(David Hodgson) Rationality + Consciousness Free

Rationality
+
Consciousness
=
Free Will
PHILOSOPHY OF MIND
Series Editor
David J. Chalmers, Australian National University
Recently Published in the Series

Thinking Without Words
Jos Luis Bermdez
Simulating Minds
The Philosophy, Psychology, and Neuroscience of Mindreading
Alvin I. Goldman
Supersizing the Mind

Embodiment, Action, and Cognitive Extension
Andy Clark
Perception, Hallucination, and Illusion

William Fish
Phenomenal Concepts and Phenomenal Knowledge

New Essays on Consciousness and Physicalism
Torin Alter and Sven Walter
The Character of Consciousness

David J. Chalmers
The Senses
Edited by Fiona Macpherson
The Contents of Visual Experience

Susanna Siegel
Consciousness and the Prospects of Physicalism

Derk Pereboom
The Sources of Intentionality

Uriah Kriegel
Rationality + Consciousness = Free Will

David Hodgson
Rationality
+
Consciousness
=
Free Will
David Hodgson
1
3
Oxford University Press, Inc., publishes works that further
Oxford Universitys objective of excellence
in research, scholarship, and education.
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 ofces in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Copyright 2012 Oxford University Press
Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016
www.oup.com
Oxford is a registered trademark of Oxford University Press
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, electronic,
mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data

Hodgson, David (David H.)
Rationality + consciousness = free will / David Hodgson.
p. cm.
Includes bibliographical references.
ISBN 978-0-19-984530-9 (alk. paper)
1. Free will and determinism. 2. Responsibility.
I. Title. II. Title: Rationality plus consciousness equals free will.
BJ1461.H53 2011 123'.5dc22
2011009507
1 3 5 7 9 8 6 4 2
Printed in the United States of America

on acid-free paper
Contents
Preface ix
Introduction 3
The problem of free will and responsibility 3
Compatibilism 4
Haldanes argument 5
How I will proceed 6
1. Foundational Beliefs 10
Can I be certain that I exist? 10
The need for language 14
Experiences and the external world 15
Foundational beliefs 18
2. Truth and Rationality 20

Truth 20
Relativity of truth? 24
Rationality 26
Fallacies and biases 29
Stichs argument 32
A legal example 34
Core assertions about truth and rationality 36
3. Plausible Reasoning 37
Formal and informal reasoning 37
Induction 40
Bayes Theorem 43
Illustration of Bayes Theorem 45
Levels of cognitive processes 48
Core assertions about plausible reasoning 53
4. Consciousness and Decision-Making 55

Dual aspects 55
Characteristic features of conscious experiences 59
Subjectivity 60
Qualia and unity 62
Neural correlates of consciousness 66
The efcacy of conscious experiences 68
Three questions 72
Rule-determined processes do not need consciousness 74
Core assertions about consciousness and decision-making 77
5. Gestalts and Rules 79

The argument outlined 79
Laws and rules 85
The Game of Life and computation 89
Tricks of consciousness 92
Some further thoughts 95
Core assertions about gestalts and rules 97
6. How Gestalts Promote Rationality 98

Evolutionary origins 99
Aesthetic judgments 102
Plausible reasoning 108
Conclusion 111
Core assertions about how gestalts promote rationality 113
7. Science and Determinism 114

A lawful universe 115
Quantum mechanics 118
The free will theorem 121
Explanation of the theorem 124
Implications 126
Time and the block universe 128
Core assertions about science and determinism 132
8. Neuroscience and Conscious Choice 133

Science and the brain 133
A general picture 136
The Cartesian theatre 138
The scale and nature of quantum effects 140
Libet, Gazzaniga and Wegner 148
Core assertions about neuroscience and conscious choice 151
vi Contents
9. Indeterministic Free Will 153
Will and responsibility 154
Comparison with Kane 159
Agent-causation 162
Compatibilism 166
Assessment of compatibilism 171
Does luck swallow everything? 173
More about luck 177
Core assertions about indeterministic free will 179
10. Value Judgments 180

A different philosophical approach 181
Natural imperatives 185
Absolute imperatives 187
Prima facie imperatives 190
No reasonable irreconcilable differences 192
Why be moral? 196
Good, evil and beauty 197
Community practices and laws 198
Legal systems 201
Capacity for reasonable value judgments 203
Core assertions about value judgments 204
11. Responsibility and Retribution 206

Responses to wrong conduct 206
Overview 210
Australian criminal law 213
Retribution as a restriction on State compulsion 216
Why retribution should be maintained 218
Philosophical bases for retribution 223
The future of retribution 225
Core assertions about responsibility and retribution 228
12. The Big Picture 230

The scientic account 230
An experienced universe 231
Constraint, empowerment and guidance 234
Religious belief: a subject for rational enquiry 236
A value-embedded universe 237
Where do we come from? 239
Contents vii
Where are we going? 240
Can more specic beliefs be supported? 243
Potential for evil and good 244
Core assertions about the big picture 246
Appendix A
Why Bayes Theorem Works 249
Appendix B
Against Fundamentalism: Biblical Morality 251
Abraham and Isaac 251
The Passover 252
The Promised Land 253
The New Testament 253
References 255
Index 261
viii Contents
Preface
In my previous book The Mind Matters, published in 1991, I argued that our conscious
mental processes make a contribution to our decisions and actions over and above the
contribution made by physical processes of our brains. That book supported the view
that human beings do have free will in a robust and indeterministic sense; but left
much unsaid about what that contribution was.
In following years, I published articles in which I sought to develop and clarify my
ideas about free will, and also to relate them to questions about our responsibility for
our actions and the justication of retributory punishment in criminal law.
The idea at the heart of the present book, namely that conscious beings can respond
appositely to whole feature-rich experiences that do not engage with laws or rules of
any kind, was rst developed at some length in an article entitled Three Tricks of Con-
sciousness published in Journal of Consciousness Studies in 2002. Parts of chapters 5
and 6 of this book are derived from that article.
In 2007 I wrote a short article in which I attempted a systematic summary of my
thinking about free will, with a view to answering Galen Strawsons very persuasive
argument (referred to in chapter 9 of this book) against the very possibility of free will
and responsibility. This article was published by Times Literary Supplement in July of
that year under the title Partly Free.
It occurred to me that this article could provide a framework for a book bringing
together ideas I had been developing since 1991, and I set about writing a book based
on the article.
As with my previous book, much writing was done on my forty-minute train
journeys to and from my workplace. In addition, I was able to take a period of long
leave from my judicial work in 2009. One month of this was spent as a visiting
fellow at the Philosophy Program of the Research School of Social Sciences at the
Australian National University. There I greatly beneted from discussions with
David Chalmers, Daniel Stoljar, Declan Smithies, Michael Titelbaum, Martine
Nida-Ruemelin, Noa Latham and others, and also from discussions at seminars I
gave there.
ix
The book was also helped by very detailed comments from two reviewers for Oxford
University Press. I am also grateful for comments from Alan McCay, Jeff Gordon and
Julia Roy.
My greatest debt is to my wife Raewyn, without whose encouragement and support
I could not have undertaken this project and brought it to completion.
x Preface
Rationality
+
Consciousness
=
Free Will
This page intentionally left blank
Introduction
TH E PROB LE M OF FR E E WI LL AN D R E S PON S I B I LITY
In Red, the nal lm of Krzysztof Kieslowskis Three Colours trilogy, neighbours throw
a stone through a retired judges window, and the judge muses, I wonder what Id do
in their place. After a pause, he says, The same thing. The girl whos with him asks,
Youd throw stones? The judge replies, In their place, of course. And that goes for
everyone I judged. Given their lives, I would steal, Id kill, Id lie. Of course I would.
And he explains why his life took the course it did: All that because I wasnt in their
shoes, but mine.
Is it the case, as this exchange suggests, that everything we do is the inexorable play-
ing out of the role in life given to us by our genes and our circumstances, by nature and
nurture? Or is it the case that we have real alternatives and real choices about what we
do, so that our actions and our lives are not entirely determined by the luck of our
genes and circumstances, but are inuenced by decisions we can freely make and for
which we are truly responsible?
The problem of free will and responsibility is as old as philosophy itself, and has been
the subject of countless books and articles. There are those who think this problem is
now resolved, because science shows that we human beings are physical systems that
change over time in accordance with the laws of nature that govern all physical systems,
and thus cannot do other than as determined by those laws. According to these views,
whatever we do is either the inevitable product of our genes and our circumstances, or
of chance occurrences, or of some combination of the two; and there can be nothing in
our genes or circumstances (or indeed in chance or the laws of nature) for which we can
be ultimately responsible. Thus, everything ows from things that happened before we
were born or are otherwise outside our control, so that we cannot have free will in any
meaningful sense and cannot be truly responsible for our decisions and actions.
Some who think along these lines claim that these views provide a sympathetic ac-
count of human conduct, which humanely avoids the judgmental approach of blam-
ing people for their wrong actions (which, on these views, they cant help doing). They
say in particular that criminal conduct should properly be regarded as an illness to be
3
treated rather than wrongdoing to be punished, and that any idea of retribution is
based on primitive urges for revenge and should be abandoned.
I disagree with these views. In this book I set out a systematic and cumulative argu-
ment for the view that, while the engagement of our genes and our circumstances with
laws of nature makes an enormous contribution to the way we are and the way we
behave, we nevertheless do make real decisions that are not themselves wholly deter-
mined by that engagement and/or by chance, but are in a fundamental way truly down
to us. I argue that these choices are inuenced by input from our conscious experiences
that is not determined by laws or rules of any kind, but rather is the exercise of a capac-
ity that our consciousness gives us; and that we can thereby be partly responsible not
only for what we do, but also for the way our own characters turn out.
I will also suggest that the idea of retribution is far from primitive and inhumane, but
is rather a foundation of human rights. In short I say that if we do not punish people
because they are guilty, there is less reason to refrain from punishing people if and because
they are innocent. If it is regarded as acceptable that government ofcials treat citizens in
any such manner as appears to be most benecial, irrespective of whether persons so
treated have done anything to deserve that treatment, the way is left open for practices
like putting political dissidents into prisons or mental asylums. Respect for human rights
requires that, with limited exceptions, governments refrain from interfering with the
freedom of citizens unless the citizens have acted in breach of a publicly stated law, in
circumstances where they are responsible for the breach and can fairly be regarded as de-
serving punishmentin which case it becomes permissible to impose punishment that
is reasonably proportionate to the criminality of the conduct in question.
I also believe that, to make the most of life, one needs to recognise ones own respon-
sibility for ones own conduct and (at least partially) for ones own character. Denial of
responsibility does not necessarily result in fatalism, but it is not (I contend) conducive
to efcacious aspiration for virtue and achievement, or resolution to seek them in
ones own life.
COM PATI B I LI S M
So far Ive written as if there are just the two stark alternatives of (1) views denying free
will and responsibility and (2) views, such as that advocated in this book, asserting our
capacity to make decisions that are not wholly determined by engagement of our genes
and circumstances with laws of nature (or by some combination of such engagement
and chance). However, there is also a third class of views, to the effect that free will is
compatible with determinism: according to these views, it is fully consistent with our
having free will and responsibility that our decisions and actions be wholly determined
by engagement of our genes and circumstances with laws of nature.
4 rationality + consciousness = free will

These compatibilist views are advocated by many philosophers, perhaps a majority
of contemporary philosophers who have addressed the problem of free will and
responsibility; and the question of whether or not free will is compatible with deter-
minism has given rise to extensive and sophisticated philosophical debates. I will not
be entering into the detail of these debates, because of what I see as the strength of
considerations in favour of the account of indeterministic free will and responsibility
given in this book.
However, in chapter 9 I will compare my account with compatibilist views, and say a
little about the arguments for the position that free will is compatible with
determinism.
HALDAN ES ARGU M E NT
One of the traditional arguments in favour of free will is the argument attributed to
biologist J. B. S. Haldane to the effect that we cant have reasonable grounds for believ-
ing determinism to be true, because if determinism were true we would simply believe
whatever we were caused to believe by deterministic processes, which govern our be-
liefs just as they govern everything else.
Two difculties with this argument have been recognised.
One is that it assumes that reasonable grounds for belief cannot be deterministic,
whereas it might be thought that our abilities to perceive accurately and to draw rea-
sonable inferences depend for their reliability upon processes that unfold regularly in
accordance with laws of nature and/or rules of logicthat is, on processes that are
deterministic. Thus, it can be argued, determinism supports the reasonableness of our
beliefs rather than the reverse.
The other is that the argument is purely negative: it doesnt give any positive reason
to believe in free will. Certainly, the argument doesnt explain how or why indetermin-
ism or free will could provide any sounder basis than determinism for reasonable
belief. It may be contended that to suggest any basis for reasonable belief other than
deterministic law-governed processes is to appeal to something mysterious or even
magical, a black box of which no explanation can be given.
This book may be regarded in part as a reworking of this traditional argument, in a
way that deals head-on with these two difculties.
As regards the rst difculty, I carefully consider the nature of human rationality, and
I argue that essential to this rationality is the ability to engage in plausible reasoning; that
is, in reasoning in which the premises or data do not entail the conclusions by virtue of
applicable rules for good reasoning, but rather support them as a matter of reasonable
albeit fallible judgment. And I offer an explanation as to how that kind of reasoning
could work in a way that is reasonable yet not determined by rules or laws of any kind.
Introduction 5
As regards the second difculty, I argue that there are very strong grounds for accept-
ing that human beings have the ability to make reasonable decisions about what to do
and what to believe (and to deny this would be to make all intellectual endeavours
pointless). Careful consideration of this ability shows that it requires both conscious-
ness and also the capacity to make choices that are reasonable yet not determined by
laws or rulesso that we have every reason to believe that we have this very capacity,
and thus that we have free will. I also suggest this capacity is no more mysterious than
consciousness itself.
My conclusion is that free will, so understood, is necessary for the ability that we ac-
tually have to make reasonable decisions, and thus is necessary if there is to be any
point to intellectual endeavours.
HOW I WI LL PRO CE E D
So in this book I argue that its reasonable to believe we have free will, in the sense that
we make real choices that are not themselves wholly determined by our genes and
environment and/or by chance, but are in a fundamental way truly down to us. Of
course, scientic discoveries about how the world works and how our brains work are
relevant to this view, and some would say they are incompatible with it; and I will
have regard to these discoveries. But my argument will take account also of wider
considerations.
Basic to my argument are the ideas (1) that in order to arrive at reasonable and justi-
ed beliefs we must use the abilities we have that are best able to produce such beliefs
and such justication; and (2) that these abilities are captured by the concept of ratio-
nality, which I understand as not being limited to intellectual ability but as including
all those perceptual, intellectual and emotional capabilities by use of which we can (at
least sometimes) make reasonable decisions about what to believe and what to do.
In relation to many questions it is rational to rely on the scientic method of formu-
lating and testing hypotheses, in order to arrive at reasonable beliefs about the world;
but it is important to appreciate that, although this method is a very important and
reliable way to arrive at reasonable beliefs, it is not the only way. For one thing, as I will
show, the scientic method itself presupposes the truth of some prior beliefs; and be-
liefs based on the scientic method cannot be more reliably true than the prior beliefs
on which the validity of the scientic method depends. I will argue that these prior
beliefs have rational support that does not depend on the scientic method.
Rationality is thus more fundamental than the scientic method; and it is also wider,
in that it can be applied to questions that cannot be addressed by the scientic method,
such as questions as to the nature of rationality itself, and also moral questions about
how we should conduct ourselves and what things we should value.

Coming now to the scheme of the book, in the rst three chapters I address questions
concerning the rational basis of our beliefs, at a level more fundamental than the sci-
entic method.
I start in chapter 1 by considering what are my own most certain and most funda-
mental beliefs, and I suggest they are the most reasonable starting points for addressing
questions such as these. Like Descartes, I begin by considering what if anything I can
be absolutely certain about; but unlike Descartes, I nd that I cannot be certain about
my own existence, but only that conscious experiences occur. As I explain, this is partly
because of the need for a language with which to express and communicate beliefs of
any complexity, and thus the dependence of complex beliefs on prior beliefs about
language and language use; and this in turn requires prior beliefs about language users
and a world in which language is used.
I begin chapter 2 by saying a little about truth. A central aim of my enterprise is to
arrive at and justify true beliefs about free will and responsibility, on the way to achiev-
ing some understanding of these matters; so it is important to make explicit my con-
ception of what counts as truth. And then I address some general considerations
concerning rationality. As I have already suggested, rationality is presupposed in any
undertaking to arrive at and justify true beliefs, and it is useful to articulate what is
involved in this presupposition.
In chapter 3, I consider one very important aspect of our rationality, namely our abil-
ity to engage in plausible reasoning; that is, as mentioned above, reasoning in which
the premises or data do not entail the conclusions by virtue of applicable rules, but
rather support them as a matter of reasonable albeit fallible judgment. I argue that
even the scientic method depends on plausible reasoning, and that plausible reason-
ing cannot be fully explained in terms of rules for good reasoning.
In the following three chapters, I argue that our consciousness contributes to our ratio-
nal decision-making in ways that are not wholly determined by rules or laws of any kind,
and thus that our decisions can be both rational and not determined by a combination of
pre-existing circumstances (including our own characters) and laws of nature. I conclude
that we do make real choices that are not themselves wholly determined by our characters
and circumstances or by chance, but are in a fundamental way truly down to us.
In the rst of these three chapters, chapter 4, I argue that consciousness must have a
role to play in our decision-making and that this role is not one that could be per-
formed by rule-determined processes (if it were, consciousness would be a superu-
ity). I identify the three distinguishing features of conscious processes that I contend
contribute to decision-making, namely subjectivity, qualia and unity.
In chapter 5, I offer a specic and straightforward proposal as to what the role of
consciousness is, namely to enable an organism to be responsive to circumstances
grasped as wholes, not just to their constituent features that engage with laws or rules.
Introduction 7
And I support this proposal by an original argument of mine, to the effect that in our
conscious experiences we grasp and respond to whole feature-rich gestalts, which do
not as wholes engage with any applicable laws or rules.
In chapter 6, I consider in more detail how the grasp of gestalts contributes to rea-
sonable decision-making, in particular (1) by non-human animals that have conscious
experiences somewhat like ours, (2) in making aesthetic judgments, and (3) in plausi-
ble reasoning generally. My conclusion is that indeed our reasonable decision-making
utilises conscious experiences in which we grasp and respond to whole feature-rich
gestalts in ways that are not determined by laws or rules.
Having reached that point, in the next two chapters I consider how this conclusion
stands with science in general and neuroscience in particular.
In chapter 7, I consider whether the general picture of the world suggested by the
physical sciences is compatible with my conclusion. I argue that the best contemporary
physical theories support two propositions favourable to my position, namely that
causation is not local and that the world is not deterministic. I also argue that the so-
called block universe view, denying reality to the passage of time and suggesting an
unchanging (and in that sense deterministic) universe, is not reasonably tenable.
In chapter 8, I give an account of how the brain could work consistently with my ap-
proach, and consider whether this is compatible with what neuroscience tells us about
the brain. To the extent that neuroscience seems to leave no room for any contribution
from conscious experiences over and above their neural correlates, I contend this is a
manifestation of current neurosciences inability to give a satisfactory account of con-
sciousness generally, and does not count strongly against my approach.
In the next three chapters I discuss how my views on conscious decision-making
support ideas about free will, responsibility and retribution.
In chapter 9, I draw together my conclusions about free will and responsibility.
I distinguish my position from other views generally favourable to free will, including
views that free will is compatible with determinism, and views that suggest it is a spe-
cial kind of causation, called agent-causation, that gives us free will; and I show how my
position answers the luck swallows everything argument against responsibility.
In chapter 10, I further advance my support for ideas of retribution by developing the
idea that we can justly be praised or blamed (and even punished) for what we do, con-
sidering whether there can be any truth in value judgmentsbecause if there is not, how
could it be truly just to praise or blame anyone for anything? I also consider whether
we have the capacity to make reasonable judgments about these matters. In chapter 11,
I give reasons why I contend it is important to maintain ideas of responsibility and
retribution in the criminal law.
Finally, in chapter 12, I consider what my approach to free will suggests about our
universe and ourselves, concluding that our universe is not one of blind pitiless indif-

ference, as Richard Dawkins would have it; but that, on the other hand, there is no
sound basis for accepting the literal truth of any religion, and indeed that fundamen-
talist religion of any kind has the potential for great evil.
As I proceed, I will highlight and number core assertions, and I will collect at the end
of each chapter the core assertions of that chapter. This will I hope help to make clear
the development and cumulative effect of the case made out in this book.
This book is a greatly expanded version of an argument previously set out in an ar-
ticle entitled Partly Free published in Times Literary Supplement in July 2007. I believe
it constitutes an original, coherent and plausible approach to a group of age-old un-
solved philosophical problems, and the book is intended as a contribution to philoso-
phy. However, I have tried to write it in such a way as to make it fully understandable
by non-philosophers who are interested in questions such as whether or not we truly
have free will, and whether (and if so to what extent) our responsibility for our con-
duct is affected by our very different genetic and environmental backgrounds.
Introduction 9
1
Foundational Beliefs
CAN I B E CE RTAI N THAT I EXI ST?
Like Ren Descartes in the seventeenth century, I would like to start my enquiry with
something I can be absolutely certain about. Descartes asked himself what he could
not doubt, and answered to the effect that he could not doubt that he was asking this
question, leading him to be certain about both his thinking and his existence: I think
therefore I am.
As a young university student, I was quite impressed by the argument that the one
thing I could be certain about was my own existence and experiences, and that every-
thing else, including the existence of a world independent of me, and the existence of
other persons who also had experiences somewhat like mine, was uncertain, supported
at best by inconclusive inference. But I never embraced the view (called solipsism) that
I and my experiences are all that there is; and I never believed that there was any real
doubt about the existence of the outside world and other people, because I thought
the inference that these things existed was pretty strong. But I also thought this line of
argument tended to support Descartess starting point.
Later on, I realised that Descartess argument could not establish very much about
the nature of the I that was said to be doing the thinking and that therefore existed.
Certainly it could not establish that I was an immaterial soul or self, distinct from my
body, or something that could survive the death of my body. I still felt sure that for
there to be thinking there had to be a thinker; but I saw no compelling reason why the
thinker had to be non-material, or even had to persist beyond any particular occasion
of thinking.
So I came to accept Descartess premise I think as supporting no more than that
thinking is going on by something, with the possibility that this something could have
no more substance or longevity than the particular thinking in question.
Later again, I came to the view that even this formulation cannot be a satisfactory
starting point as the most certain of my beliefs. This is because its truth is conditioned
on the truth of certain assumptions, so that one cannot justiably be more certain
about this formulation than one is about those assumptions. There are two main rea-
sons for this.
10
First, this formulation distinguishes something doing the thinking from the thinking
itself, and thus presupposes that there is a thinker, albeit possibly an ephemeral one,
which is not itself merely part of the thinking. But perhaps the apparent thinker is
mere illusion, or a construction that is part of the thinking or at least not distinct from
the thinking. Perhaps there is no subject of experience that can be distinguished in a
satisfactory way from the content of the experience. This idea can be traced back to
Humes view of the self as a bundle of perceptions, and nds recent expression in the
views of people like the philosopher Daniel Dennett1 to the effect that our notion that
there is a self doing the thinking is a device selected by evolution to help the brain make
sense of its own processes. Now I disagree with these views, but they need to be ad-
dressed on a rmer basis than is appropriate at this stage of the enquiry.
Second, and more fundamentally, even the formulation thinking is going on by
something is an assertion the full signicance of which can be appreciated only
through a quite sophisticated understanding of the language in which the formulation
is expressed; so that the truth of this assertion is conditioned on the truth of assump-
tions about the correct use of this language and the correct use and meaning of expres-
sions within this language. How can I be certain that thinking is going on by something,
unless I have sufcient understanding of this expression to appreciate precisely what it
does assert and what it doesnt assert? This in turn requires an understanding of the
relevant language and how it is used. It could be argued that all that is necessary is to
understand just a few simple concepts; but while I will argue that some concepts are
sufciently simple not to raise questions concerning subtleties of language use, I con-
tend that the assertion thinking is going on by something does not have that kind of
simplicity. It raises difcult questions about exactly what thinking is (as distinct from
other kinds of mental activity), and exactly what it is for thinking to be by something.
These questions can be explored only through careful use and understanding of the
relevant language. And language presupposes a community of language users, and
practices as to the use of language by that community. So a belief that thinking is going
on by something cannot be justiably held with any more assurance than a belief that
there is a language in which words thinking and by something have a place and a
meaning.
The same considerations apply, perhaps with greater force, to the assertion I think,
because of difcult problems, mentioned above, associated with the meaning and sig-
nicance of I. That is, in addition to the problems raised by the question whether
there is a thinker of any kind distinct from the content of the experience, there are
problems raised by the question of exactly what is this I, what is the subject of
experience.
1
Dennett (2003), 24555.
Foundational Beliefs 11
So far in this chapter, Ive been explaining why something like Descartess I think
therefore I am is not my most certain belief. What then is my most certain belief? I will
set to one side beliefs that there is something rather than nothing and that things happen:
the beliefs I am concerned with are beliefs with some content about what it is that exists
and what it is that happens, and about how and why things exist and happen. The
belief of this kind of which I am most certain is:
1.1. Conscious experiences occur.
Of course, Ive expressed this belief in the English language, and this expression
depends on there being such a language and on appropriate use of the words con-
scious, experiences and occur. But the idea expressed can be considered as quite a
primitive idea not depending on subtleties of language use. Understanding of the idea
does not presuppose sophisticated understanding of language use concerning prop-
erties of conscious experiences or kinds of conscious experiences or the existence of an
experiencer, or concerning distinctions between types or modes of occurrences.
Im not absolutely certain that non-human animals have conscious experiences,
although (as will appear shortly) I rmly believe that at least mammals and birds do.
Assuming that they do, I believe there is a sense in which an animal knows that an expe-
rience it has is occurring, for example that a pain it feels is occurring, even though of
course it could not express this in language. It is in that general sense that I am certain
above all else that conscious experiences occur.
When I feel a severe pain, I am certain that this pain is occurring. Whether it is
correct to call it pain, or to say it is my pain, or even my conscious experience, are
matters depending in part on language and its correct use; and I cannot justiably be
more certain about these things than I am about the existence of language and of lan-
guage users, and about the way language is used by them. But so long as I dont attempt
to categorise types of conscious experience, and so long as I dont attempt to identify
an experiencer, my belief in the occurrence of a pain that I feel can be considered as an
acceptance of an idea that does not depend on subtleties of language use; and taken in
that sense, my belief that conscious experiences occur is more certain than any other
belief I have. And while I have selected pain as my rst example of conscious experi-
ences, the conscious experiences, of whose occurrence I am certain, extend to the
whole range of conscious sensory, perceptual, cognitive and emotional occurrences
that make up my ongoing conscious experiencing (although, I repeat, I do not include
the categorisation of these experiences as part of this most certain belief).
Of course, the questions of what is consciousness, and what is its place in the world,
are extremely difcult and controversial questions; but in this most basic belief I am
not asserting anything about these questions. I am merely saying, about such conscious

experiences as I happen to have, that these occur, whatever they may be and however
they may be correctly described.
This is the most certain of my beliefs, if only because there is a theoretical possibility
that all my conscious experiencing could be some kind of a dream or illusion; although
of course without language I could not say or believe anything useful about dreams or
illusions, or contrast them with experiences that are not dreams or illusions but are
experiences of reality. And in fact I do not believe for a moment that there is any real
possibility that my conscious experiencing is no more than a dream or illusion.
Because this is the most certain of my beliefs, one important requirement that I look
for in other beliefs is that they be consistent with and cohere reasonably well with this
belief. It is conceivable that I could be caused to doubt or modify or even abandon this
belief by some other consideration or combination of considerations; but this hasnt
happened yet and I dont think it will happen. And although this belief might seem
trite, vague and general, I think it is most important, because it is a belief about
something that science has so far failed adequately to accommodate or explain. I will
elaborate on this in chapters 7 and 8.
I also make the more general comment that, although I have gradations of condence
in my beliefs, and although in general terms I am starting in this book with near-certain
beliefs and then progressing to less certain beliefs, I am not saying that my most certain
beliefs are not revisable, or even that I would not revise beliefs of which I am presently
more certain because of considerations associated with or derived from beliefs of which
I am presently less certain. I am aiming for consistency and coherence; and although,
where I nd some lack of consistency or coherence, I generally look to modify beliefs of
which Im presently less certain rather than beliefs of which Im presently more certain,
this will not necessarily be the course I nally take. My ranking of beliefs in terms of
certainty or assurance is itself a belief that is no more than provisional.
There is one further belief that I have about conscious experiences, in which I have
the highest condence and which does not depend on distinctions that require sophis-
ticated understanding of language; and that is my belief that some conscious experi-
ences change. Visual and auditory experiences, for example, are generally not static but
are changing. Whether this is a matter just of changes occurring in the experience
itself, or whether it indicates changes occurring in something that is being experienced,
involves matters depending, in part at least, on categorisation and language; but the
belief in change in the experiences is a primitive idea not depending on such niceties.
As with my rst belief, it can be said that there are questions concerning change that
are extremely difcult and require careful analysis through language: for example, the
relationship between change and the passage of time. However, my belief that some
conscious experiences change asserts no more than that, among those conscious expe-
riences that I happen to have, some at least are not static or constant, but rather change.
I believe there is a sense in which an animal can know that an experience it has is
changing, for example that sound it hears is not constant, even though of course it
could not express this in language. It is in that general sense that I have high certainty
that some conscious experiences change; and I would challenge anyone to come up
with any reason for doubting this.
So I give as my second belief:
1.2. Some conscious experiences change.
This too might be considered trite, vague and general; but I think it too is important,
particularly because it is not adequately accommodated or explained by a prominent
scientic view about the universe and nature of time. I will return to this in chapter 7.
TH E N E E D FOR LANGUAGE
Now, once I attempt to go beyond the beliefs that conscious experiences occur and that
some conscious experiences change, I nd I cannot formulate and consider beliefs in
any useful way without drawing distinctions that need to be identied with some pre-
cision in language. There is a sense in which I, like non-human mammals and pre-
linguistic human beings, can have quite complex beliefs about the world that do not
depend on their being expressed in language. For example, a lion may believe in the
existence of its prey that it sees, and a baby may believe in the existence of a toy it is
grasping, and these beliefs would not depend on language. But for me to achieve
anything constructive in formulating my beliefs, and even in thinking about them,
I have to use language and use it correctly; and this in turn depends on my having
an understanding of how language is used by a community of language users.
My experience of life, my interactions with other people, my observations of their
use of language, and my own use of language, give me strong reasons to believe in the
existence of language and in the existence of communities of language users, of one of
which I am a member. There is no good reason for me not to believe in these things,
and this belief is necessary if I am going to make sound progress in expressing or even
thinking about my beliefs. So I think it is reasonable to place next in my hierarchy of
fundamental beliefs the following:
1.3. There is language, and there are communities of language users, of one of which I am
a member.
This belief does not say much about the characteristics of members of the language
community that includes me, other than that they have the ability to use the language

that I use. But it is a starting point for further beliefs in which I can be similarly
condent.
The members of the language community that includes me are all human beings,
one of many species of animal life on this planet. I condently believe that I am a
human being, one of a large number of human language users; although that is not to
say that this is all that I am or that this describes exactly what is most important about
what I am; or that this is precisely what I am referring to when I talk or think about
myself as a subject of experience and action. An assertion that I am a human being
reserves judgment on whether or not I may be an immortal soul, or at least a spiritual
being of some kind, or a conscious subject or self that is not wholly dependent on my
existence as a physical creature. So:
1.4. My community of language users consists of human beings, a species of animal life,
and I am a human being.
EXPE R I E NCE S AN D TH E EXTE R NAL WOR LD
Human beings have many characteristics in common, in terms of physical constitution,

appearance, behaviour, and so on. The language that the members of my language
community share can express and describe many features of ourselves and the world
in which we live, including conscious experiences of the general kind that are the object of
my primary belief. Because of our similarities, and because it is reasonable to believe
that language developed and was used (and continues to develop and be used) to ex-
press and communicate things that language users could understand, it is reasonable
to believe, with very high assurance, that these language users, and indeed all normal
human beings, have conscious experiences of this general kind. But it is also reasonable
to believe that the actual experiences that give me certainty that conscious experiences
occur are experiences that I have; and that while other human beings have experiences
that are of the same general kind as the experiences I have, their actual experiences are
experiences that each of them has and I do not have.
1.5. The actual experiences that give me certainty that conscious experiences occur are
experiences that I have, and other human beings also have experiences of the same general
kind.
There also seem to be many features of the world in which we human beings live that
are expressed and described in language, but are not themselves experiences of human
beings, although they seem to be objects of experiences and potentially causative of
experiences.
There is a vast philosophical literature concerning the place of conscious experiences
in the world and their relationship with other aspects of the world; and there are many
possible views.
One, called idealism, is to the effect that everything that exists is a kind of conscious
experience or somehow reducible to conscious experience. Another, called materi-
alism or physicalism, is to the effect that everything that exists is physical (of the nature
of matter, energy and the like), and that conscious experiences are correctly consid-
ered as just one way of regarding or describing certain physical things or states or
processes. Both of these views are species of monism, which holds that there is one
basic mode of existence. Other views fall into the broad category of dualism, which
holds that there are two basic modes of existence, that of conscious experience (or the
subjective) and that of the physical world (or the objective), the latter of which can
cause conscious experiences but is fundamentally distinct from conscious
experiences.
There are two broad kinds of dualism: the dualism of substance associated with
Descartes, drawing a distinction between immaterial souls or selves which have expe-
riences, and material brains and bodies and other things that dont; and dual aspect
theories, according to which at least some things or systems may have both a material
or physical aspect and also a mental or experiential aspect. And within dual aspect the-
ories, there are those that in effect say that there is a dual aspect to everything (panpsy-
chism) and those that say that the capacity for experience is an aspect only of things
like the central nervous systems of humans and animals, and thus has emerged only
with the development of such systems.
And there are other theories that may be considered as denying all the above; in
particular, theories saying that the physical and the experiential are manifestations of
something else common to both (neutral monism), some of which assert that the dis-
tinction between the physical and the experiential is just one of many distinctions of
similar importance that can be drawn concerning what exists.
Table 1.1 shows the general relationship between these views.
Some of the distinctions between some of these views are not sharp. In particular,
there are versions of panpsychism that recognise gradations of the experiential aspect
of things, from the full consciousness of human beings down to some minimal con-
sciousness or potential for consciousness, and these versions can come very close to
emergence theories. And the distinction between neutral monism and dual aspect the-
ories is also not sharp, lying mainly in the degree of emphasis given to the distinction
between the physical and the experiential.
I cant attempt to resolve the differences between these views at this stage of my
enquiry or with the assurance required here. But I can make two preliminary observa-
tions about them, and express three further beliefs that I hold with high condence.

Table 1.1 Views on the relationship between the experiential and the physical
Monism Dualism
Idealism Substance Dualism

Everything is experiential Some things are experiential and other things are
physical
Physicalism
Everything is physical Dual Aspect Theories
At least some things have both a physical
Neutral Monism aspect and an experiential aspect:
Everything is neutral as between the subdivided into panpsychism (everything has an
experiential and the physical experiential aspect as well as a physical aspect)
and emergence theories (everything has a
physical aspect but only some things have an
experiential aspect)
First, and contrary to some versions of idealism, Im condent there is an external

world that is not just part of the conscious experiences of individual human beings.
Features of this external world are perceived similarly by different human beings, and
in that sense are common to and accessible in common by different human beings, and
so can be considered objective rather than being or depending on conscious experi-
ences of any individual human being; and they are treated as such in our language.
I am not here asserting that these features must be wholly independent of conscious
experiences, merely that they are not just part of the subjective experiences of individual
human beings but are objective in the sense of being accessible to and appearing much
the same to different human beings.
There is an argument sometimes advanced in favour of idealism to the effect that we
cannot conceive of anything except as being a possible object of conscious experience;
but I dont think this in fact supports idealism. However, it does support something
like Kants view that we cannot know about how things truly are in themselves, but
only about how they seem to us, with our particular perceptual and cognitive capac-
ities; so that, for example, it may be said there is no such thing as colour in the external
world, but only features that induce experiences of colour in normal human beings.
I will say more about this later.
So I state my sixth and seventh fundamental beliefs:
1.6. There is a world external to the conscious experiences of individual human beings,
features of which are accessible in common by human beings and describable in
language.
1.7. Our language may describe at least some of these features by reference to how these
features seem to normal human beings, and this may not adequately describe how they are
in themselves, independently of any experience of them by us.
Second, I am certain that human beings who are awake and functioning normally have
conscious experiences like mine, and I believe with high condence that other mammals
and also birds have the capacity for conscious experiences somewhat like mine, because
of their similarities to human beings in physical constitution, origins and behaviour. I
believe with decreasing condence that much the same is true of reptiles, amphibians,
sh and invertebrates. I think it unlikely that either plants or more primitive forms of life
have that capacity. I believe with very high condence that anything that is not alive, such
as a rock, does not have conscious experiences in the same sort of way as humans and at
least some other animals have experiences. Thus with high condence I believe:
1.8. Some entities or systems (including humans and some non-human animals) have
conscious experiences more or less like mine, and some (including rocks) do not.
This is not an outright rejection of panpsychism. I am not asserting that there could not
be some primitive experiential aspect to non-living things, merely that they are not con-
scious in the sense of having experiences that are in any way similar to the experiences
that human beings have, and that (I believe) some non-human animals also have.
There are some scientists and philosophers who contend that some non-living sys-
tems, such as computers, are or may be conscious in this sense. In chapters 4 to 6 I
will argue for views indicating that computers, in their present form, could not be
conscious; and that the exponential increases in computing power that are expected
to occur in coming years will not make them conscious. So in that crucial respect, I
suggest, computers are more like rocks than like human brains.
FOU N DATIONAL B E LI E FS
1.1. Conscious experiences occur.

1.2. Some conscious experiences change.
1.3. There is language, and there are communities of language users, of one of which I am
a member.
1.4. My community of language users consists of human beings, a species of animal life,
and I am a human being.
1.5. The actual experiences that give me certainty that conscious experiences occur are
experiences that I have, and other human beings also have experiences of the same general
kind.

1.6. There is a world external to the conscious experiences of individual human beings,
features of which are accessible in common by human beings and describable in
language.
1.7. Our language may describe at least some of these features by reference to how these
features seem to normal human beings, and this may not adequately describe how they are
in themselves, independently of any experience of them by us.
1.8. Some entities or systems (including humans and some non-human animals) have
conscious experiences more or less like mine, and some (including rocks) do not.
2
Truth and Rationality
In the previous chapter, I put belief in the existence of language and language users
towards the top of my list of beliefs, because there were strong reasons for this belief
and because this belief was necessary for any useful consideration and elaboration of
beliefs about the world. Much the same applies to my belief in truth and rationality.
I am aiming for an understanding of free will and responsibility, an understanding that
includes beliefs about these topics that are as close as possible to the truth: otherwise,
so it seems to me, the whole enterprise would be empty. And I can be condent that my
beliefs are true or close to the truth only if they are supported by reason, and if I believe
in the reliability of reason. If I did not believe in reason, and in particular in my own
rationality, I could have no sound basis for condence in any of my beliefs; and again,
there would be no point in my addressing these questions.
TR UTH
I am aiming for true beliefs, and I am expressing my beliefs as assertions in the English
language. What is it that makes such assertions true or not true?
For a start, I accept what is called the minimalist statement of what truth is, to the
effect that an assertion such as x is y is true if x is y: I do not think this much can rea-
sonably be denied. But more needs to said about truth, because the minimalist state-
ment does not answer the question, what is it for x to be y? Is it something about a
reality that can be distinguished from the assertion x is y, or is it rather that the asser-
tion itself is justied in some way other than by its relationship to some other reality?
And if the latter, is there something universal about the relevant justications, or are
they relative to different times and places and interests? There is an enormous philo-
sophical literature on the topic of truth, and it is beyond the scope of this book to
discuss it in detail;1 but I think it important to my overall case to make it clear what my
views on truth are.
1
A useful summary of some of the issues can be found in Blackburn (2005).
20
My basic position is that what can make an assertion true is its relationship to reality,
albeit that in some cases the reality itself may be dependent on or partly constituted by
our theories, our concepts, and our language. Of course, we can justiably believe that
an assertion is true only if it is justied by reasons that support the existence of the ap-
propriate relationship to reality; but I say that it is this relationship to reality that can
make the assertion true, and that while other justications may support the reason-
ableness of belief in the assertion, they are not constitutive of its truth.
What then is the relevant relationship to reality? It is sometimes said that it is corre-
spondence with reality. I agree with this, but only if correspondence with reality is un-
derstood in a particular way. Consistently with the minimalist statement of what truth
is, I say that an assertion is true if what it asserts is fact or reality (something that is
and/or was the case, and/or occurs and/or occurred): truth, I say, is correspondence with
reality, with the required correspondence being one of assertion or meaning.
2.1. Truth is correspondence with reality by way of meaning; that is, an assertion is true
if what it asserts is fact or reality.
Now meaning is not always straightforward, so neither is truth. Indeed, meaning is

sometimes dened in terms of truth (more precisely, truth-conditions, or the circum-
stances in which the statement whose meaning is in question would be true), suggest-
ing there may be circularity in my approach. However, as the following discussion
conrms, this does not preclude a commonsense grasp of the interrelated concepts of
truth and meaning.
Language developed initially in dealing in a commonsense way with everyday mat-
ters, and it is in relation to such matters that statements generally have their clearest
meaning and may be considered unambiguously true or not true.
Even in relation to everyday matters there may be problems, in so far as there are
misconceptions implicit in language concerning everyday matters, although such
problems are rarely signicant. Take for example the assertion:
In Sydney this morning the sun rose at 6.30 am.
This assertion could be taken as implying that the phenomenon described occurred in
Sydney itself, and that it did so by reason of the movement of the sun rather than the
rotation of the Earth; and to the extent that it was so taken it would be false. But gener-
ally it would not be taken that way and, assuming the phenomenon did occur at the
time asserted, the assertion would be recognised as unambiguously true.
But language is not limited to dealing in a straightforward way with everyday mat-
ters; and when language is applied to other than everyday matters, truth may only be a
Truth and Rationality 21

matter of degree. Further, our statements do not always assert straightforward states of
affairs and/or occurrences: many of our assertions are about ideas rather than things,
or assert matters of interpretation or evaluation, or contain implications or subtexts
that complicate what is being asserted. And in many cases the reality to which our
assertions must correspond, if they are to be true, is to a considerable extent shaped or
even constituted by our theories, our concepts, and our language; so that their truth
may depend in part on the adequacy of the relevant theories, concepts and/or language
to reect reality, as well as on correspondence with the reality so reected.
There do appear to be natural kinds in the world, that is, natural categories of things
that exist independently of the conceptual and linguistic activities of humans, includ-
ing elements such as hydrogen and gold, and species of living things such as lions and
human beings; and in relation to them truth may be straightforward. But much of the
world that we describe in language is divided into categories by theories and other
conceptual schemes invented by humans, rather than simply existing as such natural
kinds. Even the categorisation of ordinary middle-sized objects that are the most fa-
miliar constituents of our world often depends at least in part on human intentions
and purposes, and thus on concepts and language. This is the case with artefacts such
as chairs or pens; and it is also the case with many natural objects, such as vegetables or
minerals. What this means for a correspondence theory of truth is that the truth or
falsity of an assertion may depend not only on the relation of the assertion to reality as
conceptualised in the language of the assertion, but also on the adequacy of this
conceptualisation.
This is so, for example, in relation to historical explanations of social concepts such
as democracy. The statement The Magna Carta was signed in 1215 asserts a straight-
forward occurrence, and may be considered unambiguously true if what is asserted
actually occurred as asserted. However, this is not the case with the following
assertions:
1. The Magna Carta was the foundation for the development of democracy in England.
2. The Magna Carta has nothing to do with democracy: it was just a document resolving
particular disputes between the monarch and members of the aristocracy.
In each case, what is asserted is not a straightforward occurrence or state of affairs, but
rather ideas that themselves depend on language, interpretation and evaluation. In
relation to assertions like these, correspondence with reality is not a straightforward
matter, but depends in part on reasonableness in use of concepts and in interpretation
and evaluation; so that there may be a large grey area where statements are not unam-
biguously true or not true. The view can be taken that, although these two statements
are contradictory, there is an element of truth in each of them.

The same can apply in relation to concepts that have a role in scientic theories, such
as mass and energy. The assertion contained in the equation E = mc2 is unambigu-
ously true if what is taken to be asserted is a fact about or within the theory of relativ-
istic physics of which it is part: whether it otherwise accurately asserts an actual state of
affairs depends on whether this theory itself adequately deals with the relationships
that exist in the world between those features of the world that are represented by theo-
retical concepts such as mass, energy, speed and light.
Some physicists assert that such theories are no more than models to enable predic-
tions of observations, rather than being assertions about reality; but I believe the better
view is that such theories do assert relationships as actually existing, and do so with
greater accuracy and greater assurance as they are improved over time. Relativistic
physics approximates very closely to reality, at least in dealing with certain aspects of
reality, and the assertion E = mc2 can be considered as making an assertion about an
actual state of affairs and, so considered, as being at least approximately true.
Sometimes, our concepts and language are so inadequate that reality can best be
described by a number of assertions that are apparently inconsistent. For example,
prior to about 1900, scientists would have condently asserted:
All waves are periodic processes extended in space.

No particles are periodic processes extended in space.
Therefore nothing can be both a wave and a particle.
However, by 1930 it was clear that photons and electrons displayed wave-like properties
under some experimental arrangements and particle-like properties under other exper-
imental arrangements. There was no single concept in our language that adequately
expressed what they were; and the best description that could be given of them in ordi-
nary language was that they were in a sense both waves and particles. This apparent
inconsistency did not prevent this from being the description in ordinary language that
was closest to the truth; and according to the pioneering quantum physicist Niels Bohrs
principle of complementarity, such inconsistent descriptions were acceptable in cases
where the incompatible properties could not be displayed simultaneously.
Where (as with many philosophical issues) an assertion is about matters remote
from the practical situations that were the main concerns of language users as language
evolved, our language may be less than fully adequate to express the reality of the
matter, so that the application of the concepts of truth and falsity becomes problem-
atic. This may be so in relation to matters such as consciousness and free will, the
topics of this book.
It is probably so in relation to matters such as the nature of our existence, the exis-
tence and role of any superhuman consciousness or intelligence or purpose, the exis-

tence of an afterlife, and so on, which I will touch on in the nal chapter. There are
deep mysteries about the universe and about our place in it. In relation to these mat-
ters, I think it is unrealistic to expect to be able, with our concepts and language, to
express the full reality in a straightforward way. It may be that the best we can hope to
do is to approximate to the truth by way of poetry or allusion, or by allegory or meta-
phor; and sometimes even by many allegories and metaphors, which may not all be
consistent with each other.
Thus:
2.2. Where assertions are about ideas or interpretations or evaluations, or where lan-
guage is inadequate to express the reality under consideration, truth may be a matter of
degree; and sometimes reality can best be expressed by poetry or metaphor, or even by a
number of inconsistent metaphors.
It is partly for this reason that I believe some philosophical controversies are arid: for
example, disputes about whether the world really consists of facts not things, as the
young Wittgenstein asserted, or whether it really consists of states of affairs, or pro-
cesses, or events. In relation to these matters, the question is not so much truth as such,
but rather the relative merits of different categorisations, each of which may have
advantages and disadvantages.
However, ideas and language can develop, and this may in time allow more accurate
expression of what were previously mysteries. To some extent this has occurred with
quantum mechanics, as understanding has improved; although there are still para-
doxes for which our language seems to be inadequate. So:
2.3. Developing understanding may extend the range of language, so that it can better
express more of reality.
R E LATIVITY OF TR UTH?
Problems such as those I have been discussing have led some writers to argue that it is
pointless to postulate some reality outside our assertions to which our assertions cor-
respond if they are true, because reality itself is so much shaped and constituted by our
ideas; and therefore these writers reject correspondence theories of truth, and adopt
relativistic views of truth.2 Generally, these views are something like the models
For a persuasive and moderate version of this approach, see Putnam (1981), 4954; and a more
2
extreme version, in Stich (1990), is considered later in this chapter.

approach advocated by some physicists, to the effect that true statements are those
which are part of or t well with a theory that works, in that it makes sense of the world
and enables one to deal effectively with it and achieve desired results in acting in it.
I contend that these approaches are self-refuting, and also make rational appraisal of
beliefs impossible.
They are self-refuting because they deny correspondence to reality to themselves; so
that their own truth can be no more than that they themselves somehow work in
practice or are part of some wider theory that works in practice. They cannot even
assert that this is how the word truth is, as a matter of reality, used in our language;
only that the assertion that this is how the word is used is itself an assertion that works
or is part of a theory that works.
Advocates of these approaches may argue that this does not amount to self-refuta-
tion, because they are quite content with a view about truth, and about the way the
word truth is used in our language, that does not itself correspond to any other reality,
so long as this view works or is part of a system that works. I do not think that this
position can reasonably be maintained. For one thing, it is surely a question of fact
whether or not the word truth is used in that way in our language; yet according to
these views, whether or not it is true that this is how the word is used depends not on
this question of fact, but on whether the assertion that it is so used works or is part of
a theory that works. Thus, advocates of this view must say that, even if it is not the fact
that this is how the word is used (as I believe it is not), it may nevertheless be true that
it is so used.
In addition, these approaches make rational appraisal of beliefs impossible because,
according to them, such appraisal must depend on coming to a view as to whether the
beliefs in question work in practice or are part of a theory that works in practice.
Consideration of whether something works in practice must generally be undertaken
by some combination of observation and inferences, with inferences requiring the use
of language and assessment of whether various assertions used in the course of draw-
ing inferences are or are not true. However, on relativistic views of truth, these assess-
ments will not be by way of consideration of what state of affairs actually exists and/or
existed, and/or what actually is occurring and/or occurredbecause what would con-
stitute the truth of any assertion relevant to whether something works in practice can
only be that this assertion itself works in practice or is part of a theory that works in
practice. And so on ad innitum.
What I say is that the truth of beliefs must ultimately be grounded in reality, if these
problems are to be avoided:
2.4. Relativistic views of truth must be rejected because they are self-refuting, and make
rational appraisal of beliefs impossible.

I will say more about this later in this chapter, in considering a position advocated by
the philosopher Stephen Stich.
RATIONALITY
I now consider rationality.

I have extensive experience of perceiving accurately, and of drawing inferences from
what I perceive that turn out to be correct. When my perceptions and inferences turn out
to be wrong, as they sometimes do, I can generally identify reasons for this that are con-
sistent with my having the ability to perceive accurately and draw reasonable inferences.
I engage in dialogues with other people that conrm that generally they too perceive ac-
curately and draw reasonable inferences; and I nd that these dialogues themselves con-
tribute to my understanding of the world and my ability to draw reasonable inferences.
I have become aware of the vast knowledge and understanding of the world accumu-
lated by human beings, some of which I can understand and check for myself, and
some of which I take on trust because of the consensus of qualied people and/or the
successful application of this knowledge.
I am of course also aware of many defects of reason and of many examples of irratio-
nality and folly, both of myself and of other persons. However, as before, this does not
cause me seriously to doubt the rationality of human beings or of myself. Reasons for
irrationality can often be identied; and the ability to recognise mistakes and learn
from them is an important part of rationality.
So I have strong reasons for believing in the rationality of human beings including
myself, and in particular in that aspect of rationality that gives us the capacity to nd
out about the world and to make sound judgments about what to believe. These rea-
sons for belief in human rationality do not suggest a narrow view as to what consti-
tutes this rationality. Consistently with these reasons, I understand rationality as
extending to all those capabilities that contribute to reasonable human decision-mak-
ing. I certainly do not assume that rationality is restricted to any narrow category of
intellectual capabilities; and the arguments of this book will support the view that ra-
tionality includes not only perceptual and intellectual capabilities, but also emotional
capabilities that contribute (among other things) to reasonable evaluative judgments.
2.5. Rationality is to be understood as extending to all those capabilities that contribute

to reasonable human decision-making.
As well as giving human beings the capacity to nd out about the world and to make
reasonable decisions about what to believe, I contend that rationality also gives us the
capacity to make reasonable decisions about what to do. I believe with high condence:

2.6. I and other human beings are rational, particularly in having the ability to make
reasonable decisions about what to believe and what to do.
I dont think many people will seriously dispute this, although later in this chapter
I will consider an argument to the contrary effect.
We all live our lives in the belief that we have this ability, and our experiences of
our own decisions and those of other people generally conrm this. Of course ev-
eryone makes mistakes, and people do and believe silly things; but generally reasons
for this can be found, reasons that are consistent with the existence, in most persons
at least, of the ability to make reasonable decisions. Most importantly, while it is
often appropriate to question the rationality of particular beliefs or ways of reason-
ing, the whole enterprise of formation and justication of beliefs must depend on ulti-
mate condence in rationality. Any intellectual enquiry must assume that those
engaged in the enquiry have the ability to make reasonable decisions about what to
believe. The rejection of this assumption would make all intellectual endeavours
pointless.
In saying this, I am not of course saying that what we believe always depends on deci-
sions we make. In many respects, we believe things without making any decisions to
believe them, and we may be quite unable to make a decision to believe otherwise: in
those cases, belief is not a matter of decision at all. What I am saying is that when there
are inconclusive reasons supporting a belief, and particularly when there are reasons
for and against a belief, we do make decisions as to what conclusion is to be drawn, and
thus as to what to believe about the particular matter in question. This happens often
in my work as a judge. I have to come to a decision as to what happened between the
parties to a case, and there is often conicting inconclusive evidence about this. When
I make my decision as to what is established by the evidence, I come to have beliefs
about the happening of various events, beliefs that I may hold with varying degrees of
assurance. I dont think it can be denied that these beliefs are the result of decisions
made by me, on the basis of the evidence I have heard.
Ultimately my condence in my beliefs must depend on my condence in my own
rationality; but then, my own rationality can also give me condence in the rationality
and reliability of other human beings. In relation to matters that I cannot perceive for
myself, or cannot understand, it may be very reasonable to believe things just because
they are asserted by those who appear to be in a position to know them, particularly if
there is a consensus among these persons.
2.7. My condence in my beliefs depends ultimately on my condence in my own ratio-

nality, but my rationality gives me good reason for condence in things asserted by those
in a position to know them, particularly where there is consensus.

It follows from what I have said that I am prepared to believe some things even though
I have not perceived them myself and/or do not understand the scientic theories or
other reasoning on the basis of which inferences are drawn from observations. But
whether or not I am prepared to believe anything in those circumstances depends on a
prior judgment by me that it is reasonable to do so, having regard particularly to my
assessment of the reliability of the persons reporting what they have observed and/or
asserting what inferences can be drawn, and of the extent to which those reports and
assertions are compatible with my other beliefs and with reports and assertions made
by other people.
For the most part, I believe it is reasonable to accept what is asserted by scientists about
matters within their eld of expertise, especially where there are no signicant dissenting
views. It is widely accepted that the scientic method of experimentation is a most reli-
able method of nding out about the world and making sound judgments about what to
believe. This method proposes that one should formulate a reasonable hypothesis for
testing, and devise repeatable tests that will produce a predicted result if the hypothesis is
true and may not produce that result if the hypothesis is false; so that such tests either
refute the hypothesis or give it some positive support. By repeated use of this procedure,
science has progressively built up a body of well-supported hypotheses.
I accept that this is a very reliable method for nding out about the world; but I con-
tend it is not the only way beliefs can have rational support. This method itself depends
on prior beliefs of the kind I am now discussing, and also on reasoning of the kind
I will discuss in the next chapter. Indeed, even within science itself conclusions are
reasonably drawn without experimentation, for example in relation to what may be
called historical science (including some work in geology, biology, and astronomy, as
well as palaeontology and archaeology).3 Many of the questions I consider in this book
go beyond what science can address by experimentation, but they can still I believe be
the object of rational enquiry and belief. So:
2.8. The scientic method of experimentation is a very reliable method for arriving at
well-supported beliefs, but it is not the only way beliefs can have rational support.
However, it is sometimes reasonable to question what is asserted by scientists, espe-

cially where it does not have strong experimental support. There are two particular
areas of science where I think it is reasonable to doubt what is asserted by many scien-
tists, despite their superior understanding of relevant scientic issues.
First, there is time. Many theoretical physicists assert on the basis of relativity theory
that the passage of time is an illusion and that the past and future exist (in other regions
3
See Cleland (2002).

of space-time) no differently from the present; and some even say that time travel to
the past is possible. And second, there is the human mind, and in particular conscious-
ness. Many scientists in the various elds of science relating to the brain assert that the
mind is no more than the brain, and that our conscious processes have no causative
role beyond that of associated physical processes. In chapters 7 and 8 I will give reasons
why I do not accept these views on trust, and in fact believe they are wrong.
FALLACI E S AN D B IAS E S
Before leaving my discussion of truth and rationality, I should acknowledge that there is
extensive psychological research (much of it based on pioneering work of Daniel Kahne-
man and Amos Tversky)4 showing that in some respects people generally reason rather
badly, in fairly predictable waysthat various cognitive illusions, biases, and fallacies
pervade human reasoning. I will briey discuss some of these results, and also look at an
example5 of philosophical arguments that have been advanced on the basis of these
results against ordinary human rationality, and even against truth as a cognitive virtue.
I will start with two well-known psychological experiments.6
The rst illustrates the conjunction fallacy: the failure to recognise that the likeli-
hood of a conjunction of events or states of affairs can never be greater than the likeli-
hood of any element of that conjunction. This is the famous case of Linda.
Linda is thirty-one years old, single, outspoken and very bright. She majored in
philosophy. As a student she was deeply concerned with issues of discrimination
and social justice, and also participated in antinuclear demonstrations.
Experiments were conducted in which subjects were asked to rank, in order of prob-
ability, a number of statements about Linda. Perhaps the most telling result concerned
142 subjects, asked to rank in order of probability the statements Linda is a bank teller
and Linda is a bank teller and is active in the feminist movement. Eighty-ve percent
ranked the latter as more likely.
The second is an example of a selection task. Subjects were presented with four cards like
those shown in gure 2.1, with one half of each card masked; and were asked to determine
which of the cards they needed to see fully in order to answer decisively whether or not it
was true that, for these cards, if there is a circle on the left there is a circle on the right.
In one experiment, only ve out of 128 university students gave the correct answer,
namely (a) and (d). The two most common wrong answers were that one must see (a)
and (c), and that one need only see (a).
4
A useful collection on this is Kahneman et al. (1982).
5
In Stich (1990).
6
Given in Stich (1990), 47.

(a) (b) (c) (d)
Figure 2.1 A selection task, with the masked part of each card shown in black
There are many more categories of pervasive error that have been researched: I will
give a list of some of them.7
1. Hindsight illusion: the tendency to overestimate what could have been anticipated
by others, to view what actually happened as inevitable, and to misremember ones
own contrary predictions.
2. Overcondence and underqualication: the tendency to be overcondent about ones
judgments, particularly in relatively difcult tasks and in areas where one is not
qualied.
3. False-consensus bias: the tendency to believe that ones own behaviour and responses
are typical and appropriate.
4. Framing bias: the tendency for logically inconsequential changes in how problems
are stated to dramatically affect preferences and choice.
5. Salience and vividness: the tendency to overvalue aspects of ones environment that
are salient or vivid.
6. Conrmatory evidence bias: the tendency to ignore information that would discon-
rm held beliefs or expectations.
Now this research does show that there are deciencies, fallacies, and biases in much
human reasoning. Undoubtedly human reasoning does not always, as good reasoning
should do, comply with all rules of logic (I will use that word here in a wide sense as
including probability theory and mathematics, where these are relevant). Nor does it
always take into account relevant biases.
However, the considerable literature on this research tends itself to overlook a vital
point, which I will discuss in the next chapter. It is impossible to conduct all reasoning
by using rules of logic, aided by observation. In much signicant reasoning, the most
important steps (both in selecting premises and in moving from premises to conclu-
sions) depend on our instinctive informal rationality. Use of rules of logic will help
avoid error, but in realistic situations these rules often cannot of themselves support
non-trivial conclusions.
7
Taken from Sharp (1995), 8693.

Thus while ordinary human rationality is fallible and corrigible, we must use it:
Certainly, once we move beyond what is directly yielded by observation and/or appli-
cation of rules of logic, we can rely only on our ordinary fallible human rationality in
order to assess the justication for our beliefs. (Indeed, although this is not necessary
for my point here, I would contend that even reliance on observation, and on rules of
logic, must ultimately be justied by ordinary human rationality.) The best we can do
is to seek to achieve the highest standard of human rationality, being careful to avoid
fallacies and biases, applying logic to avoid detectable errors, and subjecting our rea-
soning to scrutiny and debate. In the extensive discussions of psychological research
into the deciencies of human reasoning, it is rare indeed to nd any acknowledge-
ment of the plain fact that the researchers and authors themselves are relying ultimately
on their own ordinary human rationality, and are also appealing to the ordinary human
rationality of their audience.
I do not think this can reasonably be denied. It might be said that what we (and these
authors and researchers) are relying on are the reasons for the conclusions reached,
rather than on human rationality as such. But reasons have to be evaluated, and are
evaluated differently by different people; and each person, in evaluating reasons, can
only rely on his or her ordinary fallible human rationality.
A further consideration is that, even where ordinary human reasoning is shown to be
fallacious, examples such as the above tend to conrm that it is still worth attending to,
in order to see if it is in fact capturing some insights into realistic situations, which may
not be captured by the reasoning that demonstrates the fallacy. Thus, although the 85
percent of subjects erred in their response concerning Linda because they committed the
conjunction fallacy, they nevertheless made the reasonable assessment that a person with
Lindas background and character, who had a job as a bank teller, was likely to have an
active interest outside her employment, and that this interest was likely to involve femi-
nist issuesand that is the kind of assessment which is more relevant in ordinary life
than a ranking of probabilities of categories that has no bearing on any practical task.
All in all, the psychological case does not show that we should not value ordinary
human rationality, much less that we should disregard it, but only that we should be
critical and aware in our use of it. Thus, while I have condence in my own rationality
and that of other persons, I know that my rationality is fallible and far from ideal, and
that the same is true of other persons. Each person has strengths and weaknesses in his
or her capacity for nding out about the world and making sound judgments about
what to believe. Any attempt to nd out about the world and to make sound judgments
is assisted by taking account, so far as one can, of these strengths and weaknesses in
oneself and others, and of the fallacies and biases to which we all are subject; and also by
engaging in dialogue with others, respecting their views and arguments, and being ready
to have regard to these views and arguments in forming and revising ones own beliefs.

STICHS ARGU M E NT
Psychological demonstrations of fallacies in human reasoning have also been opposed

by arguments to the effect that there are conceptual constraints on how badly people
can reason, and that evolutionary considerations also place a substantial limit on irra-
tionality. Broadly, the conceptual argument is to the effect that people could not under-
stand what other people mean by their assertions unless there is a minimum shared
rationality; and the evolutionary argument is to the effect that natural selection must
have preferred cognitive systems that are rational and tend to produce true beliefs, over
systems that do not have these advantages.
In his book The Fragmentation of Reason,8 philosopher Stephen Stich mounts a philo-
sophical case against these arguments; and in so doing attacks ordinary human rationality
as a reliable basis for reasoning and beliefs, and even attacks truth as a cognitive virtue.
Stich shows that (unsurprisingly) the conceptual argument does not establish that
human beings are ideally rational, or even that their rationality could not be improved
by adopting quite different reasoning strategies: at best, it shows that human beings
have to be tolerably good at reasoning in the particular way in which human beings
happen to reasonwhich could be far from the ideal way of reasoning.
He also shows that natural selection in evolution suggests no more than that our
reasoning capacities are fairly well adapted to making decisions conducive to survival
and reproduction; and that it does not guarantee that these capacities are ideal, or even
that there may not be radically different kinds of rationality that are as good as or even
better than the ordinary human rationality which we enjoy.
So Stichs philosophical project does conrm that we should take seriously the errors
and fallacies of human reasoning demonstrated by science, in particular by psychologi-
cal research of the kind referred to above; and as noted there, I do not question this.
However, Stich goes further than this, and suggests that this psychological work and
his own arguments mean that we should not even give provisional credence to the
deliverances of ordinary human rationality, and that we certainly should not give ordi-
nary human rationality any special place in our cognitive endeavours. He gives the
following clear and succinct statement of his position:9
Some writers have been tempted by the Wittgensteinian idea that epistemic assess-
ments must come to an end with the criteria embedded in our ordinary concepts
of cognitive evaluation. But surely this is nonsense. Both our notions of epistemic
evaluation and (more important) our cognitive processes themselves can be evalu-
8
Stich (1990).
9
Stich (1990), 2021.

ated instrumentally. That is, they can be evaluated by how well they do at bringing
about states of aairs that people do typically valuestates of aairs like being able
to predict or control nature, or contributing to an interesting and fullling life.
Stich goes on to conclude that truth is not to be taken seriously as a cognitive

virtue.
In the book, Stich mounts a case for this position which, in its details, reads well and
persuasively; but I believe his arguments are seriously awed, because they miss a
bigger picture that should be apparent to self-critical human rationality. In
particular:
1. As I will argue in the next chapter, there are not and (I believe) never will be purely
formal and/or empirical criteria for good reasoningso in any ultimate assessment of
reasoning we have to fall back on some kind of instinctive informal rationality, albeit
one that has regard to our vulnerability to fallacies and biases and that is assisted by such
processes as peer review and debate. This is essentially what I understand to be the
Wittgensteinian idea that Stich dismisses as nonsense; but his own suggestion that there
be instrumental evaluation is put forward by him (along with all the contentions in his
book) for appraisal by his readers on the basis of arguments that appeal to just this
instinctive informal rationality. That is, Stich does not in his own arguments appeal to
instrumental evaluation of his ideas, but only to the general rationality of the reader.
2. The instrumental evaluation of cognitive processes that Stich advocates depends
upon assessments both of the efcacy of those processes to produce certain states of
affairs and of the values of such states of affairs and of alternatives. If Stich is to be
consistent, then these assessments themselves, and every step in these assessments,
have to be measured, not against the criterion of truth or even of rationality, but against
the very same criteria of efcacy and valuesand so on ad innitum, unless at some
stage there is an appeal to what can be accepted as true on the basis (at least in part) of
our instinctive informal rationality. If it be said that Stich is insisting on instrumental
evaluation only of notions of epistemic evaluation and of our cognitive processes, not
of every step in making these instrumental evaluations or of other more everyday
enquiries, then I would point to his rejection as nonsense of the idea that epistemic
assessments must come to an end with the criteria embedded in our ordinary concepts
of cognitive evaluation: this commits Stich to instrumental evaluation all the way
down, and Stich does not suggest that everyday enquiries are exempted from his stric-
tures and his proposal. And if it be said that Stich does not propose instrumental
assessment of values, but merely an appeal to values we happen to have, the problem is
not reduced: the questions of what are the values we happen to have, and indeed who
are the relevant we, must on his view be subject to the same instrumental evaluation,

with the same consequent innite regress. Stich does address arguments that his posi-
tion involves circularity (at 14549), but only to combat the suggestion that it involves
a bootstrap fallacy: he does not show how he can avoid the necessity of the ultimate
appeal to informal rationality that he dismisses as nonsense.
3. Stichs conclusion that truth is not to be taken seriously as a cognitive virtue is self-
refuting (in asserting that the truth or otherwise of this conclusion itself is not a matter
to be taken seriously). Stich would presumably say that he does not care whether his
conclusion is true or not, so long as it is appropriately conducive to states of affairs that
people value. But if he said this, I would nd it difcult to believe him: his whole book
seems directed towards making out a case for the truth of this and other propositions.
A review of the book quoted on the back cover praises it as having all the appearances
of an honest commitment to the pursuit of truth; and so it seems to meyet to be
consistent, Stich himself must deny this. Indeed, he must go so far as to say that he does
not care whether or not an assertion that his conclusion is conducive to states of affairs
that people value is true or not, so long as this assertion in turn is so conduciveand so
on ad innitum.
4. Stich considers the application of the concept of truth to beliefs; but he takes beliefs
to be brain states, thereby maximising the problems of specifying exactly what kind of
correspondence with reality is required for truth. However, consistently with my ear-
lier discussion, I say the central application of the concept of truth is to assertions (or
beliefs considered as assertions); that an assertion is true if what is asserted is fact
(something that actually exists and/or existed and/or occurs and/or occurred); so that
truth is indeed correspondence with reality, with the required correspondence being
one of assertion or meaning.
My approach makes truth dependent inter alia upon the adequacy of the vehicle of
assertion, generally a natural language; and as noted above, where (as with many philo-
sophical issues) a statement concerns a matter remote from the practical situations
that were the main concerns of our language as it evolved, our language may be less
than fully adequate to express the reality of the matter, so that the application of the
concepts of truth and falsity can become problematic. None of this seems to me to
throw into question either our commonsense grasp of the concept of truth or its cen-
tral importance as a cognitive virtue.
A LEGAL EXAM PLE
As an illustration of all this, consider a criminal trial in which a person has been
charged with murder, and it can be accepted (on the basis of overwhelming evidence)
that the victim was killed by being shot once at close range by a person who was the
only other person in the vicinity.

