GARVEY, James. Companion - Philosophy - Mind - Continuum PDF

The Continuum Companion to Philosophy of Mind
The Continuum Companions series is a major series of single volume companions

to key research fields in the humanities aimed at postgraduate students,
scholars and libraries. Each companion oers a comprehensive reference
resource giving an overview of key topics, research areas, new directions and
a manageable guide to beginning or developing research in the field. A dis-
tinctive feature of the series is that each companion provides practical guidance
on advanced study and research in the field, including research methods and
subject-specific resources.
The Continuum Companion to Continental Philosophy, edited by John Mullarkey

and Beth Lord
The Continuum Companion to Locke, edited by S.-J. Savonious-Wroth, Paul
Schuurman and Jonathan Walmsley
Forthcoming in Philosophy:
The Continuum Companion to Aesthetics, edited by Anna Christina Ribeiro
The Continuum Companion to Berkeley, edited by Bertil Belfrage and Richard
Brook
The Continuum Companion to Epistemology, edited by Andrew Cullison
The Continuum Companion to Ethics, edited by Christian Miller
The Continuum Companion to Existentialism, edited by Jack Reynolds, Felicity
Joseph and Ashley Woodward
The Continuum Companion to Hegel, edited by Allegra de Laurentiis and Jerey
Edwards
The Continuum Companion to Hobbes, edited by S. A. Lloyd
The Continuum Companion to Hume, edited by Alan Bailey and Dan OBrien
The Continuum Companion to Kant, edited by Gary Banham, Nigel Hems and
Dennis Schulting
The Continuum Companion to Leibniz, edited by Brendan Look
The Continuum Companion to Metaphysics, edited by Robert Barnard and Neil
A. Manson
The Continuum Companion to Political Philosophy, edited by Andrew Fiala and
Ma Matravers
The Continuum Companion to Plato, edited by Gerald A. Press
The Continuum Companion to Pragmatism, edited by Sami Pihlstrm
The Continuum Companion to Socrates, edited by John Bussanich and Nicholas
D. Smith
The Continuum Companion to Spinoza, edited by Wiep van Bunge
The Continuum Companion to Philosophy of Language, edited by Manuel
Garcia-Carpintero and Max Kolbel
The Continuum Companion to the Philosophy of Science, edited by Steven French
and Juha Saatsi
The Continuum
Companion to
Philosophy of Mind
Edited by
James Garvey
Continuum International Publishing Group
The Tower Building 80 Maiden Lane
11 York Road Suite 704
London SE1 7NX New York, NY 10038
www.continuumbooks.com
James Garvey and Contributors, 2011
All rights reserved. No part of this publication may be reproduced or transmi ed

in any form or by any means, electronic or mechanical, including photocopying,
recording, or any information storage or retrieval system, without prior
permission in writing from the publishers.
British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.
ISBN:HB:0826431887
978-0-8264-3188-2
Library of Congress Cataloging-in-Publication Data

The Continuum companion to philosophy of mind / edited by James Garvey.
p. cm.
Includes bibliographical references.
ISBN:978-0-8264-3188-2
1.Philosophy of mind. I.Garvey, James, 1967-
BD418.3.C6565 2011
128'.2dc22 2010036913
Typeset by Newgen Imaging Systems Pvt Ltd, Chennai, India

Printed and bound in Great Britain
This book is for V
This page intentionally left blank
Contents
Acknowledgements ix
Contributors xi
How to Use This Book xv
Introduction xix
1Problems, Questions and Concepts in the Philosophy of Mind 1

Ian Ravenscro
2Consciousness 35
Daniel D. Hu o
3The Mark of the Mental 54
Fred Adams and Steve Beighley
4Substance Dualism 73
T. J. Mawson
5Physicalism 92
Barbara Montero
6Folk Psychology and Scientific Psychology 102
Barry C. Smith
7Internalism and Externalism in Mind 133
Sarah Sawyer
8The Philosophies of Cognitive Science 151
Margaret A. Boden
9Representation 171
Georges Rey
10Mental Causation 190
Neil Campbell
11Personal Identity 203
E. J. Lowe
12Embodied Cognition and the Extended Mind 220
Michael Wheeler
vii
Contents
13Current Issues in the Philosophy of Mind 239

Paul Noordhof
Glossary 280
Chronology 319
Research Resources 327
Notes 330
Bibliography 341
Index 375
viii
Acknowledgements
I relied on a large number of people for help in pu ing this volume together.
First and most importantly, I am very grateful to all of the contributors. Some
provided advice and read bits of the manuscript, and their suggestions always
resulted in improvements. I am particularly in the debt of those who saved
me and stepped in to do some last-minute writing. Each delivered good, solid
philosophy in record time. You know who you are. So thanks are owed to:
Fred Adams, Steve Beighley, Margaret A. Boden, Mark Cain, Neil Campbell,
Adam Ferner, Daniel D. Hu o, Dale Jacque e, E. J. Lowe, Tim Mawson,
Barbara Montero, Isabella Muzio, Paul Noordhof, Dan OBrien, Dimitris
Platchias, Ian Ravenscro , Georges Rey, Constantine Sandis, Sarah Sawyer,
Barry C. Smith, and Michael Wheeler.
David Avital, Carly Bareham, Tom Crick and Sarah Douglas at Continuum
are all very nearly equally excellent thank you all for your help and for
pu ing up with a lot. Thanks are also owed to comrades at Crisis, Kim Hastilow,
Ted Honderich, London Street Rescue, Justin Lynas, Anthony OHear, and my
associates at UCLU Jitsu.
I am particularly grateful to Judy Garvey for her unwavering support.
ix
Contributors
Fred Adams is Professor of Philosophy, and Chair and Professor of Linguistics

and Cognitive Science at the University of Delaware. He has published over 100
articles or reviews in philosophy and cognitive science. He is co-author (with
Ken Aizawa) of The Bounds of Cognition (2008), co-editor (with Leemon McHenry)
of Reflections on Philosophy (1993), and is editor of Ethical Issues for the 21st
Century (2005) and editor of Ethical Issues in the Life Sciences (2007).
Steve Beighley is on the Neurophilosophy Track at Georgia State University.

His research focuses primarily on animal minds, specifically on primitive
communication and emotions.
Margaret A. Boden is a Fellow of The British Academy, and Research Professor

of Cognitive Science at the University of Sussex. She is the author of The Creative
Mind: Myths and Mechanisms (second edition, 2004), Mind as Machine: A History
of Cognitive Science (2006), and Creativity and Art: Three Roads to Surprise (2010).
Her earlier books included Purposive Explanation in Psychology (1972) and
Artificial Intelligence and Natural Man (1977). She has two children and four
grandchildren, and lives in Brighton.
Mark Cain is Assistant Head of the Department of Religion and Philosophy,

Oxford Brookes University. He is author of Fodor: Mind, Language and Philosophy
(Polity, 2002) and The Philosophy of Cognitive Science (Polity, forthcoming).
Neil Campbell is Professor of Philosophy at Wilfrid Laurier University. He

is the author of Mental Causation: A Non-Reductive Approach (2008) and A Brief
Introduction to the Philosophy of Mind (2005). He also edited Mental Causation
and the Metaphysics of Mind (2003) and Freedom, Determinism, and Responsibility
(2003). He has published over 25 articles in philosophy journals.
Adam Ferner works for the Royal Institute of Philosophy and is studying for a
Ph.D. in Philosophy at Birkbeck College, University of London. His thesis is on
animalism, animals, and artefacts.
Daniel D. Huo is Professor of Philosophical Psychology at the University of

Hertfordshire. He is the author of The Presence of Mind (1999), Beyond Physicalism
xi
Contributors
(2000), Wi genstein and the End of Philosophy (2006) and Folk Psychological Narra-
tives (2008). He is also the editor of Narrative and Understanding Persons (2007),
Narrative and Folk Psychology (2009) and co-editor of Folk Psychology Re-Assessed
(2007). A special yearbook issue of Consciousness and Emotion, entitled Radical
Enactivism, which focuses on his philosophy of intentionality, phenomenology
and narrative, was published in 2006.
Dale Jacquee is Lehrstuhl ordentlicher Professur fr Philosophie, Schwerpunkt

theoretische Philosophie (Senior Professorial Chair in Theoretical Philosophy),
at Universitt Bern, Switzerland. He is the author of numerous articles on logic,
metaphysics, philosophy of mind, and aesthetics, and has recently published
Philosophy of Mind: The Metaphysics of Consciousness (2009), Ontology (2002), On
Boole (2002), David Humes Critique of Infinity (2001), and Wi gensteins Thought
in Transition (1998). He has edited the Blackwell Companion to Philosophical Logic
(2002, 2006), Cambridge University Press Companion to Brentano (2004), and for
North-Holland (Elsevier) the volume on Philosophy of Logic in the Handbook of
the Philosophy of Science series (2007).
E. J. Lowe is Professor of Philosophy at Durham University, UK. His books

include Locke on Human Understanding (Routledge, 1995), Subjects of Experience
(Cambridge University Press, 1996), The Possibility of Metaphysics (Oxford
University Press, 1998), An Introduction to the Philosophy of Mind (Cambridge
University Press, 2000), A Survey of Metaphysics (Oxford University Press,
2002), The Four-Category Ontology (Oxford University Press, 2006), Personal
Agency (Oxford University Press, 2008), and More Kinds of Being (Wiley-
Blackwell, 2009).
T. J. Mawson is Fellow and Tutor in Philosophy at St Peters College, Oxford.

He is author of Belief in God (Oxford University Press, 2005) and Free Will
(Continuum, forthcoming). He keeps the list of publications on his website,
h p://www.philosophy.ox.ac.uk/members/tim_mawson, more or less up to
date.
Barbara Montero is Associate Professor of Philosophy at the City University of

New York at the College of Staten Island and The Graduate Center. She is the
recipient of fellowships from the National Endowment for Humanities and the
American Council of Learned Societies. Most of her research has focused
on one or the other of two dierent notions of body: body as the physical or
material basis of everything, and body as the moving, breathing, flesh and
blood instrument that we use when we run, walk, or dance. She has published
numerous articles is currently writing a book, to be published by Oxford
University Press, on expertise and awareness.
xii
Contributors
Isabella Muzio has a Ph.D. in philosophy from University College London and
teaches philosophy at The Open University. She is writing a book about our
knowledge of our own emotions.
Paul Noordhof is Anniversary Professor of Philosophy at the University of

York. His main work on causation, mental causation, self-deception, belief and
the will, imagination, and consciousness has been published in Philosophy and
Phenomenological Research, Mind, Mind and Language, Proceedings of the Aristote-
lian Society, Australasian Journal of Philosophy, Synthese and Analysis. He is also
reviews editor of Mind and joint editor of a collection of papers entitled Cause
and Chance (with Phil Dowe) published in the Routledge International Library
of Philosophy Series. He received a three year Major Leverhulme Research
Fellowship in 2006 for research into the connection between consciousness
and representation for a book provisionally entitled Cement of the Mind (under
contract with Oxford University Press). The full presentation of his work in
causation will occur in another book, A Variety of Causes (also under contract
with Oxford University Press).
Dan OBrien is Research Fellow at Oxford Brookes University, Honorary

Research Fellow at Birmingham University, and Associate Lecturer with the
Open University. His publications include An Introduction to the Theory of
Knowledge (Polity, 2006), Humes Enquiry Concerning Human Understanding:
Readers Guide (Continuum, 2006), Gardening and Philosophy: Cultivating Wisdom
(Blackwell, 2010).
Dimitris Platchias is a lecturer at the University of Glasgow. He is the author of

Phenomenal Consciousness: Understanding the Relation between Neural Processes and
Experience (Acumen, 2010) and has co-edited Representationalism: Contemporary
Readings (with Fiona Macpherson, MIT Press, forthcoming) and Hallucination
(with Fiona Macpherson, MIT Press, forthcoming).
Ian Ravenscro is Associate Professor of Philosophy at Flinders University.

He is co-author (with Gregory Currie) of Recreative Minds (Oxford University
Press, 2002), author of Philosophy of Mind (Oxford University Press, 2005) and
editor of Minds, Ethics, and Conditionals (Oxford University Press, 2009).
Georges Rey is Professor of Philosophy at the University of Maryland at

College Park. He has wri en numerous articles in the philosophy of psycho-
logy, as well as a book, Contemporary Philosophy of Mind: A Contentiously Classical
Approach, and was an editor (with Barry Loewer) of Meaning in Mind: Fodor and
His Critics. Much of his work is available at his website: h p://sites.google.com/
site/georgesrey.
xiii
Contributors
Constantine Sandis is a Senior Lecturer in Philosophy at Oxford Brookes

University and New York University in London. He is the editor of New Essays
on the Explanation of Action (Palgrave Macmillan) and (with Timothy O Connor)
A Companion to the Philosophy of Action (Wiley-Blackwell). His is currently
completing a monograph called The Things We Do and Why We Do Them.
Sarah Sawyer is Senior Lecturer at the University of Sussex. Her research

interests are based around the nature of and the connections between thought,
language and knowledge. Her published work primarily concerns content
externalism, justification, fiction and singular thought. She is also the editor of
New Waves in Philosophy of Language (2010).
Barry C. Smith is Professor of Philosophy and Director of the Institute of

Philosophy at the School of Advanced Study, University of London. He has
been Head of the School of Philosophy at Birkbeck College and held visiting
positions at Simon Fraser University, the University of Califonia at Berkeley
and the Ecole Normale Superieure in Paris. He is the editor of the Oxford
Handbook of Philosophy of Language (with Ernest Lepore), Knowing Our Own
Minds (with Crispin Wright and Cynthia Macdonald) and Questions of Taste.
Michael Wheeler is Professor of Philosophy at the University of Stirling. Prior

to joining the Stirling Department in 2004, he held teaching and research posts
at the Universities of Dundee, Oxford, and Stirling (a previous appointment).
His doctoral work was carried out at the University of Sussex. His primary
research interests are in philosophy of science (especially cognitive science,
psychology, biology, artificial intelligence and artificial life) and philosophy
of mind. He also works on Heidegger. His book, Reconstructing the Cognitive
World: The Next Step, was published by MIT Press in 2005.
xiv
How to Use This Book
To help you find your way into this book, here is a short overview of each of
the main sections.
Introduction
Here you will find a very short take on the recent history of contemporary
philosophy of mind, its Cartesian roots, and a few words about recent con-
ceptual shi s, as well as a consideration of the general tone of the volume and
the choice of contents. The Introduction also contains a short summary of each
of the books main chapters, and this might serve as a springboard into the
rest of the book.
Problems, Questions and Concepts in the Philosophy of Mind
This overview by Ian Ravensco contains a mix of introductory and advanced

material which fleshes out the basic problems and questions in the philosophy
of mind as well as the answers currently in favour, alongside the relevant
objections and replies. A beginning student might read this to get a feel for the
philosophy of mind generally. It could also serve as a refresher for advanced
students and researchers.
Current Research
Here you will find eleven original essays wri en by experts in the field. Not
only do they provide overviews of large sub-topics in the philosophy of mind,
the authors take a stand and argue for their own positions. Its this combination
which makes the essays of interest to researchers at dierent levels. Again, there
is an overview of each chapter in the Introduction.
Current Issues in the Philosophy of Mind
This section, wri en by Paul Noordhof, aims to follow up from the previous
essays with reflection on cu ing-edge thinking in the philosophy of mind. This
xv
is not soothsaying. Instead, Noordhof provides an overview and explanation

of very recent trends, new questions just being asked by contemporary
researchers, and recent empirical findings which might shape the course of
future enquiry.
Glossary
This part of the volume includes short definitions and longer treatments
wri en by experts in various subfields of the philosophy of mind. Each entry
concludes with references to works which might be consulted for further infor-
mation. The aim of course is to provide a good resource for someone who
is reading this companion but might not be familiar with one or more of the
technical terms used by the authors. It also might help someone just ge ing on
with the study of the philosophy of mind who encounters a dicult word or
concept in a book or article some of the terms do not appear elsewhere in
this book. The glossary contains a number of clever distillations of dicult
concepts, so it might also be read independently, just for interest.
Chronology
This timeline is a reference for students and researchers who require speedy
access to a date or title as well as a bit of context for it. If you encounter an
unfamiliar philosopher, book or paper, have a scan of the chronology to get a
feel for where the topic fits in the history of mind. The entries up to 1950 contain
a sentence or two of explanation. Beyond that date we are too close in time to
be sure of the relevance, meaning or impact of a paper or book.
Research Resources
This section includes a list of resources on the Web, associations and research
centres, and periodicals devoted to the philosophy of mind anything of a
practical nature which might help a budding or seasoned researcher.
Bibliography
The bibliography, compiled by Adam Ferner, includes details of all the

works mentioned in this book. Additional books and articles not specifically
mentioned by the contributors but which might nevertheless be of interest to
those studying the philosophy of mind are included as well.
xvi
What Now?
You might acclimatize yourself by having a look at the Introduction in

particular, consult the summary of the main contributions to this volume. Old
hands can then dip in and out of the essays according to interest. If you are new
to the philosophy of mind or just want to have an overview of it with a fresh
pair of eyes, I strongly suggest that you read Ravenscro s introductory essay
and move on from there.
xvii
Introduction
The introduction to nearly every companion volume to a philosophical topic

does three things. First, it tells the reader that the subject ma er in question is
undergoing a renewal or revival of some sort. For one reason or another its a
hot topic, and conceptually the good times are rolling. Second, although many
such companions exist, the introduction claims that this one genuinely serves
a new and important purpose. Third, the editor confides that a much larger
number of topics would have been included, but, alas, certain hard decisions
were forced by fierce publishers, as well as the unyielding constraints of time
and space. What remains is a fine but by no means ideal table of contents. In
no other genre are cheers so quickly followed by sober apologies. I promised
myself that I would never stoop to those three clichs, but now I cant help it.
They actually fit.
Renewal
It turns out that the philosophy of mind really is undergoing a remarkably

lively renewal. We have already had one rebirth in the seventeenth century,
and were probably somewhere in the midst of another one now. Early philo-
sophical reflection on the mind was by turns dominated by Greek thought or
Roman thought or Scholastic thought until Descartes shook us free of all of
that and entangled us in something else entirely.
One can actually witness the shi in Meditations on First Philosophy, as
Descartes quite plainly distances himself from Aristotle and the rest. Here is a
representative passage: What then did I formerly think I was? Undoubtedly
I judged that I was a man. But what is a man? Shall I say a rational animal?
Assuredly not . . .. From Descartes new point of view, Aristotle gets it almost
entirely wrong. Theres nothing essentially animal about what he is, and some-
thing more than rational is needed to give a clear description of it. Descartes
characterizes himself as a thinking thing: a thing that doubts, understands,
arms, denies, wills, refuses; that imagines also, and perceives. Its a long and
durable list, and with it Descartes derailed hundreds of years of thought about
what it is to be human, simultaneously turning philosophical a ention and
fashion to substance dualism. If modern reflection on mind was born with
Descartes, contemporary philosophy of mind was born relatively recently in
xix
Introduction
an eort to find something be er than dualism. The philosophy of mind has

had a second renaissance, and we find ourselves in it now.
Some argue that one specific line of thought changed everything. Ryle and
those who followed him might have been the ones to usher in a new era of
reflection on the mind simply by finally and explicitly renouncing mind-body
dualism. Certainly early scientific psychology aimed to avoid dubious, unreli-
able talk of private mental states. The wholesale rejection of the so-called
Cartesian ghost in the machine might have stimulated the philosophy of mind
in a backhanded way. The implausibility of both dualism and behaviourism
made room for fresh thinking about the reality of mental life, or so the story
goes. Or maybe Chomskys work on language acquisition and the unconscious
complexities behind it created the space needed for new reflection on the mind.
Perhaps Turings work and the promise of artificial intelligence nudged us into
fresh thoughts about mental life. Maybe the bare possibility of understanding
cognition well enough to create or just model it was enough to shake us free of
Descartes. Possibly a combination of these and other thoughts did it. Wherever
you point the finger, something started happening in the 1950s. The ground
really did shi . There was room for philosophers of mind to make a break from
Descartes at last.
Others are less definite about the precise origins of contemporary thoughts
about the mind. Theyll say that Ryle, Chomsky, Turing and others were just
some of the many a ershocks caused by a deeper, amorphous conceptual
reorganization which got its start a very long time ago. New thinking about
the mind in the last half century or so has had more to do with our growing
scientific picture of the universe and our place in it than something as minor
as the rise of behaviourism or even computing. Perhaps Descartes could take
seriously the notion of a soul existing apart from the body an unextended
thing located precisely nowhere yet somehow running the show but we no
longer can. A er all, Descartes grew up in a world which had only just started
to assimilate the notion that Earth was not the centre of the universe.
As our understanding of such things as biology, evolution, psychiatry and
physics expanded, the old view of mind just didnt fit with the rest. How do you
bolt a soul onto an evolved organism? What use is a thinking substance to
theorizing about our psychological states? Where do you put soul stu when
your ontological inventory seems complete with ma er and energy?
Nearby thinkers will point to the catalyst of the discoveries in specific
sciences. They might say that whats energized the rebirth in reflection on the
mind are new thoughts and facts turned up by computing, neuroscience,
experimental psychology, and on and on maybe even something as unlikely
as quantum mechanics. Contemporary philosophy of mind has always made
more use of empirical discoveries than most of its philosophical cousins. Meta-
physicians are largely unmoved by whatever pops up in particle accelerators.
xx
Introduction
Philosophers of mind, however, are as likely to march forward with thought

experiments as they are with empirical experiments citing psychopathologies,
split-brain patients, mirror neurons, childhood development studies, neuro-
science, facts about computing, and on and on. No one doubts that contem-
porary philosophy of mind is enlivened by a connection to empirical enquiry,
but some argue that it owes its status as a hot philosophical topic to the need
for careful reflection on the constant supply of new discoveries turned up by
the study of both the brain and the body.
A Useful, Argumentative Companion
Reasonable people can argue about what caused the revival in reflection on
mind, but few really doubt that something new is underway. Philosophers have
turned their a ention to a huge number of new or nearly new problems in the
past few decades, and by most lights they are still finding new topics, new
arguments, new theories and distinctions, maybe even new answers. Its no
surprise that there is a large number of books on the philosophy of mind
even many valuable companion volumes but this one really does aim to do
something at least a li le dierent from the rest. For one thing, all of the com-
panions in the Continuum Series have the slightly unusual aim of being of use
to people operating at many levels of enquiry. They arent just for beginning
students or old hands. The hope is that theyll serve as useful desk references
for people at many dierent stages, from classroom work to highly technical
reflection and research. To this end you will find a number of sections designed
to help beginning students hit the ground running as well as sections to assist
advanced researchers move on through further thoughts on various subjects.
For more details, have a look at the section called How to Use This Book.
However, what really marks this volume out against many other compan-
ions or supplements to the philosophy of mind, I think, is that it is a piece of
the philosophy of mind, not just a report on it. Thats been the intention from
the start. The contributors not only scout the relevant territory with a view
to ge ing the reader up to speed with who said what, they also pitch their
essays at fellow researchers. They say what they think and argue for their own
views or against claims made by others. The pleasing result is a companion
which doesnt just nod in agreement or politely show you around. Instead,
youll read philosophers ge ing on with the job, doing philosophy, arguing,
jostling, persuading, objecting, judging and generally trying to get the truth
into clear view. The result might not be consistent or coherent in fact Im
sure theres disagreement somewhere among all the contributors but their
thoughts are interesting and worth reading, and their work will certainly
stimulate further thinking.
xxi
Introduction
Topics
This book doesnt cover everything or even almost everything that ma ers in
the philosophy of mind. I used to cringe a bit when I read similar apologies for
an obvious truth in other books: if only we had more space to cover this and a
bit more time to consider that. No book could possibly include everything
which ma ers, so why make excuses? Never again. Editorial decisions are almost
always painful judgements one thing ruled in and twelve perfectly respect-
able and interesting and important things ruled out. In most cases Ive been
guided by something more than my own feeble thoughts and limited experience:
Ive made repeated use of a number of helpful and patient advisors who
know much more about the mind than I do. The topics covered in this book
from the main essays to the glossary entries all made it in at the expense of
something else based on the closest thing to a consensus I could get from
many people. Thats not to say that I hereby abdicate responsibility if the
mix of topics could have been be er, thats entirely my fault. I also gave the
contributors a free hand to approach topics more or less as they liked, given
the general constraints of the series, and maybe firmer editorial guidance
would have resulted in a more comprehensive or balanced volume. Then again,
giving experts the opportunity to scout the territory as they see it has many
recommendations too.
I also had to consider how the contributions might hang together, given who
agreed to write about what, and that means that some topics could only appear
within others, even though they might deserve star billing alongside a dierent
mix of papers. So, for example, theres nothing specific on qualia or the first
person or intentionality, but each one of these important subjects turns up
again and again in dierent papers. I also thought it might be good to consider
not just large, abstract questions in the philosophy of mind but also some very
specific, narrow problems. The hope is to convey something of the sweep of the
philosophy of mind as it is really practiced. To give you a feel for the topics
under consideration, well now glance at each of the main papers.
Overview of the Contributions
There are eleven central essays in this book. Some take up broad topics like the
nature of consciousness and the mark of the mental. Others deal with particular
theories of mind, such as physicalism and dualism. Still others examine specific
sub-topics such as mental causation and personal identity. The central essays
are bookended by two more general pieces. The first, by Ian Ravenscro , sets
the tone with an introductory overview of the philosophy of mind. The last, by
Paul Noordhof, considers some of the possible future directions the philosophy
xxii
Introduction
of mind might take. Well begin with Problems and Questions in the Philo-
sophy of Mind by Ian Ravenscro and briefly summarize the rest.
Problems, Questions, and Concepts in the Philosophy of Mind
Ian Ravenscroft
Ravenscro identifies four broad areas of research in the philosophy of mind as
it is practiced today: metaphysical issues concerning the relation between the
mental and the physical, epistemological questions about our knowledge of our
own minds and the minds of others, themes associated with the influence of the
behavioural and cognitive sciences, and methodological issues concerning the
right approach to the study of mind. His main focus is the philosophy of minds
main focus: the mind-body problem. He works through various theories of
mind: substance dualism, reductive and non-reductive physicalism, the super-
venience relation, eliminativism, and instrumentalism. Closely connected to
the question of the relationship between the mind and body is a set of problems
having to do with mental representation. Ravenscro briefly takes up the
representational theory of mind as well as several theories of content. Next, he
considers mental causation and the specific sense in which certain problems
arise for physicalist theories of mind along with some possible solutions to
them. Finally, he considers various answers to what might be contemporary
philosophy of minds understanding of its own central question: how can
phenomenal consciousness exist in a purely physical universe? Jacksons
knowledge argument and several replies to it are considered.
In the end he argues that even if we havent go en past certain apparently
intractable problems, at least we have a more sophisticated set of tools and con-
cepts to help us understand them than at any other time in our history. Maybe we
have a good grip on what it is that we dont know, and thats a kind of progress.
Consciousness
Daniel D. Hutto
Hu o begins by admi ing that there is no clean, clear and neutral account of
what we mean by consciousness. He goes on to do what many others do: pin down
his topic with examples, alongside certain nearby expressions which seem to
strike a chord: what its like, for example. He also lists some of the characteristic
features of conscious experience mentioned in this connection, such as phenom-
enality, intentionality, subjectivity, unity, temporal extension and self-awareness.
He then takes up reductively naturalistic explanatory frameworks which by
turns equate consciousness to something else brain states, functional states,
xxiii
Introduction
and so on. Working through the main arguments for and against non-reductive
naturalism, Hu o se les on the so-called hard problem of consciousness: how
could this functional state or that representational property ever give rise to
conscious experience? The best we can do, it is sometimes concluded, is hope
for a specification of the non-reductive relations which hold between conscious
and other properties. The main replies to this thought are considered and judged.
In the final section, Rethinking Metaphors of Mind, Hu o considers a new,
or anyway dierent sort, of reply to the hard problem and the explanatory gap
between mind and world: the a empt to explain away the dierences which get
in the way of the explanatory identities we posit. He takes up Denne s analysis
of our a empts to understand consciousness, turning eventually to reflection
on representationalism and enactivism. In the end, Hu o concludes that a sat-
isfying naturalistic understanding of consciousness will require a network of
theories operating at dierent levels. More than this, whats needed is an under-
standing of which theories work best at which level no doubt alongside fresh
thinking about consciousness itself.
The Mark of the Mental

What is a mind? Whats the dierence between having a mind and not having
one? Adams and Beighley take up these questions in detail, considering several
candidates for what picks something out as mental as opposed to physical or
merely non-mental. They consider single property views which hold that all
mental states are mental because they share one property. Adams and Beighley
work through points for and against incorrigibility as a mark of the mental as
characterized by Rorty and Denne . They also consider work on intentionality
in the writings of Crane, Dretske and Tye. All such views are found wanting.
Next the authors examine a so-called system view, a type of view which
says there is a single set of properties that all minds must have, but not every
state that is part of the system must possess these properties themselves.
Searles conception of consciousness is considered in this connection, and objec-
tions are raised.
Their own account of the mental, a version of the systems view, is articulated
and defended in the final section. Beginning with a conception of the function
of mind, they go on to argue that mental systems share a cluster of properties.
To count as mental such systems must first of all possess non-derived meaning.
They must also do more than just carry information the states of a truly
mental system must rise to the level of meaning. States of a mental system must
also be capable of misrepresentation. Finally, they say that to count as mental,
a system must exhibit intentional behavior.
xxiv
Introduction
Substance Dualism
T. J. Mawson
In this spirited defense of substance dualism, Mawson gets right to it. He
assumes, along with almost everyone else, that there is physical stu which one
might pick out in part with paradigm cases: the stu which makes up tables,
chairs, stars and so on. If you take it that only this kind of stu exists, you are a
physical substance monist. Substance dualism, however, says that there is this
physical stu and another type of stu as well. According to the dualist, the
other type of stu is in essence capable of thought in the broadest sense of
that term, a property most definitely not had by physical stu.
He examines objections to substance dualism. There is of course the point
that any monism has the advantage of simplicity over any dualism, but can we
do be er than Ockhams razor? Mawson considers two sets of problems. First,
there are notorious diculties associated with identifying souls. What makes
one dierent from another, and how can we know anything about souls other
than our own? The second set of problems has to do with perhaps the loudest
objection to dualism: troubles with understanding the alleged interaction
between physical and mental stu. Mawson is unmoved by both sorts of
problems, and he gives replies to each.
He goes on to consider three reasons to believe that substance dualism
is true. Dualism lines up well with certain commonsense intuitions having to
do with personal identity, freedom and the qualitative features of conscious
experience, but physicalism has more than a li le trouble with each one.
He concludes with a Moorean argument which forces a choice between the
simplicity of physicalism on the one hand and the truth of those commonsense
intuitions which favor dualism on the other.
Physicalism
Barbara Montero
Just what is the main thesis of physicalism what does it mean to say that
everything is physical? Montero takes the ma er up in detail, working through
the meanings of the words everything, is, and physical. She begins by
considering dierent approaches to defining the scope of physicalism, conclud-
ing that we should understand everything in the most inclusive way possible:
everything, whatsoever, is physical. Next, she focuses on the relation between
the fundamentally physical properties and higher-level properties such as
mental properties. How should we understand the is relation? Pu ing other
possibilities to one side, her answer is couched in talk of upward determination
with some provisos, worlds which duplicate the fundamental physical properties
xxv
Introduction
and laws of our world also duplicate all the properties of our world. What,
then, is the physical? Again Montero considers a number of answers, plumping
finally for a negative characterization: the physical is the fundamentally
non-mental, non-divine, and non-normative.
Folk Psychology and Scientic Psychology
Barry C. Smith
We see other people as acting in accordance with beliefs and desires its a
large part of how we understand other people as people, how we see them as
taking a course of action deliberately and with an end in view, and its how we
explain and predict what they do. The sciences of the mind and brain have
revealed a great deal about what makes us tick in other senses. We have a
grip on the cognitive states and mechanisms which also serve to explain our
behavior. Smith considers the relationship between these two views that we
have of ourselves, primarily through the work of Davidson, Denne and Fodor.
He claims that a good account of folk psychology should provide a rational
explanation of action, accommodate the causal ecacy of the mental, and make
room for the dierence between first- and third-person ascriptions of mental
states. Following a careful account of folk psychology and scientific psycho-
logy, as well as a brief wave at eliminative materialism, the work of all three
philosophers is judged on the basis of these criteria.
Smith maintains that even if we can find an account of folk psychology which
satisfies those three requirements, we are still le with a number of questions.
Exactly how do we succeed when we ascribe beliefs and desires to others? What
role does consciousness play? How do the emotions jive with belief-desire psy-
chology? He concludes by pointing towards some potentially fruitful answers.
Internalism and Externalism in Mind
Sarah Sawyer
There is a large debate between internalists and externalists concerning the very
nature of mental properties is it just whats on the inside that counts? In this
paper Sawyer wades in by first giving an account of both views. Internalism,
roughly, holds that no two individuals could dier psychologically without
diering physically. Externalism, roughly, holds that individuals could be
physically identical but diverge psychologically, given certain external dier-
ences in, for example, their physical or social environments. Well-known
thought experiments owed to Putnam and Burge are considered. Sawyer gives
an account of the many possible forms of each kind of view which seem to
fall out of such reflections.
xxvi
Introduction
She goes on to outline metaphysical considerations having to do with

naturalism and mental causation which are thought to score points in favor
of internalism. Next, she considers epistemological claims concerning certain
features of self-knowledge which are thought to undermine externalism. In
adjudicating between the various objections and replies throughout her paper,
Sawyer emerges as an externalist, and she says something about her reasons
in a concluding section.
The Philosophies of Cognitive Science
Margaret A. Boden
As her title suggests, Boden maintains that theres much more to cognitive
science than merely the science of cognition there are many thinkers working
on various research programs producing a wealth of insights and her paper
captures something of the breadth of the subject. She takes up functionalism
first, as many consider it the core philosophy of the field, and examines several
possible variations in the work of Putnam, Fodor, and Denne . An interest
in neuroscience led many from functionalism to connectionism and parallel
distributed processing. Boden scouts the relevant objections and replies. She
then takes up the contribution cognitive science has made to understanding
the computational processes associated with representation.
Thoughts about representation, and in particular the senses in which inten-
tionality might be understood in terms of situatedness or embodiment, can lead
to reflection on the extended mind. Boden considers Clarks and Chalmers
claim that our minds are somehow partly located in the external world, work-
ing through some diculties for the view along the way. She then considers
the nearby notions of embodiment, enactiveness and phenomenology as they
appear in the continental tradition, as well as the influence this has lately
had on cognitive science. She concludes by examining a large number of con-
tributions cognitive science has made to central features of our understanding
of consciousness, the mind and life. In a concluding note Boden makes a case
for the claim that cognitive science really has provided not just good questions,
but satisfying answers in the philosophy of mind.
Representation
Georges Rey
At least some mental states are representational they stand for, refer to
or are about something else. How might such states fit into our general
understanding of the world? Rey considers the potential of the computational/
xxvii
Introduction
representational theory of thought to make sense of such states within a

physicalist framework.
Following a short characterization of the view, he considers various prob-
lems raised by our representational capacities which have to do with their ref-
erential opacity, and with the detection of non-local, non-physical properties.
He considers two general strategies for providing an account of intentional
content. One might go internalist and think that meaning is some sort of
internally specifiable state in the head: an image or stereotype, or an inferential
role; or one might opt for one of many externalist approaches: historical causal
theories, co-variation locking theories, or teleofunctional theories. Rey works
through points, for and against each, and concludes with some remarks on the
prospects for combining the two sorts of approaches.
Mental Causation
Neil Campbell
Mental events seem to stand in causal relations to physical events : my hopes
and fears apparently cause my smiles and frowns. Questions about mental
causation have a very long history, but many contemporary philosophers of
mind who are drawn to some version of non-reductive physicalism face a new
version of it. Motivations for non-reductive physicalism seem solid enough.
The anomalousness of the mental leads many to the view, while others are
persuaded by the innocent thought that mental events or properties are
multiply realizable in dierent physical forms. Whatever the motivation,
epiphenomenalism the una ractive possibility that the mental has no eect
on the physical seems only a few steps away.
Some worries about mental causation are raised against Davidsons anoma-
lous monism. Other objections are couched in terms of Kims principle of caus-
al-explanatory exclusion. Campbell considers both objections in close detail.
In his final section he argues that these objections depend on dubious meta-
physical assumptions about the nature of events. The objections, he concludes,
are either misguided or question begging.
Personal Identity
E. J. Lowe
Against the backdrop of some useful clarifications of both the identity relation
and the notion of a criterion of identity, Lowe takes up the question of a criterion
of personal identity. He follows Locke in thinking that before we can establish
a criterion of identity for persons, we have to say what kind of thing persons
xxviii
Introduction
are. We run into a certain sort of trouble here, however, because philosophers
have come up with a very long list of candidates: immaterial substances, material
substances, phases of substances, bundles, transcendental entities, and even mere
fictions. Lowe speculates that our very status as persons and in particular a
certain fact about the first-person pronoun is part of the problem.
He goes on to consider Lockes so-called memory criterion of personal
identity, because it is the first explicitly formulated criterion and, no doubt, the
one which has had the most influence. He considers Reids objection to Lockes
view and some possible modifications which might side-step it, alongside
further objections and replies. He concludes with a consideration of some
alternatives to variations on Lockes criterion, including the possibility that
personal identity is primitive and simple. This last, suggestive possibility goes
some way towards explaining why formulating a criterion of personal identity
is so dicult for us.
Embodied Cognition and the Extended Mind
Michael Wheeler
Wheeler describes the hypotheses of embodied cognition and the extended
mind as two stopping-o points in the flight from the Cartesian view of intel-
ligent action. In a nutshell the Cartesian view has it that the mind guides action
in a manner largely conceptually independent of the facts of embodiment.
However, recent thoughts about how we actually solve problems in the world
have made the Cartesian account less and less a ractive. Wheelers aim is to
examine the move not just away from the Cartesian picture but from embodied
cognition to cognitive extension.
He first sheds light on the embodied cognition hypothesis by marshalling
examples from recent work in cognitive science. Once a case has been made
for a certain conception of intelligent action, Wheeler argues that we face a
philosophical choice between a radical body-centrism and a new sort of
functionalism. Building up an argument from parity for the extended mind
hypothesis, Wheeler argues that the functionalist option is more a ractive. He
concludes by re-enforcing an aspect of the parity argument.
Paul Noordhof
In this contribution Paul Noordhof brings us right up to date with an account of
how three large topics central to the philosophy of mind have developed over
the last few decades. He takes up recent developments in our understanding of
xxix
Introduction
physicalism how we now characterize it and how this characterization aects

our understanding of mental causation. A discussion of phenomenal conscious-
ness follows, and Noordhof considers the so-called explanatory gap and vari-
ous a empts to put dualist intuitions to one side. He discusses a feeling, recently
sinking in, that these eorts are doomed, and he goes on to examine the role
that an appeal to representational properties might play in the debate. In a
final section, Noordhof examines new approaches to our understanding of
intentionality and the normativity of the mental.
xxx
1 Problems, Questions and
Concepts in the Philosophy
of Mind
Ian Ravenscroft
Introduction
Philosophy of mind has a long and distinguished history, but it is not the aim of
this chapter to provide an historical overview of philosophical investigations of
the mind. Rather I intend to elaborate on what I take to be the most significant
themes in contemporary philosophy of mind. It is worth noting, right at the
outset, that philosophy of mind has been one of the most dynamic perhaps the
most dynamic areas of English-speaking philosophy over the last half century.
Very roughly, research in this field falls into four broad areas. (Needless to
say, a tangled web of connections exists between these areas, rendering the
boundaries to some extent arbitrary).
1. Metaphysics of Mind: The nature of the mental, and in particular the study
of how the mental relates to the physical, is one of the most enduring
philosophical puzzles and has been at the forefront of contemporary work
in philosophy of mind. This problem is o en simply called the mind-
body problem, and will take centre stage in what is to come. Within the
broad area of the mind-body problem are a host of specialized issues
including mental causation, mental representation and consciousness.
Other specialized metaphysical issues in the philosophy of mind include
perception, memory, action and intention. Very o en these more special-
ized issues are connected with questions in other areas of philosophy.
For example, discussions of moral responsibility o en involve claims
about whether or not an action was intentional.
2. Epistemological Issues: The mind raises special epistemological questions
concerning how, and to what the extent, we have knowledge of our own
mind and the minds of other people. Over the last few decades philo-
sophical work on these questions has been influenced by work in cogni-
tive and developmental psychology and evolutionary biology.
1
3. Behavioural and Cognitive Sciences: In addition to the vigorous development

