or generalize, in the same waya problem which Wittgenstein brought
to philosophical saliencebut there is no reason to think it inevitable that
they will not. For one, corresponding to the innate quality-spaces that Quine
once posited for the sensory modalitiesin order to account, for example, for
why we see the visible spectrum as falling into basically six colors, rather than,
say, either three or thirty-threethere may even be innate aesthetic quality-
spaces, which would help to ensure that many of our more basic aesthetic
impressions were at some level mutually conformable, registered at roughly
the same levels of specificity.

Third worry. A phenomenological account of aesthetic properties naturally

faces general skepticism about the existence of phenomenology. There are a
number of philosophers who deny the existence of qualia across the board,
that is to say, intrinsic, qualitative properties of conscious experience, more
familiarly known as the ways things seem or appear to us. The implications
for the aesthetic sphere are plain. If there are no qualia of low order, such as
sensory qualia, then there are presumably no qualia of higher order, such as
aesthetic qualia. If there are no taste, smell, color, touch, or timbre qualia,
there are certainly none, we may be sure, for grace, delicacy, melancholy,
garishness, effeteness, or balance. And thus a phenomenological account of
aesthetic properties would be precluded from the outset.
Probably the best-known basher of qualia is Daniel Dennett.

for insightful remarks on the doxastic versus imaginative involvement of concepts in aesthetic
experience.
For illuminating discussion of the role of such paradigms in aesthetic discourse, see Vaida,
The Quest for Objectivity. However, the relativism that Vaida claims follows from variability in
paradigm preference among aesthetic judges is not as unavoidable as she makes it out to be.
 Aesthetic Properties, Evaluative Force, and Sensibility 325

I want to take a brief look at his best-known bash, which occurs in a paper
called Quining Qualia.
Dennett tries to show that people can be mistaken about their own qualia
in such manner as to suggest that we arent really entitled to think there
are qualia at all to be right or wrong about. One of his central thought-
experiments in support of this is the story of Chase and Sanborn, coffee
tasters at Maxwell House. Chase and Sanborn both used to like the house
coffee, and now neither does. Chase claims it tastes the same to him, though
he now grades it lowhe has apparently become a more sophisticated
drinkerwhile Sanborn claims it tastes different to him, though were it
still to taste the same to him he would grade it as high as everhis taste
buds have apparently deteriorated. The claims of both claimants depend in
part on the reliability of their memories. Now Dennett asks whether we
can take their claims at face value. Might one or both of them simply be
wrong? Might their predicaments be importantly the same and their apparent
disagreement more a difference in manner of expression than in experiential
or psychological state? That is, might it not just be a matter of how
they interpret their conditions and histories, rather than a difference in the
conditions and histories themselves?
Focusing on Chase, Dennett sketches two possibilities. One, his coffee-
qualia have remained the same, while his reactions toward them have altered.
Two, his coffee-qualia have altered gradually over the years, so that his
memory of the original quale has become unreliable, while his reaction to the
original quale, could he recover it, would be the same. Testing might not be
able to resolve which was the case, Dennett maintains, because the alternative
hypotheses might not differ in any observable consequences. This is especially
so if we consider hypotheses midway between these two, with admixtures
of both.
Dennetts argument concerning this and similar examples is thus that the
postulation of qualia is without justification because there may be no way
to distinguish empirically, in a case of qualia ostensibly altered over time,

Daniel Dennett, Quining Qualia (1988), reprinted in William Lycan (ed.), Mind and
Cognition (Oxford: Blackwell, 1990); all page references are to that reprinting. For other criticisms
of Dennetts case for quining qualia, ones largely orthogonal to mine, see Georges Rey, Dennetts
Unrealistic Psychology, Philosophical Topics 22 (1994): 25990; Eric Lormand, Qualia! (Now
Showing At A Theater Near You), Philosophical Topics 22 (1994): 12756; and Bredo Johnsen,
Dennett on Qualia and Consciousness: A Critique, Canadian Journal of Philosophy 27 (1997):
4781.
Quining Qualia, 527.
326 Aesthetic Properties

between actually altered qualia and alteration in the memory of qualia pre-
viously had. Since there may be no way in which the difference between the
hypotheses could ever be empirically manifested, even for the subject, the
hypotheses cannot really differ in content, and thus the notion of qualia in
terms of which they are framed must be bogus. I have two comments on this
argument.
My first comment is that the argument seems at base a verificationist one,
albeit of a sophisticated kind. But if one is not convinced generally by veri-
ficationist arguments aimed at the collapse of other distinctions which seem
to be real and graspablefor example, the distinction between truth and
ideally warranted assertabilitythen one will not be overly moved by the dif-
ficulty to which Dennetts thought experiment points. We might just have
to accept that our qualia might change in such a fashion that we could not
always be certain that they had done so, if the circumstances in which the
changes occurred were not favorable to such determination. But this would
show neither that the idea of such a change was incoherent, nor that, were we
to undergo such changes, we would never be in a good position to think that
we had.
My second comment is that the whole issue of memory for sensory qualia is
a more problematic one than Dennett realizes, as Diana Raffman has usefully
emphasized in recent work. Psychological research concerning categorical
perception has revealed the existence of definite limits to the fine-grainedness
of conceptual representations of sensory data. It appears we can detect or dis-
criminate sensory impressionse.g. shades of color, nuances of toneof a
sort much finer than we are able to categorize and label internally. As a res-
ult, accurate memory of such impressions is likely impossible, as is any further
conceptual processing of them. But of course that does not impugn the fact
that we have received those impressions, impressions of a surprisingly high
degree of specificity.
To make the case concrete, imagine that you are noting a difference in
comparing two color patches only a discriminable shade or two apart. Then

Language, Music, and Mind (Cambridge, MA: MIT Press, 1993), and On the Persistence
of Phenomenology, in T. Metzinger (eds.), Conscious Experience (Paderborn: Schoningh, 1995):
293308.
This effectively takes the wind out of the sails of Dennetts closing moral regarding qualia: it
would be a mistake to transform the fact that inevitably there is a limit to our capacity to describe
things we experience into the supposition that there are absolutely indescribable properties in our
experience (Quining Qualia, 544). We have no need of any illicit transformation in order to arrive
at that supposition; it just seems to be empirically true.
 Aesthetic Properties, Evaluative Force, and Sensibility 327

imagine that your qualia registrations alter in the direction of lesser refine-
ment or greater spread, so that the difference between the patches disappears,
with both now appearing more or less the same. By hypothesis, this cannot
be equally well interpreted as a case of qualia persistence with alteration in
memory of qualia, since there simply are no memories of qualia at the des-
ignated level of specificity. Thus, rather than the supposition of qualia being
bogus it would appear to be Dennetts thought-experiments that are bogus,
insofar as they posit empirical equivalence between hypotheses of qualia shifts
and hypotheses of qualia persistence with shifts in memory-of-qualia, giv-
en that, at least for highly specific sensory qualia, we simply have no such
memories at all.
The other main strand of argument against qualia in Dennetts paper is
that the doctrine that sharply distinguishes qualia from reactions to qualia is
simply not sustainable. Dennett holds the root error in qualia thinking to be
the conviction that
what counts as the way the juice tastes to x can be distinguished . . . from what is a mere
accompaniment [of that taste] . . . One dimly imagines taking such cases and strip-
ping them down gradually to the essentials, leaving their common residuum, the way
things look, sound, feel, taste, smell to various individuals at various times, independ-
ently of how those individuals are nonperceptually affected and independently of
how they are subsequently disposed to behave or believe . . . the fundamental mistake
is supposing there is such a residual property to take seriously . . .
The man who likes the taste of beer, the man who dislikes it, and the man
who is indifferent to it, suggests Dennett, do not taste the same taste; and a
man who dislikes a smell that he used to like now plausibly smells a differ-
ent smell. The idea is that the hedonic response to a taste or smell is arguably
part of what that taste or smell is for the individual in question, so that the
notion of a core, evaluatively neutral, smell or taste, common to the exper-
ience of subjects with opposed hedonic reactions to it, is simply a fiction.
Dennetts thought-experiment involving the taste of cauliflower is perhaps his
most vivid presentation of the matter:
Imagine now the cauliflower cure: someone offers me a pill to cure my loathing for
cauliflower. He promises that after I swallow this pill cauliflower will taste exactly
the same to me as it always has, but I will like that taste! Hang on, I might reply.
I think you may have just contradicted yourself. But in any event, I take the pill
and it works. I become an instant cauliflower-appreciator, but if I am asked which of

Quining Qualia, 521.

328 Aesthetic Properties

two possible effects the pill has had on me [i.e. the taste quale has stayed the same
but the reaction to it has changed vs. the taste quale has changed] I will be puzzled,
and will find nothing in my experience to shed light on the question. Of course I
recognize that the taste is (sort of) the samethe pill hasnt made cauliflower taste
like chocolate cake, after allbut at the same time my experience is so different now
that I resist saying that cauliflower tastes the way it used to taste.

It is significant, I think, that the more compelling of Dennetts thought-

experiments suggesting that qualia and reactions thereto cannot ultimately
be distinguished are ones featuring gustatory qualia, i.e. tastes in the literal
sense. But they seem much less compelling when colors, say, are substituted
for tastes. Suppose someone has a clear hedonic reaction to certain colors, so
that he is, say, tickled pink by pink and made really blue by blue. And then
suppose he is suddenly freed of his chromohedonia, courtesy of a little pill,
so that he now registers pink and blue with no noticeable affect in either dir-
ection. We can imagine him now greeting them inwardly with a mere shrug.
Is there any plausibility in the suggestion that the phenomenal appearance of
those colors will have changed for him? Is there any reason not to credit the
claim he would presumably make, that the colors look the same to him, but
now affect him differently as far as his mood is concerned? Consider a timbre
one is attending to aurally, first finding it neither appealing nor unappealing,
and then gradually coming to regard it as quite attractive. Is it plausible to
doubt that there is a constant sensory impression, a quale, in ones experience
as ones attitude toward it undergoes a shift?
I suspect that visual and auditory qualia give different results in these
thought experimentsones more supportive of the separability of quale and
reaction theretothan do gustatory and olfactory ones, for several reasons.
With colors and timbres as opposed to tastes and smells, the factor of physical
satiety generally does not come into play, and the hedonic effects involved
are generally not as pronounced. There is also an obvious sense in which
we have more distance on colors or timbres than on smells and tastes,
which perhaps enables us to distinguish them more clearly from reactions that
accompany them. If we mentally compensate for these differences when we
review Dennetts tales of ups and down in the liking for tastes and smells I
think we will be less inclined to concede that those ups and downs make for
change in those tastes and smells themselves.

Ibid. 535.
 Aesthetic Properties, Evaluative Force, and Sensibility 329

VI

Fourth worry. Well and good, one might respond to what I have said so far.
The real problem about aesthetic attribution, one may insist, still remains.
It concerns the fact of continuing and irresolvable differences in aesthetic
attributions among even ideal critics, that is, the notable absence of conver-
gence in judgments attributing aesthetic properties to works of art among
even optimally prepared and positioned perceivers. For it is that fact, above
all, that threatens aesthetic realism as regards attributions to works of art.
The thought is basically this. Whatever the truth about the descriptive
content of substantive aesthetic terms, that is, whether such content is given
by qualitative impressions of distinct character or constellations of low-level
perceptual features, and especially if such terms have a significant evaluative
dimension, whether that is a matter of sense or of conversational implication
in relevant contexts, we just cannot reasonably expect to find convergence in
the aesthetic judgments of even ideal critics. And that is because of the evident
diversity of sensibilities among even ideal critics, as even Hume recognized
250 years ago. Ideal critics, like ordinary folk, sort themselves out into
sensibility-types or sensibility-groups, and possibly very many of them.
If so, then realism about real-world, or thick, aesthetic properties, that
is, properties denoted by substantive aesthetic terms of criticism with their
evaluative aspects intact, would appear to be a vanishing prospect. That there
might always be more stripped-down, thin, aesthetic properties in the off-
ing that warring critics might be brought to acknowledge would seem to offer
only small consolation. For there would be no truth about whether or not a
given abstract sculpture was, say, graceful, because some ideal critics, of sens-
ibilities A, B, and C, find it so, in virtue of reacting positively to the core
impression presented by something that was a candidate for being described
as graceful, while other ideal critics, of sensibilities D, E, and F, do not, in
virtue of reacting negatively to that same core impression.

This is the basic thrust of both Goldman, Aesthetic Value, ch. 1, and John Bender, Realism,
Supervenience, and Irresolvable Aesthetic Disputes, Journal of Aesthetics and Art Criticism 54
(1996): 37181.
Goldman expresses this worry with special force: The prior issue is whether ideal critics
as specified above in terms of their characteristics . . . will agree in their aesthetic judgments . . . if
ideal critics disagree in their aesthetic judgments, then the [realist] account will have us ascribing
incompatible properties to the same artworks. Works will be both graceful and not graceful,
330 Aesthetic Properties

Furthermore, supervenience, a pillar of property realism, would also be

imperiled, since aesthetic properties would no longer supervene on a works
intrinsic and relational propertiesits structural features plus its artistic
contextbecause differing sensibilities among ideal critics would prevent
aesthetic properties from emerging on that basis alone. True, one might
reply that supervenience of aesthetic properties would still hold, though in
a narrow, sensibility-relativized form. Instead of gracefulness simpliciter we
would need to speak of, say, gracefulnessq , which would amount roughly to
being disposed to appear graceful to ideal critics of a given sensibility class,
comprising sensibilities A, B, and C. But there would be a hollow ring to
settling for so qualified a supervenience of properties.
At any rate, in order to begin to come to terms with the implications of
sensibility diversity for aesthetic realism we need to look at what sensibilit-
ies in this context might consist in. We should at the outset recognize the
possibility not only of a diversity of sensibilities, but of a diversity of kinds of
sensibility. There may, I suspect, very well be two basic kinds of sensibility

powerful and not powerful. The issue of agreement among ideal critics is crucial, since the [realist]
account becomes incoherent without such consensus (Aesthetic Value, 289).
See Bender, Realism.
Anti-realists about aesthetic properties will of course propose a different accommodation with
the facts of critical divergence, eschewing talk of supervenience entirely: We can now see how
nonrealists will modify the relational account of aesthetic properties . . . to avoid the ascription of
incompatible properties that plagues the realist. According to the modified account, when I say that
an object has a certain aesthetic property, I am saying that ideal critics who generally share my taste
will react in a certain way to its more basic properties (Goldman, Aesthetic Value, 37.)
Bender, in the course of arguing against the position on aesthetic supervenience adopted by
Goldman in earlier articles, draws a somewhat different conclusion: Goldman seems to think that
contrary aesthetic properties of the sort that generate irresolvable disputes, properties with opposed
evaluative implications, can each be supervenient upon the nonevaluative base properties of a
work . . . I propose to argue that Goldman is mistaken . . . [for Goldman] the truth of our aesthetic
judgments is to be construed not as a matter of whether all ideal critics would respond as we have,
but of whether ideal critics whose tastes we share would so respond . . . [This] produces a relativized
supervenience: relative to a specific standard of taste or set of values there can be no change in a rational
critics evaluation of a work without some change in its nonevaluative properties . . . But, I contend,
a relativized supervenience of this sort is no supervenience at all. Supervenience is a metaphysical
dependency relation asserting that changes in supervenient properties arise only with changes in
relevant base properties. The constraint we are left with after relativization is nothing more than the
rather trivial, and epistemic, constraint of consistency upon rational judgments (Realism, 3723).
It seems to me that the relativization that ultimate nonconvergence among ideal critics induces
need not take the form that Bender envisages. For instead of relativity to standards which critics
accept we can see it as relativity to sensibilities which partly constitute those critics. The supervenience
base would need to expand, of course, to include a particular critical sensibility, and the aesthetic
property defined by reference to it would then be a correspondingly much narrower one, implicitly
indexed to ideal critics of that stripe. But we would then still have something recognizable as
property supervenience, and not just a rational consistency constraint on judgments of any sort.
 Aesthetic Properties, Evaluative Force, and Sensibility 331

at play here, which we can label perceptual sensibility and attitudinal

sensibility. What I have in mind would parallel the distinction between phe-
nomenal impressions and evaluative reactions in my earlier analysis of evaluat-
ively charged aesthetic properties. A perceptual sensibility would be a
disposition to receive phenomenal impressions of certain sorts from various
constellations of perceivable non-aesthetic features, while an attitudinal sens-
ibility would be a disposition to react to phenomenal impressions of certain
sorts with attitudes of favor or disfavor. There is no need to assume that an
attitudinal sensibility is necessarily a fixed or inborn matter; it might indeed
generally have a strong culturally formed component. Furthermore, we would
expect that a persons attitudinal sensibility, in regard to given phenomen-
al impressions, would clearly be modifiable over time, especially insofar as
such sensibility was culturally, rather than physiologically, ordained or con-
ditioned.
Suppose then, for simplicity, that there were three sensibilities of each type.
That gives nine sensibility types with both dimensions taken into account:
these would be combined, or perceptual-attitudinal, sensibilities. So in prac-
tice, assuming ideality in all other respects, there could still very well be nine
distinct profiles of aesthetic response, corresponding to nine combined sens-
ibility types, with resulting disagreements in all individual cases. The question
naturally arises of telling to what aspect of divergence in sensibility we should
attribute any particular disagreement. How could we determine, in a particu-
lar case, whether disagreement was rooted in differences in perceptual sensib-
ility, or differences in attitudinal sensibility, or both? What might show, more
generally, that judges differed in perceptual sensibility instead of attitudinal
sensibility or vice versa?
However, though we should certainly be open to the idea of possible
diversity in perceptual sensibilities, that is, propensities to receive aesthetic
impressions from the same perceptual configurations, it is by no means as
clear that there are such among ideal critics as that there is a diversity in
attitudinal sensibilities, that is, propensities to like or dislike given aesthetic
impressions. So if we can separate descriptive matters from evaluative ones

We may note that the thesis of the ineliminable evaluativity of aesthetic attributions and the
thesis of the non-convergence in aesthetic attributions among ideal judges because of irreducible
differences in sensibility, though related, only come to the same thing if all differences in sensibility are
a matter of differing attitudinal sensibilitiesinvolving different reactions to the same phenomenal
impressionsrather than differing perceptual sensibilities, that is, differing dispositions to receive
such impressions from the same low-level perceptual features.
332 Aesthetic Properties

in this arena, as I have been suggesting that we in principle can, there

may be more hope for ultimate convergence in at least descriptive aesthetic
attributions among ideal critics than we might think, even if diversity of
perceptual sensibilities is, of course, theoretically possible.
But what if that theoretical possibility is realized after all? What if it turns
out that there are roughly as many distinct perceptual sensibilities as there are
blood types? Well, presumably one would then be well advised to find ones
perceptual as well as attitudinal sensibility group, and stay tuned. But that
would not alter the fact that suitably relativized aesthetic properties, at any
rate, would still be there for the having and experiencing.
There are two questions it is natural to pose at this point. First, why does
it matter whether there are groups of perceivers among whom one is likely
to find convergence in aesthetic judgments, given that differences of sens-
ibility clearly do induce divergences, even under ideal conditions and in the
last analysis, among aesthetic perceivers? Why should one trouble to identify
ones sensibility-group, of any sort? An obvious answer, though perhaps there
are others, is that aesthetic recommendations from critics belonging to ones
own sensibility-group will be of greater practical worth, having more predict-
ive value as regards ones own aesthetic satisfactions than recommendations
from other critics.
Second, is belonging to one such sensibility-group better than belonging
to another, and if so, how? A positive answer to this query might emerge as
a consequence of addressing the crucial though largely overlooked problem
raised by Humes celebrated essay on taste, namely, that of explaining why
judgments of ideal critics should rationally interest all perceivers, even those
far from ideality, and why such perceivers have reason to strive for greater
ideality in their own aesthetic dispositions insofar as those dispositions are
subject to change. But this is not the place to develop that answer.

VII

Before concluding this essay I turn briefly to a critique which John Bender has
offered of my earlier defense of aesthetic realism, one that gives the concerns
highlighted under this last rubric a vigorous expression. After providing a

I attempt to do so in Humes Standard of Taste: The Real Problem, Journal of Aesthetics

and Art Criticism 60 (2002): 22738 (and Ch. 21, this volume). An answer along similar lines can
be found in Goldman, Aesthetic Value.
 Aesthetic Properties, Evaluative Force, and Sensibility 333

summary of my attack on one sort of aesthetic antirealism, the main points of

which were recounted in the opening section of this essay, Bender notes that
I try to block such antirealism, which sees aesthetic properties as simply eval-
uative reactions to non-aesthetic features, by offering an alternative relational
analysis of such properties in terms, roughly, of dispositions or propensities
of an artwork to produce in appropriately backgrounded perceivers certain
holistic phenomenal impressions. However, Bender is not happy with that
analysis, and finds wanting the defense of aesthetic realism it underpins. I here
focus on what seem to be his two main criticisms.
First, evaluatively neutral aesthetic attributions, Bender claims, are as dis-
putable, and as relative to critical taste, as evaluatively charged aesthetic attri-
butions:
But what if the application of purely descriptive aesthetic terms exhibits the same
kind of relativity to the tastes or standards of particular judges [as does applica-
tion of manifestly evaluative aesthetic terms]? Will there not be irresolvable disputes
among critics over the most appropriate way to describe or interpret a works aes-
thetic content? . . . There can be as much critical indeterminacy concerning wheth-
er a musical passage is sad or resigned as there is in judging it disunified or only
uninhibited . . . I suggest that if aesthetic properties are dispositions to afford phe-
nomenal impressions or looks, as Levinson claims, these impressions will be variable
to a not insignificant degree and will reflect the judges sensibilities just as surely as his
or her evaluative reactions do.

My response is as follows. Of course there might be as much divergence

among ideal critics regarding neutral aesthetic content as there is among them
in regard to their evaluative reactions to such content. That is, diversity of
what I above called perceptual sensibilities might have as much to do with
non-convergence in aesthetic judgments among ideal critics as diversity of
attitudinal sensibilities. But it seems there is, at least so far as anyone has
shown, little reason to think this is the case.
To take Benders illustrative example, it is easy to imagine that ideal crit-
ics disagreeing over the application to a stretch of music of charged terms
like disunified or uninhibited will never be brought to agree on one rather
than the other, because of their differing hedonic reactions to or evaluat-
ive attitudes toward the music in question. But it is less easy to imagine,
given our experience with these matters, that more sustained attention to

Realism, 374.
334 Aesthetic Properties

the music, more sensitive appreciation of its style, more heightened aware-
ness of its historical antecedentsand perhaps also further elucidation of the
concept of musical expressivenesswould not serve to bring such critics to
agree that the music was better described as resigned than sad or vice versa.
Second, there are insufficient grounds, Bender claims, to believe there is a
distinctive phenomenal impression associated with every substantive aesthetic
term, isolatable from whatever evaluative force the term possesses, and shared
or shareable by ideal critics who yet differ in their evaluative reactions to it:
Levinson has given too little description of these impressions for us to eliminate
the alternative: that they vary with taste. After all, is it clear that when I describe a
painting as gaudy and you describe it as merely intensely chromatic that the painting
is affording both of us the same phenomenal impression? . . . I find a wine searingly
acidic while you find it refreshing and zingy. Do we nevertheless share some singular
phenomenal impression of the wines acidity?

My response to this is implicit in my discussion of Dennetts attack on

qualia. Bender is seeking to motivate the same skepticism as that to which
Dennett gave full rein, skepticism as to whether qualitative registrations can
be distinguished from hedonic reactions to, or evaluative attitudes toward,
such registrations. But I have already indicated how such skepticism can and
should be resisted. I would add only that the fact that we are sometimes unable
to make such introspective distinctions with confidence is hardly reason to
think that we are in principle, or even usually, unable to make them.
A last remark, as regards taste in the literal sense, the focus of Benders
second example: people just do seem to differ significantly in their baselines
in such matters, more so than with other sense modalities. What that means is
that peoples gustatory responses might first need to be roughly calibrated in
some way if their experiences are to be meaningfully compared. For example,
some persons may only decisively register sweetness when there are two
spoons of sugar in their tea, while others require just a quarter of a spoon.
Differences of that sortthat is, differences in sensitivity to non-aesthetic
perceptual features, such as sweetness or aciditymay be as likely to account
for one taster finding wine searingly acidic while another finds it refreshingly

See my Musical Expressiveness, in Pleasures of Aesthetics (Ithaca: Cornell University Press,

1996), 90125, and Musical Expressiveness as Hearability-as-Expression (Ch. 6, this volume).
Realism, 375.
This is leaving aside, of course, the more extreme matter of variation in registering the
bitterness of substances such as phenol-thio-urea, which to a quarter of humanity does not taste
bitter at all.
 Aesthetic Properties, Evaluative Force, and Sensibility 335

acidic as either differences in the overall phenomenal impression afforded by a

wine or differences in the preferences of tasters for degree of perceived acidity.

VIII

Clearly, the objectivity for aesthetic properties defended in this essay is not
one that accords them a transcendent status, independent of human reactions.
What has been defended is rather objectivity as contingent but stable inter-
subjective convergence in judgments among qualified perceivers.
Recent briefs for anti-realism about aesthetic properties give little reason
to think there are irresolvable differences of aesthetic characterizationeven
real-world or thick aesthetic characterizationsamong qualified perceivers
for a given artform in all, or even most, cases. The evidence is rather that this is
so only in some cases, ones which on that account call attention to themselves,
standing out as they do from the boring norm of widespread, unheralded
agreement in such matters among those with adequate experience.
Would any competent art lover or critic demur from aesthetic judgments
such as that the opening of Beethovens Ninth Symphony is dark and fore-
boding? That Austens Emma is witty and clever? That Rembrandts The Jew-
ish Bride is heartfelt and tender? That Mondrians Broadway Boogie-Woogie
is vibrant and exuberant? That Brancusis Bird in Space is sleek and elegant?
That Schlemmers Bauhaus Staircase is cool and airy? That Hitchcocks Ver-
tigo is tragic and disturbing? That Kirchners Street, Dresden is lurid and ali-
enated? That Duchamps In Advance of a Broken Arm is whimsical and arch?
That Vareses Ionisation is raucous and irreverent? That Hellers Catch-22 is
anarchic and hilarious? That Plaths Daddy is bitter and ironic?
Even if irresolvable disagreements among appreciatively ideal observers,
stemming from differences of attitudinal or perceptual sensibility, persist
in a fair number of cases, precluding realist interpretation of aesthetic
attributions, nothing precludes realist interpretation of aesthetic attributions,
or interpretation of them as assertibly true or false, in the majority of cases. An
aesthetic realist, it seems, can rest reasonably content with that.