First, suppose that the only defence is that the accused did not do this, it must have
been someone else. Now I believe it is clear beyond argument that the prosecutions
assertion that the accused shot the victim is either true or not true; that the proper task
of the court is to try to come to a correct decision, that is, a decision which makes a true
assertion about what happened; and that the legal obligation imposed on witnesses to
try honestly to give true evidence is appropriate and important. I believe it would be
absurd to suggest that truth was not the primary cognitive virtue here, or that its mean-
ing here was obscure; and particularly absurd to suggest that the court or witnesses
should rather be concerned to make assertions conducive to states of affairs that people
value.
Next, suppose instead that the accused raises self-defence, claiming that he be-
lieved it was necessary to shoot the victim to save his own life. This defence requires
consideration of whether the accused believed that the measure adopted to defend
himself, in this case shooting, was necessary, and of whether this measure was rea-
sonable, having regard to the nature of the threat. The competing assertions of pros-
ecution and defence in this case thus may involve not just plain matters of
unadulterated fact but also substantial matters of interpretation and evaluation; and
in those circumstances it may possibly be that the facts could fall within a grey area
such that the prosecutions assertion (that the shooting was not in self-defence) is
neither absolutely true nor absolutely not true; so that within that grey area, one
could say that the assertion had a degree of truth only. Even then however, if the facts
are on either side of this grey area, the prosecutions assertion will be true or not
true; and there will be a conviction if the court is satised beyond reasonable doubt
that the shooting was not in self-defence, that is, that the facts are on the prosecu-
tions side of the grey area.
Finally, suppose the defence is insanity; that is, according to current New South
Wales law, that the shooting occurred when the accused, by reason of a disease of the
mind, did not know what he was doing or did not know it was wrong. Now the defence
involves not just substantial matters of interpretation and evaluation, but also the ap-
plication of language to subject matter for which it may not be completely adequate.
That is, there are real questions whether the expressions disease of the mind, did not
know what he was doing and did not know it was wrong are accurate expressions of
the reality to which they are being applied. So again the facts may fall within a grey
area, within which the assertion of the accused person that this was a case of insanity
is neither absolutely true nor absolutely not true; but if the facts are outside this grey
area, then, I believe, notwithstanding any deciencies of language, the accused persons
assertion will be true or not true.
In each of the alternatives, despite indeterminacies of language, truth is the crucial
cognitive virtue; and it has to be sought through ordinary human rationality.

So my contention is that Stichs book, with its rejection of both truth and ordinary
human rationality, is itself an illustration of the pitfalls of not taking truth and human
rationality seriously enough. And I conclude:
2.9. Rationality is fallible, and needs to be subject to scrutiny and debate in order to
minimise errors, fallacies and biases; but ultimate reliance on ordinary human rationality
is justied and is necessary for all intellectual endeavours.
COR E ASS E RTION S AB OUT TR UTH AN D RATIONALITY
2.1. Truth is correspondence with reality by way of meaning; that is, an assertion is true
if what it asserts is fact or reality.
2.2. Where assertions are about ideas or interpretations or evaluations, or where lan-
guage is inadequate to express the reality under consideration, truth may be a matter of
degree; and sometimes reality can best be expressed by poetry or metaphor, or even by a
number of inconsistent metaphors.
2.3. Developing understanding may extend the range of language, so that it can better
express more of reality.
2.4. Relativistic views of truth must be rejected because they are self-refuting, and make
rational appraisal of beliefs impossible.
2.5. Rationality is to be understood as extending to all those capabilities that contribute
to reasonable human decision-making.
2.6. I and other human beings are rational, particularly in having the ability to make
reasonable decisions about what to believe and what to do.
2.7. My condence in my beliefs depends ultimately on my condence in my own ratio-
nality, but my rationality gives me good reason for condence in things asserted by those
in a position to know them, particularly where there is consensus.
2.8. The scientic method of experimentation is a very reliable method for arriving at
well-supported beliefs, but it is not the only way beliefs can have rational support.
2.9. Rationality is fallible, and needs to be subject to scrutiny and debate in order to
minimise errors, fallacies and biases; but ultimate reliance on ordinary human rationality
is justied and is necessary for all intellectual endeavours.

3
Plausible Reasoning
FOR MAL AN D I N FOR MAL R EASON I NG
I have set out reasons why I believe in ordinary human rationality, and why I believe
that this rationality is a necessary basis (and assumption) of all intellectual endeavours,
including philosophy and science. In this chapter, I argue there are not and never will
be purely formal criteria for good reasoningso that in any ultimate assessment of
reasoning and beliefs we have to fall back on some kind of instinctive informal ratio-
nality. This is a precursor to my argument in later chapters that our rationality depends
at least in part on conscious processes that are neither random nor wholly determined
by rules of any kind.
An important aspect of our rationality is that we reach and justify conclusions as to
what to believe and what to do on the basis of reasons, some of which serve as the
premises or data on which our conclusions are based.
In some cases, the relationship between a conclusion and the premises or data can be
expressed in terms of precise and conclusive rules; so that the conclusion can be reached
and justied mechanically, in the sense that it can be reached and justied as deter-
mined exactly by engagement of the rules with the premises or data which, along with
the rules, constitute the reasons for the conclusion.
This kind of reasoning is called formal reasoning or algorithmic reasoning. The most
obvious examples of formal reasoning are to be found in the elds of logic, probability
theory, mathematics and articial intelligence (although I do not of course suggest that
all reasoning in these elds is formal reasoning). Given the premises or data, no ele-
ment of judgment is involved in this kind of reasoning; and although the rules in ques-
tion here are rules of inference or computational rules rather than the laws of nature
investigated by physicists, we know that machines, whose operations are for practical
purposes wholly determined by laws of nature, can be constructed in such a way as to
give effect to these rules of inference. Indeed, all the operations of computers can be
considered as being in a sense processes of formal or algorithmic reasoning.
However, most human reasoning is not of this type, on the surface at least. Most
human reasoning is not overtly algorithmic: it does not overtly proceed precisely as
37
determined by rules of logic and/or probability and/or mathematics, or any other rules
that could be incorporated into a computer program. When we are trying to make a
reasonable decision as to what to believe or what to do, very often the reasons we see
for and against alternative beliefs or actions are inconclusive, and there is an apparent
gap between reasons on the one hand and decisions about what to believe and what to
do on the other.1 A decision then cannot be made by overt mechanical application of
precise and conclusive rules, but rather requires the exercise of reasonable albeit fallible
judgment. This kind of reasoning is called informal reasoning or plausible reasoning;
and most of the reasoning we encounter in everyday life, as well as in scientic and
philosophical writings, is of this kind.
To the extent to which there are rules operating overtly in this kind of reasoning, they
are generally not the precise conclusive rules of formal reasoning, but rather are rules
that apply ceteris paribus (other things being equal) and thus can give way to opposing
considerationswith no exact specication of what opposing considerations would
be sufcient to defeat them. Such rules are one example of what have been called2 the
soft constraints of factors relevant to human decision-making.
Clear examples of plausible reasoning supporting decisions as to what to believe can
be found in the judgments given by judges in deciding the facts of court cases they have
heard. Very often, there are contemporary documents that to some extent record what
happened, and there is sworn oral evidence given by witnesses in court. There may be
conicts between oral evidence given by different witnesses, and some or all of this evi-
dence may conict to a greater or lesser extent with contemporary records. Opposing
accounts of what people did may appear to conform to how one would expect people
to behave in the circumstances in question, or to deviate to a greater or lesser extent
from such expectations. The behaviour of the witnesses in giving their evidence may
give an impression of honesty or dishonesty, or of good or poor memory. Evidentiary
conicts of these kinds are not, and I say cannot be, resolved by explicit formal reason-
ing. The conicting considerations are inconclusive and of different kinds, and there is
no way in which a judge can overtly apply precise rules of formal reasoning to them in
such a way as to determine a conclusion. And yet, I contend, reasonable albeit fallible
decisions are made.
Much the same applies to decisions as to what to do. These decisions may require
decisions as to what to believe, but they may also require resolution of motivational
reasons, that is, a determination as to which desire or inclination is to be followed.
David Hume contended that reason has a very limited role in this latter process, being
limited to (1) determining beliefs about facts on which desires may be based and (2)
1
Cf. Hodgson (1999), 2056, and Searle (2001), 1317.
2
Horgan and Tienson (2006).

adjusting means to desired ends: he contended that nothing can oppose or retard the
inuence of a passion, but a contrary impulse, and that what prevails is always the
strongest passion. However, that contention involves the question-begging assumption
that passions or desires, like forces in Newtonian physics, are commensurable and
operate automatically, so that there can be a strongest passion that prevails, without
the intervention of any reasoning process.
If one applies Humes assumption to consciously addressed decisions about what to
do, made on the basis of consciously felt desires, then this assumption is plainly false.
There is no common scale on which (say) a feeling of hunger can be consciously and
overtly measured against a feeling of obligation to carry out a promised task, so as to
establish which is the stronger. If desires such as these conict, the outcome is deter-
mined not by any measurable preponderance of one desire over the other, but seem-
ingly by reasoning to a decision that takes account of the different characters of the
conicting desires. Humes assertion could be true only if applied not to consciously
addressed decisions what to do, made on the basis of consciously felt desires, but
rather to some process that is supposed to determine our actions otherwise than by
decisions made on that basis: desires must be understood as referring, not to con-
sciously felt desires taken into account in consciously addressed decisions, but to
something else operating mechanistically outside of consciousness. Hume did not
support such a view by any argument; and in this and following chapters, I will be
arguing against it.
Thus, I say that decisions as to what to believe or what to do, resolving inconclu-
sive reasons, cannot be made by overt calculation or computation or any other way
involving the overt mechanical application of conclusive rules: decisions of this
kind can only be made by the exercise of reasonable albeit fallible judgment. Such
plausible reasoning is inconclusive, but may still be rational; and I will be suggest-
ing that the need for it could not be eliminated even by perfect knowledge of and
ability to use all the rules of good reasoning, or by reliance on the scientic
method.
3.1. An important part of our rationality is the ability to engage in plausible reasoning,
in which premises or data do not entail conclusions by virtue of applicable rules but rather
support them as a matter of reasonable albeit fallible judgment.
Plainly, plausible reasoning is required for many matters that science cannot address;
but also, importantly, the scientic method itself depends heavily on plausible reason-
ing. It depends on plausible reasoning at least in (1) the formulation of hypotheses to
be tested, (2) the devising of experiments to test them, and (3) the selection of which
unrefuted hypotheses should be provisionally accepted (because although experiments
Plausible Reasoning 39
can refute general assertions about the world, they cannot give them positive support
without the aid of plausible reasoning).
The second of these three areas requiring plausible reasoning may be considered as
being distinct from the other two, as it is concerned with practicalities and techniques
more than with questions of what to believe. The rst and third of them are more di-
rectly concerned with questions of what to believe, and are interdependent: they give
rise to the problem of induction, which I will consider in the next section. I will suggest
they mean that one cannot reasonably be any more condent about the truth of any
general assertion about the world than one is about the reliability of the plausible rea-
soning necessary to support it.
So I say:
3.2. If plausible reasoning could not support reasonable decisions about what to believe,
science would be impossible.
I N DUCTION
One important aspect of plausible reasoning is the ability to reason by induction, that
is, the process of supporting general statements, such as hypotheses or theories stating
laws of nature, by reference to particular or singular statements, such as reports of
observations or experiments.
If one seeks to justify the assertion All ravens are black by reference to numerous
observations of black ravens and no observations of ravens that are not black, that is
an example of induction or inductive inference. Most actual uses of induction are
more complex than this, but they all involve the same basic process. The experimental
or observational evidence for any scientic hypothesis or theory, if expressed as state-
ments, comprises a large number of particular or singular statements, which are par-
ticular instances that follow from the hypothesis or theory.
However, it is not merely experimental or observational evidence that may make it
reasonable to believe that a hypothesis or theory is true or approximately true. For one
thing, there are other criteria for selecting hypotheses or theories that may reasonably
be accepted, such as the criteria of simplicity, explanatory content, and coherence with
other theories. These criteria are the subject not of denite and conclusive rules that
could be used in formal reasoning, but rather of ceteris paribus rules or soft con-
straints appropriate for the reasonable but fallible judgments of plausible reasoning.
The application of criteria like these is required for the formulation of hypotheses to
be tested, the rst of the three areas referred to above where the scientic method de-
pends on plausible reasoning (which may also be considered as one type of abductive
reasoning, referred to later). As pointed out by Hilary Putnam, selection of that kind is

necessary because at any given time innitely many mutually incompatible hypotheses
are each compatible with any nite amount of data.3
A striking illustration of this underdetermination of theory by data is given by Nelson
Goodmans new riddle of induction. Goodman4 notes that our past experience that all
emeralds we have observed are green supports the hypothesis that all emeralds are
green. He then supposes there is a predicate grue, where a thing is grue either if it is
examined before AD 2000 and is green, or if it is not so examined and is blue. Then, in
1990, the same evidence would equally support both the hypothesis All emeralds are
grue and the hypothesis All emeralds are green. Now, many reasons can be given why it
would be rational to believe the latter hypothesis and not rational to believe the former
hypothesis; but Goodmans point is that it is not possible to identify denite rules the
mechanical application of which could determine those predicates (like green) that can
reasonably be used as the basis of induction, and those (like grue) that cannot.
Thus the scientic method requires that criteria like simplicity, explanatory content
and coherence with other theories be applied, using plausible reasoning, in order to
formulate hypotheses to be tested.
Then, if a scientist (having applied such criteria) comes up with a hypothesis that if X
occurs, then Y also occurs, he or she can devise experiments where particular cases of X
occur, and it can be observed whether or not Y then occurs. If Y does not occur, and if the
experiment is properly conducted, then the hypothesis is refuted. If there are many occa-
sions when Y does occur and none when it does not, then the hypothesis is not refuted
and indeed has some positive support. The question then is, how much support?
The third of the areas mentioned above, the selection of which unrefuted hypotheses
should be provisionally accepted, depends both on the selection of reasonable hypoth-
eses to be tested and on the assessment of the strength of the support given by experi-
mental conrmation. And just as there are no rules the mechanical application of
which determines those predicates that can reasonably be used as the basis of induc-
tion and those that cannot, there are also no rules the mechanical application of which
determines the strength of the support given by experimental conrmation.
One illustration of this is given by Carl Hempels paradox of conrmation. Hempel
points out5 that the statement All ravens are black is formally equivalent to the state-
ment All non-black things are non-ravens; and that therefore anything that conrms
the latter statement must also conrm the former statement. A red pen is an instance of
something that is non-black and a non-raven, and the true statement about such a pen
that this red thing is a non-raven would conrm All non-black things are non-ravens
and thus would conrm All ravens are blackwhich seems absurd. Now, as with
3
Putnam (1979), 352.
4
Goodman (1965), chap. 3.
5
In Hempel (1965), 1120.
Goodmans new riddle, many reasons can be given why it is reasonable to regard the
observation of a black raven as giving some support to the hypothesis All ravens are
black but the observation of a red pen giving no support to it; but Hempels point is
that it is not possible to identify rules the mechanical application of which determines
when there is conrmation or what is the strength of that conrmation.
All the above is additional to another problem of induction identied by David Hume
in the eighteenth century, namely that it depends on an assumption that the future will
resemble the past.6 A similar point was made by Karl Popper7 in terms of an innite
regress: if there is a principle of induction that can, in combination with particular
statements, justify general statements, then this principle must itself be a general state-
ment, requiring a principle of induction to justify it, giving rise to an innite regress.
Both Hume and Popper also applied their argument to inferences from particular
statements to the probability of general statements. As Popper points out (par. 81), a
statement that a hypothesis is probably true cannot be derived from particular state-
ments unless the latter are combined with some principle of induction which is itself
accepted as true or probably true; so there is still an innite regress.
Popper suggested one way of dealing with this problem, namely to say that the
method of science is that of conjecture and refutation; and that science can justify the
rejection of general statements about the world, but cannot give even the slightest posi-
tive support to the truth of such statements.
One difculty with this approach is the theory-laden character of particular observa-
tions: it is now generally accepted that we see (and hear and feel) the world as we do in
part because of our beliefs about the world and our consequent expectations.
Apparently, according to Popper, we have to take general beliefs on faith unless and
until they are refuted; but Popper offers no way of choosing what to believe from the
vast range of conicting unrefuted general statements about the world. And this in
turn would suggest that, in so far as our particular observations depend on relevant
general beliefs, we cannot rely even on our particular observations.
Another difculty is that Popper does not show how or why his own general asser-
tions, including those about the scientic method, should be given any higher status
than that of an unrefuted scientic conjecture; and thus as assertions whose truth has
(and can be given) no positive support whatsoever.
I think most of us accept, contrary to Popper, that science gives us good reason to
believe that certain general statements about the world are probably at least approxi-
mately true; and that the progress of science delivers closer approximations and greater
probabilities. But the arguments about induction show that this can only be by way of
6
Hume (1748), 57.
7
Popper (1959), 29.

plausible reasoning that cannot be formalised, that is, reduced to or explained in terms
of overt compliance with rules for good reasoning. Induction, like other forms of plau-
sible reasoning, involves an element of judgment that does not depend on overt ap-
plication of logical rules or any other precise conclusive rules.
So:
3.3. Induction does provide good grounds for believing that some general statements
about the world are probably at least approximately true, but only by way of informal
plausible reasoning.
3.4. Induction and other forms of plausible reasoning cannot be reduced to any kind of
algorithmic process overtly using discovered or invented rules for good reasoning.
Although in this discussion I have focussed on induction, similar arguments apply to

other kinds of plausible reasoning, such as reasoning by abduction or reasoning by
analogy. Abduction, also called inference to the best explanation, is the selection of the
hypothesis that best explains one or more particular observations. Analogy uses known
similarities between things and/or events and/or states of affairs (say, that two things
are both ravens) to justify conclusions about further similarities between them (that
they are both black). As with induction, there has been no identication of rules for
good reasoning, the mechanical application of which can determine whether particu-
lar cases of reasoning by abduction or by analogy are reasonable or unreasonable: these
types of plausible reasoning also involve an element of judgment that cannot be wholly
explained by reference to logical rules or any other precise conclusive rules.
BAYE S TH EOR E M
Notwithstanding arguments of the kind outlined above, over the centuries there have
been arguments to the effect that plausible reasoning can be formalised. Indeed, some
people have claimed that plausible reasoning must be capable of being formalised if it
is to be relied on: its been said that if a conclusion does not follow from premises or
data on the basis of rules of logic or other rules of good reasoning, then it is not sup-
ported by them at all.
The basis of the most signicant attempts to formalise plausible reasoning, whether
by way of induction or otherwise, is Bayes Theorem, an equation of probability theory
devised in the eighteenth century by the Reverend Thomas Bayes. Bayes Theorem is
soundly based on plausible axioms of probability;8 and it concerns the effect of par-
ticular pieces of evidence on the probability of a hypothesis.
8
See Robertson and Vignaux (1993), and appendix A to this book.
One has to start with three prior probabilities, that is, probabilities prior to determin-
ing the effect of the evidence: the prior probability of the hypothesis itself; and the prior
probabilities of the piece of evidence in each of two supposed circumstancesrst, if
the hypothesis is true, and second, if the hypothesis is not true. Then, the theorem gives
a formula according to which the evidence changes the probability of the hypothesis.
Although the theorem can be expressed in various ways, I think it is easiest to under-
stand and apply if it is expressed in terms of odds (rather than direct probabilities) and
what have been called likelihood ratios.
Let us suppose that, prior to taking account of some evidence, the probability that a
hypothesis is true is 0.6 and the probability that it is not true is 0.4, so that the odds
that it is true rather than not true prior to taking account of the additional evidence,
called the prior odds, are 6:4. Let us suppose that there is then evidence, the probability
of which if the hypothesis is true is 0.8, and the probability of which if the hypothesis
is not true is 0.6. (Note that these probabilities need not add up to 1: the evidence could
in fact be very probable or improbable whether the hypothesis is true or not.) This evi-
dence gives what is called a likelihood ratio, in favour of the hypothesis being true, of
8/6. The odds version of Bayes Theorem then says that you multiply the prior odds by
the likelihood ratio, and this will then give you the odds that take into account the evi-
dence, called the posterior odds. In this case, the posterior odds are 6:4 8/6, that is, 8:4
or 2:1. Odds of 2:1 give a probability of 2/3 or 0.67 (to two decimal places).
In this case, then, the evidence has increased the probability that the hypothesis is
true from 0.6 to 0.67.
So, Bayes Theorem can be stated as follows:
Posterior odds = prior odds likelihood ratio, where:
1. Odds are a measure of the comparative probabilities of a hypothesis and its negation
such that odds of x:y for a hypothesis give probabilities of x/ (x + y) for the hypoth-
esis and y/ (x + y) for its negation.
2. Prior odds are the odds as between the hypothesis and its negation prior to taking
into account the further evidence.
3. Posterior odds are the odds as between the hypothesis and its negation after taking
into account the further evidence.
4. Likelihood ratio is the ratio between the probability of the further evidence if the hypoth-
esis is true and the probability of the further evidence if the hypothesis is not true.
Bayes Theorem can be applied successively as further pieces of evidence are consid-
ered: the order of application is immaterial. The posterior odds after each application
of the theorem become the prior odds for the next.

Very often, at least in scientic experiments, the probability of the piece of evidence
if the hypothesis is true is clear enough: it is probability 1 or certainty. However, the
other two prior probabilities that one needs in order to apply Bayes Theorem, namely,
the prior probability of the hypothesis and the prior probability of the piece of evi-
dence if the hypothesis is not true, are generally doubtful. If these prior probabilities
are not somehow established, then one has to either estimate them or arbitrarily set
them at 0.5. In real situations, it is rarely rational to set probabilities arbitrarily at 0.5;
so we are left with the necessity of having to estimate at least two prior probabilities in
order to apply Bayes Theorem.
The problem with this is that we generally cannot make such estimates with the con-
dence of accuracy that would be required if we were to have condence in the conclu-
sion given by Bayes Theorem. In fact, one can sometimes estimate the probability
sought by applying Bayes Theorem with just as much condence of accuracy as for
estimates of the prior probabilities. In general, an informal estimating of probabilities
is a pre-condition for applying Bayes Theorem; so that Bayes Theorem cannot avoid
the need to rely on our informal rationality. What this theorem and its derivatives can
do is to promote consistency among various estimates of probability: otherwise, it
depends on, and cannot assist in promoting, the reasonableness of those estimates, and
thus cannot eliminate the need for judgments of reasonableness not determined by
rules of reasoning.
3.5. The application of Bayes Theorem depends on the reasonableness of estimates of

prior probabilities, and thus the theorem cannot eliminate the need for informal
rationality.
I LLUSTRATION OF BAYE S TH EOR E M
All this can usefully be illustrated by an example, based upon a simplied version of
the facts of a decided court case, namely the New South Wales Court of Appeal deci-
sion in Stewart v. Ng [1999] NSWCA 387.
In that case, the defendant performed a tubal ligation on the plaintiff in 1993. In 1995,
the plaintiff became pregnant. She sued the defendant, claiming inter alia that the op-
eration was performed negligently. There was evidence, which we may assume the
Court accepted, that the failure rate for properly conducted operations was about 1 in
350. There was also evidence that the defendant was a distinguished surgeon, and evi-
dence from the defendant that the operation was performed correctly. The trial judge
found for the defendant, and this decision was upheld on appeal. Although Bayes
Theorem was not explicitly applied by either court, the facts can be used to illustrate
how the theorem works.
The hypothesis to be tested was that the operation was not performed properly, and the
alternative hypothesis was that it was performed properly. Because the defendant was
a distinguished surgeon, one could start with the view that, if nothing else whatever
was known about the operation, there was a high probability that it was performed
properly. One might suppose that a distinguished surgeon would get such an opera-
tion right at least 99 times out of 100, and so might take the prior odds against the
hypothesis to be 1:99.
One piece of evidence about the particular operation was that the plaintiff became
pregnant after two years. The probability of this happening if the operation was not
performed properly we might take to be high, let us say 9/10 or 0.9. The probability of
it happening if the operation was performed properly was very low, the gure given in
evidence being 1/350, or about .00286. The likelihood ratio provided by this evidence
in favour of the hypothesis was thus 9/10 1/350, or 315/1. Multiplying the prior odds
by the likelihood ratio, we get odds of 315:99 in favour of the hypothesis. The probabil-
ity of the hypothesis, taking into account just this rst piece of evidence, thus becomes
315/414 or about 0.76.
Accordingly, if there were no evidence whatsoever about the particular operation,
perhaps because no one had any recollection of it, one might come to the conclusion
that it was more probable than not that the operation was not performed properly, so
that the civil onus of proof was discharged in favour of the plaintiff.
But there was also the evidence from the defendant himself that this particular op-
eration was performed properly.
I note that, if this evidence had been based solely on the defendants usual practice,
rather than specic recollection of the particular occasion, this may have had little
signicance, because the defendants usual practice may already have been taken into
account in assuming the prior odds of 1:99 against. However, if the Court found the
evidence of practice impressive, this might possibly justify adjusting the prior odds;
and if the prior odds were adjusted to something approximating to 1:315, the Court
might nd the civil onus not discharged.
In this case, however, the evidence of the defendant was that he did specically recall
the particular operation, and that he recalled demonstrating all the steps in it to his
Table 3.1 Effect of rst piece of evidence
Prior odds against improper operation are 1:99

Likelihood ratio provided by evidence (occurrence of pregnancy) in favour of improper
operation is 315/1.
Posterior odds = prior odds likelihood ratio
= 1:99 315/1 = 315:99.

assistant, to whom he was teaching the operation; so that this evidence could carry
independent weight.
One might think that the probability that a busy surgeon would remember a particu-
lar operation some years later to be quite low; although if it was one of a small number
demonstrated to a particular assistant, this could increase that probability.
However, we may take it that the probability of such evidence being that the opera-
tion was performed properly would be higher if the operation had in fact been prop-
erly performed than if it had not been. Let us say that the court was impressed by the
defendant, and considered it very unlikely that he was lying or mistaken, and estimated
that the evidence that the operation was recalled and was performed properly to be
four times more likely if the operation had been performed properly than if it had not
been.
(This might be worked out in more detail, as follows. Assume that the probability of
recollection is 1/5, and that the probability that this witness would give false evidence
to his advantage is 1/16 and to his disadvantage is nil. Then, the probability of the evi-
dence that the operation was done properly, if the operation was in fact done properly,
is the sum of 4/5 1/16 [he did not recollect but said he did] and 1/5 1 [he did recol-
lect and truthfully said he did it properly], that is, 1/4. The probability of this evidence,
if the operation was not done properly, is the sum of 4/5 1/16 [as before] and 1/5 1/16
[he recalled doing it improperly but said he did it properly], that is, 1/16. That gives the
suggested likelihood ratio against the improper operation hypothesis of 1/4.)
This evidence then provides a likelihood ratio against the hypothesis of 1/4. Applying
this ratio to the previous odds of 315:99, which have become the prior odds for this
step, we get the nal posterior odds of 315:396. The probability of the hypothesis thus
is reduced to 315/711 or 0.44.
Thus, on the whole of the evidence assessed on this basis, the civil onus of proof of
the balance of probabilities is not discharged.
This example both illustrates the limitations of Bayes Theorem, and also its
benets.
Starting with the limitations, plainly the theorem requires estimates or guesses in
relation to at least some of the numbers used. The statistic of a failure rate for properly
Table 3.2 Effect of second piece of evidence
Prior odds for improper operation are now 315:99.

Likelihood ratio provided by evidence (eyewitness account) against improper operation
is 1/4.
Posterior odds = prior odds likelihood ratio
= 315:99 1/4 = 315:396.
conducted operations of about 1 in 350 was actually given in evidence in the case, and
presumably had some reasonably sound basis. However, the other numbers used by me
(the initial odds of improper operation of 1:99, the probability of pregnancy of 9/10 for
an operation not performed properly, and the likelihood ratio of 1/4 provided by the
eyewitness account) are no more than educated guesses; and the pertinent comment
can be made that a judge could have no more condence in guesses like these than in
an overall judgment as to whether it was shown, more probably than not, that the op-
eration was not properly performed.
This illustrates, as suggested earlier, (1) that an informal estimating of prior proba-
bilities is a pre-condition for applying Bayes Theorem; (2) that we generally cannot
make estimates of prior probabilities with the condence of accuracy that would be
required if we were to have condence in the conclusion given by Bayes Theorem; and
thus (3) that Bayes Theorem cannot avoid the need to rely on our informal
rationality.
The example also illustrates what I think are two benets of this kind of reasoning.
First, even if, as is sometimes the case, the educated guesses concerning prior odds
and likelihood ratios are on no sounder ground than a judgment as to the overall
result, the process of applying Bayes Theorem will at least disclose whether or not
there is consistency between all ones educated guesses and ones judgment. If inconsis-
tency were disclosed, then it would be necessary to reconsider each of the guesses and
also the judgment, because at least one of them must be wrong; so that an error may
thus be discovered and corrected.
Second, use of Bayes Theorem is a way of ensuring that one is not misled by spuri-
ous arguments based on statistics. In this case, it might have been contended that the
failure rate statistics of 1 in 350 were more than sufcient to discharge a civil onus of
proof. The process of applying Bayes Theorem will cause one to consider what other
material there is pertinent to the ultimate question for decision, and the theorem pro-
vides a framework for bringing all the material together.
3.6. Bayes Theorem can promote consistency among various estimates of probability,
and combat spurious arguments based on statistics.
LEVE LS OF CO GN ITIVE PRO CE SS E S
Arguments of the kind Ive briey set out here have been developed at length by
the philosopher Hilary Putnam, leading him to the following conclusion, with which
I concur:9
9
Putnam (1983), 198; Putnam (1981), 174200.

3.7. Human rationality cannot be formalised without formalising complete human per-
sonality, and possibly not even then.
I dont think the rst part of this statement should be controversial: certainly I have
not found any answer to Putnams case in support of it. But most scientists and scien-
tically minded philosophers would I believe dismiss the possibility raised in the
second part.
They maintain that the world, including human beings, changes over time either
precisely as determined by laws of nature, or else randomly within probability param-
eters precisely determined by laws of nature; and the processes that constitute rational
decision-making by human beings must proceed in the same way as other processes.
On this view, what gives rationality to these processes is that they occur in brains whose
structure and operations have been selected by trial and error over millions of years of
evolution, and thus have proved themselves to be conducive to the survival and repro-
duction of the organisms that had them; so that they can reasonably be considered as
being adapted to the formation of reasonable beliefs about the world and reasonable
decisions about what to do. Human rationality could then be formalised by sufcient
specication of the relevant properties of one or more highly rational human brains
and of the laws of nature that determine their processes, and/or (if, as seems a reason-
able assumption, these processes are of the nature of computational information-pro-
cessing) by sufcient specication of the computational systems instantiated by these
brains and the computational rules of these systems.
There have been important developments of this line of thought, based on David
Marrs identication of three levels of cognitive processes,10 which Marr named (1) the
computational level, (2) the algorithmic level, and (3) the implementational level.
The rst (top) level is that of the task given to and performed by the cognitive pro-
cess. In the case of a computer, it would include the input of a task into the computer,
and the computers display of the outcome of computation; while in the case of a
person, it would include the explicit reasoning of that person. Marrs naming of this
level as the computational level might be considered inappropriate, in that it is not the
level at which computational programs are applied. This top level could perhaps better
be called the overt level: it might be likened to a swan that is visible gliding across the
water, while the lower levels might be likened to the swans legs that are furiously pad-
dling out of sight beneath the surface. In the case of human cognition, this level can be
identied with the level of conscious reasoning.
The second (middle) level is that of algorithmic information processing, exem-
plified by the running of programs of computer software. It is the level at which
10
Marr (1982).
computational algorithms are put into effect, so as to process information in ac-
cordance with the rules of the programs being run. Marr himself, and also many
philosophers,11 have conceived of this level as operating as one or more digital
discrete-state systems, that is, systems having discrete successive well-defined rule-
determined states (like digital computers); and sometimes the notion of algorith-
mic information processing is considered as limited to systems of that kind.
However, some philosophers12 have contended that, if Marrs three-level analysis is
applied to human beings, the so-called algorithmic level could be operating as one
or more analog systems, that is, systems changing continuously as determined by
rules engaging with continuous variables (like, for example, the weather), rather
than having discrete successive states as with digital systems. Accordingly, if the
terminology of the algorithmic level is retained, in my view it needs to be under-
stood in a broad sense, as extending to analog systems of that kind.
The third (bottom) level is that of physical implementation, exemplied by the hard-
ware of computers and the physical brains of human beings, on which, on this ap-
proach, the middle-level programs run. For cognitive systems to work, the programs of
the middle level must be compatible with the hardware of the bottom level, and must
also be suited to performance of the tasks carried out at the top level.
This identication of three broad levels of cognitive systems does not of course ex-
clude the identication of further levels within these three levels. Within the bottom
level, for example, a distinction can be drawn between the level of microphysical pro-
cesses described by quantum mechanics, and the level of larger-scale physical events
such as the transition of a transistor from one of two states to the other of them, or the
ring or non-ring of a neuron of a human brain. Within the middle (algorithmic)
level, distinctions can be drawn between the level of machine language, that is, of pro-
grams in a code that correlates directly with events at the level of implementation; the
level of assembly language, that is, of programs using convenient coded abbreviations
to designate chunks of the code of machine language; and the level of compiler lan-
guage, that is, of programs in which words of a natural language are used to designate
the steps of a program.13 On some views of human cognition, it is at a level like that of
compiler language, within the algorithmic level of cognition, that representations of
things in the world are processed in accordance with computational rules.
On the basis of this three-level approach, it can be contended that even if, in human
reasoning, what happens at the top (overt) level cannot be formalised as a process that
complies strictly with precise exceptionless rules, this is nevertheless consistent with
11
Such as Jerry Fodor: see Fodor and Pylyshyn (1988).
12
Such as Terry Horgan: see Horgan and Tienson (1996).
13
See Hofstadter (1980), 29099.

processes at the middle level unfolding strictly as determined by computational rules,
and processes at the bottom level unfolding strictly as determined by laws of nature;
and that the effectiveness of human rationality is fully explained by the evolutionary
tests passed by the hardware of the bottom level and the computational rules of the
middle level.
Before leaving my brief account of this three-level approach to human cognitive pro-
cesses, I should mention two broad ways in which this approach has been developed.
On the one hand, there is what has been called the classical paradigm, which likens the
human mind to a digital computer or a multiplicity of digital computers; and on the
other hand, there is the connectionist paradigm (or connectionism), that seeks to ex-
plain human cognitive processes in terms of neural networks whose operations are
inuenced by varying strengths of connecting nodes of the network.
Connectionism is a theory concerning the implementational level of cognitive pro-
cesses (one of its attractions being that the human brain itself seems to be a kind of
connectionist system); and it has been argued that connectionism is consistent with
digital processes operating at the middle algorithmic level, because that level is indepen-
dent of the implementational level.14 However, while it is true that there is some degree
of independence of the algorithmic level from the implementational level, exemplied
by the ability of some computer programs to run on different computer hardware, it is
plain that the requirement that programs of the algorithmic level be compatible with
the implementational level does mean that the implementational level places constraints
on the algorithmic level.15 And one particular development of the connectionist ap-
proach, namely that of Terry Horgan and John Tienson,16 explicitly proposes that, in
human cognition, the middle (algorithmic) level be considered as operating not as one
or more discrete-state (digital) systems, but rather as a dynamical (analog) system.
There is a considerable literature on the relative merits of the classical paradigm and
the connectionist paradigm.
For example, Horgan and Tienson argue (in this respect, consistently with what I
have written above about plausible reasoning) that human cognition is too rich and
varied to be described by precise exceptionless rules applying to representations of
things in the world; and they go on to contend (1) that representational-level cognitive
processes rather conform to multiple soft constraints of ceteris paribus rules, (2) that
precise exceptionless rules operate only at sub-representational levels, and (3) that
human cognition is therefore best understood in terms of a dynamical system realised
in a neural network.
14
Fodor and Pylyshyn (1988).
15
Oaksford and Chater (1998), 29.
16
Horgan and Tienson (1996), (2006).
On the other hand, Jerry Fodor and Zenon Pylyshyn have argued17 to the effect that,
even if connectionism obtains at the level of implementation, human cognition must
be classical at the algorithmic level, because only classical digital architecture can ac-
commodate appropriately organised representations of the world. Daniel Dennett18
argues that results supporting the connectionist paradigm have actually been produced
by virtual neural networks simulated on standard digital computers, and that digital
approximations of analog connectionist systems have all the powers necessary for the
explanation of human cognitive processes. And Roger Penrose19 contends that digital
systems are more reliable, in that they do not require 100 percent accuracy in recording
and transmitting information (and thus are not subject to the instability and magni-
cation of inaccuracies to which analog systems are subject); and he too points out that
digital systems can simulate analog systems (including if necessary their instability and
inaccuracies) to any desired degree of approximation.
It is beyond the scope of this book to enter into the relative merits of the classical
paradigm and the connectionist paradigm; and indeed it may be that the operation of
the human brain is in fact neither exclusively digital nor exclusively analog. A common
feature of all these approaches is that where there are inconclusive reasons operating at
the top level of human cognitive processes, a single outcome is achieved by virtue of
rule-determined processes at lower levels, the relevant rules being computational rules
operating at the algorithmic level and laws of nature operating at the implementa-
tional level. According to these approaches, what may appear to be a conscious rational
judgment, resolving inconclusive reasons, is an expression of conclusive rule-deter-
mined processes at lower levels: the explanation as to why inconclusive reasons are
resolved in one way rather than another is provided by those rule-determined pro-
cesses, with no further efcacy in relation to that resolution being accorded to the
conscious processes at the top level. The argument of this book, however, is that it is
more reasonable to accept that conscious processes at the top level make a positive
contribution to the resolution of inconclusive reasons, a contribution that is not wholly
governed by rule-determined processes at lower levels. Thus I will contend that, while
human rationality can be partly explained in terms of evolution-selected rule-deter-
mined processes at the algorithmic and implementational levels, it cannot be fully ex-
plained in terms of rule-determined processes of any kind, whether they be analog or
digital or some combination of the two.
The main arguments of this book will focus on positive reasons for this view, rather
than upon purely negative arguments like that of the Haldane argument referred to in
17
Fodor and Pylyshyn (1988).
18
Dennett (2003), 1067.
19
Penrose (1989), 17074, 18182, 403.

the introduction. However, at this point I will raise two Haldane-type negative
considerations against there being an explanation of human plausible reasoning wholly
in terms of lower-level rule-determined processes.
First, such an explanation would introduce a vicious circle into the justication of
plausible reasoning. If we cannot rely on our plausible reasoning as the conscious ra-
tional non-algorithmic process it seems to be, and on associated feelings of assurance,
then any condence we could have in plausible reasoning would have to depend on the
belief that it is supported by computational processes whose reliability is assured by
the evolutionary tests they have passed. Yet this belief would itself have to depend on
extensive plausible reasoning, giving rise to a vicious circle.20
Second, it would mean that disagreements in matters of plausible reasoning could
not be addressed rationally. Such disagreements would arise from differences in un-
conscious processes that are ultimately inscrutable to conscious appraisal. Accordingly,
so long as identiable fallacies were avoided, there would be no rational basis on which
to address the question whether one process of plausible reasoning was preferable to
another.
One nal note. I think it is clear that formalisation of human plausible reasoning in
terms of rule-determined processes at lower levels, if achievable, would not give pre-
dictability of decision-making. If human brains and such computational systems as
they may instantiate were analog systems, then prediction of their development over
time would require 100 percent accuracy in specication of initial conditions, and also
solution of the many-body problem of classical physics; and even if they were wholly
digital systems, their sheer complexity would be such as to preclude predictability. So
it is not predictability that I am challenging, but rather the universality of rule-deter-
mination in rational processes.
COR E ASS E RTION S AB OUT PLAUS I B LE R EASON I NG
3.1. An important part of our rationality is the ability to engage in plausible reasoning,
in which premises or data do not entail conclusions by virtue of applicable rules but rather
support them as a matter of reasonable albeit fallible judgment.
3.2. If plausible reasoning could not support reasonable decisions about what to believe,
science would be impossible.
3.3. Induction does provide good grounds for believing that some general statements
about the world are probably at least approximately true, but only by way of informal
plausible reasoning.
20
Cf. Nagel (1997), chap. 7; Plantinga (1993), chap. 12.
3.4. Induction and other forms of plausible reasoning cannot be reduced to any kind of
3.5. The application of Bayes Theorem depends on the reasonableness of estimates of
prior probabilities, and thus the theorem cannot eliminate the need for informal
rationality.
3.6. Bayes Theorem can promote consistency among various estimates of probability,
and combat spurious arguments based on statistics.
3.7. Human rationality cannot be formalised without formalising complete human per-
sonality, and possibly not even then.

4
Consciousness and Decision-Making
The previous chapter showed (1) that an important part of our rationality is the ability
to engage in plausible reasoning, in which premises or data do not entail conclusions
by virtue of applicable rules but rather support them as a matter of reasonable albeit
fallible judgment; and (2) that plausible reasoning cannot be reduced to any kind of
However, there was left open the question whether plausible reasoning might never-
theless be achieved wholly by brain processes which unfold as determined by laws of
nature and/or computational rules, and which are rational because the structures sup-
porting these processes and any computational systems they instantiate have been se-
lected by evolutionary trial and error.
In this and the next two chapters, I argue against this possibility, contending that our
consciousness contributes to our rational decision-making in ways that are not wholly
determined by laws or rules of any kind.
I start in this chapter by saying something about consciousness, and identifying three
characteristic features of conscious processes that I will contend contribute together to
decision-making: namely, subjectivity, qualia and unity. Then I argue that conscious
experiences must have a role to play in our decision-making, and that this role is not
one that could be performed by rule-determined processes. In the following two chap-
ters, I give an account of just what I contend this role is.
DUAL AS PECTS
It is necessary for me rst to say a little more about consciousness and its general rela-
tionship to brain processes. This is another topic that has generated an enormous
amount of philosophical writing and controversy, to which I cannot do justice in this
chapter. However, it is important to my case to make it clear what my own position is
on some signicant issues.
It will be recalled that my most fundamental beliefs included beliefs that con-
scious experiences occur; that the actual experiences that give me certainty that
conscious experiences occur are experiences that I have, while other human beings
55
also have experiences of the same general kind; and that there is a world external to
the conscious experiences of individual human beings, features of which are acces-
sible in common by human beings and describable in language. Now, human brains
are a part of the world external to the conscious experiences of individual human
beings, so a question is raised as to the relationship between the brain of a person
(which is part of the objective external world) and the conscious experiences of that
person (which are not part of the objective external world).
Science has progressively explained more and more of what happens in terms of the
operation of laws of nature in the objective external world; and this has made it impos-
sible for me to believe in a dualism of substances, according to which human beings
have immaterial minds, distinct from their physical brains, which have conscious ex-
periences and make choices and initiate actions. (I elaborate on this in chapter 8.)
However, I believe that a living human brain is not just a physical object, but is rather
an entity with two distinct aspects, namely a physical or objective aspect as a brain, and
a mental or subjective aspect as a conscious mind: that is, I adopt a view that can be
classied as a dual aspect theory, within the classication outlined in chapter 1.
Thus I say that a living human brain is properly considered as being a brain-and-
mind, a physical-and-mental entity rather than a purely physical entity. I believe also
that the processes that constitute conscious experiences have both a physical or objec-
tive aspect as physical occurrences in the persons brain (occurrences that are features
of the objective external world), and a mental or subjective aspect as the persons
conscious experiences (occurrences that are not features of the objective external
world). And I believe that much the same is true of the brains of those non-human
animals that have conscious experiences.
Dual aspect views of this kind have been strongly challenged by various arguments
and points of view.
One general line of argument is to the effect that there is nothing special about the
distinction between conscious experiences and objective occurrences: while it is true
that some processes can be at once both conscious experiences and also physical occur-
rences in the persons brain, it is also true that other processes can be at once (for ex-
ample) both the digestion of food and also molecular occurrences in a persons
stomach. In both cases, it is said, there are just different levels of description of the
same thing, in the same sort of way that the same occurrences can be both global
changes in a physical object (such as increasing heat or transition from a solid to a
liquid state) and changes in the behaviour of molecules of the object (such as increas-
ing speed).1
1
Cf. Searle (1984), 2122.

My contention is that the distinction between conscious experiences and brain events
is more than just a distinction referable to different levels of description, like the dis-
tinction between global occurrences such as digestion, or the transition from a solid to
a liquid state, and their corresponding molecular occurrences. This is because in the
case of those distinctions, both elements of the distinction are features of the external
world, accessible in common to human beings generally, and in that sense objective
rather than being or depending upon conscious experiences of individual human
beings; whereas, consistently with what I wrote in chapter 1, I do not consider con-
scious experiences to be part of the external world in that sense, but rather to be subjec-
tive and accessible, as conscious experiences, only to the person having them.
I will say more about subjectivity later in this chapter, and about why I see the sub-
jectivity of conscious experiences as an important feature that distinguishes them from
features of the external world. It is not just that the content of conscious experiences
has a character such that there is something it is like to have them.2 In addition, I will
be contending that, in all cases of conscious experiences, (1) there is a subject that has
the experience, and (2) the experience is experienced only by this subject and is inter-
dependent with this subject.
The unavailability of one persons experience for experience by another person is,
I contend, of a completely different order from (say) any supposed unobservability of
neural events, or of abstract concepts such as the gaseousness of gases. As regards the
former, the difculties of observation are just practical difculties, that are to some
extent progressively being overcome; and in any event, such difculties as there are
apply to the person whose neural events they are as much as to anyone else, which is
not the case with conscious experiences. As regards the latter, abstract concepts are not
observable by anyone in the same way as things or occurrences; whereas conscious
experiences are experienced as occurrences (and in that sense observed) by the subject
that has them, but not by other people.
I should also make it clear that at this stage I am not asserting the position, suggested by
Thomas Nagel, that full understanding of relevant objective facts could not carry with it a
full understanding of conscious experiences, including their what it is like character: I am
not at this stage of my argument excluding the possibility3 that facts of conscious experi-
ence are necessitated by objective physical factsthat given the relevant objective physical
facts, conscious experiences could not be other than they are. But my ultimate position,
which I seek to make good later in this book, is inconsistent with this possibility in at least
one respect, namely in that I say that the outcome of decisions made by people, taking into
account conscious experiences, is not necessitated by objective physical facts.
2
Nagel (1974), (1986).
3
Supported in Stoljar (2006).
Consciousness and Decision-Making 57

Another line of argument against a dual aspect view is to the effect that the
conception of conscious experiences that it adopts is scientically unsound and
misconceived.
An extreme view, advocated by Patricia and Paul Churchland,4 is to the effect that
such a conception of conscious experiences is bound up with an unsound folk psy-
chology about human conduct, according to which human beings decide what to do
on the basis of conscious appraisal of consciously apprehended reasons; and that this
folk psychology is so at odds with scientic explanations of what happens in the world
that it is reasonable to expect that it will be wholly discredited and displaced as scien-
tic understanding of the brain progresses. Then, the conception of conscious experi-
ences will be so changed that they will no longer be considered as subjective private
occurrences having a different character from the occurrences of the external world.
A somewhat less extreme view is that of Daniel Dennett,5 to the effect that investigation
of the brain and of conscious experiences should proceed by addressing just the data that
can be addressed scientically. This data includes reports by persons about experiences
that they have, and other behaviour expressive of experiences; but it does not include the
conscious experiences themselves, which are of their nature unveriable and thus cannot
be the object of scientic investigation. Thus the investigation of the brain needs no as-
sumption either that conscious experiences actually occur or do not occur, and conscious
experiences can be considered simply as a kind of assumption or postulate that helps to
make sense of the reports and behaviour, rather than as actual occurrences.
I think the arguments in support of positions such as these, while they may suggest
useful scientic methodologies, do not make out a case that could require me to give
up or substantially modify my fundamental beliefs that conscious experiences occur,
that I have conscious experiences that are mine (and not directly accessible to anyone
else), that other people have conscious experiences that are theirs (and not directly ac-
cessible to anyone else), and that in these respects conscious experiences are funda-
mentally different from objective occurrences in the external world.
So I say it is reasonable to accept that conscious experiences are actual occurrences,
and that their character as conscious experiences is importantly different from their
character as physical occurrences in the brain.
4.1. There is a fundamental distinction between two aspects of the processes that consti-
tute the conscious experiences of human beings, namely a mental or subjective aspect (as
conscious experiences) and a physical or objective aspect (as physical occurrences in the
brain).
4
P. S. Churchland (1986); P. M. Churchland (1990).
5
Dennett (1991), 7284.

This gives rise to further questions as to how conscious experiences are related to their
corresponding brain processes, and how, if at all, conscious experiences as such have
some causal role in the determination of what happens in the world. Before embarking
on those questions, however, I need to say more about characteristic features of con-
scious experiences that could be related to their causal role.
CHARACTE R I STIC FEATU R E S OF CON SCIOUS

EXPE R I E NCE S
Conscious experiences have three characteristic features that could be relevant to their
causal role, and that seem quite different from causally relevant features of physical
occurrences: namely, subjectivity, qualia and unity.
First, subjectivity. Conscious experiences are experiences had by a conscious subject,
and indeed can reasonably be considered as being constituted by the interdependent
existence of a subject and the contents of the experience.6 They are not like objective
features of the world, equally available for observation by anyone in a position to ob-
serve them. When I have an experience of pain, only I feel my pain: my pain is the
content of an experience of which I alone am the subject; and while other people may
know from my behaviour and from surrounding circumstances that I am in pain, the
pain itself is not available for experience or observation by others.
Second, qualia. Some conscious experiences have features or qualities that go beyond
the causally operative physical features that seem to give rise to them: the look of colours,
the feel of pain, and so on. Experiences of that kind have been called qualia (the plural
of the Latin word quale, meaning what sort or what kind), because such experiences have
a particular quality that distinguishes them from other kinds of experiences, and also from
physical processes including the brain processes that are associated with the experiences.
Thus, an experience of the colour green has a quality quite different from an experience of
the sound of a violin, or even from an experience of the colour blue; and there seems to be
nothing about wavelengths of electromagnetic radiation, or physical brain processes caused
by visually encountering them, that captures or explains the actual look of a blue sky.
And third, unity. A conscious experience is a unity in the sense that many features are
experienced all-at-once by the subject. This is particularly striking in the case of visual
experiences, in which a subject is aware of many features of an observed scene, and
generally grasps them all-at-once in a whole unied experience or gestalt. How this
happens, when different features such as shapes and colours are processed by different
parts of the brain, is itself something of a mystery, which has been called the binding
problem of consciousness.
6
Cf. Honderich (1987), 445.

So I repeat:
4.2. Conscious experiences have three characteristic features that could be relevant to
their causal role, namely subjectivity, qualia and unity.
In the next section I will consider subjectivity in more detail, and then I will go on to
consider qualia and unity.
S U BJ ECTIVITY
In discussing experiences in chapter 1, I was not at that stage prepared to state, as one of my
most certain beliefs, that conscious experiences necessarily had a subject, even an ephemeral
subject that might last no longer than each experience. However, it will be recalled that I
soon came to refer to experiences that I have and experiences that other persons have; and in
fact I have no doubt that experiences of the kind that are the object of my most certain
beliefs are experiences had by someone or something. This is part of my understanding of
what a conscious experience is: for example, I take the concept pain to refer to pain as felt
by someone or something, and the same is true of other concepts referring to conscious
experiences. It could be that my understanding of the correct use of language is wrong in
this respect (but I dont think it is), or that language in this respect enshrines a misconcep-
tion (but I dont think it does, and Im not aware of any good argument that it does).
So:
4.3. A conscious experience always has a subject that has the experience, and is consti-
tuted by the interdependent existence of the experiencing subject and the content of the
experience.
In the case of human beings, the subject is, I believe, generally best thought of as the
person or human being having the experience, in his or her capacity as an experiencing
subject. In ordinary use of language concerning subjects of experience, it is generally a
person as a whole (or animal as a whole) that is identied as the subject: we do not say
that a persons brain has this or that experience, but that the person has this or that
experience.
But when one closely considers just what the subject of experience is, questions arise as
to whether it is, exactly, the whole human being or animal, or something other than this.
We tend to think of a persons conscious experiences as somehow going on inside
that persons head, not least because of the way such experiences are associated
with brain processes. (I will say more about this later.) This gives rise to a (falla-
cious) tendency to look for a subject of experience that is something other than the

whole person, something of the nature of a little person or homunculus located
inside the head.
Further, although in ordinary language the subject is taken to be a person as a whole,
not all of a person is necessary to constitute a subject. I see a tree. I could still see it if I
lost my arms and legs. I could presumably still see it if the life-support given to my
head and brain by the rest of my body were to be provided articially. I could still have
a similar experience if nerve processes similar to those from my eyes to my brain were
to be provided articially. This suggests that the minimum requirements for human
consciousness are something less than a whole human being; and that for many pur-
poses the subject can reasonably be considered as constituted by persons brain-and-
mind, rather than the whole human being.
However, I dont think any of this justies regarding the subject of human experience
as generally being other than the person or human being, in his/her capacity as an ex-
periencing subject.
There are cases where some kind of human consciousness may seem to occur that is
not integrated into the ordinary consciousness of the person, so that the identication
of the subject of those conscious experiences as the human being could be question-
able. I have in mind here dissociative states, hypnosis, and so-called split brains. For
instance, in relation to the third of these examples, there have been cases where per-
sons, whose connections between the two hemispheres of the brain have been cut in
order to treat severe epilepsy, have exhibited behaviour suggesting the existence of two
centres of consciousness. Although such cases indicate the possibility of the existence
of a subject other than the (whole) person, I dont think they show it is generally incor-
rect to identify the subject with the person.
Another reason for questioning the identication of the subject with the person is
that some people think of themselves as subjects that could possibly exist without a
body or even a brain. People do ask what will happen to them after they die, and this
question is understood as relating to them as subjects of experience, and is not gener-
ally dismissed as nonsense. The great religions make assertions about an afterlife, or
about reincarnation, and these assertions are widely discussed and disputed rather
than being summarily dismissed as nonsense. In saying that the subject is the human
being, I do not mean to suggest that it is impossible that a subject, continuous with a
human subject, could exist in another form.
Nor am I saying that the only possible subjects of experience are humans and ani-
mals. It might be argued that our concept of consciousness is such that only behav-
ioural criteria can warrant the ascription of that concept, and only humans and animals
can meet those criteria. However, I contend that the correct ascription of the concept
of consciousness depends, not on behaviour, but on whether anything is in truth expe-
rienced. I say it is conceivable that conscious experiences could occur even if there is no

person or animal having them, for example in an articially created system, or in an
afterlife; and that in any such case there would be a subject of experience. While we are
clearly a very long way from creating an articial system that could have an experience,
and while I do not myself believe there is an afterlife in the sense of some kind of
straightforward survival of death, the questions whether such a system could be cre-
ated and whether there is an afterlife, in either case with an associated subject of expe-
rience, are not resolved by mere conceptual arguments.
One nal problem with the identication of a subject is that it could be taken to sug-
gest that the subject is not just an experiencer, but also an originator of action that is
wholly in its control; whereas in truth unconscious processes play a large, if not always
decisive, part in determining what a human being does. I do not intend to suggest, by
identifying a subject of experience, that this subject has a kind of control over what a
person does that is free from unconscious inuences. However, I will later be arguing
that the subject (in the case of human beings, the person in his/her capacity as an ex-
periencing subject) does contribute a conscious input into the origination of action
that is not wholly determined by unconscious processes and laws of nature.
So:
4.4. In the case of human beings, the subject is the human being, in his/her capacity as
an experiencing subject; but this reserves judgment on what it is about human beings that
gives them this capacity, on problematic cases such as dissociative states or hypnosis or split
brains, on whether such a conscious subject can exist in another form, and on questions of
control.
QUALIA AN D U N ITY
I have said that a conscious experience is constituted by the interdependent existence

of an experiencing subject and the content of the experience. I have said a little about
the subject of experience in the previous section, and will now say something more
about the content of experience, and in particular the two features of this content pre-
viously identied, namely qualia and unity.
Some of our conscious experiences are experiences by means of which we perceive
features of the external world; and it will be apparent from what I have previously said
that I do draw a distinction between those features of the external world that we perceive
and the content of the experiences by means of which we perceive them. This distinction
is further conrmed by the following considerations.
The existence of qualia as a feature of conscious experiences has been challenged on
the basis that such things as colours are properties of features of the external world that
we perceive, rather than being properties of our experiences.

Now I accept that the external world does include objects that are coloured in the
sense of having properties that cause experiences of colour in normal human beings,
and that these properties are associated with objective measurable quantities available
in common for observation and measurement by different people. However, this does
not necessarily mean that colour as such exists in the external world, independently of
the qualia of colours that are features of our experiences; and it certainly does not
mean that qualia do not exist as features of our experiences.
And of course this argument does not apply at all to qualia like pain that are not
perceived as properties of features of the external world, or to qualia of experiences
that do not provide reasonably accurate perception of the external world, such as
dreams or hallucinations. Qualia such as these can have no kind of existence outside of
the conscious experiences themselves.
A more thoroughgoing line of argument against an approach such as mine, identify-
ing experiences as having contents that include qualia and other features experienced
all-at-once, is that this approach is conceptually misconceived. In a book very much
inuenced by Wittgenstein,7 Max Bennett and Peter Hacker have argued to the effect
that conscious experiences are not private occurrences with contents inaccessible to
anyone other than the person having the experience, but just processes by which per-
sons perceive and act in the world. They contend that It is mistaken to suppose that the
subject of experience has access or privileged access to his own experience (at 295);
and that, in the case of visual perception, to see an object is neither to see nor to con-
struct an image of an object (at 141).
I believe these assertions are fundamentally mistaken, and I need to take a little time
considering the nature of the contents of conscious experiences to explain why.
Bennett and Hackers latter assertion occurs towards the end of a discussion of the
binding problem of consciousness. Having argued that there is no binding problem as
generally understood, because there is no sense in which features of a perceived object
have to be combined in the brain, the authors continue (at 13334):
Above all, to see an object is neither to see nor to construct an image of an

object. The reason why the several neuronal groups must re simultaneously
when a person sees a coloured three-dimensional object is not because the
brain has to build up a visual image or create an internal picture of objects in
the visual eld. . . . Since seeing a tree is not seeing an internal picture of a tree,
the brain does not have to construct any such picture. It merely has to be func-
tioning normally so that we are able to see clearly and distinctly. It does not
have to take a picture apart, since neither the visual scene nor the light array
7
Bennett and Hacker (2003).

falling upon the retinae are pictures. It does not have to put a picture back to-
gether again, since what it enables us to do is to see a tree (not a picture of a tree)
in the garden (not in the brain).
Now I agree that to see an object is not to see an internal picture of the object, pri-
marily because that characterisation involves misuse of the word see; but I contend it
is beyond any doubt that we do see objects via privately experienced images con-
structed by our brains, which are part of the contents of our conscious experiences.
I look at the setting sun just before it disappears below the horizon. In my visual eld,
there is something that is orange and apparently circular. What is it? Bennett and
Hacker would say it is the sun itself, not an image of the sun. However, if it is literally
the sun itself, it is not the sun as it is when there is this something in my visual eld, but
as it was eight minutes earlier. Suppose the sun disintegrates during this eight-minute
period. Then, if the orange something is literally the sun itself, rather than some kind
of image or visual representation of the sun, the sun both exists (it is right there, right
now) and does not exist (it has disintegrated) at the time when I am aware of the
orange something. Further, the events or processes that are constitutive of there being
this something in my visual eld are much more closely linked to events or processes
in my brain than to the most relevant events or processes on the sun: the latter events
or processes took place 150 million kilometres from me and eight minutes earlier.
The orange something must surely be an image of the sun constructed by my brain,
though not in the sense of a picture of the sun that would itself have to be seen. Indeed,
it would not be correct to say that I see this image, and it is not any kind of thing that
exists independently of being part of my conscious experience. It is unlike a picture in
that it is immediately present to me, and in that its existence is entirely dependent on
my existence as an experiencing subject immediately acquainted with it. But it is an
image nonetheless, in the sense that it is not the sun itself but rather is a visual repre-
sentation to me of the sun.
Suppose that prosthetic eyes are developed, and that the output from these eyes to
the optic nerves can be articially controlled by computational processes (not depend-
ing on light entering these eyes) so as to be indistinguishable from the output that re-
sults from actually looking with these eyes at the setting sun. The person using these
eyes will then undoubtedly have a visual experience of the setting sun (of the nature of
virtual reality), through what can reasonably be called an image of the setting sun,
constructed by the persons brain from the processes articially induced in the persons
optic nerves. And if so, it cannot be doubted that when the person actually looks at and
sees the setting sun, so that similar processes are induced in the optic nerves by the
light from the sun entering the eyes, the persons brain constructs a similar image of
the sun and it is by means of that image that the person can see the sun.

To give another example, if I see a tree in my garden, it seems to be xed in one spot,
despite normal eye movements. But when I push with my nger against my right eye-
ball, and jiggle it, I seem to see two trees, one still stable as before and the other jiggling
as I jiggle my eyeball. Before I jiggled my eyeball, the three-dimensional image of the
tree constructed by my brain was stable because my brain had the ability, by complex
information-processing, to compensate for changes caused by my eye movements to
the patterns of retinal nerve excitation. However, my brain does not compensate for
changes to the patterns of nerve excitation caused by external jiggling of my eyeball, so
the image present to my consciousness associated with the patterns of nerve excitation
on the retina of my right eye can no longer be fused with that from my left eye into a
single stable three-dimensional image; but it jiggles about so that there are now two
two-dimensional images, one stable and one jiggling. It does not merely seem as if there
is something jiggling, there is something tree-like that is jiggling. What is the jiggling
tree, if not an image of the tree constructed by my brain, based on the jiggling patterns
of nerve excitation on the retina of my right eye?
Turning to the other claim made by Bennett and Hacker, that It is mistaken to sup-
pose that the subject of experience has access or privileged access to his own experi-
ence, this claim may seem plausible so long as it is denied that the brain constructs
images of which we are immediately aware. But once it is accepted that the brain does
construct such images, it becomes undeniable that the subject of experience alone is
aware of these images (just as I alone am aware of the jiggling tree); and this can rea-
sonably be called privileged access to such images.
So I dont think the conceptual arguments of Bennett and Hacker give reason to
doubt that conscious experiences are private occurrences, or that the distinction be-
tween subjective and objective aspects of the processes that constitute conscious expe-
riences is of fundamental importance.
My discussion of Bennett and Hacker also supports a belief as to the relationship
between contents of experience, in the case of perception of features of the world, and
objects of experience, that is, the features of the world so perceived. My example of the
jiggling tree powerfully conrms to me that the contents of experience (here, the jig-
gling image) are not the same as the object of experience (the tree itself). The contents
of experience do however include features (for example, the appearance of the trunk,
branches, green leaves) that can be considered features of the tree itself, in the sense
that they can be observed in the same way by anyone with normal vision; as well as
including at least one feature that cannot be considered a feature of the tree itself (it is
jiggling, not stable). In the case of accurate perception, the contents of the experience
will thus include features that can in that sense be considered features of the perceived
object; but always with the difference that the features as experienced, being part of the
contents of experience, exist interdependently with the experiencing subject, whereas

the features of the observed object itself are objective, not interdependent with any
subject, and potentially observable by anyone with normal vision.
4.5. The content of a conscious experience can, in the case of accurate perception of
features of the world, include features that can in a sense be considered features of the
part of the world so perceived; but this content is not itself the part of the world perceived,
and the subject of the experience exists interdependently with and has sole access to this
content.
N E U RAL COR R E LATE S OF CON SCIOUS N E SS
I have expressed the view that conscious experiences are the same occurrences as cer-
tain processes in a persons brain, and this is in fact a prevalent view among neurosci-
entists and philosophers.
It is reasonable also to believe that there are law-compliant correlations between the
physical and the mental aspects of the brain processes that constitute conscious experi-
ences; that is, that there are regularities, which could be considered laws of nature of a
kind, such that physical brain processes of a particular type Xp are associated with cor-
responding conscious experiences of a particular type Xe, such as feeling a stabbing
pain or seeing something blue.
In perception, for example, there are physical processes linking an object of percep-
tion with brain processes, and ultimately with the conscious experiencing of the object;
and although the content of the experience is no doubt affected by the interests and
expectations of the experiencer, the reliability of perception must depend on types of
conscious experiences being linked in a lawful fashion with types of features of per-
ceived objects, through types of physical processes occurring in the brain.
That is not to say that the links and correlations between physical brain processes
and conscious experiences must always be simple and straightforward. I have referred
to David Marrs three-level analysis of cognitive processes; and consistently with that
analysis, there may be one set of laws linking processes at the bottom level of physical
implementation with processes of the middle algorithmic level, and another set of laws
linking processes at the algorithmic level with the conscious processes of the top overt
level of cognition. And just as the same computer program can run on different hard-
ware, it may be that some conscious experiences are correlated by such laws directly
with computation-like programs of the brain at the middle level, rather than with
physical brain processes at the bottom level. Thus, similar conscious thoughts had by
different people (say, thinking about Vienna) might be correlated by such laws with
similar processes of computation-like programs in their brains; but the processes of
those programs might in turn be correlated by further laws with different physical

processes of their brains, where these further laws permit of the similar programs run-
ning on different brain hardware.8
So while I hold the following belief, it is to be understood that the correlation be-
tween types of physical processes and types of conscious experiences may be indirect,
in the way I have suggested and perhaps in other ways.
4.6. There are types of physical processes in the brains of humans and other animals that
correlate with types of conscious experiences, in accordance with laws of nature.
This view is also a prevalent view among neuroscientists and philosophers; and it is
associated with what is called the supervenience of the mental on the physical, according
to which what mental processes occur depends in a lawful fashion upon what physical
processes occur. I do not altogether agree with this understanding of supervenience. I do
accept that there is no change in mental processes without a corresponding change in
physical processes. But as will be seen, I contend it is no more correct to say that what
mental processes occur depends upon what physical processes occur, than it is to say
that what physical processes occur depends on what mental processes occur; so that it
is better to assert correlation between the physical and the mental than to assert that
one depends on the other.
A distinction can be drawn between two broad classes of laws correlating types of
brain processes and types of conscious experiences.
First, there must surely be laws that correlate types of physical brain processes with
types of individual features of conscious experiences, such as feelings of pain or visual
experiences of a colour; so that whenever a person has, say, a visual experience of the
colour blue there are types of processes occurring in the persons brain. It seems,
for example, that there are types of patterns of neural activity that correspond with
visual experiences of various kinds, and other quite different types of patterns of neural
activity that correspond with feelings of pain. Thus, a pattern of type A could corre-
spond with a green patch in a certain part of the visual eld, a pattern of type N could
correspond with a vertical edge in another part of the visual eld, and so on.
These laws thus link types of qualia with types of brain processes that are at least
necessary (if not sufcient) for the existence of those types of qualia; and they could be
called qualia laws. There has been some progress in the attempt to identify what types
of brain activity correspond with what types of conscious experience; but this has
generally been in terms of such things as location in the brain, connections to other
parts of the brain, relationship to sensory inputs and motor outputs, rates of neural
8
Cf. Hilary Putnams version of the theory of mind called functionalism: see Putnam (1975), xii
xiv, 292300.

ring, types of neurotransmitters involved, and such like. It falls far short of identify-
ing any type of physical occurrences in the brain as being the objective aspect of any
type of conscious experience or as being necessary and/or sufcient for the occurrence
of any type of conscious experience; and thus it falls far short of formulating laws link-
ing types of physical occurrences with types of conscious experiences. And there has
been no explanation of why some types of physical occurrences, and not others, should
have this kind of link to conscious experiences, much less any understanding of what
function this link may have, or any idea how to reproduce it articially.
Second, there must be laws that correlate types of brain processes with the binding
or chunking into a single experience of many features of what is experienced; so that
whenever these types of processes occur, individual qualia correlated by qualia laws
with types of brain processes, generally occurring in spatially extended and separated
parts of the brain, are combined into a single unied experience. These laws could be
called binding laws. Some progress has been made in identifying types of processes
associated with this chunking of qualia, such as a pattern of coordinated 40Hz oscilla-
tions of the activity of relevant neurons; but this falls even further short of providing
understanding or explanation of how this produces unied experiences.
TH E E FFICACY OF CON SCIOUS EXPE R I E NCE S
Having outlined some views about conscious experiences, I can now come to the main
points to be made in this chapter, namely that conscious experiences can contribute
positively to human rationality, and in particular to plausible reasoning; and that it is
most unlikely that this positive contribution is precisely as determined by laws or rules
of any kind.
Our brains undoubtedly have a prodigious capacity for unconscious information-
processing, by means of brain processes that do not require or depend upon any con-
scious experiences.
I remember this being brought home to me some years ago when I rst looked at a
random dot stereogram through coloured spectacles. At rst I just saw the random
dots somewhat blurrily arranged on a at surface; but then gradually I came to see in
stark clarity what appeared to be a spiral shape rising from the page towards me. Plainly
I had not consciously worked out which dots matched which, or what spatial congu-
ration the matched dots indicated: it had all been done by unconscious processes of my
brain, which must have required very elaborate computations.
Of course, much the same thing happens all the time in our ordinary visual percep-
tion. Not only are there the computations that give us three-dimensional vision, but
also there are the elaborate computations that provide stability of viewed scenes de-
spite voluntary movements of head and eyes. And similarly, there are elaborate

computations that our brains perform in order to enable us to do such things as to
balance and to catch balls.
Probably even greater computational virtuosity is required for language use. Our use
of language generally involves conscious perception and action; but underlying the
conscious processes must be a vast amount of unconscious computation that enables
the memory store and techniques that give us the capacity to use language, to be
brought to bear and given effect to on particular occasions. And many more examples
can be given to demonstrate that our brains have enormous computing capabilities
that do not require consciousness in order to be effective.
Most of us cant consciously plug directly into this computing power, as indicated by
the clumsiness and fallibility of our efforts to carry out simple mental arithmetic.
There are however reports of people who appear have done so to some extent, not
necessarily to their advantage. In his book The Man Who Mistook His Wife For a Hat,
for example, Oliver Sacks wrote9 about autistic twins who enjoyed exchanging multi-
digit prime numbers. It would appear that the condition that disadvantaged them in
their social interactions may have given them the ability to make use of their brains
computational capacity in a way that is not available to most people.
Now if optimal decisions on matters important for our survival and reproduction
could be made without a positive contribution from conscious experiences, it might be
expected that evolution would have ensured that decisions be made by using just this
prodigious unconscious computing capacity, particularly when our conscious pro-
cesses seem clumsy and fallible by comparison.
And yet: we are so constituted that, whenever in life we are faced with a novel situation
requiring some signicant decision or action, our conscious minds are automatically
brought to bear. This happens for example if I am driving a car thinking about other
things, and I become aware of a challenging situation that requires ongoing monitoring.
My conscious attention is automatically engaged to focus on this exigency: I cannot do
otherwise than apply my conscious attention fully to the situation facing me.
Similarly, whenever we have a signicant decision to make about what to believe or
what to do, we do not make it without addressing the question consciously. In the
previous chapter, I argued that an important part of human rationality is the capacity
to engage in plausible reasoning, in which the premises or data support the conclusion
not by virtue of applicable rules of good reasoning but as a matter of reasonable albeit
fallible judgment. I suggest it is clear that:
4.7. The judgment that is required for plausible reasoning is generally a judgment that is
addressed and adopted consciously.
9
Sacks (1986), 185203.