of the philosophy of mind over the last half century there has been an
extraordinary growth in the behavioural and cognitive sciences. Philo-
sophers have influenced and been influenced by these fields, which
include computer science, neuroscience, developmental and cognitive
psychology, evolutionary biology, and economics.
4. Methodological Issues: What is the correct approach to the philosophical
study of mental phenomena? Is there room for a priori (or armchair)
investigations, or are all the issues empirical? If all the issues are empiri-
cal, is there anything le for philosophers to do in this area, or should
we just humbly abandon the field to the behavioural and cognitive
sciences? Some have argued that the role of the philosopher of mind is
integrative or synthetic. There is the task of integrating the various aspects
of one cognitive science into a satisfying whole, and also the task of
integrating a range of cognitive sciences into a comprehensive vision of
human mentality. Kim Sterelny has called these the internal and external
integrative projects respectively (Sterelny, 2003, pp. 35).
So extensive has been the research over this period, and so enormous
the resulting literature, that any author of an overview like this one is forced
to make hard decisions about what and what not to include. I will focus
on contemporary approaches to what I earlier described as one of the most
enduring philosophical puzzles: the mind-body problem. Even within this
restricted domain I will be forced to make some drastic editorial decisions.
For example, there are important approaches to thinking about the mind-body
problem over which I vault without a glance. (I flinch, but I do not glance.)
In particular, there is an important strand of mid-twentieth century philo-
sophy of mind which includes on the one hand the logical positivists and on
the other the later Wi genstein and Ryle, that I do not discuss at all. This is
not because I believe these thinkers to be unimportant, but because I believe
that I can adequately develop an account of contemporary philosophy of
mind without paying these figures direct a ention. Their influence on contem-
porary philosophy of mind was considerable, but it was channelled through
figures such as Smart and Denne , about whose views I do have something
to say.
In the next section I outline some of the key moves that have been made on
the mind-body problem, beginning with Descartes because, in striking ways,
his problems have turned out to be our problems. Among other ma ers, reduc-
tive and non-reductive physicalism and the idea of supervenience are sketched
in that section. Most contemporary philosophers of mind endorse some version
of the representational theory of mind: they think of the mind as an organ for
2
developing and manipulating representations. The third section briefly surveys

some of the diculties which arise when trying to understand the nature of
mental representation. The mind-body problem challenges us to account for
the close connections which minds apparently enjoy with their associated
bodies. Those connections apparently include causal connections: the state of
my body causally influences the state of my mind, and vice versa. The fourth
section is devoted to an exploration of the way the problem of mental causation
emerges in the context of physicalism. The fi h section takes up the problem
of phenomenal consciousness which presents the mind-body problem in
especially sharp relief, and brings together many of the issues discussed in this
overview. I allow myself a brief conclusion.
Mind and Body
Many problems in the philosophy of mind emerge because we are deeply

commi ed to two distinct ways of thinking about ourselves as human beings.
Our commitments are in tension, forcing us to reconsider some of our most
cherished ideas about what it is to be human. On the one hand, we think that
human beings are closely connected with physical bodies living in a physical
world. (I will have a lot more to say about what closely connected might mean
in subsequent sections.) Our bodies are assemblages of atoms and energy,
existing in a space-time manifold and answering to the same laws as all other
assemblages of atoms and energy. We are also biological creatures, having many
features in common with other mammals and sharing a common ancestor with
all the living creatures on Earth. On the other hand, we think of ourselves
as having minds. We have an array of mental states and dispositions. We get
hungry and we get hurt. We fall in and out of love. We are prone to anger or
sympathy. We have beliefs and preferences. We can perceive our environment
through a number of modalities and can remember the past. We can learn and
we can reason. We represent the way the world was, is, and may come to be.
We even represent ways the world cannot be. And, most mysterious of all, we
are conscious.
These conceptions of ourselves are in tension because it is not obvious how
all of the features of mental states to which we are commi ed can be squared
with our conception of ourselves as closely connected to physical bodies. I will
set the scene for an exploration of these features by discussing in what sense
we are closely connected to physical bodies. In particular, I will discuss the
relationship between mental properties and brain properties. We will see that
every available answer to the question of the relationship between mind and
body throws up new challenges.
3
Dualism
The great French philosopher, mathematician and scientist, Ren Descartes,

articulated a theory of the relationship between mind and body which is
now called interactive substance dualism (or Cartesian dualism) (Descartes,
1637/1985). There is no agreed definition of the term substance; however, for
present purposes I will follow David Armstrong (1968, p. 7) and take a sub-
stance to be something which could exist alone in the universe. Importantly,
substances have properties. Thus the Sun is a substance because we can imagine
a universe containing nothing but the Sun. One of the Suns properties is its
having a mass of 2 1030 kg. Having a mass of 2 1030 kg is not a substance
because the universe could not contain nothing but the having of a mass
of 2 1030 kg: there would have to be in addition something which had
that mass.1
Interactive substance dualism is the doctrine that there are two fundamen-
tally dierent kinds of substances in the world: non-physical mental substance
and physical substance. Human minds are mental substances; human bodies
are physical substances. Descartes believed that mind and body interact: infor-
mation about the bodys environment is sent via the sensory organs to the brain
and from there it passes to the mind; instructions on how to respond to the
environment are sent from the mind back to the brain which then orchestrates
the bodys movements. Descartes did not regard the brain as merely a conduit
through which information passed to and from the mind: he allowed that the
brain played a role in processing perceptual signals and organizing motor
responses. However, Descartes did insist that all higher cognitive functions,
especially reason and language, are activities of the non-physical mind
(Descartes, 1637/1985).
For Descartes the close connection between mind and body or more
specifically, between mind and brain is causal. Our bodys being in a certain
states (e.g. damaged) causes our mind to be in a certain state (pain). Similarly,
our minds being in a certain state (pain) causes our body to be in a certain
state (withdrawing from the source of damage). Famously, Princess Elizabeth
of Bohemia, one of Descartes correspondents, pointed out an acute tension
in Descartes views (Anscombe et al., 1954, pp. 2745). On the one hand,
Descartes argued that the mind and body are fundamentally distinct sub-
stances; on the other, he held that there are causal interactions between the
two. How can such radically distinct substances interact? Descartes was
unable to oer a persuasive reply. Elizabeths challenge is, quite properly, o en
regarded as a very serious objection to interactive substance dualism. We will
see, though, that it is a just one version of the general problem of explaining
how mental states have causal powers.
4
Elizabeth challenged Descartes to account for the causal interactions

between the non-physical mind and the physical brain. But this is not the only
way causation makes trouble for interactive substance dualism. If interactive
substance dualism is correct, then some non-physical mental event, M, caused
a physical brain event, P. But modern science strongly supports the view that
the world is physically closed; that is, it supports the view that every physical
event has a sucient physical cause.2 Consequently, it seems very likely that
there exists a prior physical event, P*, which is a sucient cause of P. So,
interactive substance dualism is commi ed to endorsing one of two positions,
neither of which is very a ractive.
Position 1 Modern science is wrong: the world is not physically closed.
Position 2 Event P is over-determined. That is, event M is a sucient cause of

event Ps existing and event P* is a sucient cause of event Ps existing.
Position 1 is una ractive because it amounts to endorsing substance dualism

in the face of our best theories of the world. Position 2 is una ractive because
it involves postulating a very large number of causes which makes mental
properties a likely target of Ockhams razor. The threat of over-determination
is something we will come across again in the fourth section. To foreshadow:
physicalists about the mind run into their own version of the over-determination
problem.
Descartes thought of the mind as a non-physical substance which had
mental properties. But this way of conceiving of dualism is not compulsory.
Interactive property dualists accept that there is only one kind of substance
in the world physical substance but they think that there are two kinds of
properties in the world physical properties and non-physical mental proper-
ties. According to this view, some special physical objects have non-physical
mental properties. The living human brain is the obvious and perhaps
only example of a physical object which has non-physical mental properties.
According to interactive property dualism, human brains can instantiate a
certain physical property lets call it Q. Q causes the brain to have a non-
physical mental property, M. M in turn may cause a further physical brain
property to be instantiated say property P.
Clearly, Princess Elizabeths problem arises with interactive property
dualism: exactly how do physical and non-physical properties interact? In
addition, the claim that the world is physically closed generates a problem for
interactive property dualism. For, according to physical closure, the physical
brain property P has a sucient physical cause; for example, it was caused by
5
physical brain property P*. Consequently, interactive property dualism is

commi ed to endorsing one of two uncomfortable positions:
Position 1 Modern science is wrong: the world is not physically closed.
Position 2 Property P is over-determined. That is, property M is a sucient

cause of the existence of property P and property P* is a sucient cause of
the existence of property P.
As we saw earlier, neither denying physical closure nor endorsing over-

determinism is a ractive. One way for the property dualist to avoid the
problem of over-determination is to endorse epiphenomenal property dualism.
On this view, physical brain properties cause non-physical mental properties,
but not vice versa. Mental properties are idle wheels, driven by the physical
engine of the brain but driving nothing. To adopt a political metaphor, they are
aristocratic properties, relying for their existence on hardworking physical
properties but doing no work themselves. Since mental properties have
no causal impact on the physical realm, no over-determination of physical
properties occurs. However, as we shall see in the fi h section, epiphenomenal-
ism comes at a heavy price.
In addition to his views on the causal relations between mind and body,
Descartes also held that mental states are essentially objects of introspection;
that is, they are essentially conscious. Two powerful intellectual movements
unseated this idea in the twentieth century. The first was psychoanalysis.
Freud sought to explain a range of neurotic symptoms in terms of unconscious
desires (e.g. see Freud, 1917/1991). The aim of psychoanalytic therapy is to
bring such desires into the light of consciousness. The second was the rise of
cognitive psychology in the 1950s and 1960s. Chomskys account of syntactic
processing postulated rich informational structures which play a central role
in understanding and producing grammatical sentences, but which are not
accessible to consciousness (Chomsky, 1994). Postulating such structures
became standard practice in cognitive psychology. David Marr (1982) and Irvin
Rock (1983), for example, postulated a range of unconscious informational
structures in their work on vision.
It is commonly supposed that consciousness poses a special problem to
physicalist theories of the mind. How, Colin McGinn asked, can technicolour
phenomenology arise from soggy grey ma er? (McGinn, 1991, p. 1). We will
look at the challenge consciousness poses to physicalism in the fi h section.
For the moment, it is worth noting that consciousness also poses a serious
problem to dualism. How, we might ask, can technicolour phenomenology
arise from non-physical soul stu ? Even if we accept that our mental states
are entirely revealed to us by introspection, it is apparent that consciousness
6
remains a mystery. At best we have introspective access to the contents of

consciousness; introspection tells us nothing about the structures and processes
which make consciousness possible. Descartes might reply that there are no
such structures and processes: thoughts just are conscious entities. But this
seems unsatisfactory. We deserve an account of how the universe came to con-
tain entities which simply are conscious. The demand for such an account is
especially pressing once we accept that humans evolved from animals which
did not have conscious experiences.
Reductive Physicalism
I will now leap ahead 300 years to the rise of reductive physicalism (also called
the identity theory) in the 1950s. The theories of mind which populate the
intervening decades are neither unimportant nor uninteresting; however, the
problems on which I will focus in the next three sections mental representa-
tion, mental causation and consciousness take their modern forms in the con-
text of the physicalist theories of the mind which have their origin in reductive
physicalism. To a first approximation, reductive physicalism identifies mental
properties with brain properties. Crucially, the brain properties with which
mental properties are identified are held to be physical properties. To use an old
example from the 1950s an example whose details should not be taken too
seriously the property being in pain is identical to the property having c-fibre
activity.3 It is important to stress that reductive physicalism proposes type
identities. The claim is not merely that every instance of pain is identical to an
instance of a physical property; rather, the claim is that all instances of the
type pain are identical to instances of the type c-fibre firing.
Early proponents of reductive physicalism took the property identities
discovered by science as their model. For example, J. J. C. Smart drew an
analogy between the identity of pain and c-fibre firing (on the one hand) with
the identity of water and H2O (on the other). Crucially, the discovery that water
is H2O was the outcome of a process of scientific investigation; it is not some-
thing that can be discovered by conceptual or linguistic analysis. Similarly,
Smart thought that identities between mental properties and brain properties
would be discovered by scientific investigation. It is no objection to reductive
physicalism that pain and c-fibre activity dont have the same meaning, nor
is it an objection that the proposed identities cannot be discovered a priori. (See
the seminal papers by U. T. Place [1956], H. Feigl [1958] and Smart [1959].)
Non-reductive Physicalism
In the 1960s, two separate developments drove many philosophers to con-
clude that reductive physicalism is mistaken. The first development lead
to functionalism; the second to anomalous monism. In 1967 Hilary Putnam
pointed out that mental properties are, at least in principle, multiply realizable.
7
Mental property M is multiply realized if and only if some instances of M are

identical to instances of property P1, whereas other instances of M are identical
to instances of property P2 (P1 P2). It might be, for example, that while pain
in humans is identical to c-fibre firing, pain in dolphins is identical to d-fibre
firings. It may even be that the pain I am experiencing now is identical to an
instantiation of brain property B1, whereas the pain you are feeling now is
identical to an instantiation of brain property B2 (B1 B2). The possibility of
multiple realizations presents a challenge to reductive physicalism because it
raises the possibility that mental property M cannot be identified with brain
property B. Non-reductive physicalism takes the possibility of multiple realiza-
tions entirely seriously, and claims only that each instance of a mental property
is identical to an instance of a physical property.
Reductive physicalism has an easy answer to the question What do all
instances of pain have in common?. According to reductive physicalism, every
instance of pain is an instance of c-fibre firing. But this answer is not available
to the non-reductive physicalist who denies that every instance of pain has to
be identical to an instance of c-fibre firing. So what answer can non-reductive
physicalism advance?
Functionalism is, in eect, an answer to the question What do all instances
of mental property M have in common? which is compatible with non-
reductive physicalism. According to functionalism, mental properties are
characterized by their causal roles. Pain, for example, is the property which is
caused by bodily damage and causes withdrawal from the source of damage;
has important causal links to anxiety and to desire; and, in conjunction with
certain beliefs, can lead to particular behaviours (e.g. if I believe that applying
ice to the damaged part of my arm will reduce my pain, I will go the refrigerator
and look for some ice).
Putnam (1967) observed that dierent physical properties can occupy the
same causal role. Consider the property of being a thermostat. A number of dif-
ferent physical properties can occupy the causal role characteristic of being a
thermostat. Similarly, a number of dierent physical properties can occupy
the causal role characteristic of pain. It may be that in humans the causal role
characteristic of pain is occupied by c-fibre firings whereas in dolphins it is
occupied by d-fibre firings. Or it may be that in me the characteristic causal
role of pain is occupied by an instance of brain property B1, whereas in you it
is occupied by an instance of brain property B2 (B1 B2). So functionalism
is compatible with the multiple realization of mental properties and is an
important way of elaborating non-reductive physicalism.4
Functionalism is one of the theories which rapidly replaced reductive phys-
icalism in the 1960s. The other was Donald Davidsons anomalous monism.
Davidson begins by endorsing three principles which, prima facie, are inconsis-
tent: (1) the principle of causal interaction: mental and physical events causally
8
interact; (2) the principle of the nomological character of causality: causal rela-
tionships always fall under strict laws; (3) the principle of the anomalousness
of the mental: there are no strict laws relating mental and physical events (see
Davidson, 1970, pp. 2234). If, as (1) requires, there are causal relations between
mental and physical events, then by (2), there must be strict laws relating
mental and physical events. And yet by (3) there are no such laws. Davidson
oers an ingenious resolution of this (apparent) inconsistency a resolution
which appeals to non-reductive physicalism. However, before turning to
these ma ers I will briefly explain why Davidson takes the mental to be
anomalous.5
Davidson correctly observes that we do not a ribute mental states one by
one; rather, we a ribute extensive complexes of mental states (Davidson, 1970,
p. 221). For example, if I observe Jones walking towards a mailbox with a le er
in his hand, I will a ribute to him not only the desire to post a le er, but also
the belief that placing a le er in a mailbox is the way to post it; that the red
object nearby is a mailbox; that walking is an eective means of covering
the distance between his present location and the mailbox; etc. But there is a
very large number of sets of mental states that would account for Jones action.
(Jones believes that his le er is about to explode; he desires to prevent the
explosion; he believes that the mailbox is in fact a bomb disposal device placed
there by MI5 . . .) According to Davidson, we select one set of mental states (or
a small number of such sets) from this vast range of possibilities by applying
a principle of charity: we assume that the target is rational. But the notion
of rationality is a normative one that is not found in the physical sciences.
Consequently, we cannot expect to find laws linking the mental realm with
the physical realm.
It is important to note that Davidson is not claiming that the a ribution of
mental states requires that the target be perfectly rational. On the contrary, he
explicitly claims that the a ribution of minor cognitive slips is only possible
because we assume that people are by and large rational: Crediting people
with a large degree of consistency . . . is unavoidable if we are to be in a
position to accuse them meaningfully of error and some degree of irrationality
(Davidson, 1970, p. 221).
Lets now turn to Davidsons resolution of the inconsistency mentioned
above. According to the anomalousness of the mental, there are no strict laws
linking a mental event, described in psychological language, with a physical event,
described in physical language. But if non-reductive physicalism is true, every
mental event token is a physical event token, and therefore has a description in
purely physical terms. Consider an event E which has both a mental descrip-
tion, M, and a physical description, P. By the anomalousness of the mental there
is no strict law linking E, described as M, with some other physical event. But
there maybe a strict law linking E, described as P, with some other physical
9
event. It is in virtue of their physical realization that mental events exhibit

lawful relationships with other physical events, and engage in causal relations
with both other mental events and other physical events. So non-reductive
physicalism provides us with a resolution of the tension we have noted between
the three principles. To that extent non-reductive physicalism is supported
(see Davidson, 1970, pp. 2235).
Supervenience
According to both reductive physicalism and non-reductive physicalism,

mental properties depend on physical properties. It is in virtue of having a
certain set of physical properties that a person has their mental properties. The
notion of supervenience has been developed to articulate and explore what
depends on means in this context. There are a number of dierent ways
of conceiving of supervenience; one perspicuous approach says that property
P supervenes on property Q if, and only if, fixing the distribution of Q fixes the
distribution of P. For example, if we fix the distribution of the property of being
H2O we thereby fix the distribution of the property of being water; consequently,
the property of being water supervenes on the property of being H2O. (God did
not fix the distribution of H2O and then get on with the job of fixing the dis-
tribution of water; by fixing the distribution of H2O, God thereby fixed the
distribution of water.) An alternative way to describe supervenience is in terms
of variation: property P supervenes on property Q if, and only if, there can be
no variation in P without variation in Q. In a free market, the price of fish
depends on the demand for, and supply of, fish. There can be no variation in
the price of fish without a variation in either the demand for or supply of fish.
In other words, the price of fish supervenes on the demand and supply of fish.
(For extended discussions of supervenience, see the essays in Kim, 1993c.)
The property of being water is sometimes called the supervenient property
(relative to the property of being H2O), and sometimes called the higher level
property (relative to the property of being H2O). Conversely, the property of
being H2O is sometimes called the subvenient property (relative to the property
of being water), and sometimes called the lower level, or base, property (relative
to the property of being water).
Reductive physicalism posits the supervenience of mental properties on
physical properties. If the type identity theory is true, the distribution of c-fibre
firings fixes the distribution of pain; that is, pain supervenes on c-fibre firings.
Alternatively, there can be no variation in the painfulness of a persons experi-
ence without a variation of their c-fibre firings. Non-reductive physicalism also
posits supervenience relations between mental properties and physical proper-
ties. However, if mental properties are multiply realized, it will not be the case
10
that mental properties supervene on simple neurological properties like c-fibre

firing; rather, mental states will supervene on complex conjunctions of physical
properties. Pain, for example, will supervene on human c-fibre firings and
dolphin d-fibre firings and so forth.
Brain states have both wide and narrow properties. The brains narrow prop-
erties are its intrinsic properties the properties the brain has irrespective of its
relations to other entities. They include the spatial relations of its anatomical
parts and the spatio-temporal distribution of the neurotransmi ers it contains.
The brains wide (or broad) properties are the properties it has in virtue of its
relations to other entities. Being caused by the sound of a piano is a wide property
of some brain properties. With this distinction in place, the question arises as to
whether mental properties supervene exclusively on the narrow properties of
the brain, or on a mix of wide and narrow properties. Many philosophers have
denied that all mental properties supervene exclusively on narrow brain prop-
erties. The case for including wide properties in the subvenient base is some-
times advanced by appealing to thought experiments involving Swampman
(Davidson, 1987a). Swampman is the outcome of an extraordinary conjunction
of chance events in the Florida Everglades lightening hi ing a swamp where
anaerobic bacteria have produced just the right combination of amino acids.
Lets say that this event happens right now. By an amazing coincidence, at this
moment Swampman has a brain exactly like my brain in all narrow respects.
However, while I have a memory of Sydney Harbour Bridge, Swampman does
not. This is because a necessary condition on being able to remember Sydney
Harbour Bridge is being in the right causal relations to it, and while I am lucky
enough to have those relations, Swampman, over there in the Everglades, is
not. We have therefore a case of mental variation without variation of intrinsic
brain properties. It seems that the subvenient base of at least some mental prop-
erties includes wide properties. This is a theme to which we will return in the
Mental Representation and Mental Causation sections.
Eliminativism, instrumentalism and the intentional stance
I will close my selective history of the mind-body problem by briefly discussing

three further contemporary views of mental states: eliminativism, instrumen-
talism and the intentional stance. Eliminativism is the doctrine that mental
states dont exist. According to eliminativism, the ontological status of mental
states is akin to the ontological status of phlogiston. Arguments for eliminativ-
ism o en begin with folk psychology. Very roughly, folk psychology is a theory
of human psychology possessed by all normal human beings older than about
five (Ravenscro , 2010). According to many philosophers and psychologists, it
is by deploying folk psychology that we understand ourselves and other people
11
as psychological agents. Now folk psychology posits a number of mental states

beliefs, desires, pains, imaginings, etc. Our commitment to these states should
stand or fall with our commitment to folk psychology. (The analogy with phlo-
giston is o en pressed here. Scientists accepted the existence of phlogiston
because the phlogiston theory was, at the time, the best theory of combustion
available. When Lavoisier replaced the phlogiston theory with the superior
oxidation theory, the rational grounds for believing in phlogiston were
removed.) If it turns out that folk psychology is a poor theory, we should aban-
don the mental states it posits. Eliminativists insist that folk psychology is
indeed a poor theory and conclude that mental states dont exist. The claim
that folk psychology is a poor theory is typically defended by identifying its
explanatory weakness and by arguing that it cannot readily be reduced to
neuroscience (see, especially, Churchland, 1981). Many counter-arguments have
been oered against this kind of eliminativist argument. Prominent objections
include a acks on the idea that folk psychology is a poor or inadequate theory
e.g. see Horgan et al., 1985), and a acks on the claim that our commitment to
the existence of mental states should stand or fall with the success of folk
psychology (e.g. see Kitcher, 1984 and Von Eckardt, 1995).6
Like the eliminativist, the instrumentalist does not admit mental states into
the ontological fold. But unlike the eliminativist, the instrumentalist still values
mental states; in particular, mental states are regarded as indispensible instru-
ments of prediction. Instrumentalism is not, though, especially a ractive. For
we can ask how it is that positing mental states is predictively successful, and
the obvious answer is that positing such states is successful because they are
real. (Compare: Why is atomic theory so successful? Because the states over
which it quantifies are real.) Daniel Denne is o en regarded as an instrumen-
talist (e.g. see Fodor, 1990a), although he would, I think, regard that label as at
least partly misleading. Like Davidson, Denne stresses the normative charac-
ter of belief and desire a ribution. According to Denne , in the majority of
cases we predict peoples behaviour by adopting what he calls the intentional
stance:
Here is how it [the intentional stance] works: first you decide to treat the
object whose behavior is to be predicted as a rational agent; then you figure
out what beliefs that agent ought to have, given its place in the world and
its purpose. Then you figure out what desires it ought to have, on the same
considerations, and finally you predict that this rational agent will act to
further its goals in the light of its beliefs. A li le practical reasoning from
the chosen set of beliefs and desires will in many but not all instances
yield a decision about what the agent ought to do; that is what you predict
the agent will do. (Denne , 1987b, p. 17)
12
Denne oers the example of a chess playing computer program (Denne ,

1971). Humans can most eectively play against a chess program by taking the
intentional stance towards it. From that stance we find ourselves a ributing to
it beliefs like It thinks Im going to sacrifice my rook and desires like It wants
to move its queen into a stronger position. Of course, when we examine the
program we do not find internal states with those contents; rather we find an
algorithm which determines and ranks possible moves. Denne insists, though,
that the question of the reality or otherwise of mental states does not turn
on whether we can make straightforward identifications of brain states with
mental states (Denne , 1987b, 1991b). He calls any system whose behaviour
can be predicted from the intentional stance an intentional system. From the
intentional stance we a ribute mental states to humans and other intentional
systems, and once we grasp those mental states we can detect pa erns in the
behaviour of those systems. For example we see that, despite considerable
dierences in history and environment, Joseph and Josephine have something
in common something which we capture with phrases of the form ___ really
wants a cappuccino. Armed with that a ribution, we can predict with some
accuracy the kinds of behaviour Joseph and Josephine will undertake. Accord-
ing to Denne the pa erns we detect from the intentional stance are real
they are in the world to be detected by any intelligence capable of adopting
the intentional stance. (Compare: the pa erns described by Keplers laws of
planetary motion are there in the world to be detected by anyone capable of
adopting the astronomical stance.) There is a sense, then, in which Denne
is indeed a realist although he is not what he has somewhat disparagingly
called an industrial strength realist (Denne , 1991b, p. 42).
Earlier I raised the question of how, if mental states ascriptions arent
literally true, appealing to mental states is a useful strategy. Denne s own
response is that treating each other as rational agents works because we are, by
and large, rational. And the explanation of our being rational agents in turn
appeals to natural selection: rationality confers evolutionary fitness. However,
it is an open question whether natural selection is likely to drive the evolution
of largely rational creatures. (For discussion, see Stich, 1990.)
Mental Representation
It is widely accepted that at least some mental states are about, or represent,
states of aairs. This is sometimes expressed by saying that many mental states
have content. Prominent among the mental states that are widely assumed to
have content are the propositional a itudes, which include beliefs, desires,
fears, hopes and wishes. The name propositional a itude derives from the
13
idea that such states consist of an a itude (e.g. of belief) towards a proposition
which represents a state of aairs. Thus I can believe that the cat is wearing
a hat and hope that Dr Suess is amusing. Some other kinds of mental states have
contents, for example, perceptions. It is controversial whether the subjective
feels of mental states like perceptions and sensations are representational.
(That is an issue to which we turn briefly in The Knowledge Argument
subsection.)
In this section I sketch some of the key issues surrounding mental repre-
sentation, and some of the key theories of content. I will frequently refer to
the syntactic and semantic properties of mental states. The syntactic proper-
ties of a mental state are those of its narrow properties in virtue of which it
engages in cognitive processes. They are sometimes referred to as a mental
states shape. (Think of how subway tokens engage with the turnstile mecha-
nism in virtue of their narrow properties like shape and mass.) The semantic
properties of a mental state are those properties it has in virtue of its
representational properties. Truth and falsity are semantic properties par
excellence.
The representational theory of mind
The representational theory of mind claims that propositional a itudes are

representations, and that cognition involves sequences of representations that
bear appropriate semantic properties to one another (e.g. see Sterelny, 1990).
The computational theory of mind is a well-known form of the representational
theory of mind. Computational processes take one or more states as input and
yield one or more states as output. Crucially, the transformational processes
involved are only sensitive to the input states syntactic properties. (A string of
1s and 0s in a computers CPU may represent the velocity of a subatomic
particle or the number of servings of French fries consumed in Montreal last
fall: the semantic properties of the string make no dierence at all to how the
computer handles the string.) If the computational processes are properly
arranged, the outputs bear appropriate semantic relations to the inputs; for
example, the transformations may be truth-preserving. According to the com-
putational theory of mind, propositional a itudes are computational states with
syntax and semantics, and mental processes involving propositional a itudes
are computational processes (e.g. see Fodor, 1975, 1980). Obviously, any repre-
sentational theory of mind requires a theory of content a theory of how
mental states acquire their representational properties. In the Theories of Con-
tent subsection below, I provide brief sketches of two fundamental approaches
to content.
14
Theories of content
Recent discussion of mental representation has largely focussed on two broad

approaches to the issue of mental content: (1) conceptual role approaches and
(2) information-theoretic approaches.
Conceptual Role Approaches

Beliefs form highly structured causal networks, and propositions form highly
structured inferential networks. For any causal network of beliefs, C, we can
identify an isomorphic inferential network of propositions, I. The isomorphism
between C and I allows us to map each element of C onto an element of I. In
other words, we can assign a proposition to each belief. That proposition is the
beliefs content (e.g. see Block, 1986b). However, this approach to content faces
at least three obstacles:
(1)The approach assumes that humans are largely rational; without such an
assumption there would be li le interest in finding isomorphisms
between the causal structure of our corpus of beliefs and the inferential
structure of a set of propositions. However, there is an extensive body
of work in psychology which suggests that humans are not especially
rational. (For details, and extensive discussion of the philosophical
consequences, see Stich, 1990.)
(2)There exists a very large number of inferential networks isomorphic
to any one causal network; consequently, there will be no unique assign-
ments of contents to beliefs.
(3)The proposal in eect says that the content of a belief is dependent upon
its causal relations, whereas intuitively the causal relations of a belief
are dependent upon its content.
An important variant of conceptual role semantics is sometimes called the

map theory:
[T]he proposal is that we match the head states that are beliefs with possible
states of the world by the rule that each state of the head gets assigned the
possible state of the world which is such that if it were the way things
actually are, the behaviour that head state causes would realize what the
subject desires. (Braddon-Mitchell et al., 1996, p. 181)
This view requires that the basic unit of semantic interpretation is an agents
entire corpus of beliefs and desires. We assign to the agent those beliefs which,
were they true, would bring about behaviour that satisfied the agents desires;
15
and we assign to the agent those desires which would bring about behaviour
leading to their satisfaction if the agents beliefs were true.7
This proposal faces very similar diculties to the proposal considered
previously:
(1)It assumes that agents are in fact rational.

(2)It leads to the threat of non-unique assignments of content since, as
remarked in the Non-Reductive Physicalism subsection above, there
is a very large number of belief and desire sets capable of causing any
given behaviour.
(3)It makes content dependent on causal role rather than vice versa.
Information-Theoretic Approaches
A variety of relations exist between a thought and states outside the body.
According to externalist theories of content, a special subset of those relations
determines the thoughts contents. Which external relations are the content-
conferring relations? Many contemporary philosophers stress information-
bearing relations.8 Back in the 1980s Fred Dretske proposed that a thought, T,
is about a state of aairs, S, in virtue of carrying information about S (e.g.
see Dretske, 1981). That is, T is about S if, and only if, the probability of S given
T is 1. Smoke means fire because the probability of there being fire, given
there is smoke, is 1; similarly, my sheep thought is about sheep because the
probability there is a sheep present, given I have a sheep thought, is 1.
Ingenious though this suggestion is, it is immediately confronted by a pair of
problems. The first is semantic promiscuity: Dretskes view entails that meaning
is superabundant. There are a very large number of states of the world which
carry information about other states of the world. Smoke carries information
about fire; tsunamis carry information about earthquakes; tides carry informa-
tion about the relative positions of the Earth, Sun and Moon. It follows that
meaning is not an especially psychological notion it is not a unique feature
of minds and mind-generated artefacts. The second problem is o en called the
disjunction problem. Under certain conditions I will have a sheep thought in the
presence of a goat. (I might mistake a goat for a sheep on a particularly gloomy
a ernoon.) If that is the case, the probability of there being a sheep present
when I have a sheep thought is less than 1. Indeed, it may be the case that the
probability of there being a sheep or a goat present given I have had a sheep
thought is 1. In that case my sheep thought is about sheep-or-goats rather than
about sheep.
One important line of response to the disjunction problem appeals to
natural selection. Frogs snap their tongues at flies. Say that the frog tokens
the concept FLY when a fly is in its visual field, and (ceteris paribus) tokens of
16
that kind cause tongue snapping. We naturally think that FLY is about flies.
However, small boys sometimes throw BBs (lead pellets) at their pet frogs,
eliciting tongue snappings; that is, frogs misrepresent BBs as flies. Lets assume
that the probability of there being a fly-or-BB, given that the frog tokens FLY
is 1. It follows that the frogs FLY concept is about flies-or-BBs, not flies. In
response, the teleological theory of content introduces the idea of the biological
function of the frogs FLY tokens. It is because the modern frogs ancestors
tokened FLY in responses to flies that they were able to survive and reproduce.
The existence of the modern frog is dependent upon ancestral tokenings of
FLY in response to flies rather than inedible BBs. The theory of natural selection
gives us a principled way of saying what FLY tokenings are for, and that in turn
allows us to distinguish between appropriate tokenings which successfully
represent flies, and inappropriate tokens which misrepresent BBs as flies. (See
especially Millikan, 1984 and Papineau, 1984.)
Jerry Fodor rather doubts that Darwin is going pull Brentanos chestnuts
from the fire (Fodor, 1990d, p. 70). In the frogs ancestral environment, small,
black, fast moving objects in the frogs visual field were almost always flies.
An ancestral frog which responded to small, black, fast moving objects would
have survived and reproduced just as successfully as an ancestral frog which
responded to flies. So the teleological theory of content has no way of deter-
mining whether the modern frogs FLY concept refers to flies or to small, black,
fast moving objects. The disjunction problem has returned. (For countermoves
see Godfrey-Smith, 1994a, pp. 2734.)
Very roughly, Fodors own solution to the disjunction problem turns on the
asymmetric dependence of my goat-caused sheep thoughts on my sheep-caused
sheep thoughts. Goats only get to cause sheep thoughts because sheep cause sheep
thought: no sheep-caused sheep thoughts, no goat-caused sheep thoughts. Sheep
thoughts are about sheep not goats (or sheep-or-goats) in virtue of the depen-
dence of goat-caused sheep thoughts on sheep-caused sheep thoughts (see
Fodor, 1990e). Some philosophers are concerned that Fodors approach still
faces the problem of semantic promiscuity (e.g. see Adams et al., 1994). Fred
Adams (2003, p. 161) oers the following example, which he a ributes to Colin
Allen. Kadu antelopes bite the bark of the acacia tree which in turn emits tannin
as a deterrent. Now while the tannin response evolved as a deterrent to Kadu
rather than to humans, if a human were to damage the bark of the acacia tree
tannins would be emi ed. So the acacias tannin emissions are about Kadu bit-
ings rather than damage by humans (or Kadu-bitings-or-damage-by-humans)
because Kadu bitings cause tannin emissions, and human-caused tannin emis-
sions are dependent on Kadu-caused tannin emissions. (No Kadu-caused tannin
emissions, no human-caused tannin emissions.) So Fodor seems to be commi ed
to the claim that the emi ed tannin molecules are about Kadu bitings.
17
Narrow versus wide content
In the Supervenience subsection we considered whether two individuals,

A and B, who are narrowly identical, are necessarily mentally identical. The
example of Swampman suggests that they are not since Swampman and I are
narrowly identical and yet only I can remember the Sydney Harbour Bridge.
In recent decades a closely related question has arisen concerning content.
If A and B are narrowly identical, is it necessarily the case that their mental
states share the same content? A famous thought experiment due to Putnam
suggests that it is not (Putnam, 1975).9 Consider Oscar who lives on Earth and
uses the word water to refer to the stu that flows from taps, fills the oceans,
is necessary for human life, etc. Assume also that he knows nothing of chemis-
try; in particular, he does not know that water is H2O. Now it so happens that,
in the far reaches of the galaxy, there is another planet exactly like Earth in all
respects except that the stu that flows from taps, fills the oceans, is necessary
for human life, etc, is XYZ not H2O. XYZ cannot be distinguished from H2O
except by chemical analysis, and any biological process that involves H2O
proceeds just as well with XYZ substituted for H2O. Lets call this distant
planet Twin Earth, and call the inhabitant of Twin Earth who is identical to
Oscar in all physical respects except for having XYZ molecules where Oscar
has H2O molecules, Twin-Oscar.
Twin-Oscar and Oscar are narrowly identical in all relevant respects. Never-
theless, it seems that when Oscar has the thought I want a glass of water he is
thinking about H2O, whereas when Twin-Oscar has the thought I want a glass
of water he is thinking about XYZ. It follows that content is wide: what a
thought is about supervenes not only on the thinkers brain (narrowly con-
ceived) but also on the thinkers environment. This view is sometimes referred
to as anti-individualism about content because the content of an individuals
thoughts is held to supervene on states beyond the individual. Tyler Burge has
developed and discussed a number of cases akin to the Twin Earth example in
defence of a wide ranging anti-individualism in the philosophy of mind (see,
especially, Burge, 1979, 1986a).
Say that Oscar and Twin-Oscar both think Beer is 90 per cent water. Accord-
ing to the considerations pursued in the last two paragraphs, Oscars beer
thought is about the proportion of H2O in beer; Twin Oscars thought is about
the proportion of XYZ in beer. Some theorists, however, think that there is a
sense in which Oscars and Twin-Oscars beer thoughts are the same and that,
in a certain sense, their thoughts have the same content. Se ing aside the fact
that Oscars brain contains H2O exactly where Twin-Oscars brain contains XYZ,
their brains are identical; in particular, the narrow subvenient bases of their
beer thoughts are identical. It follows that Oscar and Twin-Oscars beer thoughts
18
will bring about exactly the same behaviour and will be expressed by u ering
exactly the same sounds or by writing exactly the same marks.
Many philosophers have marked a distinction between wide (or broad)
content and narrow content. Oscars thought that beer is 90 per cent water has
dierent wide content to Twin-Oscars thought that beer is 90 per cent water,
but the same narrow content. As we have seen, wide content supervenes on
the agents wide properties; in contrast narrow content supervenes only on the
agents narrow properties. Some philosophers have argued that psychology
needs only narrow content; in contrast, other philosophers have wondered
whether narrow content is really a kind of content at all. (For a wide ranging
discussion and defence of narrow content, see Segal, 2000.)
Mental Causation
Common sense tells us that physical and mental properties causally interact:
bodily damage causes pain and pain causes wincing. Moreover, our common-
sense notions of agency and responsibility invoke mental to physical causal
relations. We have dierent moral and aective a itudes towards those whose
destructive behaviour is caused by their intentions to behave destructively
than we do towards those whose destructive behaviour is not caused by such
intentions. If it were to turn out that there are no causal relations between
mental and physical properties many of our most cherished views about
human life would have to be reassessed. Fodor makes this point in an especially
dramatic way:
[i]f it isnt literally true that my wanting is causally responsible for my

reaching, and my itching is causally responsible for my scratching, and my
believing is causally responsible for my saying . . . if none of that is literally
true, then practically everything I believe about anything is false and its the
end of the world. (Fodor, 1990b, p. 156)
We have already noted (in the Dualism subsection) that interactionist

dualism (whether of the substance or property variety) has a problem with
mental causation. It might have been hoped that embracing physicalism would
remove this diculty. No such luck. In fact, the physicalist faces several distinct
problems of mental causation (Kim, 1998, Chapter 2). Perhaps these problems
are actually dierent faces of a single problem which will succumb to a single
solution, but that has yet to be shown. In this section I briefly describe three
physicalist problems of mental causation.
19
Mental causation and the anomalousness of the mental
As we saw in the Non-Reductive Physicalism subsection, Davidson denies the

existence of psychophysical laws. If, as Davidson believes, a causal relationship
between types A and B requires a corresponding law linking A and B, the
absence of psychophysical laws seems to imply the absence of psychophysical
causation. Davidson avoids this unhappy conclusion by proposing token iden-
tities between mental states and physical states. True, the mental properties of
those physical states cannot causally impact on other physical states; however,
the physical properties of those physical states can engage in causal relations
with other physical states. At first glance, then, the problem of mental causation
is solved.
But a nagging worry remains. What we require is an account of how
mental properties causally impact on the physical world, and that is exactly
what Davidson has failed to deliver. On Davidsons account, it is the physical
properties with which mental properties are correlated that have the causal
power to influence the world the mental properties are merely riding piggy
back on the causally ecient physical properties to which they are bound.
In other words, Davidson seems to be commi ed to a form of epiphenomenal
property dualism. (For discussion of anomalous monism and mental causation
see Davidson, 1993 and Heil, 2008.)
The exclusion problem
In the Non-Reductive Physicalism subsection we noted that substance dualism

faces a problem of overdetermination. Substance dualism allows that some
physical states are caused by mental states, for example, that mental state M
caused physical state P. However, if the world is physically closed, P will itself
have been caused by a prior physical state P*. Overdetermination now threat-
ens: both M and P* are causally sucient for P. A close analogue of this problem
arises for non-reductive physicalist theories of the mind. According to most
contemporary physicalists, the mental supervenes on the physical. Say that one
of my mental states, M, supervenes on physical property P* of my brain, and
that M causes physical state P. The claim that the world is physically closed
entails that P has physical causal antecedents which include, presumably, P*.
So again over-determination threatens: both M and P* are causally sucient
for P. One way to remove the threat of over-determination is to endorse prop-
erty epiphenomenalism: mental properties supervene on physical properties
but are causally inert, but this is deeply counterintuitive. As mentioned above,
our commonsense notions of responsibility and agency appear to require that
at least some human behaviours are caused by mental properties. Jaegwon Kim
20
calls this problem the exclusion problem because mental properties appear to
be excluded from causal interactions with the physical world (see, especially,
Kim, 1998, Chapters 2 and 3.) In the Some Possible Responses to the Exclusion
Problem subsection below I will briefly canvas some responses to the exclusion
problem.
Mental causation and representational content
Sallys desire for tomatoes caused her to purchase tomatoes. If Sally had had
a dierent desire (say for peppers), she would not have purchased tomatoes:
her desires being about tomatoes is what caused the tomato purchase. More
generally, an intentional states representational properties play a role in deter-
mining its causal relations. Now we saw in the Narrow v. Wide Content above
that, according to many philosophers, the representational content of an inten-
tional state depends in part on the agents environment (water refers to H2O
in my mouth but XYZ in my Twins mouth). So if intentional states have
causal powers they have them in part because of their wide properties. How-
ever, the causal powers of an object are entirely determined by its narrow
properties. Fodor makes this point with the following example (Fodor, 1987,
Chapter 2). A quarter activates a vending machine and causes it to emit a
Coke in virtue of its mass, shape and size. Mass, shape and size are all narrow
properties. The wide properties of the coin for example, that it was minted on
a certain date or in a certain place are irrelevant to its causal properties. It is
this fact that makes counterfeit coins possible. If a counterfeiter can succeed in
making a metallic disc with the same narrow properties as a quarter, then he or
she can steal Coke from Coke machines. The upshot of these considerations is
that the representational properties of intentional states are epiphenomenal:
they have no impact on the world.
Some possible responses to the exclusion problem
In this subsection I will focus on possible resolutions to the exclusion problem.