Note that even where there is striking ambiguity as regards a works character, as for example
with De Koonings Women paintings, whose perspective on their subject seems precariously poised
between savage animosity and joyous affirmation, such ambiguity is itself an objective part of that
character.
 20
What Are Aesthetic Properties?

I SOME WORRIES ABOUT AESTHETIC PROPERTIES

Derek Matravers, in a recent essay, has attempted to show that aesthetic real-
ism is a problematic thesis. Matravers poses some pointed worries for both
the claim that there are aesthetic properties at all, and the claim that, if there
are such, they are to be conceived as higher-order perceptual ways of appear-
ing. Call the first of these the existence claim about aesthetic properties, and
the second of these the nature claim about aesthetic properties.
Matraverss worries regarding the existence claim are basically two. The first
worry is that there seem insufficient grounds to postulate aesthetic proper-
ties, above and beyond what is admitted by almost all parties to the debate,
namely, justifiable aesthetic descriptions, on the one hand, and manifest aes-
thetic responses, on the other. For non-aesthetic perceptible properties, it is
held, which both cause manifest aesthetic responses and can be appealed to
in support of justifiable aesthetic descriptions, are all we need to explain sat-
isfactorily the phenomena of aesthetic life. The second worry is that the pos-
tulation of aesthetic properties corresponding to aesthetic predicates seems to
have two undesirable consequences, one being the ascription of ambiguity to
certain psychological predicates having both ordinary and aesthetic employ-
ment, and the other being the undermining of the autonomy of aesthetic
judgments.

First published in Proceedings of the Aristotelian Society. Supplement 78 (2005): 21127.

Matravers, Aesthetic Properties, Proceedings of the Aristotelian. Society. Supplement 79 (2005):
191210.
See my Aesthetic Properties, Evaluative Force, and Differences of Sensibility, in Emily Brady
and Jerrold Levinson (eds.), Aesthetic Concepts: Essays after Sibley (Oxford: Oxford University Press,
2001), 6180 (and Ch. 19, this volume).
Matravers also raises the issue of whether it is possible to characterize in a non-circular manner
an ideal or norm-grounding perceiver on a response-dependent account of aesthetic properties,
adapting Roger Scrutons objection in this vein to my analysis of the expressive properties of
 What Are Aesthetic Properties? 337

I begin with the second worry, as it seems to me less serious than the first.
The two undesirable consequences alleged to follow from the supposition that
there are aesthetic properties are not that at all, in the first instance because the
consequence does not in fact follow, in the second instance because, suppos-
ing it does, the consequence is not in fact undesirable.
The first instance of a putative undesirable consequence turns on an argu-
ment of Roger Scrutons concerning emotional predicates like sad. It is
claimed that sad cannot refer to a different property when applied to music
than when applied to persons, because that would mean that the term sad
was ambiguous, that is, possessed of two distinct and unrelated meanings, and
thus that one could know how to apply sad to music without knowing how
to apply sad to persons, which cannot be correct. But the response to this is
that Scrutons argument sets up a false dilemma, between there being only one
property involved, to wit, sadness, in the two cases, and there being two, but
unrelated, properties involved. The possibility of there being two, but related,
properties involved is simply overlooked. Yet that possibility is consistent with
the expressive use of sad in connection with music presupposing the ordin-
ary use of sad in connection with persons, as Scruton insists it must. The
second, aesthetic propertymusical sadnessneed only be something like a
disposition of the music to be heard as if it is the ordinary expression of sadness
by a person. If musical sadness is indeed something like that, it is perfectly
clear both why sad is not aptly described as ambiguous between its ordinary
and its musical uses, and why one would not be able to apply sad in musical
contexts without being able to apply it in ordinary, or extramusical, ones.
The second instance of a putative undesirable consequence should not
trouble a proponent of aesthetic properties, not because the consequence fails
to follow, but because it fails to be undesirable. The doctrine of aesthetic
autonomy holds that one cannot know that an aesthetic judgment is true of
an object, artwork or otherwise, without direct experience of that object. But
the grounds offered for this doctrine, which has its roots in Kant, and which
has been championed over the years by Arnold Isenberg, Alan Tormey, Mary
Mothersill, Philip Pettit, and others, have never been sufficient to establish it.

music. I respond to Scrutons objection in Musical Expressiveness as Hearability-as-Expression,

in M. Kieran (ed.), Contemporary Debates in Aesthetics and Philosophy of Art (Oxford: Blackwell,
2005); Ch. 6, this volume.
See my Musical Expressiveness as Hearability-as-Expression, though I would now modify
slightly my formulations on the nature of expressive properties so as to distance my view more
clearly from one that holds such properties to be dispositions.
338 Aesthetic Properties

This is not the place for an exhaustive refutation of that doctrine, which has
become something of a shibboleth in aesthetic theory, but a brief diagnosis of
its fundamental error, I think, would be this. There is a sense of judgment in
which one cannot be said to be judging something unless one is having, then
and there, perceptual experience of it. (Some paradigm cases: I judge this sun-
set to be the most remarkable we have encountered during our vacation and I
judge this stone, because of its cut and brilliance, to be absolutely authentic.)
So in that sense, one cannot judge something aesthetically without actually
experiencing it. But in another, perfectly proper, sense of judgment, I can
be judging something which I am not perceptually experiencing, and in many
cases, justified in so judging, if I am in possession of relevant and compelling
evidence as to the characteristics of that which I am judging, whether they be
aesthetic ones or not. (Some paradigm cases: I judge the candidate, from the
dossier in front of me, to have insufficient qualifications for the post and I
judge the Adagio of Beethovens Third Symphony, on the basis of centuries
of testimony as to its expressiveness, to be an extremely sad piece of music.)
That firsthand experience of aesthetically notable objects such as artworks is
the main point of our involvement with them should not lead us to think that
only through such experience can we learn anything about how they are aes-
thetically. Nor should the fact, if it is a fact, that no one could know how
an object was aesthetically unless someone, somewhere and sometime, has had
direct experience of it induce one to think that that someone must be you.
I conclude that if the supposition of aesthetic properties, as reportable in
straightforward predications made of artworks, has any tendency to under-
mine the doctrine of aesthetic autonomy, that is all to the good.

II WHY SHOULD WE RECOGNIZE AESTHETIC

PROPERTIES?

I turn now to Matraverss first worry regarding the existence claim about
aesthetic properties. In short, it is this: What work is done by positing aes-
thetic properties, for one who already accepts aesthetic experiences, caused by
non-aesthetic perceptible properties? Speaking of a dancer whose movement a

But see the admirable efforts of Malcolm Budd, Aesthetic Judgements, Aesthetic Principles
and Aesthetic Properties, European Journal of Philosophy 7 (1999): 295311, and Aaron Meskin,
Aesthetic Testimony: What Can We Learn from Others about Beauty and Art?, Philosophy and
Phenomenological Research 69 (2004): 6591.
 What Are Aesthetic Properties? 339

spectator is prepared to characterize as graceful, Matravers puts the challenge

to aesthetic realism like this: anti-realism is ontologically parsimonious. It
seems as if the whole story can be told using only the way the spectator exper-
iences the non-aesthetic properties of the ballet dancer. If this is the case, pos-
iting an extra layer of properties would be explanatorily redundant. For
Matravers, aesthetic properties capable of pulling their metaphysical weight
must have some explanatory value, in the sense both of explaining the exper-
ience we have of them and of explaining the normativity that attaches to
attributions of them.
As regards the second demand, Matravers thinks it will be particularly hard
to satisfy, insofar as the normativity of aesthetic attributions, roughly, their
admitting of correctness and incorrectness, is offered as a reason for positing
aesthetic properties, hence cannot, it seems, also be said to be explained by
them. But I am not sure of the cogency of this observation. For if the posit-
ing of aesthetic properties is the upshot of something like an inference to the
best explanation, so that an objects possessing an aesthetic property is the best
explanation of a parallel attribution to the objects being correct, is not that
sort of two-way relationship precisely what one should expect? For a realist,
aesthetic attributions admit of being correct or incorrect because objects really
do have or fail to have aesthetic properties; but equally, the sociolinguistic fact
of there being correct and incorrect aesthetic attributions gives us grounds to
posit corresponding properties in explanation of that fact.
It remains only to suggest why aesthetic properties are the best such explan-
ation available of the normativity in question. And that is because there are
in sight really only two alternatives to an appeal to aesthetic properties. One
would be an appeal to the ensembles of non-aesthetic perceptible proper-
ties which, in conjunction with perceiver psychologies, cause aesthetic exper-
iences, but no one has ever succeeded in elucidating how such indefinitely
varying and cognitively unruly ensembles can serve to underwrite the norm-
ativity of aesthetic judgments. Two would be an appeal to those aesthetic
experiences themselves, but then unaccounted for is which experiences, or
whose experiences, make for correctness of attributions.
As regards the first demand for explanatory payoff, I maintain that on a
higher-order-way-of-appearing conception of them, aesthetic properties do
in some sense serve to explain the generation of aesthetic experiences. It is
true that aesthetic properties so conceived may not figure in explanations of

Matravers, Aesthetic Properties, 208.

340 Aesthetic Properties

aesthetic experience conforming to a combinatorial model, whereby macro

phenomena are explained by being shown to be the transparent upshot of
combinations of micro phenomena, as in the explanation of thermodynamics
by statistical mechanics, but as many philosophers have noted, it is not
clear that any such explanation of experiences, or of the phenomena of
consciousness more generally, is possible at all. Leaving aside that sort of
explanation, then, which has had remarkable success in making intelligible
relations between the microphysical and the macrophysical, aesthetic
properties can plausibly figure in explanations of aesthetic experience of a
different, everyday sort. I return to this matter below, once I have fleshed out
somewhat more my conception of aesthetic properties.
Matraverss worry on the second claim enunciated above, the nature claim
about aesthetic properties, is essentially this: assuming there are adequate
grounds to posit aesthetic properties in addition to non-aesthetic perceptible
properties, such as colors, shapes, timbres, tastes, and textures, what exactly
are aesthetic properties, and how are such properties related to the non-
aesthetic ones on which they depend?
Before attempting, in the remainder of this essay, to address that worry I
want to acknowledge only to put aside an aspect of aesthetic attributions which
for many writers looms larger than any other, namely, their putative evaluat-
ive dimension. I will be assuming that the core of an aesthetic attribution is
a descriptive content, and hence that such attributions, whatever their evalu-
ative force, also centrally purport to ascribe properties. It is those properties,
corresponding to that descriptive content, on which I am concerned to shed
light, whatever the truth may be about the evaluative dimension of aesthetic
attributions more broadly viewed.

III WHAT PROPERTIES ARE

What are properties? They are, I suggest, ways things are, or equivalently, ways
of being. Properties so understood can be seen as the natural answer to one
For instance, Thomas Nagel, Colin McGinn, Ned Block, David Chalmers, and Joseph Levine.
I attempt to support this claim in the paper which is the main focus of Matraverss discussion,
Aesthetic Properties, Evaluative Force . . . . For skepticism about the viability of isolating an
objective descriptive content in aesthetic attributions, see Alan Goldman, Aesthetic Value (Boulder,
CO: Westview Press, 1995); John Bender, Realism, Supervenience, and Irresolvable Aesthetic
Disputes, Journal of Aesthetics and Art Criticism 54 (1996): 37181; and John Bender, Sensitivity,
Sensibility, and Aesthetic Realism, Journal of Aesthetics and Art Criticism 59 (2001): 7383.
For motivation and development, see my Properties and Related Entities, Philosophy and
Phenomenological Research 39 (1978): 122. Strictly speaking, however, although properties are
 What Are Aesthetic Properties? 341

of the basic interrogative thoughts about the world, namely, How?, just as
objects can be seen as the natural answer to an even more basic such thought,
namely, What? (Some other such thoughts, and the natural answers to them,
ones which in effect recapitulate the basic categories of metaphysics, are these:
Who? Person. Where? Place. When? Time. Why? Cause. How Many? Num-
ber. How Far? Distance. How Long? Duration.) Properties, understood as
ways of being, are thus closely related to possibilities, in that if it is possible
for an object to be a certain way W, then there exists a property, being W.
But that should not lead one to think that properties are possibilia. Properties
are realia, though abstract ones. A property exists, or is actual, insofar as it is
possible for things to be corresponding ways.

IV WHAT WAYS OF APPEARING ARE

Ways of appearing are a subclass of ways of being, just as appearing is a sub-

class of being. Alternatively, since appearing is a mode of being, a way of
appearing is a way of being. The ways things standardly appear are in effect
a part of how they are. Ways of appearing are roughly equivalent to what oth-
ers call manifest properties, meaning properties that reveal their natures in and
through their appearances.
Ways of appearing are, first, ways of appearing to perceivers of a certain sort;
and second, ways of appearing in certain conditions. They are thus implicitly
perceiver-relative and condition-relative. The first implicit relativity reflects
the fact that given ways of appearing do not manifest themselves to all
sentient creatures, but only to those with an appropriate sensory-perceptual-
cognitive apparatus; the ordinary colors of things, for instance, are ways of
appearing, but they are not detectable by amoebae or amphibians. The

closely connected to ways of being, they cannot quite be identified with them. For example, red is
a way of being, but the property that corresponds to it is not red, but rather being red. A property
is a being-a-certain-way, or condition, in which a way of being is intimately involved. The fine
distinction between properties and ways of being is a reflection of the logical grammar of property
talk: whereas a thing is its ways of being, it has its properties. In any event, I will be ignoring this
nicety for the purposes of the present essay, thus treating properties and ways of being as equivalent.
See Mark Johnston, The Manifest, web manuscript (www.nyu.edu/gsas/dept/philo/courses/
consciousness97/papers/johnston): If other things and people are to be manifest to me then they
must have the natural capacity to appear as they are. That is to say that the manifest properties
of things can be themselves the ways in which things appear . . . The idea of the colors as ordinary
manifest properties is the idea that appearances which present the colors of things thereby show us
what the colors are like.
See Budd, Aesthetic Judgements . . . , 308: It is clear that what aesthetic properties an item
possesses . . . is relative to a kind of sensibility: the acceptability of aesthetic judgements made by
342 Aesthetic Properties

second implicit relativity reflects the fact that given ways of appearing do
not manifest themselves in all circumstances, but only those conducive to
such manifestation; colors again, are ways of appearing, but they are not
perceivable in the absence of light, even by appropriately equipped perceivers.
Third, although this is not a relativity but rather a specification, ways of
appearing are standardly ways of appearing in some sensory modality, such as
visual, aural, or tactile.
The upshot of an interaction between a perceiver, an object, and one of its
appearance properties, or ways of appearing, is an event, what one can call an
appearing. And a subject Ss perceiving an object Os way of appearing W can
be equated with Ss being appeared to in way W by O. But none of that is
to say that there are appearances, in the sense of introspectible mental things,
existing within the mind.

V WHAT AESTHETIC PROPERTIES ARE

My proposal is that aesthetic propertiesor at least many of what are usu-

ally classified as aesthetic propertiesare higher-order ways of appearing,
dependent in systematic fashion on lower-order ways of appearing but
not conceptually tied to them or deducible from them. Such higher-order
ways of appearingfor example, delicate, graceful, melancholy, unified,
human beings must be understood as relative to a distinctive human sensibility, with distinctively
human powers of perception, understanding and emotional response.
Aesthetic ways of appearing are not as such sensory ways of appearing, but they standardly
involve some sensory or quasi-sensory modality. Thus, if a drawn line is graceful, or a sung phrase
poignant, or a fabric silky, one can also say that the first is visually graceful, the second aurally
poignant, the third tactually silky.
See Johnston, The Manifest, for admonitions on this score.
The claim here is that many aesthetic properties are higher-order ways of appearing. But I do
not claim, conversely, that higher-order ways of appearing are all aesthetic properties, because for a
way of appearing to be an aesthetic property arguably requires further conditions. This raises the
demarcation problem for the class of aesthetic properties. What, apart from being a higher-order
way of appearing, might be required for something to count as an aesthetic property? I would here
reiterate a suggestion offered in my Aesthetic Supervenience (Music, Art, and Metaphysics (Ithaca:
Cornell University Press, 1990). Though not all aesthetic properties are inherently evaluative, or
such as to imply an aesthetic evaluation of the objects that possess them, and though not all aesthetic
properties are even value-tending, that is, such as to conduce to the aesthetic value of the objects that
possess them, what does seem true of all aesthetic properties is that they are value-relevant properties,
that is, properties it is prima facie intelligible to cite in support of an aesthetic evaluation of an object
that possesses them. So for a way of appearing to be an aesthetic property is for it to be, in addition,
(a) higher-order, or dependent asymmetrically on other, lower-order perceptual properties, and
(b) value-relevant. What about representational properties, such as depicting a horse, which are
clearly both perceptual and higher-order? Probably these should not be accounted aesthetic, because
not value-relevant, unless representing something is held to be prima facie better than representing
 What Are Aesthetic Properties? 343

balancedarise out of the lower-order ways of appearing on which they

depend in a holistic or emergent manner. And to say that a way of appearing
is a higher-order one is to say precisely that it depends, and in an asymmet-
ric fashion, on other ways of appearing, which can on that account be labeled
lower-order ones. Thus, a sculpture is delicate, say, in virtue of its dimen-
sions and contours, but it is not thus dimensioned or contoured in virtue of its
delicacy.
Consider some familiar lower-order ways of appearing, such as colors and
timbres. Colors are ways of appearing visually, timbres are ways of appearing
aurally. In more down-to-earth terms they are, respectively, looks and sounds
that objects, or as the case may be, events, possess and display. Many aesthetic
properties, I suggest, are also ways of appearing, but of a higher order. Such
aesthetic ways of appearing may also, where vision or audition is involved
in their perception, be described as looks and sounds, but of a more subtle
and elusive sort than colors and timbres. The graceful look of a physical
movement, the melancholy sound of a musical phrase, depend on the lower-
order looks and sounds of the movement and phrase in question, and arise out
of them without collapsing into them.

VI ARE WAYS OF APPEARING DISPOSITIONS?

Is a way of appearing W just a disposition to appear W, so that perceptible

properties, both non-aesthetic and aesthetic, are in effect just disposition-
al properties? No, and for a number of reasons. One reason is that ways of
appearing are clearly manifest properties, ones we directly perceive, whereas
dispositions, even dispositions to afford appearances, being inescapably rela-
tional properties, are not. A things disposition or power to cause effects or
induce changes cannot be directly perceived, but only inferred from those
effects or changes. A second reason is that something can possess a way-of-
appearing without in fact having the disposition to appear that way to relevant
observers, owing to the extraordinary circumstances of the thing in question.
An example: the suns radiation zone, the region inside the sun immediately

nothing. By contrast, stylistic properties comfortably count as aesthetic on this score, since having a
recognizable style is plausibly an artistic plus of some sort.
That the aesthetic properties of objects are not deducible from but instead emergent on their
non-aesthetic perceivable properties is a thesis one can discern in the writings of Frank Sibley,
Monroe Beardsley, and Kendall Walton on aesthetic attributions, though Sibley, of course, was
reluctant to speak of properties at all.
344 Aesthetic Properties

surrounding its core, is thought on good grounds to emit spectral red light,
and thus to be radiant red in color; yet it is not disposed to appear that way
to relevant observers, since no sighted observers of any sort can exist in the
circumstances such observation requires. Of course a colored object will
typically have the disposition to appear as it is with respect to color, but such a
disposition is a consequence of its color, and not equivalent to it.
A third reason to resist the identification of appearance properties or ways
of appearing and dispositions to appear is that the concept of a way of appear-
ing does not include anything about conditions of observation or types of
observers, though it is true that in order to judge properly of a given way of
appearing, such as a color/look, timbre/sound, flavor/taste, the conditions
of observation have to be apt, and you have to be an apt sort of observer.
Which is to say that such properties are inherently indexed to such param-
eters or implicitly relative to them, as acknowledged earlier. But that does not
mean that such parameters figure in the concept of the way of appearing itself,
that which is both directly perceived and ascribed to the object that presents
it. Appearance properties or ways of appearing contrast in that respect with
dispositions, in whose concepts such parameters figure explicitly. In other
words, appearance properties or ways of appearing, unlike dispositions, are
not as such relational properties.
The idea of red is arguably the idea of a manifest appearance, or kind of
appearance, to which certain perceivers are sensitive. It is not the idea of a
disposition of an object to produce a kind of response, such as a sensation or
feeling, in perceivers, nor is it the idea of a disposition of perceivers to have
such responses to such objects.

VII EXPL AINING AESTHETIC EXPERIENCES

We can explain someone experiencing a dancers movement as graceful by

appealing to the fact that the movement really has the way of appearing grace-
ful, the conditions being right for such a way of appearing to manifest itself,
and the persons being an apt subject for that way of appearing, that is, the
sort of subject to which that way of appearing is implicitly relative. And we

See Johnston, The Manifest.

Something like this point is made by P. M. S. Hacker in Are Secondary Qualities Relative?,
Mind 95 (1986): 18097.
 What Are Aesthetic Properties? 345

can proceed similarly if called upon to explain someones experiencing a

passage of music as anguished, or a painting as garish, or a design as balanced.
Such explanations at least seem to be ordinary causal explanations.
However, in order for them to really be so, it would have to be a contingent
matter that apt subjects for a way of appearing W in apt conditions experi-
ence objects possessing W as W. For if that were instead an a priori matter,
then the ostensible causal explanations involved would not be genuine ones,
since the explanandum and explananda would be necessarily linked. Accord-
ing to Mark Johnston, however, where perceptible properties such as colors
are concerned, these sorts of explanations are indeed genuine, since it can
be shown to be only contingent that standard subjects for a given such col-
or property P, and in standard conditions for perceiving P, perceive an object
having P as P. In my terminology, that would be to say that it is a con-
tingent, though entirely to be expected, matter that an object possessing a
lower-order way of appearing W will appear W to apt subjects in apt con-
ditions, i.e. those to which the way of appearing W is implicitly indexed or
relative.
Johnstons argument for that is complicated, part of a long debate on the
nature of secondary qualities, and difficult to evaluate. But if Johnston is right,
and there is real explanatory value vis-a-vis experiences in appeal to perceptible
properties conceived of non-dispositionally, or as manifest properties or ways
of appearing, then if many aesthetic properties are properly conceived along
the same lines, as manifest though higher-order ways of appearing to which
certain classes of perceivers are receptive, appeals to such aesthetic properties in
relation to aesthetic experiences will be similarly explanatory, and in a robustly
causal manner. However, even if Johnston is wrong, and an explanation of the
experience of a movements gracefulness by appeal to its being graceful, that
is, its possessing that way of appearing, cannot ultimately be judged robustly
causal, it does, I submit, provide some sort of explanation of the experience in
question, on an ordinary or pragmatic level. For it at least relates the experience
to something beyond itself, a stable way of appearing, which accounts for the
experiences having the character it does.
To simply hold, as Matravers would have it, that I perceive the movement
as graceful because of how I am perceiving the movements non-aesthetic
properties, provides no explanation whatsoever of the specific character of my

See Mark Johnston, Are Manifest Qualities Response-Dependent?, Monist 81 (1998): 343.
346 Aesthetic Properties

perception, amounting, it seems, just to saying that I perceive the movement

as graceful because I perceive it as graceful.

VIII WHAT MIGHT AESTHETIC WAYS OF APPEARING

INVOLVE?

A look is a visual way of appearing. But what is it for an object to look a certain
way W, where W is other than a straightforward lower-order visual way of
appearing, such as red or round, but instead something like graceful or garish?
One popular suggestion is that it is for a standard or appropriate viewer
to have a feeling F in looking at W. Thus in the case of gracefulness, the
feeling in question might be said to be a gentle, melting sort of pleasure; in
the case of garishness, a mild, jarring sort of displeasure. But if gracefulness
or garishness is conceived as a higher-order look, or way of appearing visually,
then it seems we would have to account the feeling had upon perceiving an
object, or perhaps the objects propensity to afford that feeling, as part of the
way of appearing in question.
So, are higher-order looks such as graceful and garish partly a matter of feel-
ings consequent on perception of the objects displaying such looks? More
generally, are some higher-order, aesthetic ways of appearing partly a matter
of an objects power to afford perceivers distinctive feelings?
Here is what I had to say on this vexed point in my earlier essay, the term
phenomenal impression having the same force as way of appearing:
[this realist account of aesthetic properties] presumably gains in convincingness the
more light we can throw on exactly what an overall phenomenal impression, the
suggested core of an aesthetic attribution, might involve. Though I have until now
stressed the perceptual character of such impressions, we need not deny that such
impressions might be partly affective as well. Some aesthetic impressions might, say,
be bound up with feelings consequent on apprehending an objects non-aesthetic
features.
I am now less sanguine about that attempt to encompass perceiver reactions
within aesthetic impressions, within the ways an object aesthetically appears.
An objects power to induce feelings in a perceiver, even ones closely tied

See, for instance, Derek Matravers, Aesthetic Concepts and Aesthetic Experiences, British
Journal of Aesthetics 36 (1996): 26577. An important source of this sort of suggestion is Georges
Santayana, who famously held that beauty was a projection onto an object of a perceivers feeling of
pleasure in beholding the object.
Aesthetic Properties, Evaluative Force, 69.
 What Are Aesthetic Properties? 347

to the act of perception, cannot, it seems, really be part of the objects way
of appearing. If that is so, then if the popular suggestion is accepted, and
aesthetic properties such as gracefulness and garishness must indeed be
analyzed in terms of affective reactions in perceivers, at least some aesthetic
properties cannot be regarded simply as ways of appearing.

IX ARE AESTHETIC PROPERTIES

RESPONSE-DEPENDENT?