In an earlier draft of this book, I expressed this assertion in terms of the judgment
being made consciously rather than being addressed and adopted consciously. One
reader read this as asserting both that consciousness alone generated solutions to prob-
lems of plausible reasoning, and that unconscious processes were incapable of generat-
ing such solutions. I have reformulated the assertion to make it clear that I am here
asserting neither of these things, but merely that we do as a matter of fact consciously
address and adopt our judgments of plausible reasoning. I do build on this to argue
that this enables our conscious experiences to make a positive contribution to these
judgments; but I fully accept that potential solutions to problems of plausible reason-
ing can be generated by unconscious processes.
One clear example of this is the sleep-on-it phenomenon: after we have been con-
sciously addressing a question, and then stopped doing so, it quite often happens that after
a time a new idea occurs to us. This is another example of unconscious processes display-
ing considerable computational capacity. However, what I do say is that in such cases we
generally do not adopt the new idea without consciously appraising it; and although
sometimes we nd that the new idea is a good one and should be adopted, sometimes we
nd that the new idea is not a good one, or not the best one. Indeed, I contend it would
not be rational behaviour to adopt ideas that occur to us in this way without rst con-
sciously appraising them, and I contend that no sensible person would in fact do so.
Not only are such judgments in fact generally addressed and adopted consciously,
but also there are strong reasons for holding that this conscious activity does contrib-
ute positively to the making of these judgments.
For one thing, much unconscious information-processing seems to be nely tuned
to support conscious experiences in which currently important information is pre-
sented simply and vividly, in the manner of an executive summary prepared for a de-
cision-maker in business or government.10 This is particularly apparent in the extensive
commitment of computational capacity to providing our stable three-dimensional
vision, vividly encompassing salient features of our environment. Computer scientist
Marvin Minsky once dismissed consciousness as a very imperfect summary in one
part of the brain of what the rest is doing;11 but he failed to recognize that there must
be an evolutionary advantage in having this summary. Plainly, I contend, the capacity
of our brains to provide these executive summaries has been selected by evolution
because it assists in decision-making, and particularly because it enables us to make
the judgments necessary for plausible reasoning.
Another indication that conscious experiences contribute positively to decision-
making is the fact that we have feelings like pain to motivate us. If there was no positive
10
Cf. Penrose (1987), 26667.
11
Minsky (1985), 42.

contribution to decision-making from conscious experiences, why would there need to
be any more than unconscious computation and implementation of the course of action
identied by that computation as being best suited to detecting and repairing damage to
ourselves and avoiding damage in the future? It would be absurd (even if possible) to use
pain or any other feelings to motivate a computer: a computer would, if set up so as to
carry out an effective program for detecting and repairing damage to itself and avoid-
ing such damage in the future, simply carry out that program, without need for any
additional motivation to do so. Our conscious motivation makes good sense if (and,
I contend, only if) conscious experiences contribute to decision-making.
In chapter 2, I referred to the work of Daniel Kahneman, Amos Tversky and others12
showing that our reasoning is affected by unconscious biases. The only way that these
biases can be addressed and their effect minimised is by paying careful conscious atten-
tion to them, and it is in that way that human rationality can be protected against these
biases. Unless conscious experiences make a positive contribution to decision-making,
conscious appraisal could surely not be trusted to deal with unconscious biases; and this
would conict with the sound belief supported in chapter 2 that human beings can
make rational decisions about what to believe and what to do.
So although there is no doubt that unconscious processes play a large and indispens-
able role in our decision-making, there are extremely strong reasons for holding that
part of that role is to give rise to conscious experiences which also contribute positively
to decision-making, and particularly to the judgments necessary for plausible reason-
ing. I condently assert:
4.8. A persons conscious experiences (including visual and auditory experiences,

thoughts and feelings) contribute positively to decision-making and in particular to plau-
sible reasoning.
Before leaving this topic, I should mention that there has in recent years been some
psychological research into the relative merits of conscious and unconscious decision-
making. A paper published in 200613 suggested that choice in complex matters should
be left to unconscious thought, but this suggestion has since been strongly contested;14
and the superiority of careful conscious reection over intuitive snap decisions in the
type of reasoning with which I am most familiar, namely judicial decision-making, is
I believe well established.15
12
Kahneman et al. (1982).
13
Dijksterhuis et al. (2006).
14
See for example De Wall et al. (2008); Newell et al. (2009).
15
See for example Guthrie et al. (2007).

Although I have suggested that the positive contribution made by conscious experi-
ences is to plausible reasoning in particular, I note that one of the papers I have cited
as supporting the superiority of conscious decision-making relates to logical reasoning
rather than plausible reasoning as such; and I also note that full conscious concentra-
tion is required for us (say) to carry out tasks of mental arithmetic that can easily be
carried out without consciousness by computers or electronic calculators. This I sug-
gest is because our conscious reasoning is adapted to making the informal judgments
of plausible reasoning and, as mentioned earlier, we cannot at will plug into our capac-
ity for unconscious information-processing in order to carry out tasks such as mental
arithmetic. In addition, it often happens that logical or mathematical tasks we perform
are not isolated from plausible reasoning, but rather depend for their usefulness on
prior or subsequent exercises of plausible reasoning. For example, logical puzzles of
the kind used in intelligence tests generally require plausible reasoning to select hy-
potheses that can then be tested against the requirements of the puzzle; and I suggest
also that one often uses ones general understanding of a problem, which depends on
consciousness and plausible reasoning, to satisfy oneself that one has indeed found the
answer. I will say more about understanding in chapter 6.
TH R E E QU E STION S
If one accepts, as I think one should, that conscious experiences do make a positive
contribution to our decision-making, three questions arise:
1. What is that contribution; that is, just what is it that conscious experiences contrib-
ute to decision-making that is not done by unconscious processing?
2. What is it about conscious experiences that enables them to make this contribution?
3. How do conscious experiences make that contribution?
So far as Im aware, leaving aside suggestions of the general kind to be made in this
book,16 no plausible answers to the rst two of these questions have ever been sug-
gested by scientists or philosophers (indeed, there has hardly ever been any serious
consideration given to these questions); and one of the aims of this book is to provide
and develop plausible answers to them. Scientists and philosophers have suggested a
plausible answer to the third of these questions; and although I agree with some as-
pects of this answer, it is generally developed in such a way as to preclude plausible
answers to the rst two questions.
16
Similar ideas are hinted at in Abelson (1988), Penrose (1989), 40925: Penrose (1999), 10317; and
particularly in Kauffman (2009).

The answer generally given by scientists and scientically oriented philosophers to
the third question is that conscious experiences make a contribution to our decisions
and actions, just because conscious experiences are actually one and the same as asso-
ciated physical processes of the brain: they say in effect that since the physical processes
have a causal input, so also the associated conscious processes must have the same
input. This is generally combined with the view that everything that happens, in the
physical or material world at least, happens in accordance with physical laws of nature
engaging with physical features of the world, being either wholly determined by these
laws and features, or else happening randomly within probability parameters deter-
mined by them. It is said that the physical word is closed to causal inuences that are
not themselves physical.
As mentioned earlier, I agree that conscious experiences and associated physical
brain processes are two aspects of the same processes; but I do not agree that the whole
of the causal input from these processes is to be attributed to their physical aspect, or
that the physical world is closed to causal inuences that are not physical. I explain why
in the remainder of this book.
The suggestion that conscious experiences make the same contribution to what hap-
pens as do associated physical processes has also been put in terms of the three levels
of cognitive processes, referred to earlier. What actually happens is determined by
physical laws operating at the physical level, and also (and consistently) by computa-
tional rules operating at the algorithmic level, as is the case with our computers. The
operation at the algorithmic level can be considered as having independent efcacy,
without compromising the causal closure of the physical, because the same computa-
tional operations can be carried out on different physical systemsjust as the same
computer programs can be carried out on quite different computer hardware. It is
sometimes suggested that the same may also be true of the level of conscious experi-
ences; that is, that these experiences could also be considered as having independent
efcacy, without compromising the causal closure of the physical. However, this last
suggestion has never been developed in any plausible way.
The levels approach does not explain how or why the processes that make whatever
contribution conscious experiences make to decision-making have that third level of
description at all; it does not identify what contribution it is that processes having that
third level of description make, which is not made by processes having only the other
two levels of description; and it does not explain what it is about processes that have
that third level of description that enables them to make any such contribution. Thus,
it does not begin to answer the rst two of the three questions identied above.
Indeed, my contention is that, by insisting on the causal closure of the physical, an
approach such as this precludes answers to these two questions. Since what happens at
the conscious level happens in terms of inconclusive reasons, such as the soft

constraints of ceteris paribus rules, its distinctive contribution cant be by way of com-
putational processes consistent with causal closure of the physical; and those people
who have considered this levels approach are generally driven to say that any causal
efcacy of the relevant processes has to be by virtue of the algorithmic level and/or the
implementational level, so that (inconsistently with the view that conscious experi-
ences contribute positively to decision-making) the conscious level is in fact epiphe-
nomenal, that is, it does nothing.17
4.9. Scientists and philosophers have offered no plausible suggestions as to what is it that
conscious experiences contribute to decision-making that is not done by unconscious pro-
cessing, or what is it about conscious experiences that enables them to make this
contribution.
It might be thought that the view that the physical world is closed to non-physical
causal inuences has some support from the notion, referred to earlier, that there
are psycho-physical laws that correlate the experiential and the physical aspects of
brain processes, so that an experience of the type Xe occurs whenever a physical
brain process of the type Xp occurs. Thus, if physical processes of type Xp have a role
in determining what happens, in accordance with laws of nature, then it might be
thought that the corresponding conscious experiences of type Xe must have the
same role, in accordance with the same laws of nature plus the laws linking Xp and
Xe, without affecting the causal closure of the physical. I will explain in the next
chapter why I say that, despite my acceptance that there are psycho-physical laws
correlating the experiential and the physical aspects of brain processes, the experi-
ential aspect can and does have a causal input that could not be provided by the
physical aspect.
R U LE-DETE R M I N E D PRO CE SS E S D O NOT N E E D

CON SCIOUS N E SS
There is a further clear and I believe insoluble difculty in reconciling acceptance of a

positive contribution of conscious experiences to decision-making, with approaches
asserting the causal closure of the physical, or indeed with any approach that attributes
everything that happens to rule-determined processes and/or randomness. This is that
any conclusion that can be reached wholly as a result of the engagement of general
rules with existing circumstances and/or randomness can be reached without any need
for consciousness.
17
See for example Hofstadter (1986), 654.

It is obvious that if changes occur to existing circumstances strictly as determined by
rules engaging with and operating on those circumstances (or as determined by such
rules together with some randomness), there is no need for any conscious interven-
tion: the changes just happen automatically. The circumstance that changes occurring
within the brain may not merely be physical processes proceeding automatically as
determined by laws of nature, but may also be information-processing proceeding as
determined by computational rules selected in evolution, makes no difference to this:
the information-processing as determined by computational rules will proceed auto-
matically in conformity with both the laws of nature and the computational rules, as it
does in a computer.
It is generally accepted also, following theoretical work by Alan Turing,18 that any
rule-determined information-processing that can be carried out in a nite number of
steps by any digital machine can be carried out by a universal machine, now exempli-
ed by todays computers. No reasonable basis has ever been advanced to suggest that
our computers are conscious; so again, if human rationality is provided by rule-deter-
mined information-processing that could be carried out by a universal machine with-
out consciousness, there needs to be an explanation of how conscious processes
contribute positively to human rationality, if indeed they do so. (This particular argu-
ment does not apply to analog systems of the kind proposed by Horgan and Tienson;
but as mentioned earlier, it is seriously questioned whether such systems can achieve
results that cannot be achieved by digital systems.)
Of course, some universal machines are slower than others, and it might be suggested
that consciousness enables more efcient information-processing, or perhaps is a
manifestation or by-product of a particular way in which our brains carry out rule-
determined information-processing.
As regards the former suggestion, it has been pointed out to me that the computer
Big Blue played an excellent game of chess by unconscious application of rules to
simple elements, whereas Gary Kasparov could play as good a game by conscious ap-
plication of heuristics; this indicating that consciousness can provide an economical
way of achieving a result achievable also by unconscious rule-determined information
processing. However, for this argument to work, there would need to be an explanation
as to how consciousness could make a positive contribution to Kasparovs application
of heuristics, a contribution that was itself consistent with all the processes of Kasp-
arovs brain being rule-determined. The problem remains that there is no suggestion
with the slightest plausibility as to why or how consciousness could, consistently with
the rule-determination of all brain processes, contribute to the efciency of informa-
tion-processing.
18
Turing (1958); and see also Copeland (2002).

As regards the latter suggestion, it has been put to me that its a mistake to look for
some positive contribution referable to the particular features of consciousness that
I have identied, namely subjectivity, qualia and unity: these may be mere manifes-
tations or by-products of other advantageous features of brain processes, or indeed
of conscious processes, that do function in a rule-determined way. My response to
this argument is that it too fails to offer any plausible suggestion of why or how con-
sciousness, understood in terms of subjectivity, qualia and unity, could be such a
manifestation or by-product. I do not believe there is any plausible account of con-
sciousness that does not include the three features in question; but in any event,
there is no plausible suggestion as to how rule-determined information-processing
of any kind would give rise to these features as a by-product, let alone do so in a way
that would explain how conscious processes make a positive contribution to deci-
sion-making.
The closest I have found to such a suggestion is that originally made by Nicholas
Humphrey,19 and taken up by Daniel Dennett20 and others, to the effect that, in order
for human beings to monitor and communicate some of their own mental processes
and those of other persons, evolutionary selection has developed brains able to pro-
duce simplied user-friendly accounts of these processes, in terms of the existence of
an integrated conscious subject or self that has conscious experiences, has goals and
purposes, and chooses between available alternatives. But unless these user-friendly
accounts have effects otherwise than as precisely determined by rules, this suggestion
too gives no role to conscious experiences as such. Everything would proceed auto-
matically in accordance with the rules; and the same results could be achieved by the
construction of computable representations of such user-friendly accounts, with no
need or role for actual consciousness with subjectivity, qualia and unity. This is consis-
tent with Dennetts view, referred to earlier, that conscious experiences can be consid-
ered simply as a kind of assumption or postulate that helps to make sense of the reports
and behaviour, rather than as being actual occurrences.
I should refer also to one other suggestion for a role for conscious experiences in
rule-determined information-processing, namely that advanced by Bernard Baars21 to
the effect that they provide a global workspace where information-processing carried
out by various parts of the brain can be brought together and coordinated. I believe
that conscious experiences do indeed bring together the results of information-pro-
cessing in different parts of the brainhow it does so involves the binding problem
referred to earlierbut Baars does not provide any answer to the question of why
19
Humphrey (1983).
20
Dennett (2003), 24255.
21
Baars (1997).

processes in his global workspace need to be conscious, with subjectivity, qualia and
unity. Indeed, on Baarss account, the global workspace functions at the algorithmic
level of cognitive processing, rather than at the overt level. So long as it is supposed that
whatever processes occur in this workspace are rule-determined, consciousness on this
approach still remains a superuity.
So I maintain:
4.10. If plausible reasoning proceeded precisely as determined by rules of any kind, there
could be no positive role for a persons conscious experiences in that reasoning.
Accordingly I say there is no explanation in terms of rule-determined processes of the

positive contribution of conscious experiences to our decision-making. In the next
two chapters, I propose what I contend is a very plausible explanation of that contribu-
tion in terms of processes that are not rule-determined.
COR E ASS E RTION S AB OUT CON SCIOUS N E SS

AN D DECI S ION-MAKI NG
4.1. There is a fundamental distinction between two aspects of the processes that consti-
tute the conscious experiences of human beings, namely a mental or subjective aspect (as
conscious experiences) and a physical or objective aspect (as physical occurrences in the
brain).
4.2. Conscious experiences have three characteristic features that could be relevant to
their causal role, namely subjectivity, qualia and unity.
4.3. A conscious experience always has a subject that has the experience, and is consti-
tuted by the interdependent existence of the experiencing subject and the content of the
experience.
4.4. In the case of human beings, the subject is the human being, in his/her capacity as
an experiencing subject; but this reserves judgment on what it is about human beings that
gives them this capacity, on problematic cases such as dissociative states or hypnosis or split
brains, on whether such a conscious subject can exist in another form, and on questions of
control.
4.5. The content of a conscious experience can, in the case of accurate perception of
features of the world, include features that can in a sense be considered features of the part
of the world so perceived; but this content is not itself the part of the world perceived, and
the subject of the experience exists interdependently with and has sole access to this
content.
4.6. There are types of physical processes in the brains of humans and other animals that
correlate with types of conscious experiences, in accordance with laws of nature.

4.7. The judgment that is required for plausible reasoning is generally a judgment that is
addressed and adopted consciously.
4.8. A persons conscious experiences (including visual and auditory experiences,
thoughts and feelings) contribute positively to decision-making.
4.9. Scientists and philosophers have offered no plausible suggestions as to what is it that
conscious experiences contribute to decision-making that is not done by unconscious pro-
cessing, or what is it about conscious experiences that enables them to make this
contribution.
4.10. If plausible reasoning proceeded precisely as determined by rules of any kind, there
could be no positive role for a persons conscious experiences in that reasoning.

5
Gestalts and Rules
As I argued in the preceding chapter, there is not, and I contend cannot be, any expla-
nation in terms of rule-determined processes of the positive contribution of conscious
experiences to our decision-making. In this and the next chapter I offer and support a
specic and straightforward proposal as to what this contribution is:
5.1. Consciousness enables an organism to be responsive to circumstances grasped as

wholes, not just to those constituent features that engage with applicable laws or rules.
To expand a little:
5.2. The evolutionary advantage of consciousness is that it enables an organism to deter-

mine and/or shape an apposite response to circumstances facing it that has regard, not
only to features that engage with laws of nature and/or computational rules, but also to
whole combinations of features that are particular and perhaps unique to the circum-
stances and do not as wholes engage with applicable laws or rules.
TH E ARGU M E NT OUTLI N E D
I will begin with a statement of the basic argument for this proposal, and then I will
examine aspects of it in more detail.1
In the previous chapter I suggested there are psycho-physical laws correlating physi-
cal and experiential aspects of brain processes, so that there must be a sense in which
any information carried by the experiential aspect is also carried by the physical aspect.
However, it is important to appreciate that the information as carried by the experien-
tial aspect is combined into unied experiences or gestalts: as mentioned earlier, a
characteristic feature of conscious experience is that it is a unity in the sense that many
features are experienced all-at-once by the subject, notably in the case of visual experi-
1
This original argument of mine was rst raised in Hodgson (2001), and has been developed in
Hodgson (2002b), (2007a) and (2007b).
79
ences, in which a subject grasps all-at-once many features of an observed scene in a
whole unied experience or gestalt. I say that these gestalts, in which many features are
combined all-at-once, have informational content that is manifested and made avail-
able for use only in conscious experiences.2
5.3. Gestalts experienced by a subject, in which many features are combined all-at-once,
have informational content that is manifested and made available for use only in con-
scious experiences.
In developing the argument in this and the next chapter, I will make it clear why I say
these gestalts have this informational content, and why I say this informational content
can be used. My contention will be:
5.4. Feature-rich gestalts of conscious experiences generally do not as wholes engage with
laws or rules of any kind, but can nevertheless as wholes have a causal inuence because
subjects of experience can respond appositely to them.
This account treats the processes of our brains as including some processes that have
both a physical and an experiential aspect, but it holds that the role of these processes
in the unfolding of events is not wholly determined by physical laws engaging with
their physical aspect. It proposes:
5.5. The physical aspect of conscious processes does, in conformity with physical laws,
restrict what can happen to a limited spectrum of possibilities; but in response to the expe-
riential aspect of these processes, the subject of the experiences can control what does
happen within this spectrum of possibilities.
My account does not require a self or soul distinct from the brain, which has some
input into what happens. Rather, it proposes that the physical-and-experiential system
of the brain-and-mind has the capacity to use information carried in experiences in a
way that corresponding information carried in physical processes cannot be used, and
that generally is not wholly determined by laws of nature or pre-existing computa-
tional rules. My proposal is that the physical-and-experiential system constituted by
2
As will be seen, this additional informational content goes beyond the additional information,
what red looks like, obtained by Frank Jacksons Mary when for the very rst time she sees something
red: see Jackson (1982). That information, considered in isolation, could engage at least indirectly with
whatever laws of nature engage with its neural correlates. Whats important about the additional infor-
mational content Im referring to is that its constituted by combinations of features that do not as
wholes engage with laws of nature.

the brain (1) does, by virtue of physical laws engaging with the physical aspect of brain
processes (and by virtue of corresponding computational processes and rules), con-
strain what can happen to a spectrum of possibilities; but also (2) constitutes a subject
of experience, a person, that can respond appositely to gestalts that generally do not as
wholes engage with rules of any kind; so that (3) this subject can through this response
exercise conscious control over what happens within the available spectrum of possi-
bilities, in a way that is not rule-determined.
Plainly, my proposal concerning gestalts is important to my argument, and I need to
explain why I say that gestalts do not generally, as gestalts, engage with applicable laws
or rules.
I suggest it is characteristic of laws and rules that they apply over a range of cir-
cumstances, and must engage with types or classes of things or features that different
circumstances have in common, and/or variable quantities that can engage with
mathematical rules; so that:
5.6. While laws and rules can apply to individual unique circumstances, they engage
with features of these circumstances only in so far as each of these features is of a type or
class, and/or is a variable quantity.
Laws and rules link categories (say, X, Y, Z, etc.), where these categories are types or
classes of things or features, and/or mathematical variables. In the case of computa-
tional rules, X may be a potentially recurring situation in a computational program,
and Y may be the consequential operation to be undertaken in that situation.
I readily accept that some simple gestalts, such as a visual experience of a basic shape,
may be of a type or class such that there could be evolution-selected computational
rules engaging with them; but I say that this could not generally be true of whole fea-
ture-rich gestalts of the kind we normally experience, such as visual gestalts compre-
hending many features of an observed scene, or auditory gestalts of a unique musical
performance. And it is these particular gestalts of our ordinary experience on which I
am focussing in this discussion. Applicable computational rules could engage with
these gestalts in so far as they exemplify simple gestalts of a type or class but, leaving
aside certain exceptional circumstances to which I will come, I contend they could not
otherwise engage with and support apposite responses to whole feature-rich gestalts of
our ordinary experience.
Although I am not here considering laws of a legal system, these laws also, while ap-
plying to unique circumstances, generally engage only with types or classes of persons
or places or occurrences, and prescribe types or classes of legal consequences. Occa-
sionally, a statute law species what is to happen in a particular named place or at a
named event or even to a named person; but this is exceptional, and for the most part
Gestalts and Rules 81

laws of a legal system do not identify, and thereby engage with and specify a response
to, any particular place or event or person.
Laws of a legal system may, through engaging with each of a number of features of a
particular set of circumstances, produce a legal result that is unique to that set of cir-
cumstances; and laws of nature and/or computational rules may, through engaging
with each of a number of features that are combined into a particular gestalt, produce
a result that is unique to that gestalt. Indeed, this must happen whenever a computer
program identies a unique piece of music. However, I contend that those laws of
nature and/or computational rules that might be operating so as to constrain brain
processes generally do not engage with any rich combination of features as a whole, and
in that sense generally do not engage with whole particular gestalts; and by the same
token, particular gestalts generally do not as wholes engage with these laws or rules.
Consider for example George Gershwins melody The Man I Love. (I could equally
have chosen Summertime or Embraceable You or any of a number of otherseach of
these melodies, despite its apparent simplicity, is a unique and utterly distinctive
whole.) This melody can be given a description in terms of an aggregation of general
and/or quantitative features it has in common with other melodies, including the plac-
ing of notes, pitch changes, note lengths, and so on; and these features, being general
and/or quantitative, can engage with general rules, so that at least straightforward pre-
sentations of the melody can readily be identied by application of computational
rules to a sufcient number of the common features included in this aggregation. Such
an appealing melody may also have some constituent features that can push buttons in
our emotional make-up, because these features engage with computational rules of
our cognitive processes that have been established by evolution and environment. But
the way this particular melody sounds is a whole that is unique to this melody, rather
than being just a bare aggregation of general and/or quantitative features that it has in
common with other melodies; and the same is true of the melody of each of several
two- and four-bar chunks of the complete melody. An experience of such a unique
melody, as a whole, is an example of what I mean by a gestalt (1) that has informational
content that is manifested only in conscious experiences (the way the whole melody
sounds), (2) generally does not engage with relevant laws or rules (at least on a persons
rst encounter with it), but yet (3) can be used and be causally efcacious because the
subject of experience can respond appositely to it.
This unique melody did not exist until it was composed, and neither Gershwin nor
anyone else could have been primed in advance by evolution and/or experience for a
rule-determined response to its whole exact form. When Gershwin was composing
the melody, possibilities for how it should proceed must have been thrown up by
unconscious processes, presumably processes giving effect to computational pro-
grams of his exceptional brain, which were themselves a product of his genes and

environment (and perhaps earlier choices). But Gershwin must have consciously ap-
praised these possibilities as he composed, in order to decide whether to adopt them
or modify them or look for other possibilities; and ultimately he must have con-
sciously appraised the melody itself, in order to decide whether to assent to it as his
composition or to rene it further. What I suggest is that, in appraising the possibili-
ties and the melody, he must have been inuenced by and responded appositely to
gestalts of the possibilities and of the melody and/or chunks of it, which because they
were unique and unprecedented would not have engaged as wholes with pre-existing
rules of any kind. And if so, I suggest, Gershwins adoption of the nal form of the
melody could not have been wholly pre-determined by pre-existing circumstances
and pre-existing laws or rules.
Once this melody had been composed and heard by its composer and others, there
might from this initial hearing be constituted, for the purposes of future cognitive
processes of those persons, computational rules capable of engaging with a gestalt of
that melody as a type: and it could thus be that those persons would then become dis-
posed to respond in the future to the gestalt of the melody, in effect making the melody
a type for the purposes of future computational processing. That is the kind of excep-
tional circumstance referred to earlier; that is, the circumstance where computational
rules that can engage with feature-rich gestalts may be constituted for the purposes of
future cognitive processes of a person. But that circumstance could not exist for any
person before that persons rst hearing of the melody; and I contend that rules sup-
porting apposite responses to the gestalt would generally arise only after the person has
consciously grasped and responded appositely to the gestalt in the rst place. And of
course rules engaging with this gestalt would not engage with appreciably different
gestalts that may arise from appreciably different performances of the melody, when
those different performances are rst heard.
I understand there is a computer program that can compose music in the style of
Mozart, and there may be one that can compose melodies in the style of George Ger-
shwin. Such a program could conceivably come up with melodies as appealing as The
Man I Love. But what it could never do is to appraise and rene its creations by attend-
ing to gestalts of them; because while the rules of a computational program can engage
with aesthetic standards to which its creations should comply, and can engage with all
manner of features which its creations have in common with other things, these rules
cannot engage with whole particular unprecedented gestalts, at least on their rst en-
counter with them. The point is particularly strong in the case of groundbreaking cre-
ations that defy existing aesthetic standards, such as Wagners Tristan und Isolde and
Picassos Les Demoiselles dAvignon. When creating those works, I suggest, the authors
could not have just been giving effect to computational programs of their cognitive
processes and/or applying existing aesthetic standards; but rather they must have been

inuenced, in the course they took in creating and rening these works, by their ap-
praisal of gestalts of the works and of substantial parts of them.
A computer could receive, store and process information concerning each and
every physical feature of Picassos painting (for example); and also information con-
cerning all aesthetic standards that have so far been formulated. It could readily
identify the painting, and it could possibly perform as well as or better than human
experts in determining its conformity to those standards, and also (for example) in
determining whether a painting presented to it was the original or a copy. But what
it could never do is to experience aggregations of features as whole gestalts unique to
that painting, or respond appositely to gestalts of that kind in appraising the
painting.
So I say Gershwin, Wagner and Picasso could and did make aesthetic judgments in
which they responded to whole gestalts, gestalts that had informational content mani-
fested only in conscious experiences and that did not as wholes engage with relevant
rules. And I say all of us similarly can and do make aesthetic judgments in which we
respond appositely to gestalts that do not engage with relevant rules. I say that aesthetic
judgments are one category of judgments of plausible reasoning; and my contention is
that this argument concerning aesthetic judgments can be generalised to all kinds of
judgments of plausible reasoning, and that in making such judgments we can and do
respond appositely to gestalts of conscious experiences that do not engage with any
relevant rules.
5.7. Human beings make aesthetic and other judgments in response to whole gestalts of
conscious experiences, which do not as wholes engage with relevant rules.
In that way, I suggest, conscious experiences make a contribution to decision-making

otherwise than in accordance with rule-determined processes; and thus make a contri-
bution that is not and could not be made by unconscious processes. This provides
specic answers to the rst two of the three questions raised in the previous chapter:
1. What is it that conscious experiences contribute to decision-making that is not

done by unconscious processing? Brief answer: apposite responses to whole gestalt
experiences that do not engage with laws of nature or computational rules.
2. What is it about conscious experiences that enables them to make this contribution?
Brief answer: the co-existence of unied feature-rich gestalt experiences and a con-
scious subject that can respond to them.
I should also here explain how these answers are consistent with the idea that there are
psycho-physical laws that correlate the experiential and the physical aspects of brain

processes, and why it is not the case that, because physical processes of type Xp have a
role in determining what happens, in accordance with laws of nature, any correspond-
ing conscious experiences of type Xe must have the same role, in accordance with the
same laws of nature plus the laws linking Xp and Xe.
It will be recalled that I identied two classes of psycho-physical laws, namely qualia
laws and binding laws. Thus in a particular case qualia laws might link physical brain
processes of the types Ap, Bp, Cp, and so on, with qualia of the types Ae, Be, Ce, and so
on; while at the same time binding laws might link physical brain processes of the type
Np with the binding of these qualia into a gestalt experience that combines Ae and Be and
Ce and so on. My argument is that, although individually each of Ae and Be and Ce and
so on are types with which rules can engage, and although Np is a type with which rules
can engage, the resulting gestalt (Ae + Be + Ce + . . .) is generally, because of its particu-
larity, not a type with which rules can engage. The link provided by binding laws is not
directly between types of brain processes and types of conscious experiences, but
rather between types of brain processes and types of operations or processes affecting
conscious experiences; and thus the resulting gestalt need not itself be a type with
which rules can engage.
LAWS AN D R U LE S
So that is the essence of the argument. Because of the importance to the argument of
the role of laws and rules, I will now look at laws and rules in more detail.
There appear to be regularities in the way that processes unfold in the world, and it
is common to regard these regularities as reecting laws of nature that govern or con-
strain what happens in the world. These laws have not yet been, and may never be,
formulated with precision by scientists, and those laws that have been formulated are
often expressed as relations between variables rather than in terms of what causes
what; but it is generally assumed that there exist laws, to which the formulations of
science approximate, and which are causal in the sense that, in their totality, they con-
strain what changes occur in the world, either deterministically or by restricting these
changes to some range or spectrum of alternatives.
I think it is a matter of convenience whether one considers these laws as being in
some sense distinct from the properties of the features of the world with which they
engage, such as mass, electric charge, spatial separation, direction, motion, and so on,
or as being inherent in the features themselves or in the matter, energy, elds, and so
on, which display these features. But however one regards the laws, it seems clear that
one feature they must have is generality: in order that they be laws, they must be such
as to engage with features that the different states of affairs or objects or processes or
events to which they apply may have in common and/or with mathematical variables.

And just as laws of nature must have generality, so also must the features of the
world with which the laws engage. Very broadly, laws are to the effect that when X
occurs Y will occur; and both X and Y must then be features that a number of differ-
ent states or objects or processes or events may have in common (that is, they must be
types), and/or be mathematical variables.
Generality here does not altogether exclude uniqueness. To the extent that X and Y
incorporate variables that are quantitative and susceptible to mathematical analysis,
the uniqueness of particular quantities involved in particular cases of X and Y will not
preclude engagement with constraining laws. The archetypal laws of nature, namely
the laws of physics, apply to quantities of physical properties such as mass, electric
charge, and so on, and they apply to any such quantities and any combinations of such
properties; and so it does not matter that a combination is unprecedented and will
never occur again, so long as it can be given quantitative expression. The laws prescribe
the quantitative effect of quantitative variations.
Even something as unique and as complex as a particular state of the weather may be
considered as constituted by a limited number of common features of varying quanti-
ties (temperature, pressure, etc.), so that its uniqueness does not preclude the engage-
ment of all relevant features of that state with laws of nature. In relation to the weather,
there is no reason to think that there are any other features that do not engage with laws
of nature and yet do contribute to the determination of how the system changes. (Of
course, my argument is that it is otherwise in relation to conscious experiences.)
To the extent that any X and Y are not quantitative, they must, in so far as they engage
with laws of nature as presently understood, be the same whenever they occur: any
variations within X or Y that are not quantitative will not engage with the law when X
occurs Y will occur. And the more rare and/or more complex any X and any Y are, the
less likely it is that there will be a law of nature that engages with them. At the extreme,
if X is a feature that occurs just once in the history of the universe, then a constraint to
the effect that when X occurs Y will occur would only have the one occasion for its ap-
plication and so would not properly be called a law.
Thus for example, unless there were more than one Big Bang, a constraint that ap-
plied only in the circumstances of the Big Bang could not properly be called a law,
because it would have only the one occasion for its application. But in so far as the
conditions that existed in the Big Bang are quantitative, they may have engaged with
the same laws of nature that apply at other times: it would not matter that the par-
ticular quantities involved in the Big Bang have never existed and will never exist at
any other time. Even if, for example, heat and density of mass in the Big Bang were
innite, the same laws as apply to nite values of these quantities may have applied;
and calculation of the result of application of such laws to innite values might be
possible.

In this discussion, I am making two assumptions, which I should now make explicit.
First, I am assuming that there is indeed a valid distinction between what is quantita-
tive and what is not quantitative. I take it that, in so far as any feature (and by this I
intend to include any aspect of any feature and any difference from any other feature)
can be adequately described by numbers or other mathematical variables, to that extent
the feature is quantitative; and in so far as any feature cannot be adequately so de-
scribed, to that extent it is not quantitative, and may be called qualitative. It is beyond
the scope of this book to analyse this distinction in detail: it is sufcient for my pur-
poses if the distinction is understood in general terms, and if it is accepted that some
features of the world cannot adequately be described by numbers or other mathemati-
cal objects that can be incorporated as variables in mathematical rules.
I do not suggest a narrow limit to what is quantitative: I accept for example that
variations in tones and intensities of visual experiences of colour may be quantitative,
in that such variations may perhaps be adequately described by numbers or other
mathematical objects and incorporated as variables in mathematical rules. However,
I do assume, for example, that differences such as that between a visual experience and
a feeling of pain, or that between a visual experience of a colour and a visual experience
of a line or edge, are not merely quantitative: I assume that these differences cannot be
adequately described by numbers or other mathematical objects. (It may in fact be that
all non-quantitative and non-general features are features of conscious experiences,
but my argument does not require any assumption about this.)
Second, I am assuming that laws of nature have some minimum simplicity and
broadness of application; so that, if a law is to the effect that when X occurs Y will
occur, then, in so far as X and Y are not quantitative, they must to some minimum
extent be simple and commonly occurring. Again, I do not suggest a narrow limit
on this. I do not assume that all laws of nature must be or derive directly from laws
of physics. However, I do assume, for example, that a feature-rich visual experience
is too complex and rarely occurring for it to engage as a gestalt whole with laws of
nature.
I think that both these assumptions are modest. Although some people make wide
claims about what is quantitative, I think everyone accepts that there are at least some
basic entities that cannot be adequately described or distinguished from each other by
numbers or other mathematical variables. And my second assumption is widely made,
albeit often tacitly, in science and philosophy.
What I am suggesting then is that states, objects, processes and events of the world
combine with laws of nature, so as to contribute to the determination of what changes
occur in the world, only through their general and/or quantitative features that can
engage with such laws. And in so far as features of states, objects, processes and events are
neither quantitative nor of some minimum simplicity and frequency of occurrence,

they do not have the generality that is required for engagement with laws of nature so
as to contribute to the determination of what changes occur in the world.
One view of the way causation operates in the world is that there are succeeding
states of affairs, and that laws of nature operate on each of them so as to determine
either a single line of development from it, or alternatively a spectrum of lines of de-
velopment one of which occurs at random within probability parameters determined
by the laws. (In saying this, I am not asserting that states of affairs are basic, and objects
and processes are secondary: it is simply convenient to conduct this discussion in terms
of states of affairs.) On this view, although the whole of any state of affairs, in its full
particularity, is important in the causal process as constituting initial conditions for
subsequent development, the actual line or spectrum of lines of development from
that state of affairs is determined entirely by the combination of the relevant laws of
nature and the general and/or quantitative features of these initial conditions with
which they engage.
Thus, if a state of affairs has a feature or combination of features that does not engage
with any law of nature because it does not have generality in the sense explained above
and/or is not quantitative, then that feature or combination of features cannot be ef-
cacious through engagement with the laws in determining or even inuencing what
changes in the world are brought about: the efcacy of the state of affairs in that respect
must be by virtue of those of its features that do have generality in the sense explained
above and/or are quantitative. (On the other hand, I contend it is reasonable to accept
that features or combinations of features of that kind which arise in experience are ef-
cacious to inuence what changes are brought about in the world, but not through
engagement with laws of nature.)
My earlier discussion has indicated that laws of nature are not the only laws or rules
that can be relevant to the determination of what occurs as a result of processes of
human brains-and-minds. I accept that the physical processes of the brain are instan-
tiations of computational information-processing, in which there could be computa-
tional rules operating as a result of a persons evolutionary and/or personal history;
and that these rules could engage with features that have had some signicance in these
histories.
I referred earlier to a melody pushing buttons in our emotional make-up that have
been established by evolution and environment; and I accept that in cases like that,
there must be brain processes that instantiate computational processes occurring as
determined by computational rules that engage with types of musical sounds. The
relevant types of musical sounds could be types that have had some signicance in the
evolution of human beings, or in the particular history of the person in question.
These types would not need to have law-engaging generality in the sense identied
above; but they would need to be established as rule-engaging types for the person

concerned, and they could not, for example, include a gestalt of a newly composed
melody heard by the person for the rst time.
TH E GAM E OF LI FE AN D COM PUTATION
The operation of a system in which (unlike the human brain-and-mind, as I contend

it to be) the unfolding of events is uniquely determined by laws or rules, can be illus-
trated by reference to the Game of Life. This game was devised in about 1970 by John
Conway, then a Cambridge mathematician. Its rules can be stated shortly:3
Life occurs on a virtual [and potentially innite] checkerboard. The squares are
called cells. They are in one of two states: alive or dead. Each cell has eight pos-
sible neighbours, the cells of which touch its sides or corners.
If a cell on the checkerboard is alive, it will survive in the next time step (or
generation) if there are either two or three neighbours also alive. It will die of
overcrowding if there are more than three live neighbours, and it will die of
exposure if there are fewer than two.
If a cell on the checkerboard is dead, it will remain dead unless exactly three of its
eight neighbours are alive. In that case, the cell will be born in the next generation.
Given an initial conguration of live and dead cells, everything that happens in the
game is determined unequivocally by these basic rules, which may be considered as
analogous to the laws of physics.
Features on a larger scale than the one-cell-and-eight-neighbours scale dealt with by
the rules may be causally efcacious in the sense that they are part of initial conditions
that are modied by the rules engaging with constituent elements of these features, so
as to produce further larger-scale features; but they do not themselves engage with the
basic rules so as to bring about changes. This is true even of features that appear to
unfold in accordance with larger-scale rules, because their causal efcacy in bringing
about changes is only through their properties at the one-cell-and-eight-neighbours
scale. For example, there is a ve-cell pattern called a glider (see gure 5.1) which, after
four generations in which no other live cells are encountered, results in an identical
pattern displaced diagonally by one cell.
The rules of the game dictate that the state of any cell in any generation is wholly
determined by the state of that cell and the eight adjoining cells in the preceding gen-
eration; so the rules do not in fact engage with the glider pattern as such. Of course, the
glider pattern is itself causally efcacious in the sense that it is modied by the rules
engaging with constituent elements of the pattern so as to produce further glider
3
Levy (1993), 52.

Figure 5.1 The glider in the Game of Life
patterns; but there is nothing in the game itself that recognises or responds to a glider as
such, either as cause or effect.
An outside observer may recognise a glider, and may construct a rule that a glider
will continually move diagonally across the checkerboard, progressing by one cell every
four generations, unless it encounters any other live cells. This chunking of the ve live
cells of the glider, and their adjacent dead cells, provides an observer with a useful de-
scription of what is happening, and may help the observer to understand and predict
the games unfolding; but the glider pattern itself is not causally efcacious in bringing
about any changes in accordance with the basic rules of the game, because efcacious
causation of this kind is entirely at the one-cell-and-eight-neighbours scale. The glider
pattern does not as a whole engage with the basic rules.
However, it has been shown that the Game of Life, on a sufciently large scale and
given sufcient time, can operate as a computing machine, capable of solving any com-
putable problem.4 And if the game is set up so as to operate as a computer, there is then
a role in the unfolding of the game for the chunking of cells, as in the glider pattern.
Computers run in accordance with programs that operate at levels higher than the
level of basic physical causation in computers, or the one-cell-and-eight-neighbours
level of the Game of Life; and elements of the programs must then be represented by
and engage with larger-scale features of the system, meaning in the case of the Game
of Life features such as gliders and other larger-scale patterns that can arise in the
game. Furthermore, as mentioned earlier, the same program can be realised and run
on different systems, and so to that extent the programs may be considered as operat-
ing independently of the basic rules of the Game of Life. In that sense, the rules of the
program may fairly be considered as having their own efcacy and signicance, and
also as being independent of the two basic laws of the Game of Life.
However, the rules of the program would never require or permit anything to happen
otherwise than in accordance with the basic rules of the Game; and these rules them-
4
Poundstone (1987), 197217.

selves must engage with general features of the relevant system, such as gliders and
other recurring patterns.
The kind of information-processing carried out by computers as presently understood
(or by unconscious brain processes) can be explained in terms of rules analogous to the
rules of the Game of Life and rules of programs engaging with larger-scale features.
A request for certain information or for a solution to a certain problem may be en-
tered into a computer, giving rise to a state of the computer which, we may suppose, is
unique and unprecedented. The constitution of the computer and its program is such
that rules engage with certain general and/or quantitative features of the state in ques-
tion, and of succeeding states, in such a way as to give rise ultimately to a state of the
computer by which the answer (or perhaps both the request and the answer together)
is displayed.
There is no reason to think that the computer or its program grasps (or understands
or responds to) the request or the answer as a whole, except to the extent that the entire
process I have described could be considered as amounting to such grasping. The tran-
sition from the state representing the request to the state representing the answer is
mediated entirely by the operation of general constraining rules that engage with gen-
eral and/or quantitative features of the initial and succeeding states.
This means that any novelty or uniqueness in objects, processes, or states of affairs
dealt with by the request, and thus reected in the initial state, is not taken into account
by the computer, except possibly to the extent that this novelty may be reected in
numbers or other mathematical variables with which the laws can engage: the initial
conditions may be novel, but the line or lines of development from them would be
constrained by the laws engaging only with their general and/or quantitative proper-
ties. The state representing the answer (or the request and the answer together) may
reliably indicate that the earlier novel state occurred, if it can be understood as doing so.
But for a computing machine, such understanding of the answer would at best be by
means of further processes of the same general kind.
So it seems clear that there could be no grasping by the computer of the request,
unless the entire process from request to answer is taken to be such grasping. And
while the entire process might conceivably be argued as amounting to a grasping of the
request, it could not possibly be considered as amounting to a grasping of the answer:
that would surely require at least some further process. And if the answer is not grasped,
it is implausible to suggest that the request has been.5
It is clear that there is not, in a computer or in the Game of Life, any process whereby,
in the course of the progression from the request to the answer, there is some kind of
systematic chunking of patterns other than general features with which rules can
5
Readers may detect here an echo of Searles Chinese room: see Searle (1984), 3138.

engage, so that they can be grasped as wholes by the computer or the game; much less
any suggestion of the constitution of the computer or the game as a subject that could
do this grasping. And if there were such a process, it would be exceedingly odd. There
is simply no role for such chunking or grasping in the operation of a computer or in
the unfolding of the game. Yet of course, chunking of patterns of this kind occurs all
the time in the conscious experiences we as subjects have, as we progress from ques-
tions to answers in our plausible reasoning.
A general conclusion:
5.8. In rule-determined systems, any features that are not types or variable quantities
engaging with laws or rules cannot contribute to the determination of what happens,
except as being aspects of initial conditions that are modied by the operation of laws or
rules on features that are types or variable quantities.
TR ICKS OF CON SCIOUS N E SS
When I think about our conscious processes in comparison with computational pro-
cesses or with the Game of Life, it strikes me that consciousness seems to involve two
very important tricks, that have no place in computational processes or in the Game
of Life.
First, there is the qualia trick, the trick of associating certain types of neural processes
with certain types of subjective qualitative experiences or feelings. It seems, for exam-
ple, that there are types of patterns of neural activity that correspond with visual expe-
riences of various kinds, and other quite different types of patterns of neural activity
that correspond with feelings of pain. Thus, a pattern of type A could correspond with
a green patch in a certain part of the visual eld, a pattern of type N could correspond
with a vertical edge in another part of the visual eld, and so on.
This rst trick on its own may not seem to have great causal signicance: just as the
type of neural pattern corresponding to a visual experience of green has general fea-
tures that can engage with constraining laws (like the state of individual cells and their
eight neighbour cells in the Game of Life), so also the corresponding visual experience
of green may have general features that could be appropriate to engage with such laws
(like a glider in the Game of Life). I will suggest that it is in combination with the
second trick that the qualia trick could have enormous signicance.
So second, there is the chunking trick. Consciousness also has the trick of bringing
together a multitude of general qualia into a particular global experience that is had by
a particular subject. As mentioned earlier, this trick is associated with what is called the
binding problem of consciousness: how it is that information carried by processes in
distinct parts of the brain is brought together into a unied conscious experience.

Like the rst trick, this one does not require anything other than the operation of
general laws on general features: the laws of nature seem to be such that, whenever a
type of process occurs in a region of a persons brain (perhaps a pattern including the
coordinated 40Hz oscillations of the activity of relevant neurons, which is sometimes
said to be part of the solution to the binding problem), the types of qualia associated
with other types of processes also occurring in that region are chunked together into a
whole gestalt experience of the person.
One possible objection to my delineation of two tricks is to the effect that there can
be no qualia without a subject to experience them, and there can be no subject without
chunking; and thus there cannot be two separate tricks. There is force in that objection,
but I think it is likely that there are (1) types of neural processes that give rise to chunk-
ing (such as processes like 40Hz oscillations) in accordance with binding laws, and (2)
different types of neural processes that give rise to qualia in accordance with qualia
laws, either by themselves or in combination with the processes that give rise to chunk-
ing. If the combination of processes (1) and (2) is required to give rise to qualia, then
the processes in (2) could be considered as giving rise to potential qualia that become
actual qualia when chunked into a whole experience of a subject; and with this modi-
cation the identication of two tricks could stand.
The global experience resulting from the two tricks will display general features with
which laws of nature and/or computational rules could engage, but will also display
particular and often unique combinations of these features. A visual experience, for
example, will display features of colours, shapes, patterns, and so on, that could indi-
vidually engage with laws of nature or applicable computational rules; but it will also
display particular combinations of these features, and we seem to grasp the particular
combinations as wholes, as unied gestalts. It seems clear that such a visual experience
is not sufciently simple or commonly occurring to engage as a whole with laws of
nature, except in so far as it may be quantitative; and I suggest the previous discussion
indicates that, because a visual experience is a combination of qualia, it cannot be ad-
equately described by numbers or other mathematical variables and thus is not quan-
titative. And it also seems clear that such an experience would not generally as a whole
engage with computational rules that may be operating as a result of evolutionary se-
lection and/or a persons history. Therefore, I contend, such an experience would not
generally as a whole engage with laws of nature or computational rules.
Table 5.1 shows this schematically, in relation to a grossly oversimplied account of
seeing a work of art.
Apart from the bottom-right item, every item in the table can plausibly engage with
laws of nature or computational rules, so that it could affect what happens by this en-
gagement. Further, every step or correspondence between items on the left of the table
and items on the right could be in accordance with laws of nature.

Table 5.1 Correlated brain and conscious processes in seeing a work of art
Brain processes Conscious processes
Ap: pattern a in region x Ae: red patch in bottom left of visual eld
Bp: pattern b in region y Be: white patch in bottom right of visual eld

Np: pattern n in region z Ne: vertical edge in top right of visual eld

Rp: pattern r in regions (Ae + Be + + Ne + ): gestalt experience in

x+y++z+ which all visual elements are chunked
If, as I think likely, the relationship between items on the left and items on the right
is a kind of identity (in the sense that both are aspects of the one process), rather than
item on the left causing item on the right, then all items on the right could have effects
in accordance with laws or rules through their correspondence with items on the left:
the items on the right would in that respect be like gliders in the Game of Life, that is,
features that might be regarded as themselves engaging with laws, but might also be
regarded as having effects through lower-level features with which they correspond.
But I contend that this would not generally be the case for the bottom item on the
right. There could not plausibly be a law of nature as to what happens when the gestalt
experience (Ae + Be + + Ne + . . . ) occurs: such an experience has gestalt features
which are neither quantitative nor sufciently simple and commonly occurring to
engage with laws of nature. And these gestalt features would generally not engage with
computational rules, at least unless they had previously been grasped and responded
to. Thus, any engagement with laws or rules by such an experience must generally be
1. through its constituent qualia; and/or

2. through other general and/or quantitative constituent features of the experience;
and/or
3. through its neural correlates.
And that means it must be an engagement that does not operate through any features,
produced by the chunking of the qualia, which do not engage with laws or rules.
I am not saying, as writers like Donald Davidson6 have done, that there can be no rigor-
ous application of general laws to any mental events. I accept that, to the extent that even
qualitative features of conscious experiences have generality, they could engage with and
be rigorously governed by such laws. Rather, I say that what precludes rigorous applica-
6
Davidson (1970).

tion of laws or rules to gestalt experiences is the combined effect of the qualia trick and
the chunking trick, giving rise to particular combinations of qualia which, as gestalt
wholes, lack the generality necessary for engagement with applicable laws or rules.
It is reasonable to believe that chunking has a function. It has been said that its func-
tion is to integrate information, and it may be that chunking of qualia is associated
with physical processes that do integrate information in a way appropriate to compu-
tational operations; but this would not explain the function of the chunking of qualia,
in the absence of an account of how chunked combinations of qualia are used. It is
plain that computers as presently understood could not be programmed to use
chunked combinations of qualia, as distinct from their general and/or quantitative
constituent features.
What I suggest is that the function of these two tricks is, by means of the application
of general laws to general features, to constitute a particular experience with particular
qualities, had by a particular subject, who can then select an apposite development
from a spectrum of possible developments left open by the operation of general laws
on general features. This selection is neither determined by rules nor random, but is
based on the subjects grasp of particular wholes; and it is facilitated by the way in
which combinations of qualia present enormous amounts of information simply,
graphically, comprehensively, and with salience given to important features.
SOM E FU RTH E R THOUGHTS
Its been put to me that I am proposing an extravagant form of emergentism, in which

emergent wholes are governed neither by the laws governing their constituent features
nor by laws of their own; and that my arguments are too weak to defeat the reduction-
ist alternative, according to which we grasp unique wholes by way of grasping their
constituent features.
Its also been put to me that my arguments are insufcient to defeat views such as the
connectionism of Horgan and Tienson which, it may be contended, can account for
plausible reasoning in accordance with the soft constraints of ceteris paribus rules,
giving rise to appropriate responses to unique unprecedented circumstances.
Well, I must plead guilty to proposing what some people may regard as an extrava-
gant form of emergentism: I do indeed contend that the gestalts of conscious experi-
ences can contribute to decisions in a way not determined by operation of laws or
rules, because the subject of the experiences can respond appositely to these gestalts.
I do not of course contend that the causal role of gestalts and the subjects response to
them is free of constraint from laws and rules: I say that unconscious processes, deter-
mined by laws of nature and computational rules, contribute much to human deci-
sion-making, and also that decisions must be within spectra of possibilities constrained

by the operation of these laws and rules. But I accept that, in suggesting that there is a
contribution from a subjects response to gestalts that may be apposite to those gestalts
yet not determined by laws or rules of any kind, I am making a proposal that could be
considered extravagant.
The question then is whether the arguments of this book are strong enough to sup-
port this position. My invitation to readers is that (using their own ordinary human
rationality) they assess the strength of my arguments, including those in this and the
previous chapter, and the extent to which they meet challenges such as those referred
to above. As must appear from what I have written, my own assessment is that they are
strong enough to make the view I am proposing preferable to alternative views.
There would be no point in repeating my arguments here, in order to focus on just
these particular challenges. I will however recapitulate some of them by putting a
number of questions:
1. What is it that distinguishes conscious processes (those cognitive processes that are
conscious experiences) from non-conscious processes (those cognitive processes
that are not conscious experiences)?
2. Do conscious processes function in decision-making differently from non-con-
scious processes, and if so how?
3. Do conscious processes make a contribution to decision-making beyond that made
by non-conscious processes?
4. If no to question 3, what is the explanation for the extensive cognitive resources
expended in providing executive summaries in conscious experiences, and for the
feelings that give us conscious motivation?
5. If yes to question 3, what is this contribution, and what is it about conscious experi-
ences that enables this contribution to be made?
6. If, as Horgan and Tienson contend, the top level of cognitive processing proceeds
in accordance with soft constraints of ceteris paribus rules that do not determine
the outcome of the processing, whereas the dynamical system operating at lower
levels does determine the outcome, what if any contribution does the top level
make to cognitive processing and what explanation is there for its existence?
7. If (as seems clear) aesthetic judgments depend in part upon an appreciation of the
whole aesthetic object and of relationships of constituent features of the object
with the whole, can such judgments occur without a contribution from the grasp-
ing of a gestalt of the whole of such an object, that is additional to the grasping of
constituent features?
8. If (as seems clear) there could be no pre-existing laws or rules engaging with a gestalt
of the whole aesthetic object when it is rst created, how can a grasping of a gestalt
of the whole of such an object contribute to an aesthetic judgment about it?

My contention is that my arguments provide plausible and satisfying answers to these
questions, and that alternative views do not. I would ask readers who nd my position
extravagant to consider these questions carefully, and to consider whether plausible
and satisfying answers can be provided consistently with whatever alternative view
they prefer.
COR E ASS E RTION S AB OUT GE STALTS AN D R U LE S
5.1. Consciousness enables an organism to be responsive to circumstances grasped as

wholes, not just to those constituent features that engage with applicable laws or rules.
5.2. The evolutionary advantage of consciousness is that it enables an organism to deter-
mine and/or shape an apposite response to circumstances facing it that has regard, not
only to features that engage with laws of nature and/or computational rules, but also to
whole combinations of features that are particular and perhaps unique to the circum-
stances and do not as wholes engage with applicable laws or rules.
5.3. Gestalts experienced by a subject, in which many features are combined all-at-once,
have informational content that is manifested and made available for use only in con-
scious experiences.
5.4. Feature-rich gestalts of conscious experiences generally do not as wholes engage with
laws or rules of any kind, but can nevertheless as wholes have a causal inuence because
subjects of experience can respond appositely to them.
5.5. The physical aspect of conscious processes does, in conformity with physical laws,
restrict what can happen to a limited spectrum of possibilities; but in response to the expe-
riential aspect of these processes, the subject of the experiences can control what does
happen within this spectrum of possibilities.
5.6. While laws and rules can apply to individual unique circumstances, they engage
with features of these circumstances only in so far as each of these features is of a type or
class, and/or is a variable quantity.
5.7. Human beings make aesthetic and other judgments in response to whole gestalts of
conscious experiences, which do not as wholes engage with relevant rules.
5.8. In rule-determined systems, any features that are not types or variable quantities
engaging with laws or rules cannot contribute to the determination of what happens,
except as being aspects of initial conditions that are modied by the operation of laws or
rules on features that are types or variable quantities.

6
How Gestalts Promote Rationality
To summarise the argument so far:

I accept that some cognitive processes can be both conscious experiences and
physical processes of our brains, and that there is no experiential change without a
corresponding physical change; and I accept there is accordingly a sense in which
any information contained in our experiences must be contained or encoded in
those physical processes; but I say that this information, as experienced consciously
by us, is characteristically combined into uni ed wholes or gestalts that have infor-
mational content that is manifested and made available for use only in conscious
experiences.
Circumstances encountered and perceived by living organisms do have constituent
features that are common to different circumstances, and thus could be features that
engage with laws or rules. However, each set of circumstances also has combinations of
features that are not of a type or class but are particular and perhaps unique to that set
of circumstances. When a conscious organism such as a human being experiences cir-
cumstances facing it, features of these experiences are combined into unied gestalts,
such as visual experiences combining many features of an observed scene. Since these
feature-rich experiences are combinations of qualia, they cannot be adequately repre-
sented in terms of types and mathematical variables; and since they are particular and
perhaps unique combinations of features, they do not generally as wholes engage with
applicable laws or rules.
My suggestion is that, although these gestalts do not as wholes engage with applica-
ble laws or rules, they may plausibly as wholes make a contribution to our decisions,
because we can respond appositely to them.
I do not doubt that unconscious computational processes of our brains engage with
constituent features of what we experience, and/or representations of them in terms of
types and/or mathematical variables; and I accept that these processes are essential to
our having conscious experiences at all, and are essential in other ways to the determi-
nation of our decisions and actions. But I suggest there is in addition, in our conscious
decision-making and action, a contribution from our response to the grasping in our
conscious experiences of whole combinations of features; and since this response is
98
not determined by laws or rules, it is not one that could be achieved by non-conscious
information-processing.
My suggestion is that the capacity to respond in this way to particular and perhaps
unique combinations of circumstances is advantageous, even though it must also be
fallible because its reliability is not assured by any rules or laws. This advantage ex-
plains why consciousness has been promoted by evolutionary processes. Thus:
6.1. There is a positive role for a persons conscious experiences in decision-making, be-
cause these experiences contribute to decision-making through gestalt experiences to which
the person can respond appositely, even though these experiences do not engage as wholes
with applicable laws or rules.
In previous essays on this topic Ive confessed that I cant explain how we can respond to
gestalts in ways not determined by laws or rules (because this would require a far greater
understanding of consciousness than is available at present), and Ive contented myself
with the assertion that our ability to do so is supported by the very fact that we do experi-
ence whole feature-rich gestalts all-at-once, and by many other reasons. In this book, I
want to say more about how we can do thisthat is, to give at least the beginnings of a
plausible account of how processes that are not rule-determined can contribute to rea-
sonable decisions, an account that does not appeal to anything magical or miraculous.
EVOLUTIONARY OR IGI N S
The rst question I ask is, in what ways could an all-at-once grasp contribute to rea-
sonable decision-making?
I do not suggest that animals other than human beings have free will and responsibil-
ity for their actions: I believe this requires the rational self-consciousness and capacity
for language that only human beings have. But I do suggest that the ability to grasp
whole feature-rich gestalts all-at-once and to respond to gestalts so grasped is a feature
of all consciousness; and I do believe some animals other than human beings have
consciousness, with the three characteristic features mentioned earlier. I believe this is
almost certainly true of mammals and birds, and is probably true of reptiles, amphib-
ians, shes and at least some molluscs (such as octopuses, squids and cuttlesh). Ac-
cordingly, there should be an account of how the ability to respond to a grasp of
feature-rich gestalts could be an evolutionary advantage to animals such as these.
My suggestion is that consciousness gives organisms that have it an advantage in
being able to use conscious experiences in determining what to do, that is, in making
judgments resolving the question What is to be done?, rather than merely responding
automatically to stimuli. Making those judgments may require judgments resolving
How Gestalts Promote Rationality 99

three associated questions: (1) the factual question How are things?, (2) the value
question How should things be?, and (3) the practical question What can I do about
it? I suggest conscious organisms generally use their conscious experiences in address-
ing these questions.
It could be contended that its fanciful to refer to conscious organisms making judg-
ments as to how things should be. Surely, it might be said, they just give effect to the
urges or desires they have been given by evolutionary selection. Even human beings,
according to Hume, act in accordance with their strongest desire.
I have previously explained why I do not agree with this.1 And like human beings,
non-human conscious organisms can also have conicting motivations. There can be
conicts between (1) bodily urges such as pain, hunger, thirst and desire for sex, and (2)
less direct self-regarding motives such as fear of predators, and (3) other-regarding mo-
tives such as care for offspring. When human beings face conicts of this general kind,
there is no conscious quantication of the strength of the opposing motives but rather,
apparently, a judgment as to which should prevail; and unless the resolution of the con-
ict is entirely unconscious, there must be a judgment that can reasonably be character-
ised as a judgment as to how things should be, so far as concerns what is to be done.
If (as I believe) the resolution of similar conicts in animals with consciousness is
similarly not entirely unconscious, then I suggest it is reasonable to characterise their
conscious determinations of what they are to do as involving judgments of how things
should be; and also reasonable to hold that there is no relevant strongest desire unless
and until such a judgment is made.
So I say:
6.2. Conscious organisms make judgments resolving the question What is to be done?,
rather than merely respond automatically to stimuli.
6.3. Making such judgments may require judgments resolving three associated questions:
(1) the factual question How are things?, (2) the value question How should things be?,
and (3) the practical question What can I do about it?
6.4. Conscious organisms generally use their conscious experiences in addressing these
questions.
Returning to the question of how consciousness is advantageous to animals that have

it, I suggest it is generally conducive to an organisms survival and reproduction, and
1
I think Hume-like approaches are also erroneous in suggesting a sharp distinction between judg-
ments as to how things are and judgments as to how things should be. I contend that judgments as to
how things are and judgments as to how things should be are often inextricably intertwined, and also
that judgments as to how things should be are amenable to reason, as are judgments as to how things
are. I will say more about this in chapter 10.

the survival and reproduction of its offspring, that it be able to act on the basis of
reasonable information as to how its situation is and how its situation should be.
Rule-determined information-processing is plainly very important in identifying fea-
tures of an organisms situation and assessment of their relevance to what it should
do; but I suggest that there is more than this to optimal decision-making by conscious
organisms.
Suppose for example an organisms senses give it the information that its situation
has the features A, B, C, D, E and F, among others. A previous situation in which it
encountered a predator had the features A, B, C, D, E and G, among others. In both
situations, food was available. My suggestion is that the ability to experience these
features A, B, C, D, E and F all-at-once as a gestalt could assist it to make a judgment
whether the present situation is sufciently like the previous situation with features A,
B, C, D, E and G to indicate a substantial danger that a predator is present; and also
whether the danger that a predator is present is sufcient to outweigh the desirability
of obtaining the food. Of course, either a digital computational system or a dynamic
analog connectionist system could process information comprising a vast amount of
recurring features of situations like these, and give a computed result based on that
processing; but it could not otherwise make a decision based on a judgment about the
relevance and sufciency of similarities and differences as between such situations.
Such judgments would be fallible; but could have a better chance of being correct than
a decision based wholly on unconscious information-processing of constituent recur-
ring features of each situation.
I recognise that rule-determined information-processing can apply to simple gestalts
to which organisms react on the basis of evolution-selected computational processes
of the brain: for example, the red dot on yellow ground that induces pecking in infant
herring gulls. One may also think of spider shapes and snake shapes that may trigger
fear in some mammals, including some humans. But mammals generally also have the
capacity to experience feature-rich gestalts of whatever it is that conveys such a simple
fear-triggering gestalt, and those feature-rich gestalts would not generally be the sub-
ject of rule-determined information-processing. I suggest this capacity to experience
feature-rich gestalts can contribute to an accurate assessment of whether in truth there
is danger from a spider or a snake.
Theres a saying that a picture is worth a thousand words. All the information pre-
sented by feature-rich gestalts is carried in physical brain processes, but only as a large
quantity of words, in the sense of data concerning recurring features that can be ma-
nipulated by computational processes of the brain. The feature-rich gestalts give or-
ganisms a picture which is, in the case of at least some mammals and some birds, a
three-dimensional picture of a stable world existing continuously and changing
smoothly. I suggest an accurate and comprehensive three-dimensional picture of an

object and its immediate surroundings, experienced all-at-once, can be more usefully
informative than the mere words of the data that the brain manipulates.
More generally, information that an organism has from its senses about how its situ-
ation is, and information that it has about how its situation should be, (1) may not be
accurate, (2) will generally include things that are not relevant to how the organism
should act, and (3) will generally be far from complete in relation to those things that
are relevant. Accordingly, an organism needs to be able to check the accuracy of its
information, to be selective about the information it has, and to extrapolate from in-
formation it has so as to derive other information relevant to how it should act. My
suggestion is that the ability to grasp and respond to gestalts has advantages in at least
these three respects.
Plausibly, one advantage is that it assists the organism to make judgments about
what is signicant in the information it has, and thereby to disregard things that arent
important, and to draw inferences, for example through analogies with other situa-
tions it has encountered, about further relevant matters.
Another advantage (or perhaps another aspect of the same advantage) could be that
the grasp of whole feature-rich gestalts provides support for judgments as to whether
information given by the senses is accurate information about things that are real, in
the sense of being things actually existing outside the organism and capable of affect-
ing the organism. The all-at-once experience of many features of an observed object,
and sufcient similarity of this experience with all-at-once experiences occurring on
other occasions, could (along with judgments of the type considered in the previous
paragraph) conrm to the organism that it has accurate information about a real con-
tinuing object, to be taken seriously as such.
So I say:
6.5. The ability to grasp and respond to gestalts helps an organism to check the accuracy
of its information, to be selective about the information it has, and to extrapolate from
information it has to other information relevant to how it should act.
6.6. In particular, this ability assists the organism to make judgments about what is
signicant in the information it has, and thereby (1) to disregard things that arent impor-
tant and (2) to draw inferences, for example through analogies with other situations it has
encountered, about further relevant matters.
AE STH ETIC J U D GM E NTS
In the case of judgments made by human beings, I think the importance of the all-at-
once grasp of feature-rich gestalts is most obvious in the case of aesthetic judgments,
such as judgments as to the merits of an artistic work or a piece of music or literature.