Many solutions have been proposed; I will restrict myself to briefly discussing
three kinds of solution.
The Return to Reductive Physicalism

The exclusion problem would be blocked if it could be shown that mental
property M does not merely supervene on physical property P*, but is identical
to P*. If M is identical to P* then the issue of over-determination of P by both
M and P* does not arise. (It is important to recall at this point that properties
21
are types, and so the identities we are considering are type identities.) As we saw
in the subsection on reductive physicalism, reductive physicalists propose that
mental properties are identical to physical properties. So one way to resolve
the exclusion problem is to embrace reductive physicalism. But that involves
rejecting the plausible claim that mental states can be multiply realized. (For an
important assessment of the prospects of rehabilitating reductive physicalism
as a resolution of the exclusion problem, see Kim, 1998, Chapter 4.)
Program Explanation
Frank Jackson and Philip Pe it (1988; 1990) propose that while mental proper-
ties are not causally ecacious that is, mental properties are strictly speaking
epiphenomenal they are nevertheless causally relevant because they pre y
much guarantee that causally ecacious states are present:
The property-instance does not figure in the productive process leading to
the event but it more or less ensures that a property-instance which is required
for that process does figure. A useful metaphor for describing the role of the
property is to say that its realization programs for the appearance of the pro-
ductive property and, under a certain description, for the event produced. The
analogy is with a computer program which ensures that certain things will
happen things satisfying certain descriptions though all the work of producing
those things goes on at a lower, mechanical level. (Jackson et al., 1990, p. 114)
As a consequence of the causal relevance of mental states we can have
powerful explanations of behaviour in which mental states figure, even though
those mental states are not causally ecacious. This strategy concedes some-
thing to the various problems of mental causation without giving up on the
idea that we can predict and explain behaviour by appealing to mental states.
One diculty with this view is that computer programs are causally eca-
cious: the lines of code which constitute a computer program cause (in the right
environment) lines of code in a machine language which in turn cause the
computer to behave in the desired way. So programming cant be taken too
seriously as a metaphor for causal relevance without causal ecaciousness. As
we have seen, it is widely accepted that the mental states supervene on physical
states (in the Supervenience subsection). Perhaps the idea that M programs for
P means only that M supervenes on P (see Kim, 1998, p. 74). But that does not
seem to be very helpful. When M supervenes on P, M nomologically depends on P.
In contrast, the idea of causal relevance requires the dependence of P on M. Its
not clear, then, that the program explanation idea helps us understand how
mental states might be causally relevant without being causally ecacious.
Dual Explanandum Solutions

The exclusion problem arises because it appears that the physical properties of
the brain are sucient to explain behaviour; there is not enough work for both
22
mental properties and physical properties to do, so the mental properties are
rejected as causally inecacious. Dual explanandum solutions to the exclusion
problem challenge the idea that there is nothing for the mental properties to do.
Heres an example. Say that Trudys brain instantiates physical property P*, and
that P* is causally sucient to make her arm rise. P* is not, though, sucient to
explain one important aspect of Trudys arms rising: it is not sucient to explain
the fact that Trudys arms rising was the casting of a vote. Trudys arm raising
has two aspects its physical shape and its property of being a vote casting.
The former aspects of the arm raising are fully accounted for by the presence
of P*. However, the presence of P* cannot account for the arm raisings being a
vote casting. The vote casting aspect of the arm raising was caused inter alia
by Trudys intention to vote. We can only make sense of the distinction between
what we might call mere behaviour and intentional action by allowing that
mental states are causally ecacious. Making such an allowance does not
reintroduce the problem of over-determination because P* is not sucient to
explain Trudys action being a vote casting. Solutions along these lines are
advanced in, for example, Yablo 1992 and Thomasson 1998.
One way to understand Princess Elizabeths worry about Cartesian substance
dualism (see the Dualism subsection above) is as a demand for an account
of the mechanisms which link mental and physical properties. That worry
re-emerges here: precisely how is it that Trudys intention to vote caused her
arm raising to be a vote casting? Without such an account it is tempting to
reverse the argument: there are no vote castings because, in order to count as a
vote casting, an arm raising must be caused by an appropriate intention, and
we cannot give an account of how intentions impact on the world.
Consciousness
In the Mind and Body section above, I quoted Colin McGinns question, How
can technicolour phenomenology arise from soggy grey ma er? (McGinn,
1991, p. 1). McGinn is articulating a challenge: explain how consciousness can
exist in a purely physical universe. More precisely, the challenge is to explain
how phenomenal consciousness can exist in a purely physical universe. Phe-
nomenal consciousness is the term used to refer to the subjective properties of
experiences. Thomas Nagel (1974) identified phenomenally conscious experi-
ences as those which it is like something to have. For example, there is something
that it is like to stare at a brightly lit scene, and there is something that it is
like to smell smoke on a damp autumn evening. Phenomenal consciousness is
o en contrasted with access consciousness. A mental state is said to be access
conscious if it is (a) inferentially promiscuous that is, available for use in a
wide range of reasoning tasks, and (b) readily available for the rational control
23
of action, including speech (see Block, 1994, 1995). Ned Block argues that a
mental state can be access conscious without being phenomenally conscious, or
phenomenally conscious without being access conscious. As an example of the
former he oers the mental states of philosophical zombies: creatures function-
ally identical to humans but lacking phenomenal consciousness. As an example
of the la er he oers, rather controversially, the example of suddenly becoming
aware that the refrigerators compressor, which has been humming for some
time, has stopped. According to Block, we were phenomenally conscious of the
compressors humming all along, but were not access conscious of it; it is only
when the compressor stopped that we could report that it has been humming.
Block stresses that some solutions to the problem of consciousness involve
conflating access and phenomenal consciousness. That is, the author advertises
his or her theory as an account of phenomenal consciousness, but actually
provides a theory of access consciousness.10
For present purposes I will use term the consciousness to refer (exclusively)
to phenomenal consciousness and conscious to mean to phenomenally con-
scious. A further terminological note: I will use the term qualia to refer to
the subjective properties of conscious experiences. Thus the experience of twist-
ing ones ankle has the qualia of hurting; the experience of staring at the sky on
a clear day has the qualia of blueness; and the experience of really wanting a
cigare e has the qualia of craving.11 Occasionally philosophers use qualia
in such a way that, by definition, qualia are non-physical. However, I will
use qualia in a way that leaves open the issue of whether qualia are physical
or not.
I will briefly sketch three kinds of responses to the challenge of locating con-
scious properties in the physical world. David Chalmers (2003a) has oered a
very useful taxonomy of positions in the metaphysics of consciousness. Where
appropriate, I will indicate where positions on my taxonomy map on to his.
Physicalism
The physicalist believes that consciousness is indeed a physical phenomenon.

Dierent versions of physicalism can be distinguished along a number of
dierent dimensions. I will distinguish optimistic, pessimistic and uncommi ed
versions of physicalism. These versions of physicalism are located along an
epistemic dimension: they vary in the degree to which they take the problem
of consciousness to be humanly solvable. Chalmers (2003a) distinguishes
between types A, B and C materialism (i.e. physicalism), however, the distinc-
tions among types of physicalism he makes do not map precisely onto those
I make here.
24
Optimistic Physicalism
The optimistic physicalist believes that physicalism is true and that, moreover,
human cognitive capacities are up to the task of locating consciousness in the
physical world. Two strands of optimistic physicalism can be identified.
Strongly ptimistic physicalists believe that we have already made significant
progress towards understanding consciousness in physical terms. A very wide
range of theorists fall into this category, including Gilbert Harman (1990),
Daniel Denne (1991a), Fred Dretske (1995) and Frank Jackson (2003).
Weakly optimistic physicalists believe that while we are yet to make significant
progress on the problem, there are no good reasons to believe that we wont
do so in the future. Thomas Nagel (1974), for example, argues that we are cur-
rently unable to reason from a physical description of the brain to a description
of its phenomenal properties because we lack the required concepts. Future
research, though, may one day provide those concepts.
Pessimistic Physicalism
The pessimistic physicalist believes that while consciousness is a physical
phenomenon, humans will never achieve a completely satisfying account of
how consciousness arises in the physical brain. Two strands of pessimistic
physicalism can be identified.
Strongly pessimistic physicalists believe that while consciousness is a physical
phenomenon, developing adequate theories of the emergence of consciousness
from the physical brain will forever transcend human cognitive capacities.
The relevant theories reside, as it were, in a species-wide, cognitive blind spot.
Colin McGinn holds this position (McGinn, 1991, especially Chapter1). While
he calls it transcendental realism, it is sometimes referred to as new mysteri-
anism. (The la er term is Owen Flanagans. See his 1992.)
Weakly pessimistic physicalists believe that, while we may obtain and per-
haps already have obtained an adequate physicalist theory of consciousness,
we are likely to find any such theory unsatisfying. We will grasp the relevant
physical theory, and follow each step of the physicalist explanation, but the
conclusion will not force itself upon us. An analogy: many people understand
the four dimensional theory of time, but cant shrug o the intuition that time
flows. Similarly, we may understand the physicalist theory of consciousness
but be unable to shrug o the intuition that consciousness stands apart from the
physical. Philip Pe it has articulated a view of this sort (Pe it, 2009).
Uncommitted Physicalism
The uncommi ed physicalist accepts that consciousness is a physical phenom-
enon, but expresses no view on whether or not satisfying physicalist theories of
consciousness are available. In the 1990s David Braddon-Mitchell and Frank
25
Jackson articulated a response to the so-called knowledge argument (see

below) which can be interpreted as an argument for uncommi ed physicalism
(Braddon-Mitchell et al., 1996, pp. 1345).
Anti-physicalism
The anti-physicalist denies that consciousness is a physical phenomenon.

There are a number of dierent versions of anti-physicalism, of which I will
mention four. These positions are not of merely historical interest: a number of
contemporary philosophers have defended anti-physicalism about conscious-
ness (e.g. see Lockwood, 1989; Chalmers, 1996, 2003a; and Stoljar, 2001b).
Interactive Substance Dualism

According to interactive substance dualism, consciousness is a (non-physical)
property of non-physical mental substance (see the Dualism subsection above).
This doctrine holds, in addition, that conscious properties may be caused by
certain brain properties, and may in turn bring about other brain properties.
(This is type D dualism on Chalmers [2003a] taxonomy.) In the Dualism sub-
section we noted the diculties interactive substance dualism has with mental
causation; the existence of qualia give rise to a special version of those di-
culties. How does the neural activity associated with pain bring about the
(non-physical) qualia of pain? And how does the (non-physical) qualia of pain
bring about the neural activity responsible for expressions of pain?
Epiphenomenalism about Qualia

According to epiphenomenalism about qualia, conscious properties are non-
physical properties which are caused by physical properties of the brain, but
which do not cause physical properties of the brain. In addition, epiphenome-
nalism about qualia asserts that all other mental properties are physical proper-
ties. Se ing aside any phenomenal properties which some beliefs may have,
beliefs are, according to epiphenomenalism about qualia, purely physical states.
(Epiphenomenalism is type E dualism on Chalmers [2003a] taxonomy.) We
will shortly examine a striking argument in favour of epiphenomenalism about
qualia.
Emergentism
According to emergentism, certain complex physical arrangements of ma er
(e.g. human brains) cause new, non-physical mental properties to emerge. The
emergence of non-physical mental properties from complex arrangements of
ma er is a brute fact about the world. In particular, the existence of mental
26
properties is not physically necessitated; that is, the structure of the physical
brain, together with the laws of physics, is insucient to bring about the mental
properties. The classic presentation of this view is Alexander (1920); for a more
recent discussion see McLaughlin (1992).
Phenomenal Fundamentalism
This position, once defended by Bertrand Russell (1927), holds that the intrinsic
properties of the fundamental physical entities are phenomenal (or perhaps
protophenomenal). On this view, the fundamental physical properties such as
mass are in fact relational properties among entities whose intrinsic nature is
phenomenal. (Phenomenal fundamentalism is type-F monism on Chalmers
[2003a] taxonomy.)
Eliminativism about Consciousness
We saw in the Eliminativism, Instrumentalism and the Intentional Stance

subsection that eliminativism is the doctrine that there are no mental states.
The eliminativist about consciousness advances the more limited claim that
there are no phenomenally conscious mental states. There are, very broadly
speaking, two kinds of eliminativism about consciousness.
Strong Eliminativism about Consciousness

According to the strong eliminativist about consciousness, there is no problem
of locating the conscious properties in the physical world because there are no
conscious properties. It is as foolish to debate the metaphysical status of con-
sciousness as it would be to debate the metaphysical status Aristotles crystal
spheres. One important advocate of a view of this sort is Daniel Denne , who
has argued that there is no principled distinction between processes which
develop the content of an experience and post-experiential tamperings with
that experience (Denne , 1991a; 1994). The determinate experience of a red
circle moving le at a certain speed is a cognitive illusion cast by the working
brain. This illusion is significant to us as human beings, but there is nothing
here for the metaphysician to explain, much less worry about. (For a helpful
presentation of Denne s views, see Akins, 1996.)
Weak Eliminativism about Consciousness

Weak eliminativism about consciousness denies that the experiences we group
together as phenomenal form a natural kind. Paul Griths has pointed out
that the emotions may not form a natural kind, and if they do not, it will
make no more sense to categorize psychological events into emotional and
27
non-emotional kinds than to categorize astronomical events into super- and

sub-lunary kinds (Griths, 1997, pp. 12). Similarly, Isabel Gois has argued that
the experiences we identify as phenomenal do not form a natural kind
(Gois, 2007).
The knowledge argument
In the 1980s, Frank Jackson articulated an argument for epiphenomenalism

about qualia the knowledge argument (Jackson, 1982 and 1986). I will briefly
review this argument and some of the countermoves which have been made
against it. Many of the ideas discussed in this overview come together in
assessing Jacksons argument.
(1) If physicalism is true then someone who knows everything about the
physical knows everything simpliciter.
(2) It is not the case that someone who knows everything about the
physical knows everything simpliciter.
Therefore,
(3) Physicalism is false.
(1) seems obviously true. Since physicalism is the doctrine that everything is
physical, the truth of physicalism entails that a person who knows everything
about the physical knows everything simpliciter. Jackson defends (2) by means
of a famous thought experiment. Mary is a brilliant scientist whose colour
visual system is normal but who has been raised from birth in a black and
white environment. She learns everything about the physical aspects of the
human visual system, but has never experienced the colour qualia; in parti-
cular, she has never experienced the qualia of red. Upon her release from the
black and white environment Mary is exposed to a red surface in good light
and exclaims Now I know what red looks like. It is natural to say that Mary
learnt something when she le the black and white environment; that is, it is
natural to say that she gained knowledge of the qualia of red. But if she gained
knowledge then she must have previously lacked knowledge. Thus, even
though (by hypothesis) she knew everything about the physical aspects of
human colour vision, she did not know everything simpliciter. The conclusion,
(3), follows from (1) and (2) by modus tollens.
Upon her release from the black and white environment Mary learnt
something about the qualia of red it was knowledge of the colour qualia that
escaped her when she was in the black and white environment. Since she knew
28
everything about the physical aspects of colour vision when she was in the
black and white environment, the colour qualia must not be physical:
(4)The colour qualia are not physical.
Moreover, Jackson endorses the claim that the world is physically closed:
(5)Every physical event has purely physical causal antecedents.
It follows from (4) and (5) that
(6)Qualia are epiphenomenal; that is, they have no physical eects.
A great many responses have been made to the knowledge argument

including responses made by Jackson himself. (For an excellent survey of
responses to the knowledge argument see Van Gulick, 2009.) Heres an early
response by Braddon-Mitchell and Jackson (Braddon-Mitchell et al., 1996,
pp. 1345):
(7)Epiphenomenalism about qualia is false.
Therefore,
(8)Either qualia are physical or the world is not physically closed.
However,
(9)The world is physically closed.
Therefore,
(10)Qualia are physical.
In support of (7), Braddon-Mitchell and Jackson point out how dicult it

is to make sense of the Mary thought experiment if epiphenomenalism about
qualia is true. For example, if epiphenomenalism about qualia is true, then it
is not the case that Marys exclamation Now I know what red looks like
was caused by her exposure to the qualia of red. Moreover, if we accept that
direct knowledge of the qualia of red involves a causal connection between an
instance of the qualia of red and the tokening of the relevant knowledge state,
then epiphenomenalism about qualia renders it impossible that Mary gained
direct knowledge of the qualia of redness: having the qualia of redness could
29
not have caused her to be in a state of knowledge about the qualia of redness.
(8) expresses the fact that there are only two ways to close o the argument
from the conclusion of the Mary thought experiment to epiphenomenalism by
denying either (4) or (5). Braddon-Mitchell and Jackson assert (9) because they
accept that modern science overwhelmingly supports it. The conclusion, (10),
follows from (8) and (9) by disjunctive syllogism.
Braddon-Mitchell and Jackson call their argument the there has to be a reply
reply. It is a very striking example of what I earlier called uncommi ed phy-
sicalism. Uncommi ed physicalism accepts that consciousness is a physical
phenomenon, but is uncommi ed on the question of whether or not satisfying
physicalist theories of consciousness are available. The there has to be a reply
reply concludes that consciousness is indeed physical, but oers no view on
whether humans will be able to arrive at a good understanding of conscious-
ness as a physical phenomenon.
More commi ed physicalist responses to the knowledge argument can
be grouped into bold and modest versions. Bold physicalist responses to the
knowledge argument insist that Mary learned nothing upon release from the
black and white environment. On this view, the common intuition that Mary
learned something when she finally le the black and white environment is
mistaken. In contrast, modest physicalist responses claim that, although Mary
knew all the physical facts, there were still things she had to learn about red
a er her initial exposure to red surfaces. The a raction of modest physicalist
responses to the knowledge argument is that they preserve the common
intuition just mentioned: that Mary gains knowledge when she leaves the
black and white environment. The diculty for modest physicalist replies
is explaining how Mary gained knowledge even though she already knew all
the physical facts.
A variety of modest physicalist replies to the knowledge argument exist in
the literature. Laurence Nemirow argued that Mary would have all the relevant
propositional knowledge (knowledge that) about colour qualia when she was
in the black-and-white environment, but lacked certain skills (knowledge
how) (Nemirow, 1980; see also Lewis, 1990). The intuition that Mary learns
something when she leaves the black and white environment is explained by
the fact that Mary learns the relevant skills (e.g. she can now imagine a red
surface); however, there were no facts with which she was unfamiliar prior to
her release. Jackson has advanced an ingenious argument against the skills
response. He points out that prior to her release Mary would not know
what other peoples mental lives are like. What, she might wonder, is it like
for ordinary people to look at a ripe tomato in good light? Jackson urges that
it is implausible that Mary is wondering about other peoples skills; it is facts
about other people that she is missing (Jackson, 1986).
30
Another kind of modest physicalist response turns on the idea that the same
propositional knowledge may be stored in dierent representational media.
David Lewis advanced a view of this sort by way of an analogy:
Imagine a smart data bank. It can be told things, it can store the information
it is given, it can reason with it, it can answer questions on the basis of its
stored information. Now imagine a pa ern-recognizing device that works
as follows. When exposed to a pa ern it makes a sort of template, which it
then applies to pa erns presented to it in future. Now imagine one device
with both faculties, rather like a clock radio. There is no reason to think that
any such device must have a third faculty: a faculty of making templates
for pa erns it has never been exposed to, using its stored information about
these pa erns. If it has a full description about a pa ern but no template
for it, it lacks an ability but it doesnt lack information. (Rather, it lacks
information in a useable form.) When it is shown the pa ern it makes a
template and gains abilities, but it gains no information. We might be
rather like that. (Lewis, 1983b, pp. 1312)
Lewis describes a case in which a machine possess all the relevant proposi-
tional knowledge represented in a sentence-like medium, but does not possess
all the relevant knowledge in an alternative, analogical medium. Similarly,
before her release Mary might possess all the relevant propositional knowledge
in a sentence-like medium, but only a er her release does she acquire new
representations of that knowledge in an alternative, phenomenal, medium.
Paul Churchland presses a similar point, and provides reasons for thinking
that the human brain does indeed contain two (or more) distinct representa-
tional media with only a limited ability to translate between them (Churchland,
1989a). However, a nagging doubt remains. There is something that it is like to
have a phenomenal representation in the alternative medium. And knowledge
of what it is like to have a phenomenal representation in the alternative medium
is exactly what Mary seems to be missing.
In recent years Jackson has provided his own modest response to the
knowledge argument (Jackson, 2003). His response turns on two claims
(11)Qualia are essentially representational.

(12)Mental representation is an entirely physical phenomenon.
From which it follows that
(10)Qualia are physical.
31
In defence of (12) Jackson asserts that, while we do not yet have agreement
on the correct form a physicalist theory of mental representation should take,
we can nevertheless be confident that such a physical theory is in principle
available. The idea that qualia are essentially representational is not original to
Jackson, having been articulated by, for example, Gilbert Harman (1990) and
Michael Tye (2000). On this view, qualia are distinguished from other forms
of mental representations in virtue of their unique functional roles. Pe it puts
this point succinctly:
A state will count as experiential so far as it functions in a manner typical of
experiences: it generally disposes an agent to come to believe that things are as
they are represented to be; it does not control behaviour except when it leads
to belief; it may remain in place continuing to represent things being thus
and so, even when the subject has come to believe that they are not that way
. . . and so on. (Pe it, 2009, p. 169)
An important feature of many modes of representation is that the represen-
tation need not have the property it represents. Linguistic representations are
like this. The following token
green
represents the property green but is not green. Similarly, my qualia of redness
which, according to representationalism about qualia, represents surfaces as
being red, need not itself be red. Its long been observed that, when I have a
mental image of a ripe tomato, there need be nothing red in my brain (e.g. see
Smart, 1959). It has been suggested, though, that my image has the property of
being phenomenally red of instantiating a special phenomenal property which
is sometimes referred to as phredness. Representationalists about qualia
deny this. My qualia of redness represents red without being either red or
phred. The what it is like of the experience I have when I look at a ripe tomato
in good light is entirely exhausted by the fact that it represents the tomato as
being red. There is nothing else to be explained (see, especially, Jackson, 2003.)
Jackson accepts, however, that Mary still makes an epistemic gain when she
leaves the black and white environment; that is, he articulates what I have
called a modest physicalist reply to the knowledge argument. While Jackson
asserts that Mary knew all the propositional knowledge about qualia prior to
her release, he does not claim that Mary learned nothing about qualia upon
her release. In particular, he endorses Nemirows claim that Mary acquired
certain skills upon her release.
In response to Jacksons representationalist reply to the knowledge
argument, Robert Van Gulick has argued that physicalists can successfully
respond to the knowledge argument without endorsing representationalism.
The knowledge argument contains, he submits, a number of assumptions which
32
the physicalist can challenge. (See Van Gulick, 2009.) Chalmers has raised
objections directly against representationalism (Chalmers, 2003a, p. 111). He
distinguishes between functional and phenomenal representation as follows. A
system has a functional representation of p when it responds to p appropriately.
For example, I have a functional representation of the red trac light when
I respond to it by braking. In contrast, a system has a phenomenal representa-
tion of p when the system is phenomenally conscious of p. For example, I have
a phenomenal representation of the red trac light when there is something
that it is like for me to see the red light. (This distinction is related to Blocks
distinction between access and phenomenal consciousness, introduced at the
beginning of this section.) Chalmers worry is that Mary could have full
knowledge of the functional representational properties of the qualia of red
without having knowledge of the phenomenal representational properties of
red. That is, phenomenal properties cannot be reduced to or even multiply
realized by functional properties.
Conclusion
For most of the history of the philosophy of mind, the mental realm was
taken to be exhausted by the phenomenal realm. But under the impact of
logical positivism and behaviourism in the first half of the twentieth century,
the mental was reconceived so that consciousness was only one, inessential,
aspect of the mental. (Perhaps the last great philosophical work which took
mind and consciousness to be synonymous was C. D. Broads The Mind and
Its Place in Nature [1925]). Once it was recognized that the mental was not
exhausted by the phenomenal, a change of focus took place. Central to that
change of focus was functionalism. By identifying mental states as the occu-
pants of characteristic functional roles, it became possible to identify mental
states with brain states, thus planting the mental firmly in the physical world
and providing new ways to think about the relationship between neuroscience
and psychology.
But two apparently intractable problems remain. Clearly, the problem of
mental causation cannot be functionalized away. Functional roles are causal
roles, and the functional roles characteristic of mental states include psy-
chophysical causal relations. The problem of mental causation is a challenge
to functionalism, not a puzzle that can be resolved by appealing to functional-
ism. The second apparently intractable problem is phenomenal consciousness.
Qualia are resistant to a purely functional approach. Mary knew all the func-
tional roles characteristic of the qualia of redness, and knew what states
occupy those roles. Nevertheless, it seems that there was something about
qualia which she did not know. Similarly, it seems that we can conceive of
33
philosophical zombies which are functionally identical to normal human

beings but which lack qualia.
I said in the first section that in many ways Descartes problems have
turned out to be our problems. Does that mean that we have made no progress
beyond Descartes? I dont think so. We are much clearer now on the scope of
the problems and on their complexity. And we have a much greater range of
sophisticated tools and concepts to bring to bear on them. But we should
acknowledge, with a considerable degree of humility, our historical debts. As
Newton might have said, if we can see further than our predecessors, it is only
because we are standing on the ideas they bequeathed us.
34
2 Consciousness
Daniel D. Hutto
What is Consciousness?
There is no u erly clean, clear and neutral account of what exactly is covered
by the concept of consciousness. The situation reflects, and is exacerbated by
the fact that we speak of consciousness in many dierent ways in ordinary
parlance. A consequence of our multifarious uses of the concept is that it has
proved impossible to define its essential characteristics through conceptual
analysis. We have nothing approaching a descriptively adequate philosophical
consensus of what lies at the core of all and every form of consciousness in
terms of necessary and sucient conditions that would be accepted by all
interested parties.
This is not regarded as a cause for despair. The same is true of other philo-
sophically important topics such as knowledge and causation. Despite this,
consciousness remains of pivotal philosophical interest because of its centrality
to our psychological lives and the way that it tantalizingly resists incorporation
into a fully naturalized account of the world.
Recognizing that a empts to provide a philosophically robust definition of
consciousness are likely forlorn, a standard tactic for isolating core features
of consciousness is to provide clear-cut exemplars as specimens. By means of
this strategy we might still, at least, divine philosophically important a ributes
of the quarry. Take your experience of reading these lines. Hopefully their
content is at the focal centre of your a ention, but even if so there will be a
range of other peripheral and background things of which you are consciously
aware: colours, noises, feelings. Some of these may remain present throughout
your intellectual activity while others intrude upon it momentarily, in largely
expected but perhaps occasionally surprising ways, before vanishing from
the stage. Despite such comings and goings you will not feel as if your overall
experience is ruptured or fragmented.
Conscious experience of this sort is u erly mundane and intimately familiar.
It appears to be an all-or-nothing property that pervades the waking lives of
many creatures. Human beings, cats, octopi (apparently), and spiders (perhaps)
are kinds of beings commonly thought capable of possessing consciousness
while inanimate objects, such as chairs, are not. We say of creatures or organisms
35
that they are conscious if they are awake and sentient. Evidence of this is that
they exhibit a certain degree of sensitivity or coordination with respect to
aspects of their environment. However if the case just described is taken as
our paradigm, then merely exercising capacities for such responding will not
suce for being truly conscious. It is easy to think of examples of complex
intelligent activity, sometimes of a quite sophisticated kind, that are never-
theless apparently habitual, automatic, or unreflective.
Most philosophers insist that, minimally and necessarily, to be conscious it
must also be the case that a being possess or enjoy some degree of occurrent
experiential awareness. In other words, there must be something that it-is-like
for them to be awake, sentient or intelligently controlling its behaviour. A truly
conscious being enjoys experiences that have phenomenal aspects; it feels a
certain way to be such a creature in such and such circumstances.
Experiential awareness can take dierent forms. It may be transitive in the
sense of being awareness of environmental surroundings or aspects thereof.
For example, the subject may be aware of the red speck in the centre of its visual
field. For this reason consciousness is o en regarded as being inherently
intentional, as being directed at certain objects, not others. But it seems possible
to be experientially aware in more intransitive ways too in ways that lack
directedness at specific objects. Diuse and undirected forms of consciousness
are surely possible, as is the case with moods, such as elation, calmness or
depression. Other, even more basic forms of undirected conscious experience
are also imaginable. Either way, to repeat, being conscious appears to require
being in a state of mind with a characteristic feel one in which there is
something-that-it-is-like to be in it. This is seemingly common to all forms
of consciousness; or, more cautiously, at least there are interesting forms of
consciousness that have this feature necessarily.
Conscious beings are essentially experiencers and the particular types of
experiences that they enjoy have distinctive characteristics and notable aspects
(i.e. they have specific phenomenal properties or characters). Experiencing
itchiness, for example, is quite dierent from experiencing anger. Seeing the
peculiar greenness of an aloe vera plant diers from seeing the peculiar green-
ness of a Granny Smith apple. We can specify the dierences by using illocu-
tions such as this or that shade of greenness. But this is to invoke inevitably
crude and (still) relatively abstract categories in order to pick out something
that is much more fine-grained, analog and particular.
Experiencing phenomenal characters, apparently, ma ers. Having experiences
seems to make a dierence to what is done, in line with how such experiences
are evaluated. Encountering the unusual taste and smell of durian, for example,
may evoke reveries or prompt certain other actions, depending on whether one
finds that taste pleasant or unpleasant. In line with this some are inclined to
reserve, more stringently, the accolade of being phenomenally conscious only
36
Consciousness
for those beings that exhibit a certain degree of global control over their actions
or that are capable of reporting, expressing and appraising how things appear
to them. To achieve this, it is argued that conscious beings must not only be
aware of and a end to aspects of their environment but to aspects of experi-
ential mental states themselves. Accordingly this kind of capacity implies at
least some degree of self-awareness or self-consciousness. If one accepts this,
subjects that are truly phenomenally conscious must not only enjoy experiences
with certain phenomenal qualities, they must be aware of the qualities of these
experiences. If so, those states of mind that exhibit phenomenal consciousness
do so at best only partly in virtue of having phenomenal characters.
Still, even if all conscious beings are experiencers of some or other phenom-
enal properties it may be that they experience these in more or less unified
ways. Human experience tends to integrate experienced phenomenal proper-
ties (i.e. those associated with dierent sensory modalities), continuously and
seamlessly over time. Recent empirical studies concerning the phenomenon of
ina entional or change blindness, raise doubts about the extent and degree to
which we actually experience the world in fully detailed and non-gappy ways.
Still, for human experiences at least usually it feels as if the way in which
our experiences inter-relate and change happens in coherent, well-coordinated
and expected manners.
Typical human consciousness, at least, feels as if it were, in important
respects, objective, temporally extended and unified. It involves having a coher-
ent and unified individual perspective on reality. These unique points of view
are internally complex. When we notice and a end to specific worldly features,
such as the greenness of a particular apple, this involves being able to see an
apple as something more than just the sum of its presented features. To see an
apple as something in which greenness, and other properties, might inhere is
to see it as having a continued existence over time. Experiencing a world of
objects and their features always occurs against a larger and more complex
background in which such items are systematically related to other things. To
have experience of the world, as opposed to merely having sentient capacities,
is to experience worldly oerings in a structured way.1
This entails, modestly enough, that dierent sorts of creatures may enjoy
dierent forms of consciousness. What it is like to be a human being may vary
considerably from what it is like to be a dolphin, or more famously still, what
it is like to be a bat. Indeed, even what it is like to be a particular human being
in a particular set of circumstances can dier qualitatively from what it is like
to be a particular human being in another set of circumstances.
Conscious experience is subjective at least in the sense that as Nagel (1974)
proposes it is idiosyncratic. Being phenomenally conscious apparently equates
to having a particular point of view or perspective that involves having a range
of more or less unified experiences with individual phenomenal characters.
37
Many philosophers hold that since we have no direct access to how things
appear experientially to others, it is enigmatic whether others are conscious or
what the exact character of their conscious experience is like. Thus unless it is
possible to securely infer what it is like for the other from more objective avail-
able facts, then, for all we know, even apparently sophisticated and intelligent
beings may lack conscious experience altogether, or they may enjoy experiences
that possess radically dierent phenomenal or qualitative characters from our
own. Moreover, the way that their experiences are normally integrated with
one another or unified (to the extent that they are integrated or unified at all)
may be quite alien to the way that typical human experience is organized.
Even if a fully transparent conceptual analysis of consciousness is not on the
cards, it seems that there are a number of identifiable or at least apparent
properties that are fundamental to it that make it of real philosophical interest.
Perhaps based on empirical or philosophical reflection it will be decided that
not all of these seeming a ributes are genuine; perhaps they will not all make
the final list of properties that warrant straight explanation. Nevertheless,
phenomenality, intentionality, subjectivity, unity, temporal extension, minimal
self-awareness are prima facie prominent features of consciousness that must
be either explained or explained away.
Reductively Naturalistic Frameworks
When thinking about consciousness, the mainstream tendency in contempo-

rary analytic philosophy of mind is to focus on metaphysical (as opposed to
conceptual) concerns. The working habit of those of a reductive naturalist bent
is to propose equations about mental states and their properties. For example,
they aim to provide general formula that will tell us, say, what conscious expe-
rience is by equating it to something else. In line with this agenda, a plethora
of theories of consciousness have been advanced. To mention but a few, these
include conjectures that equate consciousness with events or properties of the
neurobiological sort (Crick, 1994; Churchland, 1989b), quantum mechanical
(Penrose, 1994), functional/representational (Carruthers, 2000; Denne , 1991a,
2006; Dretske, 1995; Lycan, 1996; Rosenthal, 2000, 2005; Tye, 1996).
Such theories do not aim to provide traditional conceptual analyses. If
successful, ultimately they would tell us what is necessary and sucient for
having conscious experience (or conscious experience of a certain type) in the
same spirit that the property of being water was identified with the property of
being H20. That, of course, is a specific empirical hypothesis about a natural
kind. What is usually on oer by reductive naturalists who theorize about
consciousness is, as Putnam (1967) observes, typically not very detailed or
finished hypotheses but rather a kind of schemata for hypotheses.
38
Consciousness
Debates between naturalists of this stripe take the form of in-house assess-
ments of (and sometimes proposed adjustments to) each others headline pro-
posals. Not every framework is regarded as equally promising. For example,
the more extreme versions of behaviourism are almost universally unpopular
today. Their followers hold that any and all genuine mental phenomena need to
be identified with behaviours or dispositions to behave that allow for opera-
tional definitions, however complex, in terms of observable causes and eects.
A major criticism of this framework is that, in relying entirely on dispositions,
however complex, in order to understand the mental it lacks the essential
resources to satisfactorily account for the holistic and complex sorts of inter-
actions that occur between mental states that take place before responses
are produced. In this respect, functionalism, its natural successor, is deemed
superior because it makes space for precisely this sort of complexity.
Reductive functionalists take it that to be a conscious mental state of a certain
type equates to being a functional state of a whole organism: a state that can be
understood in terms of its wider systemic relations or teleological purposes and
that has appropriate causal relations to perceptions, other mental states and
actions. According to the analytic or commonsense version of the doctrine,
mental phenomena, including the experience of sights, sounds, pains and other
conscious mental states, are identified with specifiable higher order causal
roles. And for reductive naturalists these, in turn, are, either directly or indi-
rectly, identified with the physical states that happen to occupy, realize or fill
those roles. Rich mental activity is thus thought to take place between stimuli
and responses. Mental states, activated by environmental triggers, causally
interact in specifiable ways with each other and other bodily states, and only
then produce outward responses.
In large measure functionalisms popularity as a general framework for
thinking about mentality derives from the fact that it gives both philosophers
and psychologists the requisite apparatus and platform for positing inner, caus-
ally ecacious mental states without having to commit, in advance, to specific
details about how such mental states are physically realized or implemented.
This is useful because there appears to be an enormous stock of creatures of the
actual, terrestrial and imaginable, alien varieties that are capable of conscious
experiencing, despite the fact that they lack a physiology similar to our own.
The key functionalist insight is that not every creature which might be capa-
ble of conscious experience need have central nervous systems or brains like
ours. Consequently, we should not expect to uncover any neat, cross-species,
one-to-one correlations holding between particular types of experience and
particular types of neural or physiological states. To assume otherwise is, it is
claimed, to promote an unwarranted species-biased chauvinism.
Moreover, empirical work has revealed that brains, such as ours, are open
to re-wiring; that the neural structures underpinning certain types of mental
39
activity are highly plastic. For example, patients who have had hemispherecto-
mies in which the cortex of one hemisphere is removed have succeeded in
enlisting other parts of their brains to restore the lost functions, thus managing
to compensate. In the light of such discoveries it appears mistaken to assume
that there will be uniquely dedicated neural configurations supporting very
specific kinds of mentality. It is likely that this is true of conscious experience too.
Considerations of this sort cast doubt on strong, type-type versions of mind-
brain identity theories: those that propose straightforward identifications of
particular kinds of conscious experience with particular types of brain event.
Type-type identity theorists hope to provide class-to-class identifications of
brain events or processes that will be capable of grounding an interestingly
predictive and informative science of the mind. Such theories hope to tell us
that being a middle-A sound is identical with being an oscillation in air
pressure at 440 hertz; being red is identical with having a certain triplet of
electromagnetic reflectance eciencies; being warm is identical with a certain
mean level of microscopically embodied energies, and so forth (Churchland,
1989b, p. 53).
However if some version of functionalism is true then psychological
laws can be cast at a higher order level even though conscious states will be
variously, and thus disjunctively, realized in species specific (and perhaps even
individual-specific and/or circumstance-specific) ways. Consider that the hum-
ble carbure or can be functionally defined in terms of the abstract causal role
it fulfils. Something is a carbure or insofar as it mixes air with liquid fuel.
In theory, such devices could be made of metal, rubber, plastic, possibly even
soul-stu as long as they are capable of discharging the stated function. The
relation between a functional role and what realizes it can be one-to-many as
opposed to one-to-one. Thus a general description of the realizers of any given
type of experience would take the form of a disjunction on the right side of the
relevant equation that includes mention of a, perhaps, indefinitely long chain
of dierent kinds of instantiating states.
The trouble is that, without significant qualification, functionalism can
appear overly inclusive when it comes to saying which sorts of systems ought
to make the list of the conscious. This is illustrated by the fact that it is easy
to imagine beings that produce the appropriate outward behaviour through
functionally identical means but which plausibly lack any kind of experiential
awareness. For example, Ned Block (1978) famously imagined a scenario in
which the behaviours of a complex artificial body are orchestrated by com-
munications between members of the Chinese nation, so as to mimic in all
functionally relevant respects the responses of a human being undergoing
a painful experience. In order to achieve this feat, the Chinese citizens are pro-
vided with rules on how to respond to instructions provided by sky-based
display and are able to communicate with one another by means of two-way
40
Consciousness
radio-links so that, by working together, they are able to remotely generate just
the right kinds of responses in the body.
However outlandish the scenario, it is at least theoretically possible that the
Chinese nation could simulate human pain behaviour in the artificial body
by such means, mirroring at one level of description the ways in which such
behaviour is normally functionally produced in humans. The worry is that if
the two systems are identical in this respect then, assuming functionalism is
true, it appears we have no principled grounds for saying that one is really
undergoing the experience of pain and the other not. The situation is preposter-
ous since, intuitively, we want to ascribe the experience of pain to the individual
human being while withholding that ascription to an artificial body controlled
by the conglomerate of Chinese individuals engaged in this bizarre exercise. Yet
the only path to that verdict seems to require thinking of conscious experience
as something distinct from functional properties per se.
Indeed, this is precisely the moral that many are inclined to draw from this
thought experiment; while functionalism might reveal something important
about the processes and structures associated with experiencing it is incapable of
providing a real insight into the nature of experience itself. On these grounds
functionalists are o en accused of having no reasonable means of accommodat-
ing the phenomenal character of experience. Their proposals apparently miss
out the most important ingredient: how it feels.
Of course, it is possible for the functionalist to bite the bullet and insist, fly-
ing in the face of standard intuitions, that if a systems behaviour is generated
in the right way then it simply is conscious. So if the human is, then so too is
the body governed by the Chinese nation. But this is not the only (nor the
most convincing) line of reply. Arguably, the charge of liberalism might be
answered by adjusting the level of grain of the proposed functional analysis.
This requires a shi of a ention from abstract job descriptions crudely, what
the thing is doing to a greater focus on precise engineering details crudely,
how the thing actually does what it does (Churchland, 1989b, Chapter 2;
Flanagan 1991). Ironically, what is hailed as the chief virtue of functionalism
its capacity to abstract from specific details and its openness to the possibility
of various realizability looks to be a vice when it comes to understanding
consciousness.
Nevertheless, to assume that such objections are fatal is to underestimate
the flexibility of the functionalist approach. It is quite open for defenders of this
framework to insist, more in the spirit of identity theory, that the peculiarities
of how a system is organized and even what materials it is composed of might
ma er to having experience (or having certain kinds of experience). This is
wholly consistent with acknowledging that two systems that dier in their
lower level engineering details might be regarded as functionally equivalent at
some higher level of analysis.
41
What this possibility reveals is that, despite some traditional disagreements