These questions in effect raise the issue of the response-dependence of aesthetic

properties. A response-dependent property is one the substantive concept of
which involves the idea of a distinctive sort of response had by perceivers to
objects possessing the property. Alternatively, a response-dependent property
is a property for which there is an a priori connection between the possession
of the property by an object and the having of distinctive responses by per-
ceivers.
So, are aesthetic properties more akin to manifest sensible properties, like
colors and timbres, which are at least arguably non-response-dependent, or
to perceiver-dispositional properties like nauseatingness or infuriatingness,
which are unmistakably response-dependent? Are aesthetic properties
fundamentally a matter of how things look and sound, or are they a matter
of how perceiving those things, or their lower-order looks and sounds, makes
us feel?
Perhaps the truth is this. Some aesthetic properties really are essentially
looks and sounds, explicable apart from feelings such looks and sounds may
engender in perceivers, and thus higher-order ways of appearing. But some are
not. The latter sort of aesthetic property is instead to be analyzed on a form-
perceived-with-feeling model, and thus as inescapably response-dependent.

X T HE SPECTRUM OF AESTHETIC PROPERTIES

Properties ordinarily classified as aesthetic seem to range from ones that are
clearly response-dependent, to ones that are possibly response-dependent, to

Johnston offers the following more precise characterization: Let us say that a property, being
F, is response-dependent if there is some predicate is f which expresses the property . . . such
that some substantial way of filling out R, S, and C makes a priori and necessary [the
biconditional] x is f if and only if x is disposed to produce x-directed response R in all actual and
possible subjects S under conditions C (Are Manifest Qualities Response-Dependent?, 9.)
348 Aesthetic Properties

ones that are arguably non-response-dependent. Let us examine a number of

them and try to situate them on this dimension.
On the one hand we have what are sometimes called formal aesthetic
properties, ones that attach closely to an objects perceivable form, such as
balance. This seems a clear example of a non-response-dependent aesthetic
property. It plausibly denotes an overall way some configurations look, not a
feeling they give us on looking at them. It seems unlikely that there is a feel-
ing, whether one of balance or not, that is invariably had when contemplating
something balanced. Much the same could be said of unity, or dynamism,
or fluidity. These plausibly denote characteristic visual appearances, accessible
to sight, rather than propensities to induce distinctive feelings accessible to
introspection.
Clearer still seems the status of stylistic aesthetic properties, such as those
which correspond to the labels impressionist, fauvist, cubist, futurist in
painting. These are readily understandable as visual looks of a high order,
which are often detectable at twenty paces, presumably before any feelings
consequent on their perception could announce themselves.
On the other hand we have aesthetic properties such as human beauty and
ugliness, in which reactions of pleasure and displeasure to perceived looks seem
essential to what those properties are. Consider facial loveliness, one of course
relative to human sensibility, and perhaps also to a specific ethnic or cultural
sensibility. That property seems plausibly analyzed in terms of the affording of a
distinctive sort of pleasurable feeling, one tinged with desire, had in perceiving
the basic visual features of a face. Or take bodily sexiness, perhaps also account-
able an aesthetic property. It is remarkable, and no doubt often remarked, how
the difference between sexy and dumpy where bodies are concerned can be the
result of just a slight difference in curve at the hip, or chest-to-legs ratio, or
degree of fleshiness. But the resulting sexiness or dumpiness itself seems
Such properties, being higher-order ways of appearing, are also accessible to other sense-
modalities, such as hearing, and not only to sight. Thus, a dynamic painting and a dynamic musical
passage both present a dynamical appearance, the one visually and the other aurally.
I here have in mind stylistic properties in a thin rather than thick sense, that is, ones
understood in exclusively visual or formal terms, without regard for historical provenance; this is
signaled by denominating such properties in lower-case. Thus, a painting can be said to display
cubist style (thin sense), if not Cubist style (thick sense), if it simply looks like a Picasso or Braque
of the relevant period, regardless of when it was made or by whom.
Note that the claim here concerns specifically human beauty. It is my view that both natural
beauty, e.g. that of sunsets or lakes, and artistic beauty, e.g. that of Mozart sonatas or Mondrian
paintings, are quite different properties from beauty in human beings, not so directly tied to
hardwired responses of an evolutionarily rooted sort. Those other beauties are nevertheless also
likely amenable to a form-perceived-with-feeling model, and so response-dependent.
 What Are Aesthetic Properties? 349

undetachable from a distinctive pleasurable reaction, one in which physical

desire is even more pronounced than in the case of facial loveliness. Consider
musical tension, attributions of which seem to presuppose some experience of
psychological tension on a perceivers part, and not just the music in question
sounding a particular way. Or take, finally, sublimity, which has often been
understood to involve something like a feeling of fear or awe in perceiving an
object. Insofar as sublimity, musical tension, facial loveliness, and bodily sexi-
ness demand analysis on an object-or-form-perceived-with-feeling model, they
are clearly response-dependent properties, and not comfortably accountable
manifest ways of appearing.
What of expressive properties, perhaps the most important kind of aesthetic
property? Schumanns Rhenish Symphony, to take a musical example, dis-
plays a remarkably varied palette of expressive properties in the course of its
five movements. In particular, its third movement, Nicht schnell, presents a
tender, solicitous, and hesitant character, its fourth movement, Feierlich, a
mournful and lugubrious one, and its fifth movement, Lebhaft, one that is
cheerful, confident, and good-hearted. Are these expressive properties strai-
ghtforwardly audible ways of appearing belonging to the music, or are they
partly a matter of how we standardly feel when we hear the musics more basic
audible features of melody, rhythm, harmony, and so on? I am inclined to
hold the former position, though not with the utmost confidence.
Return, finally, to gracefulness and garishness, with which we initiated our
inquiry into what aesthetic ways of appearing might conceivably amount to.
If the distinctive sorts of pleasant and unpleasant feelings those properties
occasion are held to enter into what those properties are, so that for some-
thing to be graceful or garish is in part for it to occasion such feelings in
relevant perceivers, then gracefulness and garishness cannot be understood
straightforwardly as ways of appearing. But it is very hard to decide whether
that is so. The affording of the feelings in question might very well be cru-
cial to those attributions, but there is another possibility. The pleasant and
unpleasant feelings commonly had on viewing graceful and garish objects
may not be essential to those properties, but instead only concomitants of the
perception of those objects and the higher-order ways they appear, such ways

See Kendall Walton, Projectivism, Empathy, and Musical Tension, Philosophical Topics 26
(1999): 40740, which argues that some degree of tension in appropriate listeners is necessary for
a musical passages counting as tense.
In offering these characterizations I am addressing in each case the expressive character of the
opening strains. As the music proceeds its expressive character of course shifts and evolves, though
a fairly homogenous emotional atmosphere is in fact maintained in each of the movements noted.
350 Aesthetic Properties

of appearing being detectable apart from the feelings that commonly result
from their being perceived. So it remains, to my mind, an open question
whether gracefulness, garishness, and the like can be understood as higher-
order ways of appearing, hence as arguably non-response-dependent, or only
as partly dispositional, response-dependent properties in which perceiver reac-
tions are irreducibly implicated.

XI ARE AESTHETIC PROPERTIES CONSTITUTED

BY NON-AESTHETIC PROPERTIES?

Are aesthetic properties, or at least those which can be understood as higher-

order ways of appearing, constituted by the lower-level perceptible properties
or ways of appearing on which they depend and from which they emerge?
Are the expressive properties of Schumanns symphony noted above, for
example, constituted by its musical features, that is, its specific melodic con-
tours, rhythmic values, harmonic progressions, and timbral combinations?
The constitution relation has its home in thinking about concrete partic-
ulars, notably objects, such as a statue, which may be thought to have both
a defining form and a composing matter. A statue is composed of, or consti-
tuted by, its matter, a given mass of marble, but is not identical to that matter,
irrespective of form. The statue is one kind of thing, with a certain distinct-
ive organization essential to it, whereas the mass of matter is another kind of
thing, with no organization essential to it beyond contiguity of its parts.
But can the relation of constitution hold between abstract entities such as
properties? This seems hard to fathom, since all properties are of the nature
of formsas the Platonic view of them reminds usand not matter. Hence,
how one property could be constituted by other properties, properties that
would compose it as marble might a statue, remains obscure. What, after all,
could compose a property, in the sense of being its matter?
Let us turn, though, to another sort of concrete particular, namely events.
Here we have more of a grip on how there could be constitution relations

An early statement of the view that expressive properties, an important category of aesthetic
properties, are constituted by the non-aesthetic properties on which they depend can be found
in Alan Tormey, The Concept of Expression (Princeton: Princeton University Press, 1971), ch. 5.
Tormey writes, for example, that the tempo, dynamics, harmonic texture, melodic contour of
Ravels Pavane are not merely the grounds, warrants or criteria for asserting that the work is tender,
they are the constituents of its tenderness (p. 131).
By contrast, that a given property might supervene on, entail, or preclude other properties
seems relatively unproblematic, these being broadly logical relations.
 What Are Aesthetic Properties? 351

among them, with some events serving as the matter of others. Consider
a familiar example involving action: my signaling for a cab is constituted,
on a given occasion, and in a given cultural matrix, by my making a hand
movement of a specific sort. The actions are not identical, though intimately
related: the latter constitutes the former.
So instead of suggesting that an aesthetic property understandable as a
higher-order way of appearing is constituted by, or made up of, the non-
aesthetic perceptible properties on which it depends, what we might suggest
instead is that the event of an objects having an aesthetic property is con-
stituted by the event of the objects having the non-aesthetic properties on
which it depends. Going further, we might also suggest that the event of
the objects manifesting such an aesthetic property is constituted by, or made
up of, the events which are the manifesting of the non-aesthetic perceptible
properties on which it depends.

The events invoked here, of course, are what are usually referred to as states of affairs.
This is a bit too simple, in that reference to an appropriate perceiver has been omitted. More
fully, what might be true is that the manifesting or appearing of an objects aesthetic property to an
appropriate perceiver, that is, one who perceptually frames or categorizes the object in the right way,
is constituted by the manifestings or appearings to such a perceiver of the non-aesthetic properties
on which the aesthetic property depends.
This page intentionally left blank
PART V I
H I S TO RY
This page intentionally left blank
 21
Schopenhauers Aesthetics

I INT RODUCT ION

Arthur Schopenhauer was a nineteenth-century German philosopher, who

lived from 1788 to 1860. He was an arch-rival of Hegel, and considered him-
self Kants only true heir in philosophy.
Schopenhauers aesthetics has arguably been more widely influential than
that of any other philosopher of the past two hundred years. Not only
have subsequent philosophers, notably Nietzsche, Bergson, Wittgenstein, and
Langer, been much affected by his vision of the place and power of art, but
a wide array of writers, composers, and intellectuals have testified, either
explicitly or in their own works, to the power of that vision and of the
metaphysics in which it is embedded: Wagner, Mahler, Stendhal, Tolstoy,
Rilke, Mann, Freud, Proust, Hardy, Conrad. It would be fair to say that
Schopenhauers aesthetics has been, in the twentieth century, the artists
favored philosophy of art.
One cannot understand Schopenhauers aesthetics without some
understanding of Schopenhauers metaphysics. And it is impossible to
understand Schopenhauers metaphysics without at least a passing grasp of
the metaphysics of Kant, Schopenhauers great predecessor and chief mentor
among philosophers. Fortunately, however, it is possible to understand
Schopenhauers aesthetics without understanding Kants aestheticsthough
here too, naturally, influences and similarities can be noted. Thus, for the
purposes of this article, we can make do with thumbnail sketches of Kants
and Schopenhauers metaphysics, and then proceed to the main topic of
concern, Schopenhauers aesthetics.

First published as Schopenhauer, Arthur, in M. Kelly (ed.), The Encyclopedia of Aesthetics (New
York: Oxford University Press, 1998), 24550.
356 History

II THE METAPHYSICAL BACKGROUND

The central notion in Kants philosophy, the philosophy of transcendental

idealism, is that the empirical world, the world known in perception, is in
significant part a product of the minds active structuring of the raw data
of experience. This structuring is in terms of the forms of sensory intuition,
namely, space and time, and the categories of the understanding, such as sub-
stance, causality, and relation. This structuring, furthermore, has a necessary
or a priori status; without it, coherent experience of an objective world would
not be possible, nor would knowledge of the sort represented by arithmetic,
geometry and pure physics. Hence, the world as it is known to us cannot, in
virtue of such extensive and ineluctable structuring by the knowing mind, be
identified with the world as it is in itself. The former has features which, since
they are contributed by our very mode of knowing, cannot be features of the
latter. The world as it is in itself must thus be grantedat least negatively, as
a limiting condition of the world of experiencewhile evidently remaining
beyond the possibility of being known as such. The known, or phenomenal,
world is not all there is to reality: there is an unknowable, noumenal world, in
some way underlying the appearances constituting the phenomenal one, and
in unfathomable correspondence with it.
Schopenhauer takes over Kants transcendental idealism wholesale, but
with certain simplifications. Though accepting space and time as primary
modes of perceptual structuring, Schopenhauer reduces Kants twelve
categories of the understanding to one, causality, and regards space, time,
and causality as in fact all forms of the Principle of Sufficient Reason, that
everything must have a reason or ground for existing in just the way it
does, which applies to all appearances or representations whatsoever. The
Principle of Sufficient Reason is valid only for the world of phenomena,
including human action, and has no application beyond it. As we shall
see, Schopenhauer ultimately entertains a notion of knowing outside of
the Principle of Sufficient Reason, a knowing which finds its clearest
exemplification in art, but this remains confined to the phenomenal realm,
because it still participates in and thus is affected by the fundamental
condition of knowledge in general, namely, the distinction between subject
and object, or knower and known. For Schopenhauer, subject and object
are correlative terms: all objects are inherently objects for a subject, and all
subjects are inherently subjects cognizing objects; moreover, the characters
 Schopenhauers Aesthetics 357

of subjects and objects so correlated are always themselves importantly

parallel.
Schopenhauer refers to space, time, and causality collectively as the prin-
cipium individuationis. The existence of a plurality of distinct individuals and
the existence of a framework of spatial, temporal, and causal relations are, for
Schopenhauer, inseparable. But since space, time, and causality are aspects of
the human way of knowing, individuality and plurality are themselves arti-
facts of that way of knowing, and cannot bear on the world as it is in itself.
Noumenal reality must be nonplural, or nondivided, in nature. But can any-
thing more be said, in a more positive vein, about that reality?
This is where the most distinctive part of Schopenhauers metaphysics
emerges. Though Schopenhauer, as a Kantian, holds throughout that the
in-itself can, in the nature of things, never be known as such, there are still
experiences available at the phenomenal level which give some insight into the
nature of noumenal reality, and which ground a certain kind of extrapolation
to its character. The key to this knowledge, albeit relative, of the thing-
in-itself, of the inner nature of the world, is ones own person, and more
particularly, ones own body. A Schopenhauer puts it, we ourselves are the
thing-in-itself (WWR, ii. 195).
All other objects one knows only from the outside, as items in space and
time enmeshed in a web of causes and effects. Ones own body, on the other
hand, is known also from within, introspectively and non-inferentially. The
purposeful movements of this body, observable to all externally, are known at
the same time by the agent, internally, as acts of will. My body and my will are
one: what manifests itself outwardly as matter moving through space mani-
fests itself inwardly as agency or volition. Whats moreand this is the cru-
cial step in Schopenhauers metaphysical argumentin knowing my body
from the inside I am knowing it in a way that gives some indicationthe
most it is possible for me to have, from this side of the phenomenal fenceof
what it is like in and of itself, because certain of the forms of structuring
which operate to generate phenomenal appearances are absent. In particular,
the appearance of bodily action from within is freed of the filter of space, and
to some extent, causality, even though such an appearance is still in time, i.e.
apprehensible only successively, and still reflects, as an appearance, the essen-
tial division between knower and known.
Thus, Schopenhauer proclaims, my entire inner nature is will-
ingdesiring, striving, urgingand all the varieties of affect and emotion,
pleasure and pain, that conduce to or inhibit action. Furthermore, I must take
358 History

this as the best indication I have been vouchsafed of the noumenal character
of myself: since when the distorting forms of phenomenal appearance are par-
tially removed, the in-itself reveals itself as essentially conative, we can only
assume that were they to be all removed, the in-itself would display a char-
acter somehow further along in that direction. (Schopenhauer is careful to
caution, though, that we can have no assurance that this would be so; what
the world is in itself absolutely, apart from all knowing, remains for him as for
Kant, unknowable.) Finally, by a second extrapolation, relying in part on the
previously established nonplurality of noumenal reality, Schopenhauer draws
the conclusion that the inner nature or noumenal character of everything is,
again, to our best approximation, on the order of will as well. If all other phe-
nomena could be known by us just as immediately and intimately [as our own
actions], we should be obliged to regard them precisely as that which the will
is in us (WWR, ii. 197.). All natural phenomena, including physical forces,
are to be understood as forms of the sort of willing or agency with which we
are familiar in our own cases, and the phenomenal world as a whole as the
manifestation of a single, undifferentiated, so to speak cosmic, Will. In fact,
and serving as a kind of empirical confirmation of this metaphysical deduc-
tion, the whole of nature, organic and inorganic, according to Schopenhauer,
shows itself when suitably viewed as nothing other than a theatre in which the
universal will manifests itself in innumerable ways, and in the playing out of
which conflict, frustration, and suffering are ubiquitous and inevitable.
This brings us, naturally, to Schopenhauers famous pessimism. Human
nature is at its core essentially striving or desire. But desire is a state of
lackof not havingand is thus an inherently unpleasant and disagreeable
condition. Furthermore, the needs and wants of a given individual are
generally both internally in conflict and externally in conflict with those
of others; thus, as might be expected, the unclouded satisfaction of desire
only infrequently occurs. Finally, such satisfaction as occurs is a very minor
good, for three reasons. First, it has an entirely negative character, being just
the pleasure of momentary cessation of desiring. Second, any desire actually
fulfilled, and so extinguished, is as a rule quickly replaced by a multitude
of others, whose noisy demands soon drown out the sensation of relief just
noted. And third, in the odd event that new desires do not immediately
surge up to take the place of those which have been quelled, what one has
is nothing more than a state of perfect boredom or ennui, no more pleasant
than the more usual state of constant unfulfilledness. It should be borne in
mind that the foregoing diagnosis, which might seem to be based simply on a
 Schopenhauers Aesthetics 359

somewhat jaundiced observation of human psychology, together with some

conceptual analysis, is for Schopenhauer undergirded by the metaphysical
conviction that persons, and indeed the whole of existence, have willing as
their essential and thus inescapable nature. Thus, insofar as one remains
anchored in willing as a spatiotemporally bound individuala bundle of
strivings and cravingssuffering is virtually guaranteed.

III SCHOPENHAUERS AESTHETICS

There is, however, a means of temporary escape from this sorry condition.
It is afforded by aesthetic experience. Aesthetic experience, unlike ordinary
perception, is focused not on material particulars in space and time, but on
the perceivable essences or universals which such particulars embody, and
which Schopenhauer, following Plato, calls Ideas. In the course of focusing
on such objects the perceiver is in effect transmuted: spatiotemporally rooted,
practically oriented individuality gives way, and what Schopenhauer calls the
pure subject of knowing, the same in everyone, takes its place. In aesthet-
ic experience, one knows the world independently of the Principle of Suffi-
cient Reason, grasping not the why or how of things, but only the what:
the knowledge faculty, ordinarily the servant of the will, becomes simply the
mirrorer of Ideas, which are not interesting to the individual in which this
faculty resides.
There are thus two sides to aesthetic experience, an objective and a sub-
jective. On the objective side, there are the Ideas embodied in concrete
particulars, on which attention now rests, and whose natures are grasped in
contemplation; the aesthetic experience is thus centrally a cognitive one. On
the subjective side, there is the transformation of the perceiver from an inter-
ested bundle of willing, concerned with objects only insofar as they are related,
spatiotemporally and causally, to the satisfaction of its needs and desires, into
a disinterested beholder of Ideas, with which individual willing can have noth-
ing to do. The disengagement of the subjects will and the shift in focus from
one sort of object to another are two sides of the same coin; as noted earlier,
subject and object for Schopenhauer are always correlated.
Schopenhauer thinks of the Ideas as grades of objectification of Will, or
thing-in-itself, on the phenomenal level. Ideas are intermediary between the
nonplural Will and the plurality of spatiotemporal individuals, but an Idea
is still a representation for a subject, and thus not a thing-in-itself. Ideas are
the most direct of the wills objectifications, or manifestations for a knowing
360 History

mind, logically prior to the plurality of individuals which Schopenhauer con-

ceived as in effect arising from the refraction of the Ideas through the lenses
of space, time, and causality. The Ideas are something like the fundamental
kinds of the phenomenal world, or the essences of them. There are Ideas cor-
responding to the species of living things, the varieties of natural forces, and
the innumerable individual human characters, conceived of by Schopenhauer
as virtually each a species unto itself. Unlike abstract concepts, however, Ideas
are intuitively apprehensible; they are grasped in perception, and not through
reason or language. But how the Ideas of living things, say, can be both per-
ceptually graspable, and yet entirely nonspatial, is not something for which
Schopenhauer ever gives a satisfying explanation.
The artist or man of genius, according to Schopenhauer, is one who is
particularly gifted in two respects. First, through an excess of intellect bey-
ond what is required for the practical purposes of daily living, the artist is
enabled to perceive the Ideas in things more readily, more widely, and more
sustainedly than the ordinary man, who as a rule has just enough under-
standing to grasp things in their spatiotemporal-causal and thus will-relevant
relationship to him. Whereas to the ordinary man his faculty of knowledge is
a lamp that lights his path, to the man of genius it is the sun that reveals the
world (WWR, i. 188). Second, the artist has the ability to embody this appre-
hension of Ideas in an artwork, a perceivable object in which the Idea has been
made more vivid, more striking, more easily discerned than it was in nature,
and so capable of triggering in the ordinary perceiver the sort of will-less con-
templation that the artist had before nature unassisted. Yet Schopenhauer
stresses that the power of recognizing in things their Ideas, of divesting them-
selves for a moment of their personality . . . must be present in all men in a
lesser and different degree, as otherwise they would be just as incapable of
enjoying works of art as of producing them (WWR, i. 194.). Schopenhauer
can be viewed as taking up Plotinus response to Platos notorious charge
against art that it merely copies what is ontologically already inferior, by stress-
ing that the artist embodies in his work not the mundane, imperfect, and
atypical objecteven when this serves as model or ostensible subjectbut
rather its Idea or essence, and in such fashion that others are enabled to exper-
ience the sort of transcendence of self in viewing it that the artist achieved
unaided in his interaction with the world.
So far I have highlighted the cognitive dimension of aesthetic awareness
more than the hedonic, but Schopenhauers conception certainly includes the
latter. The pleasure of aesthetic experience on his account would seem to be
 Schopenhauers Aesthetics 361

twofold. One part, the one most stressed by him, is purely negative: tem-
porary relief from the pain of constant striving, through transcendence of the
standpoint of the individual willer. But a second part, less stressed though no
less important, is of more positive character, and connects the hedonic aspect
with the cognitive: satisfaction in contemplation of the given Idea and in the
insight thus afforded into the timeless manifestations of Will. In effect, this
is a delight in knowing, such as Aristotle and Kant, in different ways, posit
as well. The negative pleasure of relief from the cycle of want and desire is
provided equally, it seems, by any object offering an Idea for will-less contem-
plation; the positive pleasure from knowledge of an Idea, though, rather varies
with the significance of the Idea involved.
Appreciation of beauty in nature and appreciation of beauty in art are both
founded in contemplation of Ideas. In the one case Ideas, or visible essences
of willing, are simply strikingly evident in the world itself, in its unmodified
state; in the other case, Ideas grasped in the world through the extraordinary
perceptions of genius are embodied in a created objectthe distilled exper-
ience of the geniusso as to be available to those of only modest powers of
perception. At one point Schopenhauer remarks that everything is beautiful.
What he means is that since everything embodies Ideas, if only those of the
simplest sort, everything can in principle be made the object of disinterested
contemplation. But on the other hand, as he also remarks, some things are
more beautiful than others. This is because, in virtue of their forms, wheth-
er natural or man-made, or in virtue of the value of the Ideas they offer to
contemplation, objects will differ in the readiness with which they prompt
such contemplation and in the worth of the contemplation they so prompt.
Schopenhauer, with Kant, is thus one of the important sources of the idea
of the aesthetic attitude as disinterested attention that can theoretically be
brought to bear on anything, whatever its nature or degree of fashioning,
thus potentially aestheticizing the world. Schopenhauers emphasis on the
non-spatiotemporal Idea rather than the concrete particular as the object of
aesthetic attention might also be seen as a cousin to Kants notion of the dis-
interestedness of aesthetic judgment as rooted in the disconnectedness from
real existence of the object so judged.
Schopenhauer, like Kant before him, distinguishes the sublime from the
beautiful, and the account he offers is more convincing than Kants. The
difference between beauty and sublimity for Schopenhauer resides in the dif-
fering relation of the Ideas embodied to the human will, and a concomitant
difference in the mode of engagement with the object that embodies them.
362 History