I said a little about this in the previous chapter, and now I will say something more
about how gestalts contribute to judgments of that kind.
When I refer to judgments about the aesthetic merits of an object, I am referring
both to judgments made by the artist in creating the object, and also to judgments
made by other persons in appraising or simply enjoying the object; and I am referring
to judgments ranging from bare judgments of approval or disapproval through to
elaborately considered and articulated assessments of the objects merits. The artistic
object in question may be of any kindliterary, musical, visual, and so on. The argu-
ment is perhaps clearest in relation to the creation and early appreciation of ground-
breaking works such as Picassos Les Demoiselles dAvignon and Wagners Tristan und
Isolde, but it applies to artistic works and gestalts of all kinds. It will be recalled that in
the previous chapter I made reference to the Gershwin melody The Man I Love.
Undoubtedly, when we look at a painting such as Picassos Les Demoiselles dAvignon
or listen to a piece of music such as the Prelude to Wagners Tristan und Isolde, we do
grasp all-at-once some combinations of features, combinations that are particular to
the work in question and cannot (at least on the rst occasion we encounter the work)
engage as wholes with any applicable general laws or rules. It is inconceivable to me
that this grasp does not contribute to the judgments we make about the aesthetic
merits of the work.
If it did not, then such judgments could be no more than the totality of rule-
determined responses to each and every one of a multitude of constituent features
of the work that are common to many works and can engage with laws or rules; and
this would exclude judgments of aesthetic merit based on the aesthetic effect of
particular combinations of features of the work that cannot engage with laws or rules.
I believe this would make inexplicable the aesthetic merit of works such as those
Ive mentioned, which have minimal connection with anything benecial in our
evolutionary history and which, when they were rst created, outed established
aesthetic standards.
Consider what kinds of conscious and unconscious processes are involved in a per-
sons making an aesthetic judgment. Undoubtedly there are extensive and highly com-
plex unconscious processes that must take place in the persons brain if any such
judgment is to occur.
1. There are the complex pre-conscious processes that are necessary in order that
one have a conscious experience of the work. In the case of the visual arts, there are
all the pre-conscious processes that produce the appearance of a viewed object or
scene, not merely processing its conguration but also coordinating data from both
eyes so as to give depth, and compensating for eye movement so as to give stability.
These pre-conscious processes also include the processes that enable ones memory

and understanding of the world to contribute to the recognition and interpretation
of what is seen, and thus in turn contribute to how it looks.
2. There are also pre-conscious processes that give rise to the emotions that seem to be
involved in aesthetic judgments. The person will to some extent become conscious of
these emotions, but the conscious feelings result from and express much pre-conscious
activity.
3. When a person sets about making a judgment concerning what the person is expe-
riencing (seeing, hearing, etc.) and feeling, there is plainly a great deal of non-con-
scious information-processing underlying whatever conscious thoughts may occur.
In addition to these unconscious processes, there are also subjective conscious pro-
cesses that seem to be involved in the making of a judgment. In saying this, I am not of
course asserting that these subjective conscious processes are wholly distinct from ob-
jective brain processes: as noted earlier, I believe that the conscious processes and their
neural correlates are best considered as two aspects of the same processes.
1. When a judgment is made about an artistic work, the work is generally, for some
time at least, present to consciousness. One doesnt appraise a painting without actu-
ally seeing it and consciously taking it in, and one doesnt appraise a musical perfor-
mance without hearing it and attending to it.
2. Such judgments generally involve some emotions that are consciously felt. I have
said that there must be unconscious processes that give rise to emotions involved in the
judgment, but I would maintain that at least some of these emotions are present to
consciousness as the judgment is made.
3. If the judgment is a carefully considered and articulated judgment, there are gener-
ally conscious thoughts that take place in which one formulates and assesses reasons
for the judgment. Whether these conscious thoughts add anything to their neural cor-
relates and other non-conscious processes, and if so how, is a moot point, on which I
hope this book sheds some light; but I think it is undeniable that such conscious
thoughts do occur.
Many of the processes I have mentioned as taking place in aesthetic judgments engage
with laws of nature and/or computational rules and are thereby constrained by them.
It is reasonable to accept that pre-conscious processes underlying the experience of
seeing or hearing, and underlying emotional feelings, are computation-like processes
that proceed in accordance with the constraints of laws or rules which engage with the
general features of the objects, states of affairs and processes involved. Further, the re-
lationship between the neural correlates of elements of the conscious experiences, and
these conscious elements themselves, seems to be law-like, so that general features of

the elements of conscious experiences correspond in a regular way with general fea-
tures of the neural correlates.
As we have seen, it seems that there are types of patterns of activity in the brain that
correspond with types of patterns of conscious experience. It is reasonable to think
there would also be general features of conscious experiences or their neural correlates
that engage with genetic and/or learned predispositions in the brain so as to contribute
to emotions and judgment. For example, types of patterns in the visual eld suggestive
of human faces, or of types of landscapes benecial to survival,2 whether they are con-
sciously recognised as such or not, would presumably contribute to emotional re-
sponses in rule-determined ways.
Thus laws and rules can engage with all kinds of general features of pre-conscious
brain processes, and of experiences and their neural correlates; and they can engage with
every feature necessary to give rise to and uniquely specify, dene and/or determine every
possible conscious experience. Yet the conscious experiences themselves are particular
and often uniqueunprecedented and never-to-be-repeated. And although all the gen-
eral features that give rise to and specify and together determine a particular experience
can engage with constraining laws and rules, the particular experience itself and its spe-
cial features generally cannot do so, because they lack the necessary generality.
Consider a painting like Picassos Les Demoiselles dAvignon. As an object, it is unique,
particular, and unprecedented. It is constituted by basic constituents (atoms, molecules)
in a conguration that can be expressed in mathematical terms representing the rela-
tionships between those constituents; and on a larger scale, it has general features, such
as basic colours, edges, and shapes, with locations and relationships that can be ex-
pressed in mathematical terms. On the basis of these general features, computational
processes could identify the painting and produce passable reproductions of it. But we
can have a visual experience of the painting, constructed by law-governed processes
arising from the engagement of laws and rules with general features of the painting and
of our brains; so that, through the tricks of consciousness previously identied, general
features are chunked into whole feature-rich gestalts that we can experience and grasp.
We can thus experience and grasp unique qualities of the painting, such as the look
of particular faces, and the paintings overall appearance, just because these qualities
are captured by the combination of the qualia trick and the chunking trick. But I sug-
gest that the experience of these unique qualities cannot engage with laws of nature
and generally does not engage with computational rules; so either the experience is not
taken into account and does not thereby inuence our appraisal of the painting, or it
is taken into account and thereby inuences our appraisal otherwise than by engage-
ment with laws or rules.
2
See generally Dutton (2009).

When this painting was rst created by Picasso, and when it was appraised by early
viewers, we may suppose that both Picasso and the viewers were inuenced by emo-
tions triggered unconsciously by law-governed processes. But it is also certain that they
did experience gestalts that expressed qualities of the painting and various parts of it;
and the utter novelty of these gestalts conrms that any response to the gestalts could
not have been entirely the result of general rules engaging with them, whether they be
laws of nature applying to physical processes, or evolution-selected computational
rules, or aesthetic laws determining what counts as a meritorious painting. (I am not
of course suggesting that the historical context in which the painting was created is
unimportant: plainly, the readiness of Picasso and early viewers to respond positively
to the painting was due in part to that context.)
To take another example: consider the rst time Richard Wagner grasped the open-
ing bars of the Prelude to Tristan und Isolde, when he conceived them in his imagina-
tion or perhaps when he played them on his piano for the rst time. We may assume
that there were general features of this experience or its neural correlates that triggered
emotional responses in law-like ways; but if Wagner experienced the gestalt of the
unique unprecedented whole of that passage of music, Wagners grasp of that gestalt
could not engage with laws or rules, because of the originality, uniqueness and particu-
larity of that gestalt. So we must say either that Wagners grasp of the gestalt was not
taken into account by him and did not inuence his judgment about the aesthetic
merits of the passage and his consequent adoption and retention of it in his composi-
tion, or that it was taken into account, and did inuence his judgment, otherwise than
by virtue of engagement with laws or rules.
The point can be further illustrated by reference to the Game of Life. For any pattern
in the game, no matter how signicant it is for an outside observer, or how striking its
effects appear to be, to the game itself it is only an initial condition that is operated on
by rules at the one-cell-and-eight-neighbours scale, or larger-scale computational
rules. Otherwise, there is no chunking of the pattern or grasping of the pattern, in such
a way that the pattern as a whole, or some systems grasping of the pattern as a whole,
can inuence the way the game develops.
The question thus arises, are the originators or appraisers experiences of gestalts of
Les Demoiselles or of the opening bars of Tristan merely initial conditions of this kind?
This seems to me improbable in the extreme, because surely it is of the essence of aes-
thetic judgments that they are based at least in part on gestalt experiences of the artistic
work. And if such a gestalt experience is not merely an initial condition of this kindif
the gestalt has a role in that it is taken into account and thereby inuences the origina-
tors or appraisers aesthetic judgmentthen that role cannot be one in which the
outcome is wholly determined by general constraining laws, because the gestalt cannot
engage with those laws.

The point can be stated starkly in this way. It must surely be the case (1) that the
aesthetic qualities of a work of art include qualities that inhere in its particular com-
binations of features, combinations that may be unique to that work of art (and ac-
cordingly unprecedented); (2) that we do in fact grasp combinations of this kind in
our conscious experiences of works of art; and thus (3) that we base our judgments of
aesthetic merit at least in part on apposite responses to these combinations of fea-
tures. When as is often the case the particular combinations are unprecedented, they
cannot engage as wholes with pre-existing laws or computational rules governing
processes of the persons brain, so any appositeness of the response must derive from
gestalt experiences of these combinations of features otherwise than by engagement
with laws or rules.
How then could the grasping of gestalts contribute to aesthetic judgments? Ive sug-
gested that in conscious organisms generally, the grasping of gestalts contributes to
judgments as to how things are and how things should be, among other things by pro-
moting recognition of what is signicant in information they have, thus assisting dis-
regard of irrelevancies and drawing of inferences. In human beings, judgment as to
how things are include judgments as to truth; and judgments as to how things should
be include judgments as to the beauty of objects, as well as judgments as to such things
as the rightness of actions and the goodness or virtue of people.
I suggest that judgments of aesthetic merit of a work often involve judgments as to
beauty, and also judgments as to truth, especially truth that is other than direct and
literal. At least one aspect of a judgment that a work is beautiful is a judgment that it is
something that is as it should be, that it is something that should exist as it is. Another
aspect of some aesthetic judgments is that the work conveys some truth about the way
things are or the way things should be, a truth that may otherwise be unclear or even
inexpressible, doing so perhaps by highlighting or exaggerating some features, or per-
haps by providing suggestive analogies or metaphors.
For example:
1. Part of the aesthetic merit of an impressionist painting, in addition to its own in-
trinsic beauty, is that it exaggerates evanescent lighting effects in such a way as to
convey truths about the way things are, truths that are not conveyed by precise photo-
graphic images.
2. Language evolved in everyday dealings among people, and its ability to convey with
precision the way things are and the way things should be is incomplete, especially in
relation to matters remote from the matters of common interest to people involved in
the development of language. The conveying of truths about such matters can often
only be by way of allusion, analogy and metaphor; and the capacity of a literary work
to do so can be the subject of aesthetic judgment.

So:
6.7. Human judgments of aesthetic merit of a work often involve judgments of beauty,
and also judgments of truth, especially truth that is other than direct and literal; and these
judgments are assisted in the ways referred to in assertions 6.5 and 6.6.
Of course, one important aspect of the aesthetic merit of works of art or music or
literature is that they engage our emotions, and it might be thought that this must be
by way of pushing emotional buttons based on our evolutionary origins, rather than
by way of conscious judgments. I accept that this is an important part of the story;
but I contend that a substantial part of our emotional responses to these works
comes from our understanding of what the works are conveying, which in turn de-
pends on judgments of the kind Im considering, based on all-at-once grasping of
feature-rich gestalts.
I will not take this matter further here. Im not attempting to develop a comprehen-
sive theory of aesthetics, but merely to indicate how the grasp of gestalts may contrib-
ute to aesthetic judgments, in ways that are substantive, yet are neither magical nor
such as to displace the rule-based information-processing that must also contribute to
such judgments.
PLAUS I B LE R EASON I NG
Much of what Ive said about the use of gestalts by conscious organisms and about
aesthetic judgment can be applied to human plausible reasoning generally. My conten-
tion is that our conscious experiences of feature-rich gestalts contribute to plausible
reasoning, in ways such as those previously outlined; for example, in assisting us to
make judgments as to what is signicant in information we have, and in supporting
judgments that information given by the senses is accurate information about things
that are real.
I suggest that judgments as to what is signicant in information we have enable us (1)
to make reasonable generalisations from particular observations, (2) to draw reason-
able conclusions from analogies, and otherwise (3) to draw reasonable inferences from
information we have as to matters additional to that information.
The use of analogy and the drawing of inferences based on similarities have been
attacked by some philosophers. For example, Nelson Goodman in his article Seven
Strictures on Similarity3 contended to the effect that the concept of similarity is either
hopelessly vague or else superuous. His seventh stricture (at 25) was Similarity cannot
3
Goodman (1970).

be equated with, or measured in terms of, possession of common characteristics. He
pointed to the indeterminacy and ambiguity of statements that two things are similar,
unless the respects in which they are similar is specied. He continued (at 27): when to
the statements that two things are similar we add a specication of the property they
have in common, we . . . remove an ambiguity, but . . . we render . . . superuous the ini-
tial statement about similarity.
My contention is that this approach doesnt do justice to the way we reasonably
generalise, use analogies, and draw inferences. In particular, it doesnt recognise
that we use analogies by judging that two things are relevantly or sufciently similar,
and that this depends not just on identication of common characteristics, but also
(and especially) on judgments comparing whole combinations of features of each
of the things under consideration. This in turn requires the grasping of whole par-
ticular feature-rich gestalts, and judgments about what is signicant in each of
these wholes.
Judgments as to whether information given (or apparently given) by the senses is
accurate information about things that are real can also, I suggest, be assisted by the
conscious grasp of feature-rich gestalts, in the ways suggested earlier in relation to
non-human animals. Indeed, I suggest that our very concept of reality is that it in-
cludes things of which we can have all-at-once gestalt experiences; and also that the
conscious grasp of gestalts is necessary for our understanding that some of our experi-
ences are of real things that lie beyond the experiences, that is, for what is called the
intentionality of some experiences. I have suggested that truth is a property of asser-
tions, consisting of correspondence with reality by way of meaning; so that judgments
of truth require a grasp of the meaning of the assertion, a grasp of the reality about
which the assertion is made, and an ability to compare the two. I contend that this re-
quires the ability to grasp and respond appositely to gestalts.
Another way in which the grasp of gestalts may contribute to plausible reasoning is
in assisting judgments as to whether something experienced relevantly or sufciently
approximates to an objective, or to an ideal such as beauty or truth. This may be lik-
ened to what often happens in artistic performances, where the performer has a con-
ception as to how the performance should go, and seeks to approximate the
performance itself to that conception. Ive previously4 given the example of a pianist
who concentrates intensely to shape the performance: consciousness comes too late
to direct the ngers to the right keys, but not too late to shape the performance to the
pianists conception.
An important aspect of plausible reasoning is that it apparently enables us to reach
decisions that resolve conicting reasons, which are of different types and cannot be
4
Hodgson (1996b).

explicitly compared on a common scale. I adverted to this in the case of non-human
animals; and certainly in the case of human beings, we often have to resolve conicts
between self-regarding motives and other-regarding motives, which cant be explicitly
quantied and compared on a common scale. Unless these conicts are resolved en-
tirely by unconscious computation (as I contend to be most unlikely), its reasonable
to think that a contribution is made by our grasp of whole feature-rich gestalts, and
by our ability to recognize what is important in them and to compare them, and to
make judgments as to how things should be and as to approximations to how things
should be.
In making judgments of this kind, we attend to gut feelings, or somatic markers as
they have been called by neuroscientist Antonio Damasio.5 However, Damasio treats
these feelings as merely directing attention to alternatives for consideration, whereas
I contend they play an important role in resolving decisions between the alternatives,
not as themselves determining choices but nevertheless to be taken seriously in assist-
ing our judgments as to what is important.
All of these things together can give us understanding of areas of intellectual concern,
including the capacity to make sound judgments going beyond knowledge of particu-
lar facts and other facts that could be derived from them by purely logical processes. It
is that kind of understanding that mathematician Roger Penrose has contended6 is not
available to computers, and which I contend is facilitated by the capacity to grasp fea-
ture-rich gestalts.
Mathematics may be another area where the capacity to grasp and respond appo-
sitely to gestalts contributes to reasoning and understanding.
Gdels incompleteness theorem shows that, for any formal system (and thus for any
information-processing system determined by initial conditions and rules), there is a
proposition that is true but not provable by the system. Penrose has advanced elabo-
rate arguments, based on this theorem, for the proposition that there is more to con-
scious decision-making than computational processes.7 Those arguments have been
strongly criticised, and I will not discuss them here, beyond saying that no reason has
been shown why a human being, with high intelligence and mathematical and logical
expertise, would be unable to follow the force of a Gdelian argument showing that,
for a specied formal system, there is a proposition that is true but not provable by that
system. The possibility of a contradiction (on the basis that this human being does at
the relevant time itself constitute that formal system, and would yet by this proof come
to be satised of the truth of this proposition) would then be avoided by my position,
5
Damasio (1996), at 17374.
6
Penrose (1999), at 10317.
7
See generally Penrose (1994).

that conscious decision-making is not wholly rule-determined, so that human beings
cannot constitute formal systems.
More recently, computer scientist Robert F. Hadley has argued8 that the acquisition
of human-like concepts of countable and non-denumerable innities, and human-like
understanding of some geometrically motivated proofs, require conscious apprehen-
sion of the subject matter involved. In relation to the latter in particular, he stresses the
need for gestalt perceptions, unavailable to non-conscious systems. Hadley does not go
into the question whether human use of these gestalt perceptions is or is not rule-de-
termined; but in my contention, if this use were entirely rule-determined, it could at
least be simulated, and in that way given effect to, by a non-conscious system, so that
Hadleys contentions, if correct, do support my position.
So:
6.8. The grasp of gestalts contributes to human judgments as to what information is

signicant (promoting reasonable generalisations, reasonable use of analogies, and rea-
sonable inferences generally), as to whether information given by the senses is accurate
and whether something experienced relevantly or sufciently approximates to an objective
or ideal, and as to how to resolve incommensurable reasons; and it thereby assists under-
standing of areas of intellectual concern.
CONCLUS ION
I accept that my account of how conscious processes that are not rule-determined can
contribute to reasonable decision-making is far from complete, and that it leaves many
questions unanswered. In particular, I have not gone far in explaining how or why
these contributions can be apposite.
The contributions to decision-making by unconscious processes can generally be ap-
posite because they come from physical structures selected by evolution for how they
change through engagement with laws of nature, and for how they can thereby give
effect to useful computational procedures; whereas, it may be asked, if conscious pro-
cesses are not determined by laws or rules of any kind, exactly what is it about them
that can make their contributions apposite?
I have suggested that conscious experiences assist conscious subjects in making judg-
ments about such things as signicance of information, relevance and sufciency of
approximations, and resolution of inconclusive reasons. It may be said that Ive given
no account of how conscious experiences can give this assistance: how is signicance
8
Hadley (2010).

detected and assessed, how are approximations appraised, how are incommensurable
reasons resolved? Indeed, it may be asked, why, if conscious experiences can do things
as sophisticated as making these kinds of contributions, do they need physical and/or
computational underpinnings at all?
On this last question, my answer is that I do not say that the conscious aspect of the
relevant processes achieves anything on it is own: in all conscious decision-making, the
contributions of the physical and mental aspects of the relevant processes are comple-
mentary. While entirely unconscious processes can achieve useful information-pro-
cessing without any contribution from consciousness, I do not suggest that the mental
(conscious) aspect of those cognitive processes that are conscious can achieve any-
thing, independently of their physical (non-conscious) aspect. My contention is that
reasonable conscious decision-making is achieved by the whole physical-and-mental
process, not by either aspect on its own.
Otherwise, there is little more I can say on questions such as these: satisfactory
answers will indeed require greater understanding of consciousness than is available
at present. But I contend that what I have said in this and previous chapters is enough
to make out a powerful case for holding (1) that human plausible reasoning is both
rational and not wholly determined by the operation of laws or rules of any kind;
and thus (2) that human beings can make reasonable decisions that are not deter-
mined by pre-decision circumstances and laws of nature, and (3) that such non-de-
termination is not mere randomness but derives from the conscious subjects
reasonable response to feature-rich gestalts that do not engage as wholes with laws
or rules. As I will argue later, this goes a long way towards showing that we do have
free will and are truly responsible for our actions. Certainly, it shows how good sense
can be made of a third way in which events occur in the world, which is neither
deterministic nor random.
It may be objected that this does not take account of what science tells us about how
the world works, and in particular what neuroscience tells us about how the brain
works. So in the next two chapters, I set out to show that science does not provide a
substantial argument against these conclusions. In particular, in the next chapter, I
argue that physics is consistent, and indeed coheres well, with those conclusions; while
in the following chapter I complete my non-magical and non-mysterious account of
how processes of the brain and mind that are not rule-determined can contribute to
reasonable decisions, and I argue that this account is not seriously challenged by
neuroscience.
The brief conclusion of my discussion so far:
6.9. Human beings can make decisions that are reasonable albeit not determined by pre-
decision circumstances and laws of nature.

COR E ASS E RTION S AB OUT HOW
GE STALTS PROMOTE RATIONALITY
6.1. There is a positive role for a persons conscious experiences in decision-making, be-
cause these experiences contribute to decision-making through gestalt experiences to which
the person can respond appositely, even though these experiences do not engage as wholes
with applicable laws or rules.
6.2. Conscious organisms make judgments resolving the question What is to be done?,
rather than merely respond automatically to stimuli.
6.3. Making such judgments may require judgments resolving three associated questions:
(1) the factual question How are things?, (2) the value question How should things be?,
and (3) the practical question What can I do about it?
6.4. Conscious organisms generally use their conscious experiences in addressing these
questions.
6.5. The ability to grasp and respond to gestalts helps an organism to check the accuracy
of its information, to be selective about the information it has, and to extrapolate from
information it has to other information relevant to how it should act.
6.6. In particular, this ability assists the organism to make judgments about what is
signicant in the information it has, and thereby (1) to disregard things that arent impor-
tant and (2) to draw inferences, for example through analogies with other situations it has
encountered, about further relevant matters.
6.7. Human judgments of aesthetic merit of a work often involve judgments of beauty,
and also judgments of truth, especially truth that is other than direct and literal; and these
judgments are assisted in the ways referred to in assertions 6.5 and 6.6.
6.8. The grasp of gestalts contributes to human judgments as to what information is
signicant (promoting reasonable generalisations, reasonable use of analogies, and rea-
sonable inferences generally), as to whether information given by the senses is accurate
and whether something experienced relevantly or sufciently approximates to an objective
or ideal, and as to how to resolve incommensurable reasons; and it thereby assists under-
standing of areas of intellectual concern.
6.9. Human beings can make decisions that are reasonable albeit not determined by pre-
decision circumstances and laws of nature.

7
Science and Determinism
The plausible arguments presented so far might have to give way if science were
strongly against them. In this chapter and the next, I argue that science is far more ac-
commodating of free will than is often supposed.
In this chapter I will be focussing mainly on the implications of science for determin-
istic views about the universe, and I will also say something about non-local causation
(I will explain later) and about time.
I take determinism to be the doctrine that everything that happens is xed (deter-
mined) in advance. Broadly, there are two versions of determinism, which can be as-
serted either independently or in combination. One has it that earlier circumstances
and the laws of nature uniquely determine later circumstances. This version found classic
expression in the writings of the eighteenth-century French mathematician Pierre
Laplace. The other has it that past, present and future all exist tenselessly in an un-
changing block universe, in which the passage of time is merely apparent and nothing
can ever be other than it is. This version is considered by many scientists and philoso-
phers to follow from relativity theory, which treats time and space as interdependent
dimensions in a reality of four or more dimensions, in which every event of the past,
present and future has a location in an unchanging space-time continuum.
The rst version directly contradicts the conclusion, from my arguments so far, that
human beings can make reasonable decisions that are not determined by pre-decision
circumstances and laws of nature. The second version also directly contradicts this
conclusion, in that decisions are determined by the pre-decision circumstances alone
(the existence of the unchanging block universe), so that the reference to laws of nature
becomes superuous. In addition, the second version is hard to reconcile with any idea
of human beings making decisions, that is, of human decision-making being efcacious
in determining what happens.
Of course denial of determinism does not of itself give any positive support to ideas
of free will: the point has often been made that, to the extent that quantum mechanics
suggests indeterminism, it merely suggests randomness, which is inimical to free will
rather than supportive of it. And there are respectable philosophical views that deter-
minism is in any event compatible with free will. However, the account of free will that
114
I am giving is inconsistent with determinism, so it is important to my account to show
that determinism is probably false; and I set out to do this in this chapter.
A LAWFU L U N IVE R S E
I will start by briey considering the support given by science to the rst version of
determinism.
As suggested in chapter 5, there are compelling reasons to believe in the existence of
laws of nature, and in particular laws that constrain what states of affairs can exist in
the world and how those states of affairs change over time. The existence of these laws
is also supported by the scientic method of conjecture and refutation: this method
presupposes the existence of laws, because it depends on the assumption that similar
experiments will produce similar results, in accordance with relevant laws that
determine that this happens; and the resounding success of this method conrms this
presupposition on which it is based.
Particularly over the last four hundred years, one of the main projects of science has
been to discover and formulate these constraining laws; and this endeavour has re-
sulted in the formulation of laws that approximate ever more closely and with ever
more assurance to these actually existing laws, and in ever-increasing knowledge and
understanding of how things happen in the world.
A starting point for me is Descartes, who postulated that everything in the world,
apart from the immaterial minds or souls of human beings, was subject to the con-
straint of laws of nature, thus excluding, even from animals, states or activities that
were not subject to the constraint of these laws; so that a central project for science was
to discover these laws and to explain the world by reference to them.
This idea had powerful expression later in the seventeenth century in the physics of
Isaac Newton. Newtons three laws of motion were proposed as applying without ex-
ception to all physical matter, and they suggested that, given the position and motion
of any piece of matter at one time and the forces acting on it over a period following
that time, its positions and motion during that period were uniquely determined.
Newtons law of gravitation gave rules determining the quantity and direction of the
force of gravity at any point. Formulation of similar rules for all other forces would
then, apparently, complete a scheme according to which all changes over time of physi-
cal systems are determined by universal laws of nature.
Newton did not spell out what that might mean for the physical matter of the
human brain or for the dualism of Descartes. However, the eighteenth-century French
mathematician Pierre Laplace pointed out that, under the Newtonian scheme, initial
conditions plus laws of nature determined the future, which could thus be exactly
calculated by a being with sufcient information and intelligence; and he made no
Science and Determinism 115

exception for the human brain and left no room for any independent efcacy of
human choice.
During the nineteenth century, progressive developments in physics, chemistry, and
biology gave further support to the view that all matter was subject to the constraint of
universal laws. Three crucial developments were the explanation of heat, Maxwells
theory of electromagnetism, and Darwins theory of evolution.
Heat as a physical feature was shown to consist in the motion of molecules; and so-
lidity and liquidity were explained in terms of the lattice structure of molecules in
solids, and the breaking down of that structure as heat increased the motion of the
molecules; and this opened the way for these phenomena to be explained by a combi-
nation of Newtons physics and statistics.
Maxwell showed how the forces associated with electromagnetic radiation, including
visible light, could be calculated by reference to mathematical rules, analogous to
though more complicated than Newtons law of gravitation. The effect of electromag-
netic forces on matter then followed from Newtons laws of motion. This was a giant
step towards a comprehensive basic theory of physical forces, along the lines suggested
by Newton.
Darwins contribution was even more momentous, offering an explanation both of
how the huge complexity and variety of life could have arisen from simple beginnings
by changes over time constrained by laws of nature, and of how animals and humans
could have come to give the appearance of making choices, even if they consisted en-
tirely of physical matter operating under the constraint of these laws.
The twentieth century saw a number of developments conrming the universal applica-
tion of laws of nature, including the culmination of classical physics in Einsteins theory of
relativity, and the progression towards unication of physics, chemistry, and biology.
Although Newtons physics and Maxwells theory of electromagnetism were each
highly successful and internally consistent, there was some conict between them. Ac-
cording to Maxwells theory, but not Newtons, it was necessary, in order to achieve
consistency between measurements or calculations based on different frames of refer-
ence with relative motion, to make an adjustment in accordance with a mathematical
rule (the Lorentz transformation). Einsteins special theory of relativity provided for a
modication to Newtons mechanics so as to achieve consistency with Maxwells
theory, involving among other things recognition that space and time are not indepen-
dent of each other but can be regarded as existing in a combination called space-time,
in which every event has a unique location. Einsteins general theory of relativity of-
fered a deeper explanation of the gravitational force, and also accounted for certain
observations that could not be accounted for by Newtons theory. In these and other
ways, relativity theory appeared to support and advance the view that there are univer-
sal laws of nature constraining all systems.

Two giant steps towards the unication of physics, chemistry, and biology were Linus
Paulings explanation of the chemical bond in terms of quantum physics, and Crick
and Watsons discovery of the structure of DNA. The former showed how chemical
properties of matter could be explained by basic physics. The latter showed how life
itself could be explained by physics and chemistry, so that the notion that life required
some mysterious life-force, already discredited by Darwins theory, was dealt a death-
blow.
This 400-year history strongly supports the following three assertions, which I
adopt:
7.1. There are laws of nature that constrain what states of affairs can exist in the world
and how those states of affairs can change.
7.2. The laws that scientists have so far formulated approximate ever more closely and
with ever more assurance to those laws.
7.3. Science has accumulated extensive knowledge and understanding of how things
happen in the world, by reference to the laws that have been formulated.
However, to the extent that Newtonian physics and later developments might be
thought to suggest that earlier circumstances and the laws of nature uniquely deter-
mine later circumstances, this version of determinism is strongly contradicted by an
important part of twentieth-century science, namely quantum mechanics (QM). QM
strongly suggests that laws of nature do not uniquely determine how initial conditions
change over time, but generally leave open a spectrum of possible outcomes. It thereby
undermines an argument sometimes put1 that the physical world must be closed to
non-physical affectation, because otherwise there would have to be some kind of
mental force operating alongside the known physical forces. In fact the spectrum of
possibilities left open by QM are all consistent with the operation of known physical
forces, so that any selection between them would not require the application of any
force.
QM is also at odds with relativity theory, in that QM establishes that initial condi-
tions and laws of nature can have effects non-locally, that is, can have effects that
cannot be explained by inuences communicated at light speed or less. It thereby
renders highly dubious the deterministic block universe view suggested by relativity
theory, and also renders possible and plausible the existence of causal inuence by
non-local processes such as those that appear to be involved in conscious
experiences.
I will expand on both these matters in the following sections.
1
This unsound assertion is made for example in Horgan (2001).

QUANTU M M ECHAN ICS
QM is part of mainstream science, but it has implications that are subversive of many
views that are considered as scientic by many scientists and philosophers. I tried to
acquire some understanding of QM in the 1980s when I was writing my previous book,
The Mind Matters,2 and the more I understood QM the more I was amazed by its im-
plications. It still seems to me that these implications are not appreciated by most
philosophers, or indeed even by scientists in other elds, and that it is for this reason
that QM has had surprisingly little impact on the way most scientists and philosophers
think about the world.
I cannot hope to express adequately here the implications of QM, but I will try to
give some idea of why I believe they are signicant. The implications of particular
importance for this book can be identied as indeterminacy, non-locality, indetermin-
ism and observer-participation.
According to QM, the subatomic constituents of matter are of such a character that
they do not have determinate position and motion, so that there is an irreducible inde-
terminacy in their position and/or motion that is not just a reection of the impossibil-
ity of making accurate simultaneous measurements of position and motion. Rather, it
is a matter of objective fact that, say, an electron cannot be in a precise location unless
it is in a state such that its motion is wholly indeterminate (that is, it could be moving
in any direction, at any speed), and cannot have precise motion unless it is in a state
such that its position is wholly indeterminate (it could be anywhere); and that in gen-
eral terms, the more closely the state of the electron limits its location the less closely it
limits its motion.
This relationship is expressed by the Heisenberg uncertainty principle, according to
which the product of indeterminacies of position and momentum must always be at
least of the order of Plancks constant h, a very small quantity of a physical variable
called action. Because h is so small, the indeterminacies of position and motion of
objects that we can observe are undetectable; but the fact remains that relatively large
indeterminacies exist in the position and motion of the protons, neutrons and elec-
trons that make up their constituent atoms.
Even more extraordinary is the non-locality inherent in QM systems, which contra-
dicts the locality of causation (explained in the next paragraph) postulated by the
theory of relativity.
According to the theory of relativity, space and time are combined into a continuum
called space-time, and anything that happens at any location in space-time can neither
affect nor be affected by events with space-like separation from itthat is, events at
2
Hodgson (1991).

locations in space-time other than within its past light cone or future light cone. (The
past light cone of an event is all that region of space-time from which the location of
the event could be reached by travelling at light speed or less; while its future light cone
is all that region of space-time which could be reached from its location by travelling at
light speed or less.) Relativity theory asserts that neither information nor causal inu-
ences can be communicated or propagated at more than the speed of light.
A famous article3 published in 1935 by Einstein and two co-authors argued that
quantum mechanics must be incomplete, on the basis of the assumption, supported by
relativity theory, that the outcome of a measurement on one particle cannot be non-
locally correlated with the outcome of a space-like separated measurement on another
particle. A theorem formulated by John Bell and experiments conducted in the early
1980s by Alain Aspect4 have decisively refuted that assumption, and have shown that,
where particles of matter have interacted in a certain way, the outcome of space-like
separated measurements on both particles can be correlated in a way that has no expla-
nation in terms of causal inuences operating at light speed or less. (However, these
experiments did not show that any useable information could be communicated be-
tween space-like separated events, and to that extent relativitys embargo of faster-
than-light communication was not refuted: this has some relevance to the free will
theorem, and I say a little more about this point later when I discuss this theorem.)
Thus, when entities such as electrons have interacted, their physical states are corre-
lated and can remain so even when they have moved some distance apart. This correla-
tion can be such that measurement of the position of one entity will give certainty as to
the position of the other entity, and measurement of the momentum of one entity will
give certainty as to the momentum of the other entity. Since neither entity generally has
a denite position or momentum until this is measured, the only reasonable conclusion
is that the measurement of the position (or momentum) of one entity changes from
uncertain to certain the position (or momentum, as the case may be) of the other entity,
even at a time when that other entity has space-like separation from the measurement
event. This does not necessarily require a causal inuence to be transmitted from one
event to another with space-like separation from it: however, it does require effectual
non-local causal constraints as to what the combination of results can be.
A third feature of QM is indeterminism: whereas according to Newtonian physics,
precise specication of initial conditions and relevant forces determines precisely
what outcomes will occur, according to QM the most that can be determined by the
most precise possible specication of initial conditions and of relevant forces are
probabilities for various possible outcomes; and as with indeterminacy, this is consid-
3
Einstein et al. (1935).
4
Aspect et al. (1982); Hodgson (1991), 36369.

ered a fact about how things actually happen in the world, and not just a limitation
on what we can know. According to QM, what actually happens occurs randomly
within probability parameters, established by the laws of nature to which the rules of
QM correspond or approximate; and in relation to objects at the scale that we can
directly observe these probabilities generally give rise to virtual certainties conform-
ing to the rules of classical physics. There have been attempts to reconcile QM with
determinism;5 but the majority view is still that QM supports indeterminism.
These three features of QM are illustrated by the famous two-slit experiment. A beam
of electrons is red at a screen with two parallel slits, very close together, so that those
electrons that are not blocked by this screen fall on a further screen beyond it. When
electrons are detected at the second screen, they are detected as particles with quite
precise locations; but the pattern made by the impact of these electrons on the second
screen is a series of bands, explicable in terms of the electrons that passed through the
slits having behaved like interacting waves between the two screens. This interaction (or
interference) gives rise to the seeming paradox that in some areas of the second screen
electrons arrive less densely when both slits are open than when only one slit is open.
QMs indeterminism is illustrated by the fact that the location where any individ-
ual electron arrives at the second screen is not determined by initial conditions and
laws of nature; although the pattern produced by many electrons is determined, and
predictable, in accordance with the statistical laws of QM. QMs indeterminacy is
illustrated by the fact that each individual electron that arrives at the second screen,
and is detected there as a particle, has passed through both slits and has interacted
with itself between the two screens, as if it were a wave: thus, the location of each
electron as it passes the rst screen is indeterminate as between the two slits. QMs
non-locality is illustrated by the fact that the behaviour of each electron is affected
by events occurring at both slits: the nature of the electrons behaviour on passing
through one slit is affected by a space-like separated event, namely the event of its
passing through the other slit.
The fourth feature of QM mentioned earlier, observer-participation, is more elusive.
According to the mathematics of QM, there are two processes by which physical sys-
tems change: deterministic development over time that alters the probabilities of
outcomes of various observations or measurements (called process 2 by the mathema-
tician von Neumann) and indeterministic occurrences of possible outcomes when ob-
servations or measurements occur (called by von Neumann process 1, and also called
5
Such attempts are associated particularly with David Bohm; but in his last comprehensive state-
ment of his position, Bohm himself did not embrace determinism: Bohm and Hiley (1993), 32123. One
deterministic interpretation of QM is the many-worlds interpretation: in Hodgson (2002a) I explain
why I believe that interpretation is untenable. I will refer later to a more recent attempt, by Gerard t
Hooft, to reconcile QM with determinism.

the reduction of the quantum state or of the wave function). However, the mathemat-
ics of QM gives no indication of precisely what it is that amounts to an observation or
measurement, which converts what were previously probabilities or potentialities into
certainties or actualities; and the question of just what it is that does this, and how it
does it, is called the measurement problem of QM. There are various theories about
this, none of which has general acceptance; and it is a respectable albeit minority view
that measurement (and thus certainty and actuality) is achieved only when there is a
conscious grasping of an observation by a conscious observer. Whether or not that
view is correct, all theories (apart from the many-worlds interpretation of QM, re-
ferred to in note 5) admit of some degree of observer-participation, at least in the
choice of an experimenter as to what measurement to make. For example, indetermi-
nacies in the position or motion of electrons (say) can be reduced by measurement;
and it depends on the choice of an experimenter, as to what to measure, whether it is
indeterminacy of position that is reduced (thereby increasing indeterminacy of
motion) or vice versa.
The full signicance of the observer-participation of QM is unclear to me; and al-
though I believe observer-participation does have importance for consciousness and
free will, I do not explicitly appeal to it in my arguments in this book.
So QM shows, at least:
7.4. Science does not support the view that laws of nature uniquely determine how ini-
tial conditions change over time, but rather supports the view that they generally leave
open a spectrum of possible outcomes.
7.5. Locality of causal inuences, assumed by Einstein, has been decisively refuted.
TH E FR E E WI LL TH EOR E M
As mentioned earlier, despite the generally held view that QM shows that the uni-
verse is not deterministic, there have been attempts to reconcile QM with determin-
ism. A theorem recently propounded6 by Princeton mathematicians John Conway
(as mentioned earlier, the inventor of the famous Game of Life) and Simon Kochen
(one of the originators of the Kochen-Specker paradox of QM, to which I refer later),
which Conway and Kochen call the free will theorem, strongly supports the view that
there can be no such reconciliation.
The free will theorem supports a powerful challenge to the scientic credentials of
determinism, by showing, on certain well-supported assumptions, that two corner-
stones of contemporary science, namely (1) acceptance of the scientic method as a
6
Conway and Kochen (2006), (2009).

reliable way of nding out about the world, and (2) relativity theorys exclusion of
faster-than-light transmission of information, taken together, conict with determin-
ism in both its versions. Belief in determinism may thus come to be seen as notably
unscientic.
The theorem has been reported briey in New Scientist7 and has been the subject of
considerable discussion on the Internet, but otherwise has had remarkably little pub-
licity, despite what seems to me to be its considerable importance. It seems hardly to
have been noticed by philosophers. In this chapter, I will discuss the theorem in an
informal way, with the object of making its signicance understandable by people who
are not mathematicians.
Conway and Kochen make three assumptions, which they set out as axioms (and
which I will explain further shortly).
Having noted that there is an operation called measuring the squared spin of a spin
1 particle, which always produces the result 0 or 1, they state their rst axiom:
SPIN Axiom: Measurements of the squared (components of) spin of a spin 1

particle in three orthogonal directions always give the answers 1, 0, 1 in some
order.
They then discuss the Kochen-Specker paradox of QM, to which I will return. They
note the curious fact about QM (referred to earlier) that the results of remotely sepa-
rated observations can be correlated; and they note that in particular it is possible to
produce a pair of twinned spin 1 particles (by putting them into the singleton state of
total spin zero) that will give the same answers to the above squared spin measure-
ments in parallel directions. And they then state their second axiom:
TWIN Axiom: For twinned spin 1 particles, suppose experimenter A performs

a triple experiment of measuring the squared spin component of particle a in
three orthogonal directions x, y, z, while experimenter B measures the twinned
particle b in one direction, w. Then if w happens to be in the same direction as
one of x, y, z, experimenter Bs measurement will necessarily yield the same
answer as the corresponding measurement by A.
They then refer to the assumption that experimenters are free to choose between pos-
sible experiments (explaining this as meaning that their choice is not a function of the
past); and they go on to state their third axiom:
7
Merali (2006).

MIN Axiom: Assume that the experiments performed by A and B are space-like
separated. Then experimenter B can freely choose any one of the 33 particular
directions w, and as response is independent of this choice. Similarly and
independently, A can freely choose any one of the [corresponding] triples x, y, z,
and bs response is independent of that choice.
They then state their theorem:
The Free Will Theorem. The axioms SPIN, TWIN and MIN imply that the
response of a spin 1 particle to a triple experiment is freethat is to say, is not a
function of properties of that part of the universe that is earlier than this
response with respect to any given inertial frame.
Before giving my informal explanation of the theorem, I will say something about the
axioms.
The rst two axioms are well-supported conclusions of QM. Both of them follow
from the mathematics of QM, and have experimental support in that they have been
extensively tested and never falsied.
Spin is a property of particles of matter dealt with by QM, of the same nature as
polarisation of light; and according to QM, some particles of matter are such that
measurements of the squares of their (components of) spin in three orthogonal di-
rections (directions at right angles to each other) are always 1 or 0, and always add up
to 2. That is, they always give the results 1, 0 and 1 in some order, as stated by the SPIN
axiom.
The TWIN axiom deals with properties of pairs of such particles that have been cor-
related in a particular way by interaction between them and then have moved far apart
in such a way as to preserve the correlation. In such a case the mathematics of QM
indicates, and experiments have conrmed, that when experimenters measure the spin
of these particles the results are correlated in the way stated by the TWIN axiom, even
if the experiments have space-like separationthat is, even if the experiments are per-
formed at times and places such that no signal travelling at light speed or less could
pass between them in either direction.
I interpret the assumption in the MIN axiom that, because the experiments per-
formed by A and B are space-like separated, particle as response is independent of Bs
choice, and particle bs response is independent of As choice, as meaning that the infor-
mation as to what choice B makes and the information as to what choice A makes is not
available to particle a and particle b respectively before their response; and so interpreted,
this assumption is a consequence of relativity theory that, as noted above, has not been
refuted by Bells Theorem and the Aspect experiments.

EXPLANATION OF TH E TH EOR E M
Conway and Kochen commence their proof of the theorem by relying on a version of
the Kochen-Specker paradox, which shows on the basis of the SPIN axiom that the
result of measurement of spin in as little as 33 possible directions cannot exist prior to
and independently of measurement. This can be illustrated by gure 7.1, a gure cre-
ated by Jasvir Nagra, which he explains as follows:8
Imagine a cube that snuggly surrounds a sphere. On each face of the cube, we
inscribe a circle and inside each circle we draw a square that touches the circle
at the squares four corners. We divide each such square into four smaller squares
and mark the following points on the cube.
We get 33 points (9 points per face 3 faces + 1 point per edge 6 edges) on the
cube in this way. These represent 33 directions of measuring a particle.
By working through the possibilities, it can be shown that (1) it is impossible to

assign the value 1 or the value 0 to the squared spin in each of these 33 directions
without violating the requirement that measurements of the squares of components
of spin of spin 1 particles in any three orthogonal directions always gives the results 1,
0 and 1 in some order; and thus (2) that it cannot be the case that the state of a particle
and/or the universe is such that every one of these results is xed in advance of and
independently of measurement.
Figure 7.1 Illustration of the Kochen/Specker paradox
8
Quoted from Nagra (2005).

This of itself does not refute determinism in the behaviour of the particles. In par-
ticular, it could still be the case that, in relation to any combination of directions that
can be measured, the outcome is xed in advance for that combination of directions, and
will certainly occur if information as to what other directions are being measured is
available at relevant measurement sites.
What the theorem shows is that in the case of twinned particles, as postulated by the
TWIN axiom, the MIN axiom means that determinism cannot be saved in this way.
A direction that would have to give the result 1 if one combination of other directions is
measured, and 0 if another combination of other directions is measured, could be the
single direction w measured by the second experimenter B; and the MIN axiom means
that the response of b (the twinned particle measured by B) does not have information
as to As choice of which other directions to measure. Thus it must be open that the
result of measurement by B could be either 1 or 0; but which one it is could not be de-
termined in part by As choice of which directions to measure, as would be required if
determinism were to be saved in the way suggested. That is, given the rst and third
axioms, and if determinism obtains, then Bs measurement of direction w could give the
result 1, when the second axiom would require the result 0 for this measurement.
The bottom line is that, if determinism and the SPIN and MIN axioms are all main-
tained, there would have to be directions such that, if measurements were made for
those directions by experimenter A and experimenter B, the TWIN axiom would be
contradicted. The only way that both determinism and all three axioms can be main-
tained together would be to postulate that the experimenters are somehow prevented
from measuring for these directions: that is, as Conway and Kochen put it, that the
experimenters do not have free will to measure for these directions.
This is my non-mathematical explanation of the theorem; and it also suggests some
tension between the rst two axioms and the MIN axiom. It might seem that the same
problem arises, even if determinism is rejected. If the outcomes of all possible mea-
surements are not xed in advance, beyond being subject to the constraints of the
SPIN and TWIN axioms, then it might seem that faster-than-light communication of
information between the locations of experimenter A and experimenter B is required
to achieve the correlations required by the TWIN axiom, thereby contradicting the
MIN axiom.
The resolution of this tension is to be found in the non-locality of unmeasured
quantum systems, in which there are correlations between spatially separated parts of
the systems.
A simple illustration of this is given by a single particle system, for which there may
(according to the mathematics of QM) be a 0.5 probability that the particle will be
found at location X and a 0.5 probability that it will be found at a distant location Y.
Then, if it is found at X, according to the mathematics of QM the probability of it

being found at Y is zero, not because of any passage of information from X to Y but
simply because the total probability cannot be other than 1. If in such a case there is a
measurement at Y, which has space-like separation from the measurement at X, the
particle will not be found at Y. According to some frames of reference, it will be the
measurement at X that occurs rst and establishes that the particle is at X and is there-
fore not at Y; and according to other frames of reference, it will be the measurement at
Y that occurs rst and establishes that the particle is not at Y and is therefore at X. Ac-
cording to relativity theory, both points of view have equal validity, and there is no fact
of the matter whether it was the measurement at X or the measurement at Y that was
truly causative. There is nothing in QM, or in the SPIN or TWIN axioms, that contra-
dicts this.
In the same way, the mathematics of QM gives interdependent probabilities for the
results of spin measurements by experimenter A and experimenter B, and the out-
come of relevant measurements alters these probabilities in a way that has to be cor-
related as required by the TWIN axiom. The process is not a causal one requiring the
conveying of information from the location of experimenter A to the location of ex-
perimenter B or vice versa, but rather a logical one depending on the necessity that
probabilities of alternatives always add up to 1. In that way there can be a reconcilia-
tion of the TWIN axiom and the MIN axiom; but this reconciliation is not available
if the outcome of any of the relevant measurements is xed in advance, rather than
being a matter of probability only.
I M PLICATION S
Although the authors called their theorem the free will theorem, it does not directly
support any conclusion about free will. Its conclusion (that the outcomes of the mea-
surements are not xed in advance) follows if it is the case both that the experimenters
can measure spin in any direction (at least in the sense the choice of direction is not a
function of the past in any way that correlates the choice to the particles or the other
experimenter), and that (in accordance with the axioms) laws of nature correctly spec-
ify features of the outcome in the event that any such measurement is made.
In order to avoid the conclusion and thus to save determinism, without rejecting the
well-established assertions of the SPIN and TWIN axioms, it would be necessary to
claim that it is not the case that the experimenters can measure spin in any direction in
the sense set out above.
However, the causal antecedents of whatever it is that determines which measure-
ments are made by each experimenter may be effectively independent of each other
and of the causal antecedents of the particles being measured, at least unless there is a
thoroughgoing conspiracy of nature at work. For example, the directions of the

measurements to be made by experimenter A and experimenter B could be deter-
mined by each of them rolling a number of dice, at a prearranged time when they had
space-like separation, immediately before they make their measurements.
Accordingly, the required limitation on what measurements can be made would re-
quire not merely determinism, but a thoroughgoing conspiracy of nature that corre-
lated the outcomes of these space-like separated dice-rollings. This in turn would
undermine the scientic method, because it would mean that scientists cannot have
access to random samples but rather are sometimes prevented by a conspiracy of
nature from making measurements that if made would refute a hypothesis being
tested. Indeed, it would mean that the consistent experimental conrmation of the
rules of QM expressed in the rst two axioms does not happen because those axioms
truly express laws of nature, but because a conspiracy of nature prevents experimental
disconrmation of them.
So I say:
7.6. The Conway/Kochen free will theorem conrms that science does not support
determinism.
This conclusion follows in a straightforward way in relation to the rst version of de-
terminism I identied at the outset, namely that earlier circumstances and the laws of
nature uniquely determine later circumstances. The theorem shows that, on the basis
of the axioms, if it is the case that experimenters can measure spin in any direction,
then prior to measurement being made, existing circumstances and the laws of nature
can do no more than determine correlated probabilities for various outcomes. Accord-
ingly a deterministic version of QM would require either that one or more of the
axioms must be rejected or else that experimenters are somehow prevented from
making those measurements that would contradict the laws of quantum mechanics;
and that in turn would require the conspiracy of nature referred to earlier.
In relation to the other version of determinism, the block universe, the position may
seem less clear. It could be said that the block universe theory does not involve any
particular view as to how earlier states are causally linked to later states, and that the
circumstance that outcomes would have been different if earlier circumstances had
been different, that is, if experimenters had made different measurements, is entirely
unremarkable. But the block universe view still relies on the scientic method to estab-
lish laws of nature linking events in different parts of the universe; and given that the
axioms are taken as holding good in the block universe, there would be no explanation,
which did not involve a conspiracy of nature, as to why measurements never falsify the
rst two axioms. Rather, the position would be that the laws of QM reected in the rst
two axioms do not in fact hold good in the block universe, but they are never

experimentally refuted, because the block universe is such that experiments that would
refute them cannot be made.
Recognition that the Conway-Kochen theorem makes it unscientic to accept deter-
minism would not directly support free will. Quantum mechanics treats the indeter-
minism involved in the results of such measurements as being random within
probability parameters specied by the mathematics; and as has often been pointed
out, randomness is inimical to free will rather than supportive of it.
However, as I have argued in previous chapters, there are other reasons for believing
in free will associated with rational conscious decisions and actions; and refutation of
determinism supports an argument that the physical world is not closed to inuences
from rational conscious processes, and thus that there is room for free will. In the next
chapter, I will suggest a plausible account of how this works in the brain.
Finally on the free will theorem, I note that a response to the Conway-Kochen theo-
rem has been published9 by Nobel laureate Gerard t Hooft, who has for some time
been developing a deterministic version of QM.10 He accepts that the model of the
world provided by a physical theory must give credible scenarios for a universe for any
choice of initial conditions. He claims that this requirement can be satised by a deter-
ministic theory, because a deterministic theory can still give freedom to choose the
initial state, regardless its past, to check what would happen in the future; and that,
contrary to what Conway and Kochen suggest, one cannot modify the present without
assuming some modication of the past.
I do not think this response is adequate, in that it fails to recognize that the causal
antecedents of whatever determines which measurements are made, and those of the
particles being measured, may be effectively independent of each other. So t Hoofts
argument does not avoid the necessity for a highly articial and improbable conspiracy
of nature, which would undermine the scientic method.
TI M E AN D TH E B LO CK U N IVE R S E
I have suggested that the free will theorem counts against the determinism of the block
universe view as well as against rule-based ideas of determinism. However, relativity
theory does support the idea of four-dimensional space-time and does support plau-
sible arguments that time does not pass; so I should say something more about this.
Relativity theory states, and experiments have conrmed, that the measurements of
the passage of time (and indeed also mass and distance) are not independent of motion.
Indeed, it seems clear that acceleration of an object to a high speed will cause time to
9
t Hooft (2007).
10
t Hooft (2001), (2007).

pass more slowly for that object. Thus, for example, I believe there is strong theoretical
and experimental support for the hypothesis that if one twin were to travel into space,
being accelerated to a high speed relative to the Earth, and later return, that twin would
on return have aged less than a twin who remained on the Earth. If the speed achieved
by the spacecraft were close to the speed of light, and if the journey in space lasted
some years as measured by the travelling twin, this difference would be substantial. So
I do not doubt:
7.7. The rate at which time passes for an object is affected by acceleration of the object.
Relativity theory also supports the view that there is no inertial (non-accelerating)
frame of reference that has any priority or greater validity than any other, such that the
time order of events with space-like separation according to that frame of reference has
a better claim to truth or validity than their time order according to other frames of
reference. Events within the past light cone of an event are unambiguously in its past,
and events within the future light cone of an event are unambiguously in its future; but
events with space-like separation from an event may be in its past or in its future or
simultaneous with it, depending on the motion of the frame of reference from which
the assessment is made.
There is tension between relativitys denial of simultaneity of space-like separated
events, except in relation to specied frames of references, and the non-locality of
quantum systems discussed above. And plausible suggestions can be made as to a
preferred universe-wide frame of reference, for example that which coincided with
the motion of the centre of mass of the universe immediately following the Big
Bang and has not been accelerated since. However, even if one accepts that there is
no preferred frame of reference, none of this contradicts the idea that time really
passes, or that sense can be made of the present, as well as of the past and the future;
and an account can be given, consistently both with relativity theory and with the
indeterminism suggested by QM and conrmed by the free will theorem, denying
the existence of an unchanging block universe and according reality to the passage
of time.
On this account, the universe is real but developing and changing; so any question
about the state of the universe or its features is imprecise unless it has reference to a
particular stage or particular stages of that development. If that reference point is an
event that occurs at a particular time at a particular place, then:
1. as regards the past light cone of the event, denite answers can be given as to what
the state of the universe and its features were, because in those regions the state of
the universe and its features are xed;

2. as regards the future light cone of the event, the most that can be said is that the
state of the universe and its features will be within a spectrum of possibilities deter-
mined by prior circumstances and laws of nature;
3. as regards other regions of space and time, that is, those with space-like separation
from the event, any answers may need to be given on the basis of a specied frame
of reference, because otherwise the answers may be indeterminate in that they vary
according to the motion of the frame of reference.
I suggest it is not helpful to ask whether, with reference to such reference point, any
state of the universe or any of its features is real. The changing universe is real, on any
reasonable understanding of the word; but once one has given answers to questions
about the state of the universe and its features in accordance with the above three sug-
gestions, it does not add anything meaningful to say whether or not they are real: the
word real just does not have a sufciently precise meaning. To the extent that this may
seem unsatisfactory in relation to the third category of cases, it should be remembered
that the extent of indeterminacy in that category in realistic situations is minimal: the
time it takes for a signal at light speed to travel from one place on Earth to any other
place on Earth is less than one twentieth of a second, so any indeterminacy as to the
state of the universe or its features in any place on Earth with reference to any event on
Earth will not extend beyond one-tenth of a second.
The above account can also give content to the word now that is consistent with
relativity theorys denial of preferred frames of reference. On that basis, now can have
a precise meaning only with reference to an event at a particular time at a particular
place; and would have some indeterminacy in relation to regions of space and time
with space-like separation from it. The word now in fact has most application with
reference to conscious experiences of human beings; and answers to questions con-
cerning what is the state of the universe now and what features does the universe have
now will be in accordance with the above analysis.
It will be recalled that QM supports not only indeterminism, but also non-locality,
and it asserts that some events are non-local in that they occur to, and involve instan-
taneous change to, systems that are spread out in space. In order that the above account
be consistent also with this non-locality, some modication is required. I would sug-
gest the following provisional modication.
In relation to such a non-local event, there would be past and future light cones
referable to each location in time and space where an instantaneous change takes
place in the non-local system; and the above analysis would suggest that, with refer-
ence to such a non-local event, (1) what is xed are those regions of space and time
within all of these past light cones, (2) what is merely within a spectrum of possibili-
ties determined by prior circumstances and laws of nature are those regions of space

and time within all of these future light cones, and (3) what may be indeterminate
except on the basis of a specied frame of reference are all other regions of space and
time.
I will be suggesting that conscious experiences are non-local events of this kind, be-
cause they correlate with changes occurring instantaneously in different regions of the
brain; so that if relativity theory is retained to the maximum extent consistently with
QM, this modied analysis can be applied to conscious experiences.
Finally on time, there is another fundamental reason why I do not subscribe to the
view that the passage of time is an illusion, which has to do with my basic beliefs in
chapter 1, and my basic beliefs about consciousness. In a well-known account of the
block universe view, Hermann Weyl wrote:11
The objective world simply is, it does not happen. Only to the gaze of my con-
sciousness, crawling upward along the lifeline of my body, does a section of the
world come to life as a eeting image in space which continuously changes in
time.
This passage points up a dilemma for the block universe view.

It will be recalled that one of my two most certain beliefs is that some conscious expe-
riences change; and something like the Weyl position is required to accommodate this
to the block universe view: if the past and future exist tenselessly along with the present
in space-time, then the appearance of change can be given only to something moving
through space-time. Yet on the block universe view, nothing moves through space-time:
there are just different stages of things, including brains and the gaze of consciousness,
existing at different locations in space-time.
Any reasonable view of the mind and consciousness requires a close association be-
tween the gaze of consciousness and brain events: the block universe view gives brain
events a tenseless existence in space-time, so there appears to be no explanation of why
or how the gaze of consciousness could move through space-time. It does not help to
suggest that the apparent changes of experiences may be the result of a rapid succes-
sion of static experiences, like frames of a movie: this still leaves unanswered the ques-
tion why the frames are experienced in time sequence, giving rise to the appearance of
change over time.
And if there is a gaze of consciousness that moves through space-time from earlier to
later, as Weyl suggests, then the gaze (together with its physical correlates) must leave, and
thus cease to exist at, earlier times as it moves to later times, and must arrive at, and thus
come to exist, at those later times; and this is inconsistent with the block universe view.
11
Weyl (1949), 116.

Even if the changing experiences were considered to be mere illusion, as some physi-
cists suggest, there would still be physical correlates of the illusion, which would have
to change as the illusion changes. That is, there must be something about the physical
correlates that corresponds to the illusion ceasing to be as it is at one time and coming
to be as it is at a succeeding time. And if the correlates are merely unchanging events at
different locations in space-time, then surely the correlates could not correspond to
the illusion ceasing to be as it is at one time and coming to be as it is at a succeeding
time, at least unless there is a moving gaze of consciousness; and this would give rise
to the problem referred to in the previous paragraph.
So I believe:
7.8. Our experiences of the passage of time cannot be dismissed as illusions, and science
has provided no coherent account of how they could be mere illusions.
COR E ASS E RTION S AB OUT SCI E NCE

AN D DETE R M I N I S M
7.1. There are laws of nature that constrain what states of affairs can exist in the world
and how those states of affairs can change.
7.2. The laws that scientists have so far formulated approximate ever more closely and
with ever more assurance to those laws.
7.3. Science has accumulated extensive knowledge and understanding of how things
happen in the world, by reference to the laws that have been formulated.
7.4. Science does not support the view that laws of nature uniquely determine how ini-
tial conditions change over time, but rather supports the view that they generally leave
open a spectrum of possible outcomes.
7.5. Locality of causal inuences, assumed by Einstein, has been decisively refuted.
7.6. The Conway/Kochen free will theorem conrms that science does not support
determinism.
7.7. The rate at which time passes for an object is affected by acceleration of the object.
7.8. Our experiences of the passage of time cannot be dismissed as illusions, and science
has provided no coherent account of how they could be mere illusions.