and border disputes, there is scope for seemingly distinct and opposing reduc-
tive naturalistic frameworks to become more closely aligned and tightly ali-
ated. Indeed some authors go as far as to claim that the identity theory is just
an empirically special case of functionalism, one that (implausibly) locates all
mental states at the same very low level of institutional abstraction the neuro-
anatomical (Lycan, 1996, p. 59). In the other direction even avid supporters of
functionalism are quite happy to acknowledge that when you make a mind,
the materials ma er (Denne , 1997, p. 100). Importantly this thought is wholly
compatible with accepting that when it comes to understanding consciousness
handsome is what handsome does and that ma er ma ers only because of
what ma er can do (Denne , 2006, p. 17).
What this shows is that there may be ways and means of adequately dealing
with so-called absent qualia cases and that if these were the only cases that
had to be dealt with, then nothing in principle bars the development of a fully
illuminating naturalistic theory of consciousness. The real devil is simply in the
detail of deciding exactly which levels of functional analysis are most appropri-
ate for understanding various forms of conscious experience and which aspects
of the physical world are needed for instantiating such experiences. Deter-
mining that is a long term project that would be at least partly constrained by
further theoretical considerations as well as empirical findings.
Arguments For and Against Non-Reductive Naturalism
Not everyone is sanguine about the prospects of reductive naturalism. Indeed,

many are convinced that the whole style of approach is wrong for the study
of consciousness, full stop. This is because they hold that consciousness has
special properties that are distinct and irreducible to properties of any other
kind. Traditionally, this sort of view is associated with substance dualism of the
kind promoted by Descartes. According to substance dualists reality is bifur-
cated, composed of two quite distinct and fundamentally dierent substances:
the mental and the physical. Most of todays dualists are thoroughgoing
naturalists. They claim only that phenomenal properties are real, and although
wholly natural, they cannot be equated with any other kind of properties.
Rather they are primitive properties in their own right that exist alongside
representational, functional and physical properties. Hence to fully understand
how they relate to other worldly properties requires additional explanation
in terms of special fundamental laws, laws that are as basic as any others to be
discovered by a completed physics.
Understood non-reductively, the ultimate aspiration of a science of conscious-
ness is to connect first-person data to third-person data; perhaps to explain the
42
Consciousness
former in terms of the la er, or at least to come up with systematic theoretical

connections between the two (Chalmers, 1999a, p. 8). Ultimate success in this
venture would take the form of a fundamental theory which would explicate
the simple, universal laws that underwrite the principles connecting experi-
ences and informationally driven brain processes. Yet even those who are
commi ed to this project recognize that currently there is a lack of adequate
formalisms for characterizing and ge ing at experiential properties. Thus in
order for there to be a viable first-personal science of consciousness existing
methods of investigation must be substantially improved and developed, as
is promised by techniques such as neurophenomenology (for an overview
see Lutz and Thompson, 2003).
A range of thought experiments motivates belief in the metaphysical
dualism that suggests a need for research programmes of this kind. Ultimately,
reductive naturalists claim that physical features and facts exhaust all that
there is. Once the physical details of the world are in place, everything else
follows or is entailed automatically. Jacksons (1982) famous knowledge argu-
ment is designed to cast doubt on the truth of this view. It features a thought
experiment in which the central character, Mary, a super-scientist, knows every
physical fact it is possible to know. It is imagined that Mary has been confined
from birth in a wholly black and white environment and that she has only
had access to the outside world via the black and white media of television
monitors. Although she knows everything that is possible to know about
every physical detail of the world, her knowledge is, nevertheless, incomplete.
Specifically, she lacks knowledge of certain facts about colour experience.
Due to her confinement, she does not know what it is like, either for herself or
others, to experience colours. This is not something that she can learn without
being released from her monochrome prison. The moral is that if Marys factual
knowledge is incomplete, then knowing everything that there is to know about
physics is not to know everything that there is to know simpliciter. When given
a metaphysical twist, this conclusion is thought to imply that fixing the physical
facts simply doesnt fix all the facts, namely, that physicalism is incomplete.
An even more direct a empt to demonstrate the limitations of reductive
naturalism involves conjuring up zombies. Zombies are atom for atom, com-
plete physical duplicates of conscious beings; they are identical in every beha-
vioural, functional and physical detail, but despite this they are completely
lacking in conscious experiences. For zombies the lights appear to be on even
though no one is at home. This profound experiential vacancy is unnoticeable
for all intents and purposes. It is not something that you, I, or even the zombie
itself would be able to detect or articulate. It need not be supposed that zombies
are actually possible; likely they contravene existing laws of nature that apply
to our world. But this is entirely consistent with their existence as denizens of
some logically possible world (Chalmers, 1996, p. 180).
43
Conceiving of zombies seems to be no great strain. Their possibility seems

quite coherent. Thus if it is allowed that bona fide or coherent conceivability
implies metaphysical possibility then the mere possibility of zombies putatively
shows that experiential properties need not, always and everywhere, be tied to
or co-vary with behavioural, functional and/or physical properties. Experiential
properties have intrinsic features that are logically distinct from all other prop-
erties. If so, because of this conceptual wedge, all forms of reductive naturalist
theories of consciousness are false.
Moved by these considerations, non-reductionists are confident that all
a empts to understand conscious experience in terms of something else will be
unable to clear a fundamental hurdle; they will fail to deal squarely with the
so-called hard problem of consciousness. Chalmers (1996) formulation of it
takes the form of wondering for any given proposal about the nature of con-
scious experiencing (i.e. that identifies it with a certain functional organization,
representational properties, complex pa erns of neural or organismic activity,
the global access and integration of information, or what you will) how it is that
the proposed states or activities in question could give rise to or generate
conscious experience. The assumption is that no satisfactory answer is or will
be forthcoming. The very idea of making sense of the production of conscious-
ness in terms of something else is a fools errand. Thus consciousness remains
the biggest stumbling block to obtaining a complete understanding of the natu-
ral world using the standard categories oered by the objective sciences.
Notably, the hard problem is not just hard it seems impossible to solve.
Hence when it comes to explaining the relationship between the experiential
and other worldly properties the best that can be hoped for is a non-reductive
specification of relations that hold between them.
There are a number of ways reductive naturalists can respond to these sorts
of challenge. This always takes the form of trying to show that despite appear-
ances these thought experiments do not have the implications that they appear
to have. This can be handled piecemeal. For example, some hold that zombies
are not coherently conceivable, and thus nothing interesting follows for meta-
physics from the a empt to imagine them. In the case of Mary, some argue that,
despite the fact that she learns something new on leaving her room, what she
learns is not any kind of factual knowledge.
A more popular general strategy is to appeal to special features of phenom-
enal concepts that are distinct and irreducible to physical concepts, even though
phenomenal properties are not distinct from physical properties. This provides
space for insisting that even though zombies are genuinely conceivable, their
conceivability is not a faithful guide to metaphysical possibility. It also permits
one to deny of Mary that her epistemic situation, as described, need have any of
the interesting metaphysical consequences proposed. Knowing everything
44
Consciousness
there is to know under one conceptually based description does not entail
knowing the very same things under another.
What makes this line of reply plausible is that concepts of phenomenal
consciousness are apparently special in key respects. Their special features are
what systematically foster and explain the illusion that experiential properties
can come apart from all other properties, even though this isnt so. Public con-
cepts of experience, such as redness or itchiness are, some hold, recognitional
in nature. They are thought to involve perceptions of worldly properties.
In contrast, phenomenal concepts, such as seems red or feels itchy which
some hold are formed on the basis of re-enacting or having higher order per-
ceivings or believings about first order experiential states are regarded as
purely recognitional or inherently first-personally perspectival (Papineau, 2002;
Tye, 2009).
If phenomenal concepts are formed in special ways and have unique pro-
perties because of this, then it is arguable that it only seems to us that zombies
are possible (i.e. one can conceptually imagine phenomenal properties as being
distinct from all other properties even though they cannot be) and that it only
seems that Mary doesnt know all the facts (in fact she does, but she knows some
of them under a limited description or mode of presentation). If so, those who
are excited by the idea that such thought experiments damage the prospects of
naturalism are subject to a persistent cognitive illusion.2
Those who endorse the phenomenal concepts strategy place dierent bets
on the odds of closing the so-called explanatory gap (Levine, 1983). For even
if naturalists are able to put their metaphysical house in order there remains
a lingering and perturbing question of why a neural state should be the basis
of a certain kind of experience or of any experience instead of none at all. This
question suggests that there is a gap in our understanding that still appears
to remain wide open even if one denies the force of the standard thought
experiments.
One tactic for dealing with it, promoted by McGinn (1991), is to simply con-
cede that our minds are cognitively incapable of forming the relevant concepts
required for closing it (i.e. of providing a constructive, scientific account of con-
sciousness). This is so, he maintains, even though consciousness is a perfectly
legitimate natural and, indeed, wholly physical phenomenon. In principle it is
wholly explicable in physical terms even though we are, forever, cognitively
closed to understanding how this could be so. We are prevented from this
because we lack the appropriate cognitive faculties.
Top-down a empts to understand the psychophysical link between the
experiential and the physical are impeded on one side by the limits of intro-
spection. There is nothing in our experience that provides us with the means of
intelligibly understanding how experience is generated by the processes that
45
underwrite it. Bo om-up scientific approaches are similarly limited by the

perception-based methods they employ; hence they are equally unable to make
sense of the productive link between the physical and experiential. It does not
follow that the psychophysical link is inexplicable. All that is entailed is that
the missing link cannot be characterised or made intelligible in either physical
or mental terms. Given our inherent conceptual limitations, it is just that we are
forever prevented from making sense of the relation between experience and its
material substrate. The truth about the psychophysical nexus is out there but it
is permanently beyond our ken. Consciousness will always remain a mystery.
Understanding its place in the natural world is a perpetual, epistemic problem
but not a metaphysical one.
It might be wondered what, if we lack even the possibility of epistemic
assurances, could warrant this staunch faith in the truth of physicalism? What
justifies the idea that there exists an explanation of the psychophysical con-
nection, that is forever beyond our grasp? McGinns answer is that nothing
rationally justifies it, but rather it must be accepted as an article of metaphysical
faith (McGinn, 1991, p. 87).
Other reductionists reject this pessimistic a itude wholesale, believing it is
possible to solve the hard problem by changing the tools with which we cur-
rently think about consciousness (i.e. by fiddling with the concepts on both
sides of the equation until the mystery disappears). The assumption is that in
the course of time the concepts of mental and physical concepts will co-evolve
and a solution will be revealed as we bring them into correspondence (Van
Gulick, 2000, p. 94). But this promise of conceptual reconciliation seems unlikely
to be fulfilled since our quotidian concepts of experience do not develop a er
the fashion of theoretical concepts in the basic sciences. Nor does the way the
la er are developing seem likely to make them more like the former.
The most aggressive strategy for dealing with the explanatory gap and the
hard problem is to deny that addressing such how and why questions is
legitimate. Some hold that they beg crucial questions and that in defending
their proposed identity claims reductionists should simply deny that there
are two properties here (Papineau, 1993b, pp. 17980). The key when adopting
this line is to deny, from the get go, that there is any sense in playing the gen-
eration game at all (i.e. of trying to answer the question of how the physical
gives rise to experience). Defenders of this view deny that a straight solution
to the problems of consciousness is ever on the cards. But this is not because
consciousness wont reduce. Rather it is because it makes no sense to ask, in
general, how or why the mental and the physical are related given that they
are one and the same, though dierently encountered by us. Put simply, If
feelings are one and the same as brain states, then brain states dont generate
a further realm of feelings (or give rise to them, or accompany them, or
are correlated with them). Rather, brain states are the feelings (Papineau,
46
Consciousness
2002, p. 3). Accordingly the best policy for dealing with the hard problem is the
same as in dealing with taxes; avoidance is permissible but evasion is illegal.
In this case avoidance looks like the best move since as soon as you suppose
that conscious states are distinct from material states, then some very puzzling
questions become unavoidable (Papineau, 2002, p. 2).
A empts to identify conscious experience with a physical state of some kind
or other would be doomed to fail from the outset if conscious experiences in
fact shared no essential properties in common with such states. But the reply
to this charge is that while this might seem to be the case it simply isnt so. It is
entirely possible that one and the same thing may present itself to us in dier-
ent ways. There are plenty of cases in which a single referent is mistaken for
numerous distinct ones and vice versa because of misleading appearances,
names or descriptions. Noting this is all that defenders of reductive theories
require if they are to establish that their hypotheses about consciousness might
be possibly true in ways that would obviate having to deal with the problems
of consciousness.
To take this line is to hold that there is nothing more inherently absurd in
claiming that conscious experiences might equate to certain kinds of physical
happenings than there is in claiming that the Morning star and the Evening
star are the same planet: Venus. In neither case is the identity immediately
obvious or self-evident. If we allow this then there is no reason to deny, in
advance, that conscious experiences could not be identified with something
physical. To think otherwise, on the basis of appearances of dierence, is to be
under the sway of the stereoscopic or antipathetic fallacy.
While a ractively simple, nevertheless this sort of reply only goes so far. It
does nothing, by itself, to motivate acceptance of any of the proposed identity
claims; at most it makes space for their possible truth; at best it secures the
barest logical possibility of putative identity. And it does not deal with the root
problem that underpins the explanatory gap or the hard problem because it
fails to overcome worries about the intelligibility of making certain identity
claims. The bo om line is that to make their favoured identity claims credible,
it is necessary for reductionists to deal with the appearances of dierence in
some satisfactory way.
Rethinking Metaphors of Mind
Credibly establishing identities always rests on showing that what outwardly

appears to be dierent is in fact of the same kind for example, a planet, a
person, an event and so on. Establishing an identity in a fully satisfying way
depends on the possibility, in principle, of being able to explain away appear-
ances of dierence. To make an identity claim intelligible requires showing
47
that seemingly distinct things can possess all of their apparent properties with-
out tension or contradiction. Thus to make any progress on the problems of
consciousness, to render any given naturalistic equation about consciousness
truly convincing, would involve showing how the properties proposed for the
reduction could be the kinds of things that experiential states or properties
might be.
A nagging concern is that all existing reductionists proposals leave too
many important questions unanswered about the appearances of dierence.
In particular, they give inadequate answers to question, such as: Why do
experiences feel as they do? Who or what does the experiencing? How and
where does this all come together?
A deep-seated problem is that although reductive naturalists outwardly
denounce the picture of mental objects as occupying an inner sanctum of the
mind, they are inclined to take seriously questions and problems that do not
wholly make sense without presupposing this picture in some, perhaps vesti-
gial way. For example, some are tempted to ask: Where is my experience of pain
located? The sense of this question is taken to be straightforwardly akin to
the query; where is my pen located? But this leads directly to the problem of
phenomenal space, which is the problem of finding a place for the world of
experience within the world of physical space. In this context Denne is right
to ask, Now what is phenomenal space? Is it physical space inside the brain?
Is it the on-stage space in the theater of consciousness located in the brain
(Denne , 1991a, p. 130)?
Denne s analysis of the assumptions grounding the enterprise of explain-
ing consciousness is instructive. He believes that most philosophers, and many
lay folk influenced by them, conjure up images of the mind as an inner, mental
theatre complete with a self who examines various on-stage objects in the
spotlight of consciousness (pains, colours, figments of the imagination, etc.).
Those under the sway of this picture think of our verbal reports concerning
consciousness as based directly upon what the self sees on its private, inner
screen. Apparently it introspects mental items in a way similar to that in which
we ordinarily inspect everyday things such as watches or pieces of china.
Denne has done more than most to get us to critically question our thinking
on this score so as to abandon the idea that there is any such place in the physi-
cal world (and, in particular, the brain) where all the events of consciousness
come together. Rather than starting with such dangerous assumptions about
our explanandum he thinks that we have no choice but to begin our investi-
gations into the nature of consciousness by interrogating first-person reports
in a public, intersubjective context. He gives the name heterophenomenology
to this activity. While engaged in it, we, as interpreters, eectively allow the
subjects to verbally describe to us the nature of their experiences. They generate
texts about how things seem to them. They have authority concerning the
48
Consciousness
content of what is described. But what is described are best understood (at
least in the first instance) as notional worlds that are analogous to fictional
worlds, such as Sherlock Holmess London (not the real London). In being of a
like nature to such fictional worlds, The subjects heterophenomenological
world will be a stable intersubjectively confirmable theoretical posit (Denne ,
1991a, p. 81).
It follows that speech acts are the primary interpreted data for the study
of consciousness. These are reports, judgments, and beliefs that are made
concerning purported conscious experiences. The question of whether or
not what is described in these speech acts is real or fictional is le in abeyance.
Ocially, when we start investigating consciousness scientifically using this
method, we are required to begin (but not end) by focusing on the contents of
the speech acts of humans (and other possible speakers), staying studiously
neutral on what if anything lies behind them/explains their etiology.
The ontological moral Denne is inclined to draw is that although we ought
to allow subjects to have the final word in saying how they judge that things
appear to them, this in no way commits investigators to take seriously what
they describe at the level of ontology. He maintains that this is the only sound
way to take the first-person point of view as seriously as it can be taken
(Denne , 2003, p. 19). For him, interrogating such texts are our only means
of neutrally analysing the reports about what is going on in our minds. He
claims that the texts generated in these circumstances, and not something above
and beyond to which they putatively refer, are the raw material for any theory
of consciousness.
In promoting this understanding of where we must start Denne oers a
new metaphor for consciousness; the multiple dra s model. The multiple dra s
model identifies consciousness with our ability to generate a coherent text
concerning our putative mental episodes. James Joyces Ulysses is the model.
But he goes further and advances a positive reductive theory of consciousness
in terms of the ability to generate detailed, coherent serial reports. For Denne
the business of explaining consciousness boils down to explaining how the
brain is able to produce the relevant texts. By his lights we wont have explained
consciousness until we give a naturalistic account of our ability to produce
coherent speech acts through which we describe our experience of what it is
like for us to be conscious. And like all good reductionists he believes that,
Only a theory that explained conscious events in terms of unconscious events
could explain consciousness at all (Denne , 1991a, p. 454). Explaining con-
sciousness, for him, converts to explaining the capacity for a certain kind of
text production.
His task is not to explain the existence of conscious experience as it is usually
imagined to be but rather to explain how our talk about how things seem to us
is produced by underlying sub-systems. Thus he hopes to give an ontogenetic
49
explanation of how those sub-systems were formed and an account of how they
work. The essence of his proposal is captured in the following remark I am
suggesting conscious human minds are more or less serial virtual machines
implemented ineciently on the parallel hardware evolution has provided
for us (Denne , 1991a, p. 218). In line with his multiple dra s model he calls
the virtual machine that gives rise to consciousness a Joycean machine. And
he is quite aware of the limits of his theory; he notes that If consciousness
is something over and above the Joycean machine, I have not yet provided a
theory of consciousness at all (Denne , 1991a, p. 281).
There are similarities between Denne s approach and ambitious higher
order theories of consciousness those that maintain that being phenomenally
conscious requires a ending to or noticing the aspects of ones mental states
in ways that necessarily involve making reference to those aspects in higher
order acts of perception or thought. Accordingly, phenomenal consciousness
requires the use of higher order perceptions or thoughts (Lycan, 1996;
Rosenthal, 2005; Carruthers, 2000). If such higher order operations are, in fact,
partly constitutive of phenomenal consciousness then the neural basis of
experience must include machinery for inner sensing or for making theory
of mind ascriptions.3
For those who doubt that such mechanisms form part of our basic biological
equipment, Denne s account has an advantage. He does not believe that
the Joycean so ware is built-in; he regards it as the result of cultural design.
He tells us that consciousness is, largely a product of cultural evolution that
gets imparted to brains in early training (Denne , 1991a, p. 219). But critics
regard this as an admission that non-verbals, such as animals and infants, are
incapable of having experiences. Denne s response to this worry is that our
folksy intuitions regarding animal and infant consciousness are not sacrosanct.
Je isoning some of our most deeply held intuitions concerning the nature
of experience may be a price we must pay for adopting a neater criterion of
consciousness.
A deeply objectionable feature of Denne s theory, echoing the problems of
certain versions of behaviourism and functionalism, is that a complex system
would count as conscious if it produced pa erns of behaviour identical to,
say, those of yours or mine when we generate a stream of coherent u erances
that are interpretable as saying how things seem to us. Highlighting this aspect
of Denne s account, many have complained that his theory leaves out what
is critically important for understanding phenomenal consciousness: phenom-
enal qualities themselves. A capacity for experiencing such qualities, it is argued,
is logically independent from (and developmentally prior to) capacities for
propositional believing, reportage and narrative text production.
The problem for Denne s proposal and the oerings of higher order thought
theorists is that, as stand-alone accounts, they allegedly place too much emphasis
50
Consciousness
on sophisticated extras or supplements. These may be plausibly required for

having certain kinds of conscious experience but should not be confused with
the essential ingredients of a more basic, biologically grounded capacity for
phenomenal experiencing itself.
The idea that experience is more basic than such accounts suppose is further
supported by the observation that conscious experience is not primarily about
the abstract recognition and identification of objects but rather has a more
perspectival nature. This has independently motivated the conjecture that
conscious experiences must occur at an intermediate and cognitively impene-
trable level of perceptual processing; one that is neither very high nor very low
(Jackendo, 1987). In line with this, Block (2007) has suggested that the neural
basis of experience does not include mechanisms of reportage. But he accepts
that this generates a methodological puzzle for consciousness research given
that sincere first-person reports are the accepted basis for investigating the
neural correlates of consciousness.
There are various ways of making sense of what the relevant kind of pre-
conceptual, pre-linguistic form of perceptual processing involves. A popular
conjecture is that such perceptual experiences might be best understood in
representational terms. This conjecture is lent some initial plausibility by the
fact that those sentient beings which are known to be perceptually conscious
do not get by in the world by merely being responsive and reactive to its
oerings. Rather they are conscious of things as being a certain way. As Tye
observes, the idea that perceptual experiences must be contentful is most
strongly motivated by the thought that, in seeing objects, they look some way
to us, together with the further thought that an object can look a certain way
only if it is experienced as being that way. This in turn, seems to require that
the object be represented as being that way (Tye, 2009, p. 88).
Minimally, a state represents as if it presents some portion of the world as
being a certain (potentially) truth-evaluable way, for example, as being hot or
as being red. Importantly, a creature need not have sophisticated conceptual
abilities in order to be in such states of mind; it is possible that such content is
non-conceptual. All that is required for a mental state to possess content of this
kind is for it to have inherently specifiable correctness or accuracy conditions.
The intuition is that simply by instantiating the property of phenomenal blue-
ness a mental state is automatically capable of representing something in the
environment as being blue.
The most ambitious reductive variant of this idea takes it that the phenome-
nal character of an experience is exhausted by its representational content;
phenomenal properties are nothing but representational properties. Such
theories hold that the phenomenal aspects of experience are nothing over
and above taking features of the world to be a certain truth-evaluable way.
Accordingly, what it is like to be conscious boils down to representing how
51
things might be (given that how the world seems to be and how it actually is
may dier). Consequently, there can be no dierence in phenomenal character
without a corresponding dierence in representational content because
phenomenal character just is a kind of representational content. Weaker, non-
reductive versions of representationalism hold that changes in phenomenal
character lawfully co-vary with changes in content because, although distinct
from representational properties, phenomenal properties perform representa-
tional service.
Both strong and weak versions of representationalism about consciousness
face a number of serious objections (for details see Hu o, 2009). Arguably, a
major problem with all such accounts is that they a empt to understand basic
perceptual activity by illicitly importing features that in fact necessarily depend
on being a participant in sophisticated, linguistically-based practices (e.g.
having mental states with the kind of semantic content that requires assessment
by appeal to public norms and concepts, as in the a ribution of blueness to
aspects of the environment). If so, in imagining basic experiences to have
more properties than is necessary or possible for them to have, such accounts
make the opposite mistake to Denne and his followers.
Plausibly, having a capacity for phenomenal experiencing is more rudimen-
tary and fundamental than the capacity to represent the world as being a
certain truth-evaluable way. Consequently, experiencing aspects of the world
might be thoroughly non-contentful (and not just non-conceptual). Experi-
encing might not be intrinsically content-involving even though there is
something-it-is-like to experience worldly oerings in phenomenologically
salient ways.
This non-representationalist view of experience features as the central
plank of a radically enactivist approach to phenomenality; one that seeks to
understand phenomenal experience by focusing on the ways in which creatures
actively sense, perceive and engage with their environments (see Hu o and
Myin, forthcoming). Enactivists propose that the core features of experiential
properties are best explained by appeal to specific pa erns of sensorimotor
activity, through which complex self-organising systems interact with aspects
of their environment. Their slogan is: Experience isnt something that happens
in us, it is something we do (No, 2004, p. 216; see also Thompson, 2007). They
maintain that Experience is not caused by and realized in the brain, although
it depends causally on the brain. Experience is realized in the active life of
the skilful animal (No, 2004, p. 227). Thus enactivists challenge traditional
internalist thinking about the extent of the supervenience base of conscious-
ness, holding that it constitutively involves not just the brain but also bodily
and environmental features.
In pressing this idea, enactivists are critical of endeavours to understand
the phenomenal character of experience on a purely correlative basis, namely,
52
Consciousness
by looking exclusively at what goes on inside the craniums of experiencers

(a style of approach exemplified by those who seek to identify the neural
correlates of consciousness). Like those critics who are worried about the
hard problem and the explanatory gap, enactivists hold that even if relevant
mind/brain correlations are established, the fruits of such work would remain
explanatorily sour: ultimately this alone would tell us precious li le about
how or why experiences have the particular phenomenal characters or feels
that they do. By way of contrast, it is argued that charting which environment-
involving pa erns of sensorimotor interaction make a dierence to having
experiences with specific kinds of phenomenal character holds out much greater
explanatory promise.
While it is impossible to give a full and fair assessment of these proposals
here, this analysis suggests that it seems likely that a completely satisfying
naturalistic understanding of conscious experiences will require the complex
balancing and selective integration of a range of dierent theories. Achieving
this will require modifications, not only, of the ambitions and resources of
specific proposals, in tune with a be er understanding of the level at which
they best operate; it will also require a fundamental re-thinking of some of our
basic assumptions about the nature of consciousness.
53
3 The Mark of the Mental
Introduction
Whats a mind? What would it take to build one? There must be a dierence
between having a mind and not having one. What is it? This paper defends the
view that there is a dierence and tries to say what it is.
Why be interested? First, inquiring minds want to know. The question is
intrinsically interesting. It certainly seems that there is a natural divide among
biological1 systems that have minds and those that dont. If this is not an
illusion, we should be able to discover what constitutes that dierence.
Second, we commonly talk about minds and mental states. Is this just a
convenient fiction? Some people think we talk about minds because we dont
yet know the real story about why people (and animals) do the things they do.
Speaking about minds has practical predictive value in itself even if there
arent any minds. Horticulturists say that some plants like light or like dark
places, even though no one literally thinks plants have minds, have likes or
dislikes. Yet, while most people would accept that speaking of likes for house-
plants is a convenient fiction, very few think its fiction when talking about
Grandma or the kids (or even the family pet). Here, we seem more commi ed
to the a ributions of mind and mental states to people and pets being literally
true. We think these uses are literally true and our job is to figure out what
underlies that truth.
Third, science, the law, even other areas of philosophy pin important issues
on mental states. In the law, it is important to know if one who acts is legally
sane. In epistemology, some internalists about justification claim that justifica-
tion supervenes upon mental states. This requires knowing what constitutes a
mental state. Researchers interested in embodied cognition are making some
amazing claims in the scientific literature.2 We are told that the use of com-
puters, cell phones, PDAs, or even pencil and paper while doing a complex
math problem can involve cognition extending into the environment across
the boundaries of body and brain (Clark, 2009). The skin is an arbitrary
boundary, only observed by an outmoded Cartesian view of the mind. Some of
this may be true, but unless we are able to specify what counts as a cognitive
process, it will be impossible to evaluate such claims. For now, researchers are
54
able to make such bold speculative claims precisely because there is no agreed
upon account of what makes something a cognitive process. If these uses of
tools manipulate representations and information, and if thinking (cognizing)
is a kind of manipulation of representations, then why isnt tool use a kind of
cognizing?
Fourth, suppose we were going to try to build a mind. Many artificial intel-
ligence labs around the world are racing to be the first to build a computer or
a robot that can think. Governments are funding projects to build genuinely
intelligent agents to serve many dierent functions. To win the competition,
these centers must determine what it takes to build a mind, a cognitive agent.
There are, however, more and less radical views about why there might not
be a mark of the mental at all. Eliminativism asserts that there are no minds.
Mysterianism asserts that we can never know how a brain generates a mind.
We will not discuss these views here, but we do so elsewhere.3
A less radical possibility is that researchers cannot agree on what the mark
of the mental is or why people should want one. Susan Hurleys comment
is typical: Criteria of the mental or the cognitive vary widely (if not wildly)
across theorists; it isnt even clear what agreed work such criteria should
do (Hurley, forthcoming, 5). Kim too paints a dismal prospect for a unified
conception of mind, saying:
The diversity and possible lack of unity in our conception of the mental that
the class of things and their states that we classify as mental is also likely to
be a varied and heterogeneous lot . . . A question to which we do not yet have
an answer is this: In virtue of what common property are both sensory states
and intentional states mental. What do our pains and beliefs have in common
in virtue of which they fall under the single rubric of mental phenomena?
They of course satisfy the disjunctive property qualitative or intentional,
but that would be like trying to find a commonality between red and round
by saying that both red things and round things satisfy red or round. To
the extent that we lack a satisfying answer to the question we fail to have
a unitary conception of what mentality consists in. (Kim, 2006, pp. 267)
Hurley makes her point in the context of the extended mind debate, chiding
Adams and Aizawa (2008) for requiring a mark of the mental in order to decide
whether minds extend. She rightly points out that one can study the mind
without having an agreed upon mark of the mental. We know that vision,
memory, and decision making are mental, but if one wants to claim that mental
events extend beyond body and brain, one must have a notion of what makes
something or some process mental.
Kim (2006) argues that there is a dichotomy between sensory states
and intentional states. The former seem to take instances of properties as their
55
contents (sweet tastes, round feels, loud sounds), while the la er seem to take
propositions as their contents (believes that the Democrats will win; hopes that
we will get out of Iraq). For Kim, a unified account of the mental would require
somethings being the same across these types of states. Kim is pessimistic about
this possibility.4
Views of the Mark of the Mental
We will now consider a empts to overcome these sorts of worries by saying

just what the mark of the mental is. As Kim suggested, there could be a single
thing or common property that all mental states share and in virtue of which
they are mental. Call this the single property view. Alternatively, there could
be a cluster of properties that make a mind, and something may have to possess
a certain number of the properties in the cluster to be a mind. Call this the
property cluster view. Then, finally, there is the single system view that says
there is a single set of properties that all minds must have, but not every state
that is part of the system must possess these properties themselves. Some states
may be fully part of the mental system in virtue of their causal contribution to
the system properties. We will now examine single property views, property
cluster views, and finally present our own account of the mark of the mental, a
single system view.
Incorrigibility
The first single property view that we will consider focuses on the property
of incorrigibility. Rorty (1970a, 1970b, 1972) and possibly Denne (1996) want
incorrigibility to serve as the mark of the mental. Rorty is careful to distinguish
occurrent states (my foot hurts now; Im now thinking it is time for lunch) from
standing states (my background desire for self-preservation; my background
belief that global warming is a bad thing). Given this distinction, Rorty limits
incorrigibility to only first person, reportable, occurrent sensations and thoughts,
admi ing that there is no single mark of the mental for all entities customarily
called mental (Rorty, 1970a, p. 409). According to Rorty mental does not apply
to standing states.5
On this view, incorrigible states are first-person self-reports about thoughts
or sensations, for example, Im tired now, Im angry now, and so on. Rorty says:
What makes an entity mental is not whether or not it is something that

explains behaviour, and what makes a property mental is not whether or
not it is a property of a physical entity. The only thing that can make either
56
an entity or a property mental is that certain reports of its existence or

occurrence have the special status that is accorded to, e.g., reports of thoughts
and sensations the status of incorrigibility. (Rorty, 1970a, p. 414)
For Rorty, incorrigibility amounts to situations such as when the behavioural

evidence for what Smith was thinking about conflicted with Smiths own report
of what he was thinking about, a more adequate account of the sum of Smiths
behaviour could be obtained by relying on Smiths report than by relying on the
behavioural evidence (Rorty, 1970a, p. 416).
Why incorrigibility? At one point, Rorty (1970a) criticizes Armstrong for
trying to find a topic neutral way of characterizing mental states such that
they might turn out to physical states. That is, science might discover that what
was once a ributed to minds is had by brains. As we could discover that water
is H2O or lightning is electrical discharge, we could discover minds are brains,
that mental events are physical events. Rorty rejects this view on the grounds
that it would not allow a conceptual distinction between the mental and the
physical. If minds and mental phenomena were discovered to be brains and
physical phenomena, there would be nothing to ground6 the conceptual dier-
ence between the mental and the physical. Among Rortys motives is to find
a logical or conceptual dierence between the mental and the physical as
categories.
But categories of what? Not substance; this is not supposed to be a dualistic
metaphysics. Instead, the mental and the physical are dierent categories of
statements, assertions, or reports. Physical reports have the logical property
of being able to be over-ridden, while mental reports have the logical property
of not being able to be over-ridden or corrected by other physical reports, say,
a third-person report about what is happening in ones brain.
But why reports? Why not behaviour? In a sense first-person reports are
pieces of behaviour. Are they behaviour that non-mental things cant produce?
Rorty seems to think non-mental things can produce reports, but they wont be
mental reports because they dont have the logical property of not being able
to be overridden. Hence, Rorty is coming out of the behaviourist tradition. We
cant simply look inside ones head and see events as mental. We dont know
what Rorty thought of modern techniques of neuroscience, but we suspect that
he would maintain that it is only the first-person self-reports recorded from the
same subjects scanned that give us genuine knowledge of the subjects mental
states. Rorty did mention the philosophers fiction, the cerebroscope but main-
tained that reports from the cerebroscope would not override first-person
reports. If they did, Rorty is prepared to admit that the category of the mental
would lose its incorrigible status, and thus, status as mental (Rorty, 1970a,
p. 421). So, why reports? Reports are our only access to the mental as construed
by Rorty.7
57
We will present three simple objections to Rorty and then argue against
his whole approach. The first objection is the infants-and-animals objection.
On Rortys view non-lingual infants and animals dont have minds, but surely
thats not true. Of course, there have been those who have denied that animals
have minds (Descartes) or that they have conscious minds (Carruthers), but
most philosophers and scientists accept that both pre-lingual infants and non-
lingual animals have minds. Rortys view simply cannot accommodate this.
Infants and animals make noises, but they dont give first-person self-reports,
and so they are denied the incorrigible reports that are his hallmark of the
mental. They have plenty of occurrent mental events that are simply not
captured by Rortys criterion.
The second objection is what well call the Cog-objection. Denne (1996)
sides sympathetically with Rorty as he tells of the project to build an intelligent
agent named Cog at Rodney Brooks MIT lab. Cog is fi ed with cameras
for eyes, microphones for ears, and microprocessors linked in parallel for a
brain. The goal is to get Cog to think and to teach it language. Cog is a self-
reprogramming system, and at some point the reports that come out of Cog
may not match the interpretations placed on its internal states by its program-
mers. Denne agrees with Rorty: for Cog to think would be for Cog to make
incorrigible reports on its internal states.
We think at some point Cog may babble like a baby. At a later point Cog may
come out with It is hot in here. Suppose Cog has an internal mechanism that
is not unlike a quantum mechanism in this respect. To examine its internal
states is to change them. So, in eect, Cogs reports are incorrigible. Does this
mean that Cog satisfies the criteria for the mental? It would seem so, but
this seems like a limitation on us, on what we know or can know, not a break-
through in cognitive science.
Of course, Rortys ocial pronouncement had two conditions:
S believes incorrigibly that p at t if and only if:
(1)S believes that p at t,