What distinguishes the feeling of the sublime from that of the beautiful is
that, with the beautiful, pure knowledge has gained the upper hand without
a struggle (WWR, i. 202). In the case of the beautiful, the Ideas presented by
an object are either agreeable or else neutral with respect to human nature,
thus enabling the shift from interested perception to pure contemplation to
occur passively, with purely pleasurable upshot. In the case of the sublime, the
Ideas presented by an object are inimical or threatening to the human mode of
being, and so a shift to contemplation of them occurs only with the active par-
ticipation of the subject, in a partly willful manner, resulting in pleasure with
a painful undercurrent. Contemplation in the case of the sublime requires a
forcible disengagement of the will, consciously and effortfully maintained, a
free exaltation beyond the relations of the object recognized as unfavorable to
human existence. The subject experiencing the sublime focuses on the fearful
aspect of the object while at the same time inhibiting the practical responses
the will would ordinarily have toward such hostile forces. Such a subject is
aware of the antagonistic relationship of the object, e.g. a maelstrom or thun-
derstorm, to human existence generally, but suspends, through an effort of
will, the sense of threat to his personal well-being. The reward is the peculiarly
mixed exhilaration known as the sublime.
Schopenhauer provides a categorization of the different arts, according to
the grade of Ideas embodied and the quality of cognition thus afforded.
On this scheme the arts run from the lowest, architecturewhose objects
manifest primarily the simple Ideas of gravity and rigiditythrough
landscape painting, animal painting, sculpture, historical paintingwhose
objects manifest, respectively, Ideas of vegetal nature, animal life, human
body, and human characterto the highest, tragedy, for Schopenhauer a
species of dramatic poetry. Tragedy deals with the conflict of human wills
at the highest level, epitomizing the inevitability of suffering, the futility of
aspiration, and the inexorability of fate, and thus ultimately teachingwhat
Schopenhauer endorses in his ethicsresignation and denial of the will.
The only art with no place in this scheme is music, and yet for Schopen-
hauer it is perhaps the greatest art of all. It lacks a place in that scheme because,
as Schopenhauer was well aware, being non-representational it presents for
contemplation no Ideas, no perceivable objectifications of willing, and thus
seemingly provides no occasion for the transformation of the individual into a
momentarily will-less pure subject of knowledge.
Two signs, for Schopenhauer, that music has an especially intimate relation
with the deep nature of things are, first, that its effect on us is so profound,
 Schopenhauers Aesthetics 363

and second, that it is immediately understood by all. Thus, music doesnt

copy or present Ideas, concludes Schopenhauer, but rather Will itself. Music,
in all its forms, is willing made audible, in all its inner variety. Music and the
phenomenal world are in fact parallelcomplete, though different, expres-
sions of the nature of Will. Confirmation of this is the fact that structural
similarities between the two abound, and Schopenhauer is quite resourceful
in suggesting analogies between aspects of the natural world and such music-
al features as melody, bass, fixed scale positions, major and minor modes,
cadences, modulations, and the impossibility of equal temperament.
There may appear to be a difficulty in the thesis that music is a complete
alternative expression of Will, namely that of musics itself being phenom-
enal, that is, bunches of sounds or sound waves. What this suggests is that
it is music as repeatable pattern or succession, rather than as concrete event,
that is here meant, or alternatively, sticking with music as concrete sonic phe-
nomenon, that such a phenomenon is to be regarded as a microcosm of Will,
in contrast with the phenomenal world as a whole, inclusive of music, taken as
a macrocosm of Will.
Whether or not Schopenhauers grounds for postulating this parallelism
between music and world are fully adequate, there is a more pressing prob-
lem for his philosophy of music, one we may label the paradox of musics
appeal. Music confronts a listener most directly with the awful inner nature
of the world, being in effect a direct copy of the cosmic Will, the source of uni-
versal suffering, while at the same time offering no Ideas with which to engage
objective contemplation and thus afford the subject momentary relief from
willing. How, then, can music be even tolerable to us, much less immensely
appealing? How can music gratify us, if what it centrally offers is the unfiltered
spectacle of the root of all evil?
Schopenhauer provides a number of hints as to how this paradox might
be resolved, and others can be offered on his behalf. First, although music
by hypothesis confronts one directly with willing, the bane of existence, in a
pure form, it is will-in-general, rather than some particular manifestation of
willing, which thus serves to divert attention from ones own situation and
its incessant demands; contemplation of the universal will thus puts ones
individual will in abeyance. Even though in attending to music the subject
is not presented with any Ideas, disengaging practical concern in virtue of
their utter unrelatedness to the subject in spatiotemporal terms, there super-
venes a similar attitude of effectively selfless absorption in an imagethat
of Will itselfwhich has nothing to do with the phenomenally situated self
364 History

and its materially tethered needs and desires. Second, music is still a repres-
entation of willingeven of the most immediate sortrather than willing
itself, and is furthermore divested of the particulars, spatial and causal, that
characterize any real instance of willing; so emotions, even violent ones, may
be reflected in musical flow, but being stripped of their concrete motivations
and targets, they are not found personally distressing. Third, there is cognit-
ive satisfaction in knowing will more completely by confronting it in its most
transparent manifestation, that of music. Fourth, there is in contemplating
music a kind of elation in grasping ones ultimate identity with and nonsep-
arateness from the world, if in fact everything is at base the same blind will
or life energy. Finally, it can be suggested that music models the vicissitudes
of willing in ways that, however misleadingly, give impressions from time to
time of purposefulness, rightness, or closure, which thus provide real if tran-
sient satisfactions, offsetting the distress that the naked image of Will might
induce.
Schopenhauers general view of the arts other than music as vehicles for
contemplation of Ideas, affording temporary release from willful strife, is
certainly not above criticism. First, much art seems very much concerned
with unique particulars, with getting us to relish distinctive features of the
concretely real, rather than allowing us to break free of such particularity.
Second, much art appears designed to engage, rather than detach, our
passionate or willing natures, even if still preserving some distance between
art and life. Third, much art appears aimed at an active consumer, rather
than one in whom the will is to be passively neutralized, through the
presentation of objects supremely uninteresting to it, i.e. ones that can only
be contemplated. Finally, much art seems capable of providing a portion of
positive, outgoing, and unadulterated pleasure much greater than the view
is able to accommodate, with its emphasis on the negative, crabbed pleasure
of relief, however seconded by satisfactions of a cognitive sort. Underlying
this, of course, is Schopenhauers metaphysically driven overestimation of the
degree to which people are indeed the suffering slaves of their willing natures,
awash in a sea of dissatisfactions broken up only by islands of boredom.
Some of the ideas in Schopenhauers aesthetics which are of lasting import-
ance, ones largely detachable from their metaphysical moorings, are these:
aesthetic attention as in principle capable of being brought to anything; art as
a means of transcending the self and overcoming the narrow bounds of indi-
viduality; arts reward as lying partly in the cognitive insight it affords into the
nature of things; the artists essential power, that of heightened perception,
 Schopenhauers Aesthetics 365

as continuous with that of appreciators of art; an artforms value as bearing a

strong relation to its range of concerns or subject matter; and music as funda-
mentally different from, and more immediately affecting than, the other arts.

IV BIBLIOGRAPHY

Works by Schopenhauer
The World as Will and Representation, 2 vols., trans. E. F. J. Payne (Colorado: Fal-
cons Wing Press, 1958; paperback, New York: Dover, 1969). [Schopenhauers
main work. All references in the text are to this edition.]
The Fourfold Root of the Principle of Sufficient Reason, trans. E. F. J. Payne (Illinois:
Open Court, 1974). [A version of Schopenhauers doctoral dissertation, which
sets out the epistemological basis of his metaphysical system.]
Parerga and Paralipomena, 2 vols., trans. E. F. J. Payne (Oxford: Oxford University
Press, 1974). [Additional essays on various subjects.]

Works about Schopenhauer and his Aesthetics

Alperson, Philip, Schopenhauer and Musical Revelation, Journal of Aesthetics and
Art Criticism 40 (1982): 15566.
Budd, Malcolm, Music and the Emotions (London: Routledge, 1986), ch. 5.
Diffey, Terry, Schopenhauers Account of Aesthetic Experience, British Journal of
Aesthetics 30 (1990): 13242.
Hamlyn, David, Schopenhauer (London: Routledge, 1980).
Jacquette, Dale (ed.), Schopenhauer, Philosophy, and the Arts (Cambridge: Cambridge
University Press, 1996).
Janaway, Christopher, Self and World in Schopenhauers Philosophy (Oxford: Oxford
University Press, 1989).
Janaway, Christopher (ed.), The Cambridge Companion to Schopenhauer (Cambridge:
Cambridge University Press, 1999).
Knox, Israel, Schopenhauers Aesthetic Theory, in M. Fox (ed.), Schopenhauer: His
Philosophical Achievement (Brighton: Harvester Press, 1980).
Krukowski, Lucian, Aesthetic Legacies (Philadelphia: Temple University Press, 1992),
ch. 3.
Magee, Bryan, The Philosophy of Schopenhauer (Oxford: Oxford University Press,
1983).
Young, Julian, Willing and Unwilling: A Study in the Philosophy of Arthur Schopen-
hauer (Dordrecht: Martinus Nijhoff, 1987).
, Schopenhauer (London: Routledge, 2005).
 22
Humes Standard of Taste: The Real
Problem

Arguing with pleasure is a mugs game. If people say that they are having
good sex, you can hardly tell them that they should give up lovemaking
for sunsets. You can only tell audiences satisfied by Mission: Impossible
or Men in Black that there are pleasures they are not experiencing, and
then try to say what those pleasures are.

I INT RODUCT ION

1. Although Kants Critique of Judgment (1790), and especially the Analyt-

ic of the Beautiful, has long been a favored text of aestheticians of all stripes,
lately Humes Of the Standard of Taste (1757) has become perhaps even more
a la mode, at least among anglophone aestheticians. In the last twenty years or
so a large number of studies, by writers such as Peter Kivy, Carolyn Korsmey-
er, Noel Carroll, Ted Cohen, Malcolm Budd, Anthony Savile, Roger Shiner,
Nick Zangwill, James Shelley, Peter Railton, and Mary Mothersill, among
others, have been devoted to explicating and commenting on Humes celeb-
rated essay. Though all these authors have, in one way or another, cast light

First published in Journal of Aesthetics and Art Criticism 60 (2002): 22738.

David Denby, The Moviegoers: Why Dont People Love the Right Movies Anymore?, The
New Yorker, April 6, 1998, 98.
See Peter Kivy, Humes Standard: Breaking the Circle, British Journal of Aesthetics 7 (1967):
5766; Carolyn Korsmeyer, Hume and the Foundations of Taste, Journal of Aesthetics and Art
Criticism 35 (1976): 20115; Peter Jones, Humes Aesthetics Reassessed, Philosophical Quarterly
26 (1976): 4862; Jeffrey Wieand, Humes Two Standards of Taste, Philosophical Quarterly 34
(1983): 12942; Noel Carroll, Humes Standard of Taste, Journal of Aesthetics and Art Criticism
43 (1984): 18194; Richard Shusterman, Of the Scandal of Taste: Social Privilege as Nature
in the Aesthetic Theories of Hume and Kant, Philosophical Forum 20 (1989): 21129; Mary
Mothersill, Hume and the Paradox of Taste, in George Dickie, Richard Sclafani, and Richard
 Humes Standard of Taste 367

on the issues at stake, in my opinion it is Mothersill who comes closest to

putting her finger on what I will call the real problem raised by Humes solu-
tion to the problem of taste. I will return in due course to Mothersills reading
of Hume, indicating how this helps with the real problem, but also where it
falls short. I will then propose, following Mothersills lead, what I hope can be
seen as an adequate answer to the real problem.
2. What, then, is the problem of taste as addressed by Hume? I can be
brief, since the work of my predecessors on this terrain has served to make
Humes problematic in Of the Standard of Taste widely known. Hume is seek-
ing a principle to which disputes about taste, understood as judgments about
the relative beauty or artistic worth of works of art, can be referred so as to
settle such disputes, pronouncing one judgment correct and others incorrect.
Hume observes that even though we give casual allegiance to the laissez-faire
Latin dictum, de gustibus non est disputandum, or its French equivalent, chacun
a son gout, we are at the same time conscious that there are cases of glaring,
undeniable differences in beauty or artistic worth, for instance, as between
Proust and John Grisham, or Schubert and Barry Manilow, or Cezanne and
Julian Schnabel, or Picasso and Cy Twombly, which seem to support the idea
that there is, after all, a right and a wrong in such matters.

Roblin (eds.), Aesthetics: A Critical Anthology, 2nd edn. (New York: St. Martins, 1989), 26986;
Anthony Savile, Kantian Aesthetics Pursued (Edinburgh: Edinburgh University Press, 1993), ch.
4; Ted Cohen, Partial Enchantments of the Quixote Story in Humes Essay on Taste, in Robert
Yanal (ed.), Institutions of Art (University Park, PA: Pennsylvania State University Press, 1994),
14556; Nick Zangwill, Hume, Taste and Teleology, Philosophical Papers 23 (1994): 118; Ted
Gracyk, Rethinking Humes Standard of Taste, Journal of Aesthetics and Art Criticism 52 (1994):
16982; James Shelley, Humes Double Standard of Taste, Journal of Aesthetics and Art Criticism
52 (1994), 43745, and Hume and the Nature of Taste, Journal of Aesthetics and Art Criticism
56 (1998): 2938; Malcolm Budd, Values of Art (London: Penguin, 1995), ch. 1; Alan Goldman,
Aesthetic Value (Boulder, CO: Westview, 1995), ch. 2; Roger Shiner, Hume and the Causal Theory
of Taste, Journal of Aesthetics and Art Criticism 54 (1996): 23749; Peter Railton, Aesthetic Value,
Moral Value, and the Ambitions of Naturalism, in Jerrold Levinson (ed.), Aesthetics and Ethics
(Cambridge: Cambridge University Press, 1998), 59105; Michelle Mason, Moral Prejudice and
Aesthetic Deformity: Rereading Humes Of the Standard of Taste , Journal of Aesthetics and Art
Criticism 59 (2001): 5971.
Though in this essay I adopt, somewhat provocatively, the rhetoric of a real problem
concerning Humes standard of taste, I do not of course mean to suggest that the problems that
other commentators have addressed, which I survey below, are not bona fide ones, or ones that
would not have concerned Hume (and perhaps more than the one I have chosen to highlight).
What I maintain about the problem I provocatively label the real problem is that it is the problem
that any non-ideal, art-interested person should be most worried by when offered Humes solution
to the problem of taste. Let me offer an additional disclaimer. As will be apparent, this essay is not
primarily an exercise in historical scholarship, and thus what merit it has will not reside in its having
proposed a truest-to-Hume interpretation of Humes essay in light of Humes writings as a whole.
368 History

Hume finds the principle he is seeking, a rule confirming one sentiment,

and condemning another, in what he calls the joint verdict of true judges.
Analogizing perception of beauty in works of art to perception of sensory
qualities, Hume proposes that the true assessment of such beauty is formed
by perceivers who are best fitted to receive the beauty sentiment from beau-
tiful works, that is to say, perceivers who have to the greatest extent possible
removed obstacles or impediments in themselves to the production of the
beauty sentiment, which Hume qualifies as inherently pleasurable or agree-
able, by works that, as Hume views it, are naturally fitted to raise this sen-
timent in human beings. Such perceivers are Humes true judges, and the
works they prefer, ones naturally fitted to afford us substantial beauty reac-
tions, are truly more beautiful than others. Such judges are invariably more
gratified or rewarded by Proust, Schubert, Cezanne, and Picasso, than by
Grisham, Manilow, Schnabel, and Twombly, and this shows us that the
works of the former are more beautiful than, or artistically superior to, those
of the latter.
Hume identifies five obstacles or impediments to optimal appreciation,
whose complete overcoming yields a true judge: insufficient fineness of
discrimination, insufficient practice with works of a given sort, insufficient
comparative appreciation of works, insufficient application of meansends
reasoning in assessing works, and finally, prejudice, especially such as prevents
one from entering into the spirit of a work on its own terms. Put positively,
then, the standard of taste is embodied in perceivers of this optimal kind, free
of impediments to the proper operation of the beauty faculty: Strong sense,
united to delicate sentiment, improved by practice, perfected by comparison,
and cleared of all prejudice, can alone entitle critics to this valuable character;
and the joint verdict of such, wherever they are to be found, is the true
standard of taste. With this standard in hand, claims Hume, we are in a
position to do two things: one, identify those works of art that are truly
beautiful, for they are those preferred by and most gratifying to true judges;
and two, assess individual judgments of artistic beauty for relative correctness,
by seeing how closely they approximate those of true judges.

Humes crucial formulations on this point are these: Some particular forms or qualities, from
the original structure of the internal fabric, are calculated to please, and others to displease (Of the
Standard of Taste, 259) and some objects, by the structure of the mind, be naturally calculated to
give pleasure (ibid. 260). [Citations are to a reprinting of Humes essay in Alex Neill and Aaron
Ridley (eds.), The Philosophy of Art: Readings Ancient and Modern (New York: McGraw-Hill, 1995).]
Of the Standard of Taste, 264.
 Humes Standard of Taste 369

3. So for Hume the beautiful or artistically good artwork is one preferred,

enjoyed, approved, and recommended by true judges. I will not be concerned
with whether the joint verdict of true judges is best construed as an idealized,
counterfactual ruling, or as the combined opinion of actual, near-ideal crit-
ics. Nor will I be concerned to tease out the differences among preferring,
enjoying, approving, and recommending, which some have suggested Hume
is given to conflating, for even if those attitudes are importantly different,
they are usually convergent, and there is at least a default coupling between
approving and recommending, on the one hand, and enjoying and preferring,
on the other.
Two further difficulties about the shape of Humes full account of the
standard of taste I also leave entirely aside. One such difficulty concerns
Humes relativist concession that there are, after all, different species of true
judge, thus entailing some qualification on the objectivity of judgments of
artistic goodness. Ideal critics, Hume admits, will blamelessly differ in humor
or temperament, and also in cultural outlook. But then given the disposition
to favor works that answer to ones basic personality and that involve customs
with which one is familiar, differences at least in degree of approbation
accorded particular works are to be expected. A second difficulty concerns
the role of the critics moral beliefs in judging art that departs markedly from
those beliefs, and Humes somewhat surprising suggestion that ideal critics
are under no obligation to be flexible in that regard, but may condemn such
works out of hand.
What I must, however, take seriously at the outset of my inquiry is the
question of the logical status of the joint verdict of ideal critics, which Hume
proposes as the standard of taste, vis-a-vis the property of beauty. It is not per-
fectly clear in Humes essay whether he is proposing the convergent approval
of ideal critics as an identifying rule for the beautiful in art or as a conceptual
analysis of the beautiful in art. There is, in other words, some ambiguity in
the notion of a standard of taste. Does the standard of taste function as a
definition of the beautiful, or does it function rather as a principle for resolv-
ing disputes regarding the beautiful? Are true judges the standard in the sense
that being such as to elicit their disinterested pleasure is what beauty is, or is
that just how we tell what is beautiful, by using true judges epistemically, as in
effect aesthetic divining rods or geiger counters?

For an excellent discussion of this issue, see Savile, Kantian Aesthetics Pursued. Savile calls the
two possible readings of the standard of taste in relation to the joint verdict of true judges the
370 History

The latter is arguably the more reasonable interpretation of Humes dis-

course. The true judges are consistently described as reliable detectors of the
beautiful, in virtue of their alleged superior capacities of discrimination and
response, and not as constituters of the beautiful. If so, then beauty itself must
be seen more along the lines of a capacity in things suitably apprehended to
please, in accord with the structure of the internal fabric, to which capacity
the responses of ideal critics testify. So although the approval of ideal critics
is not, for Hume, what beauty amounts to, it serves as the standard of taste
because it is strongly indicative of the presence of beauty.

II THE CONCERNS OF OTHER COMMENTATORS

I now review briefly what others have raised as problems for Humes account.
I am not concerned to assess how tractable or intractable these problems ulti-
mately are, but only to note them and then set them aside for the purposes of
this essay.
Some commentators have charged Humes standard of taste with circular-
ity, on the grounds that certain marks identifying true critics presuppose prior
identification of what is truly beautiful; this might be said, for instance, of
practice and the use of comparisons, since true critics must be experienced
with good works, and must compare a given work with good ones, which evid-
ently presupposes independent identification of good works of art. A related
complaint is that some of the marks, for instance that of good sense, are
not simply descriptive, but rather ineliminably evaluative, thus making the
standard unusable in practice.
Some commentators have found that Humes theory of aesthetic response
to works of art is too causal, mechanistic, and passive, too closely modeled on
taste in the literal or gustatory sense, making it hard to see how there can be
either improvability or normativity in regard to aesthetic response.

constitutive and the evidential readings, and supplies convincing reasons why the latter should be
affirmed.
See Kivy, Humes Standard; Korsmeyer, Hume and the Foundations of Taste; Carroll,
Humes Standard of Taste.
See Savile, Kantian Aesthetics Pursued; Zangwill, Hume, Taste and Teleology; Budd, Values
of Art; Shiner, Hume and the Causal Theory of Taste. This might be called the British objection,
so regularly is it raised by commentators from the United Kingdom.
 Humes Standard of Taste 371

One commentator feels that Humes account does not adequately explain
why actual true judges are needed for consultation, since if the traits of such
judges simply lead to optimal appreciation, anyone can just strive to approx-
imate them in himself. Another commentator feels that Humes account
does not adequately explain why the joint verdict of true judges is what is
required to embody the standard of taste, as opposed to that of a single true
judge.
Some commentators maintain that Humes account is committed to differ-
ent and incompatible standards of taste, ones turning respectively on the joint
verdict of true judges, the rules of good composition, and the canon of mas-
terworks, standards that are potentially in conflict. One commentator, on
the other hand, sees Humes standard as simply the expression of entrenched
bourgeois values, involving the setting up of the taste of those who have been
educated and conditioned in a certain way as somehow more natural than that
of others.
Humes account has been faulted as too optimistic about the likelihood
of convergence among ideal critics, even ones of a given humor and culture,
given that there are arguably many more sources of variation in judgments
among ideal critics of even the same humor and culture, ones rooted in dif-
fering sensibilities or tastes. One other concern is this. The set of traits of an
ideal critic proposed by Hume is arguably significantly incomplete. Emotion-
al receptivity or openness, for example, would seem a plausible addition to the
list, as would serenity of mind or capacity for reflection. Yet those traits, ones
plausibly essential to judging works of art fairly, are not obviously comprised
in the five traits identified by Hume.
But the real problem, I suggest, with Humes proposing the verdicts of
true judges as the standard of taste is none of these. And that problem would
remain even were all the preceding concerns to be allayed. So to it I now turn.
See Carroll, Humes Standard of Taste.
See Cohen, Partial Enchantments. Cohens answers, on Humes behalf, are as follows: (a) one
cannot be sure that any given ostensibly ideal critic is in fact entirely ideal, that is to say, a perfect
instantiation of the five marks; and (b) the irreducible idiosyncrasies of individual true critics are
likely to be ironed out or neutralized in the group verdict.
See Mothersill, Hume and the Paradox of Taste; Shelley, Humes Double Standard of
Taste.
See Shusterman, Of the Scandal of Taste.
See Goldman, Aesthetic Value. For a partial response to these concerns, see my Aesthetic
Properties, Evaluative Force, and Differences of Sensibility, in Emily Brady and Jerrold Levinson
(eds.), Aesthetic Concepts: Essays After Sibley (Oxford: Oxford University Press, 2001), 6180 (and
Ch. 19, this volume).
372 History

III WHAT IS THE REAL PROBLEM?