8
Neuroscience and Conscious Choice
SCI E NCE AN D TH E B RAI N
By the end of the twentieth century, science had gone a long way towards carrying
into fruition the project of discovering laws of nature and explaining the world by
reference to these laws, in accordance with Descartess postulate that everything in the
world, apart from the minds or souls of human beings, was constrained by laws of
nature.
However, at the same time science had also challenged Descartess exemption of
human minds or souls from such constraints. Descartes had contended that immate-
rial human minds were unconstrained by laws of nature in making choices and initiat-
ing actions, and that they gave effect to these choices and actions by inuencing the
physical brain through the pineal gland. However, that view became increasingly un-
tenable, as developments in nineteenth- and twentieth-century science showed that
the making of choices and initiating of actions was associated with and (at least) much
affected by highly complex brain activity, which was subject to the constraints of laws
of nature.
Prior to Darwin, the exemption of human minds from the constraints of laws of
nature might have been supported by an argument based on the argument from design,
advanced to prove the existence of God as Creator. Just as a watch suggests a watch-
maker, the argument went, so the plethora of complex and beautifully adapted life-
forms suggests a Creator. We could not believe that a watch had come into being by
chance, simply by the operation of impersonal laws of nature; and likewise we could
not believe that a mouse (say) had come into being by chance. This argument could be
developed to support the exemption of human minds from the constraints of laws of
nature, as follows. The creation of animals thus apparently involved a kind of causa-
tion in the world, through the activity of a Creator that was not constrained by imper-
sonal laws of nature. The Creator, making human beings in its own image, could have
endowed human minds with some capacity for the same kind of causation; so that, just
as the activity of the Creator was not constrained by laws of nature, the same could be
true also of the activity of human minds.
133
That kind of argument was undermined by Darwins theory. It showed how, starting
from initial conditions which could be thought simple enough not to require a de-
signer or Creator, the operation of laws of nature could have given rise to life as we
know it, in all its complexity and variety and beautiful adaptation. It also offered a
plausible explanation of our behavioural characteristics, on the basis solely of causa-
tion conforming to the constraints of laws of nature; that is, a plausible explanation of
how the appearance of purpose and choice could have arisen in systems which in fact
change over time in accordance with laws of nature. Darwin thus gave us reason to
question our feeling of freedom and ability to choose, and reduced the weight of this
feeling as an argument against the universality of causation through the constraints of
laws of nature.
The idea that human minds are exempt from the constraints of laws of nature was
thoroughly undermined in the twentieth century, in particular by the recognition of
the importance of unconscious processes in our behaviour, continuing progress in
neuroscience, and the emergence of the cognitive sciences.
Around the beginning of the twentieth century, the work of Sigmund Freud high-
lighted the fact that much of our motivation is unconscious; and since then, it has
been impossible to maintain a view of human agency as a matter of wholly conscious
decisions based on wholly conscious motives. Whatever one thinks of Freuds partic-
ular theories, he did thereby change the understanding of the human mind. And the
nature and extent of unconscious motivation has been further explored throughout
this century, showing not merely that much of our motivation is unconscious, but
also that we have a tendency to rationalise our conduct by (unconsciously) fabricat-
ing, and then believing, plausible but untrue stories to explain why we did what
we did.
Over recent years huge advances have been made in neuroscience and in the cogni-
tive sciences. Much is now known of the details of physical processes of the brain, and
how those physical processes correlate with different types of experiences and actions.
Together with physical investigation of the brain, the cognitive sciences have thrown
light on its functional organization, giving insights into how the physical events of the
network of neurons can give rise to sensation, perception, memory, problem solving,
emotions, and actions.
As mentioned earlier, these developments have made it impossible for me to believe
in a dualism of substances, involving a distinction between (on the one hand) immate-
rial minds or souls that have conscious experiences, make choices and initiate actions,
and (on the other hand) material things, including brains, that dont. That kind of
dualist theory raises the question, exactly what are the respective contributions of the
brain and the mind to choices and actions? If one postulates that the minds willing of
intentional actions is given effect to by the physical brain, this raises the question

whether, and if so to what extent, the physical brain is involved in forming the inten-
tions in the rst place. And if the mind can form intentions without assistance from
the physical brain, how can it do so? What is the minds (non-physical) structure which
enables it to form intentions, and why does it need to have such a complicated brain to
do no more than put intentions into effect? On the other hand, if one were to say that
physical brain processes are involved in forming intentions, then it would become nec-
essary to identify the respective roles in this of the brain and the mind: if the mind
cannot form intentions on its own, exactly what can it do, what is its distinctive contri-
bution? And as before, if the mind does make this contribution, what structure does it
have to enable it to do so?
For all those reasons, I believe the following:
8.1. Our brains are constituted by the same physical elements as other physical objects,
and are constrained by laws of nature as are other physical objects.
8.2. Our choices and actions are to a considerable extent the product of activity of our
physical brains constrained by laws of nature, and much of our motivation is unconscious
or barely conscious.
So science has produced a comprehensive framework for explaining the behaviour of

all physical systems, including human brains, in terms of initial conditions changing as
constrained by laws of nature. There is a history of formulation of laws that seem to
approximate ever more closely to actual laws of nature, and that provide ever more
comprehensive explanations about how things happen in the world.
And many people believe that the laws of nature uniquely determine how any initial
conditions in the world change over time.
However, as will have become apparent, I do not believe this. As argued in chapters 3
to 6, there are powerful reasons to hold that consciousness and plausible reasoning
cannot be fully explained in terms of evolution-selected computation-like information
processing, constrained by the physical states of our brains and laws of nature; and that
conscious human choices and actions make a contribution to what happens that is not
uniquely determined by initial conditions and laws of nature.
Further, I contend it follows from the previous chapter that, although science has
made it virtually certain that there are laws of nature constraining what states of affairs
can exist and how those states of affairs change, science has not made it reasonable to
believe that such laws uniquely determine how initial conditions change over time:
indeed, science makes it more reasonable to believe that such laws generally leave open
spectra of possible changes that could occur.
Also, science makes it reasonable to accept that there can be non-local systems, that
is, systems extended in space in which changes can occur instantaneously to spatially
Neuroscience and Conscious Choice 135

separated elements of the system; and this opens up the possibility that conscious ex-
periences, corresponding to physical processes in separated regions of the brain, mani-
fest a non-local system affected by and affecting events in one of those regions.
So I contend:
8.3. The physical sciences can accommodate the existence of unied conscious experi-
ences, corresponding with events in widely separated regions of the brain, having effects,
not determined by rules, in one or more of those regions.
A GE N E RAL PICTU R E
Ive outlined what I contend is the role of gestalts in our conscious decision-making;
and this provokes the question of how our brains could work so as to make this role
possible. I will commence my answer with a general picture of how I say conscious
processes and physical processes could work together.
I accept that laws of nature constrain the processes of our brains, as they constrain
other physical processes. I accept also that our brain processes carry out information
processing that unfolds in accordance with computational rules selected by evolutionary
trial and error. However, what I say is that, in those cases where there is a contribution
from conscious decision-making, these laws and rules do not determine a unique result.
As shown in the previous chapter, currently accepted physics makes it most unlikely
that physical laws of nature determine unique outcomes; and I contend that, once one
puts aside prejudice against causation that is not rule-determined, there is no strong
reason to think that physical laws do not, in cases of conscious decision-making, leave
scope for effective contributions from responses to experiences of gestalts. And as re-
gards computational rules, they too can allow for outcomes in which there is a spec-
trum of probability-weighted alternatives, for example by means of what is called fuzzy
logic and/or the application of rules of probability, including Bayes Theorem.
My contention is:
8.4. If it were possible to abstract physical brain processes and/or computational infor-
mation-processing from the totality of conscious decision-making, they would give the
appearance of producing probability-weighted spectra of outcomes, which include the
actual outcomes of the conscious decision-making.
However, I nd implausible the idea that there are just occasional discrete selections
consciously made between discrete alternatives thrown up by unconscious processes:
this too strongly suggests a conscious subject, distinct from the physical brain, which
does something magical. I prefer the idea of a conscious subject, not distinct from

the physical brain but rather an aspect of the physical-and-mental system that is the
brain-and-mind, which uses both conscious experiences and rule-determined pro-
cesses together, on a continuous basis, to steer a course within spectra of possibili-
ties left open by laws of nature and computational rules. For the most part, I suggest,
this course would differ little if at all from that which, on the basis of the physical
processes together with laws of nature and computational rules, considered in isola-
tion, would be the most probable course. However, on occasions there could be sig-
nicant deviations; and of course, and most importantly, the cumulative effect of
successive minor deviations could be very substantial indeed.
Indeed, the cumulative effect of probable decisions can be improbable. For example,
if four independent events are required for a particular result, and each event has 0.7
probability, the probability of the result is about 0.24. So if chance can be overridden
by conscious choice, four choices of events each having 0.7 probability can bring about
a result that initially had only 0.24 probability.
In our everyday living, we are often pretty much on autopilot. Whenever we are
acting without paying particular attention, and without concentration, deliberation or
effort, conscious input does not go beyond marginal shaping or ne-tuning of actions,
coupled with readiness to do more if something arises that calls for attention, concen-
tration, deliberation and/or effort. Our conscious motivation, so far as it is operating,
runs along the same lines as our unconscious motivation. However, the cursory atten-
tion associated with this kind of activity can rapidly (and automatically) become
heightened when something signicant happens, and this can in turn lead to concen-
tration, deliberation and/or effort.
In circumstances of heightened attention, concentration, deliberation and/or effort,
our response to experienced gestalts can have a substantial impact in directing action
within the spectrum of possibilities. I earlier gave the example of my driving a car
thinking about other things when something untoward happens, in which case my
conscious attention is quickly engaged, so that my automatic driving reactions are
supplemented by my conscious grasp of the whole situation and there can be signi-
cant conscious ne-tuning of my response.
On the other hand, when we are concentrating on a task, the conscious input does
seem continuous and substantial. For example, when I am writing something, ideas are
thrown up by unconscious processes; but I am continually appraising the sense and
sound of chunks of what I am writing so as to decide whether to keep them or to alter
them or to try to come up with other ideas. My actions seem to ow from an ongoing
process with complementary contributions from unconscious processes and conscious
experiences.
I accept that sometimes we do make instantaneous snap decisions, in which there
appears to be a contribution from consciousness; and I accept that for these decisions,

the picture of continuous steering seems less appropriate. However, even for these
decisions I suggest there may be complementary contributions from unconscious and
conscious processes; although I think that in these cases the element of conscious judg-
ment must be less than where there is a more extended conscious addressing of a prob-
lem, and in some cases it may be wholly absent.
So:
8.5. Rather than conscious subjects making discrete conscious choices between alterna-
tives thrown up by unconscious processes, they use both conscious experiences and rule-
determined processes together, on a continuous basis, to steer a course within spectra of
possibilities left open by laws of nature and computational rules.
TH E CARTE S IAN TH EATR E
One argument against the efcacy of conscious experiences is that advanced by Daniel
Dennett1 to the effect that it would require a central headquarters in the brain, a so-
called Cartesian theatre, where conscious experiences occur and can have an effect on
brain processes; and there is no such location.
Dennetts arguments against the existence of any such central headquarters are to my
mind convincing, and I agree there is no such thing. For one thing, it seems clear that
the content of conscious experiences has contributions from processes occurring in
different regions of the brain: for example, it seems clear that shapes and colours are
processed in different parts of the brain, and it does not seem that there are further
brain processes in any more localised region of the brain that bring the results of these
separate processes together. Indeed, the content of conscious experiences often has
such a richness and complexity as to make it unlikely in the extreme that it could be
supported by anything less than extensive and complex processes, which could not
plausibly be conned to any one small region of the brain. I accept that consciousness
is distributed through the brain, as Dennett contends.
However, I suggest that this supports the efcacy of conscious experiences, rather
than the reverse.
I have previously referred to the binding problem of consciousness, the problem of
how it is that features processed in different parts of the brain are brought together
into a unied experience in which they are experienced all-at-once by the conscious
subject. The expression all-at-once is of prime importance here, because in many
cases there is not the slightest element of successiveness in the experience of these fea-
1
Dennett (1991), 10139.

tures: it is assuredly not the case that we experience rst shapes and then colours, or
vice versa, rather than experience all-at-once all the colours and all the shapes of the
feature-rich gestalts of ordinary everyday vision.
On the other hand, it is true that we do experience the notes of a melody in succes-
sion but nevertheless can have an all-at-once grasp of the whole melody. In this and
other ways, our experiences sometimes seem to span a period of time, to have a kind
of time-thickness that has been called the specious present or the psychological present.
However, despite the element of successiveness in experiences such as these, there is
still an all-at-once grasp of many of their features; so that these experiences illustrate
rather than avoid the problem of explaining how features that must be supported by
process in different regions of the brain are somehow present together all-at-once in
conscious experiences.
When one asks what could it be about the way the world works that supports this
phenomenon, an obvious and I suggest highly plausible possibility is the non-locality
of some quantum systems.
If the operations of the brain were truly explicable in terms of classical and relativis-
tic physics, then there could be no causation operating between space-like separated
events in the brain, and no explanation of how conscious experiences could include,
all-at-once, experiential features supported by spatially separated processes. It would
be no explanation to say that information from different regions of the brain can pro-
gressively be collected to be registered in consciousness, because the collection and
registration in consciousness would need itself to be supported by brain processes; and
there is, as noted earlier, no particular location in the brain, no Cartesian theatre, where
such collection and registration takes place.
It is QM that allows for non-local systems, in which global non-local changes can
take place as a result of events occurring at one location within the spatially extended
system; and there is no reason in QM why global non-local changes (in conscious ex-
periences) could not correlate with or result from space-like separated processes oc-
curring in different regions of the brain. The fact that there is no localised Cartesian
theatre, together with the binding problem of consciousness, thus strongly suggest that
classical and relativistic physics are not enough to explain the workings of the brain,
and in particular are not enough to explain consciousness.
And if it is the case conscious experiences are non-local systems of the kind permit-
ted by QM, this opens up the strong possibility that they may in turn have effects based
on their non-local properties.
Dennett himself can dismiss the binding problem of consciousness because he not
only denies causal efcacy to conscious experiences as such, but removes them from
the real world of brains and neurons. He treats conscious experiences as a kind of c-
tion, asserting, for example, that qualia such as pain are characters in good standing in

the ctional world of our heterophenomenology (his word), but only complexes of
dispositions in the real world in our brains.2 This of course conicts with my most
basic belief that conscious experiences occur; and I suggest this view would hardly
convince or comfort a person in excruciating pain. Such a person would hold that his
or her pain is as real as anything in our world can be, and I would entirely agree. So
contrary to Dennett, I see the binding problem as a genuine problem, with quantum
non-locality as probably being at least part of the solution.
My conclusion:
8.6. There is no localised Cartesian theatre where conscious experiences occur, and the
solution to the binding problem must probably be found in the non-local nature of quan-
tum systems.
TH E SCALE AN D NATU R E OF QUANTU M E FFECTS
On the other hand, the great success of neuroscience in explaining much of the opera-
tions of the brain in terms of physical law-governed processes, and the great success of
the cognitive sciences in explaining much of the functioning of the brain in terms of
computational information-processing, may be considered as counting strongly
against any non-rule-determined input from conscious experiences.
I referred earlier to huge advances that have been made in neuroscience and in the
cognitive sciences.
At the level of the individual neurons of the brain, a great deal is now known
about what causes them to re or not re; how electrical signals are transmitted
within neurons and then passed across the synapses to other neurons by means of
chemical transmitters; and how all this is affected by the chemistry of the brain.
Much is also known about the patterns of connections between the neurons of the
brain. Areas of the brain associated with particular functions have been identi ed:
neuroscientists know broadly what effects on brain function and behaviour will
follow from injury to specic areas, and with the aid of brain-scanning techniques
they can observe what regions of the brain are active when particular tasks are being
undertaken.
Along with physical investigation of the brains operation, the cognitive sciences have
thrown light on its functional organization, giving insights into how the physical events
of the network of neurons can give rise to sensation, perception, memory, problem
solving, emotions, and actions. The brain is considered as an information-processing
2
Dennett (1991), Part III, especially 389: cf. Hodgson (1996a).

system, which takes the information from sensory inputs, processes it in various ways,
and gives out various kinds of outputssometimes resulting in physical actions, such
as walking or speaking. There has been much theoretical and practical work on the
idea that human cognitive performance (sensation, perception, reasoning, execution
of decisions, etc.) can be reproduced or simulated by computers.
So long as no explanation of consciousness or its role is required, this scientic work
on the brain does not point to any need, or indeed any room, for input from conscious
experiences; and what is known about the physical make-up of the brain and its physi-
cal processes might be taken as suggesting that neurons and other working parts of the
brain are too hot, too massive and too wet to be materially affected by any quantum
indeterminism or non-locality.
In addition, there is the point that the only indeterminism suggested or permitted by
QM is mere randomness, and not a rational but indeterministic input from conscious
experiences. It can also be argued that, unless in every case one of the alternatives that
are possible according to QM occurs at random within the probability parameters es-
tablished by the laws of QM, then physical law would be violated, in the sense that the
statistical predictions of QM would be falsied.
There are thus three questions that need to be considered here:
1. What is it about the indeterminism of QM that could make room for indeterminis-
tic rational choice?
2. Could brain processes plausibly be affected by quantum indeterminism and/or
non-locality?
3. Is there any inconsistency between the notion of rational indeterministic choice and
the randomness and statistics suggested by QM?
Dealing rst with question 1, standard QM allows for the possibility of indetermin-
ism in the development of a system over time in two ways: rst, in the determination
of what measurements of the system are to be made, and second, in the outcomes of
measurement of the system (reductions of the quantum state).
As mentioned earlier, standard QM postulates two processes by which physical sys-
tems change: deterministic development over time that alters the probabilities of out-
comes of various measurements that could be made of the system (process 2), and the
indeterministic occurrences of possible outcomes when observations or measurements
occur (process 1, the reduction of the quantum state).
The determination of what measurements are to be made on a quantum system is
not prescribed by the laws of QM, and that determination does affect how the system
under consideration changes. Indeed, the determination of when measurements are to
be made can substantially affect the probabilities of outcomes. For certain systems,

sufciently frequent measurements can make vanishingly small the probability of an
outcome that would otherwise have substantial probability: this is the so-called
quantum Zeno effect (or a watched pot never boils).3 It could be that conscious input
into brain processes has the effect of determining when and what measurements of
quantum systems of the brain are made; and physicist Henry Stapp argues very persua-
sively that this is indeed the case, and that the quantum Zeno effect thereby plays a
signicant role in conscious decision-making.4
It might be objected that the determination of when and what measurements of
quantum systems are made must itself depend on changes in other systems (the mea-
suring systems) that are themselves constrained by the laws of QM; so that this deter-
mination cannot be an additional source of indeterminism, over and above the
indeterminism involved in the outcome of measurements. There is some force in that
objection; but the whole question of what amounts to measurement in QM is so little
understood that the objection does not justify wholly discounting the determination
of what measurement is to be made as a possible source of indeterminism.
The second and more obvious source of indeterminism is in the actual outcomes of
measurements or reductions of the quantum state. According to standard QM, these
outcomes occur randomly within probability parameters determined by the laws of
QM, so that the results of many measurements reliably reect the statistics that the
laws of QM prescribe. It could be that conscious input into brain processes has the
effect of determining which of the outcomes permitted by the laws of QM occur, either
from a measurement of one extended quantum system of the brain, or from co-ordi-
nated measurements of many associated quantum systems of the brain.
Relevant indeterminism in the brain could conceivably come from either or both of
these sources; and it could possibly also come in some other way, if advances in solving
the measurement problem of QM suggest that standard QM is not wholly accurate in
the sharp distinction it draws between process 1 and process 2.
Given that there are these possible sources of indeterminism permitted by the
laws of QM, the second and third questions I have identied arise. In relation to
question 2, I say it is by no means implausible that brain processes could be affected
by quantum indeterminism and non-locality, although there is at present insuf-
cient understanding of the brain to make condent assertions as to precisely how
this happens.
It has been established that a single photon of light can activate a rod cell in the
human retina,5 so it cannot be argued that a cell such as a neuron could not possibly be
3
Sudbery (1986), 19293.
4
Stapp (2009).
5
Schnapf and Baylor (1987), 35.

affected macroscopically by single quantum events. And various suggestions that have
been made as to how quantum indeterminism and non-locality could be signicant in
brain processes give some indication of broad outlines of possible mechanisms by
which quantum scale events could affect brain processes.6
Roger Penrose, for example, suggests that non-algorithmic rationality could be sup-
ported by non-local co-ordinated process 1 events (measurement-like reductions of
quantum states) occurring over extended regions of the brain. It seems that this would
require non-local correlations (termed quantum entanglement or coherence) extending
over substantial regions of the brain for periods of time sufcient to support conscious
experiences, say of the order of at least one-tenth of a second; so Penrose considers
whether there are any features of the brain that could support the spread and duration
of quantum entanglement to that extent. He adopts a suggestion rst made by Stuart
Hameroff that structures within the neurons of the brain called microtubules could
isolate quantum states in such a way as to do just that.
Physicist Henry Stapp accepts that classical physics provides us generally with very
close approximations to the observable properties and functioning of mind-indepen-
dent macroscopic physical entities in the world, including human brains; and that if
one is to show that QM plays a signicant role in the functioning of the brain, an ex-
planation is required as to how this can be, given what we know about the conditions
obtaining in the brain. He postulates7 that a function of brain activity is to process
information about the world coming in from sensors, in order to produce a template
for action which can give rise to appropriate action. One feature of the complex dy-
namic systems studied in chaos theory is that, while minute differences in initial condi-
tions can give rise to huge differences in outcomes, calculation of outcomes often
shows high probabilities for outcomes approximating closely to a small number of
states, called attractors. Stapp suggests8 that this may be so for the brain, with the vari-
ous attractors representing the various possible templates for action.
Stapp goes on to analyze the effect of Heisenberg uncertainties in the position and
momentum of pre-synaptic calcium ions; and concludes that this involves uncertain-
ties as to discharge of neurotransmitters, and that this in turn gives rise to myriad dif-
ferent possibilities, each of which could be expected to evolve into something very
close to one of a small number of different attractors. The selection between the attrac-
tors could then occur by a reduction of the wave packet, that is, by a process 1 reduc-
tion of the quantum state. Stapp concludes:9
6
See Penrose (1994), Stapp (1998), Eccles (1994), and Jibu and Yasue (1995).
7
In Stapp (1993), chap. 6.
8
Stapp (1998), 8.
9
Stapp (1998), 10.

It should be emphasized that this effect is generated simply by the Heisenberg
uncertainty principle, and hence cannot be simply dismissed or ignored within
a rational scientic approach. The effect is in no way dependent upon macro-
scopic quantum coherence, and is neither wiped out nor diminished by ther-
mal noise. The shower of different macroscopic possibilities created by this
effect can be reduced to the single actual macroscopic reality that we observe
only by a reduction of the wave packet.
Thus, Stapp argues, whereas classical physics has no room for consciousness and offers
no possibility of any explanation of or role for consciousness, the picture of the world
provided by QM incorporates consciousness in a natural and parsimonious way.
The late John Eccles, a prominent neuroscientist and Nobel laureate, advanced a
hypothesis that focuses on the triggering of the discharge of neurotransmitters, giving
rather more anatomical detail than Stapp.10
As is well known, the neocortex of the human brain contains thousands of millions
of nerve cells called neurons. Each neuron consists of a body or soma, bres called
dendrites through which signals are received from other neurons, and a bre called an
axon through which it sends signals to other neurons. The axon itself ends in many
branches, and these branches terminate in synaptic knobs or boutons, each of which
closely abuts the surface of a spine from a dendrite, or of the soma of the receiving
neuron: each such area of functional contact is called a synapse. When a neuron signal
reaches a bouton, discharge of neurotransmitters (or exocytosis) may occur, the prob-
ability of such occurrence being of the order of 0.25. If exocytosis occurs, neurotrans-
mitters cross the synaptic space or cleft to the receiving neuron, and there make an
excitatory or inhibitory contribution to the ring of that neuron. Whether or not a
neuron res will depend upon the total of such contributions, through all the synapses
of its dendrites and soma, which may number several thousand for a single neuron.
Something like one-half of the neurons of the neocortex have pyramid-shaped
bodies and dendrites ascending from their apexes towards the surface of the brain. As
they ascend, these dendrites become closely grouped in bundles or clusters, which
comprise dendrites from about 70 to 100 neurons. The dendrites from each neuron
have something like 2,000 spine synapses, some as many as 5,000; so that each cluster
has over 100,000 spine synapses. Eccles contends that these clusters of dendrites, which
he calls dendrons, are basic anatomical units of the cortex.
Eccles gives close consideration to the structure and contents of the pre-synaptic bou-
tons. Each bouton generally contains something like 2,000 vesicles, quantal packages of
neurotransmitter molecules, with each vesicle containing about 5,000 to 10,000 mole-
10
Eccles (1994).

cules. At any time, about 30 to 50 of these vesicles are located in a grid which adjoins the
synapse, and thus are ready for exocytosis. When a neuron signal reaches the bouton,
one (but no more than one) of these vesicles may discharge the whole of its contents
into the synaptic cleft; and this happens only with a probability of the order of 0.25, or
one in four. It is Eccles contention that this probability is an indeterministic quantum
mechanical effect, and not merely an expression of our ignorance of hidden variables.
His hypothesis is that the self (his word for the conscious subject) affects brain processes
by momentarily increasing the probabilities of exocytosis in all of the 100,000 or so
boutons of a dendron. Chapter 9 of Eccles (1994) contains calculations that suggest that
the quantities of energy, distance, and time involved in the process are sufciently small
for quantum mechanical effects to be signicant; and also that conservation laws would
not be violated. On the other hand, simultaneously increasing the probabilities of exo-
cytosis in as many as 100,000 boutons could macroscopically affect brain processes, and
thereby contribute to the realization of subjective intentions.
There are nevertheless arguments to the effect that proposals such as these cannot
overcome the problem of the scale of quantum effects. Recent calculations by Max
Tegmark have suggested that any macroscopic quantum entanglement in the brain
would be destroyed in times of the order of 1013 to 1020 seconds.11 However, Tegmarks
arguments have been criticized by Stuart Hameroff,12 while Stapp13 claims they do not
affect his position.
In any event, even if arguments such as Tegmarks were to show difculties for stan-
dard QM in accounting for any input from conscious experiences, they would not
overcome what I say is the clear inadequacy of classical physics to give a satisfactory
account of the existence and role of consciousness, or avoid the need for some kind of
non-local causation to deal with the binding problem.
Work on applying QM to brain processes continues. A Quantum Mind conference
was held in Salzburg in 2007 (following similar conferences in Arizona in 1999 and
2003), with the following manifesto:
The fundamental question of how the brain produces conscious experience

remains unanswered through classical neurocomputational explanations.
Quantum approaches to consciousness (considered highly unlikely by most
scientists and philosophers) can potentially account for difcult issues (e.g.
unconscious-to-conscious transitions, binding, synchrony, subjectivity) but
appear vulnerable to decoherence at warm brain temperatures. Since the previ-
ous Quantum Mind conference in 2003, evidence has shown or suggested:
11
Seife (2000).
12
Hameroff and Hagen (2000).
13
Stapp (2000).

1) At the molecular level, enzyme-substrate interactions and sensory
transduction in photo-, magneto- and olfactory receptors rely on
quantum mechanisms (tunneling, spin transfer and/or radical
pairs),
2) Cellular level ion channel cooperativity and brain-wide gamma
synchrony coherence appear to require non-local quantum cor-
relations among states of proteins and ions,
3) Psychoactive molecules interact with receptors through quantum
correlations,
4) Quantum spin transfer through biomolecules is enhanced by in-
creased temperature,
5) Quantum computing processes can occur at increasingly warm
temperatures,
6) Robust quantum entanglement can involve millions of atoms or
molecules.
Further,
7) Quantum time symmetry can rescue consciousness from the un-
fortunate role of epiphenomenal illusion forced by classical
neurocomputation,
And nally
8) Quantum approaches offer a possible answer to the ontological ques-
tion: what is consciousness and what is its place in the universe?14
My own conclusion is that arguments to the effect that there is no room in brain pro-
cesses for any indeterminist input from conscious experiences, and that those pro-
cesses could not be affected by quantum indeterminism or non-locality, are not strong.
It is true that current mainstream neuroscience does not suggest that unied conscious
experiences can have effects on brain processes, and does tend to suggest that indeter-
minism associated with quantum physics occurs at too small a scale to accommodate
such effects. But this is precisely why mainstream neuroscience fails to account for or
accommodate conscious experiences, and through lack of understanding of them
minimises their signicance; so these things are indicative of limitations of mainstream
neuroscience rather than inefcacy of conscious experiences.
8.7. It has not been shown that neurons and other working parts of the brain are too hot,
too massive and too wet to be materially affected by any quantum indeterminism or non-
locality.
14
See http://www.quantumbionet.org/eng/index.php?pagina=107.

This brings me then to the third question I identied; and in relation to this question,
my contention is that there is no inconsistency between the notion of rational indeter-
ministic choice, and the randomness and statistics suggested by QM. In particular, I
say that the question whether rational indeterministic causation, of the type I am con-
sidering, violates or satises the probabilistic laws of QM, raises no difculty for my
position, for a number of reasons.
First, there is, I suggest, considerable difculty in identifying what, in the case of ra-
tional choices, would actually amount to a violation of the probabilistic laws of QM.
The statistics of QM generally support high probabilities only in circumstances where
there are large numbers of comparable events. Rational decisions are generally the
outcome of highly complex and individual circumstances, and are not the kind of
thing to display statistics of the kind that would engage with the probabilistic laws of
QM. Further, where (as must often be the case in decision-making) an outcome can
require the occurrence of all of a number of requirements, highly improbable out-
comes can occur through a combination of probable events. The question whether
conformity with the statistics of QM would then require that the highly improbable
outcome not occur, or rather that one or more of the probable contributors to that
outcome not occur, is not readily answerable.
Second, even assuming that good sense can be made of the application of QM statis-
tics to rational choices, it is very plausible that the felt strength of reasons would be well
correlated with statistical laws of QM, in their application to the development over
time of the physical systems that contribute to decision-making; so that any signicant
departure from QM statistics, if they could be calculated for something as complex as
a rational decision, is highly unlikely.
Third, however, I am contending that the capacity to make rational albeit indeter-
ministic choices has been selected in evolution because, on the whole, it produces out-
comes more conducive to survival and reproduction than random occurrences within
QM statistics; so that while rational choices may not violate QM statistics, I do contend
they need not and generally do not conform to what would occur randomly within
these statistics. So it could be argued there is a theoretical possibility that even though
violation of QM statistics would not be detectable or provable, violation could in fact
occur. What I say to this is that QM can only prescribe what follows from features of
the world that engage with laws of nature, and can say nothing about what if anything
follows from features that do not engage with laws of nature. It can thus rule out some
things, and can give probabilities for others based on the features that engage with laws
of nature; so that, if it were accepted that features that do not engage with laws of
nature can have no input into what happens, what actually happens would have to be
seen as happening randomly within these probability parameters. But I have given
reasons for accepting that features that do not engage directly with laws of nature, and

do not generally engage with effective computational rules and thus do not engage
indirectly with laws of nature, namely feature-rich gestalts of conscious experience,
can have an input into what happens; and this input cannot be fully accounted for by
QM, or indeed by any system of physical laws of general application.
Accordingly, while I cannot altogether exclude the possibility that rational choices
could conict with the statistics prescribed by QM, I say that this would not in any
event amount to violation of physical law, because QM can prescribe statistics only for
the general and quantitative features of the world with which its laws can engage. And
arguments to the effect that indeterminism must prejudice rationality miss the points
(1) that our rationality seems to be non-algorithmic, (2) that the operation of QM in
the brain may permit non-algorithmic rational processes in ways such as those I have
suggested, and (3) that, if there were any ways in which quantum processes could be
used to advantage in the brain, it is likely that evolution would have found them.
So:
8.8. Even if it were the case that the statistics of rational decision-making (if it were pos-
sible to calculate them) did not completely accord with those indicated by QM, this would
not be a violation of physical law but an indication of a limitation on its applicability.
LI B ET, GAZ ZAN IGA AN D WEGN E R
It is sometimes suggested that certain results established by neuroscientists and psy-

chologists are inconsistent with the existence of efcacious conscious choice. I will
consider three prominent examples.15
First, there are the experimental results of Benjamin Libet16 (referred to earlier)
showing that unconscious processes precede some conscious decisions. In particular,
there were experiments in which participants were asked to press a button at any time
they chose, and to note the exact time shown on a large display when they decided to
initiate the movement. Readings of brain activity of these persons showed neural prep-
aration for the action some tens of seconds prior to the time noted by the participants
as the time of deciding to initiate the movement.
These experiments seem to me entirely neutral as to whether there are rational inde-
terministic conscious inputs into decision-making. The most signicant relevant deci-
sion by the participants, namely the decision to participate in the experiment and to
push the button within a short period of time, is made before the readings of brain
15
Mele (2009) contains excellent discussions of examples such as these, generally supporting my
contentions here.
16
Libet et al. (1983).

activity. In relation to the decision to actually initiate the movement, I see it as entirely
unsurprising that there is unconscious neural preparation before the person has im-
mediately available a spectrum of possibilities, which include both carrying through
without delay what has been unconsciously begun, and also steering in a somewhat
different direction (perhaps delaying a little, or even exercising what Libet himself
called free wont and not going ahead at all). And since there has already been the
deliberate conscious decision to push the button at around this time, it is unlikely that
there would be any reason why the person would not go with the ow and simply carry
through what has been unconsciously begun.
There is also another set of experimental results of Libet17 that suggest consciousness
comes too late to play a role in our ongoing activities, since there is a delay of about
half a second between the initial arrival of a novel sensory stimulus to the brain and its
experience in consciousness. However, those experiments did not concern a situation
where something is already being consciously experienced, which is adjusted by reason
of ongoing stimuli. In such cases, there seems to be little or no appreciable delay be-
tween the arrival of a sensory stimulus appropriate to adjust the ongoing experience
and the actual conscious adjustment of that experience; and this can explain why con-
sciousness can be efcacious in the shaping of actions, as in artistic and sporting per-
formances. Performers emphasise the importance of concentration, and it seems clear
that a performance by a pianist, for example, can be moving to the pianist and the
audience because the pianist is consciously responding to heard sounds and felt emo-
tions in shaping the performance.
Next, there are experimental results obtained by Michael Gazzaniga illustrating our
tendency to rationalise our conduct by (unconsciously) fabricating, and then believ-
ing, plausible but untrue stories to explain why we did what we did. Gazzaniga con-
ducted experiments with people who had undergone a procedure (referred to in
chapter 4) to control severe seizures involving the severance of the corpus callosum
joining the hemispheres of the brain, giving rise to some separation of the functioning
of the hemispheres, with the left hemisphere receiving information from and control-
ling the right side of the body and the right hemisphere receiving information from
and controlling the left side of the body. In most patients, communication was
generated exclusively by the left hemisphere. In the experiments, patients were induced
to make a choice, displayed by the left side of their bodies and using information re-
ceived only by the right hemisphere of their brains, and then asked to explain it. The
patients, using their communicating left hemisphere that was unaware of what had
actually induced the choice, sometimes came up with plausible reasons for their choice
that had nothing to do with the true explanation.
17
Libet et al. (1979).

In some experiments,18 patients were shown two pictures, one exclusively to the right
eye (and thus the left hemisphere of the brain) and the other exclusively to the left eye
(right hemisphere), and asked to select, from an array of pictures in view of both eyes,
the picture associated with the picture shown to each eye. In one experiment, a picture
of a chicken claw was shown to the right eye, and a picture of a snow scene was shown
to the left eye. From the array of pictures, the patient selected a picture of a chicken
with his right hand, and a picture of a shovel with his left hand. When asked why he
chose those items, the patient replied The chicken claw goes with the chicken, and you
need a shovel to clean out the chicken shed. Gazzanigas interpretation of this was that
the patients communicating left hemisphere, unaware that the right hemisphere had
been shown a snow scene and aptly selected a picture of a shovel, unconsciously fabri-
cated and believed a false story, consistent with the knowledge available to it, to explain
why the picture of the shovel had been selected.
Results such as these may be compared with what has sometimes been observed in
cases of post-hypnotic suggestion. Subjects under hypnosis can be given some instruc-
tion and told to follow it later on when they have awakened; and in some cases, even
quite bizarre suggestions are followed. The subjects then sometimes give plausible rea-
sons why they did what they did, reasons that of course have nothing to do with the
post-hypnotic suggestion.
In both these cases, subjects are confronted with a situation where they have just
done something that has the appearance to them of a voluntary action even though
they are not in a position to know the true motivation of the action. I suggest it is not
surprising that some subjects attempt to make sense of this seeming contradiction by
coming up with the best explanation they can think of. I contend it says very little
about whether the reasons we think we have for ordinary decisions and actions are true
reasons or mere rationalisations; and so carries very little weight against the arguments
of this book.
Third, there are experimental results of Daniel Wegner19 which he contends show
that the will is an experience fabricated from perceiving a connection between
thought and action. In one experiment, a subject and another person were given joint
control of a cursor on a computer screen (their ngertips on a board mounted on a
computer mouse) and told to move the cursor round a computer screen and to stop
moving it every 30 seconds or so. The other person caused the cursor to stop in cir-
cumstances arranged in such a way as to suggest to the subject that it was the subject
that had stopped the cursor; and the subject sometimes reported having caused the
cursor to stop.
18
Gazzaniga (1988), 1114.
19
Wegner (2002), 74.

Again, I contend that such experiments carry little weight against my position. They
merely show that we can make mistakes about our own mental processes, particularly
if concerted and skilful attempts are made to bring this about.
Questions I would put to Gazzaniga and Wegner are whether they engage in con-
scious thought in order to devise their experiments, and by what means do they seek
to combat the unconscious biases that we all must contend with in our reasoning.
Their answer, I suggest, could only be that they do engage in conscious thought, and
that they try to identify such unconscious biases in themselves and to neutralise
themwhich must itself be a conscious process. Clearly, I suggest, they must rely on their
conscious reasoning in devising their experiments and reaching their conclusions. If
they did not do so, or if their conscious reasoning was inefcacious as they suggest,
their work would be entirely at the mercy of the unconscious biases that affect us all.
8.9. Experimental results of neuroscientists and psychologists such as Libet, Gazzaniga

and Wegner are not inconsistent with the existence of efcacious conscious choice.
COR E ASS E RTION S AB OUT N E U ROSCI E NCE

AN D CON SCIOUS CHOICE
8.1. Our brains are constituted by the same physical elements as other physical objects,
and are constrained by laws of nature as are other physical objects.
8.2. Our choices and actions are to a considerable extent the product of activity of our
physical brains constrained by laws of nature, and much of our motivation is unconscious
or barely conscious.
8.3. The physical sciences can accommodate the existence of unied conscious experi-
ences, corresponding with events in widely separated regions of the brain, having effects,
not determined by rules, in one or more of those regions.
8.4. If it were possible to abstract physical brain processes and/or computational infor-
mation-processing from the totality of conscious decision-making, they would give the
appearance of producing probability-weighted spectra of outcomes, which include the
actual outcomes of the conscious decision-making.
8.5. Rather than conscious subjects making discrete conscious choices between alterna-
tives thrown up by unconscious processes, they use both conscious experiences and rule-
determined processes together, on a continuous basis, to steer a course within spectra of
possibilities left open by laws of nature and computational rules.
8.6. There is no localised Cartesian theatre where conscious experiences occur, and the
solution to the binding problem must probably be found in the non-local nature of quan-
tum systems.

8.7. It has not been shown that neurons and other working parts of the brain are too hot,
too massive and too wet to be materially affected by any quantum indeterminism or non-
locality.
8.8. Even if it were the case that the statistics of rational decision-making (if it were pos-
sible to calculate them) did not completely accord with those indicated by QM, this would
not be a violation of physical law but an indication of a limitation on its applicability.
8.9. Experimental results of neuroscientists and psychologists such as Libet, Gazzaniga
and Wegner are not inconsistent with the existence of efcacious conscious choice.

9
Indeterministic Free Will
The arguments of the previous chapters have supported the view that human beings
make decisions as to what to believe and what to do that are not pre-determined by
prior conditions and laws of nature, yet are not random but are apposite responses to
circumstances facing them; so that these decisions can be both indeterministic and
rational. These chapters have addressed the two major difculties facing an account of
indeterministic free will and responsibility, namely how to make sense of rational cau-
sation that is neither deterministic nor random, and how to reconcile that kind of
causation with what science tells us about the world. What I want to do in this chapter
is to complete my account of indeterministic free will and responsibility, to relate that
account to other views generally favourable to free will, and to show how my account
deals with objections to free will and responsibility based on arguments about luck.
The points I developed in chapters 5 and 6 were that, in making these decisions, per-
sons respond appositely to gestalt experiences that are too feature-rich to engage as
wholes with laws of nature and generally do not engage with computational rules, so
that the response is not determined by laws or rules of any kind. I would add now that
each person also has a particular combination of features; and just as it is reasonable to
accept that the content of experiences can as a whole contribute to the persons re-
sponse, I suggest it is also reasonable to accept that the person, the subject of the experi-
ences, can also as a whole contribute to the response. That is, I suggest that both the
content and the subject of experiences are not mere aggregations of general features
and/or mathematical variables with which laws of nature can engage, but are particu-
lar combinations of features which as wholes contribute to outcomes in ways that are
not determined by laws or rules.
In chapters 7 and 8, I have outlined how these contributions may be accommodated
within what science tells us about the world. I suggested that the indeterminism and
non-locality that are features of QM could allow room for an input from the conscious
aspect of conscious processes of the brain-and-mind, while at the same time I sug-
gested that this should not be taken to be an input that is discrete from and indepen-
dent of the non-conscious or physical aspect of these processes. The picture I am
suggesting is of rational decision-making by conscious subjects (persons), which
153
cannot be broken down into component parts, but which occurs within constraints
prescribed by the combination of the physical aspect, considered in isolation, and the
laws of nature.
While chapter 8 focussed on decisions by human beings, I do say, consistently with what
I wrote in chapter 6, that the brains of animals that have conscious experiences somewhat
like ours must also in a similar way permit of input from the conscious aspect of their con-
scious experiences. This gives rise to the question of when and how consciousness, and thus
this kind of contribution to what happens in the world, rst emerged. I do not suggest that
consciousness appeared suddenly at some stage in the evolution of life, for example with the
earliest birds or the earliest mammals. If it were the case that it is only in birds and in other
mammals that there is consciousness similar to ours, then I would suggest that in earlier
forms, such as reptiles, amphibians and sh, there must be a kind of proto-consciousness.
To use an expression of neuroscientist Susan Greeneld,1 I say that consciousness comes
with a dimmer switch, rather than just an on-off switch. I think that is true of our own ex-
perience, and that it is also true of the way consciousness emerged in evolution.
This gives rise to the further question of exactly when it was that the dimmer switch
of consciousness rst began to move from the off position. Was it when life-forms with
a nervous system rst emerged, or was it at some other time? This in turn engages the
issue of panpsychism versus emergence, referred to in the introduction. My position
on this is that there must always have been, in the universe, the potential for the emer-
gence of consciousness; but that there could only be anything like the actual begin-
nings of the consciousness that we enjoy when the existence of co-ordinated systems
of considerable complexity made it possible that there be inputs into what happened
that were neither rule-determined nor merely random. Scientists studying complex
systems refer to the occurrence of self-organisation in such systems, generally without
making it clear whether or not this kind of behaviour is simply the working out of
rule-determined behaviour in these systems. It may be that the very beginnings of our
consciousness can be traced back to some kind of self-organisation in complex sys-
tems, which might have been either before or after the rst emergence of life.
WI LL AN D R E S PON S I B I LITY
There is little more I can usefully say on those questions, and I now wish to say some-
thing about the will that I say is free, to some extent at least, and about responsibility.
John Searle has identied three gaps that exist between a persons reasons for ac-
tions and the actual carrying out of actions:2 (1) between reasons and prior intention
1
Greeneld (1999).
2
Or, as I put it in Hodgson (1999), areas in which reasons do not include a clincher and thus fail to
be conclusive. See Searle (2001), 6263.

to act (that is, the decision to act); (2) between prior intention to act and intention-
in-action (that is, the actual initiation of a voluntary action); and (3) between initiat-
ing intention-in-action and carrying the action through to completion. I suggest that
rational and non-random yet indeterministic processes can operate at each of these
stages.
It is perhaps in relation to the second and third of these gaps that the notion of will
is most applicablecertainly, it is generally in relation to them that we think of strength
of will and weakness of will. So far, my discussion has been concerned mainly with the
rst of those gaps, and has been particularly relevant to the formation of beliefs as to
circumstances in which we nd ourselves and as to the merits of available alternative
actions in those circumstances. However, the progression from having reasons for be-
liefs relevant to action to adoption of these beliefs is I contend an important part of
what is generally understood as free will; and now I suggest that, if it is reasonable to
accept (as I have argued) that rational and non-random yet indeterministic processes
are operative in the adoption of beliefs relevant to action, it is reasonable as well to
accept that processes of a similar kind are operative also in putting these beliefs into
effect by deciding to act and proceeding to do so.
9.1. Rational and non-random yet indeterministic processes are operative in the adop-
tion of beliefs relevant to action, and also in putting these beliefs into effect by deciding to
act and proceeding to do so.
I have said that the subject of conscious experiences is the person, the physical-and-
mental system that is the human being; and my contention is that if free will with re-
sponsibility is to be attributed to anything, it is to each person or human being.
It is I think widely accepted that most adult human beings, when conscious and
not in an abnormal mental state, can rationally make decisions about what to do,
about whether these decisions are to be put into effect, and about whether to carry
them through to a conclusion; and that the making of these decisions causes the
person to do something rather than do something else or do nothing. In that sense,
persons are widely regarded as having the capacity to recognise and respond to
good reasons, as rationally controlling their own conduct, as being free to do one
thing rather than another, and as being responsible for doing one thing rather than
another.
However, a problem arises when one considers in depth what a person is and how
decisions are made. It can be contended that persons are constituted by physical matter
undergoing physical processes that unfold in accordance with laws of nature, and that
the decisions made as to what to do and so on are just the working out of these physical
processes; so that in truth the person has no more free will to do one thing rather than
Indeterministic Free Will 155

another than has (say) a chess-playing computer in relation to chess moves it decides
to make.
In the case of decisions made by chess-playing computers, there is a sense in which
the computer makes a decision, on a rational basis, between alternatives that are open,
namely the moves that are possible according to the rules of chess. In that sense, the
computer rationally controls its conduct, is free to make one move rather than an-
other, and is responsible for making one move rather than another. But when one
considers the physical construction of the computer and the program according to
which it operates, it can be seen that (except to the extent that the program may allow
for random occurrences or might malfunction) there never is more than one move
possible, given the conguration on the chessboard and the computers physical con-
stitution and program.
I dont believe anyone would attribute free will or responsibility to a chess-playing
computer.
Some would say that that this is not because of the inevitability of the outcome, given
the computers physical constitution and program, but rather because the computer
does not have the appropriate capacities to recognise and respond to good reasons and
to control its conduct accordingly, these being capacities that human beings do have
and that are perfectly compatible with the inevitability of outcomes. I will return to
this possibility when I consider compatibilism; but I note that it does require an expla-
nation of what are the capacities, compatible with inevitability, that human beings
have but chess-playing computers do not have.
Others would see free will and responsibility as necessarily excluded, in the case of a
chess-playing computer, by the inevitability of the outcome: they would argue that this
inevitability means that the computer could not possibly have the appropriate capaci-
ties, because these capacities cannot co-exist with determination of outcomes by factors
outside the computers control (including its physical constitution and program).
Whichever view is correct, the same problem does not arise for human beings, if the
arguments of the previous chapters are accepted. These arguments suggest that con-
scious decisions and actions can be rational and apposite to circumstances facing the
person in question, yet not limited to just one possibility (or to mere randomness) by
prior conditions and laws of nature and/or computational rules. They suggest this is
so because there is a contribution to the decision or action from what the person
makes of feature-rich gestalts that do not engage with laws or rules, a contribution
that is itself apposite so that it contributes to the rationality and appositeness of the
decision or action.
As will be seen from previous chapters, I do not suggest this contribution from the
persons response to feature-rich gestalts is discrete from and independent of other
factors contributing to the decision or action: the decision or action is that of the

person, and the person, while being the subject of conscious experiences, is a physical-
and-mental whole, whose decisions and actions are at once contributed to, and limited
by, the physical aspect and laws of nature, and also contributed to by the mental aspect
(the persons conscious experiences).
What I say is that, if it were possible to abstract from the totality the operation of the
physical aspect alone, together with laws of nature, this would give the appearance of
limiting decisions to spectra of probability-weighted alternatives, with what actually
happens falling within these spectra; while from the point of view of the conscious
subject, there are (1) spectra of alternatives (which may sometimes be no more than
the alternatives of doing something or not doing it, or marginally different ways of
shaping ones actions) and also (2) consciously felt pros and cons of these alternatives.
The conscious subject steers within those spectra, for example by doing or not doing
something, or shaping an action, or deciding upon an answer to a consciously ad-
dressed problem, and the conscious subjects take on the feature-rich gestalts of con-
scious experience contributes to the selection within the spectra of possibilities. Thus
the two aspects of the process can notionally be distinguished, but my contention is
that there are not two separate processes, but one whole physical-and-mental process
producing what actually happens.
On this account, the line of reasoning that may be seen as eliminating freedom and
responsibility in the case of a chess-playing computer is not available in relation to the
decisions and actions of normal adult human beings. There are truly available alterna-
tives (the spectra of possibilities), the outcome is not inevitable given the relevant
physical circumstances and laws of nature, and the persons control of decisions and
actions is not wholly displaced and thus eliminated by the operation of factors outside
the persons control. The persons control of decisions and actions is itself to some
extent subject to factors outside the persons present control, and to that extent is qual-
ied and reduced; and some of these factors (those consisting of or derived from the
persons genes and environment) were always outside the persons control. But to the
extent that the persons control is not displaced by these factors, the person has free will
and responsibility in a robust sense.
So I contend:
9.2. We have free will in a robust sense because (1) alternatives are truly open, albeit
limited by the engagement of prior conditions (including our own physical-and-mental
state) with laws of nature, (2) the occurrence of one of the alternatives is not random but
the result of our selection between them on rational grounds, and (3) we put our selection
into effect by voluntary action.
9.3. We also have responsibility in a robust sense because (1) we make rational decisions
and put them into effect, (2) factors outside our control at the time do no more than limit

alternatives, give rise to reasons and determine how they appeal, and give rise to uncon-
scious tendencies, and (3) while those factors reduce our responsibility at the time, they do
not eliminate it.
I also contend that decision-making is a process, and that for some decisions it can be
considered a process extending over substantial periods of time. In those cases, the way
a person steers within spectra of possibilities at earlier stages of the process may sub-
stantially affect the spectra of possibilities available at later stages; and thus a persons
responsibility for ultimate outcomes is not limited to responsibility for the course
taken towards the end of the process but may extend, for example, to possibilities lost
because of the course taken at earlier stages.
My discussion so far has focussed on free will and responsibility for normal adult
human beings. It will be recalled that in chapter 6 I suggested that some non-human
animals can respond appositely to whole feature-rich gestalts, so that for those ani-
mals as well as humans such responses make a contribution to decisions and actions
that is not random yet not determined by engagement of rules with pre-decision or
pre-action circumstances. Thus, one of the requirements for my robust account of
free will is satised.
However, I suggest that for free will as generally understood, and particularly for
responsibility, there is a further requirement, namely the self-conscious rationality of
human beings. I accept that some non-human animals have the capacity to make
reasonable decisions on matters relevant to their survival and reproduction, some
have a capacity for self-awareness, and some have motivation with considerable simi-
larities to moral reasons that weigh with human beings (care of offspring, altruism
towards fellow animals, and so on). However, they do not have the exible all-purpose
rationality of human beings, supported by (and perhaps dependent on) a exible all-
purpose language.
The requirements for fully-edged free will and responsibility can usefully be con-
sidered in relation to the development of children. I would suggest that, from the time
they rst exercise some control over their actions, infants respond to feature-rich ge-
stalts; and thus for them, no less for than for non-human animals, such responses
make a contribution to decisions and actions that is not random yet not determined by
engagement of laws or rules with pre-decision or pre-action circumstances. Rational-
ity sufcient to support some responsibility gradually emerges, perhaps sufcient after
a few years to make a parents moderate discipline not merely useful but to some extent
deserved. However, the prefrontal cortex (the area of the brain associated with a per-
sons judgment abilities) develops only gradually, and is said not to be fully developed
until the early twenties; so even when a person is as old as eighteen, rationality and
responsibility is not yet complete.

In Australian criminal law, children under ten are treated as having no criminal re-
sponsibility, while a child between ten and fourteen may be held criminally responsible
if the prosecution proves that he or she knew that a voluntary criminal act performed
was seriously wrong (not merely mischievous). Young persons between fteen and
eighteen are treated as criminally responsible for voluntary criminal acts, but generally
punished very differently from adults and in particular not imprisoned in adult pris-
ons. And the approach to sentencing adopted by the courts generally recognises that
even in early adult years, responsibility is less than optimal.
Increasing knowledge of how the brain works and develops may assist in making
ner and more individualised judgments as to how and when and to what extent re-
sponsibility develops in young persons. Meanwhile, the broad judgments made by
Australian criminal law seem not inappropriate.
COM PAR I SON WITH KAN E
My account of free will and responsibility is indeterministic, and it does not depend on
any sharp distinction between causation by events and causation by agents. It may
clarify my position if I now relate it to other ways in which philosophers have in recent
times sought to maintain ideas of free will and responsibility. I will start by comparing
my views with those of Robert Kane, perhaps the most prominent contemporary pro-
ponent of indeterministic free will. I will then look at views falling under the general
description of agent-causation, and then at views that maintain that free will and re-
sponsibility are compatible with determinism.
Kanes position is most fully developed in his 1996 book The Signicance of Free Will.
However, key elements of his position are usefully summarised in a 2002 article,3 to
which I will refer initially.
Central to his conception of free will is what Kane calls ultimate responsibility, one
element of which is that nothing for which the agent is not responsible should be a
sufcient cause of the decision or act in question.4 According to Kane, ultimate respon-
sibility requires not just that the agent could have done otherwise in respect of some
relevant act, but also that the agent could then have acted voluntarily, intentionally, and
rationally in more than one way. This need not be the case in respect of the very act
responsibility for which is in question, so long as it is the case in respect of acts in the
past by which the agent formed his or her present character, from which the act in
question issued. Kane calls these self-forming actions.
3
Kane (2002b) at 40817.
4
A similar idea has been expressed by Michael McKenna in this way: an agent is an ultimate source
of her action only if no conditions external to her are sufcient for her action. See McKenna (2009),
pars. 2.2, 5.3.4.

Kane suggests that self-forming actions may occur when an agent is faced with
competing motivations, for example in cases where the agent is torn between doing
the moral thing or doing something else he or she strongly wants to do. In such
cases, he suggests, the uncertainty felt by the agent as to what to do is reected in
the indeterminacy of neural processes, whose outcome from the physical perspec-
tive can only be a matter of probability, not certainty. He gives the example of a
businesswoman on her way to an important meeting who sees an assault, and is
torn between doing the moral thing of stopping and calling for help, and advancing
her career by going on to the meeting. He contends that both alternatives are avail-
able, and that whichever she chooses, she will have acted voluntarily, intentionally,
and rationally, and will have by her choice made one set of competing motives pre-
vail over the other.
In many respects, Kanes position is similar to mine. In his 1996 book,5 he endorsed
four key contentions that I have advocated in past writings6 and continue to advocate
in this book:
1. The contention that, prior to a choice being made, an agents reasons are character-
istically inconclusive, inter alia because they are incommensurable; and that it is the
agents choice or decision which resolves the issue. Kane endorses the idea of incom-
mensurability (at 167); and (at 133) he postulates that, in situations where an agent
has to choose between alternative courses of action and has reasons or motives sup-
porting each alternative, the agent makes one set of reasons or motives prevail over
the others by deciding.
2. The contention that what the physical perspective can only treat as a chance occur-
rence may correctly be seen from the mental or experiential perspective as an agents
choice. Kane says (at 147) that from the physical perspective, free will looks like chance,
since from a physical perspective, there is just an indeterministic chaotic process
with a probabilistic outcome; whereas experientially considered, the process is the
agents effort of will and the single outcome is the agents choice.
3. The contention that the problem of free will is closely interlinked with the problems
of consciousness and of the indeterminism disclosed by QM. Kane asks (at 148),
How can a physical process of the brain be at the same time a consciously experi-
enced effort of will?; and suggests that this is just part of the mystery of how neural
rings in the brain could be conscious mental events. And (at 15051) he suggests it
is also implicated with the general problem of indeterminacy-in-nature introduced
by quantum physics.
5
Kane (1996).
6
Notably Hodgson (1991), 13335, 38994; (1999).

4. The contention that the objective probabilities for various outcomes are to some
extent reected in the subjectively felt strength of reasons; and that rational decisions
may nevertheless be made in favour of actions with lower antecedent probabilities.
Kane points out (at 177) that antecedent probabilities of available alternatives do not
necessarily indicate which of them are more or less rational for the agent to choose.
Some differences between our positions may be seen as matters of emphasis. Kane fo-
cuses on decisions about what to do, whereas I link free will with plausible reasoning
about what to believe, as well as to other decisions about what to do. Kane also focuses
on those cases where an agent is torn between alternatives, whereas I would cast the net
somewhat wider, as extending to cases where the agent sees the reasons as pretty clearly
favouring one alternative rather than another: as I pointed out earlier, four decisions
each of which had a prior probability of 0.7 can bring about a result with a prior prob-
ability of 0.24.
I think the most important difference between our positions lies in what I see as an
absence in Kane of an explanation of how it is that the occurrence of one of the pos-
sible alternatives rather than the other, which looks like chance from a physical per-
spective, can be a voluntary, intentional, and rational choice between them. He says, in
the case of his businesswoman example, that whichever alternative is chosen is ratio-
nal, because it is supported by the reasons for that alternative, and it is also voluntary
and intentional, because it is chosen and made to happen by her.
Kane adopts what he calls a teleological intelligibility theory, according to which free
actions can be made intelligible in terms of reasons and motives, and explanations in such
terms are causal explanations. He says that his account assumes that choices can be pro-
duced by reasons or motives of the agent, and that this is both needed to account for ra-
tional agency and also an assumption shared by compatibilist accounts of free agency.7
I agree with this; but it seems to me that Kanes development of this approach falls
short in relation to the selection between the alternatives open to the agent. He claims
in effect that there is a non-arbitrary judgment by the agent that one set of reasons
should prevail over the other set of reasons, but Kane gives no explanation from the
experiential perspective of why or how such a judgment occurs or what makes it non-
arbitrary. Certainly, he makes no suggestion as to how the agents reasons could explain
which way the judgment goes: he does not point to any reasons, beyond the closely
balanced reasons supporting each alternative, that explain why one and not the other
prevails. I think this also creates a difculty for him in explaining the ultimate
responsibility of the agent for the choice. Kane does not suggest there is any relevant
indeterminism operating in the occurrence of the reasons or the felt strength of their
7
Kane (2002b), 416, 42426.

appeal, so ultimate responsibility must lie in the selection of one set of reasons over the
other; and it is precisely this that is given no explanation from the experiential perspec-
tive, other than that the agent chose it and made it happen.
This absence in Kanes account has led Saul Smilansky to claim8 that, on Kanes ac-
count, whether the choice goes one way or the other is arbitrary and not under the
agents control. Now I dont think this claim is correct; but the absence of a developed
account of how the choice between alternatives is made, from the experiential perspec-
tive, makes Smilanskys charge a plausible one.
My approach lls this gap. It says that in making this kind of choice, the agent is re-
sponding appositely to reasons which do not operate by engagement with laws or rules,
so that the outcome is rational although not determined by the engagement of existing
circumstances with laws or rules. The selection is not arbitrary, because the conicting
reasons do not merely support each alternative between which a choice is to be made,
but operate in the selection itself by way of the agents apposite non-rule-determined
response to them.
This in turn raises a further possible difference between Kanes position and mine.
Kane eschews any special forms of agency or causation, and in particular distances
himself from views that have been given the description agent-causation; whereas it
may be said that my view does involve a special form of causation, one which operates
in addition to the ordinary physical causation that apparently proceeds in accordance
with laws of nature and randomness. I will return to this question after giving a brief
account of agent-causation.
AGE NT-CAUSATION
One way in which some philosophers have sought to maintain notions of free will and
responsibility is by suggesting a distinct kind of causation that is operative in decisions
and actions of human beings, called agent-causation. A recent explanation of agent-
causation is given by philosopher Randolph Clarke in Stanford Encyclopedia of Philoso-
phy,9 as follows:
On what are called agent-causal views, causation by an agent is held not to

consist in causation by events (such as the agents having certain reasons). An
agent, it is said, is a persisting substance; causation by an agent is causation by
such a substance. Since a substance is not the kind of thing that can itself be an
eect (though various events involving it can be), on these accounts an agent is
8
Smilansky (2002b), 504 n. 1.
9
Clarke (2008).

in a strict and literal sense an originator of her free decisions, an uncaused
cause of them.
Clarke goes on to identify difculties with this approach, including its denial that
events such as the agents having certain reasons cause free decisions or actions.
There is indeed a tension between the agent-causal view that agents not events
cause free decisions and actions, and the commonsense understanding that an
agents having reasons for decisions and actions does contribute causally to the
happening of the decision or action. Different exponents of agent-causal views have
sought to resolve this tension in different ways. For example, Timothy OConnor
suggests:10
The resolution of the tension, I suggest, is to understand reasons as partially

xing the relative probabilities of particular agent-causings, while not produc-
ing them. We might say that reasons inuence freely made choices by structur-
ing the agent-causal capacity, giving us varying propensities towards dierent
outcomes. On this view, our choices are embedded in the unfolding processes of
nature insofar as those processes, operating through our cognitive and motiva-
tional states, make us objectively inclined to varying degrees to just a limited
few alternatives. Yet the choices themselves are not a direct product of such
shaping forces, but of ourselves.
Clarke also identies other difculties for agent-causal views, including (1) its denial
that any event prior to the decision or action causes it, suggesting that what happens
can only be by luck or chance; and (2) the implausibility of the notion that decisions
or actions happen precisely when they do without some event that brings them about
at that time. These difculties are I think particular examples of a more general dif-
culty with agent-causal views, namely the inscrutability of the origins of decisions and
actions, according to these views: once one has identied the occasion for a decision or
action, the possible outcomes, and the propensities towards different outcomes, it ap-
pears there is nothing that can be said as to why one or other outcome occurred, apart
from saying it was caused by the agent.
It is useful here to see what Kane says about agent-causation.
Kane says that his own account does postulate agent-causation,11 in the sense that
the agent causes free decisions by making one set of reasons prevail over the other. He
says however that this does not make his view an agent-causal view, of the kind es-
10
OConnor (2005).
11
Kane (2002b), 42728.

poused by philosophers such as Clarke and OConnor, because what Kane sees as
causation by the agent does not displace or supplement causation by the events of the
agent having and giving effect to reasons. What in his view make agent-causal views
different from his is that they seek to add agent-causation, of a non-event kind, which
somehow operates over and above the agents having and giving effect to reasons; and
he squarely rejects that.
I take a similar view. I say that agents do cause free decisions, because they cause one
thing rather than another to happen, in the way I have described in this and earlier
chapters. However, I do not see that any contribution to this process by the agent is ad-
ditional to the contribution by events involving the agent, and I do not agree with agent-
causal views that assert there is such an additional contribution.
An important feature of agent-causal views is that they seek to establish agents as
originators of decisions (uncaused causes), by assuming that as substances they cannot
be effects but can be causes. I nd this assumption mysterious and objectionable. If
one takes a view about causation such that substances can be causes (so that, say, a
sculpture can be a cause of a lump on a persons head when it falls on the person), then
what justication could there be for denying that substances can also be effects (so that
the lump can be an effect of the sculpture falling, and the sculpture can be an effect of
the work of the sculptor)? I do not believe any justication for asymmetry in this re-
spect has been shown by proponents of agent-causation in the sense explained above.
Further, even if this assumption were to be accepted, I cannot see how substances can
reasonably be considered as making a contribution to the causation of events that is
somehow additional to the contribution made by events involving the substance. How
can the sculpture make some causal contribution to the lump appearing on the per-
sons head that is additional to the effects of the sculpture falling on the person? By the
same token, how can an agent make contribution to the occurrence of one choice
rather than another, that is additional to the agents activity giving rise to the choice,
including the agents having and responding to reasons? Again, I do not believe any
satisfactory answer to this has been given.
There remains the question whether my approach differs from Kanes in proposing
a special form of causation. I do, as he does not, draw a distinction between aspects
of causation, namely between (1) the aspect of causation studied by physical sci-
ences, concerning the development of physical systems in accordance with con-
straints of laws of nature, and (2) the aspect of causation that operates in the
conscious decisions and actions of conscious organisms. In that respect, I believe my
view is different from his.
However, the distinction I draw is not a sharp distinction, giving rise to mutually
exclusive processes: in the case of the latter aspect of causation, the former aspect is still
fully operative, limiting what can happen to spectra of possibilities, and giving what

are, on the basis of the physical facts alone, objective probabilities for what is to happen
within these spectra. The latter aspect of causation then involves a selection within
these spectra by the conscious organism, but with contributions to that selection being
provided by processes that are both physical and mental, that is by processes that are
both unconscious and conscious. What gives the conscious organism control is the avail-
ability of spectra of possibilities, and its capacity to make something happen within
these spectra, taking into account consciously held reasons that are inconclusive.
Thus on my account a conscious subject or agent does control (and thus cause) the
decision or action, but this does not suggest any contribution from the agent as a sub-
stance that is additional to the events of the agents activities giving rise to the decision
or action. And although the outcome is not pre-determined by pre-decision circum-
stances and laws of nature, it is not the case that these circumstances and laws rst de-
terministically give rise to probability-weighted alternatives, between which the
conscious subject then inexplicably chooses. Rather, the choice follows indeterministi-
cally from the totality of causes without there being any such intermediate stage. Thus
the choice is not inscrutable: although it is not uniquely determined by pre-decision
circumstances and laws, or indeed by the agents reasons, it is understandable as a ra-
tional decision by the agent.
9.4. Conscious decisions and actions are caused indeterministically by the totality of the
relevant unconscious and conscious processes, including the conscious subjects choice
taking into account consciously held reasons.
Finally in this section on agent-causation, I return to the point that proponents of

agent-causation are seeking an uncaused cause in order to account for free decisions.
I say it is not necessary to nd an uncaused cause. The agent, or events involving the
agent, can be seen as causing the agents decisions and as themselves being caused; but
nevertheless those decisions are determined, not by the engagement of laws with pre-
decision circumstances, but by the agents rational choices.
My position can be illustrated by an analogy. The philosopher John Fischer has writ-
ten that our behaviour may well be in the cards in the sense that we simply have to
play the cards that are dealt us.12 This has drawn the apt comment from philosopher
Kip Werking13 that it (perhaps somewhat misleadingly) suggests there is a player of the
cards distinct from the hand that is dealt, whereas in truth human beings simply are the
cards that are dealt them by genes and environment, and the cards just have to play
themselves.
12
Fischer (2006), at 129.
13
Werking (2009).

My view can be understood as accepting this, at least as being the situation when we
start out in life; but as suggesting that each of us includes, in the hand of cards that is
dealt us and that constitutes us, along with particular cards like aces, tens, jacks and so
on, one powerful and exible general-purpose card, like a joker, that can take any value
at the selection of whatever is playing the cards. The particular cards engage with cir-
cumstances and laws of nature to limit the course of the game to spectra of possibili-
ties, while the joker, our capacity for conscious decision-making, can combine exibly
with our other cards so that the whole hand can steer a course within these spectra of
possibilities.
I am not suggesting this joker is a self or soul that itself makes decisions, or that it
corresponds to any particular region of our brains: rather, it is a capacity that operates
only in conjunction with our other cards. It is however powerful and exible. So long
as our other cards are not seriously decient, for example because of mental abnor-
mality or senility or immaturity, the joker enables us to make reasonable albeit fallible
decisions about what to believe (including what to believe about right and wrong) and
about what to do, for good or ill. And these decisions in turn can (like Kanes self-
forming actions) affect what particular cards come (along with the joker) to constitute
us for the future, for better or worse. Since we are all alike in having this joker, we are
all alike in having some ultimate responsibility for what we do and for how we turn
out, again so long as our other cards are not too seriously decient.
An agent then is the hand of cards, and if one takes a view of causation that accepts
that agents can be causes, then I say there is no reason to deny that the hand of cards is
an effect as well as a cause. It thus cannot be an uncaused cause. However, because of
the exible joker, the hand of cards can cause things that are determined not by the
engagement of existing circumstances (including its own state) with laws, but by its
own selection between alternatives left open to it by existing circumstances and laws.
There are many contributing causes to this selection, so this selection is not uncaused:
it is partly constrained but not determined by existing circumstances and laws, and
otherwise is determined by the hand of cards itself in making the selection.
Thus I say that not only is agent-causation, in the sense outlined above, to be re-
jected; but that also, in its search for an uncaused cause, it is undertaking something
quite unnecessary for free will and responsibility.
COM PATI B I LI S M
As mentioned in the introduction, the view that free will and responsibility are com-
patible with determinism is advocated by many philosophers, perhaps a majority of
contemporary philosophers who have addressed the problem. It is not central to my
case in this book to dispute this view, because I am relying principally on the positive

arguments for my indeterministic version of free will, advanced in this and in previous
chapters of this book. However, I do think there are difculties with compatibilist
views, and briey discussing these difculties will I believe help clarify my own
position.
As originally proposed by seventeenth- and eighteenth-century philosophers, the
basic idea of compatibilism was that human beings have free will and responsibility so
long as they are able to do whatever it is that they want or choose to do, and thus that
free will and responsibility are excluded only when human beings are prevented from
doing what they want or choose to do by some external constraint. To the extent that
it might be thought necessary for free will and/or responsibility that a person be able
to do otherwise than what he or she actually does, it was sufcient that he or she could
and would have done otherwise if he or she had wanted or chosen to do otherwise.
Now this approach raises two questions in particular, which have given rise to much
elaboration and debate in recent times.
First, there is the question whether a person whose wants and choices are irrational,
because of mental abnormality, can correctly be considered as having free will, and as
being responsible when acting so as to give effect to irrational wants or choices. It is in
fact generally accepted that free will and responsibility require some minimum
rationality.
Second, there is the question whether a person can correctly be considered as having
free will and responsibility just because he or she can do whatever it is he or she wants
or chooses to do, if it is the case that what a person wants or chooses is determined by
circumstances outside the persons control.
In recent years, this second question has been given impetus by what is called the
consequence argument, which philosopher Peter van Inwagen states informally as
follows:14
If determinism is true, then our acts are the consequences of the laws of nature
and events in the remote past. But it is not up to us what went on before we were
born, and neither is it up to us what the laws of nature are. Therefore the conse-
quences of these things (including our present acts) are not up to us.
This argument makes use of what has been called the transfer of powerlessness princi-
ple, which in effect says that because we are powerless to change events of the remote
past and are powerless to change laws of nature, we must also be powerless to change
anything that is a necessary consequence of (and uniquely determined by) those events
14
van Inwagen (1983), 16.

and laws. Accordingly, if determinism is true, we must be powerless to do anything
other than what we in fact do; so that there are never any alternatives open to us.
And then a further step can be argued, namely that if there are never any alternatives
open to us, we cannot be responsible (in the sense of morally responsible or morally
answerable) for what we in fact do (because there never was anything else we could
have done).
One response to the consequence arguments and similar arguments is to closely
analyse what it means to say that a person has alternatives available and can do other
than he or she in fact does, and to claim that there is a meaningful sense in which a
person has alternatives available and can do other than he or she in fact does, even if
determinism is true. This has given rise to extensive and sometimes technical debate,
which is outside the scope of this book.
A second response is to say that it is not the availability of alternatives or the ability
to do otherwise that is important: rather, it is whether the person is morally responsi-
ble, and this does not stand or fall by the availability of alternatives or the ability to do
otherwise. Here the challenge is not so much to the consequence argument itself, as to
the further step of arguing from lack of alternatives to lack of responsibility.
Another way of approaching this second response is by reference to a further argu-
ment that has been the subject of considerable discussion in recent years, the direct
argument against the compatibility of determinism and responsibility (direct, because
it goes directly to responsibility without reference to the lack of alternatives). Accord-
ing to this argument, if determinism is true, our actions are the necessary conse-
quence of events in the remote past and the laws of nature; and since we are not
responsible for these events or laws, we are not responsible for their necessary conse-
quences, including our own actions. This argument builds on what has been called
the transfer of non-responsibility principle: we cannot be responsible for the necessary
consequences of things for which we are not responsible. The second response to the
consequence argument in effect attacks this principle.
One important strand in this second response to the consequence argument (and to
the challenge to the transfer of non-responsibility principle) has been an argument
originating in an article by philosopher Harry Frankfurt,15 which seeks to show that
the availability of alternatives is not necessary for free will, and particularly is not nec-
essary for responsibility. The argument uses thought experiments giving examples
where agents make decisions as to what to do, in circumstance where there is a fail-safe
mechanism that ensures that what the agent does must be in accordance with a pre-
determined scenario.
15
Frankfurt (1983).

A typical example is of a person A who, along with another B, wishes a third person
C dead. A and B plan that A will kill C. B secretly plants a device into As brain such
that, if A deviates from putting into effect the plan to kill B, the device will manipulate
As brain processes so as to cause A to kill C. As it happens, A proceeds with the plan
and kills C without any intervention from the device. It is argued that As responsibility
for putting the plan into effect and killing C is not defeated by reason of the fact that,
because of the device, he could not in fact have done anything else.
This and like examples16 are relied on as suggesting that what makes A responsible
are the actual processes that occur in his brain and not any possibility that A could have
done something different; so that what gives responsibility is As own decision-making
in accordance with As own reasons, rather than any absence of determinism.
This focus on the actual decision-making process giving rise to the action, rather
than on the existence of alternatives, bears also on the rst question identied earlier
in relation to the original formulation of compatibilism, which led to the suggestion
that for responsibility some appropriate level of rationality is required. In fact, central
to most contemporary compatibilist positions is the idea that what is necessary for
responsibility is that a person have the capacity to recognise and respond to good rea-
sons for conduct, and to control his or her conduct accordingly.
On this approach, it is not necessary that a person be ideally rational: reasonable or
adequate rationality is sufcient. A leading compatibilist philosopher, John Fischer,
has characterised what is necessary in this respect as a mechanism that has moderate-
reasons-responsiveness. And it is not necessary on this approach that the control that a
person can exercise extend to the capacity to regulate between alternatives (called by
Fischer regulative control): it is sufcient that the person have what Fischer calls guid-
ance control, so that what the person actually does is caused by the persons exercise of
his or her rational capacity.
Fischer takes the view that the argument for the incompatibility of determinism and
alternative possibilities is considerably stronger than the argument that determinism
rules out moral responsibility; and given the desirability of protecting our status as
morally responsible agents, he is inclined to adopt what he calls semi-compatibilism,
the doctrine that causal determinism is compatible with moral responsibility, even if
causal determinism were to rule out alternative possibilities.17
Before setting out my own position concerning compatibilism, I should mention
two more lines of argument that have been advanced against compatibilism and asso-
ciated views such as Fischers semi-compatibilism.
16
For other examples see Fischer (2002) 281308.
17
Fischer (2002) 3067. Fischer and Ravizza (1998) gives a strong and detailed exposition of this
approach.

1. It has been argued that there are in fact alternatives available in Frankfurt-type cases,
so that these cases do not show that the availability of alternatives is unnecessary for
free will or for responsibility. Taking the example I gave, it can be contended that A did
have an alternative available, namely that of changing his mind about killing C, and
that we consider A responsible just because he did not change his mind, as he could
have done. It is true that, if A had changed his mind, he still would have killed C be-
cause of the fail-safe device; but A would then not have been responsible, because he
had adopted the available alternative of changing his mind.
There are in turn two broad responses to this argument. One is to the effect that the
problem can be avoided by adjusting the facts, for example by proposing a fail-safe
device that detects prospective changes of mind and cuts in before there is a change of
mind by A, so that even this possible alternative is eliminated. The other is to say that
this kind of icker of freedom (as Fischer puts it)18 is insufciently robust to ground
moral responsibility, which must accordingly be based on the reasoning processes that
actually occurred rather than on the availability of such a tenuous alternative.
2. There have also been a number of arguments19 advanced that build on the idea that
if a persons conduct were to be caused by manipulation of the persons brain by an-
other person, it would be that other person (and not the manipulated person) who was
responsible for the conduct. These arguments go on to suggest that there is no differ-
ence relevant to responsibility between a persons pre-decision state caused partly by
manipulation, and a persons pre-decision state caused wholly by genes and environ-
ment. If the person is not responsible in the former case, there is no good reason for
holding that the person is responsible in the latter case.
One response to this is to attempt to limit the description of the responsibility-giving

capacity in some way, so that it does not include cases where there has been manipula-
tion. Fischer, for example, adds a requirement that the mechanism that gives a person
moderate-reasons-responsiveness must be owned by the person, this involving some
process of recognition and adoption by the person; but it can be said against this that
it is not sufcient to exclude cases of manipulation, because even this recognition and
adoption could itself be brought about by manipulation.
Another response is to say that where there has been manipulation, the manipulated
person (so long as he or she after manipulation is still exercising a capacity to recognise
18
Fischer (2002) 28890, 298303.
19
Such as the four-stage argument, the zygote argument, and so on: these arguments are helpfully
identied and discussed in Fischer (2008).

and respond to good reasons) and the person who has manipulated that persons brain
are both responsible for acts that ow from the combination of the exercise of the ca-
pacity and the manipulation.
ASS E SS M E NT OF COM PATI B I LI S M
I have very briey outlined a representative sample of arguments concerning com-

patibilism that have been raised in recent years. These arguments have given rise to
extensive discussion, with no generally accepted resolution. It is not possible for me
in this book to enter further into these debates; and as I have said, I am relying in this
book on my positive case for indeterministic free will and responsibility rather than
on challenging compatibilism. Indeed, I accept that positions such as Fischers semi-
compatibilism are plausible positions, and I do not claim to have a knock-down argu-
ment against them.
One matter favourable to compatibilism is that determinism is not the same as fatal-
ism. According to fatalism, what will be will be, no matter how much or how little we
agonise over alternative courses of action, and no matter how much or how little we
strive to achieve our goals; whereas compatibilists can accept that it is through such
things as people agonising over alternatives and striving to achieve goals that events
unfold in the world.
Compatibilists do have to say that, when we deliberate about alternative courses of
action, we are mistaken if we believe that they are truly open to us, in the sense that all
of them are truly possible, given the state of the world at the commencement of our
deliberations and the laws of nature; but they do not have to say that our deliberations
are ineffectual in bringing about the particular alternative that occurs. On the contrary,
they say that our deliberations are an important part of the causal history that brings
about the action. Similarly, compatibilists do not say that to strive to achieve our goals
is pointless, because these goals will either happen or not happen anyway. They say that
the moral struggles of a saint and the efforts of concentration of a sportsperson, for
example, are important elements in bringing about their achievements: without those
strivings, the saint may have been less virtuous and the sportsperson less successful.
There are however three general points I wish to make about compatibilism, which
are relevant to the main arguments of this book.
First, the very extensive consideration given in recent years to compatibilism ap-
pears to have arisen in part from a concern that something as important as moral
responsibility should not depend on so shaky a ground as denial of determinism.20
20
Peter Strawson called it panicky metaphysics: Strawson (1982).