(2)There are no accepted procedures by applying which it would be
rational to come to believe that not-p, given Ss belief that p at t
(Rorty, 1970a, p. 417) .
We suspect that Rorty had to mean that one knew (1) was satisfied in virtue of
a self-report. We know that Cog believes it is hot in here because he u ers
It is hot in here. If there is some other behavioural test for occurrent belief,
then Rortys theory collapses, assuming occurrent beliefs are mental states.
Condition (2) is satisfied by the nature of the internal mechanism that changes
58
the system upon any a empt to verify its current states. Of course, one could
object that unless we have another way to validate Cogs beliefs, we dont
really know that he made a report or satisfies (1). This takes us to our third
simple objection.
The third objection is that Rorty is not entitled to help himself to the idea
that any verbal u erance is indeed a report. Some Japanese cars (e.g. the 1986
300ZX) have a system that says door not closed, to warn the driver not to
drive until closing the door. Did the car make an internal report? We would
claim not. A sensor detected that the door was ajar and sent a signal to a voice
simulator that emi ed sounds interpretable by English speakers. To be a first
person report telling us the internal state of the car, the u erance must be made
with intention and purpose. These may have existed in the minds of the Nissan
engineers who designed the car but not in the car itself.
The same is true of Cog. If It is hot in here comes out of Cogs audio port,
is that a first-person self-report? We suspect that it is not any more than it is a
report coming out of the 300ZX. Reports are linguistic u erances.8 They are
intentional. They have meaning. They are for the purpose of communicating or
conveying information. This is the type of thing minds do. In essence, genuine
reports have to come from minds.
The more general objection is that Rorty has things the wrong way around.
If genuine reports come only from minds, then it cant be the logical features
of the reports, such as incorrigibility, that make them mental. There must be
something else underlying the ability to make the report that accounts for the
system having a mind, being a mental system.
Furthermore, we deny Rortys worry9 that the distinction between the men-
tal and the physical would collapse if we reject Rortys approach. True, if minds
turn out to be physical things, then the mental will be a subset of the physical,
but this doesnt mean that there is no dierence between a minded thing and a
non-minded thing. Minded things can be physical things arranged in dierent
ways from non-minded things, with dierent functions, dierent causal histo-
ries, and dierent internal and external behavioural capacities. Just because
minds may be physical does not mean there is no dierence between minds
and non-minds. The category of the mental, and therefore the meaning of the
mental, can be a subset of the physical. Gold is a kind, and it is physical. Water
is a kind, and it is physical. They are both physical, but that does not mean that
there is no significant categorical dierence between them. We see no reason to
worry that there will be no interesting conceptual dierence between the mental
and the physical if they both turn out to be physical, as Rorty seems to fear.
Once one takes seriously the notion that possibly everything is physical, even
minds, then Rortys fears vanish. If one takes seriously the thought that minds
are natural kinds, then it is up to science to discover what minds really are.
59
Intentionality
The next single property view we will consider focuses on intentionality or

aboutness. This view has a long and venerable history, going back to Brentano
(1874). There is the worry of Kim and others that there could be no single prop-
erty had by all mental states because there seems to be a fundamental divide
between sensory states like sensations and intentional states like beliefs, desires,
hopes, and fears. If Brentanos property of aboutness were only a property
of intentional states, not sensory states, that would leave sensory states out of
the category of the mental. Of course, one could adopt a single system view and
say that the system must have intentional states (aboutness), but not every
mental state contributing to the system must itself be intentional. This would
be one way to a empt to deal with the divide.
As of late a list of distinguished philosophers including Crane10 (1998),
Dretske (1995), and Tye (1995), have argued that sensory states are intentional.
If true, then there may still be hope for the single property view. All mental
states may have aboutness, and it is in virtue of this that they are mental.
Crane nicely expresses the continuity of the intentionality thesis this way:
What is common between these dierent states of mind is expressed in
Brentanos formulation: in the idea something is conceived, in the wish
something is wished. And in the sensation something is sensed . . . (Crane,
1998, p. 238)
Crane approaches intentionality for all mental states from a phenomenolo-
gical point of view, looking for a sense in which something is given to the
mind in sensation and emotion, just as something is given to the mind in
thought and experience . . . in sensation something is felt, in emotion, some-
thing is apprehended . . . (Crane, 1998, p. 243). He even goes so far as to
emphasize the priority of intentionality as a phenomenological notion (Crane,
1998, p. 249). He wants to contrast this aspect of intentionality with another
that he seems to recognize, viz. primitive forms of intentionality . . . only
remotely connected with conscious mental life, say the intentionality of infor-
mation processing, which goes on in our brain(Crane, 1998, p. 249). He rejects
the last type of intentionality as mental because it is not phenomenological.
We find it curious that one would appeal to intentionality as the mark of
the mental and then acknowledge a kind of intentionality that is not mental.
Crane notes that there may appear to be a problem in this, saying:
This would be a perverse or circular way to proceed if we did not already

have a grasp on the concept of a mind. But we do have such a grasp: it is that
concept which we try and express when we say that to have a mind is to have
a point of view or perspective on the world, or when we say that there is
something it is like to be conscious, or when we talk about the world being
60
manifest to a subject of experience, or when we talk about the world being

a phenomenon for a subject. (Crane, 1998, p. 249)
But he seems to think that he gets o the hook by appealing to the phenomeno-
logical side of conscious mental states. We dont think he makes it o the hook.
If his concept of mind is the phenomenologically conscious (what is given),
why doesnt he make that the mark of the mental? If it is not, then how does
he rule out those information processing states in the brain as being mental?
For instance, some parts of the brain detect low blood sugar or cold extremities
and reduce insulin production or constrict the capillaries to hold blood from
the extremities. Why arent these mental activities if they are intentional states
of the brain? What about unconscious desires or beliefs? Surely these are
mental states even if they lack a phenomenology. We are sympathetic with the
idea that sensory states are intentional, but probably not because they are given
phenomenologically, rather because they are the right kinds of representations.
So we are on board with Crane (1989) until the very end where he acknow-
ledges intentional non-mental states. He needs a principled way to distinguish
the mental intentional states from the non-mental ones. Crane wants to do it
via phenomenology. We think there must be a be er way.
Furthermore, since Crane defends only a weak view that sensory states
are intentional, but their qualia may not be, he acknowledges that this leaves
him open to the question of what makes qualia mental. Qualitative states
certainly seem to be mental states in good standing, and they seem to be on a
par with phenomenological states generally. So it would be strange indeed if
a mark of the mental le them out. Realizing this, Crane acknowledges that
more needs to be said (Crane, 1998, p. 251, n. 26).
We believe accounts like those of Tye (1995) and Dretske (1995) have the
advantage of making the qualitative characteristics of conscious states them-
selves representational and hence intentional, thereby leaving no dangling
qualia unaccounted for by the mark of the mental. While there are some
important dierences between their views, for our purposes here we empha-
size their similarities.11 Consider the sweet taste of sugar in ones mouth. For
Dretske (1995), the qualia of experience arise due to an indicator function in
the sensory system. Although Dretske uses the term function in his theory, we
dont believe too much weight should be placed upon that term (Adams,
2003). What is essential is that there is a type of sustained causing (Adams,
1991). That is, the structure in the brain S that indicates the presence of sugar
will cause some other brain activity or bodily movement M (say, swallowing).
When the S causes M (rather than some contrasting N spi ing) because of
the indication of sugar by S, then and only then does S acquire the function of
indicating sugar. What is important is the sustained, contrastive causing. The
structure S must be sustained in its causing some relevant eect by the fact
61
that it indicates the presence of sugar. If so, then S comes to represent the pres-
ence of sugar and makes the person in whom S does this conscious of sugar by
virtue of the qualitative experience of sweetness.
Tye (1995) defends a representational theory of the phenomenal mind on the
basis of a co-variation view of representation. The view states that something
S represents that P = df. If optimal conditions obtain, S is tokened in X if and
only if P and because P. In this definition, X is a place holder for a person and
P is a place holder for a proposition, due to the that on the le hand side of
the identity sign. The la er is a particularly bad choice because, first, proposi-
tions dont cause things. Events or instantiations of states of aairs may cause
things and may have propositional content, and may cause things because
they have propositional content, but propositions themselves dont cause
things. Second, Tye himself says the structure of phenomenal representation
is topographical (Tye, 1995, p. 120). To us this suggests that it would be far
be er to interpret P in the definition as a property instance. If so, then the
definition says that S (sensation of sweetness) represents P (sugar in the mouth)
if, under optimal conditions, S is tokened in person X if and only if there is
sugar in the mouth and because there is sugar in the mouth. Hence, the sweet
sensation represents sugar (in the mouth) because it is tokened when there is
and caused by sugar (in the mouth).
On the accounts of both Dretske and Tye, therefore, the qualia themselves
arise out of the representational role of the sensory states. It is because the states
are representational (intentional) that the qualia are as they are. Ones qualita-
tive experience of sweetness is itself an intentional state a representation of
sugar in the mouth under normal conditions.12
A further issue raised by Crane (1998) and Enc (1982) is whether intentional-
ity is a sucient condition for the mental. As weve seen, Crane embraces the
existence of two kinds of intentionality because he is willing to say that infor-
mation processing in the brain is a primitive form of intentionality but is itself
not mental. If this primitive intentionality is aboutness of the type generated
by any informational connection in the world, then it exists everywhere, not
only in the brain. Litmus papers turning pink is about a liquids being an acid.
A thermometers rising is about an increase in temperature. The falling barom-
eter is about a decrease in atmospheric pressure. This kind of informational
aboutness is everywhere in the world, and barring panpsychism it is not su-
cient for the existence of minds. So if this primitive intentionality13 is indeed
intentionality, it is not the right kind to qualify something as mental. Any
single property view of the mental that claims intentionality is the mark of
the mental is going to have to clarify what is the right kind of intentionality
to serve as the mark of the mental and explain why it is the right kind.
Fodor (1986b) introduces the notion of the detection of non-nomic properties
for precisely this type of reason. Lower organisms, such as paramecia, exhibit
62
purposive types of behaviour, such as photosensitivity. Fodor does not count

this behaviour as action, as he reserves that term for the behaviour of inten-
tional systems (Fodor, 1986b, p. 6). However, Fodor needs a principled reason
to exclude lower organisms from the class of genuine intentional systems. To
do that, he introduces a behavioural test: being able to dierentially detect
and respond to non-nomic properties. A non-nomic property is something
such that [if] objects fall under laws in virtue of possessing it, then that prop-
erty is ipso facto nomic (Fodor, 1986b, p. 10). So, having mass or momentum
are nomic, and being a crumpled shirt or a le shoe are not. In his view minds
alone can detect a non-nomic property F qua F. So his view addresses the
problem of finding the right kind of intentionality for the mark of the mental
in this way. As he puts it, the dierence between paramecia and us is that
we can respond selectively to non-nomic stimulus properties and they cant
(Fodor, 1986b, p. 11).
We are not sure how much emphasis should be placed on non-nomicity.
For in the end Fodor seems to appeal mainly to the fact that paramecia and
other lower organisms lack concepts (whether of nomic or non-nomic proper-
ties). They dont have the proper inferential mechanisms to form concepts
over wide ranges of varying stimuli, and it is what is lacking that ma ers more
than the behavioural test. At one point Fodor puts his point this way: What
distinguishes intentional systems from the rest is that, whereas weve got
perceptual categories, what theyve got is, at most, sensory manifolds (Fodor,
1986b, p. 20). This leads us to think his principled dierence between proper
intentional systems (like us) and only apparent ones (like the paramecium), is
that we have concepts and they are purely sensory systems. 14
Another person who appeals to intentionality is Fitch (2007). His view is a
single hierarchical system view where the intentionality of a mental system is
built upon a basis of nano-intentionality that starts at the level of the biological
cell. Indeed, Fitch thinks this is the dierence between computers and biologi-
cal systems. By nano-intentionality Fitch means the capacity cells have to
rearrange molecules in individual circumstances in an autonomous and adap-
tive fashion (Fitch, 2007, p. 10). It seems to us that Fitchs nano-intentionality
is placeholder for evolved biological function. If a cell or an organ has acquired
the function to do F, then the cell or organs nano-intentionality is about F-ness.
So somethings being for a biological purpose or goal is its having nano-
intentionality. When nano-intentionality involves the processing of information
in a neuron, then Fitch thinks things rise to the level of mental aboutness (Fitch,
2007, p. 12). For Fitch biologically being a neuron seems to be a necessary con-
dition of mental aboutness, as opposed to nano-intentionality of cells generally.
Still, he seems to think nano-intentionality itself is necessary for mental about-
ness.15 He also says that the relevant dierence between us and the amoeba is
that we have dedicated information-processing machinery. Although he does
63
not explain this term, we think he means something like the specialization of a
property detector or concept.
Fitch also stresses that the nano-intentionality of cells or systems responds to
the novelty of their circumstances and records, in a non-mental sense, internal
changes based upon the organisms history of interactions. He maintains that if
the instructions for responding are already encoded in the organisms DNA,
then its responding to circumstances accordingly is not nano-intentionality.
He is interested in a biological kind of learning, or so it seems to us,
that depends on processes not simply inherited or primarily fixed by genetic
inheritance.
Fitch thinks genuinely mental representations are internal mental models
or possible worlds instantiated in neuronal firings (Fitch, 2007, p. 20). The
system has to be sensitive to the model, as well as, via the model, sensitive
to the world. These models also have to direct the systems behaviour that
is contingent upon tracking both the world and the internal models. Once
models are in place, then representation and misrepresentation are possible.
Hence, Fitchs view is that these models or complex representations are the
hallmark of the mental and only occur at higher levels of organization in the
nervous systems of animals.
Consciousness
We turn now to Searles view, which is actually not a single property view but a
system view, where the system itself must possess consciousness to be mental.
It is a system view because he appeals to the thesis of the background:
Intentional phenomena such as meanings, understandings, interpretations,

beliefs, desires and experiences only function within a set of Background
capacities that are not themselves intentional . . . all representation, whether
in language, thought, or experience, only succeeds in representing, given a
set of nonrepresentational capacities. (Searle, 1992, p. 178)
Searle blocks the spread of intentionality to lower parts of the brain, or

merely information-processing parts of the brain, by insisting that only a being
that could have conscious intentional states could have intentional states at
all, and every unconscious intentional state is at least potentially conscious
(Searle, 1992, p. 132). This is Searles famous connection principle (Searle, 1990a;
1992, pp. 155.). It is with the connection principle that Searle distinguishes
unconscious mental states from other states of the brain that are not conscious
and may have nothing to do with the mind at all, thermo-regulatory states or
regulation of blood sugar levels, for example.
64
Searle (1992) cannot simply appeal to intentionality as the mark of the

mental because he thinks sensory states are mental but lack intentionality.
He accepts that both sensory states and intentional states are either conscious
or are accessible to consciousness.16 He cannot define consciousness in terms
of intentionality for similar reasons. He thinks consciousness arises out of the
neurochemical properties of the brain, and it is the neurochemical properties
that are responsible for consciousness. Consciousness itself has a cluster of
structural properties, including: dierent sensory modalities, unity, intentional-
ity, subjective feeling, figure-ground structure in experience, aspect of familiar-
ity, overflow, center and periphery, boundary conditions, mood, and dimensions
of pleasure and pain (Searle, 1992, Chapter 6).
For Searle, what makes something a mental state is that it either is conscious
or is potentially conscious. If it is an intentional state, then the connection
principle is run via aspectual shape (Searle, 1990a, p. 587; 1992, p. 156). If it
is a sensory state, like an unconscious pain, then aspectual shape is dropped
and the only appeal is to the underlying neurophysiological processes capable
of generating a conscious state, like conscious pain, and appropriate pain
behaviour (Searle, 1992, p. 165).
Searles argument for the connection principle when the unconscious state
is intentional is roughly this. Intentional states are essentially aspectual. For
example:
When you see a car, it is not simply a ma er of an object being registered

by your perceptual apparatus; rather you actually have conscious experience
of the object from a certain point of view and with certain features . . . as
having a certain shape, as having a certain color, and so forth . . . and . . .
[this] is true of intentional states generally (Searle, 1990a, p. 587).
Unconscious intentional states must be aspectual since they are intentional,

but since they are unconscious their aspectual shape is not manifest. Searle
claims that there is no aspectual shape at the level of neurons. That leaves
unconscious intentional states like any other neurophysiological state, even the
non mental ones. So to be mental, unconscious mental states must preserve
their aspectual shape. The only sense that we can give to their preserving their
aspectual shape, when unconscious, is that they are possible contents of con-
sciousness (Searle, 1992, pp. 15960). The only fact about the underlying neuro-
physiology is that it can cause conscious states that manifest conscious
intentional aspectual shapes. So the unconscious states are only mental
because they have the capacity to produce conscious states with the right stu:
aspectual shape.
No one would deny that being conscious is sucient for being mental.
Still, we worry about Searles claim that there is no aspectual shape at the level
65
of neurons (Searle, 1990a, p. 588). This may be true at the level of single neurons.
However, in macaque monkeys, at the level of single cell recordings, there
are cells that are sensitive to only goal-directed movements of other monkeys.
This seems aspectual. More importantly, Searle gives no reason why there
cannot be aspectual shape at the level of collections of neurons. A er all, he
insists that the only things that cause consciousness are the neurophysiological
properties of the brain. If these cause aspectual shape in conscious processes,
then there must be something specific in the neural structure that accounts for
the specific aspectual shape of ones conscious experience when accessed. Of
course, Searle may mean that the aspectual shape of conscious states only
exist as an emergent property in the conscious access of the unconscious state.
But since he thinks all conscious states are neural states, then at some level,
perhaps the conscious level of organization in the brain, clusters of neurons
must possess aspectual shape. They must be aspectually organized in their
firings, so to speak. To deny this would make Searle an emergent dualist,
something he would vehemently deny.
In addition, we worry that there may be mental states in good standing that
cannot be brought to consciousness. One type of case is blindsight, where an
individual who lacks a conscious presentation of a visual scene nonetheless
responds purposively by hand orientation in relation to a slit or negotiating
objects in a room (Weiskrantz, 1997). Another type of case is vision for action
(Milner and Goodale, 1995). Some actions guided by the dorsal stream in the
visual system are able to correct for visual illusions, such as Titchener Circles.
An object being reached for might appear larger than it is because of an illusion,
but thanks to dorsal stream processing ones reach is for the actual size of the
circle.17 Milner and Goodale, as well as Weiskrantz, have given examples of
states that are surely mental. They are guiding high-level purposive activity.
Yet they are states that seem not consciously accessible to the blindsight
subjects who cannot describe the scenes before them and whose dorsal stream
processing is guiding their action.
Our view
Our view is a systems view.18 We dont think that every state in a mental system
must possess a single type of property shared by all other core mental states.
Some states contribute to a systems having the property or properties that
make a system mental without themselves possessing the core property or
properties. We think that mental systems share a cluster of properties that we
will articulate below.
On our view one cannot properly say what minds are without saying what
minds are for. Minds allow organisms to track changes in their environment
66
and respond dierentially to those changes. This includes allowing them to

maintain integrity and a empt to satisfy their needs for survival. Some tracking
is coarse grained, as in sensory processing, and some fine grained, as in concept
formation. An organism that senses a source of sugar may have a sensation of
sweetness. Sugars are carbon based. So the organism would be sensing some-
thing carbon based, but not necessarily sensing it as something carbon based.
Thoughts, however, may be fine-grained, and one may be able to think that
something is a sugar and also think of it as carbon based if one has the requisite
conceptual competence.
Tracking changes in the organisms environment, both inner and outer, is done
by processing information. This requires a network of internal and external sen-
sory mechanisms that transmit or process information to the structures in the
organism that use this information to modify or control the organisms behaviour.
Following the mathematical model articulated by Shannon,19 we understand
information as involving a law-like connection of properties set against a back-
ground of locally stable conditions. Some call this level of information process-
ing done by transducer mechanisms and internal mechanisms that regulate
body temperature or blood sugar intentional (Crane, Fitch). We, however,
distinguish levels of intentionality and maintain that ge ing to the level of
intentionality that is mental aboutness requires rising above information to the
level of semantic content (Dretske, 1981, 1985b). Smoke carries information
about fire, but smoke means smoke, whether or not an occasion of its use
carries information about fire. Minds, genuine mental systems, rise to the level
of meaning. Their internal states can mean or be about things in the way that a
thought about smoke can mean smoke. This level of aboutness permits mis-
representation (Dretske, 1986).
For us, the requirement that a mind has states that rise to the level of mean-
ing is what determines that amoeba and paramecia lack minds. So we make the
same distinctions between mental and non-mental systems as Crane, Fodor,
and Fitch, but we get there by a slightly dierent route, although we are close
to Fodors view, if one downplays non-nomicity. Semantic content can be as
simple as a dedicated property detector in a lower-level mind or as complex
as propositional a itudes in human minds. Creatures with minds do things
because of the contents of their minds because of what their internal states
mean or are about (Dretske, 1985b).
We have no arguments that there could not be purely sensory systems. No
one else that weve read has such arguments either. We are inclined to think
that purely sensory systems arising in nature would not survive. They would
have to be able to do things to their benefit because of what they sense. If they
did this, they would have concepts along with wants, desires, and the means
of satisfying them. So we are inclined to the view that actual sensory systems
are conjoined to intentional systems, that is, systems with concepts.
67
We follow Dretske (1981) in distinguishing orders of intentionality. Purely

sensory systems have only the first order of intentionality, the same order
as information itself, what we called coarse-grained information processing.
So if all Fs are Gs, but not as a ma er of law, just as a ma er of coincidence,
something that carries the information that t is F does not necessarily carry
the information that t is G, even though it is G. A sensory system that detects
property F need not detect property G. Hence, even a sensory system has this
degree of intentionality. If Fs and Gs are connected as a ma er of natural law,
however, a purely sensory system that carries the information that something
is F by detecting Fs will also necessarily carry the information that something
is G. Sensory systems lack a second order of intentionality: being able to dis-
tinguish Fs from Gs. Systems with concepts, on the other hand, appropriately
equipped, would be able to make this fine-grained distinction. Grandma knew
there was water in Lake Michigan even if she didnt know there was H2O
there, perhaps because she did not have the concept of H2O. So we dierentiate
purely sensory systems from systems with concepts on the basis of orders of
intentionality.
A question that haunts us and others is why amoebas or paramecia dont
have sensory states (qualia). Everyone weve discussed would deny that such
creatures have minds largely because they dont have intentional states, con-
cepts, mental models, the ability to detect non-nomic properties, and so on.
But why dont they have sensations? For example, Fitch says they have nano-
intentionality, the ability to rearrange their parts based upon environmental
conditions, and to do so in a way not determined by their DNA alone. This
sounds like what minds do, viz. track changes in the environment, make inter-
nal changes on that basis, and then modify behaviour. So why dont they have
minds? We see Fitch denying that eukaryotic cells have minds or qualia even
though they modify their internal states in response to external conditions
and do so in novel ways. He denies that they have minds because they dont
employ mental models.
Tye (1995) and Dretske (1995) both maintain that systems with sensory states
must be harnessed to conceptual states to yield qualitative states. For instance
Tye says systems that altogether lack the capacity for beliefs and desires cannot
undergo phenomenally conscious states (Tye, 1995, p. 144). He thinks this
because he holds that qualitative sensory states lie at the interface between
outputs from sensory modules and inputs to cognitive systems, but he doesnt
really justify or explain this. Dretske too maintains that experiences are types of
representation whose function it is to supply information to a cognitive system
for calibration and use in the control and regulation of behaviour (Dretske,
1995, p. 19). For support Dretske refers the reader to Gareth Evans.
Evans suggests that conscious perceptual experience arises only when
sensory inputs are connected to behavioural dispositions (of perhaps some
68
phylogenetically ancient part of the brains motor system) and serve as input to
thinking, concept application, and reasoning (Evans, 1982, p. 158). Evans says
this just a er discussing a case of blindsight. In this case a subject dierentially
responds to an external visual stimulus while reporting no conscious visual
field. This leads Evans to say if the sensory system only provided information
to the motor system, there would be li le interest in producing qualitative
states blindsight seems to have no need of them. In blindsight visual informa-
tion is processed, but there is no conscious perceptual experience or qualia.
This leads Evans (and perhaps Tye and Dretske) to maintain that it is only when
properly connected to a conceptual system, perhaps conjoined with a motor
system, that qualitative experiences arise. Only then will there be qualitative
visual fields, auditory experiences, and so on.
Why dont amoeba and paramecia have qualitative experiences? Fitch
would say that although they have nano-intentionality, they dont have internal
mental models. Evans, Dretske and Tye would say that they dont have sensory
systems which feed into conceptual systems. Even though all would agree
that information is processed at the first level of intentionality. We think that
there is a perfectly good sense in which these organisms dont need qualitative
experiences. The dierential responses that they make to environmental
changes do amount to a low-level processing of information, Fitchs nano-
intentionality, but the processes involved can all be explained at the level of
chemistry or photo-chemistry alone, plus a bit of history of the organism in
its local environment. There are no dedicated processors no biologically
selected structures that are recruited for the purpose of tracking information
about the environment or the internal states of the organism as they respond to
environmental changes. Thus, there are no more or less permanent structures
that have the biological function of indicating to the organism changes of envi-
ronment and self.20 Thus there are no internal structures that have the function
of both informing a cognitive system and driving a motor system that serves
the needs and desires of the organism. Having such internal, dedicated infor-
mation-processing structures requires explanation that rises above the level of
local chemical reactions. The organisms use of the information generates a
qualitative sensory experience.
It may seem like magic21 that a biological structure whose function is to
deliver information from a sensory transducer to a conceptual, cognitive
system should generate a qualitative visual, tactile, auditory, or gustatory
experience. This leads someone like Searle or Block to say that qualia must
arise from the brains neurochemistry. To us thats no less magical. How do the
chemicals do it?
Along with Dretske and Tye, we maintain that the qualitative nature of the
sensory experiences is explanatory. It is because of the way things look, taste, or
feel, that we do what we do when we experience them. In blindsight cases an
69
individual may orient her hand vertically because a slit in front of her is
oriented vertically, not because of how it looks phenomenally. Whereas, when
we do the same thing with a full visual field, we do orient our hand because
of how the slot ahead looks. So the qualitative representational content of the
sensory experiences plays an explanatory role in our purposive behaviour.
Of course, how the light of qualitative states comes on is no less a mystery, even
if this is how or why it comes on in conscious sensory systems.
Since we maintain a systems view, and since we think there arent actually
any purely sensory mental systems, all actual minds must have concepts.
Concepts have a semantic cognitive content or meaning, and this content rises
to the level of non-derived semantic content. By non-derived we mean that
the content is not given by or dependent upon the mind of another. So, for
example, the first minds that arose did not derive their mental contents from
the mind of another. Internal structures had to acquire a content that was
meaningful to the organism itself. This means that the system interprets the
world on its own. This dierentiates minds from contemporary computers that
have only derived content, content supplied by the minds of the engineers who
build them, content meaningful only to the engineers, but not to the computers
themselves.22 This also means that the concepts in minds have to rise to the
second level of intentionality.
The best explanation currently available of how this goes is due to Dretske
(Dretske, 1988). Dretske explains how a structure or concept rises to the level
of non-derived meaning in the context of solving the dreaded disjunction
problem the problem of how C could come to mean just F when either Fs
or Gs cause tokenings of C. Dretskes solution to the disjunction problem
has at least two components. The first component is ge ing to the first level
of intentionality, being able to indicate that something is F without also
indicating that it is G. The symbol C must start out with the ability to naturally
mean Fs and not Gs. If all of its natural meaning is disjunctive, if it only
indicates Fs or Gs, then a disjunctive content is the only semantic content it
could acquire.
The second component is the jump to semantic content. Even if Cs indicate
Fs only, to acquire semantic content, a symbol must lose its guarantee of pos-
sessing what Grice called natural meaning. Smoke naturally indicates fire
because in the wild the two always co-occur. Hence smoke has the ability to
indicate the presence of fire. A symbol that has smoke as its semantic content
needs to become locked to smoke and permit robust and even false tokening
that means smoke without infecting its semantic content. It has to lose its
ability to always naturally indicate fire. Dretske appeals to the explanatory
relevance of the natural meaning. For Dretske, it is not just what causes Cs, but
what Cs in turn cause, and why they cause what they cause that is important
in locking Cs to their content (F).
70
Lets suppose that a squirrel needs to detect Fs (predators) to stay alive. If Fs

cause Cs in the squirrel, then the tokening of Cs indicates Fs. Dretske claims
that Cs come to have the content that something is an F when Cs come to have
the function of indicating the presence of Fs. When will that be? Every predator
is not just a predator; each one is also an animal (G), a physical object (H), a liv-
ing being (I), and so on for many properties. Hence, tokens of C conjunctively
will indicate all of these, not just Fs. Dretskes answer is that when Cs indication
of Fs (alone) explains the animals behaviour, then Cs acquire the semantic con-
tent that something is a predator (F). Hence, it is the intentionality of explana-
tory role that locks Cs to F, not to G or H or I.
For Dretske, intentional behaviour is a complex of a mental states causing
a bodily movement. So when C causes some bodily movement, M say,
the animals movement into a hole the animals movement consists of its
trajectory into its hole. The animals behaviour is its causing that trajectory.
The animals behaviour running into its hole consists of Cs causing M
(CJM). There is no specific behaviour that is required to acquire an indicator
function. Sometimes the animal slips into its hole (M1). Sometimes it freezes
(M2). Sometimes it scurries away (M3). This account says that Cs become
recruited to cause such movements because of what Cs indicate, what Cs
naturally mean. The animal needs to keep track of Fs, and it needs to behave
appropriately in the presence of Fs to avoid predation. Hence, the animals
thought content becomes locked to predators when tokens of C explain some
appropriate movement M because of Cs indication or natural meaning. Not
until Cs natural meaning has an explanatory role, does C mean that something
is F in a non-derived way.
The account does not require that the animal must be moving to be thinking.
The relevant type of causing of M by C is required only to lock the content of
C to Fs. Once so locked, the animal can think about predators even when not
moving. What is more the animal can now mistakenly think there is a predator
when C is tokened by a shadow or rustling leaves. Hence this gives a teleo-
logical naturalized account of how concepts come into existence and have
non-derived meaning. It also explains how a concepts having the content it
does can explain behaviour.
This completes our systems view of the mark of the mental. States of a
mental system make a causal contribution to the core properties that make
the system mental. For us, there is a cluster of core properties necessary and
sucient for somethings being a mental system. Hence, our property cluster
list for mental systems is:
(1)Mental systems possess non-derived meaning.

(2)Mental systems possess states that rise to the second level of
intentionality.
71
(3)Mental systems have states capable of misrepresenting (or representing

non-actual things or states).
(4)Mental systems exhibit intentional behaviour explained via the represen-
tational content for the system (sensory states explain via their qualitative
feel, and conceptual states explain via their semantic contents).23
72
4 Substance Dualism
T. J. Mawson
Introduction
Substance dualism could not have a more venerable lineage, being traceable
back through Descartes at least as far as Plato and Socrates. However, the
respect with which people treat the view has declined to such an extent in the
last few hundred years that it has recently been described as, not so much
a position to be argued against, as a cli over which to push ones opponent.
Certainly within contemporary educated circles, were one to venture the
opinion that we have souls, one should expect to find oneself held to have
propounded an extravagance only slightly less great than had one ventured
the opinion that visiting extra-terrestrial life was in part responsible for the
construction of the pyramids or that Elvis may be seen working in the local chip
shop. The most favourable response one could realistically hope for would be
the concession that perhaps, before the development of such things as com-
puter science and neuroscience, such a whimsy might have been excusable, but
even so, now souls must surely go the way of phlogiston and light-carrying
ether: onto the intellectual scrap heap.1
Here I shall advance the claim that, despite the near universality of the
assumption that the theory may be easily cast aside, within the structure of a
hylomorphic substance/property metaphysic, the only reason to suppose that
we do not have souls is that provided by Ockhams razor and even that reason
is conditional upon an assumption, albeit an assumption that it is no more my
intention to cast doubt upon here than it is my intention to cast doubt upon
the substance/property metaphysical structure within which I shall be framing
this debate.2 The assumption is that there is physical stu.
Given that there is physical stu, it would indeed be simplest to suppose
that the mind is ontologically reducible to that or to processes going on in that.
But given that, as I shall also argue, there are some reasons to suppose that
we do have souls that is, that such a reduction cannot be accomplished, so one
might find oneself, probably idly, reflecting on the fact that idealist substance
monism would oer one all the advantages of simplicity oered by physicalist
substance monism while in addition accommodating these reasons for sup-
posing we have souls. This reflection would probably be idle as there is li le
73
danger of idealist substance monism emerging, on balance, as the preferable

theory of the mind for us: the assumption that there is physical stu is held
by most of us so deeply as to be near immovable by argument. (For similar
reasons, I shall ignore neutral monism.) Thus it is that most of us will find
ourselves weighing the reasons in favour of the claim that we have souls in
the balance against the rational a ractiveness of the simpler metaphysic that
a physicalist substance monism oers. Where this balance ultimately se les is
something on which opinions will divide. All I can hope to secure consensus
on by what follows is that no substance/property metaphysic will give us
everything we want, which in itself of course is a reason to re-examine the
substance/property starting point. If one refrains from doing that however (as
I shall), one must conclude that either several of our assumptions concerning
the nature of persons (assumptions which are not held significantly less deeply
than is our assumption that there is physical stu ) are in error or the world
is more complex than physicalist substance monism allows, for we do have
souls a er all.
What Substance Dualism is
We have to start somewhere and time is pressing, so let us put onto the table
without oering argument in its favour a certain commonsense realism about
the physical world and our knowledge of it as gained through the natural
sciences. First then, let us assume that there is physical stu. This may be
characterized as stu of the sort that we suppose ourselves to encounter with
our five senses in everyday life; that our folk science describes more or less
adequately for our everyday purposes; and that our natural sciences describe
with increasing accuracy as they develop. We may define the sort of stu we
have in mind by paradigm examples of things which are made of it: this desk,
here; that star, there; and so on. In a previous century, we might have called this
physical stu simply ma er, but now we know that ma er may be converted
into energy and vice versa and we hear scientists speculate concerning quarks,
hyper-dimensional strings, and so forth as making up the more commonplace
objects that we encounter in everyday life. These are things which, while strik-
ing us as no doubt physical, do not strike us as in any obvious way material,
so, instead of ma er, we call this stu physical stu . We shall call the view
that this physical stu is all the stu that there is physicalist substance monism
or physicalism for short. Obviously one might hold that in addition to this
sort of stu, there is another type of stu as well. We shall call this second
view substance dualism.
Substance dualism is commi ed then to there being a type of stu that
resists full integration into the natural sciences. What we might call partial
74
Substance Dualism
integration will need to be allowed for to take into account psycho-physical

causal interactions, which as we shall see the most plausible substance
dualist view will wish to maintain occur. Incapacity for full integration is not
however by itself enough to characterize this second type of stu adequately.
There might turn out to be a sort of stu that resisted the sort of integration
that the substance dualist will wish to claim for his or her souls yet which it
was obvious to commonsense was nevertheless purely physical; the unity of the
natural sciences is a hope or perhaps something stronger, a regulative idea. But
it could have turned out, or could even yet turn out, to be misguided.
Similar problems would beset a empts to characterize this second type
of stu in terms of its failure to fit into the existing categories of the natural
sciences. We will not wish now to draw up a list of what properties might be
made mention of in a completed science, conscious as we are that some of the
properties of physical stu as quantum physics describes them are very dier-
ent from the properties of it as we encounter it in everyday life or as would
have been supposed to be primary and fundamental in the days of the corpus-
cularians; they are, it has been said with some understatement, spooky. But we
can issue a promissory note here and that is sucient for our current purposes:
the second type of stu in which the substance dualist believes is a type of stu
the nature of which will not be fully integrated into a completed science of
objects such as our paradigm physical objects tables, stars, et cetera because
it has properties that will not feature in that completed science and will not be
reducible in any way to properties that do so feature. The obvious contender for
the fundamental property here is the capacity to have mental properties such
as beliefs, desires, emotions, and so on. We might think of intentionality and/or
first-person privilege as their hallmarks.
The substance dualist will maintain that the essence of soul substance is that
it is capable of thought in the broadest sense of the term, and this is a property
which would, in a completed science, turn out not to be a property of stu of
the sort that makes up tables, stars and so forth and turn out not to be reduc-
ible to any such properties. Nevertheless, the substance dualist maintains,
contra the eliminative materialist, thinking is definitely going on and, given
the substance/property metaphysic within which this debate takes place, he
or she validly concludes from all of this that it must thus be going on in a sub-
stance other than a physical one soul substance, as we have been calling it.
We might say then, more or less following Descartes as we do so, that, accord-
ing to the substance dualist, it is of the essence of physical substance to have
properties of the sort that our paradigm examples of physical objects tables,
stars and so on have, which will not include thinking. (Descartes se led on
spatial extension as the essential property of physical stu; we have le this
more open; perhaps spatial extension as we ordinarily understand it will
turn out in a completed science to be a property of only some physical stu
75
[e.g. tables], a property constructed out of more basic elements.) It is of the

essence of soul substance that it is capable of thought, where thought is taken
in the broadest of senses to include all mental happenings beliefs, desires,
sensations, emotions, acts of the will, and so on. Belief in the existence of these
two types of substance is what is definitional of substance dualism.
There are various views within substance dualism about the relationship
between this soul stu and us as persons. Are we as persons simply our souls,
or do we persons have souls as one part and bodies as another? Descartes
entirely identified the person with his or her soul, and thus would have had
no objection to our talking of disembodied souls were they to continue on
a er bodily death as fully the people they had been when earlier embodied.
An alternative view is possible. Arguably it is that held by Aquinas. This is the
view that persons are to be identified with the conjunction of body and soul
and thus that where these two cease to be conjoined in the right sort of way
most obviously, perhaps, if they cease to be conjoined in any way at all as
one of the conjuncts entirely ceases to exist (e.g. the body is vaporized by an
exploding nuclear device) what survives, if anything, is not the person in his
or her entirety, but merely a part of the person. And it may be that a disembod-
ied soul part (of a former person) would not, as a ma er of causal fact, be able
to do any thinking once separated in this way from the body part with which
it had previously formed a person. One might go even further down this track
and think that the destruction of the body part would inevitably cause the
destruction of the previously associated soul part and thus the entirety of the
person. But if any of these things are so, then, according to substance dualism,
they are so as a ma er of metaphysical contingency, not necessity. substance
dualism makes it metaphysically possible for the person (Descartes) or a part
of the person (Aquinas) to survive the complete and final destruction of his
or her body, but it does not entail that this actually ever happen. It makes
it metaphysically possible that any disembodied soul would be able to have
a mental life as rich in what we might call pure mental properties (not, for
example, suering from toothache a concept which spans the ontological gap
between body and soul) as an embodied soul, but it does not entail that this
actually ever happens.3
There is a third view, interior to substance dualism, although it has not
in fact ever been propounded by anyone who believes we do have souls; this
is the view which would identify us entirely with our bodily parts. The most
plausible variant of this view would, it strikes me, have to give up on the idea
that we are fundamentally persons and thus may be pictured as having some-
thing akin to animalism: we are our animal selves; at the moment, these animal
selves happen to be in causal contact with souls and thus happen to be able to
think (and through doing so become persons), but, were such souls destroyed,
no element of our animal selves, and thus no element of us, would be destroyed;
76
Substance Dualism
we would just cease to be able to be persons. The soul is entirely inessential to

what makes us us even though it is not inessential to what makes us persons.
As it has in fact never been propounded by a substance dualist, despite its
being a potential variant of the view, I shall ignore this view in what follows.
As well as providing the materials with which dierent views of the nature
of the self and us-as-persons may be constructed, substance dualism also allows
for various views on the causal commerce between souls and bodies. For vari-
ous reasons which will become apparent as we progress, the most plausible
variant of substance dualism is interactionist substance dualism (the body
causally aects the soul and vice versa). The alternative views are psycho-
physical parallelist substance dualism (the two have no causal interchange
whatsoever); epiphenomenalist substance dualism (the body causally aects
the soul, but not vice versa); and the view again a neglected alternative in
that no-one actually holds it that the soul causally aects the body, but not
vice versa.
Reasons to Suppose Substance Dualism False
As mentioned in the introduction, as a theory about the nature of the mind,

substance dualism is more ontologically extravagant than substance monism.
Given that there is physical stu, it would be simplest to suppose that the
mind is somehow reducible to that stu, or, more plausibly, to processes going
on in certain bits of that stu: brains, presumably; mind is to brain mutatis
mutandis what digestion is to the digestive system. Given that the properties
of physical stu are by no means obvious to us and recent scientific develop-
ments have indicated that some seem to be spookier than earlier generations
of scientists would have found even imaginable and given that, from what
we already know, the brain is the most complex structure in the universe, it is
not unreasonable for us to hold out hope that a completed science would be
able to fill in the mutatis mutandis here. Of course it cannot do so yet, but these
are early days. This is, it must be conceded, a reason to suppose substance
dualism false. What reasons might we have to suppose it false beyond its
complexity relative to physicalism?
I shall consider two areas from which it is o en suggested additional
reasons for supposing substance dualism false emerge.
The first area centres on supposed problems in identifying souls, both
ontologically and epistemically. What is it that makes one soul dierent from
another and how can we ever know of souls that they are the same over time or
know of souls other than our own that they exist at all? In short, my analysis
here will be as follows: firstly, insofar as substance dualism faces problems that
parallel those faced by physicalism (as it does in addressing the issue of what
77
makes one fundamental unit of substance dierent from another and how we
know of such units that they are the same over time), these problems cannot be
reasons to favour physicalism over substance dualism and so are not properly
construed as objections to it, rather than perhaps as objections to the wider sub-
stance/property metaphysic within which this debate is taking place. Secondly,
insofar as substance dualism faces problems not faced by physicalism (as it
does in addressing the issue of how we can ever know that units of substance
other than our own exist at all), the fact that it commits one to a certain sort of
scepticism here is a reason to suppose it true, not false.
The second area of concern centres around supposed problems in explaining
the causal interaction between the two sorts of substance the substance dualist
posits. How can mind and body act on one another? Does not any answer to
this question run into insuperable problems from what we already know of
physics, for example concerning the causal closure of the physical world and
the conservation of energy? In short, my analysis here will be that the inter-
actionist substance dualist is not beholden to answer the question of how mind
and body act on one another, rather than merely assert that they do, as it is
not a commitment of interactionist substance dualism that this question will
be answerable by us. Positing that there is an interaction of this kind does not
in fact require one to contradict things which we already know of physics,
although there is potential for physics (were it to move back into a deterministic
mode) to put pressure on the claim that there is in fact interaction of the sort
posited. At the moment then, there is no reason from science to suppose
interactionist substance dualism false.
Let us go into these objections in more detail.
Problems of identication
We may sensibly ask the substance dualist what it is that makes one soul
distinct from another and predict that he or she will have li le informative to
say by way of reply. Obviously, he or she may maintain that it is extremely
unlikely that two souls will have all the same properties as one another, so he
or she may point out any two souls will in fact dier in this fashion. One
will be thinking about strawberries, another, about cream, and so on. But exact
qualitative identity between two souls is not a metaphysical impossibility gen-
erated by the nature of souls per se and even if it were somehow impossible for
two souls to have exactly the same properties, this impossibility would not
ground the numerical dierence between two souls, but rather presuppose it.
In any case, it looks as if the substance dualist should agree that there is nothing
in the nature of souls per se that prevents there being two qualitatively identical
yet numerically distinct souls, for it seems that theres nothing in the nature of
78
Substance Dualism
souls per se that prevents there being an exact duplicate of this universe. In that
universe there would consequently be a person thinking qualitatively identical
thoughts to those that you are currently thinking. That person would, never-
theless, not be you; it would be your duplicate. So, the substance dualist should
say that it is not fundamentally in virtue of their dierent properties that
dierent souls are dierent. Rather, he or she should admit that souls might in
principle dier solo numero. (They have what is sometimes called thisness.)
Need he or she be embarrassed that he or she can say no more than this? I do
not think so.
Presumably the person who believes in units of physical substance will
wish to maintain that at least with regards to some of these there is nothing in
their nature that prevents their diering solo numero too. The classic thought
experiment on this topic involves imagining a universe composed simply of
two chemically pure iron spheres, each of the same diameter, hanging in other-
wise unoccupied space a certain distance away from one another; these spheres
would be qualitatively identical to one another, yet they would be numerically
distinct. Can the physicalist substance monist say more about how these two
spheres manage to retain ontological individuality than that they do, that they
dier solo numero or have thisness? No. So the substance dualist need not feel
embarrassed about being able to say no more than this about how two souls
might retain their ontological distinctness even were they to have qualitative
identity.
As this discussion might have already indicated, this type of issue and
in fact the one we are about to go on to discuss is an artefact of believing
in substance as such, (i.e. of believing in things to which the principle of the
identity of indiscernibles does not apply of necessity). As such, this type of
issue and the one we are about to go on to discuss cannot be a reason to prefer
any theory that claims that substances exist over any other that claims that they
do. Thus it cannot be a reason to prefer substance dualism over physicalism.
Belief in substance raises certain problems at the epistemic level. Of sub-
stance dualism, it is sometimes said, souls might be swapping bodies every
few minutes but each inheriting the psychological properties of the soul that
had just vacated the body into which the new one was now moving. Were
this to be the case, no one would be able to detect these changes, yet people
(Descartes) or significant parts of people (Aquinas) would constantly be swap-
ping bodies. Furthermore, we seem to face on substance dualism a peculiarly
intractable variant of the problem of other minds: how do you know, as you
encounter another person through the medium of the physical world, that he
or she is a person at all, that he or she has a soul in the right sort of causal
connection with the body which you observe directly?
Again we may observe that the first problem aects those who believe in
substance per se and thus in substance of the physical sort; thus, whatever it
79
is a reason to believe, it cannot be a reason to believe in physicalism over