1. The raison detre of this essay is my sense that virtually all commentators
on Humes essay fail to acknowledge the question that most naturally arises in
the mind of an ordinary, skeptical art-lover in regard to Humes solution to
the problem of taste. Here is one version of it: Why are the works enjoyed
and preferred by ideal critics characterized as Hume characterizes them ones
that I should, all things being equal, aesthetically pursue? Why not, say, the
objects enjoyed and preferred by criticscall them izeal criticswho are
introverted, zany, endomorphic, arrogant, and left-handed? True, you are not
yourself introverted, zany, endomorphic, arrogant, and left-handed. But then
neither do you, by hypothesis, possess the traits of Humes ideal critics. So
why should you care what they like?
It will be remarked immediately that the traits of Humean ideal critics,
in contrast to the traits of izeal critics, are inherently desirable and widely
admired. But that does not in itself show why it will be to your aesthetic bene-
fit to acquire them and to follow up the preferences of perceivers who have
them. It will then be recalled that the traits of an ideal critic, unlike those of an
izeal critic, are not only desirable or admirable in themselves, but enable their
possessors to have superior aesthetic reactions from works with the capacity to
afford them.
But how do we know that? That is, how do we know that the traits of ideal
critics put them in a better position overall to have aesthetic experiences from
works of art? What assures us that those traits, and not others, optimize capa-
city for aesthetic response? What guarantees that the traits of ideal, as opposed
to izeal, critics are aesthetically optimific? To put the question in its most ego-
istic form, why think you will be aesthetically better off if you become ideal,
rather than izeal? True, works will appeal to you that didnt when you were
non-ideal. But then works will appeal to you, were you to become izeal, that
didnt when you were non-izeal.
2. The crucial practical, as opposed to exegetical, question concerning
Humes solution to the problem of taste is why one should care what is truly
beautiful, if one accepts Humes account of how such things are identified,
One commentator other than Mothersill who touches briefly on something like the real
problem is Ted Cohen, though without offering any answer to it: The proto-question is this: In
what sense is the response of a true judge correct? The correlative question, which seems to me to be
the unpleasantly deep and corrosive question, is whether one should be a true judge. Would one be
better to be a true judge? (Cohen, Partial Enchantments, 155).
 Humes Standard of Taste 373

to wit, through the converging verdicts of ideal critics. Why should one be
moved by the fact that such and such things are approved or preferred by ideal
critics, if one is not one oneself? What is special about truly beautiful things,
understood in Humean fashion as those which ideal critics approve or prefer?
Why does it matter what things are truly beautiful, if there are things which
aesthetically gratify you now, but which are, by hypothesis, not among the
truly beautiful?
To these queries it seems fruitless to reply that some objects, the truly beau-
tiful ones favored by ideal critics, are just naturally fitted, from the structure
of the internal fabric, to afford us the beauty pleasure, since it is clear that
other objects, those apparently not truly beautiful, are just as naturally fitted,
from the structure of the internal fabric, to afford the beauty pleasure to you
and your ilk. Why should you switch one set of gratifiers for another? Why
concern yourself with what someone else maintains is artistically better, rather
than what works for you?
Again, why should you care what critics of a given profile prefer, approve,
enjoy, or judge good, if you are of a different profile? Now its true that crit-
ics of a certain profilethey are, say, more discriminating, more practiced,
more given to making comparisons, more adept at assessing ends to means,
less prejudicedprefer works which are thereby, for Hume, truly beautiful.
But what of it? What ultimately rationalizes deference to the counsels of crit-
ics of that stripe? Are you not, it seems, rational to confine your attention to
the class of meautiful works, those gratifying to the group of middling appre-
ciators to which you belong? What is your motivation to become an ideal
critic if you are not? Presumably the ideal critic has no rational motivation
to become you, even though, were he to do so, he would be more in touch
with, and better able to appreciate, the things that you now appreciate, the
meautiful objects. So why this asymmetry? Why, in short, shouldnt everyone
just appreciate what he or she appreciates, and leave it at that?
3. I suggest that a Humean solution to the problem of taste can only
respond to skepticism of this sort by showing that there is something special
about ideal critics understood in a certain way, something about their rela-
tionship to the aesthetic sphere that makes it rational for anyone, or at least
anyone with an antecedent interest in the aesthetic, to attend to the deliv-
erances of and to strive to emulate such critics, and thus something special
as well about the objects identified as truly beautiful through winning the
approbation of a majority of ideal critics. The primary burden of a defender
of a Humean solution to the problem of taste is thus to show in a non-circular,
374 History

non-question-begging way why a person who is not an ideal critic should

rationally seek, so far as possible, to exchange the ensemble of artistic objects
that elicit his or her approval and enjoyment for some other ensemble that is
approved and enjoyed by the sort of person he or she is not. That is, such a
defender must address what I call the real problem about Humes solution.
Why should we think that what ideal critics recommend or prefer really has
more to offer aesthetically than what we already appreciate without their guid-
ance or example? Its not enough to say that ideal critics judge comparatively,
grading things as better or worse in relation to what they have already exper-
ienced. For so do we, yet our rank orderings of the same works, also based
on experience, diverge from theirs. Nor, it seems, is simply listing the other
appreciatively relevant traits of ideal critic any more conclusive. After all, we
each have as specific a set of appreciatively relevant traits, and they suit us,
evidently, to finding satisfaction in other things.
What needs to be explained is why critics of a certain sort are credible
indicators of what works are artistically best, in the sense of ones capable of
affording better, or ultimately preferable, aesthetic experiences. I will suggest
that that can only be done by putting the accent on the special relationship
such critics bear to works of unquestioned value, that is, masterpieces, whose
identification is in turn effected, though defeasibly, by passage of the test of
time.
4. Perhaps, though, we can see why ideal critics credibly serve as indic-
ators of artistically superior works merely by reflecting further on the marks
by which such works are identified. Beautiful works, says Hume, are those
naturally fitted to please us. But if an ordinary person is not pleased by
such works, in what sense are they naturally fitted to please? The answer,
it appears, is that they please if obstacles or impediments to the exercise of
their inherent power to please are removed. So perhaps the defining traits of
an ideal critic are, as Hume suggests, all of them conceptualizable as involving
the removal of such obstacles or impediments, which would help to explain
why they would recommend themselves to us in our search for better aesthetic
experiences.
No doubt some of the marks conform to that conception. The mark which
most obviously conforms would be absence of prejudice, since prejudice often
gets in the way of a works providing us the pleasure it is capable of afford-
ing us. A case might also be made for good sense, understood as the capacity

Of course the ensembles in question may very well be partly overlapping ones.
 Humes Standard of Taste 375

to employ logic or reason where required by the content or form of a work,

without which it will invariably seem less compelling.
But does delicacy of taste or fineness of discrimination conform also to this
conception? In other words, is the power of an object to reward one always
enhanced by the acquisition of greater fineness of discrimination? Perhaps
some works of art affect us more favorably if we do not maximally discrimin-
ate their elements, but instead allow them to make a more holistic impression
on us. And fineness of discrimination might in some cases be an outright
curse, if that entailed perceiving nuances beyond what even the artist would
have been aware of. A person with hyperfine color sensitivity, say, will receive
a distractingly varied color impression from canvases by Barnett Newman or
Ellsworth Kelly, where everyone else, the painter included, sees and is inten-
ded to see fields of uniform, homogeneous color. Consider next the use of
comparisons and being practiced in an artform. Though likely to be generally
advantageous for appreciation, there would seem to be cases where we would
have more rewarding experiences if we were to forgo comparisons or long
practice, reaping instead the benefits of a fresh or unconstrained approach to
the object in question. Perhaps this is true of some recent modes of art, such as
acousmatic music, performance art, and installation art.
But leave the doubts just aired aside. Label the sum total of the five traits
a cultivated taste. The fact is that even if a case can be made that a cultivated
taste is by and large well suited to exploit the ways objects are naturally fitted
to please us, and even that such a taste enables one to ascertain better the true
character of a work of art, the familiar question remains: If one is not now
a cultivated perceiver, why should one care to acquire a cultivated taste and
so be in a position to appreciate what is truly beautiful? Granted, that would
allow one to register the qualities of and be gratified by works that one was
blind to and unmoved by before. But assuming that one is deriving aesthetic
satisfaction from other works, albeit ones by hypothesis not truly beautiful,
and that one is not primarily driven, in ones aesthetic life, by the purely cog-
nitive desire to perceive things correctly, what motivation does one have to
change aesthetic programs, given the real costs of such a change, in terms of
education, training, effort, and the foregone pleasures of what one has already
come to appreciate?

The example of Sancho Panzas kinsmen might also be cited here. It is not clear that wine
tasters such as they have a better or happier oenological life than others who are not quite so sensitive
to a wines chemical composition.
This latter is an example of what economists call opportunity costs.
376 History

I conclude that even if all the traits of Humes ideal critic could be shown to
represent the removal of barriers to natural responses, or even to contribute as
well to the making of more accurate aesthetic assessments, the question would
remain, from the self-interested point of view, of whether an ordinary con-
sumer of arta non-ideal critic, as it wereyet had good reason to engage in
the effort of self-education or self-transformation necessary to appreciate the
works most favored by ideal critics.

IV MOTHERSILL ON HUME

According to Mary Mothersill, Humes essay has in addition to its text a sub-
text, and it is that subtext which, suitably amplified, provides a solution to
the problem of taste. Mothersills interpretation of Humes essay underlines
the tension between Humes official doctrine, invoking rules of composition
imperfectly embodied in ideal critics as the standard of taste, and his unofficial
doctrine, which appeals to great works of art as paradigms of artistic beauty.
The official doctrine, on the essays surface, is that there are rules of com-
position or principles of goodness which operate in the artistic sphere, but
that they are difficult to discern, which is why in disputed cases we have
recourse to the judgments of ideal critics, who have the best insight into what
those rules and principles are and how they interact. The standard of taste is
embodied in the judgments of ideal critics, because they judge in accord with
those rules or principles.
The underlying doctrine, the essays subtext, goes rather as follows: works
standing the test of time, paradigms of excellence in art, constitute the stand-
ard of taste in a given artform; there are no rules of composition with general
application to be found; and true critics are not individuals who have grasped
such non-existent rules, but rather ones who are attuned to greatness in art
and suited to identifying and explicating it for us. Mothersill plainly regards
this subtext, and not Humes ostensible proposal, as what is capable of resolv-
ing the paradox of taste.
Mothersill has the merit of asking, more than any other commentator, how
the various elements of Humes approach to the problem of aesthetic objectiv-
ity, such as the faculty of taste, the rules of composition, the profile of the
true critic, and the canon of great works of art, are best fitted together. She

See Mothersill, Hume and the Paradox of Taste.

 Humes Standard of Taste 377

is, furthermore, absolutely right to foreground the role which unquestioned

exemplars of artistic worth must play in any solution of a Humean sort to the
problem of aesthetic objectivity, if such a solution is to be able to address what
I have called the real problem that such solutions raise.
But Mothersill nevertheless fails to connect all the elements in Humes
account in the most convincing manner, declining to establish, in particular,
a strong link between masterworks as paradigms of beauty, on the one hand,
and the role ideal critics play in guiding aesthetic appreciation and settling
aesthetic disputes, on the other. She fails to optimally integrate the two main
parts of Humes solution to his problem, namely, an appeal to masterworks
that pass the test of time, and an appeal to the preferences or judgments of
ideal critics, rejecting too completely the idea that the converging judgment
of ideal critics can serve as a standard of taste with probative force, even if its
so serving is anchored in prior identification of masterworks as exemplars of
artistic value.
Obviously the masterworks themselves cannot serve directly as yardsticks of
artistic worth, since relevant similarity, say, between a given work and some
masterwork would, in the first place, itself require judgment to estimate, but
in the second place, would be no reliable measure of such worth, for relev-
antly resembling earlier successful work is neither a necessary nor a sufficient
condition of artistic success. By themselves masterworks are even more plainly
impotent to settle disputes or guide appreciation in regard to markedly origin-
al or revolutionary works of art.
In my view, only some form of artistic-value-as-capacity theory, appro-
priately coupled to a canon of masterworks passing the test of time, which is
in turn used to identify ideal critics, who then serve as measuring rods of such
value generally, is adequate to resolving the questions about aesthetic objectiv-
ity which Humes essay so usefully raises. I turn now to fleshing out a theory
of that sort, one that remains Humean in spirit, if not in all particulars.

Mothersill comes closest to what I have called the real problem with Humes solution to his
puzzle when she poses the following question: Given that most of us are dull normals, and lack
these virtues [the traits of true judges], why should we agree that such a character [as true judges
possess] is estimable and valuable? (ibid. 279).
A generalization of the idea, invoked in Section I, that beauty or goodness in art is a matter of
the capacity to give aesthetic pleasure when appropriately apprehended.
Shelley, Humes Double Standard of Taste, also interrogates the normative force of the
standard of taste, but he, unlike Mothersill or myself, is concerned to locate that force in the letter
of Humes essay. Shelley claims that the normative force of the standard, as embodied in the joint
verdict of true judges, resides in the fact that though we share a common nature with them, they
are perceptually superior to us. Their verdicts are thus nothing but the verdicts of our perceptually
378 History

V A RESPONSE TO THE REAL PROBLEM

1. On my proposal as to how to assemble the elements of Humes the-

oryin particular, ideal critics and acknowledged masterworksthere is an
answer to the real problem, an answer which remains elusive on other recon-
structions of Humes solution to the problem of taste.
I make three claims for my response to the real problem bequeathed us by
an account such as Humes. First, it addresses the issue Hume was funda-
mentally concerned with, how to reconcile differing critical opinions about
art and justify greater respect for some rather than others. Second, it assigns a
role to almost all the elements highlighted in Humes discussion of the prob-
lem, if not exactly the same role that Hume appears inclined to assign them.
Third, it offers a plausible general answer to the problem of the objectivity of
judgments about goodness in art, and in such a way that the worry about why
anyone should care what is truly beautiful or artistically better is dispelled, or
at least significantly allayed.
2. There is reason to believe, in reflecting on the nature of ideal critics
understood as identified in a certain way, that works that are approved and
preferred by that sort of perceiver are ones aesthetic best bets, that is, that they
are works most likely to provide aesthetic satisfaction of a high order. Here is
why. Artistically good artworks will be ones that are in some measure compar-
able in their rewards to those masterpieces recognized universally as aesthet-
ically outstanding. Artistically good artworks will thus be works favored and
approved by the sort of perceiver who is capable of appreciating masterworks,
who can thus gauge the extent to which the rewards of such works compare
to those which acknowledged masterpieces can, under the best of conditions,
afford. Such perceivers may be called ideal critics. Now, what characteristics
do such perceivers notably possess, that is, what characteristics do they need
in order to recognize, appreciate, and enjoy to the fullest exemplars of aes-
thetic excellence? Arguably, something like the five which appear on Humes

better selves, which would seem to have an obvious claim on us. Although this is an insightful
reading, and plausible as exegesis of Hume, it is not, I think, enough to lay to rest what I have called
the real problem. For all our old practical and motivational questions simply re-emerge. How do
you know it is in your aesthetic interest to become a perceptually superior appreciator? How do you
know you will then be aesthetically, rather than just cognitively, better off? How do you know that
beautiful works, identified as those preferred by perceptually superior perceivers, are more worth
spending time with? And so on.
 Humes Standard of Taste 379

tally, perhaps supplemented by a few others. So perceivers of that stripe

are a sort of litmus test for good art, art with superior potential to afford
valuable aesthetic experience. Thus, if one is interested in aesthetic experi-
ence at all, one should be interested in what such perceivers recommend to
ones attention.
Now an answer of this sort assumes at least three things that have not
yet been explicitly spelled out. The first is an ensemble of masterworks in
a given genre which are identifiable other than as those works upon which
the approval or preference of ideal critics devolves. The second is a reason
for thinking that masterworks in a given genre truly are pinnacles of artistic
achievement, that is, works possessing an unusual potential to afford aesthetic
satisfaction. The third is a reason for thinking that the considered preferences
of ideal critics are indicative or revelatory of what sorts of experiences really
are better, that is, ultimately more worth having. But these assumptions can, I
think, be made good.
3. I now sketch the overall shape of my answer to the question of why
ordinary perceivers should rationally be concerned to learn of, attend to, and
if possible follow the recommendations of ideal critics, an answer which mar-
shals most if not all of the elements invoked in Humes essay.
(i) The primary artistic value of a work of art, which Hume calls its beauty
or excellence, is plausibly understood in terms of the capacity or potential of
the work, in virtue of its form and content, to afford appreciative experiences
worth having.

For example, I suggested earlier that Humes list of marks of an ideal critic might reasonably
be expanded to include at least emotional responsiveness and reflective capacity. But if we open
the door to expanding the list, can we justifiably exclude any objective virtue of a cognitive or
affective sort, e.g. knowing the calculus, being kind to those in need, being sober and reliable, being
a good listener, etc.? The answer is that we can, from our present vantage point, exclude those,
since although they are admittedly virtues of some sort, they are not ones that have been found
particularly helpful in recognizing and appreciating great works of art in any artform, whereas the
marks that Hume proposes, and others that we might add, presumably have been.
See Monroe Beardsley, The Aesthetic Point of View, in The Aesthetic Point of View (Ithaca:
Cornell University Press, 1982) for a canonical statement of the artistic-value-as-capacity thesis; see
also Budd, Values of Art, ch. 1. Note my qualification of the thesis in the text by the word primary;
artistic value, as it happens, is not wholly a matter of experience-affording capacity. (See my critical
notice of Budds book, Art, Value, and Philosophy, Mind 105 (1996): 66782.) At any rate, a
more elaborate formulation of a qualified artistic-value-as-capacity thesis, tailored to the terms of
the present problem, would be this: to say that X is artistically good is to say, in the main, that X
has the capacity to give aesthetic experiences of significant magnitude when properly apprehended,
the benchmark of such capacity being provided by the aesthetic experience of masterworks, that is,
works that have robustly stood the test of time and that are strongly recommended by ideal critics,
380 History

(ii) Certain works of art, which we can call masterworks, masterpieces,

or chefs duvres, singularly stand the test of time. In other words, they are
notably appreciated across temporal barriers (that is, their appeal is durable)
and cultural barriers (that is, their appeal is wide), and are appreciated on
some level by almost all who engage with them (that is, their appeal is
broad ). It is thus a reasonable supposition that such works have a high
artistic value, or intrinsically-valuable-experience-affording potential, that
value being responsible for their so strikingly passing the test of time. Such
a supposition would be an example of what is sometimes called Inference to
the Best Explanation.
(iii) Though masterworks are thus paradigms of artistic value and incon-
trovertible proof of its existence, masterworks cannot by themselves provide
a standard of taste, that is, an effective criterion of and guide to artistic value
generally. We cannot, say, directly compare a given work of art whose value
is up for assessment with some masterwork in the same medium and judge it
to be of value to the extent that it resembles that masterwork or any other.
Artistically good works of art are good in different ways, especially if they
are innovative or revolutionary, and that is all the more true for artistically
great ones.
(iv) The masterworks, however, can serve as touchstones for identific-
ation of the sort of critic or judge who is a reliable indicator or identifier
of artistic value, that is, intrinsically-worthwhile-experience-affording capa-
cityin its varying degrees. A critic who is able to comprehend and appreci-
ate masterworks in a given medium to their fullest is thus in the best position
to compare the experiences and satisfactions afforded by a given work in that
medium to the sort of experiences and satisfactions that masterworks in the
medium, appropriately apprehended, can provide.
(v) That the experience afforded by masterworks is generally preferred by
such a critic to the experience afforded by other works of art is indicative of its
really being preferable, that is, more worth having, all told. For as John Stuart
Mill famously observed, the best, and possibly the only, evidence of one sat-
isfaction or experience being better than another is the considered, ultimate,
decided preference for the one over the other by those fully acquainted with
and appreciative of both.

ones who excel in deriving aesthetic satisfaction from works of art and in guiding others in their
appreciation.
See his Utilitarianism, ch. 2. Note that I have modified Mills decided preference criterion
slightly in the counterfactual direction by adding the words considered and ultimate. I am aware
 Humes Standard of Taste 381

(vi) Ideal critics, identified as those capable of appreciating to the fullest

masterworks in a given medium, themselves identified by their passing the
test of time, have certain notable characteristics, which underwrite or facilitate
their capacity for optimal appreciation. These characteristics are more or less
those offered by Hume in the essay in his profile of true judges, though that
general profile could reasonably be augmented in a number of respects, and
even more clearly, supplemented by more detailed desiderata defining specif-
ic profiles of ideal critics adequate to particular artforms, genres, or artistic
domains.
(vii) One thus has a reason to attend to the judgments of ideal critics even
if one is not such oneself, since one presumably has an interest in artistic value
understood primarily as aesthetic-experience-affording capacity, and in gain-
ing access to the most rewarding such experiences possible.
4. More concisely, then, the justification for attending to the recom-
mendations of ideal critics that can be constructed from elements in Humes
essay goes like this: ideal critics, that is, ones who show themselves equal to
and inclined toward the appreciation of the greatest works of art, the mas-
terworks, and who possess the cognitive/sensory/emotional/attitudinal traits
that aid in such appreciation, where such masterworks are independently, if
defeasibly, identified by the breadth, width, and durability of their appeal, are
our best barometers of the artistic value of works of art generally. But if artistic
value is centrally understood in terms of intrinsically-rewarding-experience-
affording potential, then the fact that a work X is preferred to another work
Y, all things considered, by a consensus of ideal critics, gives a non-ideal per-
ceiver, one content in his appreciation of Y, which he prefers to X, a reason, if
not a conclusive one, to pursue X, putting himself if possible in a position to
appreciate it better.
So why care what is artistically good, understood as what ideal critics prefer
and recommend? The answer is that there is reason, albeit defeasible, to be-
lieve that what ideal critics, so understood, approve is capable of giving a
satisfaction ultimately more worth having than what one gets from what one
enjoys as a non-ideal perceiver, because of (a) a criterial connection to great
works, through which individuals are recognized as ideal critics, and (b) the
implications of the preference of those who are capable of experiencing both
kinds of satisfaction, that afforded by incontestably great works of art and that

that many are skeptical of Mills criterion, even so modified, but I believe it is fundamentally sound
nevertheless.
382 History

afforded by works that just happen to please one in some measure or other, in
virtue of ones particular background or makeup.
Ideal critics are the best suited to judging the potential of such works be-
cause their artistic tastes and appreciative habits have been honed on and
formed by uncontested masterworks, whose standing the test of time is good,
if defeasible, evidence of their unusual aesthetic potential. Ideal critics are thus
reliable indicators of artistic value in works of art generally, and most import-
antly, those which have not yet stood the test of time.
Great works are ones that stand the test of time, understood in terms of
durability, breadth, and depth of appeal. Ideal critics, those with the sort
of appreciative profile that makes them optimal enjoyers, appreciators, and
explainers of great works, are the best suited to estimating works of art gener-
ally, that is, assessing their aesthetic rewards against the benchmark provided
by the great works. Such critics, in short, are our best truffle pigs as regards
artistic worth.

VI SOME OBJECTIONS CONSIDERED

1. The test of time, it might be said, is an unacceptable yardstick of artist-

ic value, since there are so many other factors, apart from a works inher-
ent potential to reward us, that enter into whether a work will enjoy artistic
longevity. But this observation can be accepted without in any way diminish-
ing the response I have sketched to what I have labeled the real problem of the
standard of taste.
First, the test of time is not proposed as a criterion of artistic value, but
only as an important, yet entirely defeasible indicator thereof. Artistic value
itself, recall, is conceived as potential or capacity to provide aesthetic experi-
ences that are worth having. Second, the test of time is only leaned on in the
defeasibly sufficient direction, not in the defeasibly necessary direction. In
other words, that a work passes the test of time is a strong prima facie reas-
on to think it has significant artistic value, but that it fails the test of time
is only a weak prima facie reason to think it lacks significant artistic value.
Many worthwhile works, we may be sure, have not passed the test of time
for social, political, and economic reasons, while others languish in obscurity
for purely accidental reasons. Their failing the test of time is, so to speak, not
the fault of those works. But passing the test of time, by contrast, is almost
 Humes Standard of Taste 383

always to a works credit. And that is all my solution to the real problem
requires.
2. One might object that the solution sketched shows only why you might
be interested in what ideal critics of your cultural-temperamental sort prefer,
but not why you should be interested, that is, why there is any practical imper-
ative for you to attend to ideal critics insofar as you are rational. In other
words, the objection goes, you might derive benefit from attending to such
critics so, but you equally well might not.
The answer to this objection is simple. The objection underestimates the
prima facie reason for benefit to you that the convergent preference of ideal
critics of your cultural-temperamental sort provides. That convergent prefer-
ence grounds much more than the mere possibility that you will be better off,
offering something much closer to a reasonable likelihood.
Of course that is indeed only a likelihood, not a guarantee. Suppose it turns
out, for example, that one of the traits needed to optimally appreciate art-
works in a given artform is a certain level of verbal facility, or a certain sense of
humor, or a certain capacity for spatial visualization. If those traits are bey-
ond youthat is, if they are ones you cannot feasibly acquirethen the
force of the reason to be interested in what ideal critics of your stripe prefer
is admittedly undercut. What this brings out is the role that an assumption of
shared human faculties plays in the argument. In other words, the conclusion
of the argument, that it is rational to care what ideal critics of your cultural-
temperamental sort prefer, can be understood to have an implicit proviso, to
the effect that you are not in fundamental respects cognitively or affectively
different from such critics. But until you find out that you are, it remains
rational to lend your ear and your mind to their counsels.
3. As we have noted, ideal critics are in a good position to assess wheth-
er experiences with certain works of art are more worth having, all told, than
experiences with others, once they have had both. Fortunately, ideal critics
are also in a reasonably good position to estimate, from their own histor-
ies of aesthetic education, whether the effort or cost of achieving these more
rewarding, ultimately preferable, experiences or interactions was worth the
reward. This is important because it is of course possible that in some cases
the answer will be no, that even though one experience is ultimately preferable
to another, the cognitive, emotional, or physical preparation required in order
to be able to have the first is sufficiently laborious or unpleasant that it is not
384 History

clearly rational to undergo such preparation, rather than that, by hypothesis

less demanding, which is required for the second. Costbenefit considera-
tions have their place, even in aesthetics.
But at this point the following objection might reasonably be lodged.
How does one know that one is not so changed, by acquiring the training
or background necessary to appreciate finer things, that ones comparative
judgments as between different experiences or interactions are not valid for
one as one was before, or for others who remain in the condition that one was
formerly in?
An answer emerges, I think, if we look more closely at the form Mills test
should take as applied to the issue at hand. The criterion of better aesthetic
experiences is basically a matter of whether you would choose to go back to
your former appreciative condition once you had arrived at your present one.
You ask yourself whether you would rather not have had the new aesthetic
experience, in light of the effort you made to obtain it. If the answer is no, that
suggests that the new experience was indeed worth having, and more so than
its predecessor.
Naturally the question you pose to yourself is only answerable by you as you
are now, and so from your present vantage point. But that does not mean it
is without probative force for you as you were before. Undergoing the change
in question was a live option for you at that time, and the knowledge that you
would be glad to have so opted cannot be irrelevant to deciding whether or
not to elect it. It is important, though, that in cases of aesthetic education of
the sort we are considering there would be no hesitation in identifying one-
self, and identifying with oneself, across such a change. That is because the
self-alteration in question is a minor and gradual one, not a radical one such
as would be involved in going from one species to another, or from one per-
sonality type to another, or from a potent to a feeble mental condition or the
reverse, as in some of the more extreme puzzle cases common in discussions of
the intelligibility of intrapersonal utility comparisons over time.
4. A final difficulty. Why, after all, spend any of ones free time with
Shakespeare, Flaubert, Titian, Bergman, or Beethoven, as ideal critics of the
respective forms of art will clearly urge one to do? Why not spend it all, say,

Of course critics engaged in such costbenefit calculations will be subject to the pull of
overvaluing the apprenticeships they have endured, and correspondingly tempted to discount their
downsides, but it is not clear how anyone else could be in a better position to weigh the costs and
benefits of the aesthetic transformations they have undergone.
 Humes Standard of Taste 385

in some combination of windsurfing, motorcycling, parenting, communing

with nature, doing good works, practicing yoga, touring Europe, exploring
Asian cuisine, and learning to master Godels proof? For those are all demon-
strably good things. Whats so special, then, about art?
In a way, this difficulty for my response to the real problem lies outside the
scope of the problem as so far conceived, where it is assumed we are deal-
ing with art-interested persons, and thus ones presumably concerned, to some
extent, to make that part of their lives as rewarding as possible. Yet the ques-
tion, why be an art-interested person at all, given all the other options that
exist for filling a life satisfyingly, is certainly a legitimate one. Though I cannot
hope to answer that question here, I suspect it might be answerable in the con-
text of a general account of intrinsic value, the nature of human lives, and our
considered visions of what we, as human beings, most want to be.