However, I say that determinism is almost certainly false; and that anyone who wants
to suggest the contrary will need to come up with a good answer to the Conway/
Kochen free will theorem. This means that the compatibilism of free will and respon-
sibility with determinism, the consequence argument, and the direct argument, are
almost certainly beside the point, because almost certainly determinism is false.
Of course, this does not touch the arguments that the world is deterministic for all
practical purposes and that, in any event, the only indeterminism suggested by QM is
randomness, or the associated argument that the rationality of human cognition must
for reliability depend on processes that are deterministic for all practical purposes; and
many of the arguments concerning compatibilism are relevant to these points. How-
ever, it does suggest that the central problem of free will and responsibility is not
whether they can be reconciled with determinism, but is rather whether the best ac-
count we can give of rational human decision-making is one that supports free will
and responsibility. And in my contention this should be approached, as I have done,
with an open mind as to whether rational processes are or are not rule-determined.
Second, compatibilism depends on the idea that the capacity to recognise and re-
spond to good reasons for conduct and to control conduct accordingly (possibly with
some further requirement to exclude cases of manipulation) is sufcient for free will,
or at least for responsibility. I mentioned earlier that a chess-playing computer has this
capacity, albeit in the limited domain of chess, yet no one would suggest that such a
computer has free will or responsibility; so compatibilism requires an explanation of
what is the capacity that human beings have but chess-playing computers do not have.
I contend that the idea that the capacity to respond to reasons and to control conduct
is sufcient for free will or responsibility can be plausible only if the relevant capacity
is understood as a capacity that is exercised consciously: if we should ever nd our deci-
sions and actions being caused entirely by unconscious processes of our brains (like
the processes of a chess-playing computer), with no control being available for exercise
by our own conscious processes, we would surely regard ourselves as not being in con-
trol or responsive to reasons. In those circumstances, no matter how reasonable the
conduct might be and no matter how apparently reasonable the unconscious processes
that give rise to it might be, we would surely regard ourselves as not responsible for this
conduct. The suggestion that, consistently with determinism, we can be in control and
responding to reasons, depends for its plausibility on an assumption that this control
and response is provided consciously; but compatibilists give no account of how, con-
sistently with determinism, this is plausible or even possible.
9.5. Compatibilism depends for its plausibility on an assumption that control of conduct
and responsiveness to reasons is provided consciously, but gives no account of how, consis-
tently with determinism, this is possible.

That is, compatibilists do not address the question of what is involved in conscious re-
sponsiveness to reasons and conscious control of conduct, as distinct from non-con-
scious responsiveness and control. Were they to do so, they would come up against
issues of the kind I have discussed in earlier chapters of this book; and if they did this
without any preconception that conscious processes must be rule-determined if they
are to be rational, they should I believe see merit in my arguments in those chapters.
Third, in my contention compatibilism cannot deal in a satisfactory way with the
relationship between luck and responsibility, and particularly cannot deal in a satisfac-
tory way with questions of degrees of responsibility or mitigation of responsibility by
reason of the causal inuence of factors outside the control of the agent. I will look at
this in the next section, where I will also discuss how my approach deals with these
matters.
D OE S LUCK SWALLOW EVE RYTH I NG?
Responsibility has been challenged by Galen Strawsons luck swallows everything ar-
gument,21 which has some similarity to the direct argument referred to earlier, but
does not depend explicitly on determinism being true and has greater focus on what
are the actual causal springs of our conduct:
1. we do what we do because of the way we are, in respect of character and motiva-

tion; so
2. we cant be responsible for what we do unless we are responsible for the way we are;
and
3. we cant be responsible for the way we are when we rst make decisions in life as very
young children (that must be all down to genes and environment); so
4. we can never become responsible (through earlier decisions) for the way we are later
in life, or for what we do.
Strawson builds on this argument to suggest that all we do is a matter of luck: luck
swallows everything. Our genetic and environmental backgrounds are pure luck, so far
as we are concerned, everything about our conduct depends on the pure luck of our
genes and environment, and people cannot be responsible for what is just a matter of
luck.
My contention will be that this argument does not run against my account of free
will and responsibility, because my account gives an interpretation of step 1 of the ar-
gument such that step 2 is not supported. By contrast, a compatibilist approach asserts
21
See Strawson (1996), (1998) and (2002).

that responsibility can co-exist with an interpretation of step 1 as asserting that the way
we are (coupled with other existing circumstances and laws of nature, for none of
which we are responsible) uniquely determines what we do; and on that interpretation
of step 1, step 2 would follow if the transfer of non-responsibility principle is accepted.
I have noted that this principle is challenged by compatibilists, and while I myself nd
this principle persuasive and indeed convincing, consistently with what I said in the
previous section I will not enter in depth into the arguments that have been advanced
on this question.
However, the Strawson argument suggests another approach which I contend sup-
ports both the transfer of non-responsibility principle, and also the view that compati-
bilism cannot deal in a satisfactory way with the relationship between luck and
responsibility.
Some people are lucky in having genetic and environmental backgrounds that are
very conducive to conduct conforming to moral and legal requirements, while other
people are unlucky in having genetic and environmental backgrounds that are very
conducive to conduct conicting with those requirements. Surely, a reasonable theory
of responsibility would in general attribute less responsibility for wrong or unlawful
conduct to those whose genetic and/or environmental background strongly predis-
posed them to such conduct, than to those whose genetic and/or environmental back-
ground did not so predispose them.
Compatibilist or semi-compatibilist theories such as that of John Fischer specify re-
quirements for responsibility that would determine when responsibility is present and
when it is absent. So far as I have found, they do not go carefully into the question of
degrees of responsibility. Indeed, it has been contended22 that these theories systemati-
cally exclude, from any assessment of responsibility, consideration of the effect of hard
social conditions (such as childhood abuse, poverty, inequality and discrimination) on
human conduct; and it may also be suggested that they exclude consideration of the
effects of genetic disadvantages (such as susceptibility to anger), which diminish but
do not eliminate a reasonable capacity to recognise and respond to good reasons for
conduct and to control conduct accordingly.
One approach that compatibilists could possibly take would be to say that if the
capacity to recognise and respond to good reasons for conduct and to control con-
duct accordingly is reduced (but not eliminated) because of substantial genetic dis-
advantages and/or hard social conditions, then responsibility is also reduced. But
this would raise the question of why should responsibility be reduced because ca-
pacity is reduced by some factors outside the agents control (substantial genetic
disadvantages and hard social conditions) and not by others (other genetic and
22
Kaye (2007).

environmental factors): there would appear to be no explanation within compati-
bilist theory.
If in response to this compatibilists were to say that degrees of capacity (for what-
ever reason) are to be reected in degrees of responsibility, the question is raised as to
why on compatibilist views degrees of capacity should matter at all. Suppose A killed
C, and is considered responsible because he had the capacity to respond to good rea-
sons and control his conduct. The question then arises as to why A did not on this
occasion exercise this capacity correctly, by responding to good reasons and refrain-
ing from killing C. According to compatibilism, any motivation to kill C may be con-
sidered as operating deterministically, in which case the greater As capacity to respond
to good reasons and control conduct, the stronger must have been the motivation
required to (deterministically) cause A to exercise this capacity wrongly and to kill C.
It is difcult to see why the fact that A is subject to stronger motivation to kill C
should increase his responsibility. If on the other hand through genetic disadvantages
and hard social conditions A had reduced capacity to respond to good reasons and
control conduct, presumably he could have been caused to kill C by weaker motiva-
tion to do so; and again it is difcult to see why the fact he is subject to weaker motiva-
tion should reduce his responsibility.
As I see it, the problem for compatibilism here is that it has no place for any idea that
distinguishes factors for which the agent is responsible from factors for which the agent
is not responsible. On the other hand, on my approach the agent is not responsible for
a decision or action to the extent that it is caused by genetic and environmental factors
outside the agents control, but is responsible for it:
1. to the extent that it is caused by the agents non-rule-determined response to gestalt

experiences, and
2. to the extent that it is caused by the agents character, to the extent in turn that the
agents character has been caused by self-forming decisions or actions, and those
decisions or actions were caused by the agents non-rule-determined response to
gestalt experiences.
Thus my approach ts well with a commonsense approach of saying that people with
a reasonable capacity for decision-making and control of actions are responsible for
their wrong actions, but that this responsibility may be mitigated if there can be iden-
tied signicant factors outside their control, such as genetic disadvantages and/or
hard social conditions, which have contributed to the occurrence of these wrong ac-
tions. If a person is affected by factors such as these, it can be subjectively harder for
that person to exercise correctly the capacity to recognise and respond to good rea-
sons and to control conduct, so that failure to do so is less blameworthy (and success

in doing so more praiseworthy) than it would be for those who are not subject to
these disadvantages.
Returning now to Strawsons argument, I have said that it does not run against my
account of free will and responsibility, because my account gives an interpretation of
step 1 of the argument such that step 2 is not supported; and I should now explain why
this is so.
In relation to step 1, I accept that our decisions and actions are subject to consider-
able pre-choice constraints. We have no alternatives outside spectra of possibilities left
open by physical circumstances and physical laws. We have no experiences that can
give us consciously held reasons for choosing within these spectra apart from
experiences that arise from pre-choice circumstances and are correlated with physical
brain processes. The way these reasons feel and appeal to us, and the tendencies to act
that these and other brain processes produce, also arise from pre-choice circumstances
and are correlated with physical brain processes.
However, I say that because our decisions are made in part in response to gestalts that
do not engage with rules, we have the capacity to make decisions that are not wholly
determined by the engagement of laws of nature with our formed characters and our
circumstances. On the view I am proposing, we can (particularly in circumstances of
attention, concentration, deliberation and/or effort) make signicant choices as to
what to do, choices that are not wholly pre-determined by pre-choice circumstances
(including pre-choice states and processes of our brains) and laws of nature and/or
computational rules, but are in part determined by our responses, as whole physical
and experiential beings, to gestalt experiences that do not engage with any laws or
rules. There is nothing external to us, or for which we are not responsible, that is a suf-
cient cause of these decisions or actions; so that we can be considered, at least in part,
the ultimate source of them and ultimately responsible for them.
In this way, I contend, we do have substantial indeterministic free will and responsi-
bility, particularly in ne-tuning our actions, in making aesthetic and moral judg-
ments, in deciding what to believe when there is conicting evidence, and in deciding
what to do when there are conicting reasons. In doing so, we are limited and inu-
enced by our formed characters, to the extent that they affect the available alternatives,
the reasons and their feel and appeal, and the associated tendencies to act, but I suggest
not otherwise; so that otherwise we are responsible, indeed ultimately responsible, and
step 2 does not follow.
Thus, in response to Strawsons argument, I say:
9.6. The sense in which it is true that we do what we do because of the way we are is that
(a) the way we are plus our circumstances plus laws of nature provide alternatives, incon-
clusive reasons, and unconscious tendencies, and also the capacity to decide between the

alternatives on the basis of the reasons; and (b) what we do is what we decide in exercise
of that capacity.
9.7. That leaves us at least partly responsible for what we do, even if we were not respon-
sible for the way we are.
The constraining effect of the way we are is limited to determining alternatives, reasons
and unconscious tendencies. Subject to that, our decisions are not constrained by any
distinguishing features of the way we are (and thus are not constrained by anything
that is due to the luck of our genes and environment), and to this extent we are truly
responsible for them.
And I also contend:
9.8. We do become partly responsible for the way we are, as our decisions, for which we
are partly responsible, come to supplement the effects of genes and environment on the way
we are.
There is no doubt that we can train ourselves to have capacities and capabilities, and
that (more generally) our decisions and actions can affect our characters. Thus, while
our genes and early environment enormously affect the way we are, decisions and ac-
tions for which we are partly responsible can also do so. And this means in turn that we
can become partly responsible for the way we are, as our choices, for which we are
partly responsible, come to supplement (by way of self-forming decisions and actions)
the effects of genes and environment on the way we are.
So I say that while our different genes and different environmental backgrounds
make life a handicap event, most of us have some capacity to modify our handicaps
and, within limits, to make our own luck and to shape our own lives.
MOR E AB OUT LUCK
In addition to Galen Strawsons argument, there are other arguments to the effect that
exercises of indeterministic free will must be a matter of luck.23
The essence of these arguments is captured by the replay argument suggested by
Peter van Inwagen (2000). It is supposed that God repeatedly replays the history of
the universe, starting from a time just before a person decides whether to do action A
or action B in an exercise of indeterministic free will. Since the pre-decision circum-
stances do not uniquely determine the outcome of the decision, it must be that in
some of these replays the person will do action A, while in other replays the person
23
See for example N. Levy (2008), McCall and Lowe (2008), and Levy and McKenna (2009).

will do action B. There is nothing to distinguish the pre-decision circumstances in any
one of these replays from those in any other; that is, nothing about the circumstances,
or the persons character or motivating reasons, or anything else. Accordingly, it is
suggested, which way the decision goes in any one of these replays must be a matter
of luck. (I should mention that van Inwagen himself does not support compatibilism,
and he advances this argument not so much to dispute indeterministic free will as to
highlight the difculties of the question of free will.)
This and similar arguments challenge the possibility that, in whichever replay we are
considering, it is the persons rational choice whether to do action A or action B that
determines the outcome. In effect they assert that in all replays, the state of the universe
immediately prior to the exercise of free will is not causally sufcient to bring about
just one out of action A and action B; and that this suggests that which of them occurs
in any replay cannot be due to the agents rational choice.
The argument has some plausibility if one assumes that the transition from pre-de-
cision circumstances, when both alternatives are open, to the occurrence of one of
them, must be an instantaneous change; and this assumption can be supported by the
argument that either the alternatives are open or they are not, so the change from the
one situation (alternatives open) to the other (alternatives not open) must be instanta-
neous. On that basis it can be said that the transition can hardly be a matter of rational
choice, because rational choices cannot occur instantaneously.
Now my account of rational choice does treat it as a process that continues right up
to the time when an outcome occurs: this process does not conclude at any earlier time,
so there is no time earlier than the actual occurrence of the outcome that pre-decision
circumstances uniquely determine the outcome. However, this does not mean that
there is just an instantaneous all-or-nothing change at the time of the outcome. If it is
the case that the time selected by God for commencing the replays is a time when more
than one outcome has substantial probability, then on my approach one would expect
that this would have to be an appreciable time before the action is actually done. On
my approach, as the person proceeds from this time to make a rational choice between
action A and action B, the person will be steering within spectra of possibilities that
progressively become available as the rational process continues; and one would expect
that in this process the objective probabilities of action A and action B, on the basis of
physical circumstances and physical laws, will vary, and perhaps will uctuate; but that
as the time when the action is done approaches, these probabilities will approach one
for the action that is done and zero for the action that is not done. Certainty would not
be achieved until the action is done, but uncertainty would reduce towards being van-
ishingly small as this time approaches.
So in my contention, the circumstance that in Gods replays we may suppose that the
decision will sometimes go one way, and sometimes the other, has no bearing on the

question whether the decision is a rational decision or just a matter of luck. The replay
argument therefore does not detract from the case, made out by this and previous chap-
ters, that such decisions are indeed matters of indeterministic rational choice, and not
luck. And what my approach does, that I think no other approach has done, is to iden-
tify something specic about the process that is both indeterministic and rational,
namely the persons apposite response to feature-rich gestalts of conscious experience.
COR E ASS E RTION S AB OUT

I N DETE R M I N I STIC FR E E WI LL
9.1. Rational and non-random yet indeterministic processes are operative in the adop-
tion of beliefs relevant to action, and also in putting these beliefs into effect by deciding to
act and proceeding to do so.
9.2. We have free will in a robust sense because (1) alternatives are truly open, albeit
limited by the engagement of prior conditions (including our own physical-and-mental
state) with laws of nature, (2) the occurrence of one of the alternatives is not random but
the result of our selection between them on rational grounds, and (3) we put our selection
into effect by voluntary action.
9.3. We also have responsibility in a robust sense because (1) we make rational decisions
and put them into effect, (2) factors outside our control at the time do no more than limit
alternatives, give rise to reasons and determine how they appeal, and give rise to uncon-
scious tendencies, and (3) while those factors reduce our responsibility at the time, they do
not eliminate it.
9.4. Conscious decisions and actions are caused indeterministically by the totality of the
relevant unconscious and conscious processes, including the conscious subjects choice
taking into account consciously held reasons.
9.5. Compatibilism depends for its plausibility on an assumption that control of conduct
and responsiveness to reasons is provided consciously; but gives no account of how, consis-
tently with determinism, this is possible.
9.6. The sense in which it is true that we do what we do because of the way we are is that
(a) the way we are plus our circumstances plus laws of nature provide alternatives, incon-
clusive reasons, and unconscious tendencies, and also the capacity to decide between the
alternatives on the basis of the reasons; and (b) what we do is what we decide in exercise
of that capacity.
9.7. That leaves us at least partly responsible for what we do, even if we were not respon-
sible for the way we are.
9.8. We do become partly responsible for the way we are, as our decisions, for which we
are partly responsible, come to supplement the effects of genes and environment on the
way we are.

10
Value Judgments
It will be recalled that in the introduction, I disputed the contention that ideas of
blame and retribution for wrongdoing should be abandoned, and I suggested instead
that the idea of retribution is a foundation of human rights. My arguments so far have
supported the view that we do have free will and responsibility for our actions in a
robust sense. As a further step in my support for ideas of retribution, I now want to
develop the idea that people can be justly praised or blamed, or rewarded or punished,
for what they do.
The idea that it is just to do these things requires that what people do can truly be
right or good, and thus deserving of praise or reward, or wrong or bad, and thus deserv-
ing of blame or punishment. All this raises the question of the status of value judg-
ments about such things as justice, right and wrong. Can such judgments be true? And
if so, does their truth consist merely in conformity with some artefact of human evolu-
tion and culture (such as a surrounding climate of ideas about how to live),1 or rather
in conformity with some objective reality that is not just an artefact of human evolu-
tion and culture? It is important for me to say something about these questions, if only
because my contention is (1) that it is just to impose punishment to an extent that is
deserved by reason of wrong conduct, for which a person is responsible, and (2) that the
person in question should accept it as being just. The signicance of these assertions is
much affected by the status of the value judgments included in them.
The position I will very briey support in this chapter is that the wrongness of some
conduct does not consist solely in its relationship with some artefact of human evolu-
tion and culture, and similarly that the imposition of deserved punishment is just in a
sense that is not wholly dependent on some artefact of that kind. My contention is that
value judgments like these may be true and, consistently with what I wrote in chapter
2, that this truth consists in correspondence, by way of meaning, with reality; and that
in this case the reality is an objective reality that is not merely an artefact of human
evolution and culture.
1
See Blackburn (2001), 1.
180
A DI FFE R E NT PH I LOSOPH ICAL APPROACH
Much contemporary philosophical writing takes a very different view. There is a

vast literature dealing with questions like these, and I cannot hope to do justice to
this literature in a short chapter. What I will do is to take one example, a recent well-
received introduction to ethics that develops a view contrary to mine, and to say
briey why I disagree profoundly with it. I will then very briey outline my own
approach.
Philosopher Simon Blackburn begins his 2001 book Being Good with an assertion on
page 1 that strikes me as objectionably question-begging. It is that the moral or ethical
environment, which he identies as the surrounding climate of ideas about how to
live, determines what we nd acceptable and unacceptable and gives us our stan-
dards. That is, at the very start of his book, Blackburn simply assumes that this climate
of ideas does not merely inuence what we nd acceptable, but determines it; so there
is no place for any appeal to reasons that go beyond the surrounding climate of ideas
and that may challenge it and cause it to develop.
This assumption pervades the whole book. A basis for it is suggested in Part I of the
book, where Blackburn puts forward the further (unsupported) assumptions that
the physical world contains only is and not ought, and that even if there were a fact
that makes ethical commitments true, we could not detect any such fact.2 He fore-
shadows that a way out of this problem for ethics will be provided in Part III of the
book.
In Part III, Blackburn considers the role of reasons in ethical thinking, and raises the
possibility that there may be Reasons (with a capital letter), that is, reasons that every-
one must acknowledge to be reasons independently of their sympathies and inclina-
tions. However, that possibility is quickly dismissed on the basis of an uncritical
adoption of Humes position that reasons proper sphere is conned to mathematics
and logic, while knowledge about the way things are is due wholly to sense experience;
and that when it comes to ethics we are in the domain of preference and choice, where
reason is silent. Blackburn continues (at 112):
This could be put in terms of a contrast between description and prescription.

Reason is involved in getting our descriptions of the world right. What we then
prescribe is beyond its jurisdiction. Reason is in fact wholly at the service of the
passions. It is just because we must act in the world that we need to know about
it: Reason is and ought only to be the slave of the passions, and can never pre-
tend to any other oce than to serve and obey them.
2
Blackburn (2001), 29.
Value Judgments 181

In following sections, Blackburn looks at challenges to Humes position by Kant and by
those he calls contractarians; but he nds these challenges inadequate, and he does not
raise any objections of his own against Humes position.
Hume is revered as a philosopher and he was undoubtedly highly rational; but his
rationality, like that of all of us, was fallible, and he did not have the benet of the chal-
lenging ideas of people like Gdel, Turing and Wittgenstein (to name but three), or of
the ndings of quantum mechanics and of contemporary neuroscience and cognitive
science. I contend that his position, as expounded and apparently adopted by Black-
burn, is simply untenable.
It is perverse to seek to conne reason to mathematics and logic: as we have seen,
most of what counts as human reasoning is plausible reasoning, in which premises do
not entail conclusions, but rather support them as a matter of reasonable judgment,
and most intellectual endeavours depend on this kind of reasoning. (Do Hume and
Blackburn really claim that their own philosophical writings are not exercises of reason?
And if they do, what do they say their writings are, and why should we take any notice
of them?) Even if one were to think it appropriate (as I certainly do not) to separate out
reason from the perceptual capabilities that give us our sense experiences, the fact
remains that we experience the world as we do in part because of our expectations and
our conceptualisations of the world, both of which in turn depend in part on theories
supported by plausible reasoning.
And while the passions can be non-rational or irrational, it is surely unreasonable to
suggest they can neither contribute to nor be inuenced by human rationality; and
neither Hume nor Blackburn provides justication for adopting such a view.
Contemporary neuroscience strongly suggests that emotion is important to our ra-
tionality: it has (for example) shown that injury to part of the brain associated with
emotion, while it may not interfere with performance of logical and other intellectual
tasks, is highly detrimental to reasonable decision-making.3 It might be suggested that
this merely shows that without passions we just do not engage in any useful activity;
but in my contention this would be far too narrow a reading of the relevant ndings of
neuroscience, and in any event it is surely part of our rationality that we do proceed to
decision-making when this is reasonably required of us.
Perhaps more importantly for ethics, the arguments of chapters 5 and 6 strongly sug-
gest that our values and our desires can be affected by plausible reasoning. Hume con-
tended that nothing can oppose or retard the inuence of a passion, but a contrary
impulse; but he gave no justication for this view, and in my contention it is clear that
reason can modify the strength or operation of passions.
3
See generally Damasio (1996).

Hume limited the role that reason could have in affecting desires to determining
beliefs about facts on which desires may be based; and he gave the example of a person
desiring fruit as of an excellent relish and ceasing to do so when convinced that the
fruit was not so. This assumes both that a fruits being of excellent relish is pure fact
with no admixture of value, and that there is a clear distinction between reasoning
about facts or descriptions, on the one hand, and reasoning about values or
prescriptions, on the other, with only the former being legitimate reasoning; but Hume
does not justify either assumption and in my contention they are erroneous.
The arguments of chapters 5 and 6 suggest that there can be plausible reasoning af-
fecting an aesthetic judgment about a work of art, and that this is legitimate reasoning
about values or prescriptions; and this reasoning can affect whether or not we nd the
work of art to be of aesthetic merit and for that reason a possible object of desire. And
much the same goes for plausible reasoning about what is morally right and wrong:
surely such reasoning can affect both what we believe to be right or wrong, and also
our feelings of obligation or inclination to give effect to those beliefs. For example I am
persuaded, reasonably I believe, by arguments given later in this chapter, not to use as
my sole criterion for what I ought to do the idea that the right act is always that with
the best consequences: this surely is reasoning affecting what I believe to be right and
wrong, and affecting consequent feelings of obligation.
Hume famously contended that one cannot derive an ought from an is; but it is
important to note that this contention is valid only for formal reasoning, and does not
hold good for plausible reasoning of the kind considered in the previous paragraph.
We have seen that plausible reasoning is both reasonable and necessary in scientic
enquiry, and in my contention this is also the case in ethics.
On all these points, I contend, Hume was wrong. And these considerations, to my
mind, render highly suspect any sharp distinction between is and ought, between
description and prescription: I will elaborate on this later in this chapter. No doubt
rationalist philosophers such as Descartes, Spinoza and Leibniz had gone too far in
appealing to rationality considered as distinct from experience, but Hume went much
too far in the other direction of seeking to circumscribe reason.
Blackburn concludes his book by offering a way out of the ethical relativism sug-
gested by his rejection of Reasons: he appeals to the activity of giving and receiving
reasons for actions and for attitudes, and thereby seeking a common point of view
from which to appraise actions and attitudes. Blackburn continues (at 134):
So is there such a thing as moral knowledge? Is there moral progress? These

questions are not answered by science, or religion, or metaphysics, or logic. They
have to be answered from within our own moral perspective. Then, fortunately,
there are countless small, unpretentious things that we know with perfect
Value Judgments 183

certainty. Happiness is preferable to misery, and dignity is better than humilia-
tion. It is bad that people suer, and worse if a culture turns a blind eye to this
suering. Death is worse than life; the attempt to nd a common point of view
is better than manipulative contempt for it.
The assertion that we know things with perfect certainty from within our own moral
perspective begs the questions who are this we and why should we (or anyone) be
satised with what is known from within the moral perspective of this unidentied
group. The best answer to the rst question, implied but not explicitly given by Black-
burn, must be that we are all those presently existing people in a community who are
willing, through giving and receiving reasons, to seek a common point of view. The
content of ethics for that community must then come down to such ethical views as
may be held in common by all of these people (or perhaps a majority, or perhaps
some elite?), including their willingness to seek a common point of view through the
giving and receiving of reasons; but with no further basis available for appraising
whatever happens to appeal to them (presumably, on the basis of their genes and
culture).
Now I accept that views held about such things as right and wrong by such people
provide evidence for what is right and wrong, and are to be respected. But to say, as
Blackburn does, that these views are constitutive of what is right and wrong, seems to
me plainly mistaken.
For one thing (at least unless Blackburn were to acknowledge as a Reason an absolute
and objective prescription to seek a common point of view through the giving and
receiving of reasons, which he does not do), it would make Blackburns questions about
progress pointless and his judgments about those questions baseless: on his view there
is nothing, outside the views that happen to be held by different people at different
times in different communities, by which to judge progress.
And surely, this would not be the view of the people in question, apart from such of
them as may be persuaded by philosophical arguments of the kind Blackburn ad-
vances: surely they would hold that, in their activity of seeking a common point of
view through giving and receiving reasons, they are seeking to nd answers to ques-
tions of what is right and wrong, and not by that activity themselves creating and im-
posing otherwise non-existent answers to these questions.
Further, Blackburns view is not, I suggest, in accordance with the ordinary meaning
of words such as right and wrong, which convey an appeal to reasons or standards that
do not consist merely in the activities or views of any particular group of people, how-
ever contemporary, numerous or well-motivated that group may be. In my contention,
Blackburn gives no sound basis for adopting a view that is contrary to ordinary lan-
guage useand indeed, I suggest, contrary to his own language use when he talks of

moral progress, and also when he vigorously attacks moral views that are in conict
with his own.
Blackburns book is just a small part of the vast philosophical literature I have re-
ferred to, but it is I think a fair sample. There are writings that develop philosophical
positions contrary to mine in greater depth than does Blackburns short introduction
to ethics,4 but all of them ultimately treat values and moral prescriptions as nothing
more than artefacts of human evolution and culture, and not as having any further
objective reality or validity. As will appear from the rest of this chapter, I accept there
are good arguments for this position; but I say that it does not do justice to the univer-
sality, validity and bindingness of moral considerations, and that the arguments for it
can be answered.
NATU RAL I M PE RATIVE S
In my experience, most human beings regard some conduct as wrong and other con-
duct as right. Our ideas about rightness and wrongness raise the possibility that there
may be some features of the world of the nature of natural imperatives, which exist
objectively and call for compliance by us whether we recognise them or not, and
whether we like it or not.
There are strong arguments against the existence of such natural imperatives.
They have no place in the scientic view of the world. They cant be the object of
straightforward observation or experiment. In so far as human moral opinions and
behaviours can be observed, these opinions and behaviours can be given an explana-
tion in terms of human evolution and culture. Although most human beings regard
certain conduct as wrong and other conduct as right, there are wide and seemingly
irreconcilable differences of opinion as to what conduct is wrong and what conduct is
right, differences that can be seen as pointing strongly against the existence of universal
objective natural imperatives.
Furthermore, it can be asked, how could the existence of such imperatives bind us,
particularly when there appears to be little justice associated with any supposed natu-
ral imperatives? The fact that often the innocent suffer and the guilty prosper can be
seen as discrediting any idea that there are such imperatives, or that we need to obey
them.
And yet, after taking all these things into account, I rmly believe that at least some
conduct is wrong, in a sense that appeals to an objective reality that is beyond human
evolution and culture. This is partly because of the universality of appeals to right and
wrong (even by those who purport to deny that these can be any more than artefacts
4
A good example is Gibbard (1990); and Blackburns views are developed in Blackburn (1998).
Value Judgments 185

of human evolution and culture), partly because of more specic reasons that appear
in this chapter, and partly because I cannot accept that my belief that some conduct is
wrong is no more than such an artefact.
Consider a judgment that it is wrong to torture a child for amusement. Do those who
deny objectivity and reality to value judgments believe in their heart of hearts that such
a judgment is no more than an artefact of human evolution and culture, and has no
more bindingness on anyone than whatever he or she may happen to attribute to such
an artefact? Can such a judgment reasonably be considered as no more than a position
arrived at by persons willing to seek a common point of view through giving and re-
ceiving reasons, so that all that could be said against people who do such things and see
nothing wrong in them is that their views diverge from a surrounding climate of ideas
about how to live, which has emerged in that way among some people? I contend that
anyone with reasonable intellectual and emotional capacities who considers the matter
carefully and conscientiously will, at least unless persuaded by philosophical argu-
ments to the contrary, or perhaps strongly conditioned by a culture permissive of such
things, (1) consider a judgment that it is wrong to torture a child for amusement to be
a matter of truth and objective reality, and (2) feel a motivational obligation to refrain
from that kind of conduct (perhaps better described in this particular case as
abhorrence).
And if, as I believe, the wrongness of such conduct is a matter of truth and objective
reality, then there must be some objectively existing feature of the world, of the nature
of one or more natural imperatives, which makes some conduct wrong and not to be
done, and the appreciation of which, by persons with reasonable intellectual and emo-
tional capacities, carries with it a feeling of obligation to refrain from that conduct. Or,
to use Blackburns terminology, there must be Reasons (with a capital letter) that ev-
eryone should acknowledge to be reasons.
What then do I say about the objections I have mentioned, in particular (1) that
moral imperatives cannot be observed or made the subject of scientic investigation,
(2) that moral opinions can be explained by evolution and culture, (3) that they are the
subject of intractable disagreements, and (4) that there is no justice and no reason to
be moral?
As for objection 1, Ive already set out and supported my beliefs that rationality is
wider than science, and that we can by plausible reasoning arrive at reasonable con-
clusions that go beyond what can be observed or experimentally tested. An assertion
that the world consists of what is and therefore contains no oughts is question-
begging: why cannot some oughts be included in what is? So I believe that if moral
requirements can be supported by rational arguments, the fact that they are not di-
rectly observable or testable does not count strongly against their existence. More
generally, I believe:

10.1. Questions concerning the existence and content of natural imperatives can be the
subject of rational enquiry.
As for objection 2, the fact that there can be explanations of morality in terms
of human evolution and culture may be seen as suggesting that the mere existence of
moral opinions and behaviour is not of itself a strong argument for the existence of
true moral imperatives; but the possibility of such explanations does not itself count
strongly against other arguments for the existence of such imperatives.
As for objection 3, I do not think differences of reasonable opinion on what conduct
is wrong and what conduct is right are irreconcilable. This is partly because I contend
there can be resolution of the conict between those views of morality that say that
what is important is what consequences our actions have, and those that say that some
things are wrong (dishonesty, injustice, etc.) independently of their consequences.
I will elaborate on this later.
And as for objection 4, I do not say that moral imperatives should be obeyed just
because they somehow exist as imperatives or laws: rather, I say there are compelling
reasons (indeed, Reasons with a capital letter) why we should act in certain ways, which
may conveniently be expressed and summarised in terms of imperatives or laws. That
is, we should do the right thing not just because there is some kind of law saying we
should, but because the reasons for doing so are compelling; and are compelling even
though justice is not assured; and appreciation of these reasons, by persons with rea-
sonable intellectual and emotional capacities, carries with it a feeling of obligation to
give effect to them. This too I will revisit later.
So I will contend:
10.2. Wrongness of certain behaviour is not just an artefact of human evolution and
culture: there are natural imperatives.
I cannot in this chapter fully develop a theory of ethics and value judgments. I will give
the barest outline of my thinking on the subject, so that it can be taken into account in
considering my position concerning responsibility and retribution. I will begin by pro-
posing what I contend to be plausible candidates to be natural imperatives, and briey
giving reasons, in addition to their intrinsic plausibility, why I believe them to be so.
AB SOLUTE I M PE RATIVE S
What I suggest as the most basic imperative is the imperative to do the right thing and
not to do the wrong thing. (I will take this as including an imperative, applying where
one either does not do the right thing or does not know what it is, to do what is more
Value Judgments 187

right in preference what is less right.) So expressed, it is not necessarily a moral
imperative: it is rather an imperative to do the most appropriate thing (or more appro-
priate thing), to do what is to be done.
I suggest this imperative, or something like it, follows from the circumstance that we
can and do make decisions between alternative courses of action that we see as avail-
able to us, and often there are reasons for and against each of the alternatives; so that
we have to decide between them, having regard to these reasons. These reasons may
include basic motivational feelings such as pain and hunger, more long-term consider-
ations of self-interest, and moral feelings based on concern for other people. Now, we
may come to have general views as to how conicts between these reasons are to be
resolved, for example that generally reasons of self-interest should prevail over reasons
based on concern for others, or vice versa. However, the development and application
of such views is really part of determining what is the right thing to do, not in a speci-
cally moral sense but in the sense of determining what is to be done; and the impera-
tive of doing the right thing, in this sense, is presupposed in the whole exercise of
developing and applying these views. Whether or not the right thing is what is morally
required depends on further considerations.
So this leads me to what I suggest as the second most basic imperative, rationally
determine what is the right thing to do. This imperative, or something like it, follows
from our ability to make decisions, our rationality, and our ability to apply our ratio-
nality in deciding what to do (here of course taking the wide view of rationality ex-
plained earlier, as including appropriate attention to emotional feelings). It is important
that we determine what is the right thing to do, so we should seek to optimise the rea-
sons on the basis of which to act. This requires us to attend to relevant information and
feelings, as delivered by our senses and emotions, to explore relevant information and
feelings, to verify by checking, to look for coherence and consistency in our beliefs, to
attend to analogies, and to seek a grasp or understanding of issues facing us.
This does not mean that one should, whenever faced with alternative courses of
action, enter into an elaborate consideration of reasons for and against each alterna-
tive: to do so would often itself be the adoption of a course of action that might be
wrong. Rather, it means that one should use ones rationality, both at times when there
is opportunity for reection (when general principles and approaches might be identi-
ed, as well as particular actions decided upon) and at times when immediate action is
called for, so as to be able to determine rationally what is to be done.
It is not to be presumed in advance that the result of this determination would be to
give pre-eminence to other-regarding moral considerations rather than to prefer ones
own interests: the rational approach must be to presuppose neither one nor the other,
but to apply ones rationality as best one can. However, I believe that this rational ap-
proach supports the more contentful imperatives that I set out below, and that perhaps

the most important role for rationality arises where these imperatives conict in par-
ticular cases and it is necessary to determine which of them should prevail.
So far, then, there is little content to the natural imperatives I am suggesting, and
nothing to make them moral rather than prudential; and it will be seen that when
I later come to suggest a number of imperatives with more specic content, I acknowl-
edge that these are not absolute imperatives but are imperatives that admit of
exceptions.
However, there is one more specic imperative that I do suggest as being absolute,
and which has a avour of morality, which I express as act out of care for all persons.
A similar idea could be expressed in terms of acting out of love or respect for all per-
sons or as treating persons always as ends and never as means; and the idea is similar
to the Golden Rule in its positive or negative form (Do to others as you would have them
do to you or Do not do to others what you would not have them do to you). However, I
think care or concern is the most appropriate concept; and I would not regard either
version of the Golden Rule as universally applicable, because it ties ones conduct to-
wards others to what one would wish or not wish for oneself; and that may be different
from what others wish or do not wish for themselves and/or what is good for them.
This imperative does not require that the care be the same for all people, or the same
for others as for oneselfmerely that no persons interests affected by ones actions can
be wholly disregarded.
One reason why I suggest this as an absolute imperative is that I am convinced that
certain things are wrong just because of the way they are detrimental to other persons,
and that it does not matter which persons are affected; so that there is no limit to the
persons about whom I should care. Another reason is that I regard my own interests as
of value, to be respected by other persons as well as by myself; and I see no reason why
the same should not be true of the interests of all persons. And another again is that
this is the view of the most rational and admirable people of whom I am aware.
So I suggest:
10.3. The overriding and exceptionless natural imperatives are:

a. Do the right thing and dont do the wrong thing.
b. Rationally determine what is the right thing and what is the wrong thing.
c. Act out of care for all persons.
As I have indicated, the third of these imperatives does not require that one give equal
weight to the interests of each person, and it does not require that there never be action
contrary to the interests of other persons. It leaves open, to be determined rationally by
having regard to other considerations, that one may prefer ones own interests, or the
interests of particular persons, over the interests of other persons; and also that one
Value Judgments 189

may act adversely to the interests of other persons, and even harm them, in certain
circumstances, for example when it is just to do so.
However, it does require that one should not act adversely to the interests of others,
and in particular should not harm them, unless there is justication for doing so; so
that it absolutely rules out seriously harming others for ones own amusement. It
would not altogether rule out practical jokes in which slight harm may be done for
amusement of persons including the victim, or taking some pleasure in the just visita-
tion of harm on a wrongdoer; but otherwise, I suggest, it is surely wrong to cause harm
to others in order to derive pleasure from doing so.
PR I MA FACI E I M PE RATIVE S
I dont think any other imperatives, with more specic content, can be absolute. They
can conict, and where they do, at least one must give way; and the outcome must be
determined by application of the second of the absolute imperatives. So I believe these
imperatives are prima facie only.
Some of these imperatives apply at all times and to all our actions, so that it is always
wrong not to comply with them unless there is justication for not doing so, justica-
tion which generally must be based on other imperatives. These generally applying
prima facie imperatives may be expressed and grouped in different ways, but I think
most would be covered by the following.
10.4. The generally applying prima facie natural imperatives are:

a. Do no harm.
b. Act justly.
c. Act honestly.
d. Full commitments.
e. Act in a life-afrming way.
The rst four of these imperatives are suggested by the basic imperative to act out of
care for all persons, although of course they are not themselves absolute, because they
can conict with each other. Other similar imperatives such as act kindly, act loyally,
act cooperatively, act tolerantly, and so on, could be added; but I think they are suf-
ciently included by implication in those I have stated. For example, the imperative to
act justly requires that one make a fair contribution to any cooperative project from
which one takes benets.
The fth of these imperatives is perhaps less clear. For most people, I think, life is a
positive, something valuable and worth having, or at least is capable of being so. For
these people at least, an appropriate attitude to life is to welcome it, to embrace it, to

celebrate it, to make the most of it and (at least if there is a belief in a giver of life) to
be grateful for it. Life is a gift to be valued and appreciated. This imperative may apply
less readily, and certainly applies less strongly, to the minority for whom life is not a
plus but a burden, whether through illness or poverty or injustice or otherwise; yet
even for these persons I think there is an imperative to make the most of what they
have in life.
If this imperative is interpreted as requiring the advancing and preference of ones
own interests, and as having the capacity to prevail generally over other imperatives, it
could be regarded as potentially undermining them. However, if it is regarded, as I
believe it should be, as just one among many imperatives to which regard is to be had,
with the outcome to be reached taking into account all relevant imperatives to the best
of ones ability, other imperatives would not be undermined.
There are other prima facie imperatives that require positive action that cannot rea-
sonably be undertaken at all times, but should be undertaken on occasion when it is
reasonable to do so. I would express them as follows:
10.5. The occasional prima facie natural imperatives are:

a. Do good.
b. Improve oneself.
The rst of these is self-explanatory. It is similar to the rst of the generally applying
imperatives, but has less universal application: I do not think it is reasonable to expect
people to trying to do good all the time, although it is reasonable to expect them to be
willing to take advantage of at least some of the opportunities to do good that present
themselves from time to time.
The last imperative, Improve oneself, is suggested by the basic imperatives. It requires
one to enhance ones own ability to do the right thing, rationally determine what is the
right thing, and do good. It requires the cultivation of virtues associated with the seek-
ing of truth, particularly in so far as the truth is relevant to ones own actions; and of
virtues associated with readiness, willingness, and ability to put decisions rightly made
into effect, and also to enlarge ones opportunities to do good. Thus it requires us to try
to become, in our rational deliberations, both interested (inquisitive, enthusiastic, self-
reliant, diligent, tenacious, and so on) and disinterested (balanced, fair-minded, self-
critical, open to appreciation of our own strengths and limitations, concerned for truth
not self-promotion or self-justication, willing to admit error, duly respectful of other
opinions, and so on). And it requires us to enhance our capacity to give effect to our
decisions and to do good, for example by improving our practical skills and courage
and sympathy, combating laziness, guarding against giving undue weight to our own
interests, and so on.
Value Judgments 191

Because most of my suggested natural imperatives are prima facie only, there is con-
siderable room for disagreement as to what they require in particular circumstances.
However, I dont think there could be reasonable disagreement about the substance of
imperatives along these lines, although there could of course be reasonable disagree-
ment about their precise formulation. At one time I thought that conscientiousness
was of prime importance, and that it could not be wrong to do what one believed was
right. I now think this gives insufcient weight to imperatives such as to Rationally
determine what is the right thing and what is the wrong thing, to Act out of care for all
persons and to Do no harm. Not to care about other persons and to do harm without
justication are I suggest such moral deciencies, and are so readily disclosed as being
so by honest and unbiased consideration, that belief to the contrary does not justify or
excuse them.
NO R EASONAB LE I R R ECONCI LAB LE DI FFE R E NCE S
With those suggested natural imperatives in mind, I can return to the two questions
left for further consideration, namely whether there are irreconcilable differences be-
tween reasonable opinions concerning moral questions, and whether moral impera-
tives can really be binding.
One basic area of disagreement between apparently reasonable moral views, that
could be considered to be irreconcilable, is that between views to the effect that the
right act is always that which has the best consequences (the act-utilitarian or act-
consequentialist criterion for rightness of actions), and views to the effect that at least
some acts are right or wrong independently of their total consequences, for example
killing innocent people, being dishonest, being unjust.5 Even within views to the effect
that the right act is that with the best consequences, questions of justice raise differ-
ences that may be considered irreconcilable, notably those between views that regard
the totality of whatever is of value as being the test of what consequences are best, and
views that regard the distribution of whatever is of value among persons affected as also
being important (and in that case, there are further difculties in how distribution is
to be weighed against totality).
5
There are other views falling within the general description of utilitarianism, which appeal to the
consequences of adopting and supporting moral rules, rather than the consequences of individual acts.
These views exist in many different forms involving many complexities and many different implica-
tions (see for example Regan 1980), and generally they do not involve the stark conict with other
moral views that is raised by act-utilitarianism. I will not here address them specically: however, I note
that they are all subject to the second difculty I identify for act-utilitarianism, namely the indetermi-
nacy of hypothetical and future consequences.

The point has often been made that any such disagreements are unlikely to be of
practical signicance, because it is very difcult to decide correctly what would have
the best consequences and, since moral and legal rules are the result of long experience
and wisdom, it is more reasonable to follow them than to rely on ones own fallible
judgment in particular cases. Further, even on views that it is the totality of what is of
value that is the test of what consequences are best, a just distribution of what is of
value is generally required, because generally less is required to improve the lot of those
who are badly off than to improve to the same extent the lot of those who are already
well off. Ideas such as these were developed at length by the prominent utilitarian phi-
losopher Henry Sidgwick in his book The Methods of Ethics.6
However, the question remains whether the right act is in truth the act with the best
consequences, or whether the right act is one that complies in the best way with natural
imperatives such as those I have been considering. J. J. C. Smart has argued that, for
someone concerned that things turn out as well as possible for everyone, it must be
irrational rule-worship to follow any more particular imperative when to do so would
not have the best consequences.7 In the 1960s I was impressed by this argument, but
I could not help feeling that the wrongness of things such as dishonesty did not consist
merely in its having consequences that in totality were worse than those of at least one
alternative, or that it was just irrational rule-worship to be honest rather than dishon-
est where to be dishonest would have marginally better consequences.
What I then came to realise (and to develop in my rst book, Consequences of Utili-
tarianism)8 was that even to correctly decide what to do on the basis of the act-utilitar-
ian criterion may not necessarily have better consequences than to decide what to do
on the basis of more particular moral or legal rules. If one correctly applies the act-
utilitarian criterion, then one must do what out of the available alternatives would have
the best consequences; but these available alternatives would only be those available to
a person applying the act-utilitarian criterion. The very circumstance that a person is
using this criterion in order to decide what to do could itself have bad consequences.
For example, good consequences can result from a persons purporting to have re-
spect for particular moral and legal rules; bad consequences can result from a persons
being dishonest, for example in pretending to have such respect when one does not; and
so there can be good consequences from honestly respecting such rules and laws, con-
sequences that cannot be achieved by a person who respects only the act-utilitarian
criterion.
6
Sidgwick (1874).
7
Smart (1961).
8
Hodgson (1967).
Value Judgments 193

This position is particularly striking in relation to a persons relationships with family
and friends. Relationships with family and friends make up a very important part of
what is of value to most human beings; and honesty, loyalty, commitment and trust are
important in these relationships. Loyalty, commitment and trust involve treating
family and friends as of more concern than people in general, and honesty requires that
one be frank and open with them. The act-utilitarian criterion on the other hand re-
quires that one treat the interests of family and friends as ultimately of no more impor-
tance than those of other persons; and although it is in accordance with act-utilitarianism
that one should generally have more regard to the interests of family and friends than
to the interests of other people, because one is in a position to benet or harm the
former more than the latter, this falls short of the mutual loyalty and commitment
(and thus preference) that is essential to the best human relationships. Thus, adoption
of the act-utilitarian criterion as the single moral rule would be prejudicial to some-
thing of great value in human affairs; so that even the correct application of act-utili-
tarianism could have worse consequences than the adoption of a morality that permits
honest commitments to and preference of particular other persons.
So I believe it is plainly not irrational rule-worship, for example, to keep a promise to
a friend or family member, where it would have marginally better consequences to
break it. If one were prepared to break promises to friends or family members in those
circumstances, this would be destructive of ones relationships with them: if they knew
this, they could not condently rely on ones promises; and if they did not know this,
then ones promises would be deceptive, and to engage in deception (however
benevolently intended) would itself tend to be destructive of good relationships. And
thus the strongest argument for act-utilitarianism as the sole moral criterion fails.
Another reason why I now believe it is not reasonable to accept act-utilitarianism
(or indeed any kind of utilitarianism or consequentialism) as the sole moral criterion
is that there is irreducible indeterminacy in the consequences of alternatives that are
not chosen, and in future consequences of an alternative that is chosen. I argued in
chapter 7 that there is irreducible indeterminism in what happens, and this means
that hypothetical consequences (of alternatives not chosen) and future consequences
(of an alternative chosen) must be indeterminate. A reasonable utilitarianism would
set no limits to the consequences that are relevant to the rightness of actions, and the
indeterminacy of consequences must increase with the (hypothetical or future) pas-
sage of time. There is theoretically a quantum probability attached to each alternative
that is possible within the hypothetical and future consequences of alternative ac-
tions; but it would not be reasonable to make the rightness of actions depend solely
upon the weighted sums of these probabilities as at the time of the action (or adop-
tion of a rule). For example, if the action (or rule) actually chosen was the best on this
test, but in fact soon caused a disaster that had little probability at the time of the

action (or rule-adoption), it is surely unreasonable to say it was the right act (or rule),
justied on the utilitarian criterion.
Thus I contend that, while there are irreconcilable differences between act-utilitari-
anism and other moral views, they are not differences that can reasonably be main-
tained, because it is not reasonable to accept the act-utilitarian criterion as the sole
moral criterion.
Another possible area of disagreement is that some people assert that certain par-
ticular moral rules are exceptionless, for example, a rule against killing innocent per-
sons. However, the maintenance of such a view requires distinctions that are difcult
to justify, between those cases where a desired result itself requires the killing of an in-
nocent person (not permissible), and those cases where the desired result does not re-
quire this but the killing of an innocent person is an inevitable side-effect (perhaps
permissible). Further, scenarios can be constructed in which the only reasonable
answer is to kill an innocent person. For example, suppose there are Siamese twins, one
of whom is terminally ill and whose condition will fatally prejudice that of the other
twin unless they are separated quickly; the separation cannot be achieved without kill-
ing one of the twins; and the terminally ill twin requests that the operation be per-
formed so as to save the other twin. I believe it would be difcult to justify any other
course.
Perhaps there could be disagreement about my suggested exceptionless imperative
Act out of care for all persons, in that some might argue that we need to care only about
persons with whom we have some relationship, or that we need not care about evil
persons; and others might argue that our care should extend to animals or all sentient
beings, and not merely to persons. Certainly, there could be reasonable disagreement
about what constitutes a person, for example at what stage a human embryo becomes
a person. Concern for animals, and sentient beings in particular, is reected in the
imperatives to Do no harm and to Do good; but I think it is reasonable to maintain a
distinction between beings with the rationality and moral concerns of human beings,
on the one hand, and other sentient beings, on the other hand. And although consid-
erations of justice may justify, and even require, that harm be done to wrongdoers, I do
not believe it is reasonable to hold that we need not care about them. If an evil person
is evil by choice, then reform is possible; while if an evil person has been born evil or
otherwise inevitably caused to be that way by genes and environment, then he/she is
not responsible and is to be pitied.
In my opinion, most if not all reasonable disagreements come down to differences
concerning objective facts, and differences as to what respective weights should be
given to imperatives about which there can be no reasonable disagreement. These dis-
agreements may be important and have large practical consequences; but they can be
rationally addressed, with prospects of resolution in particular cases. So I believe:
Value Judgments 195

10.6. The available area of reasonable disagreement about natural imperatives is not
such as to count strongly against their existence.
In particular cases, natural imperatives will often give a reasonably clear answer as to
what action is right and what action is wrong; but sometimes they will not do so, and
there may remain reasonable differences of opinion on the matter. In such cases, I be-
lieve the appropriate attitude is not to condemn those with views that are different
from ones own, but to engage in constructive debate to see if agreement can be reached;
and if despite all reasonable efforts differences remain, the only reasonable position is
one of tolerance. It may be that in some particular cases, there is in fact no single cor-
rect answer.
10.7. Reasonable differences of opinion on moral issues are possible, and when they occur
the appropriate response is constructive debate and tolerance.
WHY B E MORAL?
Turning to the second question left for further consideration, whether natural im-
peratives can really be binding, I do not appeal to the existence of God as a reason
why we should be moral and obey natural imperatives. So far in my discussion, I
have not given any reason to believe in God or anything like God; and it will be seen
in chapter 12 that such belief as I have in anything like God depends in part on my
belief in morality, rather than vice versa. Similarly, I do not support morality by any
idea of ultimate justice being accorded in conformity to how well or badly we have
followed natural imperatives.
It has been said that if there is no God, then everything is permitted. I disagree pro-
foundly with that. If our moral obligations depended on Gods commands, and noth-
ing else, I would see no reason, other than prudential reasons, to obey those commands.
It could be said that, because God created us, we owe God gratitude and obedience; but
then either the obligations of gratitude and obedience must themselves depend on
Gods command, or they must be based on some imperative that does not consist en-
tirely in Gods command. If the former, then it is a bootstrap exercise; and if the latter,
then moral obligation depends on something in addition to Gods command.
Somewhat similar considerations are raised by the question of justice. If perfect jus-
tice were considered necessary for the existence of natural imperatives, these impera-
tives would depend on breaches being punished and compliance being rewarded
precisely to the appropriate extent. But then prudential and moral considerations
would coincide exactly, and the need to have moral considerations weighing along
with and sometimes against prudential considerations would disappear. Whereas

I believe that moral considerations have claims on us that may conict with prudential
considerations, and can weigh against and prevail over prudential considerations.
I do believe, however, that if there were gross and obvious injustice visited on those
who comply with natural imperatives, so that they were clearly worse off through
complying rather than not complying, this would count against the existence and
bindingness of natural imperatives. But I do not believe that there is injustice of this
nature and degree. Although it seems, for example, that some criminals prosper, and
that people amass wealth through sharp dealing and tax evasion, I am not convinced
that their lives go better than if they had acted morally; nor do I believe that the lives
of those who generally respect moral imperatives are worse than if they did not do so.
Undoubtedly many people have suffered terribly and many continue to suffer terribly
through no fault of their own; but there is little reason to think that their lot would
have been signicantly better if they had disregarded moral imperatives, or signi-
cantly worse if they had respected them better.
I see natural imperatives as binding on us not because they are commands of God or the
universe, but because of the reasons underlying them: the very reasons that support the
existence of moral imperatives are the reasons why we should obey them, and it is appre-
ciation of these reasons that lead persons with reasonable intellectual and emotional ca-
pacities to feel an obligation to give effect to them. Although proven gross injustice would
cause me to question and perhaps reject this belief, I do not think there is proven injustice
in relation to complying with or breaching these imperatives. So I believe:
10.8. We should comply with natural imperatives, not because they are commands of
God or the universe or because justice is guaranteed, but because of the reasons underlying
them.
GO OD, EVI L AN D B EAUTY
So far Ive been arguing that questions of right and wrong can be matters of truth and
reality, not merely of human evolution and culture. I now suggest that questions of
good and evil can also be matters of truth and reality. Just as some actions are right, in
the sense of to be done, and some are wrong, in the sense of not to be done, so also
some things are good, in the sense of to be approved of (and so admired, promoted,
sought after, etc.), and some things are bad (or evil), in the sense of to be disapproved
of (or abhorred). And here I mean good or bad in an absolute sense, not merely good
or bad of their kind, like a good car or a bad apple, or good or bad because of their
consequences or tendencies.
I will not attempt to give a fully worked-out account of what things are good in this
sense, or to present substantial arguments justifying the views I express here.
Value Judgments 197

However, I do believe that right actions are good in this sense, as are persons dis-
posed to do right actions, and virtues such as kindness, honesty, wisdom and justice.
I believe that unselsh love of other persons is good. I believe that human ourishing
is good, and that certain human experiences are good, generally pleasurable experi-
ences, although I do not believe pleasure is always good. (I do not believe pleasure
taken in the suffering of others is good.)
I believe wrong actions are bad in this sense, as are persons disposed to do wrong
actions. At the extreme, such actions and persons are evil. I believe that hatred of other
persons is generally bad. I believe that human suffering is generally bad.
So I suggest:
10.9. Some things are good in themselves, such as right actions, persons disposed to act
rightly, unselsh love and human ourishing; and some things are bad in themselves, such
as wrong actions, persons disposed to act wrongly, hatred and human suffering.
A further reason why I believe that values (right, wrong, good, bad) are not mere
artefacts of human evolution and culture, but have a basis in a further objective
reality, is my belief that beauty or (more generally) aesthetic merit can be more
than such an artefact. I accept that personal taste and cultural context are impor-
tant factors in aesthetic judgments; but I do not believe this detracts from the abso-
lute worth of the beauty of the natural world, and that (for example) of Rembrandts
paintings, Shakespeares plays and poetry, and Bachs music. So I believe that the
richness and beauty of life and the natural world are good, and so also are the
richness and beauty of superior aesthetic creations of human artists, writers and
musicians.
For those who are capable of creating things of high aesthetic merit, to do so con-
forms to a number of natural imperatives, including Act in a life-afrming way, Do
good, and Improve oneself. The extent to which this could justify non-compliance with
other natural imperatives would be a matter for reasonable judgment.
I believe:
10.10. Things of beauty or (more generally) aesthetic merit are good in themselves, and
for those capable of creating such things it is prima facie right to do so.
COM M U N ITY PRACTICE S AN D LAWS
So far, I have argued that there are objectively existing moral requirements, and have
set out what I suggest to be the basic moral imperatives. I suggest that these impera-
tives do not depend upon their being adopted or implemented by the practices or laws

of particular communities, but that what they actually require is greatly affected by
these practices and laws.
The basic moral imperatives are very general; and while they give clear guidance
on many issues, in some areas they do not do so, particularly in relation to the
details of living in communities and of appropriate responses to adverse conduct
by other persons. In relation to these matters, rules of conduct comprised in prac-
tices and laws of particular communities are very significant, for example in filling
out the content of the requirements referred to in paragraphs b, c, and d of my as-
sertion 10.4 (act justly, act honestly and fulfil commitments). They are also signifi-
cant in relation to appropriate responses to adverse conduct by other persons, and
in relation to conflicts that can occur between directly seeking good outcomes, on
the one hand, and complying with particular moral requirements, on the other
hand.
I now want to consider briey issues about the relationship between basic moral
requirements and the practices and laws of communities, about appropriate re-
sponses to adverse conduct by others and, particularly in relation to these matters,
about the tension between seeking best outcomes and complying with more specic
rules.
Most human beings live in communities in which there are generally accepted moral
practices, and also laws providing more structured requirements as to their conduct
and as to appropriate responses to adverse conduct by others. Issues of praise and
blame, and reward and punishment, often arise in relation to these practices and laws,
instead of or in addition to natural imperatives of the kind I have been considering;
and I should say something about whether there is any moral imperative to comply
with them, so as to make it fair that there be positive or negative responses to compli-
ance or non-compliance.
The requirements of these practices and laws do not as such automatically operate as
moral imperatives for members of these communities, in the same way as do the natu-
ral imperatives I have discussed. Where the practices and/or laws coincide with natural
imperatives, they should of course be complied with. Where they squarely conict with
natural imperatives, the natural imperatives should prevail. For example, practices of
denying education, employment and freedom of movement to women, marriage of
and consequent sexual intercourse with female children, genital mutilation, and
honour killings, conict so plainly and grossly with the imperatives to act out of care
for all people, to do no harm, and to act justly, that to give effect to them cannot be
morally justied (and, pace Blackburn, this is so even if these practices are accepted in
a community that adopts them by those persons who are willing through giving and
receiving reasons to seek a common point of view, and even if those persons include a
majority of the women in the community).
Value Judgments 199

There remain vast areas where a communitys practices and laws neither coincide
with nor conict with natural imperatives; and in those areas, I say that there are good
moral reasons for generally complying with those practices and laws.
First, those practices and laws generally reect views reached by many people as to
what conduct is appropriate in particular circumstances, and they have been tried and
tested over time; so although such views may be wrong and/or may have become un-
suited to changed circumstances, one should not arrogantly assume that this is so. One
should have respect for practices and laws that may be the result of wise choices and
long testing, and should not reject them without good reason.
Second, the practices and laws generally contribute in various ways to the stabil-
ity and successful functioning of the community for the benefit of all its members;
and it is unfair to take the benefits of the compliance of others with those practices
and laws, and not to share in the burden of compliance. We should not be
freeloaders.
Third, the practices and laws give rise to reasonable expectations on which commu-
nity members place reliance, and which ll out the requirements of justice, honesty
and commitments; and the creation and fullment of these expectations contributes
to the success of the community. It is generally harmful to disappoint such expecta-
tions, not only by causing detriment in particular cases, but also in tending to weaken
practices that contribute to cooperation.
Fourth, in so far as the successful functioning of the community requires practices
and laws of some kind, non-compliance could damage the stability and cohesiveness of
the community.
10.11. There are good moral reasons for generally complying with community practices
and laws, having to do with respect, fairness, reasonable expectations and stability.
However, there can also be good moral reasons for not complying with some commu-
nity practices, and even for not complying with some of a communitys laws, although
in the case of laws additional considerations come into play. These reasons arise in two
main areas, namely where the practice or laws are unjust or unreasonable or otherwise
decient, and (at least in the case of practices) where variety may be preferable to dull
uniformity.
Where there are deciencies in practices that are not laws, there is generally no pro-
cedure for changing the practices other than conduct and advocacy. If it is indeed the
case that the practices are unreasonable, it must follow that the rst reason for compli-
ance does not apply: but the other three could still do so, and the advantages of non-
compliance in particular cases and of promoting desirable change would need to be
weighed against these reasons.

The desirability of variety rather than dull uniformity generally arises in relation to
practices that are not reective of natural imperatives such as do no harm, be just, be
honest and full commitments: there seems to be no moral justication for departing
from these requirements in the interests of variety. However, in relation to practices
that are not reective of natural imperatives, relating to what might be called lifestyle
choices, there may be good moral reasons justifying diversity. This may be so in the
case of occupations, hobbies, cultural and spiritual practices, living arrangements,
body decoration and at least some aspects of manners.
Sexual morality has elements of both categories. It seems to follow from the natural
imperatives that a child should not be created except in circumstances conducive to the
ourishing of the child. This tends to support conventions to the effect that persons who
are not in a stable long-term relationship should not engage in sexual intercourse that
has the potential for creating a child. Natural imperatives that we should act out of care
and respect for other persons, do no harm, behave honestly and full commitments are
also relevant to sexual relationships. In other respects, details of sexual practices may be
considered to be matters of lifestyle choices, where diversity is morally justied.
10.12. The reasons for complying with community practices may be outweighed, particu-
larly in the case of practices that are unjust or unreasonable or otherwise decient, and in
the case of practices that are mere lifestyle choices.
LEGAL SYSTE M S
Where a community has a functioning legal system, which is a reasonably just and ef-
fective system, the moral reasons for general compliance are very strong.
The four reasons I identied above in relation to community practices apply with
particular force. Laws have generally been stated or enacted on the basis of considered
opinions reached by relevantly qualied people as to their justice and appropriateness.
The contribution of such a legal system, and of general compliance with it, to the suc-
cessful functioning of a community is very great indeed; and the unfairness of taking
the benet of compliance by others, without complying oneself, is substantial. Most
members of the community order their lives in the reasonable expectation that laws
will generally be complied with, and thus are susceptible to signicant harm when
these expectations are not fullled. A great deal of cooperation and coordination of
effort that is very benecial to the community depends on the fullment of these rea-
sonable expectations. And the weakening of a legal system through signicant non-
compliance has the potential for very great harm.
There is also the further consideration that most reasonably just and effective legal
systems provide mechanisms for changes to be made to particular laws; so that where
Value Judgments 201

laws are unjust or unreasonable or otherwise decient, steps can be taken to change
them otherwise than by non-compliance and advocacy of non-compliance.
10.13. If there is an established legal system that is reasonably just and effective, there are
strong moral reasons for obeying the laws of that system.
But this is not to say that there can never be moral justication for not complying
with some laws of such a legal system.
Some breaches of some laws may be considered de minimis, for example where a
pedestrian disobeys a trafc light when it is clear that there is no trafc anywhere in the
vicinity and no possibility that anyone could be put at risk.
More signicantly, there may be circumstances in which it is morally justied to break
a law in a way that is not de minimis, on the ground that this particular law is unjust or
otherwise bad and that deance of it is an appropriate course to have it changed. Because
there are legal ways to change laws, this course would generally be morally permissible
only if other reasonable attempts to have the law changed have failed; and because the
objective is to have the law changed, the breach of it should be open and peaceful.
However, I believe this course would not be morally permissible for persons who
have undertaken to uphold the law, such as a judge. For such people, in a system that is
reasonably just, the importance of compliance with the law is such that, deliberate
breaches, other than de minimis breaches, could be morally justied only in extreme
circumstances.
10.14. If there is a law in a reasonably just legal system that is unjust or otherwise bad,
change should be sought within the system, but there may be circumstances in which it is
morally permissible to seek such a change by openly breaching the law.
Generally, I believe it is morally permissible to disobey the laws of such a system only
where the moral reasons supporting that course are strong, for example where great
harm would be caused by complying with the law in particular circumstances.
10.15. Apart from cases of de minimis breaches, it is morally permissible to disobey the
laws of a reasonably just legal system only if there are strong moral reasons to do so.
If there is a legal system in place that is not reasonably just, the moral reasons for com-
plying with it are less strong and are more readily displaced. If the system is sufciently
unjust, moral reasons may support attempts to overthrow the system and replace it;
and it may even be morally wrong for people such as judges to participate in the system
at all. I cannot here go deeply into these issues.

10.16. A legal system may be so unjust that it is morally permissible to seek to overthrow
it, and morally wrong to participate in it.
CAPACITY FOR R EASONAB LE VALU E J U D GM E NTS
I have argued there truly are moral imperatives breach of which can justly deserve
blame or punishment. It might be said, however, that I have assumed that people who
may be subjected to blame or punishment are aware of what is morally required of
them: it cannot be just to blame them or punish them for not complying with require-
ments of which they are unaware.
Australian law takes the robust approach that ignorance of the law does not excuse
breaches of it; although the law also provides a defence of insanity if an accused person
proves that, by reason of a disease of the mind, he or she did not know that what he or
she was doing was wrong.
I contend this is a reasonable approach. Most adult human beings have the capacity
to make reasonable value judgments, so that they have the capacity both to apprehend
the existence of moral obligations and to feel appropriate motivation to conform to
them; and they can properly be regarded as at moral fault if they break laws because
they do not appreciate that it is wrong, for example, to harm or cheat other people
without strong justication, or to take the benet of compliance by other people with
the laws of a community while not giving others the benet of ones own compliance.
And the last point makes it reasonable to expect that people accepting the advantages
of living in a community will take reasonable steps to ascertain what is required of
them by the laws of that community.
It may be said in addition that even those who (unreasonably) dont accept there are
moral obligations binding on themselves are, in a reasonably just legal system, given a
fair opportunity to conform their conduct to the requirements of the law: the laws are
made public, as are the possible consequences of not complying with them, and pun-
ishment is applied only to those who are proved to have breached the public laws by
their voluntary actions.
There are however people who, because of mental abnormality, are not capable of
appreciating moral and legal requirements. The law is justied in dealing humanely
with those people, as well as those who are unable to appreciate what they are doing or
to control their conduct, so as to prevent them being an unreasonable danger to them-
selves or others, without regard to questions of responsibility and desert.
10.17. Most adult human beings have the capacity to make reasonable judgments
about what actions are right or wrong, and failure to do so is not an excuse for wrong
actions.
Value Judgments 203

COR E ASS E RTION S AB OUT VALU E J U D GM E NTS
10.1. Questions concerning the existence and content of natural imperatives can be the
subject of rational enquiry.
10.2. Wrongness of certain behaviour is not just an artefact of human evolution and
culture: there are natural imperatives.
10.3. The overriding and exceptionless natural imperatives are:
a. Do the right thing and dont do the wrong thing.
b. Rationally determine what is the right thing and what is the wrong thing.
c. Act out of care for all persons.
10.4. The generally applying prima facie natural imperatives are:
a. Do no harm.
b. Act justly.
c. Act honestly.
d. Full commitments.
e. Act in a life-afrming way.
10.5. The occasional prima facie natural imperatives are:
a. Do good.
b. Improve oneself.
10.6. The available area of reasonable disagreement about natural imperatives is not
such as to count strongly against their existence.
10.7. Reasonable differences of opinion on moral issues are possible, and when they occur
the appropriate response is constructive debate and tolerance.
10.8. We should comply with natural imperatives, not because they are commands of
God or the universe or because justice is guaranteed, but because of the reasons underlying
them.
10.9. Some things are good in themselves, such as right actions, persons disposed to
act rightly, unselsh love and human ourishing; and some things are bad in them-
selves, such as wrong actions, persons disposed to act wrongly, hatred and human
suffering.
10.10. Things of beauty or (more generally) aesthetic merit are good in themselves, and
for those capable of creating such things it is prima facie right to do so.
10.11. There are good moral reasons for generally complying with community practices
and laws, having to do with respect, fairness, reasonable expectations and stability.
10.12. The reasons for complying with community practices may be outweighed, particu-
larly in the case of practices that are unjust or unreasonable or otherwise decient, and in
the case of practices that are mere lifestyle choices.
10.13. If there is an established legal system that is reasonably just and effective, there are
strong moral reasons for obeying the laws of that system.