substance dualism. How do you know that the physical stu underlying the
properties of the desk in front of you has not been swapped out by some
malign demon in the last few moments, leaving all the properties behind in
the sense of their being inherited by the physical substance which this demon
instantaneously moved in to replace that which he was removing? So this
worry generalizes to physical substance. But, having said that, its not too great
a worry. The physicalist substance monists response to this sort of worry seems
to me entirely adequate. It is indeed metaphysically possible for the substance
of the desk to be being changed in the imperceptible way suggested, or, if this
is not metaphysically possible, then that is for reasons exterior to the nature
of physical substance per se (e.g. that there can be no spirits of the right
sort). But unless we have positive reasons for supposing that such swaps are
happening, as it would be simplest to suppose that they are not, so we should
suppose that they are not. The same move, then, that both the physicalist
substance monist and the substance dualist make with respect to physical
substance, the substance dualist makes with respect to soul substance. If it
works in one area, what reason is there to suppose it will not work in the
other? None.
The problem of other minds is o en thought to particularly aect and thus
speak against the truth of substance dualism; were substance dualism true,
it is suggested, there would be peculiar diculties in our knowing that other
people exist. I shall deploy a two-pronged approach to meeting this charge:
first, I aim to show, similarly to previous objections, that, if this is a problem, it
is a problem that is faced, at least to a greater extent than is o en appreciated,
by physicalism too. Secondly, as it must be admi ed that, pace point one, it is
faced to a greater extent by substance dualism, so I shall aim to show how this
extra problem of other minds is not, in fact, one it is implausible to suggest we
face. Were substance dualism true, there would indeed be an extra diculty
in knowing that others have minds, but that is not a reason to suppose that
substance dualism is false; indeed it is a reason to suppose it true for there is
much plausibility in suggesting that we do face this extra epistemic hurdle in
coming to know that others have minds.
First, then, though it would take too much time to argue it here, the most
plausible physicalism will identify the having of a mind with physical processes
that are recondite in the extreme. For example, a crude behaviourism, whereby
being angry is simply behaving in a certain fashion, which may be specified
entirely adequately in terms of movements of the body, movements that are
suciently macroscopic for us to be able to identify them without any great
diculty, using our unassisted five senses, will not prove adequate to the
task. Rather, some neurological happenings of a certain type will need to be
called upon in the analysis of anger, but as soon as the physicalist pushes
80
Substance Dualism
the happenings which are mind-happenings, interior to the skull, then, unless
we meet people who are themselves in fMRI scanners of sucient sophistica-
tion to reveal to us these happenings, we never ourselves see the happenings
that are, on the physicalist account, being angry, or what have you. On physical-
ism no less than substance dualism, we never observe the having of minds
other than our own. How then do we know, if physicalism is true, that others
have minds?
To cut a long story short, the answer to this question is that they tell us that
they do, and we ordinarily have no reason to doubt them. Someone says that he
or she is suering, let us say, from anger. If physicalism is true, they will be
speaking truly if a certain happening is occurring in their brain; but we do not
see this happening and indeed at the current stage of science might not know
that it was their feeling of anger, even if we did see it. But, unless we have
reason to doubt them (e.g. they are performing in a play or some such), we
are surely rational, whatever the theory of mind to which we subscribe, in
believing that they are angry simply on the basis of their saying that they are.
Without taking this sort of epistemic route into knowledge of others minds, it
would be impossible for the physicalist substance monist to construct the
theories by which he identifies to his satisfaction the having of anger with the
brain happening that he could then, in principle, find to be universally corre-
lated with the tendency to report it. (This is sometimes called the privilege
that must be given to first-person reports of the mental.) But if that is so, then
this same route is open to the substance dualist.
It is true that on the substance dualist view, the actual feeling of anger is
something happening in a substance even more recondite than the inner parts
of the brain. It is happening in a soul and thus in something that could never
be revealed by investigation into the physical world however advanced fMRI
scanners became. But the same route which the physicalist substance monist
takes in everyday life, before hand-held fMRI scanners and the like become
commonplace (and which he or she will have to hold as epistemically authori-
tative even were they to do so, to accommodate the issue of privilege), is open
to the substance dualist. This is how the problem of other minds is to be over-
come whatever ones theory of mind: by taking claims to have minds as a prima
facie reason to believe minds are had.
However, moving on to the second point, it seems as if the physicalist
substance monist may argue that whatever problems he or she faces in
coming to knowledge of others minds, and however these are to be overcome,
the substance dualist must face an additional problem unless he or she posits
some direct and very reliable telepathic contact between minds as an alterna-
tive source of knowledge, which positing would itself be most implausible.
This is a true point. But does it speak against or really in favour of substance
dualism?
81
Were physicalism true, then, a er science has been completed and presum-
ing it has allowed for hand-held fMRI scanners or some such of sucient
accuracy let us call them brainoscopes one could perhaps confidently
bypass first-person reports as a source to knowledge of others minds; one
could, instead of speaking to a person, directly apply ones brainoscope to
someones skull and, on the basis of its findings, confidently report things like,
No need to speak; I see from my brainoscope that you are angry at my having
applied it to your head without first asking your permission. These reports
could be unfailingly accurate. (Note: not all physicalists believe that this will
prove possible, but we are considering the views of one who does in order to
point out the contrast with substance dualism and the extra problem of other
minds that it faces.) Let us consider a physicalist substance monist who con-
tends that, a er science has been completed, one will be in a position to know
that a certain brain state or some such may be identified with anger being felt at
having had a brainoscope applied without having been asked for permission
and, with the technology of the brainoscope properly applied, one will know
that this brain state is being had, so, one will know that the person is angry in
this way. For such a physicalist substance monist, there will then be no gap
into which a sceptical doubt may creep. It might appear that nothing similar
could happen on substance dualism. But, in fact, the substance dualist may
hold that it could. If substance dualism is right, then in a completed science this
technology might well be possible. The substance dualist of course would not
make the extra step of identifying the brain state or what have you that is
revealed by the brainoscope with the mental state, but he or she can acknow-
ledge that there might well turn out to be a perfect correlation of the sort the
physicalist we are considering anticipates our finding, and thus the substance
dualist might admit that the sort of brainoscope that is capable of bypassing
first-person reports in the manner described could well turn out to be possible.
But there is, nevertheless, it must be conceded, a gap for the substance dualist
here relative to his physicalist substance monist counterpart, a gap generating
an extra problem of other minds.
The extra problem for the substance dualist is generated because it will
always remain possible that the brainoscope is in error, even once the science
is completed and the brainoscope working (for all we know) properly, for,
according to the substance dualist, the brain state or what have you that the
completed science finds universally to be conjoined with a thought of a certain
kind (and we are supposing that this is what it will find) and that the braino-
scope correctly reports to be present in this case is not to be identified with
the thought of that kind. According to the substance dualist, one could know
everything about the physical world, yet not know without the possibility of error
what mental state a person was in (or indeed even if they were a person at all)
for there is according to substance dualism an ontological gap between the
82
Substance Dualism
physical world and the mental, a gap which may be bridged by causation, but
causation not being a conceptual relationship any particular bridge across
which may or may not hold and thus any particular judgements using which
may be in error. But now this extra problem of other minds for the substance
dualist looks more like an asset than a liability, for, as we shall see when looking
at Mary-type arguments for substance dualism, it is apparently possible that
someone might know everything about the physical world yet not know some-
thing about the mental, which appearance has to be ruled out as deceptive by
the physicalist substance monist we are considering.
Problems of interaction
The version of substance dualism on which we are focusing suggests that there
is two-way causal exchange between physical substance and soul substance.
This is o en held to generate problems for the view. First, it is suggested that it
runs contrary to a finding of physics. In particular, it looks as if the principle
that ma er/energy is conserved across a closed system such as the physical
universe must be violated if substance dualism of the interactionist sort is true.
Second, it is suggested that there is something problematic in general in any
case regardless of whatever physics might be telling us about non-physical
substances causing changes in physical ones and vice versa. We know, a priori,
that such is an impossibility.4 I do not find either of these two lines of thought
tempting.
Let us suppose for a moment, what we shall later see is in any case false,
that the interactionist substance dualist is commi ed to laws of physics being
violated. It does not seem that an objection arising from this commitment would
be any more than a restatement of the objection from the relative complexity of
substance dualism over physicalism. Obviously it would be simpler were the
universe closed and the laws of physics not violated, and that is indeed, we
have already conceded, a reason to suppose that it is so. We should not double
count this objection to substance dualism. In fact though, the interactionist
substance dualist is not commi ed to his or her souls violating natural laws.
With the advance of physics beyond determinism, another possibility arises.
The substance dualist may maintain that happenings in the brain which are
caused directly by the soul are caused in ways compatible with the preceding
brain state and the laws of nature, but these two not being such as to necessi-
tate what state emerges from them they are caused to be the particular way
that they are by the soul. That the brain be in state q, rather than state r, a er it
has previously been in state p is something which was always allowed for by
the preceding physical states (given indeterminacy), but, in fact, the substance
dualist may maintain, that it ended up in state q was caused by the relevant
83
persons soul. It is no bar to this theory to point out, if such a fact can be pointed
out (and it is doubtful that it can be), that, of any individual sub-microscopic
event where such quantum indeterminacy plausibly reigns, it seems incapable
of producing cascade eects up to the macroscopic level which results in arms
being moved and so forth. For presumably some brain state leads to macro-
scopic happenings such as arms being moved, and this is made up at the sub-
microscopic level of many such quantum happenings. So, the substance dualist
may maintain that the souls influence on the brain, in causing it (e.g. to raise
ones arm, occurs in a number of disparate tiny locations, any one of which is
perhaps not sucient, or perhaps even necessary, for the event to occur, but
which then jointly cause ones arm to rise). Those quantum happenings in the
brain which are similar in the properties they reveal to the natural sciences as
those happening in an inanimate object where they are indeed uncaused are
in fact, when they happen in the animate object that is the brain, caused by
the soul of the relevant person. The universe is not indeed causally closed, but
no laws of nature need be violated.
So, in short, even were fundamental physics to return to a deterministic
mode, the interactionist substance dualist could maintain that souls are able to
influence physical stu (and vice versa) although by doing so he or she would
be positing that the laws of physics are violated li le bits of energy come into
and go out of existence. However, within the current indeterministic paradigm,
no such violations are required as a part of the substance dualists account of
this interaction. The substance dualist may maintain that the soul operates
in the causal gaps, otherwise filled by randomness, that indeterminism opens
up. And of course even were the dominant paradigm of interpretation of the
laws of nature within the community of physicists to revert to determinism,
it would still be just a paradigm of interpretation; there would be no necessity
that the substance dualist follow it.
Of course, such suggestions on the part of the substance dualist presuppose
that in general a spiritual substance may cause a change in a physical substance
and vice-versa, and someone might hold as a ma er of principle that the only
possible relata of causation are physical events, so such a suggestion may be
ruled out in advance. But why adopt such a principle? It may be rejected by the
substance dualist as mere prejudice if argued for a priori (although of course if
argued for validly a priori, the substance dualist will need to find one or more
premises to which to object) and the substance dualist will insist that such a
principle cannot be discovered a posteriori, for the actual universe is one which
has souls operative in it, so does not follow it. Descartes himself said all that,
it strikes me, needs to be said on this issue in a le er to one of his objectors.
These questions presuppose among other things an explanation of the

union between the soul and the body, which I have not yet dealt with at all.
84
Substance Dualism
But I will say, for your benefit at least, that the whole problem contained in
such questions arises simply from a supposition that is false and cannot in
any way be proved, namely that, if the soul and the body are two substances
whose nature is dierent, this prevents them from being able to act on each
other. (Descartes, in Co ingham, vol II, 1994, p. 275).5
So, in summary: the reasons for supposing interactionist substance dualism

false and physicalism true reduce to the simplicity of the la er over the former.
Simplicity is a reason to prefer one theory over the other, but so is explanatory
adequacy, and it is far from clear that physicalism will prove adequate, as we
shall now see.
Reasons to Believe Substance Dualism True
Various arguments in favour of substance dualism have been put forward over
the last two and a half thousand years, and it would be impossible to provide
an adequate treatment of all of them in anything smaller than a substantive
book. That being so, in the space that remains for me, I wish to focus on just
three areas where, it strikes me, the substance dualist can plausibly contend
that substance dualism does be er than physicalism in accommodating various
commonsense intuitions we have about ourselves. Of course commonsense
intuitions are hardly the basis for conclusive arguments in favour of substance
dualism. A er all, if our commonsense intuitions about such issues were not
sometimes wrong, there would hardly be any point in the discipline of meta-
physics. I conclude then by discussing what weight we may in general give
to this type of argument relative to the weight we may give to the virtue of
simplicity which, it has been conceded, physicalism has over substance dual-
ism. The three areas are personal identity, freedom, and consciousness. I shall
consider them in order.
Personal identity
What is it that makes a person at a later time, t+1, the same person as existed at
an earlier time, t? Substance dualism has a simple answer: it is fundamentally
the continuity of the same soul (or, for Aquinas perhaps, the same soul and the
same body), and souls themselves do not continue in virtue of anything more
basic continuing (bodies presumably do). For the physicalist substance monist,
the issue is more complicated: there are three options. The person may be
identified with a certain set of properties (usually psychological properties
are chosen); with a part of the physical substance which makes up his or her
85
body (usually the brain is chosen); or with a combination of these (e.g. psycho-
logical properties p going on in brain b). However, none of these options seems
to oer a satisfactory theory of personal identity. There are problems peculiar
to each, but a general defect may be observed in play in their dealing with
almost all the thought experiments that are used, it is supposed, to illuminate
this issue.
So, for example, one is asked to imagine a brain bisection, a er which the
two resultant hemispheres are transplanted into separate clones of the original
body where they take up more or less functional residence. To add weight to
the situation, perhaps one of the resultant people is then tortured to death
over the next five minutes while the other is given a gin sling to enjoy. Which
of these two resultant people, if either, is the person who originally underwent
the brain bisection? one is asked. Then the details of the experiment are altered;
perhaps one of the two resultant people gets more psychological continuity and
the other more of the physical substance of the original brain. What then do
we say? For some proportions of psychological continuity and continuity of
physical substance, the physicalist must say that it either becomes ontologically
indeterminate whether a resultant person is the same as the original, or it
remains determinate, yet he or she does not know whether he or she is the
same or instead a new person inheriting some of the originals psychology
and/or brain ma er. But our commonsense intuitions about personal identity
do not allow for indeterminacy, as shown most markedly when one thinks of
these possibilities from the first-person perspective of someone about to
undergo the relevant experiment: Either I will survive or I wont; it cannot be
ontologically indeterminate in a few minutes time whether Im there or not.
But nor is there anything unknown le for the physicalist to hang a determinate
fact of personal identity from, something which again we might perhaps see
most sharply by imagining the first-person perspective: If I can know where all
the properties are going and know where all the physical substance is going,
yet still not know where I am going, then I cannot be identified with any
combination of properties or the physical substance; I must be something else,
and the only something else le (once weve swept properties and physical
substance o the table) is soul substance. This is not conclusive of course, for
one could be unbeknownst to oneself identical to some indivisible property
or indivisible bit of physical stu and thus even if one knew in advance of this
property/bit of stu where it was going to go, one would not know that in
knowing this one was knowing where one was oneself going.
However, each of these claims that one is to be identified with an indi-
visible property or an indivisible bit of physical stu would itself be most
implausible. Properties and sets of properties (whether properties of physical
substance or soul substance) are capable of multiple instantiation, and the
sorts of sets with which people might most plausibly be identified (that go into
86
Substance Dualism
psychological continuity accounts of personal identity) are themselves capable

of degrees of survival. But people are not the sorts of things which seem to
commonsense capable of multiple instantiation, or division as the most dis-
cussed variant of multiple instantiation is sometimes called. One would not
think, Maybe, in five minutes time, Ill be two people, one being quickly
tortured to death and one enjoying a pleasant drink, so that in ten minutes
time Ill be both alive and dead. And, as already observed, people are not the
sorts of things which seem capable of survival by degrees. So it is implausible
to suggest that persons are to be identified with any set of properties or any one
indivisible property. Physical substance is not capable of multiple instantiation,
but the set of bits of physical substance with which people might most plausi-
bly be identified brains are capable of division and, as already observed,
people are not capable of division. Of course there might be some genuinely
indivisible bit of physical substance within the brain an atom in the original
Greek sense which enables one to side-step this issue if one identifies oneself
with that, but to posit such a thing and to identify oneself with it would be
most implausible. Substance dualism, then, gives the best theory of personal
identity by reference to our commonsense intuitions about persons as not
being capable of multiple instantiation/division and as not being capable of
survival to a degree.6
Freedom
In everyday life, we o en suppose ourselves to have been able to do something

dierent from whatever it is that we ended up doing, even had everything else
in the physical universe up to the moment of our choice remained exactly the
same. Of course, if everything else had really remained exactly the same, then
we might wonder why we would ever have behaved dierently from however
it is we ended up behaving, but some of our choices are, we believe, whimsical
in the following way. I am oered the choice between tea and coee; I have no
preference between the two but a strong preference to have one rather than
remain thirsty and, not wishing to be like Buridans ass, I thus say, on a whim,
Tea, please. In reflecting back moments later, I believe of myself that I could
have said Coee, please instead. In a situation such as this, although perhaps
most vividly in situations where things of great moral moment turn on what
we end up doing, we suppose that the fact that we end up doing whatever it
is we do end up doing is to some extent at least down to us. I say to some
extent at least as it must be conceded that we always operate within a finite
range of options and sometimes this finitude exculpates us from at least some,
possibly all, responsibility (I agree that what I did was bad, but look at the
alternatives I faced; each was worse), but we ordinarily suppose that this finite
87
range is greater than one we do genuinely have options and, when we have
options and end up realizing one rather than another as a result of the right
sort of conscious choice on our part, we suppose that in that way the causal
and moral buck stops with us. We are in this way free agents, responsible to a
greater or lesser extent for the choices we make and thus for the shape of our
lives and the lives of those we aect.
Substance dualism of the interactionist sort gives a straightforward
and simple account of how all of this gets to be so. (That is the long-promised
reason why interactionist substance dualism, rather than, for example, psycho-
physical parallelist or epiphenomenalist substance dualism is the most plausi-
ble.) According to interactionist substance dualism, the soul, while of course
being aected by things going on in the physical world (e.g. in coming to the
beliefs that it has about that world), is not always necessitated to do what it
does by those eects; sometimes it initiates causal chains, which then impinge
upon the physical world when it could yet have initiated dierent causal chains
and thus impinged dierently, had it chosen to do so. When my soul does
so, that is me (Descartes) or a part of me (Aquinas) making a choice. The com-
monsense view of ourselves as articulated in the previous paragraph finds its
metaphysical grounding.7
Physicalism cannot ground this commonsense view. On physicalism, either
what I ended up doing was entirely causally necessitated by preceding states
extending back though time to the big bang or there was a certain amount of
randomness (uncaused-ness) involved in the causal chain that ended up with
my doing whatever it was I did. In neither case would the causal and, one
might hence think, moral buck stop with me; either the happening was
caused by factors beyond my control (for they go back to the big bang, which
is certainly beyond my control); or it was random; or it was some mixture.
Various accounts of how the moral buck might stop earlier than the causal
buck and in the right spot me have of course been advanced by physicalist
substance monists keen to accommodate moral responsibility to their world-
view. So, for example, one might say that if my body does what I want it to
do as a result of me wanting it to do that thing, then thats me being morally
responsible for the doing of that thing, and the fact that my wanting it to do that
thing rather than something else was itself caused by factors beyond my control
does not detract from that. This account is open to easy counter-arguments, but
there are of course much more sophisticated accounts. However, they all suer
from the common feature that whatever psychological states are posited as
sucient to lead to the agent being morally responsible, it seems possible to
imagine a skilled enough hypnotist inducing those states in a person and yet
we not hold such a victim of such hypnosis to any extent responsible for the
actions that then flowed from these states. In cases where we can identify causal
responsibility, moral responsibility, we think, falls straight through to it; we are
88
Substance Dualism
strongly commi ed to the causal and moral buck stopping in the same place.
Substance dualism of the interactionist sort is the view that accommodates
this strong commitment in oering a third way to causal necessitation of the
physical sort stretching back beyond our births and randomness: my actions
are caused by me (Descartes) or my mental part (Aquinas). Substance dualism,
then, gives the best theory of freedom by reference to our commonsense intu-
itions about ourselves as being the initiators of and thus morally responsible
for our actions.8
Consciousness
The classic thought experiment here concerns someone called Mary, who,
we are asked to imagine, has been brought up in an entirely black and white
room. In this room she has access to black and white science textbooks and
science is now completed. She thus learns everything there is to know about
the physical properties of colour and indeed, let us say, about the physical
properties of brains too. She then leaves the room and goes into the outside
world. For the first time, she herself sees a red apple. Is it not plausible to sup-
pose of her that she thereby learns something new: what red looks like? We
may call this new fact a fact about red qualia: what it is like to see red. From
the fact that Mary ex hypothesi knew everything about the physical qualities
of the colour red and the brain prior to leaving her room yet did not know
about this qualiatative, as we may call it, property, so we can conclude that
this qualiatative property is not a physical property of red or the brain; of what
is it a property then? The substance dualist has a ready answer: of red as it is
experienced by the soul.
There have been various physicalist responses to Mary-type thought
experiments; they tend to deny the fact that Mary comes to know about a
qualiatative property; rather, they tend to assert, she comes to have an ability
which she did not previously have, the ability to recognize red objects in a
new way.9 This however seems wrong-headed to me, for Mary plausibly will
not gain the ability to recognize red objects simply by ge ing out of the room
and seeing a red apple for the first time. She will only gain that ability once
someone provides her with information in the following manner: That apple
youre looking at, Mary, its red. In hearing someone say that, she will plausibly
gain a new ability to recognize red objects therea er, but she had already come
to know what red objects looked like prior to hearing someone say that, just by
looking at the red apple. She wouldnt say back to the person whod just said
this to her, Now, for the first time, I know what red is like; shed say something
like this: Now youve told me that that apple is red, I realize that I already
knew just by looking at it what it was that red was like, rather than what it
89
was that blue was like, and so on. But although I didnt know that it was red,
the qualiatative nature of which I knew about by looking at the apple prior to
your telling me, it was red that I had discovered something new about simply
by looking at the apple. 10
Of course the analysis presented here has been, perforce, terribly brief
(cannot property dualism deal with the issue of consciousness to which we
have recently adverted?), but, even so, it appears that the facts of personal
identity, freedom of choice and consciousness, as they present themselves to
commonsense are, when taken together, easily accommodated by substance
dualism and fail to be accommodated by physicalism. Either these facts are
not facts at all commonsense is wrong or physicalism is wrong.
Conclusion
If we suppose a substance/property structure to our metaphysic of the mind

and we suppose that there is physical stu (two suppositions I have not
called into question in anything but the most oblique way here), then reasons
of ontological economy alone would suggest that we should believe that we
do not have souls, that our mental life could in principle be explained in terms
of our physical. Such a world would be simpler to the tune of one whole
class of substance than the world posited by substance dualists. physicalist
substance monism has simplicity on its side when compared with substance
dualism, but, as we have seen, it does not seem to have anything else; there are
no other reasons for thinking substance dualism false. While simplicity is a
virtue, so is explanatory adequacy, and there are things that we have reason
to suppose a physicalism cannot explain. That current natural science cannot
explain something is of course in itself very slight reason to suppose that
future science will not be able to explain it, but there are at least three areas
where, it has been argued, we are able to detect diculties in principle. First,
the facts of personal identity as they are presented to commonsense seem
to suggest that we as persons are (Descartes) or are constituted in part by
(Aquinas) units of substance which are indivisible over time, and souls are the
best candidate for such. Secondly, the facts of freedom of choice as they are
presented to commonsense roughly that the causal and moral responsibility
bucks stop in the same place can be accommodated by substance dualism,
but not by physicalism. And finally, what we have called the qualiatative
facts of consciousness, what it is like to see things like red, are not reducible to
facts about the physical properties of colours (or indeed colours and brains),
something which again can be accommodated by substance dualism but not
physicalism. Our discussion of all these points has perforce been very brief,
but I hope sucient to suggest reasons for this analysis. If so, one might sum up
90
Substance Dualism
our findings thus: theres one argument against substance dualism (its more
complex) and three in favour (it be er explains personal identity, freedom of
choice, and consciousness). If that is so, neither substance dualism nor physical-
ist substance monism will give us everything we want, and we shall naturally
turn to considering how we should weigh simplicity against these other consid-
erations when deciding what we have, on balance, most reason to believe.
Moore has taught us that we may take any valid argument in either of two
directions, as articulating a reason to suppose its conclusion true or a reason to
suppose one or more of its premises false, and that the direction in which it
is most reasonable to take a given argument will depend on whether the
premises are jointly more obviously true than the conclusion is obviously false.
So we may give the considerations presented here some direction by finally
asking ourselves this question: Knowing now that you can only believe one,
which of the following seems more obviously right to you?
z We are persons in more or less the same way that commonsense suggests;
we have freedom of the sort supposed in everyday life; and colours
and indeed mental happenings in general have qualiatative properties.
z The world is as simple a place as physicalism suggests.11
91
5 Physicalism
Barbara Montero
Physicalism, as some see it, takes the fun out of life. In their eyes, if physicalism
is true, the pleasure of a great bo le of wine, the euphoria of that first kiss,
the thrill of a hole in one and so much more are nothing but the workings of
the brain. At the same time, physicalism is probably the most widely held
general philosophical theory of the nature of the world, and many of those
philosophers who think that physicalism takes the fun out of life still defend
it tooth and nail. But what exactly is the theory of physicalism? Here I hope
to make some headway towards understanding physicalism, the theory that
many philosophers both love and hate. In particular, I aim to arrive at an under-
standing of the thesis of physicalism that captures its essence and at the same
time can be used to ground the contemporary debate over whether it is true.
Physicalism is a view about the ultimate nature of the world along the lines
of Thaless view that all is water or Democritus view that all is atoms in the
void. But rather than pronouncing all is water or all is atoms in a void, physical-
ism pronounces that all is physical, or as it is usually phrased, everything
is physical. Of course, this isnt very informative unless you know what it
is to be physical. Indeed, each term everything, is, and physical is
open to various interpretations. In what follows, I examine each of these
components in turn.
The Domain of Physicalism
Ontology is the very general study of reality. And physicalism is typically

thought of as an ontological theory: it tells us that everything is physical. But
everything is not always taken to mean literally everything. But if it doesnt,
just how much of reality is supposed to be captured by the physicalists
net? (Here and throughout I use the term physical broadly to cover not only
physical entities and properties at the fundamental level, but also physical
phenomena, such as rocks, trees and chairs.)
How one restricts the scope of physicalism depends on ones purposes. And
since the central physicalist target is typically the mental, it is not unusual
for physicalists and their foes to simply focus on the question of whether the
92
Physicalism
mental is physical. Indeed, some may even simply refer to the theory that the
mental is physical as physicalism. It may be that this is simply intended as
shorthand for the view that everything (or some significant subset of every-
thing) is physical. Yet this shorthand can be confusing when a more encom-
passing type of physicalism is evoked to justify physicalism with respect to
the mental, such as when physicalists argue that the mental is very likely to
be physical because everything else is physical. Obviously, here the scope of
everything else is not just the mental. So what, then, is supposed to count as
everything else?
Some understand physicalism in the broadest sense possible. It is theory
about everything whatsoever, a theory that says that all reality is physical. On
this inclusive conception, physicalism implies not only that people, animals,
rocks, trees, and all other concrete objects are physical, but also that abstract
objects which on some accounts include numbers, properties, classes, rela-
tions, and propositions are all physical. Even God, if she exists, would need
to be deemed physical given the truth of this conception of physicalism.
Others think that physicalism ought to have a more restricted scope. For
example, some understand it as a theory about only the concrete world, that
is, roughly about phenomena in space or time. Physicalism, then, is true if and
only if all phenomena in space or time are physical. This understanding of
physicalism ensures that the status of the mental is relevant to the truth of
physicalism, since, whatever else they are, mental processes do seem to occur
over time. However, the existence of abstract numbers (regardless of what they
are like in other respects) would not refute such a physicalism. Jerey Poland
can be seen as defending this conception of the scope of physicalism (if we
assume, as many do, that the abstract world has no causal influence on us)
when he claims that physicalists are (or should be) concerned with what exists
in nature that is, with what can be spatially and temporally related to us, with
that with which we can interact and by which we can be influenced, and with
that of which we and the things around us are made (Poland, 2001, p. 228).
A related approach to defining the scope of physicalism is to think of physi-
calism as a theory about the empirical world, that is, about the phenomena that
we come to know via our senses, or to put it more carefully, about phenomena
that are such that our knowledge of them must be justified via our sense experi-
ence. If, as is o en thought, our senses do not justify knowledge of abstracta,
this restriction allows for the existence of non-physical abstract entities to be
consistent with physicalism. However, if abstracta are known via our senses,
then, the truth of physicalism, on this interpretation, implies that abstracta are
physical.
A more encompassing view, such as Andrew Melnyks, takes physicalism
to be a theory about the contingent and/or causal world (Melnyk, 2003). If
abstracta are not causal or if they exist necessarily, this restriction comes close
93
to the previous restrictions. However, on this view, the truth of physicalism

implies that anything that has causal powers is physical. So, for example, if
abstract numbers have causal powers, then, on this version of physicalism,
numbers would need to be physical in order for physicalism to be true. More-
over, on this understanding of physicalism, even something that has no
causal aect on us, as long as it is contingent, would need to be physical if
physicalism were true.
Should physicalism have a restricted scope? If we were to restrict physical-
ism to only the concrete world we would not be able to make sense of what
might be called physicalist structuralism. Physicalist structuralists, such as
James Ladyman, hold that the fundamental properties of physics are purely
structural, revealing only the relationships between things and nothing of the
things themselves (Ladyman et al., 2007). Thus, the fundamental physical world
on his view is entirely abstract. Moreover, Ladyman holds that since the funda-
mental physical world determines everything, there is nothing else besides
structure, or as the title of his book declares, everything must go. If we were
to hold that physicalism is a theory of only the concrete world, Ladymans
view would be physicalistic in only a trivial sense.
Melnyks restriction, however, accommodates the physicalist structuralist
(assuming that the fundamental properties of physics are either contingent or
causal). His restriction also accounts for the intuition that if our world had
undetectable contingently existing spirits cohabitating happily among them-
selves, physicalism would be false. But what would the status of physicalism
be if there were a necessarily existing God who had no causal influence on us
or the world as we know it? On Melnyks view, the existence of such a God is
compatible with physicalism. But it is not clear that it should be.
Physicalists are drawn to restricted versions of physicalism as they are
easier to defend; Occams razor notwithstanding, it is very dicult to argue for
the view that, say, no undetectable spirits exist. Nevertheless, it seems to me
that an argument for physicalism in a non-restricted sense would still count as
successful even if it does not rule out impossible-to-rule-out situations, as no
theories can do that. In all theories outside philosophy, and most theories in
philosophy, save for in the domain of skepticism, one need not present a theory
as applicable to only the knowable world. So I think physicalists as well need
not say that the scope of physicalism is only that of which we can in principle
have knowledge. If it is false about that, it is still good enough. Of course,
restricting the scope of physicalism so that the existence of abstracta, no
ma er what their nature, could not refute physicalism is a dierent issue. It
seems that physicalists who take this route have the sense that abstracta are
not a threat to physicalism. However, I think that a be er way to accommodate
this intuition is, as I shall describe in section three, to merely count them as
94
Physicalism
physical. I propose, then, that we understand everything in the most inclusive

way possible:
Physicalism: Everything, whatsoever, is physical.
The relation between mountains and molecules
When the physicalist claims everything is physical, what is being said about
everything? Typically physicalists deem something physical if its existence
depends in the right way on basic or fundamental physical properties.1 And
typically the fundamental physical properties they have in mind are the micro-
physical properties countenanced by physics, such as the property of having a
charge, of being a quark, and so forth. In the third section I shall question this
conception of the fundamental physical properties. Here, however, I want to
ask, what exactly is the relation between the fundamental physical properties
and higher-level properties, such as mental properties, which is thought to
make the higher-level properties count as physical? In other words, when
physicalists say that everything is physical, just what is meant by is?
Some hold that the relation between higher-level physical properties and
fundamental physical properties is that of explanation (Jackson, 2006; Witmer,
2006). On this view it is thought that physicalism is true if and only if every-
thing is either a fundamental physical property or law, or can be explained
in terms of such properties and laws. As such, physicalism is an epistemic
thesis about what we can explain. It may have ontological implications since
typically we think that a good indication of whether the fundamental nature
of r is p is the fact that we can explain r in terms of p. Nonetheless, such a view
is primarily an epistemic thesis.
Many philosophers, however, see physicalism as an ontological thesis, a
thesis that tells us about what the world is like, whether or not we can under-
stand how it could be like this. Physicalism, many think, could still be true
even if we never arrive at a physical explanation of, say, pain, as long as pain
is an entirely physical phenomenon. As Joseph Levine puts it, I am prepared to
maintain that materialism must be true, though for the life of me I dont see
how (Levine, 1998, p. 475). And some philosophers such as Brain Loar (1990)
and Colin McGinn (1989) have proposed theories about why we cannot
understand physicalism could be true of the mind, even though they think that
physicalism might very well be true.
To make sense of positions such as these, physicalistic dependence relations
cannot be formulated in terms of explanation. Of course, most advocates of
thinking about physicalism in terms of explanation do not mean that we can
95
provide a physicalistic explanation of pain now, nor even sometime in the

future. Rather, they think that for physicalism to be true, such an explanation
must be in principle possible. But it is usually not clear what principle is at use
here. The idea that there is an explanation that the human mind can grasp might
seem too restrictive. Why should there not be phenomena that are beyond the
grasp of human intelligence?2 However, it is dicult to grasp what it would
mean for an explanation to be possible for an ideal mind, a mind that is capable
of knowing everything.
In any event, many formulations of physicalism employ an ontological rela-
tion between lower-level physical properties and higher-level properties that is
supposed to capture the idea that higher-level properties are nothing over and
above lower level properties. For example, it is supposed to capture the idea
that a mountains height is nothing over and above the cumulative height of
the rocks, pebbles and earth that compose the mountain, and that the rocks,
pebbles and earth are nothing over and above the molecules out of which they
are composed, and that the molecules are nothing over and above the atoms
out of which they are composed, and so on.
Already, however, we run into diculties, for arent there properties of, say,
Mt. Fuji that are not dependent on the properties of the dirt that composes it?
For example, Mt. Fuji has the property of being revered in Japanese society, yet
it is not clear that the rocks and pebbles have this property or have any other
properties that would imply that the mountain should have this property.
Physicalists address this type of worry by broadening the dependence base
for the physical world. Perhaps all the properties of Mt. Fuji do not depend
entirely on the properties of its parts, but they nonetheless do, says the physi-
calist, depend on fundamental physical properties. If we set all the fundamental
physical properties of the world, we will have set Mt. Fujis property of being
revered in Japanese society since, according to the physicalist, we will have
set the Japanese peoples reverence of it as well.
But what exactly is the relationship between the properties of Mt. Fuji and
the properties of molecules? Though there is considerable disagreement over
how physicalists should explain the relationship between Mt. Fujis properties,
or other higher-level properties, and the fundamental physical properties, many
think that the relationship involves, at a minimum, upward determination, or
what is also called supervenience. Upward determination is typically expressed
as the view that any world that duplicates all the fundamental physical proper-
ties and laws of our world also duplicates all properties of our world. So it
implies that any world that duplicates the microphysical properties of our
world would duplicate Mt. Fuji, as well as all other higher-level features of
our world, including minds.
The relation of upward determination, or supervenience, is sometimes
explained metaphorically as the view that all God had to do in order to create
96
Physicalism
the world was to create the fundamental properties of physics. A er this she
could rest, as everything else came along for free.
How close does upward determination take us to physicalism? Upward
determination states that any world that duplicates all the fundamental
physical properties and laws of our world also duplicates all properties of
our world. But now imagine a necessarily existing God. A world that dupli-
cates all the fundamental physical properties of our world would also dupli-
cate such a God. Yet, intuitively, the existence of God refutes physicalism. If
this is correct, then upward determination is not a sucient condition for
physicalism.
To be sure, if this necessary God interferes freely with the workings of the
world, a fundamental physical duplicate of our world might not duplicate all
aspects of our world, for God might arrange things so that in the duplicate
world, although all the fundamental feature of the world are the same, I prefer
coee to tea. As such, upward determination would fail. However, if the role of
God were merely to set the fundamental nature of the world, merely to be the
hand behind the big bang, as it were, then a necessarily existing God would
be consistent with upward determination.
If you accept Humes view that there are no necessary connections between
distinct entities, then such a God cannot exist.3 Such a God is distinct from the
rest of the world, yet her existence is necessary, given the world. Alternatively,
one could restrict the scope of physicalism so that such a God would be consis-
tent with the truth of physicalism. But if you reject Humes view and also think
that the existence of God is incompatible with physicalism, you are led to reject
upward determination as perhaps a necessary condition for physicalism, but
not a sucient one.
The desire to find both a necessary and sucient condition for physicalism
has led some philosophers to hold that explanation plays a role in our under-
standing of physicalism a er all.4 Physicalism, as they see it, is not just the
view that everything is determined by fundamental physical properties, but
that everything is determined and ultimately explained by the fundamental
physical properties. Such a view presumably rules out a necessarily existing
God from counting as physical. And if it doesnt, such a God would seem to be
physical.
But many are content with a mere necessary condition since much of the
action in the literature on physicalism involves various arguments against
physicalism, all of which purport to show that upward determination, which
is taken to be a necessary condition for physicalism, fails to hold. For example,
the zombie argument against physicalism is intended to show that the pos-
sibility of zombies not the lumbering Hollywood variety, but creatures that
duplicate our microphysical structure yet lack consciousness implies that
consciousness is not physical.
97
Is there a way to satisfy the desire that physicalism should be both an

ontological thesis and incompatible with a necessarily existing God.
Physicalism: Any world that duplicates all the fundamental physical proper-
ties and laws of our world (and contains no other fundamental properties) also
duplicates all properties of our world and everything in our world is ultimately
constituted by fundamental physical entities.
Assuming that both immaterial souls and a necessarily existing God have
nonphysical fundamental properties, this view implies that their existence is
incompatible with physicalism, which is just what we want.
The Physical
Now we must address the question, what is the physical? When we say, for
example, that everything is determined by fundamental physical phenomena,
what are these fundamental physical phenomena? Most define the fundamen-
tal physical properties in terms of the entities and properties and perhaps
laws posited by microphysics: the fundamental physical phenomena are those
entities and properties mentioned in the theories of microphysics. But what
is meant by microphysics? Is it current microphysics? This would provide a
relatively clear position: physicalism would then be the view that all of the
fundamental properties are properties of microphysics. Unfortunately, this is a
theory that is rather dicult to accept since we know that current microphysics
is most likely neither entirely true nor complete and thus we now know that
it is most likely not true that all higher level properties are determined by the
properties of microphysics.
A more common understanding of what counts as the fundamental physical
properties in the thesis of physicalism is that they are the properties posited
by an ideal physics, a true and complete physics, or a physics in the end.
Can we formulate physicalism in terms of a true and complete physics?
Of course, we do not currently know what future physics will be like, and
therefore we cannot now determine whether physicalism is true. But perhaps
physicalism can be seen as a hypothesis that awaits scientific confirmation (or,
for that ma er, refutation). Physicalists, on this understanding, are be ing that
it is correct, but do not claim to be able to now determine that it is correct.5
I see no problem with making physicalism a thesis that awaits empirical
support. However, it seems that far from turning physicalism into a thesis
whose truth awaits empirical support, defining the physical in terms of a true
and complete physics actually seems to turn physicalism into a trivial truth.
For what is a true and complete physics, save for one that accounts for the
fundamental nature of everything? If free floating souls exist in our world, a
completed physics will, by definition, account for the most fundamental nature
98
Physicalism
of these souls. Yet neither physicalists nor their foes think that at this time in the
debate physicalism is true merely as a ma er of definition. Physicalists think
the thesis needs to be argued for and, as many hold, will ultimately depend on
what scientific investigation reveals. And their foes clearly do not think that
they are denying what amounts to, more or less, an analytic truth. It seems,
then, that physicalists who define physicalism over a true and complete physics
cannot simply mean by this a theory of everything since then their claim that
the mind is physical is trivially true. Yet, there is also reason to think that they
do not simply intend to refer to the temporal end of physics. For this physics
might still be inaccurate and incomplete; even worse, for all we know, physics
might regress. We need, then, another route to defining the physical.6
Some argue that there are phenomena that physics and perhaps scientific
investigation in its entirety does not aim to cover. Rather, physicists, they argue,
in their role as physicists, are only concerned to account for a certain class of
phenomena and souls and spirits are not in this class. As such, the truth of
physicalism becomes open to debate. The question, then, is: Are there no other
fundamental properties than those that are under the hegemony of a true and
complete physics, where what counts as being an object of study for physics
is restricted in certain ways?
This makes physicalism admirably more risky, but should we assert that
physics has identifiable limits (besides, of course, that which is by definition
unknowable)? As I see it, it makes good methodological sense to hold that
scientific inquiry should not accept a priori barriers. Certainly, it would be
reasonable to say that as things stand, government grant money ought not to
fund physics research into the properties of souls. This research would seem
to be currently hopeless. However, the claim that physics should never investi-
gate the nature of souls even if in some currently unfathomable way a physics
lab reveals signs of souls is a much stronger claim. And, indeed, it seems
that such barriers could hinder progress. In other words, it seems that a good
approach to scientific investigation is that when you discover territory that
does not conform to your map, change the map, not the territory. Such changes
might involve not only expanding our scientific ontology, but changing our
scientific method as well. For example, if standard controlled experiments fail
to reveal phenomena that we nonetheless think exist as some have claimed
could be the case with parapsychological phenomena we should try to find
a way to change the control. If we were somehow convinced that there
was a spiritual realm that was causally isolated from our world, let us try to
understand it.
Where does this leave us? I think that it indicates that, despite the consonance
of the two terms, the physical should actually not be defined over physics.
Physics is the study of the fundamental nature of the world, whatever that
nature may be. But physicalism is more discriminating about what is to count
99
as fundamentally physical. Even if fundamental acts of pure consciousness