For part of such an account, see my Intrinsic Value and the Notion of a Life, Journal of
Aesthetics and Art Criticism 62 (2004): 31929 (and Ch. 24, this volume).
This page intentionally left blank
 PART V I I
OTH E R M AT TE R S
This page intentionally left blank
 23
The Concept of Humor

I INT RODUCT ION

This essay is concerned primarily with the question of what humor is, and
with the somewhat distinct question, of what the bases or grounds of humor
are. Traditional theories of humor are reviewed, an analysis of humor is pro-
posed, and the issue of goodness in humor is briefly addressed. No distinction
is made between the humorous or funny and the ludicrous, the comical, the
witty, the satiric, the farcical, or the jokey. All are understood here as species of
humor, joined by their common production of amusement.
The three traditional theories of humor are the superiority theory, the
incongruity theory, and the relief theory, to which the names Hobbes,
Schopenhauer, and Freud are often attached as representative proponents.
Each theory is an attempt to capture the essence of the humorous, understood
either as what invariably makes something funny or as what the response to
something as funny necessarily involves. Several influential accounts, such
as Henri Bergsons and Arthur Koestlers, combine features of the three
traditional theories. It is argued below, however, that the essence of humor
is, strictly speaking, given by none of the traditional theories or their variants,
residing instead in a particular pleasurable affect, amusement, elicited through
cognition, and bearing a non-accidental relation to the behavior known as
laughter.

II THE MAIN QUESTION

What is humor, or alternatively, what makes something funny? This funda-

mental query can be interpreted in two ways, as asking either for conceptual

First published as Humour in E. Craig (ed.), The Routledge Encyclopedia of Philosophy (London:
Routledge, 1998), 5627.
390 Other Matters

elucidation or causal explanation. Taken the first way, the question effectively
is: What does it mean to say, or what are the truth conditions of saying, that
a given item is humorous or funny? In other words, how is humorous or
funny to be defined in philosophically rigorous fashion? Taken the second
way, the question becomes: What about a humorous or funny item causes or
is responsible for its being humorous or funny? On the first construal, we seek
to understand exactly when something counts as humorous; on the second,
we seek to learn in virtue of what a given item succeeds in being humorous.
These questions have not been sharply separated in traditional theories, which
sometimes seem concerned to address the one and sometimes the other. How-
ever, whether in the last analysis the questions can be sharply separated is
not entirely clear. What it ultimately comes down to is whether the identi-
fying response to humorousness can itself be identified without specification
of what occasions it, either internally or externally. This identifying response
is usually labeled amusement, but it should be understood that amusement
is here invoked in its specifically humor-related sense, and not in the sense of
entertainment or diversion generally.
Although the fundamental query has been here formulated in the objective
mode, as concerned with what it is for something to be humorous, there is the
perhaps prior question, in the subjective mode, of what it is for someone to
find something humorous. But these are plausibly related roughly as follows:
something is humorous iff it is found humorous by appropriate audiences
under favorable conditions, including cognitive, attitudinal, and emotion-
al ones. Context should make clear under which of these modes the basic
question is being pursued at a given point. Obviously some objectivity about
humorsome degree of true-or-false-ness in humor attributionsis pre-
supposed in considering the question in the objective mode, but this appears
to be justified: items are regularly and sustainedly classified as humorous, and
not just as humorous to a particular person, on a particular occasion. Humor,
though patently a response-dependent phenomenon, seems to have at least as
much objectivity as beauty or virtue.

III TRADITIONAL THEORIES OF HUMOR

There are three main philosophical theories aiming to account for the phe-
nomenon of humor: the incongruity theory, the superiority theory, and the
relief or release theory.
 The Concept of Humor 391

The hallmark of incongruity theory is locating the humorous in some

incongruity presented by or perceived in some item. The humorous item may
be itself incongruous, relative to some assumed other object, or else involve or
contain incongruity. Incongruity, or non-fittingness of items or elements one
to another, has been variously interpreted, and ranges from logical impossib-
ility or paradoxicality, through absurdity and irrelevance, to unexpectedness
and unaccustomedness, to general inappropriateness.
Incongruity theorists include Francis Hutcheson, William Hazlitt, Sren
Kierkegaard, Arthur Schopenhauer, D. H. Monro, Arthur Koestler,
Michael Clark, John Morreall, and perhaps Immanuel Kant. Kant held
the humorous to consist in the sudden transformation of a strained
expectation into nothing. The incongruity here, if any, is between the
expectation, or what it points towards, and its deflation, or what it issues
in. Schopenhauer gave incongruity theory a clearer, perhaps canonical,
formulation: the essence of the ludicrous, he claimed, lies in the incongruity
between concepts, the vehicles of abstract thought, and concrete objects,
apprehended in perception, when the incongruity is grasped suddenly. The
mismatch of thought and perception can manifest itself from either of two
directions: a single concept can be applied to two very different objects, which
only awkwardly encompasses them both (wit), or two objects originally
ranged under a given concept are subsequently realized to be fundamentally
disparate (folly). Koestlers version of incongruity theory, a descendant and
elaboration of Schopenhauers, holds that humor arises from the bisociation
(double association) of an item in respect of two different and incompatible
reference frames or interpretive matrices at once.
Recent incongruity theorists have generally held perception of incongruity
to be the core of the response to something as humorous, but not the whole
of it. Perceived incongruity is taken as necessary, but not sufficient, for the

See his Reflections Upon Laughter (Glasgow, 1750).

See his Lectures on the English Comic Writers (London: George Bell, 1819).
See his Concluding Unscientific Postscript (1846), trans. D. F. Swenson (Princeton: Princeton
University Press, 1941).
See his World as Will and Representation (1819/1844), 2 vols., trans. E. F. J. Payne (New
York: Dover, 1966).
See his Argument of Laughter (Melbourne: Melbourne University Press, 1951).
See his Act of Creation (New York: Macmillan, 1964).
See his Humour and Incongruity, Philosophy 45 (1970): 2032; and Humour, Laughter
and the Structure of Thought, British Journal of Aesthetics 27 (1987): 23845.
See his Taking Laughter Seriously (Albany: State University of New York Press, 1983); and
Enjoying Incongruity, Humor 2 (1989): 118.
See his Critique of Judgment (1790), trans. W. Pluhar (Indianapolis: Hackett, 1987).
392 Other Matters

occurrence of humorous amusement. The reasons are not far to seek. Incon-
gruity is on its face an undesirable property, and its confrontation usually fails
to elicit pleasure straightforwardly; anxiety or bewilderment or, at best, curi-
osity, are more likely results. In addition, not all pleasure taken in incongruity
appears to constitute amusement, as opposed to aesthetic or scientific or other
satisfaction.
What more, then, is required? Some theorists hold that the perceived incon-
gruity must be enjoyed for its own sake, some that it must be enjoyed as
such but not aesthetically, some that it must not give rise to negative emo-
tions (such as fear or disgust), some that it must not engage practical concerns
(as for knowledge or safety), some that it must have a tendency to issue in
laughter. Some stress the temporal structure needed for perceived incongruity
to be found humorous, while others insist that it is not perceived incongru-
ity itself that is the source of amusement, but only the consequent resolution
of such incongruity. Still others underline the fulfillment of background con-
ditions, such as being in fun, or the absence of sympathetic concern for the
object of humor.
Nevertheless, skepticism seems warranted as to whether all intuitively
accountable instances of humor, including instances of mimicry, satire,
sarcasm, slapstick, and sexual ribaldry, turn on the perception of incongruity.
It has not been demonstrated that no property other than incongruity can
sensibly figure as what one is explicitly amused by. Thus, there remains
doubt whether incongruity theory, however qualified or supplemented, can
be correct as a conceptual elucidation of humor.
Superiority theorists, who include Thomas Hobbes, Alexander Bain,
Henri Bergson, and perhaps Roger Scruton, construe humor as rooted
in the subjects awareness of superiority, in some respect, to the humorous
object. Hobbes famously declared humorous laughter to be the result of a
sudden glory in ones eminency or fortune, by contrast to another or to ones
former self. Bergson theorized the comic as essentially the encrustation of
the mechanical on the living, a falling-off from the human ideal of flexibil-
ity, suppleness, and accommodation. The subject accordingly feels superior

See M. W. Martin, Humor and the Aesthetic Enjoyment of Incongruities, British Journal of
Aesthetics 23 (1983): 7484; and Koestler, Act of Creation.
See his Leviathan (1651) London: Penguin, 1982).
See his Emotions and the Will (London: Longmans and Green, 1875).
See his Laughter (originally Le rire: Essai sur la signification du comique, 1899), trans.
C. Brereton and F. Rothwell (Mineola, NY: Dover Publications, 2005).
See his Laughter, Proceedings of the Aristotelian Society Supplement 56 (1982): 197212.
 The Concept of Humor 393

to such, takes pleasure in so feeling, and manifests his pleasure naturally in

laughter at the imperfectly human. Comedy and emotion are held to be
incompatible, comic engagement short-circuiting emotional involvement by
its very nature. In addition, comic laughter for Bergson is a social corrective,
chastising and hopefully reforming the socially undesirable rigid behavior at
which it is directed.
The best-known relief theorists are Herbert Spencer and Sigmund
Freud, who locate the essence of the humorous in the relief from psychic
constraint or release of accumulated mental energy that it occasions. Spencer
felt it important to investigate not only the features of humor, but why it is
specifically laughter that humor induces, thus necessitating a physiological
explanation. The explanation he offers emphasizes nervous tension and its
bodily manifestation when suddenly excessive or redundant. Freuds account
of the pleasure in jokes, influenced by Spencers, is striking and well worked
out, as is his extensive typology of jokes in terms of their structures and
techniques. Freud viewed enjoyment of jokes as rooted in an economy of
psychic energy, namely, that of inhibition or repression. In the case of
innocent jokes, the inhibition is against nonsense and pure play, while with
tendentious jokes, the inhibition is against display of aggression or sexuality,
but in both cases the energy of inhibition thus freed up manifests itself as
pleasure.
Whatever truth they contain, superiority and relief theories lack the gener-
ality of incongruity theory. In addition, they seem more concerned with the
concomitants or mechanisms of the humorous reaction than with its concep-
tual core as such. Thus, these competitors to incongruity theory are currently
seen as even less apt to provide an adequate answer to the basic question about
humor.
The above classification of theorists involved a good deal of
oversimplification, for strands of each of the three guiding intuitions of
reflection on humor can be uncovered in almost every major theorist. Thus,
Kant might as easily be categorized as a relief theorist as an incongruity
theorist, in virtue of the stress he placed on the animation of the body through
quick release of the tension built up in expectation of what does not arrive.
Bergson could with justice be classified as an incongruity theorist rather

See his The Physiology of Laughter (1860), in Essays on Education and Kindred Subjects.
(London: Dent, 1911), 298309.
See his Jokes and their Relation to the Unconscious (1905), trans. J. Strachey (London: Penguin
Books, 1956).
394 Other Matters

than a superiority one, the incongruity posited as definitive of the comic

being that between a human or human-like being and various quintessentially
anti-human automatisms and rigidities with which he or she is afflicted
or trammeled. Schopenhauers account, though obviously foregrounding
incongruity, includes a strain of superiority theory as well; for Schopenhauer,
the phenomenon of humor exemplifies an important truth that we have
independent reasons for acknowledging, namely, the superiority of perceptual
to conceptual modes of knowing the world, and part of our pleasure in humor
is directly in consequence of its affirmation of this truth. Spencers account,
which conceives the humorous reaction as a sudden release of nervous energy,
also posits that this results from a descending incongruity, in a manner
reminiscent of Kant. And Koestlers account, borrowing from Spencer, joins
to the postulation of a collision of incompatible frames or matricesa form
of the incongruity ideathe notion of an emotional mass and its explosive
diversion as laughter when deserted by thought, in the tradition of relief
theories.

IV THE ANALYSIS OF HUMOR

A number of considerations must be borne in mind in formulating an

adequate analysis of humor, by which is meant an answer to the What is
humor? question construed conceptually. Most of these concern the proper
relationship of humor to other phenomena, such as laughter, emotion,
pleasure, and aesthetic appreciation. First, humor and laughter are not
coextensive, i.e. not all laughter, by any means, is occasioned by humor.
Laughter can result from, among other things, tickling, nitrous oxide, organic
disorder, joy, embarrassment, or vengeful exultation. Second, not all humor
is in fact productive of laughter, even in appropriate subjects; humor may
engender amusement without any behavioral manifestation, or with only the
lesser one of smiling. Third, humor does not always produce amusement,
its characteristic pleasure, even in appropriate subjects; certain background
conditions of mood or psychic preparedness need also to be met. Fourth,
humor seems to have both a cognitive and an affective component, which are
somehow bound up together in the response.
It might be thought that, of the theories discussed, incongruity theory is
clearly aimed at the conceptual query regarding humor, with relief theory
clearly aimed at the causal one: noticing incongruity at least appears explic-
ative of what finding something funny consists in, whereas release of tension
 The Concept of Humor 395

seems rather to concern the mechanism whereby finding something funny

generates pleasure or affect. Perceived incongruity is a plausible intentional
object of amusementwhat it is directed onwhereas release of tension is
not.
But what of superiority theory? Though one is surely not amused at a quick
release of nervous energy, it seems not impossible that one might be amused at
ones evident superiority to some unfortunate other, suddenly noted, in addi-
tion to or as opposed to whatever incongruity such misfortune may present.
It is not clear that the pleasure I take in someones accidentally slipping on
a banana peel without serious harm cannot be accounted part of humor-
ous enjoyment as such, but only something distinct. It is not clear that an
items reinforcement of ones favored condition, or its deflation of expect-
ations, or its presentation of ambiguity, or its surprisingness, or its strange-
nessor some other perhaps more specific propertycannot itself be what
is relished, and even the whole of what is relished, in certain cases of amuse-
ment. If so, then it may be a mistake to regard perceived incongruity as con-
ceptually requisite to humorousness.
In light of these considerations, there seem to be two choices for proceeding
with the analysis of humor. On the one hand, if all cases of humor
can be demonstrated to involve perceived incongruity, and to do so non-
accidentally, then perceived incongruity should figure in elucidation of the
notion of humor, with apparent cases of non-incongruity-involving humor
being shown to be either cases of non-humor, or else cases of humor in
which non-humorous pleasure, derived from other sources, overshadows
what humor pleasure proper is present. On the other hand, if apparent cases
of non-incongruity-involving humor be accepted as genuine instances of
humor, then an analysis is needed which elucidates humor without reference
to perceived incongruity. The most promising way of doing so, I suggest, is
in terms of a distinctive and recognizable effect on perceivers, one that arises
through cognition of the item in question. On such a perspective, apparent
incongruity would be only the most common, but not the entirely necessary,
focus of humorousness; other properties might figure as such a focus, on other
occasions. And this, it seems, is as it should be: that incongruity is the almost
invariant basis of funniness should come as a discovery, and not simply fall
out of the mere understanding of what it means to say something is funny.
Suppose, then, believing even incongruity theory inadequate to the
conceptual query because unjustifiably limiting the possible objects of
amusement, one begins with this general idea of the funny: that which makes
396 Other Matters

one laugh through thinking or perceiving it. Then a natural refinement would
be as follows. An item x is humorous or funny iff x has the disposition
to elicit, through mere cognition of it, and not for ulterior reasons, a
certain kind of pleasurable reaction in appropriatethat is, informationally,
attitudinally, and emotionally preparedsubjects generally, where this
pleasurable reaction, amusement, is identified by its own disposition to
induce, at moderate or higher degrees, a further reaction, namely, laughter.
By these lights, the funny cannot be detached from all felt inclination,
however faint, toward the convulsive bodily expression we call laughing. The
propensity of the state of amusement to issue in laughter is what normally
identifies it as such for us, and underlies the widespread intuition that humor
and laughter, though not coextensive, are nevertheless intimately related.
The connection between amusement and laughter, then, would be this: the
mental state of amusement is largely identified by its dispositionuniversal
in humans, if ultimately contingentto issue in laughter when sufficiently
intense.
What of the idea that amusement is not amusement unless it both arises
in a certain way and has a certain intentionality? The present analysis
acknowledges this in its own fashion, for it entails that a reaction to x
is not amusement unless, in addition to being pleasurable and leading
characteristically to laughter, it comes about in virtue of cognition of x, and
is also directed on x. But pace certain theorists, the present analysis denies that
amusement has a formal object as sucha description under which an object
must be seen if it is to amusebeyond the minimal that which is amusing.
What makes x funny?, taken as a question about why speakers count x as
funny, is thus directly answered by the analysis offered. Roughly, x is funny
in that, or because, the cognition of x amuses people with an appropriate
mental set, that is, pleases them in a way that is marked by a felt inclina-
tion, perhaps very mild, to laugh at x. But What makes x funny?, taken as

As is well known, amusement, when not strong enough to issue in laughter, often issues
instead in smiling. (The intimate relationship between laughing and smiling is more transparent
in French than in English: whereas laugh is rire, smile is sourire, or in effect, sub-laugh.) But
the affect of amusement cannot be identified via the disposition to give rise to smiling at the
thought of something, since such a thought, if pleasantfor instance, the thought of a recent
promotionnaturally disposes to smiling, but without amusement necessarily being present.
If, as the above account suggests, the state of amusement is identifiable only through an
associated disposition to give rise to laughter, it would not be surprising that humor, as almost all
theorists of it have noted, requires suddenness in the perception central to its appreciation, since
laughter exhibits a related character of abrupt explosiveness, making it a natural accompaniment of
such a perception.
 The Concept of Humor 397

a question regarding what it is about x that underwrites or contributes to its

being funny or eliciting amusement, can be answered only by empirical invest-
igation or survey. There may be many factors that enter into the explanation of
an items possessing the power of humorousness, though we can be confident
that presenting appearances of incongruity will figure largely among them.
At the present stage of debate, then, it may be useful to recognize two ana-
lyses of humorousness as such, one thinner, one thicker. The thinner analysis,
which we have just been sketching, denies that there is any necessary focus
or intentional object of humor, and holds an items humorousness to be pre-
cisely its power to raise via cognition a certain pleasurable affect, identified
through its connection with laughter, in appropriate subjects. The thicker
analysis assumes that all cases of humor can be shown conceptually to involve
perception of incongruity, and so adds that to the specification of the thought
through which a humorous item must raise pleasurable affect or produce enjoy-
ment. But as we have seen, adopting an analysis of that sort runs the danger
of foreclosing prematurely on the possible objects of amusement, a reaction
which seems characterizable without recourse to a focus on incongruity.

V SOME OBJECTIONS ANSWERED

I briefly review some objections that have been raised to the thinner analysis
of humor offered above, which I propose captures the core of the concept of
humor. I will label that analysis the Affective Theory of Humor, since it loc-
ates the essence of humorousness in somethings power to raise a distinctive
affect, one identified through a felt inclination to laughter.
Objection 1: The Affective Theory of Humor as a conceptual matter
ties humorousness to bodies of a certain form, capable of supporting the
behavior of laughter, since the connection with laughter is made essential
to the affective response, amusement, by which humorousness is identified.
Reply: This is not the case. The connection of amusement and the felt
inclination to laugh is more that of a phenomenon and a reference-fixer to
that phenomenon. It is not essential to the affect of amusement that it
actually dispose creatures to laughter. Rather, amusement is identified, but

These objections were put by Noel Carroll in his article Humour, in J. Levinson (ed.), Oxford
Handbook of Aesthetics (Oxford: Oxford University Press, 2003), 34465.
Carroll, in his critique, labels it the Dispositional Theory of Humor, but I prefer my own
label, since it better conveys what is specific to the account in question.
398 Other Matters

not defined, as that pleasurable affect which, in humans, normally disposes to

laughter. And that leaves it open that amusement might be felt by creatures
physiologically unequipped for laughing, though it might be problematic
whether we could then ever verify that it was amusement they felt.
Objection 2: The Affective Theory of Humor as a conceptual matter ties
humorousness to the propensity to raise sensations or feelings of some sort,
but surely disembodied beings, devoid of physiologies, could possess humor
and be amused by humor. Reply: This seems far from clear. For consider how
reluctant one would be to say that such beings experienced the comicality of a
cartoon, or felt the humor in a situation, or responded to the wit in a joke. The
burden of proof would seem to be on the denier of an affective component
to humor to show that beings could intelligibly be said to be amused without
registering sensations or feelings of any kind.
Objection 3: The Affective Theory of Humor is at odds with the fact that
certain avant-garde artworks, for instance, modernist films of the 1960s, such
as those of Jean-Luc Godard, tend to dispose informed viewers to laughter, in
virtue of the occasions they afford for the knowing detecting of obscure allu-
sions, yet without those viewers being amused by such works or judging such
works to be humorous. Reply: True enough, but these are not cases in which
an inclination to laughter comes about merely through thought of the item in
question. Such inclination arises, rather, from a complex self-conscious cogni-
tion involving acute awareness of others and the desire to signal ones identity
as a connoisseur of advanced art.

VI INCONGRUIT Y RECONSIDERED

Even if, as argued above, incongruity is not a necessary condition or com-

ponent of humorousness, no account of humor can fail to accord it a special
status. Beyond the fact of being the most common focus of humor, its special
status may consist in the following. First, there is reason to think that superior
forms of humorthose which are most satisfying, intellectually and emo-
tionallyall rely on incongruity in one way or another. Second, there may be
categories of humor, for instance, that of jokes, which are unthinkable apart
from incongruity, even if there are categories, e.g. farce, which perhaps have
another basis. Third, the quality of incongruity-based humor may be tied to a
further feature, obviously presupposing such incongruity, namely, the nature
and extent of the resolution of the incongruity that the humor item embodies
or presents.
 The Concept of Humor 399

The pleasure afforded by incongruity-based humor seems characteristically

to require that the apparent incongruity be in some sense resolved by the
subject. Such resolution can be more or less an object of conscious aware-
ness on the subjects part, and can take various forms, including justification,
rationalization, unification, or dissolution, but it is perhaps best understood
as the grasping of the rationale of the incongruity the humor item presents.
The appreciation of incongruity-based humor can be likened to the solution
of a puzzle, though one where insight is attained in a relatively immediate and
effortless way. The resolution of the incongruity presented in a joke is easily
related to, or even identified with, the experience known as getting the joke.
In good incongruity humor, e.g. that of a clever pun, one is made to see the
why of the incongruity involved in addition to merely the what.
The idea of grasping the incongruity in a humor instance might be taken
further, in this direction. To resolve the incongruity in an item of humor, and
thus be in a position to appreciate its humorousness, is to grasp the basis of
the incongruity involved, but at the same time, an aspect of congruity as well,
often residing in the humorous vehicle itself. Without such a double grasp, of
both the fit and the non-fit involved in a piece of humor, amusement of a
high order is unlikely. Good incongruity humor offers the pleasure of finding
connections where none were thought to exist.
We may thus propose that model instances of incongruity humor involve
an underlying unification of their disparate contents, a tying together in the
humor vehicle of incongruous elements, rather than just their brute juxtaposi-
tion. In other words, the best incongruity humor always has a pivot, on which
the humor turns, one that rationalizes the apparently incongruous elements
that have been brought together.

See Tomas Kulka, The Incongruity of Incongruity Theories of Humour, Iyyun 39 (1990):
22335. Kulka argues convincingly that humor derives more from resolution of perceived
incongruity than from perceived incongruity as such. The idea that humor often turns on resolv-
ing incongruities should not be interpreted as a matter of eliminating incongruities (or making
them disappear), or as a matter of dissolving incongruities (or showing them to be illusory). That
resolving incongruities is a matter of grasping them (or clearly perceiving their bases) sounds the
right accent, since it does not, in contrast to the other formulations, do away with incongruities in
the same breath as it acknowledges them.
Monro, Argument of Laughter.
 24
Intrinsic Value and the Notion of a Life

I INT RODUCT ION

What I will investigate here is the notion of intrinsic value. Disputes about
the nature and extent of intrinsic value have been at the very heart of the the-
ory of value. The idea of something that is valuable in itself, or ultimately
valuable, or valuable for its own sake, is an intriguing one, though there has
not been unanimity on exactly how this idea should be elucidated, nor on
what properly falls under it. That the extrinsically valuable can be explained
by reference to the intrinsically valuable has perhaps been a point of general
agreement. But what may intelligibly be considered to have intrinsic value is
not a matter on which there has been substantial accord, nor is there accord
regarding what would count as adequate defense of a given claim of intrinsic
value. I note that the following have all been proposed, sometimes exclusively,
by one philosopher or another, as possessors of intrinsic value: beauty, pleas-
ure, knowledge, health, artworks, persons, nature, thought, truth, friendship,
love, virtue, communication, contemplation, happiness, joy, justice, integ-
rity, aesthetic appreciation, and right action. But it will not be my purpose
to adjudicate among candidates such as these. I am concerned rather with the
general shape of sustainable judgments of intrinsic value.
One important form that the disagreement about intrinsic value takes is
that of an opposition between an object-based view of intrinsic value and an
experience-based view of it. To those who adopt the former sort of view, it is
primarily objects or more generally, portions of the external worldthat
have intrinsic value, that are good in themselves, and whose goodness we are
sometimes fortunate enough to enjoy. To those who adopt the latter view, it is
primarily experiences that have intrinsic value and that are good in themselves,
irrespective of the objects that occasion them. Is it possible to reconcile these

First published in Journal of Aesthetics and Art Criticism 62 (2004): 31929.