10.14. If there is a law in a reasonably just legal system that is unjust or otherwise bad,
change should be sought within the system, but there may be circumstances in which it is
morally permissible to seek such a change by openly breaching the law.
10.15. Apart from cases of de minimis breaches, it is morally permissible to disobey the
laws of a reasonably just legal system only if there are strong moral reasons to do so.
10.16. A legal system may be so unjust that it is morally permissible to seek to overthrow
it, and morally wrong to participate in it.
10.17. Most adult human beings have the capacity to make reasonable judgments about
what actions are right or wrong, and failure to do so is not an excuse for wrong actions.
Value Judgments 205

11
Responsibility and Retribution
In previous chapters of this book, I have argued for a particular account of free will and
responsibility, challenging the view that what we do is the inexorable playing out of
roles given to us by our genes and environment, and supporting the view that our vol-
untary actions are in a fundamental way our own free choices and our own responsibil-
ity. In this chapter, I will be considering what these arguments suggest as to appropriate
responses to wrongful or harmful conduct; and I will focus on what are the appropriate
responses of a legal system, and in particular a system of criminal law, to conduct of that
kind. My main concern will be the question of retribution in punishment.
As with a number of other topics, I do not propose to enter in depth into the exten-
sive literature dealing with topics such as the justication of criminal punishment: my
objective is rather to consider the implications in this area of my arguments concern-
ing free will and responsibility.
R E S PON S E S TO WRONG CON DUCT
I have argued in previous chapters for the views that most adult human beings do, in
a substantial sense, have free will and responsibility for their actions; that they have
the capacity to make reasonable judgments about what conduct is right and what
conduct is wrong; that such judgments concern matters of truth and reality, not
mere artefacts of human evolution and culture; and that failure to make reasonable
judgments about what actions are right or wrong is not an excuse for wrong actions;
and accordingly I contend that it is generally reasonable and fair to hold people re-
sponsible for their wrong conductso that responses such as blame, criticism, ex-
pectation of apology and/or remorse, and so on, may be justied. And if the
206
wrongdoer is oneself, it may be appropriate that one experience feelings of guilt and
that one apologise and express remorse to persons affected, as well as resolving to do
better in future.
When the wrongdoing is serious, and particularly when some substantial harm is
done by it, more substantial adverse responses may be appropriate. Four broad kinds
of possible responses to conduct that is seriously wrong and/or harmful may be identi-
ed, namely revenge, punishment, reparation and non-punitive treatment.
1. Revenge is essentially the visiting of harm, normally by a harmed person or associ-

ates of that person, upon the perpetrator of that harm or associates of that person. The
concept is similar to that of vengeance or retaliation. The focus is on the harm initially
done, and the visitation of harm in return, for harms sake, so as to balance or recom-
pense that initial harm. Although the targeting of the person or persons who are to
suffer harm in the revenge, and the amount of harm inicted, may be affected by con-
siderations of actual responsibility for the harm that has given rise to the revenge-taking,
and of degrees of responsibility for that harm, this is not necessarily the case. For ex-
ample, revenge for a wrongdoers killing of a child may consist in the killing of the
wrongdoers child. The focus is on the harm initially done, and on somehow repaying
that harm.
2. Punishment is different from revenge in important respects. It does involve the
visiting of harm or detriment on a wrongdoer, but it is generally imposed by a person
or institution that has some authority to impose punishment, not necessarily or even
usually being the person detrimentally affected by the wrongful conduct, or associates
of that person. Punishment is generally considered as a kind of retribution for the
wrongdoing, and the harm or detriment imposed in punishment is often related to
what the wrongdoer is considered to deserve, having regard to the wrongness of the
conduct and the degree of the wrongdoers responsibility for it; although, as we will
see, it is sometimes argued that punishment is justied solely by its benecial
consequences.
3. Reparation looks to the harm suffered as a result of the wrongdoing, and in-
volves action by the wrongdoer to compensate the victim for that harm, either at the
initiative of the wrongdoer or under some compulsion from others. Questions of re-
sponsibility and consequent desert may be relevant, but are less so than in the case of
retribution: the focus is on adequate compensation for harm actually suffered by the
victim.
4. Non-punitive treatment of the wrongdoer involves steps taken with a view to en-
suring that there is no further wrongdoing, which may involve action taken to reform
or rehabilitate the wrongdoer and may involve physical restraint of the wrongdoer. The
undertaking of treatment may be voluntary; or else treatment may be imposed as a
Responsibility and Retribution 207

matter of compulsion, in which case the central difference from punishment is that
treatment is imposed not as deserved retribution, but purely on the basis of expected
benets in avoiding further wrongdoing. Questions of responsibility and desert are
relevant only if and to the extent that they inform what treatment may be expected to
be effective.
In relation to these kinds of response, different considerations may arise according to

whether or not there is an effective legal system applying to the wrongdoing. What
I want to consider is the appropriate response of an effective legal system to conduct
that it is seriously wrong and/or harmful; and in particular, what are the implications
for this of my approach to free will and responsibility.
I suggest it is clear that an effective legal system should not tolerate, much less en-
courage, the taking of revenge.
I accept that where there is no effective legal system, some private forceful response
to serious wrongdoing may be justied, particularly if done with a view to obtaining
reparation for harm actually done, but possibly also if done with a view to imposing
some detriment that, as in the case of punishment as retribution, is appropriately re-
lated to the wrongdoers desert and responsibility. However, even where there is no
effective legal system, I suggest that the taking of revenge in the sense of visiting harm,
for harms sake, to recompense for the harm done by the wrongdoer and without
regard for what the wrongdoer truly deserves, is unjustied. Certainly, for example, the
killing of a wrongdoers child as revenge for the wrongdoers killing of the revenge-
takers child could never be justied.
Where there is an effective legal system, any private forceful obtaining of reparation or
forceful imposition of detriment on a wrongdoer would generally itself be unlawful con-
duct, unless exempted; and there are strong reasons why there should not be any exemption
in the case of revenge. It is enough here to say that it would promote lawlessness, and pro-
duce a situation where the strongest prevail, irrespective of the justice of their position.
Punishment, on the other hand, may be an appropriate response by an effective legal
system. Certainly, on the basis of my approach to free will and responsibility, it may be
reasonable and fair to hold people responsible for wrong actions and therefore deserv-
ing of appropriate punishment by way of retribution. Punishment by way of retribu-
tion may also be supported by compatibilist and other approaches. I will consider this
in more detail later.
The distinction between revenge and retributive punishment is sometimes blurred,
in particular when the popular media (and sometimes victims, with the encourage-
ment of the popular media) express views to the effect that punishment cannot be
just unless the wrongdoer is made to suffer no less than the victim, and that the detri-
ment imposed on the wrongdoer is to be taken as the measure of what the judge (or

the legal system) treated the harm to the victim as being worth. In my contention, the
extent of the victims suffering can be relevant to what the wrongdoer deserves by way
of retributive punishment, but can never be determinative of it; and other factors
relevant to the degree of the wrongdoers responsibility may mean that fair retribu-
tion involves the imposition of either less or more harm on the wrongdoer than was
suffered by the victim.
Provision for reparation may also be an appropriate response of an effective legal
system. But whereas punishment is generally initiated by the State itself, the legal sys-
tems provision for reparation generally leaves it to the victim to initiate and pursue the
remedies that the system provides; so that reparation is generally the subject of civil
rather than criminal law.
Even in relation to civil remedies providing for reparation, questions of what is fair
having regard to the responsibility of the wrongdoer can arise. Particularly this is so
because liability to make reparation can arise not merely when harm is caused by in-
tentional wrong conduct, but also when harm is caused by conduct that is merely neg-
ligent, that is, (in Australian and similar law) by conduct that falls short of standards of
reasonable skill and care. While the emphasis is on compensating persons who have
suffered from the wrong action, rather than imposing a detriment on the wrongdoer,
reparation can in fact impose detriment on the wrongdoer that is nancially devastat-
ing, if the wrongdoer is not insured against the liability in question. Where the wrong-
doing is intentional, it may be considered fair that the wrongdoer rather than the
victim suffer the total adverse consequences of the wrong. Where the wrongdoing is
merely negligent, the fairness of this may be less clear. Since my focus in this chapter is
on criminal law, I will not consider this further here, except to make two points:
1. in Australia and in most countries with similar legal systems, for most types of cases
where substantial harm can be caused by negligent conduct, insurance is either
compulsory or at least common; and
2. negligence can generally be considered a wrong for which a person is responsible,
because exercise of care is under our control, as is the undertaking of tasks for which
we do not have appropriate skill (so that moral responsibility is wholly absent only
in cases where there is no lack of care but only lack of skill, and also no lack of care
in the persons judgment that he or she did have adequate skill for the task being
undertaken).
Provision for non-punitive treatment might also be an appropriate response of a legal

system. No substantial question of justication arises in relation to treatment under-
taken voluntarily by the wrongdoer. However, if treatment is imposed as a matter of
compulsion, then justication is necessary; and my contention is that as a general rule

compulsory treatment is justied only (1) if it is imposed within the parameters of
punishment that is justied as retribution, or (2) if it is imposed on persons who are
not responsible because they do not have a reasonable capacity to recognise and re-
spond to good reasons for conduct and to control their conduct accordingly.
It will be recalled that I mentioned in the introduction that some people argue that
retributive criminal punishment is inhumane, based on primitive and uncivilised
urges for revenge, and should be abandoned; and that criminal conduct should prop-
erly be regarded as an illness to be treated rather than wrongdoing to be punished. The
main focus of this chapter will be to combat that assertion, in the light of the argu-
ments of previous chapters, and to contend it is important to maintain ideas of re-
sponsibility and retribution in the criminal law.
I will begin with an overview of my position. I will then set out in more detail the
extent to which criminal law in Australia and other countries gives effect to retributive
principles, and then I will return to why I say that retributive principles should be
maintained, having regard to the arguments of previous chapters.
OVE RVI EW
The criminal justice system in my country Australia, and also in other countries
with similar legal systems, including the United Kingdom and the United States of
America, is generally regarded as serving two broad types of purposes, which are
together considered as justifying the imposition of restraints or other detriments
on offenders: retributive, backward-looking purposes, and consequentialist, for-
ward-looking purposes. These purposes guide the development of the criminal law,
and inform decisions as to when to impose such restraints or detriments and what
they should be.
The former (retributive) purposes are based on the idea that a person who has acted
criminally deserves to be punished for this conduct, and that it is just that appropriate
detriment be inicted on that person. And the idea that a person deserves punishment
for criminal conduct presupposes that the person is truly responsible for it, and is not
deprived of that responsibility because the conduct was the inevitable outcome of
things outside the persons control, such as genes and environment.
The latter (consequentialist) purposes involve no such ideas. They simply look to the
good consequences that imposition of detriment on offenders may be expected to have,
notably:
1. demonstration that certain types of behaviour are unacceptable, and deterrence of

the offender and of others from engaging in that behaviour;
2. restraint of the offender from further crime during incarceration;

3. reform of the offender;
4. placating victims and perhaps compensating them (although the latter may be con-
sidered a matter for civil rather than criminal law); and
5. reassurance of the community that they are protected and that offenders will be
punished (promoting condence in security and in the rule of law, and discourag-
ing self-help).
So:
11.1. The criminal justice system serves two broad types of purposes: retributive, back-
ward-looking purposes, and consequentialist, forward-looking purposes.
11.2. Retributive purposes are based on the idea that a person who has acted criminally
deserves to be punished for this conduct; and consequentialist purposes look to the good
consequences that imposition of detriment on offenders may be expected to have.
It will be seen that the consequentialist purposes to some extent support the response
I have classied as non-punitive treatment, as well as the response I have classied as
punishment: in particular, restraint of the offender and reform of the offender, consid-
ered in isolation, could be undertaken as non-punitive treatment. However, the conse-
quentialist advantages of restraint and reform are not, in Australia and in countries with
similar systems, generally considered as sufcient to justify compulsory treatment to
those ends, except as incidental to punishment that is justied as retribution. In general,
the imposition of punishment and/or compulsory treatment in individual cases is con-
sidered to be appropriate only if it is justied as fair retribution for criminal conduct.
There is, as I have said, a school of thought that retributive purposes of the criminal
law should be abandoned or at least de-emphasised, which has recently been given
some impetus by developments in neuroscience.1 There are two main strands to this
contention.
1. First, it is argued that the real responsibility for criminal conduct lies in the genes
and circumstances of the criminal, and free will and responsibility are illusions; and
that accordingly it cannot be just to impose punishment on the basis that a person
deserves it by reason of responsibility for criminal conduct. On this approach, crime
is an illness to be treated rather than wrongdoing deserving punishment, and retri-
bution can be no more justiable than revenge.
2. Second, it is contended that the application of retributive ideas subjects people to harsh
and inhumane treatment, which is not justied by any good consequences that might
1
See for example Greene and Cohen (2004), and material on Thomas Clarks website http://www.
naturalism.org/.

be expected. Although some harsh treatment might possibly be justied by consider-
ations of deterrence, harsh treatment imposed as retribution is not limited by such
considerations; and certainly that kind of harsh treatment cannot be justied by some-
thing as illusory, or at least dubious, as the existence of free will and responsibility.
Some who support these views contend that punishment should be abolished alto-
gether, while others, probably the majority, contend that punishment should be strictly
as justied by consequentialist considerations alone.
The position I support in this chapter is that there are good reasons for retaining
retribution as a guiding purpose of criminal law, both as a general basis for determin-
ing when detriment is to be imposed on citizens and as an important and limiting
factor in determining what detriment is to be imposed.2 These reasons include what
I contend is the fairness and reasonableness of retribution, having regard to the earlier
chapters of this book, as well as the following considerations.
1. As my discussion of the application of retributive principles will show, their most

signicant role is to legitimise, and quantify an upper limit to, the application of
State compulsion to offenders; and far from being inhumane, it is humane to refrain
from imposing detriments on persons who have not been proved to have voluntarily
breached a public law, and to impose on persons who have been proved to have
breached a public law detriments that are no greater than are considered in some
way proportionate to the criminality of the offences committed.
2. In Western societies, it is widely regarded as legitimate for governments to impose
sanctions on people who have acted voluntarily in breach of public laws and, with
limited exceptions, not legitimate for governments to impose sanctions or other loss
of liberty on people who are innocent of doing this: to impose sanctions or loss of
liberty (or compulsory treatment) on people who are not at fault in this regard,
without powerful justication, is considered a gross violation of human rights.
3. The need to prove voluntary conduct in breach of a public law before imposing detri-
ment on persons is a necessary restraint on the conduct of the State and its ofcials. If the
imposition of detriment were to be considered as justied whenever doing so has good
consequences, whether or not the person concerned has done something to deserve it,
there is no clear basis in principle inhibiting the State and its ofcials from arrest and
detention whenever they claim this has benecial consequences for the community.
4. Retribution that is proportionate to the criminality of an offence need not be con-
sidered as justifying treatment that is harsh in a way that does not contribute to
deterrence, and it is consistent with pursuing rehabilitation.
2
See Hodgson (2000).

5. This retributive approach has further indirect advantages, in particular:
a. if punishment is not generally limited to those who deserve it, no one could feel
secure that their compliance with the law would generally ensure they were not
subjected to loss of liberty or conscation of property by the State; and
b. if punishment is not made dependent on wrongdoing and proportional to the
seriousness of the wrongdoing, the law would be less respected and resort to self-
help and revenge may not be kept in check.
6. To have regard only to the consequences of whatever is done to citizens, without
regard to whether they deserve it, is to deal with them not as responsible persons
but as vehicles for treatment to be manipulated for the general good. That is not
appropriate in relation to people who are capable of acting rationally; and it does
not encourage people to take responsibility for their conduct.
It will be seen that these considerations include both consequentialist considerations

and considerations of justice and associated notions of human rights; and some who
oppose retributive principles may say that these consequentialist considerations can be
served by a broad consequentialist approach, without any appeal to retributive prin-
ciples. I will return later to the question of whether this can be done without at least an
appearance of application of retributive principles.
The relevance of earlier chapters to my position here is threefold. First, they sup-
port the view that good sense can be made of responsibility (including degrees of
responsibility), desert, and just retribution, providing a foundation for the appli-
cation of retributive principles. Second, they support the view that the application
of retributive principles is a matter of justice, not mere expediency (even though,
as I have said, some of the above considerations do go to the expediency of appli-
cation of retributive principles). And third, in combination with these consider-
ations of expediency, they undercut the contention that something as dubious as
the existence of free will and responsibility should not be relied on to justify re-
tributive punishmentbecause so long as good sense can be made of retributive
principles and they are not shown to be unreasonable, their application has sub-
stantial advantages.
AUSTRALIAN CR I M I NAL LAW
I turn now to the way the criminal law of Australia, and other countries with similar
legal systems, gives effect to ideas of responsibility and retribution.
In the rst place, this can be seen (1) in the general doctrine of mens rea or guilty
mind, (2) in the mental element involved in the very denition of many offences, and
(3) in various defences based on factors affecting mens rea.

A key principle of our criminal law is the principle of mens rea: generally, no person
can be convicted of a crime unless the prosecution proves beyond reasonable doubt
not only a guilty act, but also a guilty mind. For criminal responsibility, generally the
action in breach of the law must be conscious and voluntary; and there may be addi-
tional requirements, such as that there be belief or lack of belief as to certain matters,
intention of a particular consequence, or recklessness as to this consequence. In the case
of murder, for example, generally the act causing the death of the victim must be done
with intention to kill or inict really serious bodily injury, or with reckless indifference
to human life. Because of this requirement of a guilty mind, a person will generally not
be guilty of any crime if the act in breach of the law occurs independently of the will or
by accident; or if the person mistakenly believes the facts to be such that the act would
not have been in breach of the law.
Then there are defences which, while not negating the voluntariness of the act in
question, are regarded as negating or mitigating the guilty mind and thus criminal
responsibility. Duress (in cases other than homicide) can negate mens rea if the act was
done under the inuence of a threat of death or really serious injury. Self-defence can
negate mens rea if the act was done in self-defence, in the belief that what was done was
necessary having regard to the threat. Necessity can negate mens rea if the act was
reasonably done to avoid irreparable harm, in the belief that what was done was neces-
sary having regard to the danger.
Responsibility can also be challenged by evidence of mental abnormality. There may
be evidence of insanity within the rules established by MNaghtens Case,3 namely that
by reason of a defect of reason, from disease of the mind the person did not know
what he or she was doing or did not know it was wrong. If this defence is established,
the person concerned may be detained for the safety of the community: because re-
sponsibility is excluded by this kind of mental abnormality, the person has to be treated
paternalistically, with a view to achieving best outcomes for the person and the com-
munity. There may sometimes be evidence that for some other reason the action was
not conscious and voluntary, the so-called defence of sane automatism;4 and if this
defence succeeds, generally the person is released, because the mental abnormality in
such cases is regarded as involving only a temporary and non-dangerous exclusion of
responsibility.
Evidence of mental abnormality that does not exclude responsibility in either of
these ways, but is considered merely to diminish responsibility, is generally not rele-
vant to whether a person is guilty of a crime, although in some States of Australia a
nding of substantially diminished responsibility can reduce murder to manslaugh-
3
(1843) 8 ER 718.
4
Discussed by the High Court of Australia in R v Falconer (1990) 171 CLR 30.

ter. Otherwise, it can be relevant only to the amount of punishment that is
imposed.
The second main way in which Australian criminal law gives effect to ideas of re-
sponsibility and retribution in is relation to the punishment of offenders, if guilt
beyond reasonable doubt is established.
There is an overriding principle that punishment for an offence should not be more
than what is considered proportionate to the offence itself. In particular, a term of
imprisonment for an offence should not be increased beyond what is considered pro-
portionate to the offence, in order to extend the period of protection of society from
the risk of further crimes by the offender.5 The extent of guilt or criminality is an
important element in determining proportionality. The case of Veen v the Queen
[No 2]6 stated that in deciding what is proportionate, the need to protect society and
the antecedent criminal history of the offender may be taken into account along with
the gravity of the offence, but not so as to result in the imposition of a penalty dispro-
portionate to the gravity of the offence. In no case can the maximum penalty pre-
scribed by law for the offence be exceeded, and this maximum should be reserved for
cases falling within the worst category for that offence.
Then, particular factors mitigating personal responsibility may justify the imposition of
a penalty less than that considered proportionate to the gravity of the offence. Thus, ge-
netic and congenital disadvantages, early environmental difculties such as child abuse,
and more immediate matters such as pressing social, psychological, or nancial problems,
may be taken into account so as to justify a reduced sentence. However, some of these fac-
tors may not work unequivocally in reducing the penalty. For example, if genetic or envi-
ronmental factors have resulted in a mental abnormality which, while not eliminating
responsibility, makes the offender more prone to violence, this may diminish moral cul-
pability and thus point towards a shorter sentence of imprisonment; but at the same time
point towards a longer sentence because of considerations of deterrence (greater penalties
may be thought necessary to deter such persons from offending) and protection of society
from the risk of further offending.7 Thus the court does not look for a perfect match be-
tween desert and punishment: so long as an offender was responsible to some extent for
his or her conduct, and had a real choice as to whether or not to commit an offence, it is
reasonable that protection of society be considered as justifying a penalty limited only by
what is considered proportionate to the criminality of the offending conduct.
(It must also be acknowledged that in any event perfection in assessment of degrees of
responsibility of this kind is impossible. On my approach, a person is ultimately respon-
sible for an action to the extent that it is caused by his or her non-rule-determined
5
Veen v the Queen [No 1] (1979) 143 CLR 458.
6
(1988) 164 CLR 465.
7
Veen [No 2] at 47677.

response to gestalt experiences and to the extent that it is caused by the persons character,
to the extent in turn that the persons character has been caused by self-forming deci-
sions and actions and those decisions or actions were caused by his or her non-rule-
determined responses to gestalt experiences. To apportion responsibility accurately on
that basis, as between genes and environment on the one hand, and the person on the
other, would require little short of omniscience. Thus, while retribution that was accu-
rately aligned with ultimate responsibility and desert might perhaps be considered as
something good in itself, to be pursued irrespective of consequences, our capacity to
accurately assess ultimate responsibility and desert is so limited that I do not propose
that role for retribution: rather, I propose it as something that can be applied in a broad-
brush way, so as to limit and inform what is to be done, as a matter of fairness, in a
system that is also partly justied by its pursuit of consequentialist goals.)
Finally, I note that in Australian law, the circumstance that a long time has elapsed since
a person committed a crime does not do away with the responsibility of that person for that
crime. Even if the person has since reformed and lived an exemplary life and made valuable
contributions to society, it is still open to the State to prosecute the person and, if the person
is found guilty, for the court to impose punishment proportionate to the offence. Punish-
ment may not be supported by considerations of deterrence of the offender and reform of
the offender, but other retributive and consequentialist factors would still apply. However,
if the crime is not a serious one, the lapse of time and the subsequent conduct of the person
may make it reasonable for prosecuting authorities to take no action; and if the person is
prosecuted and convicted, these same factors may justify imposing a lesser punishment
than the maximum justied by the criminality of the offence.
Although it is outside the scope of this book to enter in any depth into the question
of what is required for personal identity over time, I suggest it is reasonable to regard a
person as remaining responsible for what the person has done for so long as the person
lives. This view is conrmed by the consideration that conscious decision-making is
often an extended process, which can last weeks or months or years; so that the impor-
tance I give to conscious decision-making is consistent with a view that it is relevantly
the same agent that is active throughout a process that can extend over a lifetime.
R ETR I B UTION AS A R E STR ICTION ON STATE

COM PU LS ION
It will be seen from the above discussion that ideas of responsibility and retribution
operate as restrictions on the application of State compulsion to citizens: they dene
circumstances in which the application of State compulsion is permissible (namely,
when guilt of an offence, including mens rea, is established beyond reasonable doubt),
and also set limits to that compulsion.

This is the general position, to which there are exceptions. Before returning to the
question of attacks on the application of retributive principles, and the relevance of
the arguments of this book, I will outline the main exceptions, because they illustrate
the strong justication required to displace the restriction of retributive principles.
I have mentioned that the result of a successful defence of insanity is that the person
concerned may be detained for the safety of the community: because responsibility is
excluded by this kind of mental abnormality, the person is treated paternalistically,
with a view to achieving best outcomes for the person and the community. The same
thing may happen independently of the criminal law, if a person is shown by reason of
mental illness to be a danger to himself or herself or others. This possibility is subject
to close controls. It is another example of the paternalistic compulsory application of
non-punitive treatment, to persons who are shown not to have the capacity for respon-
sible conduct and to be dangerous for that reason.
Another example of a paternalistic approach which is considered justiable arises in rela-
tion to carriers of dangerous diseases: in such cases, the danger to the community may be
so great and so clear that the quarantining of carriers, even against their will, is acceptable.
A further area where State compulsion is applied outside the limitations I have dis-
cussed is that of arrest and detention with a view to determining whether a person has
been guilty of an offence. Where prosecution authorities believe on reasonable grounds
that a person has acted in breach of the law and that it is necessary to arrest and detain
the person so that a trial can take place, then coercion can be applied to that extent, so
long as an early opportunity is afforded to the person to apply to a court for release on
bail. This exception is a necessary accompaniment to any system of criminal law, and
detention in these circumstances is considered justiable if and to the extent that is
necessary to ensure that a fair trial can take place (for example, to ensure that the ac-
cused faces trial, and to prevent interference with witnesses).
Next, in some cases, where carelessness can cause great harm, punishment may be
allowed for carelessness (for example, negligent driving); and in some cases, where
great harm can be caused in circumstances where intention or carelessness is difcult
to prove and where harm can and should be prevented by particular people taking ap-
propriate precautions, strict liability may be imposed, so that the prosecution does not
have to prove mens rea. This may be the case, for example, in matters relating to worker
safety, public health and protection of the environment. Generally, punishment for
offences in these categories are pecuniary penalties rather than imprisonment.
Finally, perhaps the most signicant exception is one that has arisen recently in some
Australian jurisdictions, namely the extended detention of persons convicted of seri-
ous sex offences, who have completed their sentences for their crimes but, having not
successfully completed sex offender programs during their sentences, are considered at
high risk of re-offending. I would argue that this procedure should be kept within

strict limits, being applied only to serious sex offenders and where the risk of
re-offending is clearly proved to be high. Even then, it be could be argued to be a denial
of human rights, amounting to punishment for an offence going beyond that consid-
ered proportionate to the criminality of that offence; although against this, it could be
argued that extended detention is not unfair in the case of someone who has already
committed a serious crime and has not taken the opportunity, provided by the sen-
tence for that crime, to address the risk factors that contributed to its commission.
Thus I say:
11.3. The criminal law proceeds on the basis that most adult human beings for most of
the time have the capacity (a) to make reasonable judgments about what actions are right
and what actions are wrong, and (b) to act in accordance with those decisions.
11.4. Those that do have this capacity are treated as being responsible for their voluntary
actions in breach of the criminal law, justifying punishment up to an extent considered
proportionate to the criminality of what they have done.
11.5. Those that do not have this capacity are treated paternalistically, with a view to
achieving best outcomes for them and the community.
11.6. The capacity to make reasonable judgments, and to act in accordance with them,
varies extensively in degree within the group that has this capacity; and also circumstances
in which people nd themselves can make it more or less difcult for them to exercise this
capacity and to conform to the requirements of the law.
11.7. This variability does not exclude responsibility justifying retributive punishment,
where there is voluntary breach of the law by persons judged to have the capacity referred
to above, but can properly be taken into account in determining the degree of blamewor-
thiness for illegal conduct, and thus in determining the extent of punishment.
11.8. Otherwise, the application of State compulsion to citizens requires strong
justication.
WHY R ETR I B UTION S HOU LD B E MAI NTAI N E D
As mentioned earlier, some who oppose retribution contend that punishment should
be abolished altogether, while others, probably the majority, contend that punishment
should be strictly as justied by consequentialist considerations alone.
An example of the former is the philosopher Derk Pereboom. Pereboom8 contends
that we do not have free will of the kind required for moral responsibility, and that
accordingly criminal punishment cannot be justied as retribution for wrongdoing.
8
Pereboom (2001), 17486; (2002), 47981.

He gives reasons for not accepting that punishment can be justied as moral education
or on the basis of deterrence, and concludes that the appropriate response to crime is
similar to that accorded to carriers of dangerous diseases. Just as we have the right to
quarantine carriers of severe communicable diseases to protect society, we also have
the right to quarantine a person whose criminal activity has shown that he or she is a
comparable danger to society, and to respond in less intrusive ways (such as monitor-
ing) to those whose criminal activity does not justify detention. The quarantining or
other response should be no greater and no more harsh or intrusive than is necessary
to protect society having regard to the danger posed by the criminal, and should be
combined with attempts to rehabilitate the criminal and thereby minimise the need for
continued quarantining or monitoring.
This gives rise to the question of how the law would support such a system. Since the
rationale for intervention is the protection of society from a danger that justies inter-
vention, the approach most consistent with Perebooms basic position would be to
empower State ofcials to quarantine or otherwise limit or monitor the activities of
people whenever it is established that they are a danger to society, to the extent that the
ofcials consider necessary having regard to the danger.
I suggest that a society with that system would be a nightmare of insecurity and un-
certainty, as well as injustice. Leaving aside for the moment the question of injustice,
which Pereboom may reject in this context, the imposition of quarantining or moni-
toring (whatever that meant and however it was to be put into effect) would not even
be conditioned on a persons having actually done anything, so that people could not
order their lives in such a way as to be assured of freedom from this kind of interven-
tion. In the absence of any requirement for intervention other than established danger,
or any basis for determination of the nature and extent of the intervention other than
what was needed to protect society, there would be no principled and transparent basis
on which the State ofcials could determine what was to be done, and on which the
appropriateness of their decisions could be reviewed.
Suppose for example it was established that a person posed a danger of breaking into
houses and stealing property from them. Would this justify quarantining the person,
and if so for how long? Should it be for some denite period, or should it be indenite,
until some other ofcial determined that there was no longer a danger that justied
quarantining? No doubt expert evidence would be available to assist ofcials in making
these decisions; but even with the assistance of expert evidence, there would be no
principled, transparent and reviewable basis on which such a decision could be made,
particularly a decision as to the period of any detention, when the gravity of what (if
anything) had actually been done by the person would be relevant only to assessing the
danger, and the only question would be what response is justied to protect society
from that danger.

More generally, I suggest it would be grossly unjust for the law to authorise the sub-
jection to coercive treatment of a person (having the capacity for reasonable decision-
making and control of conduct) otherwise than through application of due process
directed at the apprehension and coercive treatment of persons who have committed
crimes by voluntarily breaching publicly known laws, or the subjection of such a
person to any greater coercive treatment than is justied by reference to what the
person has done. I suggest it would be grossly unjust for a person to be kept in custody
merely because State ofcials think the person to be detrimental to the good of society,
where this person has done nothing against any clearly stated law. Disregard of this
principle has led to such abuses as the connement of dissidents in mental institutions
in the former USSR. Even if it were said that Perebooms denial of responsibility re-
moves the basis for any consideration of justice, it would plainly be undesirable that no
one could feel secure that compliance with the law would ensure he or she was not
subjected to loss of liberty by the State.
Having regard to these considerations, it is clear (even on consequentialist grounds
alone) that on Perebooms approach, the law should, in its application to people who
have the capacity for reasonable decision-making and control of conduct, authorise
State ofcials to quarantine or otherwise limit or monitor the activities of those people
only if they have actually engaged in certain conduct dened by the law; and indeed
Pereboom himself accepts this.9 This would go some way towards enabling people to
order their lives so as to be assured of freedom from this kind of intervention, although
it would not do so adequately unless the law also required that the conduct was en-
gaged in voluntarily, and that the other requirements of mens rea were satised. Fur-
thermore, there would remain the problem of the lack of any principled, transparent
and reviewable basis on which ofcials could make their decisions as to the nature and
extent of any intervention, unless the law went further and linked the nature and extent
of any intervention in a principled way to what the person had actually done.
Thus, many of the features of a retributive system would be needed, so that compul-
sory treatment would be imposed within parameters justied by reference to what the
person had done; and in my contention, the only reason then for not categorising the
treatment as punishment would be an absolute commitment to the idea that there is
no such thing as responsibility and desert. And of course, having regard to the argu-
ments of this book, my contention is that such a position is incorrect.
Turning to the other view that, while punishment should be maintained, it must be
based purely on consequentialist principles, two approaches are possible. One would
be to say that each application of punishment must itself be justied in terms of its
consequences, and the other would be to say that the system of punishment should be
9
Pereboom (2001), 17677.

so justied. My contention is that the former approach is unreasonable, while the latter
approach supports the application of retributive principles, at least so long as good
sense can be made of ideas of responsibility and desert.
The former approach would be vulnerable to the same points as Perebooms approach
unless the law, in its application to people who have the capacity for reasonable deci-
sion-making and control of conduct, (1) authorised the application of punishment only
where a person had actually engaged in conduct dened by the law and the require-
ments for mens rea were satised, and (2) related the extent of punishment in a prin-
cipled way to what the person had done. So again, many of the features of a retributive
system would be needed. In addition, this approach would be vulnerable to arguments
to the effect that general deterrence and reassurance of the community would be
achieved so long as the person punished appeared to be guilty; and if an innocent person
was punished (whether by mistake or deliberately), this could be considered a bad thing
only if the overall consequences were worseinjustice as such would not count for
anything. Thus, if the punishing of an innocent person happened to have better conse-
quences than the exoneration of that person, because the harm to the individual was
outweighed by the general deterrent effect and the placating of the community, the
consequentialist purposes of this act of punishment would be served. However, I sug-
gest that this would be an injustice that most people would nd wholly unacceptable.
A more reasonable consequentialist approach would look to consequentialist justi-
cation of the system of punishment.
As we have seen, there are powerful consequentialist reasons for having a system that
so far as possible ensures that only those who are truly guilty are punished, proportion-
ately to their guilt, in particular to promote condence in the justice system and the
assurance that law-abiding citizens are not punished; and to limit punishment to those
who need and can benet from deterrence and reform. It is those people who commit
crimes unaffected by insanity or other disabling mental conditions who may be de-
terred by punishment and threats of punishment, and whose rehabilitation may be as-
sisted by appropriate punishment; while those who act dangerously because of insanity
or other disabling mental conditions are most usefully regarded not as criminals but as
persons who may need to be restrained and treated to protect the community.
Thus again many features of a retributive system would be adopted in a system with
consequentialist justication. However, it could be claimed that the system would not
be retributive, because offenders would not be regarded as deserving of punishment,
and because there would be no harsh treatment not justied by good consequences;10
and that elimination of retributive ideas would enable proper attention to be given to
things that really matter, namely what are the genetic or environmental causes of crime
10
See Greene and Cohen (2004), 1783.

and how are criminals best treated to free them from those causes and to ensure they
do not commit crimes in the future.
As I have indicated earlier, I do not advance retributive principles as justifying the
imposition of harsh treatment that cannot be related to the pursuit of consequential-
ist goals; and in particular, I do not regard them as justifying the imposition of harsh
treatment that could not reasonably be expected either to contribute to deterrence of
the offender or others or to advance the reform and rehabilitation of the offender.
I have suggested that retribution accurately aligned with ultimate responsibility and
desert might be considered as something good in itself, to be pursued irrespective of
consequences; but I say that human beings can do no more than make broad approxi-
mate assessments of responsibility and desert, which can support judgments as to
what is fair and permissible in pursuit of consequentialist goals, but are not suf-
ciently reliable and accurate to support harsh treatment entirely unrelated to the pur-
suit of such goals. And I fully support attention being given to the genetic and
environmental causes of crime, and to how criminals are best treated to free them
from those causes. In these respects, I do not disagree with this broad consequentialist
approach.
What I do disagree with is the assertion that it would be an advantage that offenders
not be regarded as deserving of punishment. In my contention, the characterisation of
punishment as something that is deserved, together with the linking of its nature and
extent with the criminality of what was done, contributes signicantly to the effective-
ness of a system of punishment, even if this is assessed on a purely consequentialist
basis.
I suggest that it encourages people to address their conduct rationally and to control
their conduct accordingly, if they are treated as responsible, and as deserving to be left
alone if they conform to the law and as deserving to be punished if they voluntarily
break the law. I suggest that conformity with the law is supported if it is regarded as
something that is morally right, and to be done for that reason, not just for reasons of
prudence and avoidance of detriment, and if deliberately breaking the law is regarded
as wrong and deserving of punishment, again not merely something to be avoided for
prudential reasons.
I suggest also that the support of a system of criminal law by the community is en-
hanced if it is seen as dealing fairly with people, according to what their conduct de-
serves: victims and their associates will more readily be placated and discouraged from
self-help, and also, I suggest, offenders will more readily be rehabilitated if they recog-
nise that their conduct was their own responsibility and that the punishment they re-
ceive is deserved and fair and not something to be resented.
I have already suggested that it would be most undesirable to authorise State of-
cials to address on a purely consequentialist basis the question of how much punish-

ment is appropriate for various kinds of conduct, and that for that reason the law
should link the nature and extent of any intervention in a principled way to what the
person to be punished had actually done. I contend that ideas of retribution and
desert are required if there is to be a principled, transparent and reviewable basis on
which to determine what is the appropriate punishment for various types of conduct,
and in particular what nature and degree of detriment is appropriate as a deterrent to
the conduct in question.
Accordingly, I suggest that once it is accepted, as I do accept, that ideas of retribution
and desert do not justify the imposition of harsh treatment that could not reasonably
be expected either to contribute to deterrence or to advance the reform and rehabilita-
tion of the offender, there are strong consequentialist advantages to the retention of
these ideas in the criminal law.
11.9. There are good reasons for retaining retribution as a guiding purpose of criminal
law, both as a general basis for determining when detriment is to be imposed on citizens
and as an important factor in determining what detriment is to be imposed.
PH I LOSOPH ICAL BAS E S FOR R ETR I B UTION
The consequentialist advantages of maintaining ideas of retribution and desert in

criminal law have led some writers to argue that, even though there is no such thing
as free will or responsibility for conduct, we should maintain the illusion that there
is.11 For my part, I dont think that retribution could be justied if there is no free will
or responsibility for conduct, and I dont think its a good idea to act on the basis of a
pretence. It is best always to try to ascertain the truth, and to guide our conduct and
our practices on the basis of what we reasonably believe. Accordingly I would not sup-
port this approach, to the extent that it advocates maintenance of an illusion.
A more common approach is to rely on compatibilist thinking, and to argue that,
although the world is deterministic for all practical purposes, nevertheless free will and
responsibility exist and are compatible with this determinism. One prominent legal
theorist, Stephen J. Morse,12 argues that responsibility is explained by our capacity to
grasp and be guided by good reasons, and that this is so despite the truth of determin-
ism. Another, Michael Moore,13 argues that the persistence and ubiquity of retributive
impulses are signs of the moral reality that retribution is a good thing and a primary
aim of the law; and he supports the view that it is sufcient for culpability and respon-
11
For example, Smilansky (2002a).
12
Morse (2000).
13
Moore (1997).

sibility that a person have the capacity to act rationally and to respond to reasons; and
that a person is excused if sufciently compromised in his or her rational capacity or
coerced to act against his or her wishes.
Consistently with what I said in the chapter 9, I accept that compatibilism is a defen-
sible view. Certainly, even if the world is deterministic for all practical purposes, there
still would be an important distinction between those persons who are rational in their
behaviour, and those who by reason of mental abnormality are not rational or not fully
rational; and this distinction supports retributive ideas, in that those who are not ra-
tional are less likely to be deterred by the threat of punishment, and are more appro-
priately dealt with by treatment, and if necessary connement, than by punishment.
And I would accept Stephen Morses advocacy of the capacity to grasp and be guided by
good reasons as supporting our responsibility concepts and practices, in a way that can be
adopted without necessarily having to deny determinism. A person who has that capacity
may be considered to have a fair opportunity to conform his or her behaviour to the re-
quirements of the criminal law,14 and thus fairly to be subjected to punishment if he or she
does not do so. This approach is also consistent with that adopted by Nancey Murphy and
Warren Brown,15 suggesting that the downward causation of a person acting for reasons
can support attribution of responsibility, whether or not determinism is true.
However, consistently with my discussion in chapter 9, I have concerns about this
approach. I think the commonsense notion of capacity to grasp and be guided by
reasons depends heavily on the assumption that this capacity is exercised by con-
scious decision-making in which conclusions are reached on the basis of consciously
grasped reasons that may be incommensurable and inconclusive, and which thus call
for conscious resolution.16 Determinism, and particularly a deterministic under-
standing of neuroscience, strongly challenges this assumption, suggesting the real de-
cisions are by rule-determined processes that dont require consciousness, thus
tending to undermine this approach. Further, this compatibilism remains vulnerable
to Galen Strawsons argument considered above: if we and all our actions are ulti-
mately and completely the products of our genes and environment, how can anything
we do be other than the inevitable result of things outside our control? And compati-
bilism has difculty in accounting in a systematic way for degrees of responsibility.
11.10. The capacity to grasp and be guided by good reasons is a reasonable basis for ret-
ribution and justice, but this capacity depends on conscious decision-making not readily
explicable by compatibilist views.
14
This approach was advocated by the jurist Herbert Hart: see Lacey (2007), 237.
15
Murphy and Brown (2007).
16
See Hodgson (2005).

Thus, while I see compatibilist approaches as supporting the retention of retribution,
I contend that a reasonable and principled application of retributive principles is more
strongly and reliably supported by the indeterministic version of free will and respon-
sibility presented in earlier chapters of this book.
TH E FUTU R E OF R ETR I B UTION
I do not suggest that the use of retributive principles by Australian and similar legal
systems is ideal and cannot be improved. I expect that science (and in particular neu-
roscience) will make signicant contributions to the development of the criminal law.
However, I do suggest that this should not be to the prejudice of the application of
retributive principles in punishment.
It is to be expected that increasing knowledge of how the brain works will assist in
devising programs for rehabilitation of offenders and programs for addressing genetic
and environmental factors contributing to criminal conduct, and will contribute to the
development and implementation of educational strategies to discourage criminal con-
duct. However, I contend that central to both rehabilitation and education is belief in
and acceptance of personal responsibility for conduct, and recognition and acceptance
of appropriate moral standards of right and wrong. If science were permitted to under-
mine belief in and acceptance of personal responsibility for conduct and/or belief in the
importance of moral considerations, then this would be highly counterproductive.
To the extent that science assists in identifying brain conditions that involve particu-
lar risks of criminal behaviour, and in devising methods to minimise these risks, this
could properly be taken into account in determining what sentence to pass on a person
who has committed a criminal offence (up to the limit considered proportionate to the
criminality of the offence), and in guiding the treatment of that person during the
period of that sentence. And it could properly provide a basis for offering voluntary
treatment to persons who have not committed an offence. However, for reasons I have
given, I would oppose its being used to justify detention and/or compulsory treatment
of persons who have not yet committed any criminal offence, other than (subject to
strict limitations) persons shown through mental abnormality to lack capacity to rec-
ognise and respond to good reasons for conduct and to control their conduct accord-
ingly, and also shown for that reason to be a danger to themselves or others.
One currently operating example of what I see as an appropriate interplay of re-
tributive principles and science is the use of special procedures in the case of property
offences prompted by drug addiction. Addiction is not generally regarded as elimi-
nating responsibility for crimes undertaken to satisfy a craving for drugs; and accord-
ingly, I suggest that crimes committed by reason of addiction should properly be
treated as wrongdoings to be punished, in accordance with principles I have discussed.

Punishment is deserved and thus is legitimate, and an appropriate upper limit to
punishment can be set having regard to the criminality of the conduct. However, ad-
diction may be considered as providing motivation for criminal conduct that is very
strong and hard to resist, and so may mitigate responsibility and thereby suggest lesser
punishment; and if the offender demonstrates willingness to undertake and complete
measures supported by science to overcome addiction, the prospects for rehabilita-
tion may be good, and the need for restraint and personal deterrence may not be
great. There are procedures in Australia that in certain circumstances permit an of-
fender before sentencing to undertake effective measures to deal with addiction; and
if the offender completes such measures and demonstrates a reasonable probability of
remaining drug-free, then the factors supporting reduced punishment may operate
very strongly indeed. I agree with these procedures; but I contend that the recognition
of criminal responsibility, together with the possibility of commensurate punish-
ment, is still appropriate, both to promote acceptance of responsibility by the of-
fender himself or herself, and to provide an incentive to the offender to carry the
treatment program through to completion.
Science may also be expected to inuence the development of the law concerning the
attribution of criminal responsibility, particularly in the case of those affected by
mental abnormalities. However, there is an underlying difculty to applying neurosci-
ence in this area. The categories used by the criminal law are not those of neuroscience,
but rather are pre-scientic commonsense categories, which have their source in what
has been disparagingly called folk psychology. Categories such as willed or voluntary
action, belief, and intention (of alleged offenders), and consent (of alleged victims), are
often central to determining criminal liability; and they are not matters that are sus-
ceptible to scientic proof or even description. Even when what is in issue is a question
of mental abnormality, the categories used by the law are non-scientic categories such
as disease of the mind, knowing what one is doing, knowing that what one is doing is
wrong, and substantially diminished responsibility.
These categories used by the law presuppose an active conscious agent who is
generally responsible for conduct; whereas neuroscience focuses on brain func-
tions in terms of physical causes and effects, and seeks to provide explanations of
things and processes that contribute to the functioning or malfunctioning of the
brain. And there is not at present any accepted theoretical framework that links
and reconciles these two approaches. Further, if I am right in my contention that it
is desirable that notions of personal responsibility and retribution be maintained
by the criminal law, then it is to be hoped that the criminal law will continue to use
categories that give effect to these notions; that is, categories which, like those
presently used, presuppose active conscious agents generally responsible for their
actions.

Reconciliation of these folk-psychological categories with the categories of neuro-
science is unlikely to be achieved any time soon. Thus I would expect that the contri-
bution of neuroscience in this area will for some time be limited to giving more
detailed and accurate accounts of relevant aspects of brain functioning, which can
then be used in commonsense reasoning to arrive at conclusions that may guide
future developments of the law. In particular, I would hope and expect that the law
will retain the presumption that persons of sufcient maturity are generally respon-
sible for their conduct; and I would also expect that, if this presumption is to be re-
butted by scientic evidence in particular cases, then this evidence will need to identify
some brain abnormality and the effects of that abnormality, on the basis of which a
commonsense conclusion can be reached as to whether and if so to what extent re-
sponsibility is excluded.
However, the combination of neuroscience and common sense may enable the law to
provide a more systematic approach to questions of this kind.
The present category of insanity is presumably intended to capture those cases where
responsibility for conduct is effectively absent due to a brain abnormality that is of
indenite duration and/or difcult to treat; so that, if that brain abnormality is such as
to make the person dangerous, there will be the option available of indenite detention
in order to protect the public. Neuroscience may well assist in achieving a better deni-
tion of these cases than is provided by the MNaghten rules.
The present category of sane automatism is presumably intended to capture those
cases where responsibility for conduct is absent either without there being any brain
abnormality, or else because of a brain abnormality that is temporary and/or easy to
treat and/or not such as to make the person a danger; so that it is reasonable to
excuse the person altogether without there being any need for extensive future re-
straint or even monitoring. Again, neuroscience may assist in better dening these
cases.
Where the law recognises that murder should be reduced to manslaughter in cases
where there is substantially diminished responsibility, neuroscience may contribute to
a better denition of when such diminished responsibility is to be recognised. Again,
in this area also it might be appropriate to distinguish between those cases where the
abnormality that causes this diminished responsibility is of indenite duration and/or
difcult to treat, and those cases where it is temporary and/or easy to treat and/or not
such as to make the person a danger; and neuroscience may assist in providing rules to
distinguish these cases. This approach may have some application to those cases where
addiction contributes to criminal conduct.
Another area where neuroscience may contribute to better legal rules concerns the
criminal liability of children and young people. As mentioned earlier, in Australia a
child under ten cannot be criminally liable, while a child between ten and fourteen can

be criminally liable if the prosecution proves that he or she knew that the conduct in
question was seriously wrong. Children over the age of fourteen are subject to the same
presumption of responsibility as adults, although until they are eighteen they are dealt
with by separate courts and, if custody is considered appropriate, imprisoned in differ-
ent institutions than adults. Neuroscience may contribute to a more systematic ap-
proach to the criminal responsibility of children and young people.
So my nal assertion on retribution is:
11.11. Science can be expected to make signicant contributions to the development of the
criminal law, but it should not do so in such a way as to put an end to the application of
COR E ASS E RTION S AB OUT R E S PON S I B I LITY

AN D R ETR I B UTION
11.1. The criminal justice system serves two broad types of purposes: retributive, back-
ward-looking purposes, and consequentialist, forward-looking purposes.
11.2. Retributive purposes are based on the idea that a person who has acted crimi-
nally deserves to be punished for this conduct; and consequentialist purposes look to
the good consequences that imposition of detriment on offenders may be expected to
have.
11.3. The criminal law proceeds on the basis that most adult human beings for most of
the time have the capacity (a) to make reasonable judgments about what actions are right
and what actions are wrong, and (b) to act in accordance with those decisions.
11.4. Those that do have this capacity are treated as being responsible for their voluntary
actions in breach of the criminal law, justifying punishment up to an extent considered
proportionate to the criminality of what they have done.
11.5. Those that do not have this capacity are treated paternalistically, with a view to
achieving best outcomes for them and the community.
11.6. The capacity to make reasonable judgments, and to act in accordance with them,
varies extensively in degree within the group that has this capacity; and also circumstances
in which people nd themselves can make it more or less difcult for them to exercise this
capacity and to conform to the requirements of the law.
11.7. This variability does not exclude responsibility justifying retributive punishment,
where there is voluntary breach of the law by persons judged to have the capacity referred
to above, but can properly be taken into account in determining the degree of blamewor-
thiness for illegal conduct, and thus in determining the extent of punishment.

11.8. Otherwise, the application of State compulsion to citizens requires strong
justication.
11.9. There are good reasons for retaining retribution as a guiding purpose of criminal
law, both as a general basis for determining when detriment is to be imposed on citizens
and as an important factor in determining what detriment is to be imposed.
11.10. The capacity to grasp and be guided by good reasons is a reasonable basis for ret-
ribution and justice, but this capacity depends on conscious decision-making not readily
explicable by compatibilist views.
11.11. Science can be expected to make signicant contributions to the development of the
criminal law, but it should not do so in such a way as to put an end to the application of

12
The Big Picture
I have now completed the main argument of this book. It supports an original concep-
tion of the interplay of rationality, consciousness and free will, which happens to be
very much in accord with common sense, and which is consistent with and coheres
well with what science has actually established about the world. However, this concep-
tion is very different from what is widely thought to be the scientic view of the world,
which treats the world as changing precisely as determined by laws of nature, or as
determined by some combination of these laws and chance; and it does have wider
implications for an overall picture of the workings and signicance of the world. In
particular, it is relevant to whether one takes what might be classied as a physicalist
view of the world, and also to what is to be made of religious thinking and attitudes.
In this nal chapter, I want to explore these implications. In many respects, what I say
is highly speculative; but I think it is important to show that there are rational alterna-
tives to physicalist and mechanistic views of the world, on the one hand, and non-
rational religious views (and especially non-rational fundamentalist views) on the
other. So I believe a brief exploration of these issues, in the light of the arguments of
this book, is an appropriate nale to it.
TH E SCI E NTI FIC AC COU NT
Science has developed quite a comprehensive account of our origins, beginning with
the Big Bang about thirteen billion years ago. This account tells of the evolution of the
universe from that beginning and of the emergence and evolution of life, in terms of
the matter and energy and laws of nature that have existed since the Big Bang.
Very briey, an account is given of how the simplest atoms, atoms of hydrogen, were
formed, how irregularities in the distribution of matter coalesced into stars, how progres-
sively more complex atoms came to be cooked by nuclear processes in the stars, and how
planets including our Earth came to be formed. Then, on our Earth about four billion years
ago, simple life-forms emerged; and from these life-forms evolution produced, through
random mutations and selection on the basis of success in proliferation of genes, myriad
life-forms including those now existing on our planet. We are one of those life-forms.
230
The account science gives of this history in terms of matter and energy and laws of
nature, and in terms of natural selection for proliferation of genes, is extremely de-
tailed, persuasive and comprehensive; and so far as it goes and subject to what I say
later about potentialities, I believe it is substantially correct. However, this account has
the limitations noted in earlier chapters, that it gives no account of consciousness,
values or free will; and of course, it does not explain its own raw materials of the Big
Bang, matter, energy and laws of nature.
12.1. The scientic account of our origins, in terms of matter and energy and laws of
nature, is substantially correct, so far as it goes; but it gives no account of consciousness,
values and free will, and it cannot explain its own raw materials.
Further developments of the scientic account can be expected to provide more ac-
curate formulations of laws of nature, further explanations of constraints concerning
what exists and how it can change, and fuller accounts of how life emerged from non-
life and how evolution proceeded. While these developments may give deeper explana-
tions of the Big Bang and of matter, energy and laws of nature, I do not expect they will
ever completely overcome the limitations I have identied.
Even if scientists nd and formulate a single law that accurately expresses a basic
law of nature that underlies and explains all other laws of nature, and which con-
strains what things can exist and how they can change, this would still leave unex-
plained why there is this law and why there is anything on which this law operates.
More importantly, it would not explain consciousness, values and free will: con-
sciousness and free will because, in my belief, they involve more than matter and/or
energy and/or anything else existing and changing as determined by laws of nature,
and values because they cannot be investigated by the scientic method of conjec-
ture and refutation.
AN EXPE R I E NCE D U N IVE R S E
In chapter 7, I referred to a feature of quantum mechanics (QM) that I called observer-

participation. At its simplest level, this reects the idea that the ultimate constituents of
matter do not have determinate position or motion unless and until they are measured
as being in a particular place or moving in a particular way; but there are also reason-
able views according to which measurement itself requires participation of a conscious
observer, with the consequence that everything that exists outside conscious experi-
ences is best considered in terms of potentialities for causing conscious experiences of
various kinds, rather than in terms of things, states of affairs and/or processes that exist
otherwise than as potentialities.
The Big Picture 231

In saying that these are reasonable views, I am neither adopting them nor contradict-
ing my basic belief (1.6) that there is a world external to the experiences of human
beings, features of which are accessible in common by human beings; but rather, I am
elaborating on my explanation that this external world may not be exactly as it appears
to us, with our particular perceptual and cognitive capacities. QM suggests a view of
this external reality very different from our perception of it, as having quite abstract
non-local features that are in effect decoded in conscious perception into what are to
us the meaningful features of location, motion, spatial extension, shape, colour, and so
on, which we attribute to things and states of affairs and processes in the external
world. This does not mean that those things and states of affairs and processes do not
really have these features at all, but rather that, when they are considered as existing
independently of being perceived, the features they do have are more accurately con-
sidered as abstract features adapted to producing in human beings the conscious expe-
riences of location, motion, spatial extension, shape, colour, and so on, and that it is
convenient to refer to those abstract features in terms of the substantive features that
are actually experienced.
Now since on that view our experiences do bring this element of meaningfulness to
what are otherwise abstract features, it may not be entirely unreasonable to propose
that our experiences also bring determinacy to what otherwise exist merely as
potentialities.
The view that measurement does not occur unless and until there is some perception
by a conscious observer has been subject to various quite powerful criticisms, which
I considered in section 14.3 of The Mind Matters but ultimately, in section 15.4, found
to be less than conclusive. One objection that I did not explicitly mention there is the
objection that, on this view, nothing could have existed other than as potentialities for
experience during the whole history of the universe until the emergence of conscious-
ness, when for the rst time there would have been some determinacy brought about,
together with the elimination of myriad incompatible potentialities that had been
multiplying for billions of years.
In fact I do not nd this scenario altogether implausible. I think that a universe that
is not at least capable of being experienced exists only in a most tenuous sense, if it can
be considered as existing at all; or to approach this another way, I think that some-
thing that is essential, if a universe is to exist in a way that truly counts as existing, is
that there be a capacity for some feature of this universe to be grasped and responded
to. If all there was to a universe were some aggregation of features that existed and
changed in conformity to some aggregation of laws (with or without randomness),
and that can never be grasped and responded to in some substantial way, it seems to
me that such a universe might just as well not exist, and cannot be considered as exist-
ing in a way that truly counts as existing. I think that for a universe to exist in a way

that is signicantly different from non-existence, it must at least have the potential to
give rise to something with the capacity to grasp and respond to features of the
universe.
The kind of response I have in mind here is not just an effect determined as con-
strained by the laws of the universe, because that is exactly what would happen in
those universes that might as well not exist. The response must be in some sense oc-
casioned by and apposite to what is grasped, yet not something that follows exactly as
dictated by the laws of the universe. This is of course what I have suggested to be the
case for human beings exercising free will in making rational choices; but I am also
suggesting here that a capacity for responses of this kind may be a wider feature of the
universe.
This capacity may conceivably have existed in or in relation to the universe prior to
the emergence of conscious life, in something having similarities to traditional con-
ceptions of God; but I think any reality concerning God is likely to be more elusive
than this. If this capacity rst arose in the universe with the emergence of conscious
life, it would follow that, on the view of QM I am considering, it would only be then
that the potentialities constituted by the external world would have given rise to actu-
alities of experience which in turn would exclude and eliminate potentialities in the
external world that were incompatible with them.
This would not mean that suddenly, on the rst experience being had by whatever
individual organism was the rst to have any kind of experience in which some aspect
of the universe could be grasped and responded to, a single history extending back to
the Big Bang was for the rst time established by reason of the elimination of potenti-
alities associated with all other histories. The determinacy established by any experi-
ence would on this approach be no more than what was necessary for that experience;
and the less discriminating the experience, the less would be its capacity to reduce in-
determinacies in the universe and to eliminate incompatible potentialities and histo-
ries. It is reasonable to suppose that consciousness as it rst occurred had only a small
capacity to discriminate, and thus only a small capacity to reduce indeterminacies and
to eliminate potentialities. If that is correct, then there would be a reasonable possibil-
ity that the elimination of potentialities incompatible with the history of the universe,
as it is now supposed to be, could have occurred progressively up to the present and
would still be occurring.
Whether or not the conscious grasping of features of the universe is necessary to give
determinacy those features, as one view of QM suggests, I suggest it is reasonable to
believe that our universe is and always has been suitable for supporting the capacity for
features of the universe to be grasped and responded to; and that this has come to be
realised in conscious entities that can experience features of the world, including
ourselves.
The Big Picture 233

12.2. An essential feature of our universe is its suitability for supporting a capacity to
grasp and respond to features of the universe, and thus for giving rise to conscious entities
that experience the universe or aspects of it.
Consistently with this, I also suggest that consciousness, and particularly rational con-
sciousness such as that of human beings, is very important in the existence and func-
tioning of the universe. There is of course a sense in which conscious systems like
human beings may be small beer in the cosmic scheme of things, as has been suggested
by the philosopher J. J. C. Smart;1 namely, in terms of physical quantities like size, mass,
energy, distance, and so on. Yet even in purely physical terms, size isnt everything; and
it is widely accepted that the human brain is the most complex physical system known
in the universe. More importantly, I think there is a vast difference between a universe
without observers and a universe with observers. Although I dont agree with those
idealist thinkers who say that we can only conceive of a universe as observed, so that an
unobserved universe is inconceivable, I do say that a universe without observers would
be pointless in a way that our universe is not. In addition, there is the possibility sug-
gested by QM and outlined above that participation of observers is an essential feature
of our universe; and there is also the view of many thinkers that our universe happens
to be ned-tuned in just such a way as to permit the emergence of conscious intelli-
gence. I think it is reasonable to believe that consciousness is very important in the
scheme of things.
CON STRAI NT, E M POWE R M E NT AN D GU I DANCE
In the case of human beings, a new system that has the capacity to experience and re-
spond to aspects of the universe comes into existence each time a new person comes
into existence, and it does so regularly, in very similar physical circumstances. So it is
reasonable to accept that there are laws that mediate the emergence of these systems, to
the effect that a system having such and such characteristics will be conscious and thus
will have the capacity to experience and respond to what is experienced. These laws
would not be of the same kind as the laws of nature that simply constrain what can exist
and what changes can occur.2 Rather, they would mediate the emergence of conscious
systems and empower them to bring about changes in ways that are not simply the
effect of constraining laws engaging with pre-change circumstances, with or without
randomness, but are selected by the system.3
1
Smart (2005) at 62.
2
I have called these laws C-laws or laws of constraint: Hodgson (2001).
3
I have called these laws E-laws or laws of empowerment: Hodgson (2001).

12.3. In addition to laws of nature that constrain what can exist and changes that can
occur, there are also laws of nature that mediate the emergence of conscious systems and
empower these systems to bring about changes that are not simply the effect of constraining
laws.
I have suggested in earlier chapters that the changes that conscious systems can make
are not simply as determined by constraining laws, but nevertheless are apposite re-
sponses to aspects of the universe that are grasped by the systems; and this raises the
question of how these systems can determine just what the responses should be. In the
case of human beings, I have suggested that they make choices by exercising the infor-
mal rationality exemplied by what I have called plausible reasoning; and while I am
condent about the informal rationality of human beings, for the reasons I gave in
chapters 2 to 6, and while I have given some account of how I understand that informal
rationality works, it is still puzzling what it is that can render choices, which are not in
accordance with laws or rules, other than arbitrary or capricious.
I believe the explanation must lie in the existence of principles that provide guidance
to conscious systems in responding to what they grasp, but without prescribing pre-
cisely what that response is to be. (If they did prescribe an exact response, they would
be no different from constraining laws; and in any event, I have argued that conscious
systems respond to gestalts with which general rules or laws do not engage, so that laws
of nature cannot prescribe these responses exactly.)
I have already suggested that there are natural imperatives and values of goodness
and beauty embedded in the universe; and what I now suggest is that these imperatives
and values are associated with principles that provide some kind of pull or attraction
or call upon conscious systems, and thereby give guidance to these systems in deter-
mining responses.4 The more primitive the conscious system, I suggest, the more in-
denite would be its grasp of these imperatives and values, and the more indenite
would be the pull or attraction or guidance that they provide.
In the case of human beings (and perhaps also other mammals and birds), one im-
portant aspect of their responses to their experiences concerns judgments they make
as to what to believe; and this would appear to require guidance not just towards what
is right and/or good and/or beautiful, but also towards what is true. It is, I believe, the
pull or call of this value that is in play when we exercise techniques of our informal
rationality, such as making judgments that have regard to incommensurable reasons
and/or to analogies that do not depend on identity or quantitative assessment of
common features.
4
I have called these principles G-laws or laws of guidance: Hodgson (2001).
The Big Picture 235

So I suggest:
12.4. There are laws or principles, which are associated with natural imperatives and
with values of goodness and beauty and truth, and which provide guidance to conscious
systems by pulling or calling them to these imperatives and values.
R E LIGIOUS B E LI E F: A S U BJ ECT FOR RATIONAL

E NQU I RY
What then does my account suggest about religious belief?

In chapter 2, I asserted my belief (2.5) in human rationality and also my belief (2.7)
that our rationality can be applied to questions beyond the reach of the scientic
method. Consistently with this, I say that religious questions can be addressed ratio-
nally, and that religious beliefs may possibly have rational supportalthough in rela-
tion to some religious questions, it may be that the only rational belief is that there is
no answer that can reasonably be accepted.
Not only do I believe that religious questions can be the subject of rational enquiry,
but I also believe we should not adopt, and certainly should not act on, any beliefs on
religious matters without subjecting them to rational scrutiny. In chapter 10, I asserted
my belief (10.3.b) that we should use our rationality to decide what is the right thing to
do and what is the wrong thing to do. Beliefs on religious matters can have a strong
bearing on these decisions: indeed religious beliefs can lead people to think they are
justied in taking actions that seem grossly wrong to others who do not share these
beliefs. And so I say that we should be rational in addressing religious questions, par-
ticularly in so far as they inuence judgments as to right and wrong, no less than in
addressing other questions that may be relevant to our conduct.
12.5. Religious beliefs can and should be subject to rational scrutiny to no lesser extent
than other beliefs.
A further reason for circumspection in adopting religious beliefs follows from another
belief I stated in chapter 2, namely my belief, included in my assertion 2.7, that lan-
guage may be inadequate to express the reality under consideration, and my associated
belief that this is so in relation to matters going beyond the practical situations that
were (it may be surmised) the main concerns of language users as language evolved.
Religious issues are largely remote from matters of everyday experience, and from mat-
ters that can be investigated scientically. They are also issues in relation to which
language may well be inadequate. Accordingly:

12.6. Many religious issues are to a considerable extent beyond our present ideas and lan-
guage, and to that extent literal and unambiguous truth about them is not to be expected.
A VALU E-E M B E DDE D U N IVE R S E
A view held by some scientists and others as to the nature of the universe has been elo-
quently expressed by Richard Dawkins:5
The universe we observe has precisely the properties we should expect if there
is, at bottom, no design, no purpose, no evil and no good, nothing but blind
pitiless indierence.
On this view, all our ideas, feelings and values have a complete explanation in a com-
bination of laws of nature and Darwinian selection. They require no more than this to
account for their existence, and they are no indicator that there is any more to the
universe than the raw materials on which the laws operate and the products of that
operationthese products including ourselves, our ideas, our feelings and our values.
What Ive written in earlier chapters shows why I do not accept this view.
My basic belief concerning the existence of conscious experiences means that I am
looking for a conception of the universe in which conscious experiences have a place;
and for reasons given in chapter 4, a conception of the universe as fully explained by
the operation of laws of nature on physical things, events and states of affairs does not
explain conscious experiences, or allow for any non-trivial place or role for conscious
experiences. A satisfactory conception of the universe requires at least the possibility of
an explanation of how it initially included the potential for the emergence of conscious
experiences and, once they have emerged, of what is their place and role in the work-
ings of the universe. These explanations are excluded by the conception of the universe
suggested by Dawkins.
Further, as explained in chapter 2, I believe in the existence of rationality which is not
wholly explicable in terms of processing of information in accordance with laws of
nature and/or computational rules; and as explained in chapters 4 to 6, I believe that at
least part of the explanation for this is that subjects of experience can respond ratio-
nally to whole conscious experiences (gestalts) in ways not wholly determined by the
operation of constraining laws. These views give informal rationality and conscious-
ness a substantial role in the working of the universe, and require a conception of the
universe that is consistent with their having such a role.
5
Dawkins (1995), end of chap. 4.
The Big Picture 237

Next, as explained in chapter 10, I believe there are values embedded in the universe,
including natural imperatives concerning the rightness and wrongness of choices
made by conscious subjects, and objective values of goodness and beauty.
I also believe there is evil, and that the workings of the universe involve much that is
bad. I do not believe this can be explained away by reference to wrong choices made by
human beings or other conscious subjects: in particular, I believe that the operation of
natural selection among living things makes it inevitable that there will be enormous
suffering through disease, hunger, exploitation and predation. But I believe that for
most human beings most of the time, and indeed probably for most conscious subjects
most of the time, existence as conscious subjects is better than non-existence; and this,
together with the beauty of the universe in general and our Earth in particular, leads
me to believe that good is stronger than evil.
Finally, there is the point that the universe seems precisely ne-tuned to make pos-
sible the emergence of life, without which it seems consciousness as we know it could
not occur. Life would be precluded by the minutest changes in properties of matter or
laws of nature.
Putting all these things together, I think it is reasonable to believe that there is some-
thing fundamental about the universe that has to do with the potentiality for con-
sciousness and conscious-based rationality, the existence of natural imperatives and
the values of goodness and beauty, the unfolding beauty of the universe itself, and the
predominance in it of good. I think it is also reasonable to regard all this as indicative
of a universe that is in some sense benevolently creative.
12.7. The universe is not one of blind pitiless indifference but one in which values of
goodness and beauty are embedded and which is in some sense benevolently creative.
From that, it might be thought a small step to personify or anthropomorphise these

features of the universe, so as to arrive at something like a traditional conception of a
monotheistic God. To take such a step may lead to ways of thinking about these fea-
tures of the universe that are easier to assimilate and manage than the vaguer and more
abstract ideas I have been suggesting, and that may operate more effectively to inspire
respect for what is good about the universe and to inspire conduct that is appropriately
responsive to that good. However, it is a step I do not take, except possibly to the extent
of regarding a person-like God as an allegory or metaphor for a more subtle and elu-
sive reality.
Once one postulates a God that is literally a person or like a person, existing in and/or
apart from the universe, one is led into blind alleys of questions: Did God create the
universe? How then was God created? Do natural imperatives apply to God? Does God
know what I am doing and thinking? If so, how? And so on. We can be pretty sure the

universe exists. It is unlikely we will ever know why there is something rather than noth-
ing, although we can make progress in understanding what is this something that exists,
and why this something is as it is. And it is as part of this enquiry that I have come to
believe that the universe is benevolently creative rather than one of blind pitiless indif-
ference; but this does not enable me to say anything useful on questions such as these.
WH E R E D O WE COM E FROM?
There is a very large and very beautiful painting by Paul Gauguin in the Boston
Museum of Fine Arts, called Where Do We Come From? What Are We? Where Are We
Going? It is the one work that I remember clearly from my visit there a few years ago,
and that would draw me back again. The painting depicts people at various stages of
life in a sumptuous Tahitian landscape: on the right a baby and three young women,
on the left another young woman and an old woman apparently near to death, and in
between two boys taking fruit, a man scratching his head in puzzlement, and women
in intimate conversation and contemplation. There are animals and birds, and a stone
idol. The painting suggests that, yes, we are human beings in a world of earth and water
and plants and animals, but it also conveys a powerful feeling of spirituality and other-
worldliness. And it conveys the mystery of the three questions of its title.
My discussion in this book has covered most of what I can usefully say about Gauguins
central question, What are we? Is there anything I can say about his other two questions?
Part of the story of where we come from is that told by science; but what the scientic
story misses is those features of the universe I have been considering in this chapter.
Some of what I think are crucial features of the universe that are disregarded by the
scientic story are:
1. its suitability for supporting a capacity to grasp and respond to features of the uni-
verse, and (possibly) its dependence upon that capacity for its existence otherwise
than as mere potentialities;
2. the role of consciousness in the existence and functioning of the universe;
3. the embedding in the universe of natural imperatives and values of goodness,
beauty and truth; and
4. what I believe to be the caring creativity of the universe.
The other general limitation of the scientic account that I noted was that it could not
explain its own raw materials. There is force in the argument, put by some scientists, that
explanations for our existence in terms of creation by God is not satisfactory because it
seeks to explain the existence of something for which we can give some explanation by
reference to something for which we can give no explanation whatever; and that the
The Big Picture 239

simple materials and laws by reference to which science explains our existence are far less
in need of explanation than is something as complex as a God that was not created by
anything else. However, the questions remain why there is something rather than noth-
ing, and why this something is just as it is.
I do not expect there will ever be an answer to the rst of those questions, but prog-
ress may be made on the second of them. However, what I am trying to do is to go as
far as I can on another question, namely just what is this something that exists and how
does it work; and on that basis, my conclusion on the question Where do we come from?
can be shortly expressed as follows:
12.8. We are the product of a universe in which the capacity to grasp and respond to its
features is important, in which natural imperatives and values of goodness, beauty and
truth are embedded, and which is caringly creative; and the nature of which can be illu-
minated by the metaphors and allegories of religion.
WH E R E AR E WE GOI NG?
Turning to Gauguins third question, Where are we going?, it is pertinent rst to con-
sider whether there is any purpose to our existence.
I have not embraced the idea that we are the creation of a person-like God existing
in or separate from the universe, and the ordinary concept of purpose suggests a pur-
pose held by a person or at least something person-like. Further, even if a person-like
God had some purpose in mind for us in bringing us into existence, it is not clear that
this of itself would mean that we must or should recognise and adopt Gods purpose
as our own purpose.
However, I have supported the view that we are the product of a universe in which
natural imperatives and values of goodness, beauty and truth are embedded and which
is caringly creative, and the nature of which can be illuminated by the metaphors and
allegories of religion. In those circumstances, I think it is not unreasonable to believe
there is some purpose to our existence, in the sense of a purpose (in some extended
meaning) of this caring and values-embedded universe, and/or in the sense of a pur-
pose that we, as a matter of morality, should adopt as our purpose.
12.9. There is some purpose to our existence, in the sense of a purpose of a caring and
values-embedded universe and/or in the sense of a purpose that we, as a matter of moral-
ity, should adopt as our purpose.
One possible such purpose was suggested earlier in this chapter, in the requirement for
a universe of interest that aspects of it be grasped and responded to; and this suggests

a purpose for conscious entities of experiencing and responding to and appreciating
the universe. A related possible purpose is suggested by the discussion in chapter 10
concerning what is good and beautiful, namely to enrich the universe by contributing
to the creation, ourishing and appreciation of what is good and beautiful. There
could be another possible purpose, if there is some kind of existence beyond this life of
ourselves or of some kind of consciousness related somehow to ourselves, namely a
purpose associated with preparing for and/or enhancing that kind of existence.
12.10. Our purpose may be to experience and respond to and appreciate the universe,
and to enrich the universe by contributing to the creation, ourishing and appreciation of
what is good and beautiful, and possibly also to prepare for and/or enhance some other
kind of existence.
So what can I say about where we are going, if anywhere?