were part of the domain of physics (as the physicist Eugene Wigner claimed
were required to explain the collapse of the wave function) they should not
count as physical. But if physics is not our guide to what counts as physical,
what, then, is?
Physicalism is an ontological thesis, but it is an ontological thesis that is
supposed to capture the sentiments of those who call themselves physicalists
while presenting a thesis that those who think of themselves as opposing physi-
calism will reject. And thus we are looking for an understanding of physicalism
that classifies free floating minds, a God that is not determined by anything
other than God, and fundamental, irreducible norms all as nonphysical.7 I think
that we can achieve this if we merely define the fundamental, physical pro-
perties negatively, that is, in terms of the types of properties it excludes.
The fundamental physical properties, then are the fundamental non-mental,
non-divine and non-normative properties.
But why should those and only those be excluded on a physicalistic con-
ception of the world? While most philosophers would agree that physicalism
does indeed exclude those sorts of properties, what exactly it ought to exclude
is somewhat of an open question. For example, some but not all see vitalism as
anti-physicalistic as it posits a fundamental life force. But in any event, as long
as one makes it clear at the outset what types of fundamental properties are
to count as non-physical, we have a framework around which debates over
the truth of physicalism can proceed.
Filling the framework in, here is the theory of physicalism we have
arrived at:
Any world that duplicates all the fundamental non-mental, non-normative

and non-divine properties and laws of our world (and contains no other
fundamental properties) also duplicates all properties of our world, and
everything in our world is ultimately constituted by fundamental non-
mental, non-normative and non-divine entities.
This way of understanding physicalism may be somewhat of a mouthful,

but it seems to capture the spirit of physicalism since it is inconsistent with
the existence of such things as immaterial souls and mental properties that are
over and above the physical (even if their existence follows from necessity given
the physical domain.) Or rather, it is inconsistent with such things as long as
they count as fundamental. If, however, they are determined by (but do not
determine) non-mental, non-celestial and non-normative properties, they count
as physical, as they should.
But doesnt this leave us with just a disparate list of properties that are to
count as non-physical? Physicalism, according to Frank Jackson, is the very
100
Physicalism
opposite of big list metaphysics. Rather, he sees physicalism as highly

discriminatory, operating in terms of a small set of favoured properties and
relations (Jackson, 1998, p. 5). To be sure, the list of properties excluded by
via negativa physicalism is hardly large, yet one might still want to know what
unifies the nonphysical properties besides the fact that there are a number of
people who call themselves physicalists who simply dont like them. Why is
it that physicalists do not like these properties?
Why should certain properties, such as fundamental properties that are
mental, count as non-physical? I think that certain properties have been deemed
physically unacceptable because they hint at a world that was created with
us in mind. If mental phenomena were fundamental, being, for example, part
of the original brew that was set in motion in the big bang or as emerging as
something extra along the way, mentality would have a place of prominence
in the world. And this, I think, for many, suggests the existence of a God who
was looking out for us. This hint, however, is not an implication, and anti-
physicalists can be atheists. However, I think that non-physical properties
have go en their bad reputation because on many accounts of God, these are
the sorts of properties that would exist, if God were to exist. And the reputation
remains, even when its origin is forgo en.
As should be the case, if you are a theist, you will reject physicalism, as
defined. However, this physicalism does not seem to take all the fun out of
life. If physicalism is true, the pleasure of a great glass of wine need not
be merely something going on in your brain. Rather, if physicalism is true,
such pleasure is determined by neural properties and ultimately fundamental
non-mental properties, but it is as real as anything. But is physicalism true?
This question, alas, I shall need to save for another discussion.
101
6 Folk Psychology and
Scientic Psychology
Barry C. Smith
Rational beings have a propensity to recognize other rational creatures. We

anticipate and respond to others, not as moving bodies but as agents: minded
creatures with motives for action. To treat their behaviour as action is to see it
as governed by their intentions; intentions fixed by their beliefs and desires. It
is because we see agents as acting in the light of their beliefs and desires that are
able to make sense of their behaviour in rational terms, and it is this tendency
to see peoples behaviour in rational terms that lies at the heart of our everyday,
folk psychological understanding of others. At the same time, psychologists
and neuroscientists have made great progress in uncovering the cognitive states
and mechanisms that explain human capacities for perception, speech and
action. So how do these two, very dierent types of explanation fit together?
This essay explores the relations between folk psychology and scientific psy-
chology by considering the views of three very dierent philosophers of mind
and where they stand on this issue. The philosophers are Donald Davidson,
Daniel Denne and Jerry Fodor.
Common Sense Psychology
When we explain peoples behaviour in intentional terms, we take their beha-

viour to be the upshot of mental states that gave them reasons to act. Citing
individuals wants and wishes, thoughts and feelings, hopes and fears, helps
to explain both what they are doing and why they are doing it. We routinely go
in for such explanations as part of our psychology of other people. Central to
such commonsense, or folk psychology, is ascription of particular beliefs or
desires that belong to a larger pa ern of a itudes and actions that makes sense
of peoples acts and u erances. For example, when we notice someone leaving
the room discreetly, we may say, She le the room because she wanted to
escape from the party unnoticed. Ascribing to her this desire is one part of a
larger explanation that would include her beliefs about her surroundings and
her beliefs about what she has to do to satisfy her desire. (Notice that simply
102
having a desire to escape the party unseen would not lead her do anything
unless she also believed she could leave the party unseen by exiting swi ly from
the room.) Similarly, we could have cited a belief, She tiptoed past the drawing
room because she thought no one would hear her leave. Again, the explanation
is partial, and we are assuming that she had a desire to leave without anyone
hearing her. We may mention a belief or a desire even though it is beliefs and
desires together that constitute someones reasons for acting, and in citing these
reasons we are oering rational explanations of an agents behaviour.
The beliefs and desires posited to explain behaviour, are hypotheses that can
be re-worked in the light of further evidence. The beliefs ascribed must make
sense in the light of other beliefs, and similarly, desires must make sense in the
light of further desires. Thus it is part of our competence in giving such expla-
nations that we stand ready to adjust our a ributions of belief and desire if they
are not consistent with what else it makes sense to ascribe to an individual on
the basis of what they say and do elsewhere and at other times. The overall
picture must make sense of the individual as by and large a rational thinker,
and these relations between beliefs and beliefs, between desires and desires,
and between beliefs, desires and actions, are logical or rational relations.
Propositional Attitudes and Intentional Actions
We call beliefs and desires propositional a itudes because an agent can take
dierent a itudes such as believing, desiring, hoping, or fearing towards
the same proposition, for example, that the war on terror will continue for a
long time. Alternatively, one can take the same a itude towards dierent prop-
ositions. It is the logical relations between propositions believed that ensures
the rational connections between propositional a itudes. For example, if we
believe that the US president J. F. Kennedy was assassinated we must also
believe that he is dead, as well as believing that there was someone called
J. F. Kennedy who was president of the United States, etc. If Oswald intended
to kill Kennedy then he must have wanted Kennedy dead and believed that
by firing the gun he could bring about that outcome (i.e. he must believe the
death of Kennedy would satisfy that desire). If he didnt want Kennedy dead
(but wanted merely to frighten him) then we can suppose his intentional action
is incorrectly described as his killing Kennedy even if the shot he fired acciden-
tally resulted in Kennedys death. Perhaps he was genuinely surprised that
Kennedy died a er all, it was a very improbable shot. His subsequent surprise
would make no sense if he all along desired to kill Kennedy and believed that
pulling the trigger would result in Kennedys death. Even if Oswalds observ-
able behaviour, his li ing the rifle, aiming and pulling the trigger was the same
in both scenarios, his surprise would be a reason for supposing that he didnt
103
intend to kill Kennedy, and for denying that we should call his action a killing.
By contrast, his lack of surprise and satisfaction would make rational sense of
supposing he did intend to kill Kennedy, and that killing is the correct descrip-
tion of his action. The correct description the description under which an
action is intentional depends logically on the intentions of the agent: the par-
ticular beliefs and desires he or she has. These propositional a itudes rational-
ize the agents behaviour, enabling us to see it as an intentional action. Thus if
the intentional description of the action were dierent, the a itudes we ascribe
to the agent would have to be dierent too. In courts of law whether someones
action should be described as murder or manslaughter depends on whether
they had a premeditated intention to kill before they acted as they did. The
behavioural event is the same but whether it counts can be intentionally
described as an act of murder or an act of manslaughter depends on which
mental states it is correct to a ribute to the agent (i.e. what his reasons were in
behaving as he did).
Beliefs and Desires as Part of an Intentional Network
We make sense of peoples beliefs and desires in the light of further beliefs and
further desires we have grounds to a ribute to them, based on what they do
and say elsewhere and at other times. And we aim for the most consistent over-
all interpretation of someones actions and u erances by constructing a net-
work of intentional a itudes and actions that makes best sense possible of their
overall behaviour. We constantly rework and revise our portrait of someones
mental life in the light of further evidence. We find ourselves equipped to do
so without any explicit training. We operate quite instinctively in forming views
about other peoples states of mind. Premack and Woodru (1978) coined the
term theory of mind for this set of abilities. It is not an explicit theory, of course,
but it amounts to a tendency or ability on the part of normal human thinkers
one we develop fully by about the age of four to make hypotheses about the
beliefs, desires, hopes and fears of fellow humans and to explain and, to some
extent predict, their actions. It is the nature and status of this everyday folk
psychology that we will now examine.
Folk Theories
Folk psychology plays much the same role in dealing with others, as folk
(or naive) physics plays in our understanding of physical objects and forces.
We know that folk physics is not literally true. We, the folk, say that the sun
goes down behind the hill even though it is the earth that goes round the sun.
104
We say that cold water cools the hot water in our bath, even though what
actually happens is that the hot water heats up the cold and thus loses its kinetic
energy. In these cases, the proper scientific explanation replaces the rough and
ready generalizations of folk theorizing about how the world works.
Is the same true of the generalizations of folk psychology? Will they eventu-
ally be replaced or revised when we learn more from the science of the mind?
Will neuroscience gradually replace the false but appealing assumptions of
the folk? If that is not the model of the relation between folk psychology and
scientific psychology, will belief-desire psychology be vindicated or refined
by scientific psychology? The worry is that if folk psychology is not reducible
to scientific psychology we seem to have competition between two explana-
tions of the same purposeful behaviour and perhaps only one of them can be
genuinely explanatory.
To ask whether the particular claims and generalizations of folk psychology
are true we need to know what they commit us to, and to know what kind of
explanation folk psychology provides. As we have already seen, folk psycho-
logical explanations of behaviour are rational explanations: the giving of reasons as
to why people do what they do. But such explanations are also causal explanations.
The reasons we cite the beliefs and desires that make sense of people acting as
they do are also the causes of their behaviour. When we say Charlo e le the
party because she wanted to catch her train, the because is used in a causal
sense. (Contrast this with, She broke the law because she parked on a double
yellow line.) But if folk psychological explanations are intended as causal expla-
nations explanations of what brought about certain events the key question
is whether folk psychological explanations are literally true of human agents?
The eliminativist threat to common-sense psychology
Are our descriptions of our own and others behaviour as rational actions
really true: is behaviour really the upshot of beliefs and desires at work in us?
Or are we actually caused to behave as we do by something entirely dierent:
configurations of neural firings in the brain that have nothing like the neat
structure of belief and desire? What aspects of reality are we picking out when
we make ascriptions to people of beliefs and desires, citing these states as the
causes of their actions?
Let us look at the eliminativist challenge. What if states like belief, desire, hope
and fear, cited in our common sense psychological framework, turn out at
no level of organisation to be among the causes of human behaviour? Our
everyday psychological scheme would misrepresent our internal states and
activities, just as conceptions of the world that mentioned witches, ether,
or phlogiston, in earlier thinking, misrepresented the nature of reality. Paul
105
Churchland has suggested that there may be some reason to accept the legiti-
macy of this challenge to common sense psychology because of the failure of its
theoretical concepts to line up with the categories of neuroscience. The former
may have to be eliminated by, rather than reduced to the categories of a fully
mature neuroscience of cognition and action (see Churchland, 1978).
This is the eliminative materialists option. And even if we reject it, the threat
it poses is real enough. Consider the following passage from Brian Loars Mind
and Meaning (1981):
If it were to turn out that the physical mechanisms that completely explain
human behaviour at no level exhibited the structure of beliefs and desires,
then something we had all along believed, viz. that beliefs and desires
were among the causes of behaviour, would turn out to be false. Naturally,
we would continue to use the belief-desire framework to systematize
behaviour, but that should then at the theoretical level have the air of
fictionalising and contrivance. (Loar, 1981, pp. 145)
This fictionalising and contrivance oers a form of irrealism about psycho-

logical talk, where we are either in error when we speak this way, or else merely
using such talk without any pretention to describe what is real. We can call
this option, non-factualism about psychological states.
Realism about folk psychology
If we are to resist non-factualism and adopt a realist construal of folk psycho-

logical descriptions of peoples mental lives, we need to answer a number of
questions. What does an account of folk psychology our common sense ascrip-
tion of beliefs and desires to one another to explain our behaviour in rational
terms commit us to? What would be required of any successful a empt to
vindicate folk psychology? What aspects of reality are picked out by the notions
it uses? Can they give us an adequate philosophical conception of the mind?
If they can we need to say what relation obtains between this level of mental
description and the levels of description invoked by cognitive psychology and
neuroscience.
There are three key desiderata any satisfactory folk psychology:
(1) It should provide a rational explanation of certain behavioural events as

actions.
(2) It should accommodate the causal ecacy of the mental.
(3) It should accommodate first- and third-personal ascription of beliefs,
desires.
106
We have touched on the rational and causal requirements already, and we

will return to them below. But now let us consider the third requirement. Any
adequate account of folk psychology has to acknowledge that the very same
psychological states (beliefs and desires) that we a ribute to others can also
be a ributed to ourselves. It is part of our psychological self-knowledge that
we currently know what we are thinking and what we want. And yet the basis
for ascribing psychological states to ourselves diers from that on which we
ascribe such states to others. Ascriptions of beliefs and desires to others depends
on evidence and inference: we observe peoples behaviour and a ribute states
that, together with other background beliefs and desires we are prepared to
ascribe to them, make rational sense of their behaviour; while in our own case,
we ascribe a itudes to ourselves from the first-person point of view without
relying on evidence or inference. We do not have to observe our own behaviour
in order to know what we currently thinking or what we want. We just know.
We know our own minds best, be er than we know the minds of others and
be er than they know our minds. So there is an asymmetry in grounds for making
first-personal and third-personal psychological ascriptions to someone.
Folk psychology and self-knowledge
We ascribe psychological a itudes to ourselves without basing our ascriptions

on evidence. We do not have to theorize about ourselves. (Although there are
occasions on which we try to figure out what we really want, or what our
motives were in acting as we did.) We are authoritative in our self-ascriptions,
but not similarly authoritative about someone elses mental states. Ordinarily,
others know best what they think and what they want. This is a further epistemic
asymmetry between first- and third-person perspectives on the mind.
Any satisfactory account of our psychology must recognize that we ascribe
beliefs and desires to others and to ourselves: that there are first-person and
third-person a ributions of beliefs and desires and asymmetries between these
two modes of a ribution. The grounds of our psychological self-knowledge are
very dierent from that of our knowledge of other minds, so any decent account
of our everyday folk psychology will have to make room for a satisfactory
account of our psychological self-knowledge.
Scientic Psychology
The science of psychology aims to explain a vast range of human cognitive

capacities and abilities that underpin and make possible our propensity for
perception thought and action. As currently practiced, it is unlikely to satisfy
107
key requirements (1) and (3) on a satisfactory folk psychology. The various
accounts it gives of the cognitive states and processes that subserve specific
abilities for vision, language, audition, and motor-control are states not known
to agents first-personally. Nor are they ordinarily known third-personally.
Instead, we should think of them as states of sub-personal mechanisms not
states we a ribute to persons posited by a theory of the internal cognitive
mechanisms that subserve particular capacities. Such content-bearing cognitive
states do not provide rational explanations of the behaviour in which we display
our abilities and capacities We see distance and depth in the visual field, but
the psychological states and processes responsible for this aspect of the visual
scene do not give the sighted person reasons to see things this way: they causally
explain why creatures with stereoscopic vision do see things this way in virtue
of the content of those states. The common element that such underlying states
of our cognitive systems share with the psychological states posited by folk
psychology is that both are content-bearing or representational states; but the
cognitive states invoked by scientific psychology do not have propositional
contents that sustain logical and hence rational relations to one another.
Scientific psychology, in particular cognitive psychology, extends the domain
of content-bearing psychological states without extending the domain of
rationally governed states, thus marking a division within the mind.
Given this division within the mind between dierent kinds of mental states
and the dierent kinds of psychological explanations in which they feature,
how can a psychology that rationalizes behaviour in terms of beliefs, desires and
intentions accommodate an underlying sub-personal psychology that explains
capacities exhibited in the same behaviour? Does the science underpin or
undermine the folk psychological we give of ourselves and others?
We shall look first at a purely a priori defence of the legitimacy of folk
psychology. This view rejects the idea that common sense psychology could
be answerable to scientific psychology by denying that there could be a science
of the mind.
This is the interpretationist view of the mental set out by Donald Davidson.
Davidsons Philosophy of Mind
Is the commonsense (or folk) psychology we use to make sense of one anothers
verbal and non-verbal behaviour vulnerable to scientific challenge? Could
findings in scientific psychology or neuroscience show that a scheme we use
for explaining human action is fundamentally flawed and mistaken?
An a priori defence of commonsense psychology, if successful, would render
it immune to scientific challenge. This is the view of the mind proposed by
Davidson, who sees mental life as constituted and exhausted by the application
108
of a priori principles of interpretation used to make sense of one anothers

behaviour in rational terms. It is this a priori characterisation of the principles
of commonsense psychology that renders further empirical investigation of
its standing irrelevant. For Davidson, these principles characterize what it is
to be minded in the first place.
Levels of description
According to Davidson, our understanding of the mind depends wholly on the

concepts and categories we use to ascribe mental states to one another. We are
gradually inculcated into a practice of ascribing beliefs and desires to one
another, and to seeing each other as rational agents engaged in purposeful
activity. The practice of ascribing or reporting our own or others beliefs plays
a constitutive role in the nature of those beliefs; in an account of what beliefs
fundamentally are. For Davidson, the theory of belief is a theory of belief-
ascription.
The mental is constituted and exhausted by the intentional categories we
have for ascribing states of mind to one another. We ascribe such states to make
sense of one anothers actions in rational terms and our psychological idiom
cannot be reduced to the vocabulary of the physical sciences. Our everyday
psychological talk describes a normative structure of a itudes and actions that
makes rational sense of peoples behaviour. For Davidson it is the rational rela-
tions between the a itudes (and actions) that cannot be captured in purely
physical terms. Having a particular belief requires us to have others logically
connected with it. But these rational connections have no echo at the physical
or neural level. Physical or neural states do not rationally or logically require
the existence of other physical states. The irreducibility of intentional to physi-
cal vocabulary does not, however, mean that there are irreducible mental
entities. There is just one set of entities (events) and two vocabularies to describe
it: the intentional and the physical, with some of the events we describe in
physical terms being also describable intentionally. Events are mental events
just in case they can support intentional descriptions: an event is an action if
an only if it can be described in a way that makes it intentional. (Davidson,
1980, p. 229)
But what are the constraints on whether an event is intentional or not, that
is, on whether it can be described in intentional terms? Not just any physical
event is a mental event. We need a way of deciding which physical events are
mental events, and this is se led by which physical events sustain intentional
re-description as mental events. Such intentional re-descriptions are introduced
by interpretations of bits of the agents behaviour as rational actions: an episode
of behaviour for which a rational explanation can be given, an explanation that
109
gives the agents reasons for behaving as she did. The reasons cited are given
in terms of what people believe, desire and intend.
Notice that the target of rational explanation is not behaviour itself, where
behaviour is construed as bodily movement, but action: intentionally described
episodes of behaviour or bodily movements that amount to an agents perform-
ing an action for a reason. It is these intentional actions as observable aspects
of mind for which we give rational explanations. Equally, actions as part of
the mind of a creature only come into view when we interpret that creature
as a rational agent acting on beliefs and desires. Thus rational interpretation
constitutes its own explananda by coming to see certain bodily movements
as actions. When a bodily movement is interpreted as an intentional action
part of someones mind it is the very same event that can be described in
physical and mental terms: all we have are two descriptions of the same thing.
The mental is an intentional level of description of otherwise physical events.
As Davidson puts it: events are mental only as described (Davidson, 1980,
p. 215). Events are particular, datable unrepeatable occurrences, and all events
are physical (i.e. physically describable). Some of these events are also mental,
that is, correctly describable in mental terms. When we re-describe a episode
in a persons physical history a bodily movement in intentional terms,
describing it as the action of an agent intentional under a certain description,
we see it as part of her mental life. And in treating certain behaviours as part of
someones mental life as actions undertaken we thereby introduce beliefs
and desires as the agents reasons for performing those actions, and at that
stage we retrospectively identify those mental states with the physical or
neurological states that are the causes of the bodily movement in question.
Keeping the mental and physical levels of description separate, Davidson
argues for a non-reductive physicalism he terms anomalous monism:
(a) Some mental events interact causally with physical events.

(b) Events related as cause and eect fall under strict deterministic laws.
(c) There are no strict psycho-physical laws (and no strict psychological
laws).
The mental events that cause bodily movements are identical with physical
events: an event described in mental terms can also be described in physical
terms, and it is under its physical description that it can instantiate a strict
causal law couched in purely physical vocabulary. The physically described
event, which is also described in mental terms, enters into causal relation
with another event described in physical terms. Events are related as cause
and eect when they have descriptions that instantiate causal physical laws.
So even though there are no strict laws for predicting or explaining psycho-
logical phenomena, mental events can still cause physical events. The mental
110
event that is the cause of some piece of behaviour has a description under
which it instantiates a causal law that links physical events of that kind to
physical events of some other kind. The singular causal statement linking
As reasons on a particular occasion to what he did will be true only if it is
backed by a general causal law that relates events of that kind when physically
described.
Davidson and the Causal Requirement
It may seem that it is only at the physical level that causality occurs, and so the
mental may seem causally irrelevant or epiphenomenal. Physically described
events would continue to have the eects they do whether or not they were
describable in mental terms. But care is needed here if we are not to misunder-
stand Davidsons position. Davidson believes in the supervenience of the
mental on the physical: there can be no mental dierence without a physical
dierence, and any two events alike in all physical respects will be alike in all
mental respects. So when one event, physically described, causes another,
physically described, and the first is also describable in intentional terms as a
mental event and the cause of the resulting behaviour, it could not have been
that very event if it did not have that intentional description; for if it had been
dierent in some mental respect (i.e. not having any mental description), it
would have had to be dierent in some physical respect on pain of violating
supervenience, and then it would have to be a dierent event all together.
Whenever physically described events are also describable in mental terms,
those events count as mental events and cannot be otherwise unless things
had been physically dierent. So the challenge to causal ecacy that suggests
things could have taken place physically whether the mental was present
or not fails, and Davidson can still claim that the mental events can causes
physical events.1
Anomalousness and the Holism of the Mental
At the outset, each a empt at explanation is a move within an interpretative

scheme, an initial description of someones action that makes sense only in the
light of certain beliefs and desires that give the agent reasons for acting. And
the ascription of these beliefs and desires in turn makes sense only in the light
of further propositional a itude ascriptions:
Beliefs and desires issue in behaviour only as modified by other beliefs and
desires, a itudes and a endings, without limit. Clearly this holism of the
111
mental realm is a clue to the autonomy and to the anomalous character of

the mental. (Davidson, 1980, p. 215)
Each of these further rationally related a itudes must be justified in the

light of what the person says and does elsewhere and at other times. When
interpreting a creature, each piece of evidence on behalf of a given interpre-
tation retains its evidential standing just in case it forms part of an overall
interpretative theory that incorporates more and more of a persons speech and
behaviour within an overall intentional scheme: Every case tests a theory and
depends on one (Davidson, 1980, p. 221). In this way, the accumulating evi-
dence in favour of a given interpretation is part of the constitutive fabric of the
interpretation itself. And the question of the correctness of any single inten-
tional description of an agent depends on the cogency of the overall intentional
network of beliefs and desires to which that particular description belongs.
Whether a given intentional description is correct depends not on its fit with
some isolable physical or neurological fact about the creature, but on whether
we have respected the prevailing conditions for interpretation and satisfied
the interpretative principles of rationality and charity (see below), and whether
that interpretation enjoys the best possible fit (up to indeterminacy) with the
overall facts of that agents physical history. Thus what it is for an agent to have
a particular belief (or desire, etc.) is for him or her to be apt to be ascribed that
belief (desire, etc.) in the course of giving an interpretation that makes the best
sense possible of his or her total history and behavioural conduct.
A Priori Principles of Interpretation
The principles that ensure the coherence of a persons a itudes and actions,
from the point of view of an interpreter, are the principles of rationality and
charity. Beliefs should make rational sense in the light of other beliefs to
which they are logically related by the contents of those beliefs; and the actions
we see people as performing should make sense in the light of the beliefs and
desires we are prepared to ascribe to them and which give them reason to do
and say what they do and say. In addition to rationality, the other constraint on
correct interpretation is the principle of charity, which ensures that a persons
beliefs about their surroundings should be, by and large, correct by the inter-
preters lights. To be interpretable, most of an agents ordinary beliefs about the
world around her should be true. And the correct interpretation of a persons
behaviour will be the one that ensures the best fit with these two principles.
The principles of rationality and charity must guide us in the course of
building up a portrait of someones overall propositional a itudes and actions.
They do so when we ascribe particular beliefs, desires and meanings that make
112
sense of someones actions and u erances, in the light of further beliefs, desires
and other meanings that comport with the persons overall behaviour and their
surroundings.
Rationality requires that if there are two interpretations, each consistent
with everything the person does and says, we should favour the one that makes
their network of beliefs, intentions and actions, more and not less rational. The
principles of practical reason suggest that people will by and large do what
they believe will secure their fondest wish at that moment, assuming no coun-
tervailing beliefs and no countermanding desires. For you to do otherwise
would make no sense to us as interpreters. Charity requires interpreters to
make sense of people by a ributing to them beliefs it would make sense for
them to have given their current surroundings. Generalizing, we should inter-
pret people charitably by ascribing to them beliefs that are largely true (by our
lights). That is, we should not gratuitously ascribe to someone a bizarre belief
about what is going on around them, but assume, charitably, that they are like
us in having reliable beliefs about their current surroundings. Notice that the
principles of rationality and charity interact: to ascribe to someone an outland-
ish belief may result, through the holistic connection between a itudes, in a
scheme of interpretation for that person that makes him less rather than more
rational. On the other hand, if ascribing to someone veridical beliefs about his
current surroundings would make his behaviour less rationally explicable than
ascribing a false belief to him, we must forgo charitable interpretation on this
occasion. By and large, Davidson thinks the conditions for interpretation show
that there is a large degree of truth and consistency in the thought and speech
of an agent.2 The coherence of belief is guaranteed by the joint application of
charity and rationality. If there is not enough coherence in someones beliefs we
would not be justified in regarding them as rational agents at all. Even irratio-
nality assumes that a prior rational standard is operative and that a person
is going against it on this occasion. To assume that no rational standard is
operating is to see a creature as non-rational.
As interpreters we must always a empt to see peoples a itudes and actions
as making rational sense by our lights. This is not a subjective view of rationality.
What enables me to see pa erns of rationality in the a itudes and actions of
others is the rationality at work in my own thought and talk. The standards of
rationality that enable me to make sense of others are the standards that enable
me to make sense simpliciter; this is what enables other to make intelligible
sense of me. Thus, there is just one standard of rationality for all or at any rate
for those who are mutually intelligible.
Rationality and charity are relative a priori principles: they are not principles
designed to get at the (independently constituted) facts of someones mental
life; rather it is only when the intentionally described states of a creature con-
form to them that creatures count as having mental states. What it is correct to
113
say about someones psychology is answerable to these principles: the mental

is constituted and exhausted by interpretations governed by principles of
rationality and charity. In eect, they are synthetic a priori principles telling us
what minds are. However, a creatures possessing mental states is not depen-
dent on the existence of an actual interpreter who interprets the creature: it
suces that the creature exhibits behaviour that would sustain intentional
re-description in terms of a network of a itudes and actions that would make
sense of the creatures behaviour in rational terms. Wherever a creature could
be interpreted in accordance with the principles of charity and rationality the
creature counts as minded.
No account is given of how ordinary interpreters produce the interpretations
they do. But limits on what interpreters can know about an agent are made
explicit by what a fully informed theorist of interpretation could ascribe to the
agent on the basis of the non-intentional evidence about his behaviour and
physical history, and deployment of the intentional vocabulary under the
control of the a priori principles of rationality and charity. The theorist of inter-
pretation can provide a justification for the particular ascription of beliefs,
desires and intentions that an ordinary interpreter goes in for.
Linguistic and Non-Linguistic Interpretation:

Belief and Meaning
When interpreting other people, much of our information about what they
think comes from what people say. But to find out what they are saying we have
to know what their words mean. Even in everyday speech we can still wonder
whether other people are using words in the same way we are. So to confirm
that others mean the same thing as us by their use of words, or to adjust for the
dierence, we have to resort to interpretation. Interpreting speech is part of
interpreting a persons overall behaviour. Thus, in addition to ascribing beliefs,
desires and intentions to an individual we also have to assign meanings to her
words, and ensure the best holistic fit with her linguistic and non-linguistic
behaviour. In this way the interpretative principles are also at work in justifying
our ascription of meanings to peoples words. But if we have to know what
someone means to know what they believe and cannot know what they mean
without first interpreting them ascribing them beliefs and desires we go
round in a circle. We must either solve for two unknowns simultaneously or
find a way to break into the circle. Davidson proposes to do the la er by focus-
ing on the case of a speaker A holding a sentence S true under certain circum-
stances. We can know the speaker holds S true without knowing what S means,
and we can see the holding true as depending on what the speaker means by S
and what she believes to be the case. Now if we charitably interpret a speaker
114
as having a true belief about the prevailing circumstances (believing what we

do), we can take those circumstances as the truth conditions for sentence S
and thus know what someone believes and means. When we have interpreted
enough sentences we will see them as having parts in common, and by assign-
ing meaning (truth-aecting properties) to their parts, we can figure out what
new sentences composed of those parts mean and under what conditions they
are true. This will enable us to both interpret and test our interpretations with
respect to the further things people say. In this way, the principle of charity
enables us to break into the hermeneutical circle. By knowing the conditions
under which the speaker holds sentence S holds true, the interpreter can reason-
ably assume these are the conditions in the speakers language for that sentence
S to be true (its truth conditions).
At times it will be necessary to assume the interpreted subject has false
beliefs, for there are occasions on which a speaker asserts a sentence that we
interpret as being true under quite dierent circumstances. So we must either
revise the ascription of truth conditions to the sentence or assume that the
speaker believes falsely that those conditions do in fact obtain. Belief and
meaning are thus inextricable for Davidson: if people dont assent to a sentence
we expect them to, we can either interpret them as believing what we do and
adjusting the meanings we assign to their words, or we interpret them as
meaning what we do and interpret their beliefs uncharitably. Either way such
interpretations must bring the meanings, a itudes and actions a ributed to a
person into a rationally coherent pa ern that makes sense of the totality of
their linguistic and non-linguistic behaviour.
Davidsons position satisfies the rational and causal requirements of a
satisfactory account of folk psychology, but can it also leave room for self-
knowledge?
An Essentially Third-Person Epistemology of Mind
This is an essentially third-personal or a ributionist account of the mental.

Once we have assured ourselves that the principles of interpretation governing
our ascriptions of a itudes have been observed there is nothing more to say
about someones intentional states. The beliefs a person has are beliefs they
can be ascribed as having by a fully informed interpreter. Beliefs and other
propositional a itudes are exhausted by the criteria we have for a ributing
them to one another. As Davidson puts it, beliefs, desires, hopes and fears are
just those states whose contents can be discovered in well-known ways.

If other people or creatures are in states not discoverable by these methods,
it can be, not because the methods fail us, but because these states are not
115
correctly called states of mind they are not beliefs, desires, wishes or
intentions. (Davidson, 1986b, p. 160)
In commenting on Davidsons third-personal view of the mental, Michael Root

writes:
Other minds, on Davidsons view, are what we get when we interpret the
behaviour of others. Bodies are what we have before we interpret their
behaviour. (Root, 1986, p. 294)
Is there room on such an account for a first-person point of view? The

idea that the mind of a person depends on what it is correct for an interpreter
to ascribe to that person risks losing sight of the first-person point of view all
together. Does the mind exist only in the eye of the interpreter? Not exactly.
The best interpretation we can give is one that portrays the beliefs and desires
that make sense of that persons own way of seeing the world, that reflect her
view of how things are and that explain her reasons for acting, and in this sense,
arguably, we capture that persons own point of view. What would it be to see
her as acting in accordance with our view of her reasons, leaving out how she
saw things: such an interpretation would get things about her wrong.
But what of the subjects own knowledge of her mental states? She doesnt
have to interpret her own mind to know what she is thinking. Subjects know
their own minds without interpretation. Davidson uses this point to explain
an important feature of the first-person perspective: namely our first-person
authority about our current mental states. First-person authority is the author-
ity a subject enjoys in her judgements about her own mental states. What sort
of authority is it? Well, when a subject takes herself to have a particular belief or
desire she is typically right. Her knowledge of what she is currently thinking is
eortless and groundless. The subject doesnt base her judgement on anything
else: what she says about what she is currently thinking is typically correct and
needs no further justification. Others are not similarly authoritative about what
another person is thinking. There is no presumption that when they claim
someone is in such and such mental states they are usually right. Notice, how-
ever, that a subjects authority about what she is currently thinking or desiring
does not amount to infallibility: there are occasions on which we are entitled
to overrule what the subject says about herself; she may be self-deceived or
simply un-self-knowing.
What confers first-person authority on a subjects judgements about her
own mental states? For Davidson, the advantage comes from an evidential
asymmetry between the grounds one has for ascribing thoughts to oneself and
those one has for ascribing thoughts to others: unlike me, others cannot simply
pronounce about my mental states without grounds; they rely on observation
116
and evidence to ground their claims. So in ascribing thoughts to you I have to

rely on what you say and do in order to figure out what you think. In my own
case when I say what I think I dont have to first work out what my words
mean: I just know. I take one less step and am therefore at an epistemological
advantage. On this view I know what I think because I know what I mean when
I say what I think. I literally speak my mind (see Davidson, 1984a and 1987a).
But such an account of how I know my own mind does nothing to explain
how I know what I mean without interpreting my words. It is here that
Davidson insists that I cannot but know what my words mean.3 This may be
true, but it doesnt tell us anything about the nature of such knowledge or
what makes it available to the subject (see Smith, 1998).
Davidson against Scientic Psychology
Notice that on this a priori defence of common sense psychology li le or nothing

can be said about the relation between the intentional level of description and
the cognitive or neural levels posited by research in empirical psychology. The
relation is one of imposition of intentional descriptions on the levels below:
the levels are levels of description, not levels of organisation in the organism.
Nothing discovered scientifically at the levels below can have anything to con-
tribute to our understanding of the mental given that rationality and charity are
constitutive and exhaustive of the mental terms we use to identify minds.
Davidson does not eschew all empirical data, but he insists that it is a ma er
of purely a priori reflection to determine which empirical details bear on the
nature of mind. Which creatures have propositional a itudes? According to
Davidson, The question is not empirical: the question is what sort of empirical
evidence is relevant to deciding when a creature has propositional a itudes
(Davidson, 1982, p. 317). Put this way, there is still room for an a priori, philo-
sophical dispute about the type of empirical evidence that bears on the correct-
ness of psychological ascription. However, this view forces us to treat psychology
not as a science but as part of philosophy. In eect, Davidson leaves no room
for scientific psychology: as far as levels of description of reality go there is
just physics and the folk. Claims at both of these levels give us true descriptions
of reality.
Notice that a realist who maintains an a priori defence of belief-desire
psychology may simply refuse the challenge to substantiate its constructs at
any lower level. The claims of commonsense psychology, Davidson will say,
are answerable to no other criteria than their own: what makes any particular
folk psychological explanation we give of someones behaviour true is a ma er
of whether we have applied those criteria correctly, not an objective internal
ma er about what states they are in.
117
Let us turn now to a position that respects our interpretations of one

another in terms of beliefs and desires but which makes room for cognitive
psychology.
Dennett on the Intentional and Other Stances
The intentional stance
For Denne , beliefs and desires go together with the idea of an intentional
system a system whose behaviour can be predicated by ascribing it beliefs
and desires and assuming it will act rationally. To treat a creature as if it were
a rational agent is to adopt an intentional strategy, to take up the intentional
stance towards it. According to Denne , this means we ascribe it the beliefs
it ought to have given its place in the world and its progress through it.
Likewise, we ascribe it the desires it ought to have given its place in the world
and its purposes. We then predict what it will do given the a ributions
we make to it and the assumption that it is rational. (When we dont succeed
we may have to modify particular a ributions, or give up the assumption of
rationality.)
Taking up the intentional stance is a strategy to predict how the creature
(or system) will behave. But note that to predict is not necessarily to explain.
Denne s analogy is the chess-playing computer. We predict its moves by
assuming it has certain beliefs and goals it wants to get its queen out early;
it thinks Im weak on the le flank and treating it as a rational opponent. But
in some sense, it doesnt really have beliefs and desires. This is just a heuristic,
or useful assumption that helps us to predict the computers moves so we can
compete against it. If we want to know what actually accounts for and explains
its behaviour, we have to drop down to what Denne calls the design level by
adopting the design stance.
The design stance
At this level predictions are in line with the functioning components of the parts
of the system and what they are designed to do. For example, the chess-playing
computers workings are predictable by programmer on the assumption that
everything is functioning properly and working as it should do at the physical
level. The program may not be all we need to know in order to know the beha-
viour of the machine. It may not be working as it was designed to do because
physical components may have failed. This requires us to turn to the next
level down by adopting the physical stance towards the system.
118
The physical stance
To know about the workings of the system at the physical level is to adopt
the physical stance towards its workings. At this level we expect to achieve
physical law-like predications. For Denne , the physical and the design stance
(descriptions of the functional organisation and mechanical working of the
system) oer not only predictions but also genuine explanations of the beha-
viour of the system. However, when we adopt the intentional stance, we are
only involved in predication, not explanation.
Dennetts Original Position
Just as its not literally true to say the chess playing computer has beliefs and
desires that cause it to act, so its not literally true of us either. We simply ascribe
beliefs and desires to one another and assume people will act rationally. In
this way, we succeed in making fairly reliable predictions about one anothers
behaviour. But at the design level, the level of our inner organisation studied
by cognitive psychologists, there may be nothing like beliefs and desires
causing us to act. Belief-desire talk does not describe our innards. Denne is
a realist about our internal functional organization but a non-factualist, or
instrumentalist about beliefs and desires. Beliefs and desires are a ributed to
us by others to make sense of our behaviour, but they dont feature in our
inner workings, and they dont causally explain our behaviour. To do that
we need a scientific psychology that addresses the functional architecture of
the mind, and at that sub-personal level (another of Denne s distinctions about
systems smaller than persons which lack the properties assigned to persons),
we will have no use for belief-desire talk.
The trouble with this view, as Brian Loar put it in the quote above, is that
our continuing to use the belief-desire framework to systematize behaviour
. . . [would then] have the air of fictionalizing and contrivance. But Denne
wants to reject this portrayal of his view. He wants to maintain that he does, in
some sense, still believe in the reality of beliefs and desires. But in what sense?
Dennetts Revised View
Like Davidson, Denne is an a ributionist about the mind; they both hold
an essentially third-personal view of belief-desire psychology. But their views
dier. Davidson is a realist about beliefs and desires. He thinks that belief-
desire psychology gives us genuine explanations but not predictions of human
agency. Denne , on the other hand thinks that belief-desire psychology gives us
119
predication but not explanation. The dierence between these views hinges on
the rival conceptions of rationality. According to Davidson, rationality is uncod-
ifiable: there are no psychological or psycho-physical laws. Rational explana-
tion is always post hoc and partial and what counts as rational is just what
makes sense to us as rational creatures. By contrast, Denne thinks rationality
is law-like although rationality is at best an approximation of our behaviour.
The real explanations are found at the levels below.
For Denne the laws of folk psychology are at best rough and ready approxi-
mations (or idealizations) that help us to predict and interact with others. Hence:
Folk psychology is best seen not as a sketch of internal processes, but as an idea-
lized, abstract, instrumentalistic calculus of prediction. (Denne , 1987a, p. 48)
Beliefs and desires are calculus-bound entities, or logical constructs. So