 Intrinsic Value and the Notion of a Life 401

two outlooks, and in general, to illuminate what underlies similar disagree-

ments concerning the intrinsically valuable?
I believe so. What I want to explore is the possibility that the notion of
a life provides the mediating link between these contrasting object-centered
and experience-centered views of intrinsic value. For lives comprehend both
objects and experiences, configured and interrelated in various ways. More
specifically, my idea is that sustainable judgments of intrinsic value are those
which, either explicitly or implicitly, take a life as subject. What, we may ask,
most fundamentally, is intrinsically valuable? The answer I return, in general
terms, is this: lives being certain ways.
The reason a suggestion like this helps to resolve the conflict between the
two views referred to above is that the notion of a life is a broader notion
than that of either experiences, on the one hand, or external objects (or more
broadly, states of affairs), on the other. Lives include items of both sorts, e.g.
the feeling of yesterdays warm bath and the really existing bathtub in ones
bathroom, the contemplation of the Velasquez and the physical being of the
canvas in the National Gallery London, the hope of tomorrows tennis victory
and the fact of having won or lost, the suspicion of spousal infidelity and the
persistence of the unjustly suspected faithful spouse.
That a life is a certain way is thus a far-reaching notion. It comprises not
only subjective experiences but objective states of affairs, spreading out to
encompass, at the limit, perhaps the whole world in which a life is situated.
My conjecture is that what is held to be of intrinsic value must ultimately be
groundable in lives being certain ways or having certain characteristics. I am
not concerned in this essay to identify specific ways in which lives can be that
are intrinsically good; my aim, rather, is to simply make plausible the idea that
sustainable attributions of intrinsic value must be to lives, or more precisely,
to lives being certain ways.
A number of puzzling cases concerning value that have concerned philo-
sophers over the years are, at bottom, problems about how to understand
intrinsic value. Pleasure seems to be a good thing, perhaps an unqualifiedly
good thing. Yet the pleasure a wicked seducer or vile sadist derives from his

A lifes being a certain wayroughly, its having a certain very complex propertyis a state of
affairs of the whole life, albeit one closely connected, in complicated fashion, to the myriad states of
affairs that consist in portions of the life being certain other and simpler ways. Furthermore, a life
and a lifes being the way it is, though intimately related, are distinguishable, if only barely: while the
life might be seen as the locus of a viable claim of intrinsic value, it is the lifes being the way that it
is that is, strictly speaking, the subject of such a claim of value. That said, I do not intend in what
follows invariably to observe the very fine distinction between a life and a lifes being the way it is.
402 Other Matters

activity seems, on reflection, a bad thing. Beauty is usually considered an unas-

sailable good, and thus it seems that a world awash in verdant pastures and
first-rate Cezannes is better than one strewn with sputtering slime-pools and
second-rate Utrillos. But if there were no people or sentient beings of any
kind, it is not clear whether there would remain any difference in quality
between the worlds. Happiness is often thought to be the ultimate good, and
thus if two people appear to be equally happy from the insidethat is, if
their experiential streams are qualitatively comparable and equally satisfying
to their ownersthis might be thought to settle how valuable each persons
life is. Yet if one and not the other is, as it turns out, gravely deceived as to his
or her achievements in the world, or in his or her relations to those near and
dear, then the value equivalence of their happy existences is called into doubt.
Seeing lives being certain ways as the central locus of intrinsic value helps
explain our considered reactions in the above three cases. Pleasure as such is
an empty abstraction; any actual pleasure is a particular sort of pleasure taken
by a particular person in a particular thing in particular circumstances. Thus
when we ask whether some pleasure is good in itselfroughly, whether it
ought, for its own sake, to existwe are forced to take into account who is
having it, in what it is being had, and how and why it is being had. But this
is, in effect, to assess the goodness of a slice of life, which invariably opens out
into an assessment of the whole life in which it figures. Beauty is a value, to
be sure, but inasmuch as it is a response-dependent property par excellence, it
is not surprising that out of all connection to sentient beings whose lives can
be enriched through experience of it the value of beauty is probably nil, and
its claim to being an intrinsic value not sustainable. Finally, if we focus on the
life of the undeceived achiever, rather than on her experiential state, that life
can be judged intrinsically better than that of her only-apparently-achieving
doppelganger, because it stands in a truer relation to reality, despite their lives
being characterized by the same degree of felt happiness.
The idea that subjectively equivalent experiences, one of which is in touch
with reality and one of which is not, are not equally valuable, is by now a
familiar one. It is expressed, for example, in Robert Nozicks well-known
experience machine thought-experiment, and on a grander scale, in the
recent film The Matrix. An experience is valuable, we want to hold, not only
in virtue of its qualitative character, but in virtue of its connection to the
world, such as its standing in some sort of true relation to it. For example, the

See his Anarchy, State and Utopia (Oxford: Blackwell, 1974).

 Intrinsic Value and the Notion of a Life 403

thrill of climbing Mt Everest and believing that you have done so is perhaps of
some value, we want to say, but it is vastly more valuable if your experience is
a veridical one, that is, if you have actually done so rather than merely being
wired up to think you have done so. The more general form of this insight,
which goes beyond experiences, is that lives that are the same in qualitative
character, or viewed from the inside, are not necessarily of equal value, viewed
from the outside, or as they really are.
Naturally some skepticism is in order as to what is established by thought-
experiments of the Nozickian sort. What is quite clear is that one prefers
to contemplate the fact that one has really achieved something than the fact
that one has not despite appearing to have done so. But perhaps ones having
really achieved something has no additional value of its own, and we are only
deceived into thinking it does by a sort of Humean slippage, whereby the rosy
glow we feel in contemplating the achievement spills over, illicitly, onto the
achievement itself? To the question, Is feeling you have climbed a mountain
while having done so more valuable, in itself, than feeling you have climbed a
mountain while not actually having done so?, most of us will readily answer
Yes. But does this show that your having climbed the mountain has great-
er value than its only seeming, perfectly and undetectably, as if you had, or
does it show only that your contemplating the feeling of having climbed the
mountain in the belief that your feeling is veridical is more satisfying than
your contemplating the alternative? The difficulty, in short, is to tell when
the results of reflection on a hypothetical situation testify to the qualities of
the situation as opposed to the qualities of the act of reflection. At any rate,
reliance on Nozickian thought-experiments, which I cannot here forgo, pre-
supposes that one can indeed tell. And I think one can, if one carries out the
reflection scrupulously and honestly enough.

II THE IMPORTANCE OF INTRINSIC VALUE

Why, though, should we care about intrinsic value, either what it means, or
what has it, or whether there is any at all? I remind you of some reasons.
First, almost everyone is interested in having a good life, in being happy,
in having compelling reasons to live, and there seems to be an intimate rela-
tionship between those things and somethings having intrinsic value. On the
view defended here, that intrinsic value has something fundamentally to do
with lives, and the sorts of lives that humans enjoy in particular, this is not at
all surprising.
404 Other Matters

Second, intrinsic value looks to be required as a foundation for value gener-

ally. That is to say, it seems reasonable to think that all non-intrinsic value has
its valuableness on loan, or with a promissory note attached, so that if noth-
ing were intrinsically valuable, nothing would be valuable at all. Some have
attempted to evade this conclusion, proposing that value might not ultimately
rest on what is intrinsically valuable, values instead forming a kind of web of
mutual support, with no sort of value serving as foundation of the value of the
whole. While this picture has its attractions, as does a similar one concerning
the structure of knowledge, it is a picture that, like an Impressionist paint-
ing, only looks good at a distance. Viewed close up, I suggest, its attractiveness
dissolves.
If we ask ourselves why we value something, we try to identify what we
value it for, and then we ask why we value that, and so on, until we get to
something regarding which the question no longer makes sense, that is, some-
thing we evidently regard as intrinsically valuable. We are not satisfied, in
other words, when we query our valuing, until we arrive at something that
strikes us as valuable tout court, something we simply regard the world as bet-
ter for containing, something pursuit of which seems self-justifying. Where
values are concerned we need what Wittgenstein called bedrock, something
against which our spade turns. And that seems to mean intrinsic value.
Third, most ethical theories seem to presuppose a conception of that which
is intrinsically valuable, valuable in itself, or valuable for its own sake. Moral
good is often explained in terms of non-moral good, for example, as the dis-
position or propensity to produce such non-moral good. Thus utilitarianism
clearly regards pleasure as intrinsically good, and attempts to construe moral
value in terms of its maximization. Kantian ethics, for its part, views willing
in accord with duty, or treating others as ends, as the foundation of morality,
because that, by its lights, is the only thing that is intrinsically good. And vir-
tue ethics holds that being a certain kind of person, or having certain kinds
of motivations, is the basis of moral worth, presumably because those, too,

There appear to be counterexamples to this claim, baldly stated. For example, could there not
be a world with only two things, x and y, such that the value of x was conditional on the value of y,
and the value of y conditional on the value of x? Then both x and y, it seems, would be valuable,
while neither would be intrinsically valuable. But even if such a world be admitted, something
intrinsically valuable is clearly present in it as well, namely, the complex of x and y, which exists
insofar as x and y do.
This short brief for the inescapability of intrinsic value is not meant to deny that local,
day-to-day justification of action appeals to all sorts of values, and in such fashion as to bring to
a halt, on the local level, further demands for justification. But it is otherwise, I suggest, with the
global justification of action, on the level of plan of life or way of living as a whole.
 Intrinsic Value and the Notion of a Life 405

are regarded as intrinsically good. So ethical theory appears not to be com-

plete without an adequate conception of intrinsic value and how it might be
assessed.

III CONCEPTIONS OF INTRINSIC VALUE

It is time to examine more closely what we mean, or should mean, by saying

that something is intrinsically valuable. The natural interpretation of intrinsic
value is that it is the value a thing has considered by itself, apart from the
circumstances into which it enters, or the wholes to which it belongs. The idea
is that something intrinsically valuable has its value in itself, in virtue of what
it is, and not in virtue of its connections or relations to other things.
However, there are at least two different possible construals of such in
itself value. One construal of in itself value we can label self-contained value:
something has self-contained value if it would be valuable even were there
nothing in the world but it, or if it would be judged good, or something that
ought to exist, entirely on its own. This is more or less equivalent to G. E.
Moores famous isolation test of intrinsic value, first articulated in his Prin-
cipia Ethica and then leaned on in a number of subsequent writings. As Moore
puts it, it is necessary to consider what things are such that, if they existed
by themselves, in absolute isolation, we should yet judge their existence to be
good.
A second construal of in itself value we can label persistent value:
something has persistent value if it remains valuable, and to just the same
degree, regardless of the situation in which it is embedded or the context in
which it is viewed. This conception of intrinsic value is acknowledged, for
example, in the following remark of Christine Korsgaard: Since intrinsically
good things . . . are thought to have their value in themselves, they are thought
to have their goodness in any and all circumstances. Persistent value is thus
roughly the same as unconditional value, or the value something possesses
irrespective of, and unaffected by, situating, framing, or contextualization.
What is the relationship between self-contained value and persistent value?
Is self-contained value perhaps a species of persistent value? Only, it seems, if

See G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903), 187. For
discussion, see Monroe Beardsley, Intrinsic Value, Philosophy and Phenomenological Research 26
(1965): 117.
Christine Korsgaard, Two Distinctions in Goodness, Philosophical Review 92 (1983): 16995
(p. 171).
406 Other Matters

one takes the null context, where a thing is the sole occupant of a world, to be
a kind of context. But it is better to keep the notions of self-contained value
and persistent value separate. Something judged valuable in complete isola-
tion might not be judged valuable when placed in a context. (For its perceived
value might change when viewed in that context.) And something judged
valuable in any context, or irrespective of context, might not be judged valu-
able outside of all contexts. (For its perceived value might depend on some
contrastive relation or other to a surround.)
A difficulty in applying either of the above criteria in practice, however, is
that very little of normative interest is wholly metaphysically detachable from
the context in which it occurs. That is because almost anything we can con-
ceive has part of its being or identity tied up with its relations to other things.
And yet application of the tests for self-contained and persistent value seem
to require precisely such bloodless conceptual surgery, by which an item is
extracted from its setting and putatively grasped as it is in itself.
Consider complex pleasures, or enjoyments, as Moore called them, such
as pleasure in the sinuous and unearthly weirdness of the melodic lines in
Ligetis instrumental music, or pleasure in the career frustrations of the only
modestly talented Madonna. The objects of these pleasures figure in them
essentially, so that they would not be the pleasures they are without having
the objects they have. Such pleasures cannot have self-contained value, since
the supposition that they exist in total isolation, alone in the universe, verges
on incoherence. How can such a pleasure be separated from its surrounding
context without leaving a gaping wound behind? (The problem is akin to
that faced by Shylock in trying to claim his pound of flesh from the body of
the hapless Antonio, the eponymous Merchant of Venice.) And the persistent
value of such pleasures is also at the least unclear, given indeterminacy regard-
ing how such pleasures are to be conceptually detached from their original
contexts and reinserted into new ones. Yet it does seem as if we can, nonethe-
less, make judgments of intrinsic value in such cases. For instance, the first of
the above pleasuresin the music of Ligetiis presumably harmless, and so
likely intrinsically good, while the secondin the setbacks of Madonnais
mildly vicious, and so likely not intrinsically good.
A further difficulty in applying the isolation test of intrinsic value is that the
candidate under consideration is often incompletely and inadequately spe-
cified. For instance, if we ask whether a man feeling unmixed pleasure, con-
sidered entirely by itself, is a good thing, we may justifiably want to know,
before judging: What man? What kind of pleasure? Taken in what object?
 Intrinsic Value and the Notion of a Life 407

Preceded by what? Succeeded by what? And so on. Sustainable judgments

of intrinsic value, it seems, are not judgments of quite general state-of-affairs
type, like someones being pleased, but rather either judgments of particular
states of affairs, like Bill Clintons being pleased at being elected president in
November 1996, or else judgments of narrowly specified state-of-affairs types,
like a man of a certain position and character being pleased at having fairly
and squarely won high office.
Without seeking to minimize the difficulties posed by the metaphysical
undetachability and incomplete specification of candidate items for intrinsic
value, I will assume that we are still usually able to apply the tests for self-
contained and persistent value in at least a rough-and-ready way. I will also
assume that each of the tests of intrinsic value I have sketched captures
something important about intrinsic value, and thus that what is, so to
speak, really intrinsically valuable will exhibit both self-contained value and
persistent value. It is thus that combined criterionof self-containedness and
persistencethat should be understood as in force from here on.
Now one might be tempted, in view of the difficulty of applying the self-
containedness test for intrinsic value, because almost everything has relational
properties that make it what it is, to consider instead relying solely on the
persistent value test. But the workability of the self-containedness test can be
shored up by distinguishing sharply between, on the one hand, relations to
external matters that identify a thing as what it is, and on the other hand,
relations to external matters that are necessary for the thing to continue in
existence once identified as a particular thing. Thus, relations to Velasquez,
Spain, and the history of European painting may contribute to the identi-
fication of The Toilet of Venus and help to make it the particular it is, but there
is a sense in which, having been so constituted, the external matters to which
the painting is related, and to which it owes its identity logically, could fall
away or cease to exist, and yet the painting would continue, in isolation, to
be, with its apparent self-contained value thus coherently assessable. In other
words, to apply the self-containedness test for intrinsic value to a culturally
embedded and implicated entity one grants the relations to outside objects,
persons, and events necessary in order to constitute the thing as a cultural
object, but then imagines everything else subsequently going out of existence,
and then estimates the value of the lonely entity that remains. I submit that
this is an intelligible exercise of thought. However, and unsurprisingly, when
carried out, no cultural object, however splendid in its originating context,
passes the test. Hence no cultural object, even the greatest work of art, has any
408 Other Matters

intrinsic value, though of course it may have an immense instrumental value,

consisting largely in the intrinsic value of the engagements and experiences it
underwrites.
In any event, our inability to completely isolate almost anything of norm-
ative interest from its context because of its metaphysical entanglement with
that context gives us strong reason to think that it is complexes, or things-in-
relation-to-one-another, that are the only plausible candidates for bearers of
intrinsic value. And lives, I want now to suggest, are complexes of just the
right scope to sustain judgments of intrinsic value on the combined concep-
tion I have urged, on which intrinsic value is value that is both persistent and
self-contained. Lives being certain ways turn out to be the sorts of complexes
that, on contemplation, strike us as good in themselves, unaccompanied by
anything else. They are also the sorts of complexes that, having been judged to
be good, do not subsequently reveal themselves not to be good when consider-
ation is widened to include an embedding context.
The judgment that a lifes being a certain way is good is as robust as a
judgment of value can get, one called into question neither by varying the
context in which it is viewed nor by being viewed out of all contexts whatso-
ever. By contrast, the problem with anything narrower than a lifesuch as an
object, an experience, a pleasureserving as a locus of intrinsic value is that
the perceived values of all such things are liable to overturning once we con-
nect them to their situational contexts. But a life is arguably something broad
enough to include everything that could make a difference, such as related
lives being certain ways, relevant states of the world being as they are, relevant
history being as it is, constitutive experiences exhibiting a certain character,
and so on.

IV INTRINSIC VALUE AND FINAL VALUE

Another intuition we have regarding intrinsic value, apart from its being value
that something has in itself, seems to be that it is something worthy of valuing
absolutely, for its own sake, or as an end. That is to say, it looks as if there
is some kind of connection between a things having intrinsic value and a
things being something one should ultimately pursue, endorse, or endeavor

For further discussion of the intrinsic and instrumental values of artworks, see my Art, Value,
and Philosophy, critical notice of Malcolm Budd, Values of Art, Mind 105 (1996): 66782.
Whether a life is in fact simply coextensive with the world in which it occurs is something I
address in Section VII.
 Intrinsic Value and the Notion of a Life 409

to realize. Christine Korsgaard, though, in a much-discussed paper, has argued

vigorously that end or final value should not be assimilated to intrinsic value,
on the grounds that intrinsic value is a matter of the basis or locus of the value in
question, its residing in or depending entirely on the thing itself, whereas end
or final value is a matter not of a values basis or locus in a thing but of how the
thing is being valued, or the manner in which it is being valued. That may well
be. However, the suggestion made a moment ago was not that intrinsic value
and final value were precisely the same thing, but rather that things possessing
intrinsic value were things worthy, and preeminently so, of being valued finally
or as ends. Whether they are perhaps the only things so worthy I will leave open.
My suggestion is that intrinsic value and final value, though not identical, are
more closely related than Korsgaard is inclined to allow.
Its true that what may sensibly be valued as an end may not have intrins-
ic valuefor we may value as an end something that does not have value in
itself, but only in relation to other things, as perhaps when we value a keep-
sake. But if one says that the keepsake is valued as an end, though not regarded
as intrinsically valuable, valuable only for its associations or history, then in
what sense, really, is one valuing it as an end, or for its own sake? Isnt one
rather valuing it precisely for the sake of those associations or that history, and
thus evidently not for its own sake? Or is valuing as an end supposed to come
apart from valuing for its own sake? But then what does valuing as an end
amount to?
The mark of valuing as an end is sometimes held to be an attitude of irre-
placeability adopted toward the object valued as an end. Thus it is said that
in valuing a wedding ring, say, in such a way one values precisely it, and not
just something about it that some qualitatively indistinguishable object might
provide. But if what one values about it is, say, the objects ability to link one
to a loved one, root one in the past, or anchor ones identity, its not surpris-
ing that an attitude of irreplaceability should be adopted toward the object
in question, since evidently only that object can secure the benefit just articu-
lated. That is, nothing without its unique historical properties can do so. And
that benefitconnection to the past, anchoring of personal identity, linkage
to a loved oneis arguably something that goes to making a life intrinsically
more valuable. Thus, once more, with only a little probing under the surface,
something valued as an end yet not regarded as intrinsically valuable ends up
disclosing something that can plausibly be regarded as of intrinsic value.
Cases where something without intrinsic value is sensibly valued as an end,
I suspect, are all cases where something intrinsically valuable is, so to speak, in
410 Other Matters

the wings. In the case just discussed, what is intrinsically valuable is the preser-
vation of certain life-enhancing memories or connections to ones pastor
more exactly, ones lifes having a certain character as a resultsomething
which retention of the keepsake helps to effect.
Perhaps good health can serve as another example here. Good health might
reasonably be valued as an end, yet be held not to be intrinsically valuable,
because it is only valuable in conjunction with acceptable conditions of life.
Thus if one were being tortured, with no hope of escape, one might wish for a
weak heart, with the prospect it offered of a fatal heart attack, which would
bring ones suffering to a close and provide deliverance. Still, this doesnt
so much show that different things are picked out by the final and intrinsic
value detectors as that there just needs to be an adjustment of scope to get the
right object of value into view: what is perhaps intrinsically valuable, and also
perhaps reasonably valued as an end, is good-health-provided-acceptable-life-
conditions.
If what one claims to value as an end does not seem to be intrinsically valu-
able, most likely something has been left out of the description of what one
so values. If X is the ostensible object of valuing as an end despite not seem-
ing to possess intrinsic value, most likely there is some Y such that it is really
X-in-relation-to-Y, or X-given-Y, or X-in-conditions-Y, that is the true object
of valuing as an end. Thus, even if something not intrinsically valuable can
sensibly be valued as an end, it seems there is always a fuller description of
that something that one must be prepared to give in order to make sense
of the final valuing in question. And that fuller description of the object of
final valuing, I suggest, will generally point to some larger whole, a complex of
which the object is a part, that is arguably of intrinsic value.
In any event, if we reflect on what should be valued as an end, if push comes
to shove, the natural answer is: that which we take to have intrinsic value.
For only then, it seems, are we justified in valuing it, as we say, for its own
sake. In other words, valuing as an end something that one regards as intrins-
ically valuable seems self-legitimating, whereas valuing as an end something

Of course, even that may not be intrinsically valuable. That is to say, even good-health-under-
acceptable-life-conditions may only be instrumentally valuable, one might argue, because of the
intrinsically valuable activities, experiences, or achievements it makes possible. But the point in the
present context is just that such a good is rightly seen as finally valuable only in the measure it is
seen as intrinsically valuable.
 Intrinsic Value and the Notion of a Life 411

that one recognizes not to be intrinsically valuable seems odd, and at the least,
in need of special pleading.

V CL ARIFYING THE THESIS

It is time to look more closely at the form a defensible thesis about lives and
intrinsic value should take.
First, what notion of a life is presupposed in the idea of lives as the central
focus of intrinsic value? Human lives? Intelligent lives? Animal lives? Con-
scious lives? The answer, I think, is richly sentient lives, that is, lives involv-
ing, however rudimentarily, awareness, interests, and points of view. These
are, then, not just any lives, such as those of bacteria, or roundworms, or
tulips, but lives of creatures to whom things matter, or who care how things
are. Thus, in the thesis that lives being certain ways are the fundamental bear-
ers of intrinsic value, lives should be understood, if not otherwise indicated,
as richly sentient lives.
Second, is to claim that lives are the only proper subjects of attributions
of intrinsic value an oblique covert way of advancing the substantive claim
to the effect that life is the only intrinsically valuable thing? No. The thesis I
defend is not the claim that all life is intrinsically valuable, nor does the thes-
is defended enjoin a uniformity of attitude toward living things. As regards
richly sentient but non-human creatures, such as dogs, one can sensibly wish
their lives to be a certain way, at least in part for the sake of those creatures
themselves. As regards non-sentient or only minimally sentient organisms,
however, say, trees, flowers, and ladybirds, one cannot justifiably wish their
lives to be a certain way for their own sakes, but only for the possibilities of
experience by richly sentient creatures that their existence affords.
Note that what I have just affirmed is that the notion of intrinsic value lacks
purchase except in relation to the lives of richly sentient creatures. But I do
not make the same claim for value simpliciter, for that notion may have pur-
chase even for non-sentient lives, e.g. those of plants, which arguably have
needs, which arguably implicates an idea of flourishing, which is an evaluat-
ive notion. So value may be intelligible in relation to life of any kind, even if
intrinsic value, which seems to presuppose a life in which things matter to the
subject of the life, is not.
412 Other Matters

VI THE VALUE OF EXPERIENCES IS NOT INTRINSIC

VALUE

We need now to get clearer on why it is lives rather than experiences

that are the proper subjects of attributions of intrinsic value. Why arent
experiencesat least when understood widely, as including characterization
of the haver and the external objects of the experiencealso loci of viable
claims of intrinsic value? For example, a good man taking pleasure in a good
piece of music, or an honest researcher synthesizing a new compound of
cobalt? Are these experiences not intrinsically valuable? And if so, cannot such
experiences, and not only the lives in which they figure, count as bearers of
intrinsic value? Here are two reasons for a negative answer.
First, when seen in the context of the life in which they take place, such
experiences may strike us as of lesser, or even no, value. Suppose that the
former experience mentioned abovethe musical oneoccurs right after
the mans wife has died; we may feel it is better if someone does not take
pleasure in anything at such a time. More generally, the reason why experi-
ences taken out of life context cannot be seen to have a value unconditionally
is that the perceived value of an experience depends on how it is situated in a
life, that is, on when and why and to whom it occurs, a point that has been
made by a number of contemporary philosophers. This makes it the case
that experiences cannot pass the persistent value test for intrinsic value. Thus,
synthesizing a new compound of cobalt, or discovering a proof of Goldbachs
conjecture or a cure for herpes, will be of differing goodness in a life depend-
ing on when in the life it occurs, e.g. in ones youth or ones primeand
on how it occurs, e.g. with what relation to ones efforts. Second, where
such judgments appear immune to reversal or deflation by a widening of focus
to encompass a life as a whole, this is invariably because the subject of the
experience has been rather fully specified, e.g. as a good man of a certain age
and marital condition and stage of career and recent accomplishments and
current objectives, and so on and so forth. But clearly, that is tantamount to

See, for instance, C. I. Lewis, An Analysis of Knowledge and Valuation (La Salle, IL: Open
Court, 1946); Michael Slote, Goods and Virtues (Oxford: Oxford University Press, 1985); David
Velleman, Well-Being and Time, Pacific Philosophical Quarterly 72 (1991): 4877.
See, for instance, Slote, Goods and Virtues, and Bernard Williams, Persons, Character and
Morality , in Amelie Rorty (ed.), The Identities of Persons (Oxford: Blackwell, 1976), 197216.
 Intrinsic Value and the Notion of a Life 413

specifying the life the subject of the experience is living, thus in effect making
that life the locus of intrinsic value, and the claim of intrinsic value the claim
that that lifes being a certain way is intrinsically valuable.
The points just recalled about the interrelated values of experiences and
lives can be found, to a surprising extent, in an underappreciated work of
C. I. Lewis, his Analysis of Knowledge and Valuation, a good thirty years or
more before they were articulated by contemporary ethicists. Here are some
representative quotations:
A life good on the whole . . . is something whose goodness or badness is at no moment
immediately disclosed, but can be contemplated only by some imaginative or syn-
thetic envisagement of its on-the-whole quality. (p. 483)
A life that begins badly and ends well is better than one that begins well and ends
badly. (p. 488)
We can quite well imagine that two lives might be made up of constituents separ-
ately comparable in their immediate and momentary values, and yet that one might
be better than the other. (p. 495)
The final evaluation of any particular experience is evaluation of it as contributing
to a whole of experiencethe whole lifewhich it enters as a constituent. (p. 503)

The value of an experience taken by itself, in other words, is subordinate to

the value that it contributes to a broader whole, the life into which it enters
and of which it is a part. The value attaching to a whole life is not the sum of
the values of its parts assessed in isolation, for a life, as both Moore and Lewis
stressed, is an organic unity, a series of internally connected, mutually qualify-
ing episodes, whose value depends not only on the value of the episodes that
make it up, but on their order of occurrence and their relations to one anoth-
er. It is thus a whole lifes being a certain complex way, and not its component
experiences having certain characters, that is the real bearer of intrinsic value.
In no way, furthermore, can experiences, unlike lives, serve as the loci of
all sustainable claims of intrinsic value, even if the difficulties just aired about
regarding them as ever bearers of intrinsic value could be overcome. And that
is because although experiences, widely construed, are a significant part of
a life, they are not the whole of it. As underlined most notably by Thomas
Nagel, not all of a life is experiential. Certain aspects of a lifesuch as the
faithfulness of ones spouse, the state of ones reputation, or ones degree of
privacy in ones homemay lie outside what one has experience or know-
ledge of, and other aspects of ones lifesuch as ones intellectual legacy,
the realization of ones projects, or the flourishing of ones childrenmay
414 Other Matters

outrun ones lifetime entirely. Yet those non-experiential aspects of a life

being a certain wayfor instance, ones good name not being tarnished after
ones death, ones not being spied on by hidden surveillance cameras during
private acts, or ones daughter achieving success after one is gonecontribute
importantly to the intrinsic value of that life. As Nagel puts it, a mans life
includes much that does not take place within the boundaries of his body
and his mind, and what happens to him can include much that does not take
place within the boundaries of his life. The notion of a human life, in other
words, cannot be entirely recuperated as the notion of a series of experiences,
even if such a series forms the core of such a life, because some ways a life is are
not experiential in any sense.
None of this is meant to deny the fact that the value of a life significantly
depends on the character of the temporal parts of the life, most notably, the
experiences it contains. Clearly, if a given life had very different temporal
parts, it would almost inevitably have a quite different value. But though the
value of a life significantly depends on the character of its temporal parts, that
does not mean that the value of the life depends on the value of those com-
ponent parts. For as we have seen, the contextual embedding of those parts,
and not just the values those parts might be held to possess in isolation, is
crucial to the value of the whole they comprise. Which is just to say, once
more, that the value had by the temporal parts of a life, such as experiences,
fails to qualify as intrinsic value. Such parts have at most prima facie intrinsic
value, meaning that they are likely to contribute, in the direction they already
manifest, to the properly intrinsic value of the life as a whole.