I think there are reasons to think that there may be some mode of existence beyond
our lives on this Earth, in particular the nature of the universe as value-embedded and
caringly creative, the transcendent qualities of some works of art and music and of
some persons, and the idea that there should be some kind of justice for undeserved
suffering.
12.11. There may be some mode of existence beyond our lives on this Earth, having regard
to the nature of the universe as value-embedded and caringly creative, the transcendent
qualities of some works of art and music and of some persons, and the idea that there
should be some kind of justice for undeserved suffering.
However, I think it is unlikely in the extreme that there is anything like a straightfor-
ward survival of death of individual persons, either as disembodied spirits or through
reconstitution of the body and brain or through reincarnation. If I were to survive
death either as a disembodied soul or in a reconstituted body, what version of me
would this be? Would it be me as I was in the prime of life, whatever that may be, or as
I was just before death, perhaps suffering from senile dementia? What I am, at any time
of my life, is very much bound up with my physical and psychological make-up at that
particular time; and I cannot identify any particular subject of my experiences that can
be separated out from the various ways I have been at different times of my life.
These considerations apply even more strongly against any notion of reincarnation
as a different person or animal. The character traits of any other person or animal
would be those constituted by the physical and psychological make-up of that person
or animal, and I cannot make any sense of how I, or any aspect of me, could be contin-
ued in any other person or animal.
The Big Picture 241

In appendix B I give moral reasons why I do not accept anything like a traditional
conception of heaven and hell, which supposes that human souls are divided into two
groups, one of which enjoys eternal rewards in heaven and the other of which suffers
eternal punishment in hell; and this would in any event be ruled out if there is no
straightforward survival of individuals after death. This consideration could be seen as
removing all force from one factor I mentioned in favour of the view that there may be
some mode of existence beyond our lives on this Earth, namely that there should be
some kind of justice for undeserved suffering. If there is no straightforward survival of
individuals, how can there be justice for those who have died after a life of undeserved
suffering, or indeed for those who have lived an enjoyable life despite their evil actions
causing suffering to others?
There is no easy answer to these questions, but I do not think this consideration
removes all force from the idea that justice calls for some kind of existence beyond
our lives on this Earth. However, it is plain from what I have said that I believe any
such existence must be a mode of existence quite different from our present lives, a
mode of existence the nature of which cannot be fully captured by our ideas or our
language.
I believe that if there is some mode of existence beyond our present lives, it might
have some or all of the following features:
1. in so far as any individuality of particular subjects is preserved, it could be an indi-

viduality that somehow sums up or comprehends the totality of the life of this
subject, and thus has a different relation to time than we do as human beings in
this life;
2. there could be a breaking down of the separateness of individual subjects;
3. there could be some kind of merging into whatever it is about the universe that is
caringly creative and in which values are embedded;
4. the quality of this existence could itself be affected by what has happened in this life,
by reason of what has been created or preserved or destroyed and/or by reason of
the responsibility for this of particular individuals.
Beyond these very general suggestions, I feel I can contribute nothing useful, so
I conclude:
12.12. There is no straightforward survival of death of individual persons, either as a

disembodied spirit or through reconstitution of the body and brain or through reincarna-
tion; and any existence beyond our lives on this Earth must be a mode of existence quite
different from our present lives, a mode of existence the nature of which cannot be fully
captured by our ideas or our language.

CAN MOR E S PECI FIC B E LI E FS B E S U PPORTE D?
It could be said I am arrogantly relying on my own ideas here, and ignoring a great
body of evidence that does make it reasonable to believe in more specic assertions
about God: the authority of holy books, the authority of prophets, the authority of
great religious institutions, the consensus of large numbers of people, and the persua-
siveness of the content of teachings of religions. And the resurrection of Jesus.
Holy books are said to be authoritative because they are inspired by God, and proph-
ets are said to be authoritative because they are in communication with God. Such
arguments are agrantly circular, seeking to prove the existence of a God that inspires
holy books and communicates with prophets, by means of these supposed inspirations
and communications. They are also highly implausible. There is no good reason to
think that either the authors of holy books or the prophets were other than human
beings like us, relying for their beliefs on their own perceptions and reasoning. If
people today claim to be authoritative sources of religious truth because they are in-
spired by God or because they are in communication with God, their claims are rightly
regarded with scepticism. I see no reasonable basis for any different view concerning
the authorship of holy books or the authority of prophets.
Ive heard it suggested that it is reasonable to accept that the Christian Bible is the
word of God, unless and until the contrary is proved. But well-documented internal
inconsistencies in the Bible, and the existence the holy books of other religions convey-
ing quite different messages, take away any rational basis for assuming that the Bible
has the imprimatur of the one true God, and thus as being literally true, unless and
until the contrary is proved. Having regard in addition to the circumstance that the
Bible was written by human beings like us, I say that the only reasonable course is to
assess its credibility in the same way as that of other ancient texts and other holy
books.
As regards the authority of religious institutions and the consensus of large numbers
of people, these may have been factors supporting the reasonableness of beliefs in his-
torical times when there was a broad consensus on religious beliefs throughout whole
communities, a consensus that was also in accord with the teachings of highly respected
religious institutions which had a monopoly on learning and scholarship in those
communities, and when there was little if any means of knowing much about other
communities and institutions. But today we know there is no consensus, and no good
reason for giving particular weight to views promulgated by any particular religious
institutions: there are believers and non-believers, and among believers and religious
institutions there are different and conicting belief systems. And what we now know
about the history of great religious institutions such as the Roman Catholic Church
can only inspire suspicion of their reliability as sources of religious wisdom.
The Big Picture 243

And as for the persuasiveness of the content of religious teachings, although there is
much in the moral teachings of the religion with which I am most familiar, the Christian
religion, that I nd persuasive, there is also much in its teachings that I nd unpersuasive.
And the same is true of other religions, to the extent that I am familiar with them. I elabo-
rate on this in appendix B, where I give reasons why I see aspects of the Christian Bible as
being grossly immoral, destroying its reliability as a text of authority.
Thus, I do not think it is reasonable to rely on the authority of holy books, the au-
thority of prophets, the authority of religious institutions, the opinions of large num-
bers of people or the content of religious teachings, in order to justify beliefs about
God that do not otherwise have the support of evidence or rational arguments. And
I believe:
12.13. On rational scrutiny, no current religion can be accepted as literal and unambigu-
ous truth, such as could make different beliefs false; but all great religions may qualify as
metaphors with degrees of truth.
POTE NTIAL FOR EVI L AN D GO OD
I suggest in appendix B that some Old Testament stories have the potential to inspire
evil. And history shows that acceptance of the literal truth of religious beliefs has
indeed inspired great evil. Particularly has this been so when the acceptance of the lit-
eral truth of a belief system of a religion has involved the belief that those with differ-
ing views are wrong and to be condemned.
Obvious historical examples are the Crusades and the Inquisition. In the former,
wars were started because of a belief that a particular part of the world should not be
in the hands of people with beliefs different from those of the aggressors. In the latter,
people were imprisoned, tortured and killed for professing beliefs different from those
of ofcial church teachings.
There are also more contemporary examples, some of which cannot be dismissed as
mere aberrations of extreme minorities.
I understand that it is a not uncommon belief among Muslims that it is appropriate that
Muslims who renounce their faith should be severely punished, or in some cases even
killed, and I also understand that there are places where this is put into effect. This view
does not seem to be strongly condemned by the majority of inuential Muslims. Certainly
it should be so condemned, and inuential Muslims who are in any position to do so
should be denouncing such an evil belief and striving to ensure that it is eliminated.
There are still people in the Middle East who believe that there is part of the world
that has been given by God to a particular group of people in which they are included,
and that forceful disposition of other people from that land has been and is not only

sanctioned by God but also a religious duty, and who strive to give effect to these be-
liefs. Although these beliefs are apparently professed only by a minority of their group,
it has been and remains an inuential minority whose views are not rmly repudiated,
as they should be, by others in the group.
More generally, any belief that characterises one group of people as the elect or as saved
or as having the one true religion, and other groups as indels or otherwise less worthy, is
highly dangerous, particularly if combined with a belief that those in the former group have
the support of God and know what the will of God is. So although many people who be-
lieve in the literal truth of their religion are the most admirable of people who live exem-
plary lives, I think there is great potential for evil in the practice of religion on that basis.
Christians and Jews might argue that this has no application to them: they do not sup-
port genocide; they do not demonstrate violently when their religion is said to be violent:
they do not advocate terrorism or the slaying of indels; they do not condone the killing
of apostates. Well, I think it does apply to them. I think its reasonable to believe that
failure of Christians and Jews to squarely repudiate the morality of Old Testament stories
can have signicant consequences, both in their own conduct and in holding back their
challenge to evil beliefs derived from Islam. It is undeniable that moral beliefs inuence
conduct, and its reasonable to think that beliefs that demonise the enemy and justify
extreme violence against the enemy contribute to violence in the world.
The truly worrying thing about fundamentalist Christians and Jews is not that they
believe things happened that did not happen, like the Flood or the walls of Jericho
tumbling down, but that their beliefs must mean they admire the God portrayed in
these stories and the morality of that God. And this worry applies with similar force to
those Christians and Jews who do not believe in the literal truth of these stories, but
nevertheless believe them to have a kind of truth in telling us about the nature of God
and Gods dealings with human beings, and about morality.
I was particularly struck by the reference in Dawkinss book The God Delusion6 to a
study carried out by Israeli psychologist George Tamarin. In this study, reported in
1966, he presented Joshua 6:2021 (telling of how Joshua and his followers killed every
man, woman, child and animal in the defeated city of Jericho) to 1,066 Israeli school-
children, aged eight to fourteen, across a broad spectrum of Israeli social and economic
classes. He asked them the question Do you think Joshua and the Israelites acted
rightly or not? Their answers were categorized as follows: A means total approval,
B means partial approval or disapproval, and C means total disapproval. Sixty-six
percent of responses were A, 8 percent B, and 26 percent C. The A group made com-
ments such as that In my opinion Joshua was right when he did it, one reason being
that God commanded him to exterminate the people so that the tribes of Israel will not
6
Dawkins (2006), 25557.
The Big Picture 245

be able to assimilate amongst them and learn their bad ways. Even the disapproving
groups included comments such as that I think Joshua did not act well, as they could
have spared the animals for themselves.
This is just one study, carried out over four decades ago. But it has strong implica-
tions about the beliefs of the adults who inuenced these children, and who might not
themselves have given such frank expression to their views; and my search of the Inter-
net has disclosed no reference suggesting error by Tamarin, or any different results
obtained in more recent studies.
Despite all this, it remains the case that religion has also inspired the most admirable
of conduct by many people and has also inspired some of the greatest artistic, literary,
architectural and musical achievements of humankind. And as I suggested earlier, a
view that personies or anthropomorphises features of the universe that are benevo-
lently creative may lead to ways of thinking about these features that are easier to as-
similate and manage than the vaguer and more abstract ideas I suggested, and that may
operate more effectively to inspire respect for what is good about the universe and to
inspire conduct that is appropriately responsive to that good.
So I readily accept that the adoption and practice of religious beliefs can be valuable:
it may not only help individuals to nd meaning in their lives and to live good lives, but
may also promote a sense of community and common participation in a worthwhile
and uplifting endeavour. But it is important that any leap of faith to religious belief be
consistent with rationally held beliefs about the world and particularly about morality,
and that any beliefs contrary to reasonable morality should be rejected. So I say that it
cannot be reasonable to accept any one religion as providing literal and unambiguous
truth, such as could render different beliefs false; but that all religions may qualify as
helpful metaphors providing ways of relating to what is benevolently creative about
the universe and what is good and beautiful about the universe.
12.14. Practice of any particular religion on the basis that it is literally and unambigu-
ously true, that different beliefs are false, and that people with different beliefs are to be
condemned, is not rationally justied and has the potential for great evil; but practice of
religion not as providing literal truth, but as one way of relating to what is good, beautiful
and benevolently creative about the universe, can be valuable.
COR E ASS E RTION S AB OUT TH E B IG PICTU R E
12.1. The scientic account of our origins, in terms of matter and energy and laws of
nature, is substantially correct, so far as it goes; but it gives no account of consciousness,
values and free will, and it cannot explain its own raw materials.

12.2. An essential feature of our universe is its suitability for supporting a capacity to
grasp and respond to features of the universe, and thus for giving rise to conscious entities
that experience the universe or aspects of it.
12.3. In addition to laws of nature that constrain what can exist and changes that can
occur, there are also laws of nature that mediate the emergence of conscious systems and
empower these systems to bring about changes that are not simply the effect of constraining
laws.
12.4. There are laws or principles, which are associated with natural imperatives and
with values of goodness and beauty and truth, and which provide guidance to conscious
systems by pulling or calling them to these imperatives and values.
12.5. Religious beliefs can and should be subject to rational scrutiny to no lesser extent
than other beliefs.
12.6. Many religious issues are to a considerable extent beyond our present ideas and lan-
guage, and to that extent literal and unambiguous truth about them is not to be expected.
12.7. The universe is not one of blind pitiless indifference but one in which values of
goodness and beauty are embedded and which is in some sense benevolently creative.
12.8. We are the product of a universe in which the capacity to grasp and respond to its
features is important, in which natural imperatives and values of goodness, beauty and
truth are embedded, and which is caringly creative; and the nature of which can be illu-
minated by the metaphors and allegories of religion.
12.9. There is some purpose to our existence, in the sense of a purpose of a caring and
values-embedded universe and/or in the sense of a purpose that we, as a matter of moral-
ity, should adopt as our purpose.
12.10. Our purpose may be to experience and respond to and appreciate the universe,
and to enrich the universe by contributing to the creation, ourishing and appreciation of
what is good and beautiful, and possibly also to prepare for and/or enhance some other
kind of existence.
12.11. There may be some mode of existence beyond our lives on this Earth, having regard
to the nature of the universe as value-embedded and caringly creative, the transcendent
qualities of some works of art and music and of some persons, and the idea that there
should be some kind of justice for undeserved suffering.
12.12. There is no straightforward survival of death of individual persons, either as a
disembodied spirit or through reconstitution of the body and brain or through reincarna-
tion; and any existence beyond our lives on this Earth must be a mode of existence quite
different from our present lives, a mode of existence the nature of which cannot be fully
captured by our ideas or our language.
12.13. On rational scrutiny, no current religion can be accepted as literal and unambigu-
ous truth, such as could make different beliefs false; but all great religions may qualify as
metaphors with degrees of truth.
The Big Picture 247

12.14. Practice of any particular religion on the basis that it is literally and unambigu-
ously true, that different beliefs are false, and that people with different beliefs are to be
condemned, is not rationally justied and has the potential for great evil; but practice of
religion not as providing literal truth, but as one way of relating to what is good, beautiful
and benevolently creative about the universe, can be valuable.

Appendix A
Why Bayes Theorem Works
To supplement what I said about Bayes Theorem in chapter 3, I will try to convey some un-
derstanding of its validity, and of how and why it works, with reference to my example from
that chapter; rst, with some attempt at rigour, and then, in a more simple and intuitive way.
The former exercise uses tables.
We start with a table based on the likelihood ratio given by the rst piece of evidence, namely
the occurrence of the pregnancy. This sets out the probabilities of that event (pregnancy) oc-
curring or not occurring, conditional in each case upon the truth of the hypothesis (improper
operation) or its negation (proper operation), with one column dealing with each of these two
assumptions.
Each column adds up to 1, because for the gures within each column it is assumed that the
hypothesis dealt with in that column is true, giving it probability 1. However, we know that
the two hypotheses cannot both be true; and we have assumed that the odds between them,
prior to taking the evidence into account, are 1:99 against the improper operation. So to arrive
at the absolute probabilities for the rst piece of evidence and its negation, prior to it being
ascertained whether pregnancy occurs or not, one should adjust this table so that it shows the
correct probabilities for the hypothesis and its negation, rather than the probabilities based on
the impossible assumption that both alternatives are true.
The probability of the improper operation, reecting the prior odds, is 1/100, while that of
the proper operation is 99/100. So to get to the absolute probabilities for the rst piece of
evidence and its negation, we make proportional adjustments to the rst column so that the
gures add up to 1/100 rather than 1 (that is, we multiply each gure by 1/100), and to the
second column so that the gures add up to 99/100 rather than 1 (that is, we multiply each
gure by 99/100). This gives the conditional probabilities shown in table A.1.
We see from the rst row of table A.2 that the total probability of pregnancy following the
operation is 9/1000 + 99/35,000, that is (315 + 99)/35,000 or 414/35,000; and we see from the
second row that the total probability of no pregnancy following the operation is 1/1,000 +
34,551/35,000, that is, 34,586/35,000. These probabilities add up to 1, as they should.
What happens to these gures when it is ascertained that the pregnancy in fact occurs?
The probability represented by the total of rst row of table A.2 becomes 1, and the proba-
bility represented by the total of second row becomes 0. That means that, without altering
the proportions, the total of the rst row should be adjusted to 1 instead of 414/35,000, and
the total of the second row becomes 0. To do this, we must multiply each entry in the rst
row by 35,000/414, and each entry in to second row by 0. In that way, we get a new table, the
rst row of which shows the adjusted probability of the hypothesis, having regard to the
evidence.
That is, the pregnancy that has in fact occurred has a probability of 315/414 of having oc-
curred in association with an improper operation, and a probability of 99/414 of having oc-
curred in association with a proper operation.
249
Table A.1 Conditional Probabilities of Evidence
Improper Operation Proper Operation
Pregnancy 9/10 1/350

No pregnancy 1/10 349/350
Total 1 1
Table A.2 Absolute Probabilities of Evidence
Pregnancy 9/1000 99/35,000

No pregnancy 1/1000 34,551/35,000
Total 1/100 99/100
Table A.3 Adjusted Probability of Hypothesis on the Basis of the Evidence
Pregnancy 315/414 99/414

No pregnancy 0 0
All this amounts to the same thing as getting a likelihood ratio of 315/1 (the ratio between the
two entries in the rst row of table A.1) and multiplying it by the prior odds of the hypothesis
1:99 (reected in the totals of the columns in table A.2), giving posterior odds of 315:99, and
thus probabilities of 315/414 and 99/414 as shown in the rst row of table A.3. This accords
with the result of the rst stage, when I originally discussed this example.
The same process can then be repeated with the second piece of evidence (the eyewitness
accounts), with the result indicated earlier.
One can also give a simple and intuitive account of the whole process, in this case using the
eyewitness evidence rst. The initial odds against an improper operation are assumed to be
1:99. The eyewitness evidence makes it four times more unlikely that the operation was im-
proper, so the adjusted odds are 1:396 against. The statistical evidence makes it 315 times more
likely that the operation was improper, so the nal odds are 315:396 against, giving a probabil-
ity for the improper operation hypothesis of 315/711 or 0.44.

Appendix B
Against Fundamentalism: Biblical Morality
In chapter 12, I contended that religious fundamentalism has the potential for great evil, and
cannot be supported by reference to the authority of holy books. Here, I support that conten-
tion by considering one example of a holy book, the Christian Bible, and suggesting that its
authority is undermined by what I see as the gross immorality of many of its stories.
I will illustrate this by rst considering a story that is common to three great religions, Juda-
ism, Christianity and Islam, namely the story of Abraham and Isaac; and I will then consider
two other stories from the Old Testament. And then I will go on to consider whether the au-
thority of the Bible can be restored by its account of Jesus, because of the persuasiveness of
some of his moral teachings and the evidence for the resurrection.
The immorality of these and other accounts in holy books is I suggest a powerful reason to
reject religious fundamentalism. It is also a matter to be taken to heart by believers who are not
fundamentalists. When I received my non-fundamentalist Christian education, I was told that
the New Testament gave a better picture of God and morality than the Old, but stories such as
those I consider below were presented to me as stories of Gods dealings with his special
people, and I was given no warning of their gross immorality.
AB RAHAM AN D I SAAC
In the Old Testament story of Abraham and Isaac (Genesis 22), God tells Abraham to offer his
son for a burnt offering. Abraham builds an altar, prepares wood for a re, ties Isaac up, lays him
on the altar, and takes a knife to kill him. Only then, an angel tells Abraham not to harm Isaac,
and Abraham sees a ram caught in a thicket; and Abraham sacrices the ram instead of his son.
This story is generally presented as an example of meritorious sacrice and obedience to
God, carrying also the message that even if what God requires seems difcult to understand,
God will make sure it turns out for the best. To my mind, its message about both God and
Abraham is abhorrent, and has the potential for great evil.
About God, it says that God expects obedience to Gods command to kill an innocent child,
where there is no discernable reason for this except that it would please God (!) to have the
child killed and to have obedience shown in this way; that God expects followers to respect a
God who would be pleased to have an innocent child killed for no better reason than this; and
that God would without good reason subject an innocent child to a terrifying ordeal.
About Abraham, it says that he had respect for such a God to the extent that he would, on
the basis of such a capricious order, kill an innocent child. And that is quite apart from the
point that Abraham, as a human being with no more than our capacities for perception and
reasoning, could not have had any reasonable basis for believing in the existence of a God who
would have such expectations or issue such an order, or for believing that such an order had
actually been issued to him.
251
Even if Abraham had seen a great face in the sky speaking to him and had heard the words
spoken, it would have been more reasonable for him to believe this was a dream or hallucina-
tion than to believe that a God, conceived of as good, would have such expectations and would
issue such an order. And even if Abraham was justied in believing that what he saw and heard
was not a dream or hallucination, the reasonable conclusion for him to reach would have been
that this supernatural phenomenon was a manifestation of evil not of good.
Its been suggested that Abrahams willingness to sacrice the child he loved is admirable.
But that assumes Isaac was Abrahams to sacrice; whereas in truth no person belongs to an-
other in that way. It has also been suggested that killing an innocent child is not wrong if God
has commanded it. But that assumes it is Gods command that makes things right or wrong;
whereas, as I have argued, there would be good moral reasons to obey Gods commands only
if morality had force independently of Gods commands. And this suggestion also ignores the
point that human beings only have their perception and their reasoning to ascertain whether
there is a God and if so what its commands are; and reason is strongly against there being a
God who would issue such commands.
So this story is about a God unworthy of respect, and an Abraham who was prepared to do
something grossly immoral, to kill an innocent child, for no good reason that he could have
apprehended. And it has the potential to inspire great evil in its message, apparently accepted
by some people today, that it is OK to kill innocent people if you believe God has told you to.
TH E PASSOVE R
Another story in the Old Testament recounts how God killed all the rstborn children of a
group whose ruler Pharaoh was oppressing the Israelites, in order to induce Pharaoh to free
them.
In Exodus 11, Moses predicts that God would kill all the rstborn of Egypt, from the rst-
born of Pharaoh that sitteth upon his throne, even unto the rstborn of the maidservant that
is behind the mill. Exodus 12 tells of how God went ahead and did this, avoiding killing any
Israelites by passing over houses where blood had been placed on the lintel and the two side
posts. The morality of the God depicted in this Passover story, who kills children (including
the rstborn of the maidservant that is behind the mill) to put pressure on a leader to achieve
a worthy political outcome, seems no different from the morality of suicide bombers and
other Islamist terrorists.
Its been said that what God did was a last resort, to free good people from enslavement by
bad people, after the Egyptians had been given every chance to act on less extreme incentives.
But suicide bombers regard their objectives similarly, and what is most appalling about them
is not their objectives, but their targeting of innocent and powerless people, people like the
rstborn of the maidservant behind the mill. And this attempted justication of what God is
said to have done in the Passover story conrms the implication from the story that Christians
and Jews, like the Israelites, are good, with God on their side and with just causes against bad
people, causes for which it must surely be right to act as God acted.
Its also been said that God moves in mysterious ways, and we should not presume to judge
God. But God (accepting this concept) gave us the ability to reason about moral issues; and as
I said earlier, we cannot assume the Bible is inspired by God. The fact that the Bible, which is
supposed to enlighten us on moral issues, contains messages so plainly contrary to reasonable
morality, is a powerful reason for thinking its vision of God is in places a awed one, created
by fallible human beings.

TH E PROM I S E D LAN D
A further story in the Old Testament recounts how the one true God, who lovingly created all
of humankind, favoured one group of human beings over others to the extent of giving the
favoured group land, and instructing them to slaughter the people who previously occupied
it. Deuteronomy 20 recounts Gods instructions to the Israelites on what to do when they
defeat cities. In the case of cities that are not in the land God has given them, they are to kill all
the men and take for themselves the women and children; and in the case of cities that are
in the land God has given them, they are to kill everyone. Joshua 6: 2021 recounts that when
Joshua and his followers took Jericho, they utterly destroyed all that was in the city, both man
and woman, young and old, and ox, and sheep, and ass, with the edge of the sword, as
instructed by God.
The ghting of a war of aggression because of a belief that God authorised it is bad enough;
but the subsequent killing of the defeated people is simply appalling. The morality of the God
of these passages seems no different from the morality of the Serb ethnic cleansers in Sre-
brenica (in those cases where only the men are to be slaughtered), or the Nazis of the Holo-
caust (in those cases where everyone is to be slaughtered).
Its been said that the people of Jericho were evil and could not be permitted to corrupt the
Israelites; but this is just what Serb ethnic cleansers and Nazis would say. Theres no reasonable
basis for believing that the people of Jericho were any worse than any other group of people at
the time, much less that their conduct could justify the slaughter of every man, woman and
child. And again, this attempted justication conrms the implication that we are good and
they are evil, so these are things we may do to them but they must not do to us.
TH E N EW TE STAM E NT
Finally, there is the Bibles account of Jesus. There is much about the teachings of Jesus that
suggests extraordinary moral insight, and the resurrection, if it occurred, would be strong evi-
dence for his divinity; so that (it might be said) at least the New Testament, as the primary
account of the life of Jesus and his teachings, must be considered as authoritative.
The New Testament presents Jesus as having given central importance to caring about other
human beings, over and above complying with particular rules of behaviour, however well-
entrenched those particular rules may be; and as including among the human beings about
whom one should care not just ones friends but also ones enemies. It presents him as a person
who was particularly concerned for the lowliest of society. It presents him as attacking hypoc-
risy and self-righteousness, and recommending that those condemning the behaviour of
others should have regard also to their own shortcomings. I do not know if these ideas were
entirely new with Jesus; but I dont believe they were widely held views at the time, and they
seem to me to be wholly admirable.
And then there is the resurrection. The Bibles accounts of various parts of the story are
second- or third- or fourthhand; but they have some historical credibility. In addition, there is
the extraordinary history of a religion growing from the apparent defeat and death of its
founder, which leads one to look for an explanation of how the early followers of Jesus were
able, against all odds, to promulgate the divinity of Jesus and the idea that his death and resur-
rection were for the salvation of humankind.
But for me this is not enough to show that the Bible can be considered authoritative in its
account of Jesus or his resurrection.
Against Fundamentalism 253

One important reason for this is that, quite apart from conicts and inconsistencies in the
biblical accounts, there are aspects of the Bibles presentation of Jesus that are so contrary to
my other beliefs that I cannot accept them.
Jesus is presented by the Bible as a continuation of the revelation of the divine contained in
the Old Testament, and as so regarding himself, thus associating him with the view of God ex-
emplied by these stories. For example, the Bible presents the Last Supper as a celebration of
the Passover, suggesting that Jesus condoned the killing of children to persuade Pharaoh to
release the Israelites.1 And there are other aspects of the Bibles account of Jesus that seem to
me to be morally abhorrent, in particular his endorsement of a stark view of heaven and hell,
according to which all humankind would be separated into two groups, the sheep destined for
life eternal and the goats for everlasting punishment (Matthew 25: 3146).
I think its obvious there is a continuous spectrum of human character, from very good to
very bad, with most of us somewhere in the middle. The idea that a line will be drawn, so that
those just on one side would be destined for eternal bliss, and those just on the other side
would be destined for eternal suffering, is arbitrary, unjust and abhorrent. And if it is said that
the criterion is not merit but faith in Jesus, this would be worse, because there are good and
honest reasons for not having that faith and many persons wont have had the slightest op-
portunity to have that faith. And all this is quite apart from the powerful considerations (1)
that where one ends up on any spectrum of merit is enormously inuenced by genes and en-
vironment, even if, as I do believe, we have some capacity through free will to modify our
handicaps in life; and (2) that this stark scheme of heaven and hell has no intelligible place for
infants and the mentally ill.
I am not here suggesting that it must be unreasonable to believe in the divinity of Jesus or in
his resurrection, but merely that in its presentation of Jesus, the Bible cannot reasonably be
considered authoritative, much less as having divine authority. Accordingly, even in relation to
the New Testament and its account of Jesus, fundamentalism is to be rejected.
1
I do not mean here to be critical of everyone who celebrates the Passover. What I am objecting to
is the combination of acceptance of the Bible story of the Passover as literally true, or as conveying
some substantial truth about Gods dealings with humankind, and unqualied celebration. I think
celebration of the surrender of Japan in 1945 should be tempered by acknowledgment of the killing of
innocent people in Hiroshima and Nagasaki; and in that case, it is arguable that what was done was the
least of the available evils. By contrast, on the Bibles account of the Passover, Gods resort to terrorism
(as opposed for example, to either not hardening Pharaohs heart, or targeting Pharaoh and his army)
was gratuitous. The Bibles account of Jesus giving instructions for the Last Supper suggests no reserva-
tions either about the understanding of the Passover suggested by the Biblical account of it, or about its
morality so understood.

References
Abelson, R. 1988. Lawless Mind. Philadelphia: Temple University Press.

Aspect, A., J. Dalibard, and G. Roger. 1982. Experimental test of Bells inequalities using time-
varying analyzers. Physical Review Letters 49, 18047.
Baars, B. 1997. In the Theater of Consciousness. New York: Oxford University Press.
Bennett, M., and P. Hacker. 2003. Philosophical Foundations of Neuroscience. Oxford:
Blackwell.
Blackburn, S. 1998. Ruling Passions. Oxford: Oxford University Press.
Blackburn, S. 2001. Being Good. Oxford: Oxford University Press.
Blackburn, S. 2005. Truth. London: Allen Lane.
Blakemore, C., and S. Greeneld, eds. 1987. Mindwaves. Oxford: Blackwell.
Bohm, D., and B. Hiley. 1993. The Undivided Universe. London: Routledge.
Churchland, P. M. 1990. Eliminative materialism and propositional attitudes. In Lycan 1990.
Churchland, P. S. 1986. Neurophilosophy. Cambridge, MA: MIT Press.
Clarke, R. 2008. Incompatibilist (nondeterministic) theories of free will. In The Stanford En-
cyclopedia of Philosophy, ed. E. Zalta. URL = http://plato.stanford.edu/entries/incompati-
bilism-theories/.
Cleland, C. E. 2002. Methodological and Epistemic Differences Between Historical Science
and Experimental Science. Philosophy of Science 69, 47496.
Conway, J., and S. Kochen. 2006. The free will theorem. Foundations of Physics 36(10), 1441.
URL = 1441. URL = http://arXiv:quant-ph/0604079v1.
Conway, J., and S. Kochen. 2009. The strong free will theorem. Notices of the American Math-
ematical Society 56(2), 22632. URL = http://arXiv:0807.3286v1.
Copeland, B. J. 2002. The Church-Turing thesis. In The Stanford Encyclopedia of Philosophy,
Fall 2008 edition, ed. E. Zalta. URL = http://plato.stanford.edu/entries/church-turing/.
Damasio, A. 1996. Descartes Error. London: Macmillan.
Davidson, D. 1970. Mental events. In Foster and Swanson 1970.
Dawkins, R. 1995. River Out Of Eden. London: Weidenfeld and Nicolson.
Dawkins, R. 2006. The God Delusion. London: Bantam Press.
Dennett, D. 1991. Consciousness Explained. New York: Allen Lane.
Dennett, D. 2003. Freedom Evolves. London: Allen Lane.
De Wall, C. N., R. F. Baumeister, and E. J. Masicampo. 2008. Evidence that logical reasoning
depends on conscious processing. Consciousness and Cognition 17, 62845.
Dijksterhuis A., M. W. Bos, L. F. Nordgren, and R. B. van Baaren. 2006. On making the right
choice: the deliberation-without-attention-effect. Science 31, 10057.
Dutton, D. 2009. The Art Instinct. New York: Bloomsbury.
Eccles, J. C. 1994. How the Self Controls its Brain. Berlin: Springer.
Einstein, A., B. Podelsky, and N. Rosen. 1935. Can quantum-mechanical descriptions of physi-
cal reality be considered complete? Physical Review 47, 77780.
Fischer, J. 2002. Frankfurt-type examples and semi-compatibilism. In Kane 2002a.
255
Fischer, J. 2006. The cards that are dealt you . Journal of Ethics 10, 10729.
Fischer, J. 2008. The direct argument: you say goodbye, I say hello. In Trakakis and Cohen
2008.
Fischer, J., and M. Ravizza. 1998. Responsibility and Control. Cambridge: Cambridge University
Press.
Fodor, J., and B. Pylyshyn. 1988. Connectionism and cognitive architecture: a critical analysis.
Cognition 28, 371.
Foster, L., and L. Swanson, eds. 1970. Experience and Theory. Amherst: University of Massachu-
setts Press.
Frankfurt, H. 1983. Alternative possibilities and moral responsibility. Journal of Philosophy 80,
25778.
Gazzaniga, M. 1988. Mind Matters. Boston: Houghton Mifin.
Gibbard, A. 1990. Wise Choices, Apt Feelings. Oxford: Oxford University Press.
Goodman, N. 1965. Fact, Fiction and Forecast. New York: Bobbs-Merrill.
Goodman, N. 1970. Seven strictures on similarity. In Foster and Swanson 1970.
Greene, J., and J. Cohen. 2004. For the law, neuroscience changes nothing and everything.
Philosophical Transactions of the Royal Society London B 359, 177585.
Greeneld, S. 1999. How might the brain generate consciousness?. In Rose 1999.
Guthrie C., J. Rachlenski, and A. J. Wistrich. 2007. Blinking on the bench: how judges decide
cases. Cornell Law Review 93(1), 10141.
Hadley, R. F. 2010. The essential role of consciousness in mathematical cognition. Journal of
Consciousness Studies 17, 2746.
Hameroff, S., and S. Hagen. 2000. Quantum computation in the brain? Decoherence and bio-
logical feasibility. Paper presented at Toward a Science of Consciousness conference,
Tucson. URL = http://u.arizona.edu/~hameroff.
Hempel, C. G. 1965. Aspects of Scientic Explanation. London: Macmillan.
Hodgson, D. 1967. Consequences of Utilitarianism. Oxford: Oxford University Press.
Hodgson, D. 1991. The Mind Matters. Oxford: Oxford University Press.
Hodgson, D. 1996a. Nonlocality, local indeterminism, and consciousness. Ratio 9, 122.
Hodgson, D. 1996b. The easy problems aint so easy. Journal of Consciousness Studies 3, 6975.
Hodgson, D. 1999. Humes mistake. In Libet et al. 1999.
Hodgson, D. 2000. Guilty mind or guilty brain: criminal responsibility in the age of neurosci-
ence. Australian Law Journal 74, 66180.
Hodgson, D. 2001. Constraint, empowerment and guidance: a conjectural classication of
laws of nature. Philosophy 76, 34170.
Hodgson, D. 2002a. Quantum physics, consciousness, and free will. In Kane 2002a.
Hodgson, D. 2002b. Three tricks of consciousness. Journal of Consciousness Studies 9(12),
6588.
Hodgson, D. 2005. Responsibility and good reasons. Ohio State Journal of Criminal Law 2(2),
47183.
Hodgson, D. 2007a. Partly free. Times Literary Supplement, July 6, 1516.
Hodgson, D. 2007b. Making our own luck. Ratio 20, 27892.
Hofstadter, D. 1980. Gdel, Escher, Bach: An Eternal Golden Braid. Harmondsworth: Penguin.
Hofstadter, D. 1986. Metamagical Themas: Questing for the Essence of Mind and Pattern. Har-
mondsworth: Penguin.
Honderich, T. 1987. Mind, brain and self-conscious mind. In Blakemore and Greeneld 1987.
Horgan, T. 2001. Causal compatibilism and the exclusion problem. Theoria 16, 95116.
256 References
Horgan T., and J. Tienson. 1996. Connectionism and the Philosophy of Psychology. Cambridge,
MA: MIT Press.
Horgan T., and J. Tienson. 2006. Cognition needs syntax but not rules. In Stainton 2006.
Hume, D. 1748. An Enquiry Concerning Human Understanding.
Humphrey, N. 1983. Consciousness Regained. Oxford: Oxford University Press.
Jackson, F. 1982. Epiphenomenal qualia. Philosophical Quarterly 32, 12736.
Jibu, M., and K. Yasue. 1995. Quantum Brain Dynamics and Consciousness: An Introduction.
Amsterdam: Benjamin.
Kahneman, D., P. Slovic, and A. Tversky, eds. 1982. Judgement under Uncertainty: Heuristics and
Biases. Cambridge: Cambridge University Press.
Kane, R. 1996. The Signicance of Free Will. New York: Oxford University Press.
Kane, R., ed. 2002a. Oxford Handbook of Free Will. New York: Oxford University Press.
Kane, R. 2002b. Some neglected pathways in the free will labyrinth. In Kane 2002a.
Kauffman, S. 2009. Physics and Five Problems in the Philosophy of Mind. URL = http://
arXiv:0907.2494.
Kaye, A. 2007. The secret politics of compatibilist criminal law. Kansas Law Review 53,
365427.
Lacey, N. 2007. Space, time and function. Criminal Law and Philosophy 1, 23350.
Levy, N. 2008. Bad luck once again. Philosophy and Phenomenological Research 77(3), 74954.
Levy, N., and M. McKenna. 2009. Recent work on free will and responsibility. Philosophy
Compass 4(1), 96133.
Levy, S. 1993. Articial Life. Harmondsworth: Penguin.
Libet, B., A. Freeman, and K. Sutherland, eds. 1999. The Volitional Brain. Thorverton: Imprint
Academic.
Libet, B., C. Gleason, E. Wright, and D. Pearl. 1983. Time of conscious intention to act in rela-
tion to onset of cerebral activities (readiness potential): the unconscious initiation of a
freely voluntary act. Brain 106, 62342.
Libet, B., E. Wright, B. Feinstein, and D. Pearl. 1979. Subjective referral of the timing for a
conscious sensory experience. Brain 102, 193224.
Lycan, W. S., ed. 1990. Mind and Cognition. Oxford: Blackwell.
Marr, D. 1982. Vision. San Francisco: W. H. Freeman.
McCall, S., and E. Lowe. 2008. The determinists have run out of luckfor a good reason.
Philosophy and Phenomenological Research 77, 74548.
McKenna, M. 2009. Compatibilism. In The Stanford Encyclopedia of Philosophy,ed. E. Zalta.
URL=http://plato.stanford.edu/entries/compatibilism/.
Mele, A. 2009. Effective Intentions: The Power of Conscious Will. New York: Oxford University
Press.
Merali, Z. 2006. Free willyou only think you have it. New Scientist, May 6, 8.
Minsky, M. 1985. Science Digest, October.
Moore, M. 1997. Placing Blame: A General Theory of Criminal Law. Oxford: Oxford University
Press.
Morse, S. 2000. Rationality and responsibility. Southern California Law Review 74, 25168.
Murphy, N., and W. Brown. 2007. Did My Neurons Make Me Do It? New York: Oxford Univer-
sity Press.
Nagel, T. 1974. What is it like to be a bat? Philosophical Review 83, 43550.
Nagel, T. 1986. The View from Nowhere. New York: Oxford University Press.
Nagel, T. 1997. The Last Word. New York: Oxford University Press.
References 257
Nagra, J. 2005. Conways proof of the free will theorem. URL = http://www.cs.auckland.ac.
nz/~jas/one/freewill-theorem.html.
Newell, B. R., K. W. Wong, J. C. H. Cheung, and T. Rakow. 2009. Think, blink, or sleep on it:
The impact of modes of thought on conscious decision making. Quarterly Journal of Ex-
perimental Psychology 2, 70732.
Oaksford, M., and N. Chater. 1998. Rationality in an Uncertain World. Hove: Psychology
Press.
OConnor, T. 2005. Freedom with a human face. Midwest Studies in Philosophy 29, 20727.
Penrose, R. 1987. Minds, machines and mathematics. In Blakemore and Greeneld 1987.
Penrose, R. 1989. The Emperors New Mind. Oxford: Oxford University Press.
Penrose, R. 1994. Shadows of the Mind. Oxford: Oxford University Press.
Penrose, R. 1999. The Large, the Small and the Human Mind. Cambridge: Cambridge Univer-
sity Press.
Pereboom, D. 2001. Living without Free Will. Cambridge: Cambridge University Press.
Pereboom, D. 2002. Living without free will. In Kane 2002a.
Plantinga, A. 1993. Warrant and Proper Function. New York: Oxford University Press.
Popper, K. R. 1959. The Logic of Scientic Discovery. London: Hutchinson.
Poundstone, W. 1987. The Recursive Universe. Oxford: Oxford University Press.
Pribram, K., ed. 1998. Brain and Values: Is a Biological Science of Values Possible? Mawhah,
NJ: Erlbaum.
Putnam, H. 1975. Mind, Language and Reality. Cambridge: Cambridge University Press.
Putnam, H. 1979. Mathematics, Matter and Method. Cambridge: Cambridge University Press.
Putnam, H. 1981. Reason, Truth and History. Cambridge: Cambridge University Press.
Putnam, H. 1983. Realism and Reason. Cambridge: Cambridge University Press.
Regan, D. 1980. Utilitarianism and Co-operation. Oxford: Oxford University Press.
Robertson, B., and G. A. Vignaux. 1993. Probability: the logic of the law. Oxford Journal of
Legal Studies 13, 45778.
Rose, S., ed. 1999. From Brains to Consciousness. Penguin: London.
Sacks, O. 1986. The Man Who Mistook His Wife For a Hat. London: Picador.
Schnapf, J. L., and D. A. Baylor. 1987. How photoreceptor cells respond to light. Scientic
American, April, 3239.
Searle, J. 1984. Minds, Brains and Science. London: BBC.
Searle, J. 2001. Rationality in Action. Cambridge, MA: MIT Press.
Seife, C. 2000. Cold numbers unmake the quantum mind. Science 287(5454), February 4, 791.
Sharp, L. 1995. Cognitive heuristics and law. Bulletin of the Australian Society for Legal Philoso-
phy 20, 71100.
Sidgwick, H. 1874. 1981. 7th ed. The Methods of Ethics. Indianapolis: Hackett.
Smart, J. J. C. 1961. An Outline of a System of Utilitarian Ethics. Melbourne: Melbourne Univer-
sity Press.
Smart, J. J. C. 2005. Comments on Hodgson. Journal of Consciousness Studies 12(1), 5864.
Smilansky, S. 2002a. Free Will and Illusion. New York: Oxford University Press.
Smilansky, S. 2002b. Free will, fundamental dualism, and the centrality of illusion. In Kane
2002a.
Stainton, R., ed. 2006. Contemporary Debates in Cognitive Science. Malden, MA: Blackwell.
Stapp, H. 1993. Mind, Matter and Quantum Mechanics. Berlin: Springer.
Stapp, H. 1998. Pragmatic approach to consciousness. In Pribram 1998.
Stapp, H. 2000. Decoherence and quantum theory of mind: closing the gap between being
and knowing. URL = http://physics.lbl.gov/~stapp/stapples.
258 References
Stapp, H. 2009. Philosophy of Mind and the Problem of Free Will in the Light of Quantum
Mechanics. URL = http://arXiv:0805.0116v1.
Stich, S. 1990. The Fragmentation of Reason. Cambridge, MA: MIT Press.
Stoljar, D. 2006. Ignorance and Imagination. New York: Oxford University Press.
Strawson, G. 1986. Freedom and Belief. Oxford: Oxford University Press.
Strawson, G. 1998. Luck swallows everything. Times Literary Supplement, June 26, 810.
Strawson, G. 2002. The bounds of freedom. In Kane 2002a.
Strawson, P. 1982. Freedom and resentment. In Watson 1982.
Sudbery, A. 1986. Quantum Mechanics and the Particles of Nature. Cambridge: Cambridge
University Press.
t Hooft, G. 2001. How does God play dice? URL = http://arXiv:hep-th/0104219v1.
t Hooft, G. 2007. On the free-will postulate in quantum mechanics. URL = http://arXiv:quant-
ph/0701016v1.
Trakakis, N., and D. Cohen, eds. 2008. Essays on Free Will and Moral Responsibility. Newcastle:
Cambridge Scholars Publishing.
Turing, A. 1958. Computing machinery and intelligence. Mind 59, 43360.
van Inwagen, P. 1983. An Essay on Free Will. Oxford: Oxford University Press.
van Inwagen, P. 2000. Free will remains a mystery. Philosophical Perspectives 14, 119.
Watson, G., ed. 1982. Free Will. Oxford: Oxford University Press.
Wegner, D. 2002. The Illusion of Conscious Will. Cambridge, MA: MIT Press.
Werking, K. 2009. URL = http://people.wm.edu/~ktwerk/cards.doc.
Weyl, H. 1949. Philosophy of Mathematics and Natural Science. Princeton: Princeton University
Press.
References 259
This page intentionally left blank
Index
abduction 43 beliefs 7, 13, 32, 155, 214

Abelson, R. 72, 255 Bell, John 119, 123
Abraham and Isaac 251252 Bennett, Max 6365, 255
actions 155; see also self-forming actions biases 3031, 71, 151
act-utilitarianism 192195 Bible 243245, 251254
addiction as mitigation 225226 Big Bang 86, 129, 230, 233
aesthetic judgments 84, 96, 102107, 183, 198 Big Blue 75
agent-causation 8, 159, 162166 binding laws 68, 85
algorithmic level 4952, 66, 7374 binding problem 59, 6364, 138140, 145
algorithmic reasoning see formal reasoning Blackburn, Simon 20, 181185, 199, 255
alternatives 168172, 176177, 193194; see blame 3, 180, 199, 203, 206
also spectra of possibilities blind pitiless indifference 237239
analog systems 50 block universe 8, 114, 117, 127132
analogy 43, 101102, 108109 Bohm, David 120, 255
animals (non-human) 12, 18, 99102, 115, brains 18, 5051, 56, 133151
154, 158, 195 Brown, Warren 224, 257
apposite responses 80, 8384, 111112,
156158, 162, 179, 233, 235 capacity for reasonable moral
approximations 109; see also guidance (laws of) judgments 203
argument from design 133 cards (we are dealt) 165166
arrest 217 care 189
artefact (of human evolution and Cartesian theatre 138140
culture) 180, 185187, 197, 206 causal closure (of the physical) 73, 117
articial (consciousness, intelligence) 37, 62 causation see agent-causation, conscious
Aspect, Alain 119, 123, 255 causation, event-causation, physical
assembly language 50 causation, uncaused causes
assertions 21 ceteris paribus rules 38, 51, 7374, 9596
attention 137 chance see luck
attractors 143 change 1314, 131132
authority (of holy books, prophets, religious character 159, 175177; see also self-forming
institutions) 243244, 251254 actions
autistic twins 69 Chater, N. 51
axons 144 chess 75
chess-playing computers 156157, 172
Baars, Bernard 7677, 255 children 158159, 227228
Bayes Theorem 4348, 136, 249251 Christianity 243245, 251254
Baylor, D. A. 142 chunking trick 9295, 105
beauty 107109, 198, 235236, 238241 Churchland, Patricia 58, 255
261
Churchland, Paul 58, 255 Damasio, Antonio 110, 182, 255
Clark, Thomas 211 dangerousness 203, 215, 219, 221
Clarke, Randolph 162164, 255 Darwin, Charles 116, 133134
classical paradigm 5152 Davidson, Donald 94, 255
Cleland, C. E. 28, 255 Dawkins, Richard 9, 237239, 245, 255
climate of ideas 181 deception 194
cognitive science 134, 140141 decision-making 27, 3739, 49, 55, 6877, 84,
Cohen, J. 211, 221, 256 109112, 136, 153154, 158, 166, 169, 172; see
colour 6263, 67, 87, 138 also judicial decision-making
commitment 190, 194, 199, 201 defences to crime 213214
community practices 198201 degrees of responsibility 173176, 215
compatibilism 45, 114, 156, 166175, 223224 degrees of truth 2224, 35
compiler language 50 deliberation 137
compulsory treatment 209211, 225 dendrites 144
computational level see overt level Dennett, Daniel 11, 52, 58, 76, 138140, 255
computational rules 37, 4952, 55, 7376, 81, Descartes, Rene 10, 12, 16, 115, 133, 183
88, 91, 93, 136137, 153 deserving 4, 180, 203, 207, 213, 222
computers 18, 5051, 7172, 75, 8384, 9092; desires 3839, 100, 181182
see also chess-playing computers determinism 8, 114128, 167168, 171172,
concentration 137, 149 223224
conrmation 4142 deterrence 210, 219, 221223
connectionism 5152 De Wall, C. N. 71, 255
conscious causation 164 differences of opinion (as to morality) 185,
conscious contribution (to decision- 187, 192196
making) 7077, 79113, 135138, 144150, digestion 5657
153, 156158, 224 digital systems 5052
conscious control 165, 172173, 224 Dijksterhuis, A. 71, 255
conscious experiences 1218, 5674, 131132, 139 diminished responsibility 214215, 227
consciousness 7, 29, 5577, 135, 141, 230231, dimmer switch (of consciousness) 154
237240 direct argument 168, 172
consequence argument 167168, 172 discrimination against women 199
consequentialism see utilitarianism dissociative states 61
consequentialist purposes (of DNA 117
punishment) 210213 dual aspects 1617, 5558, 72, 80
conspiracy of nature 126128 dualism 1617, 134135
constraint (laws of) 234235 duress 214
content (of experience) 11, 57, 59, 62, 65 Dutton, Denis 105, 255
control (of conduct) 157, 162, 165, 169, 222;
see also conscious control Eccles, John 143145, 255
Conway, John 89, 121128, 255 efcacy (of consciousness) see conscious
Copeland, B. J. 75, 255 contribution (to decision-making)
correspondence theory (of truth) 21 effort 137
court cases 3435, 4548 Einstein, Albert 116, 119, 255
Crick, Francis 117 electromagnetism 116
crime as illness 34, 210211 embryos 195
criminal law 159, 206228 emergence 1617, 154
Crusades 244 emergentism 95
262 Index
emotion 26, 104, 108, 182 Gershwin, George 8284, 103
empowerment (laws of) 234235 gestalts 8, 59, 7985, 98112, 136137, 139, 153,
ethics 181187 156158, 237
ethnic cleansing 253 Gibbard, Allan 185, 256
event-causation 159, 164 gliders 89, 92, 94
evil 195, 197198, 238, 244246, 251253 global workspace 7677
evolution 6970, 75, 99102, 116, 133135, God 133, 177178, 196197, 233, 238240,
147, 187 243245, 251253
executive summaries 70, 96 Gdel, Kurt 182
exocytosis 144 Gdels incompleteness theorem 110
expectations 200201 Golden rule 189
extended detention 217218 Goodman, Nelson 4142, 256
external world 17, 5657, 6266, 232 goodness 107, 180, 191, 197198, 235236,
238241
fail-safe device 168170 Greene, J. 211, 221, 256
fallacies 2931, 53 Greeneld, Susan 154, 256
fallibility (of reason) 3031, 38, 69, 99, 101, 182 guidance (laws of) 235236
family 194 guilt 207
fatalism 4, 171 gut feelings 110
ne-tuning (of universe) 238 Guthrie, C. 71, 256
Fischer, John 165, 169171, 174, 255256
icker of freedom 170 Hacker, Peter 6365, 255
Fodor, Jerry 5052, 256 Hadley, Robert F. 111, 256
folk psychology 59, 226227 Hagen, S. 145
forces 115117 Haldane, J. B. S. 5, 5253
formal reasoning 37, 4950; see also Hameroff, Stuart 143, 145, 256
computational rules, logic, rules of handicaps 177
inference hard social conditions 174176, 215
Frankfurt, Harry 168, 256 harm 190, 201
Frankfurt-type cases 169170 Hart, Herbert 224
free will 8, 112, 153179, 180, 206, 230231 hatred 198
free will theorem 121128, 172 heat 56, 116
free wont 149 heaven 242, 254
Freud, Sigmund 134 Heisenberg uncertainty principle 118, 143144
friends 194 hell 242, 254
functionalism 67 Hempel, Carl 4142, 256
fundamentalism 230, 245246, 251254 Hiley, Basil 120, 255
fuzzy logic 136 Hodgson, David 38, 79, 118, 119, 120, 140, 154,
160, 194, 212, 224, 234235, 256
Game of Life 8992, 94, 108 Hofstadter, Douglas 50, 256
gap (between reasons and decisions) 38, 155 Holocaust 253
Gauguin, Paul 239240 holy books 243244; see also Bible
Gazzaniga, Michael 149151, 256 homunculus 61
genetic and environmental causes of Honderich, Ted 59, 256
crime 221222, 225 honesty 190, 193194, 198, 199, 201
genetic disadvantages 174176, 215 Horgan, Terry 38, 5051, 75, 9596, 117,
genocide 245 256257
Index 263
human beings 1518, 154 Kahneman, Daniel 29, 71, 257
human rights 212213 Kane, Robert 159164, 166, 257
Hume, David 11, 3839, 42, 100, Kant, Immanuel 17, 182
181183, 257 Kasparov, Gary 75
Humphrey, Nicholas 76, 257 Kauffman, Stuart 72, 257
hypnosis 61, 150 Kaye, A. 174, 257
hypotheses 3841 killing 195
kindness 190, 198
idealism 1617 Kochen, Simon 121128, 255
ignorance of the law 203 Kochen-Specker paradox 121122, 124
illusions 13, 131132
images 6465 Lacey, Nicola 224, 257
immorality (in Bible) 244246, 251254 language 7, 1115, 2124, 69, 107
imperatives (natural) 185192, 235236, 238 Laplace, Pierre 115116
implementational level 4952, 66, 7374 laws 7988, 9496, 198201; see also laws of
incommensurability (of reasons) 3839, nature, legal systems, psycho-physical
160, 224 laws, rules
inconclusive reasons 3839, 160, 224 laws of nature 37, 55, 7374, 8588, 93,
indeterminacy 118, 120, 130, 194 114117, 133137, 153, 155157, 167
indeterminism 118120, 141148, 153, 194 legal systems 8182, 201203, 206, 208209
induction 4043 Leibniz, Gottfried 183
inevitability 156 Les Demoiselles dAvignon 8384, 103,
informal reasoning see plausible reasoning 105106
information 7980, 82, 95, 98; see also levels of cognitive processes 4952, 6667,
informational content 7374, 96
informational content 80, 98 levels of description 5657
Inquisition 244 Levy, Neil 177, 257
insanity 35, 214 Levy, S. 89, 257
inscrutability (of reasons) 163165 Libet, Benjamin 148149, 257
intention 155, 214 life (and consciousness) 18, 154
intentionality 109 life-afrmation 190191
Islam 244245, 251 lifestyle choices 201
light cones 119, 129131
Jackson, Frank 80, 257 light speed 119, 123, 129130
Jesus 243, 253254 likelihood ratios 4447, 250
Jibu, M. 143, 257 locality of causation 118
jiggling tree 65 logic 181182; see also formal reasoning
joker analogy 165166 Lorentz transformation 116
Joshua 245246, 253 love 189, 198
Judaism 245246, 251254 Lowe, E. 177, 257
judgment 3739, 43, 6970, 84, 99102, 161; loyalty 190, 194
see also aesthetic judgments, luck 153, 163, 173179
decision-making
judicial decision-making 38, 71 machine language 50
justice 180, 184, 187, 190, 192, 196203, manipulation arguments 170171
208209, 213, 220222, 241242; see also many-worlds (interpretation of
unfairness QM) 120121
264 Index
Marr, David 4950, 66, 257 Oaksford, M. 51, 258
materialism see physicalism objective aspect 1617, 56
mathematical variables 81, 8587, 98 objects (of experience) 15, 6566
mathematics 110111, 181182 observer-participation 120121, 231234
Maxwell, James Clerk 116 OConnor, Timothy 163164, 258
McCall, Storrs 177, 257 odds 4447, 249
McKenna, Michael 159, 177, 257 ordinary human rationality 3036, 37, 96
meaning 21, 109 oughts 183184, 186; see also imperatives
measurement problem (of QM) 121 (natural)
measurements (of quantum overt level 4952, 66
systems) 141143
mechanisms for change 201202 pain 12, 59, 60, 62, 67, 71, 87, 139140
Mele, Alfred 148, 257 panicky metaphysics 171
melody 8283, 139 panpsychism 1618, 154
mens rea 213214 particles 23
mental abnormality 166, 203, 214215, 217, particularity (of experiences) 82, 85, 93, 95,
225227 103, 105107, 153
mental arithmetic 69, 72 particularity (of persons) 153
mental element of offences 213214 passions see desires
Merali, Z 122, 257 Passover 252, 255
microtubules 143 Pauling, Linus 117
Minsky, Marvin 70, 257 Penrose, Roger 52, 70, 72, 110, 143, 258
MNaghten rules 214, 227 Pereboom, Derk 218221, 258
monism 1617 personal identity 216
Moore, Michael 223, 257 physical causation 162, 164165
moral feelings 186, 188 physicalism 1617, 230
moral progress 183185 physical occurrences 56
Morse, Stephen 223224, 257 physical sciences 8, 114132
Moses 252 Picasso, Pablo 8384, 103, 105106
motivational feelings 7071, 96, 100, picture (worth a thousand words) 101102
158, 188 Plancks constant 118
Murphy, Nancey 224, 257 Plantinga, A. 53, 258
plausible reasoning 5, 7, 3754, 55, 69,
Nagel, Thomas 53, 57, 257 108111, 135, 182183, 186
Nagra, J. 124, 258 Popper, Karl 42, 258
natural kinds 22 Poundstone, W. 90, 258
necessity (as defence to crime) 214 practical jokes 190
negligence 209, 217 predictability 53
neural correlates (of consciousness) 6668, prefrontal cortex 158
74; see also psycho-physical laws premises (of reasoning) 37
neurons 140141 probability 4248, 137, 145, 147148, 160, 178,
neuroscience 8, 112, 133151, 182, 211, 225228 194, 249250
neutral monism 1617 programs 4951, 73, 156; see also
Newell, B. R. 71, 258 computational rules
Newton, Isaac 115117 promised land 253
non-locality 117121, 129130, 135136, promises 194
139140, 143, 146 proportionality (of punishment) 215216
Index 265
psychological present 139 responsibility 8, 112, 153179, 206228; see
psycho-physical laws 6768, 74, 79, 85 also ultimate responsibility
punishing the innocent 221 restraint (of offenders) 210
punishment 4, 180, 199, 203, 206228 retribution 8, 180, 206228
purpose (of existence) 240241 revenge 207208
Putnam, Hilary 24, 41, 4849, 67, 258 rightness (of conduct) 107, 180186, 222, 235236
Pylyshyn, Zenon 5052, 256 Robertson, B. 43, 258
rocks 18
qualia 55, 59, 6266, 76 rule-determined processes 5153, 7477, 84,
qualia laws 6768, 85 101, 104105, 224
qualia trick 9295, 105 rules 7991, 9496; see also computational
quantum entanglement 143, 145 rules, laws, rules of inference
quantum mechanics (QM) 117131, 139148, rules of inference (or of reasoning) 3739,
153, 160, 172, 182, 231234 4043, 55
Quantum Mind conference 145146 rule-worship 193194
quantum Zeno effect 142
quarantine 217, 219220 Sacks, Oliver 69, 258
Schnapf, J. L. 142, 258
randomness 34, 7475, 95, 114, 119120, science 112, 114132, 230231
141142, 147148, 172 scientic method 6, 28, 115, 121122, 127128
rationalisation 150 Searle, John 38, 56, 91, 154155, 258
rationality 5, 6, 2636, 158, 169, 182183, 186, seeing 6466
188, 236, 237; see also ordinary human Seife, C. 145, 258
rationality selection (between alternatives) 161162
Ravizza, M. 169, 256 self-defence 35, 214
reality 2025, 109, 129130, 180, 186, 197 self-forming actions 159160, 166, 177
reason (as slave of the passions) 181183; see self-organisation 154
also rationality semi-compatibilism 169
reasons 37, 155, 163164, 176, 187, 197, sexual morality 201
223224; see also inconclusive reasons Sharp, L. 30, 258
reasons-responsiveness 155, 169 Sidgwick, Henry 193, 258
reduction (of the quantum state) 121, signicance (of information) 102, 107109
141143 sleep-on-it phenomenon 70
reform 211, 222; see also rehabilitation Smart, J. J. C. 193, 234, 258
Regan, Donald 192, 258 Smilansky, Simon 162, 223, 258
rehabilitation 222, 225 soft constraints 38, 51, 9596
relativistic views (of truth) 2426 solipsism 10
relativity theory 114, 116117, 119, 122123, somatic markers see gut feelings
126, 128131 space-like separation 118119, 123127, 129130
religious belief 9, 230, 236246 space-time 114, 118119
reparation 207208 spectra of possibilities 8081, 96, 117, 130,
replay argument 177179 135, 137, 157158, 165, 176177
representations 50 spin (of particles) 122126
respect 189, 200 Spinoza, B. 183
responses to wrong conduct 206210; see split brains 61, 150151
also revenge, retribution, reparation, treat- stability 200201
ment (non-punitive) Stapp, Henry 142145, 258259
266 Index
State compulsion 212, 216218 ultimate responsibility 159162, 166, 176
statistics (QM) 142, 147148 uncaused causes 164166
steering a course 137, 157158, 178 unconscious information-processing 6877,
Stich, Stephen 24, 3234, 259 98, 112, 137
Stoljar, Daniel 57, 259 unconscious motivation 134
Strawson, Galen 173177, 224, 259 underdetermination of theory by data 41
Strawson, Peter 171, 259 understanding 72, 110
subject (of experience) 1011, 5962, 7985, unfairness 200201
92, 93, 95, 145 unity (of conscious experiences) 55, 59,
subjective aspect 1617, 56 6264, 76, 138140; see also binding
subjectivity 55, 57, 5962, 76; see also subject problem, chunking trick
(of experience) user-friendly accounts (of brain
Sudbery, A. 142, 259 processes) 76
suffering 198 utilitarianism 192, 194; see also
supervenience 67 act-utiliarianism
survival (of death) 61, 241242
synapses 144 value-embedded universe 237241
value judgments 8, 180203
Tamarin, George 245246 values 230231, 235236, 238; see also beauty,
Tegmark, Max 145 goodness, rightness, truth
template for action 143 van Inwagen, Peter 167, 259
tendencies 176177 vengeance see revenge
terrorism 245, 252 vicious circle 53
The Man I Love 8283, 103 Vignaux, G. A. 43
tHooft, Gerard 120, 128, 259 virtual reality 64
Tienson, John 38, 5051, 75, 9596, 257 voluntary conduct 214
time 1314, 2829, 128132 von Neumann, John 120
tolerance 190, 196
torn decisions 160161 Wagner, Richard 8384, 103, 106
torture 186 Watson, James 117
transfer of non-responsibility principle 168, 174 waves 23
transfer of powerlessness principle 167168 Wegner, Daniel 150151, 259
treating persons as ends 189, 213 Werking, Kip 165, 259
treatment (non-punitive) 207211; see also Weyl, Hermann 131, 259
compulsory treatment where are we going? 240242
tricks of consciousness 92; see also chunking where do we come from? 230231,
trick, qualia trick 239240
Tristan und Isolde 8384, 103, 106 will 154159; see also free will
trust 194 wisdom 198
truth 7, 2026, 3236, 107109, 180, 186, 197, Wittgenstein, Ludwig 24, 3233, 63, 182
235236, 239241 wrong conduct see responses to wrong
Turing, Alan 75, 182, 259 conduct
Tversky, Amos 29, 71, 257 wrongness 180196, 236
two-slit experiment 120
types 81, 85, 86, 92, 98 Yasue, K. 143, 257
Index 267

(David Hodgson) Rationality + Consciousness Free

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(David Hodgson) Rationality + Consciousness Free

Uploaded by

Copyright:

Available Formats

Rationality

Recently Published in the Series

Supersizing the Mind

Perception, Hallucination, and Illusion

Phenomenal Concepts and Phenomenal Knowledge

The Character of Consciousness

The Contents of Visual Experience

Consciousness and the Prospects of Physicalism

The Sources of Intentionality

Rationality + Consciousness = Free Will

Copyright 2012 Oxford University Press

Published by Oxford University Press, Inc.

Library of Congress Cataloging-in-Publication Data

Printed in the United States of America

2. Truth and Rationality 20

4. Consciousness and Decision-Making 55

5. Gestalts and Rules 79

6. How Gestalts Promote Rationality 98

7. Science and Determinism 114

8. Neuroscience and Conscious Choice 133

10. Value Judgments 180

11. Responsibility and Retribution 206

12. The Big Picture 230

TH E PROB LE M OF FR E E WI LL AN D R E S PON S I B I LITY

4 rationality + consciousness = free will

6 rationality + consciousness = free will

8 rationality + consciousness = free will

CAN I B E CE RTAI N THAT I EXI ST?

1.1. Conscious experiences occur.

12 rationality + consciousness = free will

1.2. Some conscious experiences change.

14 rationality + consciousness = free will

EXPE R I E NCE S AN D TH E EXTE R NAL WOR LD

Human beings have many characteristics in common, in terms of physical constitution,

16 rationality + consciousness = free will

Idealism Substance Dualism

First, and contrary to some versions of idealism, Im condent there is an external

1.1. Conscious experiences occur.

18 rationality + consciousness = free will

Now meaning is not always straightforward, so neither is truth. Indeed, meaning is

In Sydney this morning the sun rose at 6.30 am.

Truth and Rationality 21

22 rationality + consciousness = free will

All waves are periodic processes extended in space.

Truth and Rationality 23

extreme version, in Stich (1990), is considered later in this chapter.

24 rationality + consciousness = free will

Truth and Rationality 25

I now consider rationality.

2.5. Rationality is to be understood as extending to all those capabilities that contribute

26 rationality + consciousness = free will

2.7. My condence in my beliefs depends ultimately on my condence in my own ratio-

Truth and Rationality 27

However, it is sometimes reasonable to question what is asserted by scientists, espe-

28 rationality + consciousness = free will

Truth and Rationality 29

30 rationality + consciousness = free will

Truth and Rationality 31

Psychological demonstrations of fallacies in human reasoning have also been opposed

32 rationality + consciousness = free will

Stich goes on to conclude that truth is not to be taken seriously as a cognitive

Truth and Rationality 33

A LEGAL EXAM PLE