VII LIVES AND WORLDS

We must now address a crucial question: Does a specification of a life

adequate to assigning it an intrinsic value in fact implicate the whole world
to which it belongs? I have suggested that a lifes being a certain way W is the
only possible subject of a sustainable judgment of intrinsic value. But it might
be, first, that in order for us to assess adequately the value of a lifes being some
way W, W must be so detailed that specifying it is in effect specifying a whole

An inkling of this insight can be found, again, in C. I. Lewis: A life is bounded, not by the
physical limits of it, but by its horizon (An Analysis of Knowledge and Valuation, 503).
Thomas Nagel, Death, in Mortal Questions (Cambridge: Cambridge University Press,
1979), 6.
 Intrinsic Value and the Notion of a Life 415

life, and second, that specifying a whole life is in effect specifying a whole
world, because of the manifold relations, both spatial and temporal, that lives
bear to their surroundings.
If that is so, it provides an easy route to the conclusion that a life or a lifes
being a certain way is the only possible locus of a sustainable judgment of
intrinsic value. For the values of whole worlds surely pass the double test of
intrinsic value I have adopted, namely self-containedness and persistence: the
value of a world is both self-contained, there being nothing outside a world,
and the value of a world is not susceptible to upset by reframing, there being
nothing more encompassing than a world. But then if a fully specified life
effectively implicates the world to which it belongs, the value of such a life will
clearly be intrinsic as well.
If possible, however, it would be better to show that a life is a substantial
enough entity that even if it is not understood in such a way as to determine
a whole possible world, judgments of value attaching to such lives are suf-
ficiently resistant to overturn if contextualized by filling out the rest of the
world in which they occur, that they may still serve as subjects of sustainable
judgments of intrinsic value.
Think of what a life arguably involves or includes. It enters into your life,
say, that you have a son and that he will attend Harvard Medical School; that
you live in a country founded in the late eighteenth century and colonized
centuries before that; that you speak a language dating at least to Shakespeare;
that your solar system boasts of Saturn and its rings; and so on. There are
innumerable tendrils, in time and space, that connect you to your world, most
of which can be seen to figure in what your life, broadly viewed, is.
It seems that a constructive dilemma might thus be sketched on this matter.
Either fixing a life effectively fixes, in qualitative respects, the whole world to
which it belongs, or it does not. If it does, the value of such a lifeor of the
lifes being the way it iswill clearly be both self-contained and persistent. If
it does not, then given the wide extent of livesthe objects, actions, historical
antecedents, future consequences, relations to others, relations to an environ-
ment, etc., they containwhat would remain to be fixed in a world that is
not fixed by the full elaboration of the life must be so remote, so disconnec-
ted from the life in question, as to have little power to overturn its perceived
value by variant specifications thereof. Thus, once again, the value of such
a life being the way it is would seem to be securely both self-contained and
persistent.
416 Other Matters

VIII THE VALUE OF BEAUT Y NOT INTRINSIC

I come now to the bearing of my thesis on certain matters in aesthetics. Moore

famously claimed that a beautiful world without sentient creatures is, to some
extent, intrinsically more valuable than an ugly world without sentient
creatures. But on what grounds might he have held this? I cannot see any
that do not surreptitiously bring in our contemplation of the existence of
such a world. On the thesis I have defended, in a world with no sentient
creatures, but containing beautiful natural objects, there can be no intrinsic
value, because there are no richly sentient lives and ways such lives are, hence
no intrinsic goodness in such lives being such ways.
However, an undiscovered and unexperienced beautiful planet, or other
natural object, may contribute intrinsic value to a world such as ours, which
does contain richly sentient lives, because it is arguably good that such lives
include the possibility of experiencing that thing, even if its existence is never
discovered and experience of it never comes about. Its existence makes such
lives richer, if only slightly, in the way possibilities do, even when unrecog-
nized. Compare the way in which ones life in a city with good libraries and
cinemas is in virtue of that richerbecause containing more real possibil-
itiesthan life in a city without them, even if one is somehow unaware of
their existence.
If we consider a world with just grass, at first blush that seems intrinsically
better than a world without grass. It may thus seem as if any life is a bearer
of intrinsic value, and not just richly sentient life. But this first intuition, I
suggest, is a corrupt one, in which pleasure in contemplation of the grassy, as
opposed to the barren, sentience-free world has entered into and distorted the
assessment. Again, the visual beauty of the Shenandoah River is not intrins-
ically good, since if there are no lives into which experiences of beauty enter
or can enter, then there is no intrinsic value present. But an undiscovered
and unexperienced Shenandoah River would still contribute to the intrinsic
value of a world in which there were lives it could enrich, and would enrich, if
discovered and experienced.
Consider next a beautiful painting, such as Velasquezs Toilet of Venus,
which on the view proposed here is not intrinsically, but only instrument-
ally, good. A world with just Velasquezs Venuswaiving for the moment
the difficulties noted earlier in rendering that supposition coherentis not
intrinsically better than a world containing nothing. The painting without
 Intrinsic Value and the Notion of a Life 417

anyone to behold it is, perhaps, just as beautiful, even supposing its beauty
to be indexed to human perceptual capacities, but it is not intrinsically good.
Only the organic whole including it and beholders capable of appreciating its
beauty is good in that way, that is, in itself.
The point could be taken further: a painting like Velasquezs Venus fails
to be intrinsically good not only because its goodness requires people who
might experience its beauty, but also, and massively, because its goodness
requires a whole biological and social and historical framework in terms of
which the paintings meaning and appeal are constituted. Still, a world with
human subjects and an undiscovered Velasquezs Venus would, like the world
with human subjects and an undiscovered beautiful planet or an undiscovered
Shenandoah River, be intrinsically better than a world without it. Or so I claim.
Consider finally a great piece of music, such as Brahmss Fourth Symphony.
Though as a music lover of the first order I am tempted to make an exception
of music, and to accord works on the level of Brahmss Fourth Symphony
intrinsic value, out of the immense fondness and respect I have for them,
the truth is, they have none. Even the satisfying experience of such a work
does not, strictly speaking, have intrinsic valuethat is, self-contained and
persistent value. What most likely does have intrinsic value, however, is that
there be lives which contain satisfying experiences of Brahmss symphony and
which are characterized in various further ways which assure that the ostens-
ible goodness of those satisfying experiences is not undermined or undercut.

IX CONCLUDING REMARKS

I have proposed that richly sentient lives being certain ways are the prime, and
very likely the only, subjects of sustainable claims of intrinsic value. I have
defended that proposal mainly by highlighting its superiority to proposals
which accord that role instead to external affairs being of certain sorts or
experiences being of certain characters, and by showing how lives are of just
the right scope to have their ostensible intrinsic values pass the tests of self-
containedness and persistence.
If what I have argued here is correct, we now know what form sustainable
claims of intrinsic value must take. They must say something to the effect
that some richly sentient lifes being a certain way is intrinsically good or bad.
And if what I have argued here is correct we now know what we should ulti-
mately want, unconditionally, and for its own sake. Not that certain objects
418 Other Matters

exist, nor that certain experiences occur, nor that certain impersonal states of
affairs obtain, but that richly sentient lives, our own and those of others, be
certain ways.
Naturally, that is not all we want to know. It is not even what we most want
to know. We most want to know what are the ways that make such claims
true, that is, what are the ways such lives should be. But there is, so far as I
know, no algorithm for that. Figuring it out is, rather, the work of a lifetime.
And not all of that work is philosophical.
 Index

Acconci, Vito 29 Browne, Nick 153 n. 18

Adair, Gilbert 67 n. 21 Bruce, Graham 161 n. 30, 168 n. 36, 169, n.
Alkan, Charles-Valentin 139, 140 40
Allen, Woody 177 Budd, Malcolm 49 n., 94 n. 6, 96, 18695,
Andre, Carl 116 200 n., 2023, 243 n. 5, 338 n., 341 n.
Arcimboldo, 126 11, 367 n. 2, 370 n. 8, 379 n. 23, 408
Aristotle 61, 137 n. 7
Ax, Emmanuel 231
Cage, John 27, 65
Bach, Johann Sebastian 104, 196, 213, 221, Campion, Jane 180
3189 Carlos, Walter 177
Bacon, Francis 27, 65, 255 Caravaggio 256
Bain, Alexander 392 Carroll, Noel 14 n. 4, 1517, 42 n. 4, 52 n.
Baker, Nicholson 2645 16, 130 n. 2, 155 n. 25, 161, 181 n., 308
Baldessari, John 29 n. 9, 310 n. 14, 366 n. 2, 370 n. 7, 371
Ball, Hugo 114 n. 6 n. 9
Balzac, Honore 283 Casati, Roberto 7880
Barnes, Annette 276 n. 4 Chanko, Kenneth 179 n. 47
Barry, Robert 29 Charlton, William 42 n. 6
Bartok, Bela 132 n. 5 Chatman, Seymour 14650
Baxandall, Michael 71 Chicago, Judy 255
Beardsley, Monroe 54 n. 23, 190 n., 1967, Clark, Kenneth 258 n. 5
343 n. 15, 379 n. 23 Clark, Michael 391
Beethoven, Ludwig van 5, 84, 104, 1345, Cohen, Ted 221 n. 5, 289 n., 367 n. 2, 371 n.
141, 192, 210, 213, 214, 2146, 218 10, 372 n.
Bellini, Vincenzo 213 Collingwood, R. G. 36
Bellotto, Bernardo 256 Coltrane, John 64, 69, 70
Bender, John 197 n. 17, 329 n. 30, 330 n. 32, Cometti, Jean-Pierre 13 n.
3325, 340 n. 8 Copland, Aaron 155, 181
Berg, Alban 70 Courbet, Gustave 258, 2689
Berger, John 258 n.5 Crowther, Paul 60 n. 7
Bergson, Henri 389, 392, 393 Currie, Gregory 19, 225, 47 n., 130 n. 3,
Bernini, Gianlorenzo 256 147 n. 6, 308 n. 10, 311 n., 316 n. 5,
Bernstein, Leonard 162 320 n.
Black, Max 296 n. 14, 298 n. 18, 301 n.
25
Bloom, Paul 2, 13 n., 26, 303, 37, 73 Dali, Salvador 255
Boghossian, Paul 1056 Danto, Arthur 28, 275 n.
Bordwell, David 144 n. 1, 146 n. 4, 154 n. Davidson, Donald 6, 28894
Borges, Jorge Luis 2846 Davies, David
Botticelli, Sandro 255 Davies, Stephen 14 n. 5, 1718, 212,
Brahms, Johannes 92, 209, 227, 231, 234, 49 n., 53 n. 25, 86 n. 12, 92 n., 97100,
417 103 n., 185 n., 194 n. 11, 220 n. 4,
Brancusi, Constantin 118 n. 10 225 n.
Branigan, Edward 146 n. 5 Debussy, Claude 192
Bratman, Michael 58 n. 4 Degas, Edgar 255
Bresson, Robert 1745, 1801 Deigh, John 39 n.
Brewster, David 127 De Kooning, Willem 256, 335 n.
Bronzino 268 Delacroix, Eugene 256
420 Index
Demme, Jonathan 189 Hacker, P. M. S. 344 n. 17
Denby, David 366 n. 1 Haines, Victor Yelverton 2 n.4
Dennett, Daniel 3258, 334 Halbreich, Harry 57 n.
De Sousa, Ronald 39 n. Hanlon, Lindley 174
Dewey, John 36 Hawkins, Coleman 70
Dickie, George 28, 54 n. 23, 197 n. 16 Haydn, Franz Joseph 132 n. 5, 226, 234
Dokic, Jerome 7880 Hazlitt, William 391
Donatello 255 Hemingway, Ernest 163 n.
Dostoyevsky, Fyodor 445 Hermeren, Goran 276 n. 4, 315 n. 3, 317 n.
Duchamp, Marcel 28, 35 12
Dubuffet, Jean 256 Herrmann, Bernard 16770
Dvorak, Antonn 210 Higgins, Kathleen 188 n., 191 n. 8 , 221 n. 5
Hills, David 294 n. 10, 297 n. 16
Hirsch, E. D. 294
Eaton, Marcia 53 n. 17 Hitchcock, Alfred 69, 1646, 16770,
Eco, Umberto 275 n. 1712
Eisenmann, Peter 73 Hobbes, Thomas 389, 392
Eisenstein, Sergei 68 Hodges, Johnny 64 n. 14, 70
Eldridge, Richard 200 n., 243 n. 6 Holden, Stephen 180 n.
Ellington, Duke 1845 Holiday, Billie 64 n. 14
Elster, Jon 3, 5674, 139 n. Honegger, Arthur 57 n.
Engstrom, Anders 290 n. 3 Hudson, Hugh 1734
Estes, Richard 265 Hume, David 8, 329, 332, 36685, 403
Hutcheson, Francis 391
Faure, Gabriel 213 Huysmans, J. K. 120 n. 12
Feagin, Susan 53 n. 18
Fellini, Federico 1623, 1701, 181 Ingres, J. A. D. 256, 257, 268
Fodor, Jerry 286 n. 9 Iseminger, Gary 308 n. 11, 309 n. 12
Fowles, John 149 n. 11 Isenberg, Arnold 297 n. 15
Franck, Cesar 319 n. Ives, Charles 132 n. 5
Freud, Lucian 255
Freud, Sigmund 300 n. 23, 389, 393
Jarre, Maurice 1723
Jobim, Antonio Carlos 217
Gaut, Berys 21 n. 22, 53 n. 20, 146 n. 4, 323 Johns, Jasper 28, 113
n. 22 Johnson, Mark 297 n. 16
Gehry, Frank 73 Johnsen, Bredo 325 n. 24
Gerome, Jean-Leon 257 Johnston, Mark 341 n. 10, 342 n. 13, 344 n.
Getz, Stan 5, 217 16, 345, 347 n.
Glass, Philip 116 Johnson, Samuel 68
Godard, Jean-Luc 398 Jones, Peter 366 n. 2
Goldman, Alan 49 n., 321 n. 17, 329 n. 30, Judd, Donald 116
332 n. 36, 340 n. 8, 371 n. 13 Juilliard Quartet, 231
Gombrich, Ernst 1267
Goodman, Nelson 52 n. 15, 66, 91, 222, 235,
283 n. 9 Kafka, Franz 304
Gorbman, Claudia 145 n., 155 n. 24, 182 Kalinak, Kathryn 155 n. 22, 161 n. 29
Gordon, Robert 39 n. Kandinsky, Wassily 27
Goya, Francisco 256 Kant, Immanuel 8, 355, 356, 358, 361, 366,
Gould, Glenn 231 391, 393, 394
Gracyk, Ted 367 n. 2 Karl, Gregory 108 n.
Greenspan, Patricia 39 n. Katz, Alex 265
Grice, Paul 295 Kazan, Elia 162
Griffiths, Paul 40 n.3 Keats, John 299 n. 22
Guarneri Quartet 231 Kieran, Matthew 6, 2647
Gurney, Edmund 121 n. 14, 199, 220 Kierkegaard, Soren 391
n. 2 Kittay, Eva 301 n. 25
 Index 421
Kivy, Peter 49 n., 54 n. 21, 86 n. 12, 98, Moore, G. E. 9, 405, 406, 413, 416
1278, 138, 223, 366 n. 2, 370 n. 7 Moran, Richard 291, 297 n. 17, 300
Klimt, Gustav 265 Morreall, John 43 n., 53 n. 19, 391
Koestler, Arthur 389, 391, 394 Morris, Robert 29
Koolhaus, Rem 73 Mothersill, Mary 366 n. 2, 367, 371 n. 11,
Korsgaard, Christine 405, 409 3767
Korsmeyer, Carolyn 366 n. 2, 370 n. 7 Mozart, Wolfgang Amadeus 68, 812, 132 n.
Kosuth, Joseph 29 5, 193 n., 213, 221
Kozloff, Sarah 155 n. 23 Mulvey, Linda 258 n. 5
Kubrick, Stanley 1779
Kulka, Tomas 399 n. 21 Nagel, Thomas 4134
Kupka, Frantisek 27 Nancarrow, Conlon 196
Nead, Lynda 258 n. 5
Lakoff, George 297 n. 16 Neill, Alex 43 n., 45 n.
Lamarque, Peter 42 n. 5, 278 n. Newcombe, Anthony 4, 134, 1356
Laredo, Ruth 231 Newman, Barnett 67
Leddy, Tom 14 n. 6 Nielsen, Carl 84
Levine, Sherrie 116 Nietzsche, Friedrich 225 n., 227 n. 15
Lewis, C. I. 412 n. 10, 413, 414 n. 12 Nochlin, Linda 257
LeWitt, Sol 29 Novitz, David 291 n. 6
Ligeti, Gyorgy 406 Nozick, Robert 402, 403
Livingston, Paisley 7 n. 10, 306 n. 5, 307 n. 8, Nussbaum, Martha 39 n., 266
309 n. 12, 310 n. 15
Lodge, David 299300
OKeefe, Georgia 256
Lopes, Dominic 242 n., 243 n. 8, 268 n.
Oppy, Graham 14 n. 5, 19 n. 18
Lormand, Eric 325 n. 24
Orff, Carl 1756
Lucie-Smith, Edward 258 n. 5
Lumet, Sidney 69
Lyons, William 39 n. Packer, Mark 52 n. 14
Lysippus, 255 Panksepp, Jaak 5 n., 2226, 2289
Pasnau, Robert 78 n. 4
McGinn, Colin 320 n. Pearlstein, Philip 255
McLaren, Norman 126 n. 21 Perec, Georges 56, 668, 116 n.
Machaut, Guillaume 69 Pettit, Philip 316 n. 5
Madonna 406 Picasso, Pablo 71, 255
Magritte, Rene 255 Pink Floyd 224
Mahler, Gustav 134, 1356, 213 Plato 359, 360
Malick, Terrence 1757 Plotinus 360
Mankoff, Robert 248 n. 18 Pletnev, Mikhail 231
Marais, Marin 179 n. 49 Pollock, Griselda 258 n. 5
Mark, Thomas 68 n. 22 Pollock, Jackson 65
Martin, M. W. 392 n. Poussin, Nicolas 256
Marx, Chico 83 Prokofiev, Sergei 177
Mason, Michelle 367 n. 2 Preminger, Otto 1601
Matravers, Derek 86 n. 12, 94 n. 6, 323, 336, Puccini, Giacomo 301
33840, 345, 346 n. 19 Putnam, Hilary 24
Maus, Fred 4, 108 n., 132 n. 4, 1345, 136,
141 Radford, Colin 46 n., 49 n., 188 n., 221
Mendelssohn, Felix 211 n. 5
Meskin, Aaron 338 n. Racine, Jean 68
Meyer, Leonard 191 n. 8, 198, 199, 230 Raffman, Diana 326
Michelangelo 701 Railton, Peter 316 n. 4, 367 n.2
Mill, John Stuart 380, 384 Raksin, David 1601
Miller, Richard 316 n. 4 Ramos, Mel 255
Monro, D. H. 391, 399 Rauschenberg, Robert 27, 28
Monteverdi, Claudio 1745 Ravel, Maurice 72, 350 n. 27
422 Index
Ray, Nicholas 1612, 1667 Stecker, Robert 14 n. 3, 15, 19 n. 17, 20 n. 20,
Reich, Steve 116 212, 1002, 104 n., 304 n., 306 n. 5,
Reinhardt, Django 64 n. 14 310 n. 15
Resnais, Alain 144 n. 2 Steinberg, Leo 258 n. 5
Rey, Georges 325 n. 24 Stella, Frank 113
Ridley, Aaron 49 n., 91 n. 1 Stern, Josef 293 n. 8
Rimsky-Korsakov, Nicolai 2278 Storr, Anthony 225 n., 227 n. 5
Robbe-Grillet, Alain 68 n. 24, 163 n. Stravinsky, Igor 70, 192, 196, 213
Robinson, Jenefer 40 n. 2, 49 n., 53 n. 19, 87
n., 1001, 108 n. Tatum, Art 64 n. 14
Rochberg, George 70 Tchaikovsky, Peter 319 n.
Rockwell, John 65 n. 17 Thomasson, Amie 2, 335, 36
Roelofs, Monique 323 n. 22 Thompson, Kristin 144 n. 1
Rosen, Charles 214 n. Thurber, James 248 n. 18
Ross, Harold 248 n. 18 Tiomkin, Dmitri 164
Rota, Nino 171, 181 Titian, 256
Rubinstein, Arthur 72 Tobey, Mark 62
Tolhurst, William 302 n. 2
Santayana, Georges 346 n. 19 Tolstoy, Leo 27
Sarraute, Nathalie 68 n. 24 Tormey, Alan 91 n. 1, 350 n. 27
Satie, Erik 1756, 196 Trivedi, Saam 311 n.
Savile, Anthony 191 n. 8, 194 n. 11, 221 Tsukerman, Slava 179 n. 49
n. 5, 306 n. 5, 367 n. 2, 369 n., 370 Twain, Mark 1845
n. 8 Turner, Mark 297 n. 16
Schaper, Eva 42 n. 4 Tyner, McCoy 70
Schiele, Egon 258, 269
Schoenberg, Arnold 70, 192 Updike, John 259 n.
Schopenhauer, Arthur 8, 35565, 389, 391,
394
Schubert, Franz 4, 1412, 2048, 209, 213, Vaida, Iuliana Corina 318 n. 13,
218, 222 324 n. 23
Schumann, Robert 213, 222 n. 7, 349, 350 Van Dongen, Kees 245
Scorsese, Martin 161 Vangelis 1734
Scriabin, Alexander 5, 63 n. 13, 2301, 234 Vasari, Giorgio 71
Scruton, Roger 49 n., 88 n. 14, 92 n., 105, Vaughan, Sarah 64 n. 14
1067, 121 n. 15, 258 n. 5, 266, 319 n., Vaughan Williams 834, 213
336 n. 3, 337 Velasquez, Diego 268, 401, 407, 4167
Shakespeare, William 140 Velleman, David 412 n. 10
Shelley, James 367 n. 2, 371 n. 11, 377 n. 21 Vermazen, Bruce 60 n. 7, 91 n. 1, 934,
Shiner, Roger 367 n. 2, 370 n. 8 1001
Shusterman, Richard 276 n. 3, 366 n. 2, 371 Von Wright, Georg 287 n.
n. 12
Sibley, Frank 78, 60 n. 7, 197 n. 17, 3157, Wagner, Richard 184, 189, 301
343 n. 15 Walhout, Donald 188 n.
Sinatra, Frank 64 n. 14 Walton, Kendall 4, 47 n., 53 n. 20, 94 n. 6,
Sircello, Guy 258 n. 5 103 n., 108 n., 115 n., 130 n. 3, 149 n.
Slater, Harley 46 n. 11, 158 n., 240, 245 n. 13, 246 n. 16,
Sloboda, John 22930 263 n., 316 n. 5, 343 n. 15, 349 n. 25
Slote, Michael 412 n. 10 Warburton, Nigel 2 n. 4
Smetana, Bedrich 138 Warhol, Andy 28, 35
Snow, Michael 116, 144 n. 2 Webern, Anton 196
Solomon, Robert 39 n. Webster, Ben 70
Sontag, Susan 264 n. Weir, Peter 1723
Sparshott, Francis 49 n. Welles, Orson 1589
Spencer, Herbert 393, 394 Wesselmann, Tom 255
Spielberg, Steven 160 Wieand, Jeffrey 366 n. 2
 Index 423
Williams, Bernard 412 n. 11 Yanal, Robert 45 n.
Wilson, George 130 n. 1, 1514, 172 n. 41, Young, Lester 60 n. 8, 64 n. 14, 70
179 n. 48
Wittgenstein, Ludwig 20913, 217, 2189, Zangwill, Nick 316 n. 6, 317 n. 10, 367 n. 2,
276 n. 2, 2945 370 n. 8
Wollheim, Richard 5, 60 n. 6, 656, 23951,
256