although we find it useful to a ribute such intentional properties to thinkers,
they correspond to no real joints in nature. At best they are pa erns in the
behaviour of creatures perceived by creatures like us, creatures who engage
in interpretative practices, ascribing beliefs and desires to one another. This
may lead us to think that there really are no such things as beliefs according
to Denne . But this is too quick. On the revised view, he stresses that
any object whatever its innards that is reliably and voluminously

predicatable from the [intentional] stance is, in the fullest sense of the word,
a believer. (Denne , 1987a, p. 15)
There is no science of intentional psychology; there is no scientifically

respectable property, which all thinkers who are ascribed a given intentional
property must share. There will still be room, however, for scientific psycho-
logy to explain the cognitive sub-systems of creatures that give rise to the
behaviour we characterize, heuristically, in those intentional terms. But the
very disparate nature of the actual inner causes of behaviour in dierent
creatures undermines the explanatory pretentions of folk psychology. So
Denne s moral is that
folk-psychology is idealized in that it produces its predictions by calculating

in a normative system. It is abstract in that the beliefs and desires it a ributes
need not be presumed to be intervening states of an internal behaviour-
causing system. (Denne , 1987a, p. 52)
However, this diagnosis is not oered as a reason to replace or dispense with

the categories of intentional psychology; it is simply that we mustnt project
120
details from the intentional level of folk psychology onto the levels of organisa-
tion below. But we are still allowed to ask the question:
Exactly what feature must we share for [a given belief ascription] to be true
of us? More generally . . . what must be in common between things truly
[italics mine] ascribed an intentional predicate such as wants to visit China
or expects noodles for supper? (Denne , 1987a, p. 43)
Denne s answer is
a shared property that is visible, as it were, from one very limited point of
view: the point of view of folk psychology. Ordinary folk psychologists have
no diculty imputing such useful but elusive commonalities to people. If
they then insist that in doing so they are postulating a similarly structured
object in the head, this is a gratuitous bit of misplaced concreteness, a
regre able lapse in ideology. (Denne , 1987a, p. 55)
The point of view of folk psychology is the perspective from which we

make use of the intentional strategy, ascribing creatures (and computers) beliefs
and desires under the idealising assumption that they are rational agents. The
assumption of rationality, however, is only an ideal; the myth of our rational
agenthood structures and organizes our a ributions of belief and desire
(Denne , 1987a, p. 43).
A belief is just a useful but elusive commonality; but either these intentional
ascriptions pick out real properties of thinkers or we must succumb to the air
of fictionalising and contrivance. For there are just two kinds of answers to the
original question of what thinkers who are truly ascribed the same intentional
property (like believing or desiring something) must have in common. The first
answer says they are disposed to do and say such and such, and to judge so
and so, an answer that requires further a ributions. The second kind of answer
cites some underlying micro-property or functional property that they all
share. The first type of answer belongs to commonsense psychology and the
second to scientific psychology. Denne s mistake is to have stranded himself
between both, denying the strict literalness of commonsense talk and failing to
revise or vindicate it in scientific terms. This leaves an embarrassing gap when
it comes to explaining why many predications from the intentional stance are
so reliable. But this is the very heart of the ma er. What do all thinkers who
share a given thought have in common? For Denne , the new answer is that
beliefs and desires are real paerns in behaviour, but these are real pa erns
discernible from (and only from) the intentional stance (Denne , 1991b)
Minds are pa erns seen by (and only by) other minds.
121
Is this a satisfactory realism about beliefs and desires, a satisfactory account

of folk psychology? It acknowledges a rational structure of beliefs and desires.
However the believes and desires that are our reasons for action cannot be
the causes of it. For beliefs and desires are real pa erns in behaviour and it is
hard to see how pa erns exhibited in behaviour can also be the causes of that
behaviour. Denne s revised view has no satisfactory way to accommodate the
causal constraint.4
We now turn to a strongly realist view of the mental that takes seriously
the causal powers of beliefs and desires by rejecting the interpretational view of
folk psychology in favour of scientific account of beliefs and desires as internal
behaviour-causing states.
Fodors Intentional Realism
Taking up the challenge where Denne s view gives out, Jerry Fodor asks
what all thinkers who share a given thought have in common if its not their
neurophysiology:
How much (and what kinds of similarity between thinkers does the
intentional identity of their thoughts require? This is, notice, a question one
had be er be able to answer if there is going to be a scientifically interesting
propositional a itude psychology. (Fodor, 1986a, p. 9)
Fodor advocates the view that scientific psychology will vindicate common-
sense psychology. The problem is to understand how a science of the mental
can explain what we do without replacing or reducing our common sense
concepts.
Fodor aims to steer between these two dangers reduction (not likely) or
replacement (not palatable) by insisting that the causal laws of belief-desire
psychology that capture intentional generalizations about individuals have
to be explained by the computational and syntactic laws that implement them.
The dierence between the intentional and computational laws is not a ma er
of levels of description but levels of organisation within a creature. The compu-
tational laws govern the mechanisms that mediate the connections between
intentional causes and their behavioural eects. They are many and varied, so
no strict reduction of the intentional laws is possible. However, the lack of
reduction doesnt loosen all connection between the levels, for although there
are intentional causal laws that explain why we think and act as we do, we
are always entitled to ask, for any such law, how do those causes bring about
those eects in that individual.
122
The special sciences
This strategy is a general feature of the special sciences5, according to Fodor.

For any special science law L which says that Fs cause Gs we can always ask
why they do, and so long as the law in question is not a basic law of physics,
there will always be some further explanation of it in terms of the properties
of F-instantiations that render them sucient to bring about something with
the properties of G-instantiations. In the case of intentional psychology, the
mediating mechanisms that ensure that states with certain intentional contents
bring about certain behavioural eects relate to mental representations and
work according to computational operations defined over them. The reason
that law-like explanations in intentional psychology hold is ultimately a ma er
of underlying computational laws, though we can expect dierent laws (and
dierent mechanisms) to be operative in dierent thinkers. All that is required
is that there be some computational account of the mediating mechanisms that
explain why those intentional states engender those behavioural eects in the
thinker. So a law of intentional psychology applies to an individual thinker
in virtue of another law (a computational law) that implements the higher
law, and therefore shows why those law-like connections are sustained in the
thinker.
Unlike Denne , Fodor takes the laws of intentional psychology that govern
the relations of propositional a itudes to one another and to behaviour to
be true of us in virtue of our in-head states. Propositional a itudes like beliefs
and desires are inner states with semantic and causal powers. For Fodor,
psychological a itudes like believing and desiring are computational relations
to mental representations with propositional contents. The computational laws
that implement the laws of intentional psychology govern the causal connec-
tions among mental representations, keeping the causal relations in step with
relations of content between those mental representations. Hence, when mental
representation A entails mental representation B, the computational laws
should ensure that a tokening of A brings about a tokening of B, thereby show-
ing why law-like connections between psychological a itudes involving A
and psychological a itudes involving B are sustained in creatures like us. To
sustain a psychological law one has to have computationally organized innards
that mediate the connections between mental states that respect the relations
of content between those states.
Fodor wants his story about folk psychology to be based on three key
assumptions:
(1)The laws of folk psychology are intentional laws that cite contents.
(2)The content (semantics) of intentional states is informational and external.
123
(3) Laws of intentional psychology are implemented by computational

laws (mechanisms).
Are they compatible? First, it is a consequence of Fodors implementation

story that semantics doesnt cross implementation boundaries. This is because
the computational laws of mental processing that implement the higher level
intentional laws are purely syntactic. Although to preserve what needs to be
explained and to secure the widest application to thinkers, we have to frame
our generalizations in content-using terms at the intentional level (2).
What do all thinkers who share a particular belief have to have in common?
They are not required to be functionally identical: to make just the same infer-
ences or to have just the same accompanying states. This is extremely unlikely
in all but molecule for molecule duplicates (Twin-Earth twins). So what else
do they share? Pre y much they will be subsumed by the same intentional
psychological laws (1) (i.e. they all satisfy the same causal regularities picked
out by an intentional covering law). But what makes this true? Fodors answer
will be dierent for dierent thinkers since the vast disjunction of intervening
states of mind and cognitive computational mechanisms are, for him, typically
quite heterogeneous. He can agree with Denne that there is no reason to sup-
pose possessors of particular beliefs must be ultimately in some structurally
similar internal condition (Denne , 1987a, p. 55, italics mine) narrowly construed;
though Fodor has independent reasons for thinking they must each be in some
internally structured condition. This is the Fodors commitment to mental rep-
resentations whose syntactic structures follow the conceptual contours of the
content of those representations: the language of thought hypothesis.
But his appeal to a language of thought doesnt require thinkers to share
type identical mental representations; it is enough that each thinker have
some syntactically structured vehicle to play its content-bearing role. The com-
monalities among thinkers are to be found not at the level of computational
psychology, but rather at the higher intentional level. All that such thinkers
share is the property of satisfying the same laws of intentional psychology; what it
is to satisfy this property is explained dierently for dierent thinkers. Thus
Fodors argument for a content-using psychology rests on the claim that inten-
tional psychological laws remain indispensable in securing generalizations
across individuals who may have no physical or computational states in com-
mon. Many dierent internal mechanisms serve to implement the same higher
level laws, but all that unifies these diverse creatures is the intentional general-
izations the intentional causal laws they fall under. So the scientifically
interesting commonalities for psychology are not to be found at the level of
computational psychology, the level addressed by scientific psychology, but
the highest and relatively observational level of description at which mental
states are represented as having intentional content: the level at which we
124
predict behaviour by finding generalizations across individuals such as when

they have particular beliefs and particular desires, they will do such and such,
ceteris paribus. Folk psychology is to be vindicated by science by simply treat-
ing the laws of intentional psychology as (special) scientific laws. It is only in
so far as there are intentional causal laws that there can be a science of psychology
at all. But then if vindication is not achieved by recourse to the levels below,
in what sense is vindication really illuminating? Fodor will say that there has
to be some grounding of the higher level in the levels below. For the laws of
intentional psychology, treated as scientific laws, albeit ceteris paribus laws,
have to be implemented computationally. There are intentional laws because
there are mechanisms in nature that sustain (implement) those laws in crea-
tures like us. This is what explains how content-bearing states can have their
causal eects on behaviour. But scientific vindication comes at the higher level,
and not at the level of computational psychology with its Fodorian commitment
to computational operations defined over mental representations (sentences in
the language of thought). This is a distinct level of scientific theorizing, thinks
Fodor, with its own motivations.
We are le with an unexplained mystery: why should creatures with such
dierent innards all satisfy the same intentional laws? On Fodors ocial
story, our intentional generalizations quantify over a variety of computational
mechanisms that sustain law-like behaviour. So although a belief of a given
type will have to be a computational state of an agent, it is a mistake to think
that for any such belief there is one kind of computational state it has to be.
Thinkers with dierent internal organizations will still be answerable to the
same higher level psychological laws; which means that laws that capture those
generalizations cannot be formulated at the computational level. However,
Fodor does think that creatures with very dierent sets of beliefs and desires
will still be subject to the same intentional generalizations and so to the same
psychological laws. So commonalities preserve the claim that one is thereby
picking out genuine properties of individuals that generalize across cases.
Moreover, what they must have in common to fall under the same intentional
taxonomy is a ma er for science to decide, not a priori philosophy.
The urge to secure the scientific status of propositional a itude psychology
may lead to revision of what Fodor calls granny psychology: the everyday
use of belief and desire psychology which is pragmatic, context-sensitive and
vague. These aspects will not aect the causally explanatory laws of intentional
psychology, whose terms have been adapted to fit the generalizations we can
explain in computational terms, though these features may cause problems for
it. Two big problems are Frege cases where someone has two representations
with the same content that each have dierent behavioural eects (e.g.
Oedipus mother and Jocasta for Oedipus) and twin cases, where one
acts similarly with respect to mental representations with dierent contents
125
(e.g. Elm/Beech, or H20 and XYZ). Fodors answer is that both these cases mostly
dont occur, and when they do we can explain what has gone wrong. In Frege
cases, the dierent syntactic forms of the representations explains how they
aect computational mechanisms dierently: semantics and syntax come apart.
Semantics and syntax march in step in twin cases, but because of oddities of
the environment, distinct representations with dierent contents trigger the
same computational mechanisms, which is extremely rare.
There will be minor revisions in our everyday intentional idiom but not
massive revision, or else the laws of intentional psychology would not work
as well as they do, nor would the computational mechanisms operating in us
really be implementations of those laws and thus sustain intentional regulari-
ties. In other words, Fodor is holding out for the claim a hostage to empirical
fortune that there is a class of computational syntactic processes that makes up a
natural domain for psychological explanation. This will be the domain for creatures
like us who satisfy the same intentional generalizations we do, and the relation
between the computational underpinnings and the level of intentional states,
though contingent, will be reliable and explicable. Certainly, all believers and
desirers have this much in common: they must all have mental representations
which are syntactic vehicles for the contents of those states. This is the indepen-
dently motivated claim for the language of thought which does generalize across
thinkers however dierent their beliefs and desires. So regardless of their
specific mental states, and the particular psychological explanations we give
of their behaviour, they will all satisfy the same general explanatory pa erns
of acting in such a way as to do what they believe will secure for them their
fondest wish. And the claim that the internal mediating mechanisms of thinkers
can dier from individual to individual will still have to leave room for the
claim that they have enough overlap to ensure that the syntax of their mental
representations, whatever they are, mediate the same broad behavioural similarities
among thinkers:
The syntax of the mental representations which have the facts that P in their
causal histories [and so are about P] tend to overlap in ways that support
robust behavioural similarities among P-believers. (Fodor, 1994, p. 53, italics
and brackets mine.)
The trick is to find in each of us mentalese sentences, which are the causes
of the sorts of behavioural proclivities that the laws of psychology say that
P-believers share (Fodor, 1994, p. 54). For if no such overlapping set of causes
exists it is hard to see what relation of constraint between the levels Fodors
view imposes. And if it turned out to be none at all, we should have le Fodor
no be er o for a story about the relationship between folk psychology and
scientific psychology than Denne .
126
On the other hand, a Fodorian view which found commonalities, that is

to say overlap, among the heterogeneous class of syntactic vehicles of
thought would challenge Denne s picture by foisting on him a language
of thought among the otherwise diverse sub-systems of thinkers, a language of
thought story which Fodor argues is essential in order to have conceptually
organised thoughts the objects of intentional psychological a itudes. Claims
to intentional realism rest on a detailed working out of this issue.
Folk Psychology and the Full Extent of Mental Life
Even if we succeed in providing a philosophically satisfying characterization

of our common sense psychology in line with the key requirements given
above, several further questions arise:
(a) How do we succeed in ascribing beliefs and desires to one another?

(b) What role does consciousness play in a conception of the mind framed
in terms of belief-desire psychology?
(c) What role is there for including the emotions in belief-desire psychology?
The issue in (a) concerns not what makes the ascription of a belief to an indi-
vidual true, but rather how we go about ascribing beliefs and other a itudes in
the first place. What equips us to make such ascriptions? What provides us with
the means to a ribute mental states to other creatures? We shall briefly look
at two empirically dierent accounts of how we a ribute mental states to one
another, namely the theory-theory and the simulation accounts of our theory of mind.
So far as (b) is concerned, consciousness has had li le if anything to do with
the issues we have been discussing so far. It is a central aspect of the human
mind and a full and final theory of human psychology will have to account
for it, but it plays a much less prominent role in our folk psychology. We know
very li le about the conscious character or experiences of other people and yet
we are still adept at figuring out their reasons for action.
Finally, under (c), we surely need to accommodate the emotions as part of
folk psychology. A er all, in our everyday dealings with one another, we use a
wide repertoire of emotional terms such as anger, fear, humiliation, jealousy,
joy, happiness, envy, longing, and loneliness to help explain and predict one
anothers behaviour. And yet, when we turn to philosophical descriptions of com-
mon sense psychology, we have, until very recently, seen li le or no acknow-
ledgement of the role emotions play in our mental lives. Can a psychology that
rationalizes behaviour in terms of beliefs, desires and intentions leave room for
emotions to play more than a merely disruptive role as sand in the mecha-
nism? Let us end by looking at each of these issues in turn.
127
Our Mentalizing Abilities
How do we make ascriptions of propositional a itudes to one another?

Although Davidson and Denne have li le or nothing to say about this, Fodor
assumes that folk psychology is a theory that we (perhaps unconsciously)
deploy by having implicit, perhaps innate knowledge it. This view of our
mentalizing about others states of mind is called the theory-theory. Notice,
however, that theory-theory yields no insight into how we a ribute mental
states to ourselves. For we surely dont need to theorise about our own beha-
viour in order to know what we are currently thinking
An alternative view is that when we a ribute mental states to others we
start from the first-person perspective and use ourselves and knowledge
of our own psychology to simulate the mental states of others. This is called
simulation or simulation theory.
(O)I observe you doing F.

(S)I consider what mental states M I would be in were I to do F.
(Sim)I a ribute mental states M to you when you do F.
We judge what mental states would give us reason to behave in a certain

way and then use our own simulated mental states to understand others
mental states. To simulate what someone might do, or think, we imagine our-
selves in their shoes, see what we would do, think or feel in those imagined
circumstances, and use our own reactions to understand them. We take o-line
simulated inputs, note our reactions and read o others mental states from
our own. However, our mind-reading abilities also have to take into consider-
ation the ways other people dier from us if we are going to simulate them
accurately. We need to adjust for what we know about those people and what
they are like (e.g. fearful, hasty, excitable, etc.) But to do this dont we need to
appeal to a general theory of mind to make the relevant adjustments, and
doesnt this simply bring back in the theory-theory? For this reason, some have
suggested that we need a mixed system of simulation and theory-theory (see
Davies and Stone, 1995). Notice too, that while simulation gives central and
prior place to the first-person perspective, it oers no account of our knowledge
of our own minds; it merely takes it for granted.
Mirror-neurons and direct perception of others mental states
Some have objected that we do not need to theorize about others to know what
they are up to and to know their mental states. The mirror neuron system in
humans and monkeys activates the same populations of neurons in the pre-motor
128
cortex when performing an action or observing someone else performing the

same action. Such representation is before and below full mentalizing, and does
not provide to knowledge of someones intentions in acting but the mirror
system does code for the goal of a movement performed by self or other, and
thus enables a creature to recognize not just bodily movements but motor goals
in behaviour. The Davidsonian interpreter tells us nothing about how we go
from observing mere behaviour to choosing the intentional states to explain
that behaviour as action. The mirror system may explain, in part, how to bridge
that gap. Arguably, without this level of neural matching of anothers goal with
ones own potential for goal-directed movement, we could not even identify in
someones behaviour the targets for intentional explanation (see Gallese, 2006).
Others believe that instead of theorizing about others mental states we have
direct awareness of others minds. Following Wi genstein we might say that
we see someones mind in their face (see Gallagher, 2008). How we do this is
not clear. Are mental states visible, or just the expression of mental state?
Are the psychological states we see read o or read into the face and bodily
movements of others? Are psychological states something we see in the face
of others? Knowledge of other minds may not be inferential. We may see
expressions of mindedness as sorrow, or pensiveness, or desire. But seeing-as
when perceiving expressions of mind may not amount to direct unmediated
perception of minds. Besides, we also predict and explain what people do when
we are not perceiving them, so the mental states we are said to see must be
integrated into a rational structure of intentional states that explain more that
perceptual appearances, and so a further story about our conception of the
mind will take us beyond its perceptible aspects.
Consciousness
When we talk about folk psychological concepts like belief and desire
concepts with both first and third person applications all mention of con-
sciousness is missing. Davidson makes no use of it, nor of perceptual experience.
Experience does not give grounds for believing what we do. Instead, he argues
that the only thing that can justify a belief is another belief. And the only things
that can bring about beliefs in non-rational ways are the causal impacts the envi-
ronment has on us: sensation plays a crucial role in the causal process that
connects beliefs and the world (Davidson, The Myth of the Subjective in his
2001, p. 46). When we think about what we see we form beliefs about what
we are looking at. So Davidson pays no a ention to consciousness. No doubt
we have it but it is not playing an important epistemological or metaphysical
role in characterizing mental states. We know a lot about other peoples minds
while knowing li le or nothing about their conscious experiences.
129
Similarly, Fodor takes consciousness to be a mystery about which, he says,

we know precisely nothing. We can give naturalistic accounts of intentional
content and of the causal (computational) powers of such intentional states.
But we can say nothing so far about what consciousness is or about what gives
rise to it.
Consciousness is tackled by Denne alone, but his interpretative, third-
personal stance means he struggles to accommodate it. Heterophenomenology
replaces phenomenology as it features in the familiar picture of mental life.
In the familiar picture we have three stages: brain states give rise to experience
and following the appearance of experiences in consciousness, subjects report
on their experiences.
Denne claims we can do without the middle level and simply se le for
brain states giving rise to reports on supposedly pre-existing experiences. What
is more, Denne provides evidence against the existence of any fixed facts
about the middle level arguing that there is no fixed place or time where all
the sensory inputs come together in the brain to create a conscious experience.
Instead, subjects oer reports about experience and the content of these reports
will changing depending on timing: the precise point at which one asks the
subject to make a reflective judgement. We collect these dierent judgements,
not necessarily aware of how they may conflict. This leads Denne to his
multiple dra s model of consciousness (see Denne , 1991a).
Emotions
How do emotions, conceived as primitive feelings, fit into our everyday

(rational) psychology? Three strategies suggest themselves:
(a) Minimal accommodation: emotions play no real role in psychological

explanations but are just disruptive accompaniments (like sand in the
mechanism).
(b) Less minimal accommodation: reduce them to, or re-construct them
from existing categories in the mental inventory (as sensations or percep-
tions, beliefs, desires, a mixture of the two, or as uncompleted actions.
See Prinz, 2004 for a perceptual treatment.)
(c) Full accommodation: emotions belong in a separate category of sui gen-
eris mental states which play a proper role in everyday psychological
explanations. (Requires revision to folk psychology; see Elster, 2003.)
Option (a) has it that emotions play no real role in explaining others. The
virtue of this strategy is that it would cause minimal mutilation to existing
philosophical accounts of everyday or folk psychology. But it neglects the role
130
of emotions in explaining others and ourselves and giving us knowledge of the

world and others. The most conservative version of option (b) would invoke the
analogy between emotions and sensations. Emotions, like sensations, are both
objects of knowledge and sources of knowledge: they are narcissistic in telling
us something about ourselves and outward looking in telling us something
about world beyond; they have a characteristic felt quality; episodic (flash of
anger, pang of grief, feeling of euphoria); we undergo rather than undertake
them; they are an accompaniment to other (rational) mental states. (For more
on these options, see Smith, 2002b.)
Are emotions just sensations? No; emotions can be dispositional as well as
occurrent; they need not announce their presence in the mind in order to exist
(cf. exceptions such as humiliation, joy, elation). Besides, the a empt to identify
emotions and sensations is based on a false view of both that treats them as
private and ineable. But how would we identify such states in ourselves or
others? Being in a mental state is one thing: knowing which state one is in is
another. On what basis do subjects classify their privately felt sensations or
emotions? What generates the phenomenological taxonomy? And how is it
related to the taxonomy for others? There is a need once again to acknowledge
both the first and third personal dimensions of the mental, while recognising
the dierences. We need an account of the emotions that respects the tie between
the first and third person perspectives on the mind.
We need option (c): emotions are a separate class of mental states, though
they enjoy close connections with almost all other categories of mental state.
This makes them central and indispensable to folk psychology. (The folk always
knew this from common sense: common sense is one thing; the philosophers
view of common sense is another.) How do emotions relate to, or impact on,
other kinds of mental states? One view is to treat them as motivating states with
expressive or a itudinal content but no rational dimension. However a cognitiv-
ist view of the emotions sees them fi ing into larger pa erns of rationality in
the mind. Particular emotions are appropriate or inappropriate, they are propor-
tionate or not, they are to be trusted or doubted. Many of ones own assessments
of the emotions assume they are guides to action, judgements of a sort and
so can count as evaluations. This makes them centrally important to a well-
balanced mental outlook (see Wollheim, 2005). They can sponsor and be spon-
sored by other propositional a itude states: what I believe, what I want, and so
on. They can be mental dispositions but also occurrent phenomena. They play
long-term roles in our lives and short-term disruptive roles too. But perhaps we
should make a distinction here between ones deeper feelings and ones simply
being emotional. We still need a good account of how emotions fit into an
overall commonsense psychology by means of which we a empt to understand
ourselves and others as creatures with minds, as perhaps the only creatures who
can understand themselves and recognize the minds of others.
131
We have seen that folk psychological explanations appear to float free of

any appeal to the underlying cognitive states and processes that sustain our
capacities for perception, action and language. And yet a persons thoughts,
wants and wishes are not entirely independent of the perceptual, cognitive and
aective states they are in. So there is scope for both for a re-examination of
the way philosophers characterize folk psychology and of what people really
do appeal to or depend on in making ascription of mental states to others.
Additionally, we need not contemplate seriously the eliminativist option in
trying to reconcile the lived experience of our inner lives with the findings of
neurobiology and neuroscience, for these disciplines too must make sense of
the experience, thoughts and reflections of subjects at the personal level. They
may cast light on why our experience has the form and character it is, and
what happens when the underlying mechanisms break down, but it cannot
dispense with the level at which fixes on the mental states it is interested in
explaining. Thus a non-reductive cognitive neuroscience and a non-reduced
but rich and detailed folk psychology must eventually be aligned.
132
7 Internalism and Externalism
in Mind
Sarah Sawyer
Internalism and Externalism: The Basics
The individuation conditions of psychological properties is the topic of this

chapter.1 There are two opposing views: internalism and externalism. According
to the former also known as individualism psychological properties are
individualistically individuated, which is to say that their instantiation by an
individual depends entirely on the individuals intrinsic physical make-up.
According to the la er also known as anti-individualism psychological
properties are anti-individualistically individuated, which is to say that their
instantiation by an individual depends not only on the individuals intrinsic
physical make-up, but in addition on objective relations she bears to objective
properties in her environment. If a psychological property is individualistic,
then its associated content is said to be narrow; if it is anti-individualistic, then
its associated content is said to be broad.
Internalism, then, is the view that psychological properties supervene locally
on physical properties: no two individuals could dier psychologically without
diering in some intrinsic physical respect. Externalism rejects this local super-
venience thesis, maintaining in contrast that individuals could be exactly alike
with respect to their intrinsic physical properties and yet dier psychologically
if, for instance, they were related to relevantly dierent environments. Both
internalism and externalism are consistent with global psycho-physical super-
venience, the claim that no two worlds could dier psychologically without
diering physically. Local supervenience entails global supervenience (since
worlds can be construed as individuals), but not vice versa. The local super-
venience thesis is, therefore, the stronger claim, and the question of its truth
lies at the heart of the internalism/externalism debate.2
This paper provides an overview of the prevailing issues concerning the
debate. In the second section I distinguish various kinds of externalism and outline
some considerations in their favour. In the third section I discuss various forms
of internalism. In the fourth section I deal with metaphysical considerations
concerning naturalism and mental causation that have motivated internalism
133
and been thought to tell against externalism. In the fi h section I deal with epis-
temological considerations concerning the direct, non-empirical, authoritative
nature of self-knowledge that have been thought to tell against externalism.
I then conclude briefly in the sixth section.
Kinds of Externalism
Dierent considerations are thought to be relevant to the individuation con-

ditions of dierent kinds of psychological property. Consequently, one might
embrace externalism for certain kinds of psychological property but inter-
nalism for others. In this section I introduce a number of considerations that
favour externalism and catalogue various resulting kinds of externalism. The
kinds of externalism fall into two broad camps: externalism about concepts
expressed by predicative terms, which I will call predicative externalism,
and externalism about concepts expressed by singular terms, which I will call
singular externalism.
Predicative externalism
The most widely recognised consideration in favour of externalism emerges

from reflection on counterfactual scenarios in which a subjects intrinsic physi-
cal make-up is hypothesized to remain constant while the broader physical
environment in which she is embedded is hypothesized to dier.3 Such an envi-
ronmental dierence, it is urged, would be responsible for a dierence in the
subjects psychological states precisely because non-intentional causal relations
to objective properties in ones environment partly determine what one can
represent in thought. For example, a subject S, related in the right kind of non-
intentional way to silver, might have various thoughts involving the concept
silver, such as that silver jewellery is cheaper than gold jewellery. She may be
unable to distinguish (either practically or theoretically) various other actual
or possible metals from silver and may well acknowledge this. Nevertheless,
she possesses the concept silver because she is related in the right kind of
way to silver, and hence can think various things about silver by means of that
concept. Now suppose S had lived in dierent circumstances, circumstances in
which there was no silver for her to be related to either directly (via perception)
or indirectly (via other people). In such a situation S would be unable to think
about silver as such because there would be nothing to ground her possession
of the concept silver. How could she have acquired the concept? Suppose instead
that she had been related to one of the actual or possible metals that she is
unable to distinguish from silver. Call this metal twilver. In such circumstances,
134
S would have been related to twilver in just the same way as she is actually
related to silver, and hence it is plausibly the concept twilver that S would have
acquired. Consequently, where S thinks that silver jewellery is cheaper than
gold jewellery, counterfactual S thinks instead that twilver jewellery is cheaper
than gold jewellery. The dierence in representational content between the
belief S has and the belief S would have lies in the dierence between the objec-
tive properties to which she is related (silver) and would be related (twilver)
respectively. If these considerations are persuasive, then what determines the
representational content of a subjects beliefs goes beyond her intrinsic physical
make-up and her discriminative capacities (which are hypothesized to be
identical in the actual and the counterfactual scenarios alike) and depends in
addition on the objective properties to which she is related.4
This kind of thought experiment is taken by many to establish externalism
specifically with respect to natural kind concepts: concepts that carve nature at
its joints and feature in the true final set of scientific theories: concepts such as
(perhaps) quark, electron, hydrogen, water, heart, tiger, planet.5 However, reflection
on two further kinds of counterfactual scenario favours a more general exter-
nalism. The first draws upon the possibility of incomplete linguistic under-
standing6; the second draws upon the possibility of non-standard theory7.
I outline each in turn.
First suppose that a subject S has a wide range of ordinary beliefs a ribut-
able by means of the term game: she believes that some games are more fun
than others, that chess is a game, that children like party games, and so on.
However, she believes in addition (and mistakenly) that games must involve
at least two people, a point she would readily accept correction on if her
mistake were pointed out to her. Next consider a counterfactual scenario
in which her intrinsic physical make-up is hypothesized to remain constant
while her linguistic community is hypothesized to dier. In the counterfactual
scenario the term game is defined and standardly used to apply to games
that involve at least two people. Since game and game involving at least
two people mean dierent things, the word-form game in the counterfactual
scenario has a dierent meaning than it does in the actual situation. In like
fashion, the concept expressed by the word-form diers in the actual and the
counterfactual situations. In the actual situation the word-form game expresses
the concept game and includes in its extension games such as solitaire and
patience. In the counterfactual situation, in contrast, the word-form game
expresses a dierent concept that does not include in its extension either soli-
taire or patience. Consequently, S may in fact believe that pass the parcel is a
game, but had she been a member of the counterfactual linguistic community
she would have possessed a distinct concept, believing instead that pass the
parcel is a shgame, say. Once again, Ss intrinsic physical make-up and classifi-
catory capacities are identical in the actual and the counterfactual situations,
135
and the dierence in representational content between the belief she has and
the belief she would have lies beyond her intrinsic physical properties, this time
anchored by the classificatory practices of the wider linguistic community of
which she is and takes herself to be a part.
Behind this counterfactual scenario lies a certain understanding of linguistic
meaning according to which the conventional linguistic meaning of a term
(roughly its dictionary definition) is a complex abstraction from communal
rather than individual use. Linguistic meaning is determined by actual and
possible agreement among the most competent users, where the most compe-
tent users are those to whom others do and would defer if a question about an
individuals use were to arise. On this view, understanding the meaning of a
word is not an all-or-nothing thing, but rather comes in degrees. And it is the
possibility of understanding a word incompletely that allows for the dierence
in linguistic meaning in the actual and the counterfactual situations to be con-
sistent with there being no dierence in intrinsic physical make-up between
actual and counterfactual S. The dierence in linguistic meaning is then taken
to imply a dierence in concept expressed.
The final consideration that favours a general externalism trades on the fact
that even a subject with a full understanding of the linguistic meaning of a term
can doubt whether the dictionary definition that reflects that meaning correctly
characterizes the things referred to by that term. Thus suppose a subject S has a
full understanding of the term sofa and yet comes to wonder whether sofas
are really religious artefacts and not pieces of furniture made for si ing. Her
proposed theory about sofas is false, but this need not compromise either her
full understanding of the term sofa or her ability to think with the concept
sofa; rather, it reflects a strange view about the nature of sofas thought of as
such. Now hypothesize a counterfactual situation in which Ss false theory
is standard and true of a dierent yet superficially indistinguishable class of
entities (call them safos).8 The linguistic meaning of the term sofa in the
actual situation diers from the linguistic meaning of the term sofa in the
counterfactual situation even though the entities referred to are superficially
indistinguishable. This is because the actual linguistic community and the
counterfactual linguistic community have agreed upon dierent characteriza-
tions of the relevant entities. Moreover, the concept expressed by the term
diers in the two situations because the entities referred to dier: in the actual
situation they are sofas (pieces of furniture made for si ing), whereas in the
counterfactual situation they are safos (religious artefacts). Consequently, while
actual S believes that sofas are religious artefacts, counterfactual S believes that
safos are religious artefacts.
Behind this counterfactual scenario is a certain understanding of the dier-
ence between the linguistic meaning of a term and the concept expressed by
that term. The linguistic meaning of a term goes beyond individual use and is
136
grounded in communal use, as mentioned above. Communal use may well

change over time, and hence the linguistic meaning of a term may well change
over time. (Dictionaries are plausibly updated in part to reflect such changes
in linguistic meaning.) But the concept expressed by a term may well remain
unaltered even while the linguistic meaning of that term changes. This will
happen, for instance, when entities of a given kind are identified through per-
ception and then characterized. The concept will be anchored to the entities
through perception, whereas the linguistic meaning will reflect received views
about the entities, and this characterization may well need updating as investi-
gation proceeds and even while the concept remains unchanged. It is the fact
that we can be mistaken in our characterizations of the things we perceive that
allows for non-standard theory to be entertained, and this in turn grounds a
general form of externalism. In the sofa/safo case, of course, Ss theory would
prove false under empirical tests and hence would not lead to a change in
linguistic meaning. Cases where proposed theories are adopted, however,
would lead to corresponding changes in linguistic meaning. This is what allows
us to make sense of genuine theoretical disagreement about a class of entities
thought about by means of the same concept, and grounds constancy of refer-
ence through theory change.
Thus far I have distinguished two broad kinds of externalism: natural kind
externalism, based on noting subjects relations to natural kinds; and social
externalism, based on the possibility of incomplete linguistic understanding
and the possibility of theoretical doubt. Both are kinds of what I have called
predicative externalism since they concern concepts expressed by predicative
terms. Natural kind externalism has gained more support than social external-
ism, but so long as we take seriously, as we must, the thought that our concepts
concern a world about which we can be in error, there is reason to adopt a
general predicative externalism.9
Singular externalism
According to singular externalism, the representational content of a subjects

thoughts about particulars (singular thoughts) is individuated partly by the
particulars those thoughts concern. This is directly analogous to predicative
externalism according to which the representational content of a subjects
thoughts about properties is individuated partly by the properties those
thoughts concern.
There are two main kinds of singular externalism: externalism about thoughts
expressed by sentences containing demonstratives; and externalism about
thoughts expressed by sentences containing proper names. To take a demon-
strative example first, suppose that actual S is looking at a particular apple, A1,
137
while counterfactual S is looking at a dierent apple, A2. Suppose further that

S and counterfactual S u er the sentence That is nutritious. It is clear that
Ss u erance (and thought) concerns A1, whereas counterfactual Ss u erance
(and thought) concerns A2. This is so even if Ss intrinsic physical make-up is
identical in the two situations. Moreover, Ss u erance (and thought) is true if
and only if A1 is nutritious, while counterfactual Ss u erance (and thought) is
true if and only if A2 is nutritious. Crucially, according to singular externalism
this dierence in truth conditions is due to a dierence in representational
content.
Parallel remarks hold for externalism concerning thoughts expressed by
sentences containing proper names. Thus if S u ers the sentence Danny is
interesting, referring to Danny Alpha, with whom she is acquainted, and coun-
terfactual S u ers Danny is interesting, referring to Danny Beta, with whom
she is acquainted, their u erances and thoughts have dierent truth conditions,
and this is consistent with Ss intrinsic physical make-up being the same in
both the actual and the counterfactual situations. Again, according to singular
externalism this dierence in truth conditions is due to a dierence in represen-
tational content.
Singular externalism is upheld by a number of people in a number of dier-
ent ways. According to Gareth Evans and John McDowell, all thoughts are
composed of Fregean senses, but singular thoughts contain de re senses which
exist only if there is an object to which they refer. Evans and McDowell advo-
cate this kind of singular externalism for all thoughts about particulars, whether
the particulars are thought about by means of demonstratives or by means of
proper names.10 According to direct reference theorists, in contrast, the thought
expressed by a sentence containing a proper name contains not a de re sense of
the object named but the very object itself.11 Here again, the existence of the
thought depends upon the existence of the object thought about. This view is
typically not extended to demonstrative thought, although in principle it could
be. A variant of the direct reference theory that accommodates Fregean insights
(about dierent ways of thinking about an object) without countenancing de
re senses holds that the thought expressed by a sentence containing a proper
name contains the object named together with a mode of presentation of that
object, but the implication is the same: the existence of the thought depends
upon the existence of the object thought about. This view could also in principle
be extended to the demonstrative case. What makes all these views forms of
singular externalism is the common claim that the content of a singular thought
is object-dependent.
Considerations that bear on singular externalism thus far parallel consider-
ations that bear on predicative externalism, as noted at the outset. But the
question of the individuation conditions of singular thoughts introduces the
138
possibility of a distinction between a singular thought and its representational

content; and this distinction has no analogue in the predicative case. The
distinction opens up the possibility of accepting that the truth conditions of
singular thoughts are object-dependent while denying that this is in virtue of
a dierence in representational content. What results is a theory according to
which the representational content of a singular thought is preserved across
intrinsic physical duplicates but can be thought of (and hence true or false of)
dierent individuals on dierent occasions. To take the demonstrative example
above, on this view, S and counterfactual S both have a thought the representa-
tional content of which is given by the open sentence is nutritious. Actual S
thinks this of A1, whereas counterfactual S thinks this of A2. The dierence in
truth conditions between the thoughts is on this view due to a dierence in
contextual application rather than representational content. To accept such a
distinction between a thought and its content is to embrace a kind of two-factor
theory of singular thought according to which the object thought about is a
constituent of the thought but is not referred to by a conceptual constituent of
the thought. On such a view the object contributes to the truth conditions of
a thought concerning it but does not aect its representational content. Hence
the view is a form of singular internalism.12
This view has been popular in the demonstrative case, but has gained li le
support in the proper name case due to the dominance of direct reference
theories. However, if one were to accept singular internalism for the demon-
strative case and in addition think of singular uses of proper names as involv-
ing a demonstrative element, then one would naturally be led to embrace
singular internalism for the proper name case too. To take the second example
above, a singular use of a name such as Danny is to be understood as involving
a demonstrative element and hence as semantically equivalent to That Danny,
which can be used to refer to dierent Dannys on dierent occasions. On this
view, S and counterfactual S both have a thought the representational content
of which is given by an open sentence something like is a Danny and is inter-
esting. Actual S thinks this of Danny Alpha, whereas counterfactual S thinks
this of Danny Beta. The dierence in truth conditions between the thoughts is
again due to a dierence in contextual application rather than representational
content.13
Kinds of Internalism
I have already discussed singular internalism above to contrast and clarify

singular externalism. Consequently I will confine my discussion in this section
to versions of predicative internalism, of which there are four primary forms.
139
Two kinds of thorough-going internalism
The most straightforward way of being a predicative internalist is to reject

outright the interpretation of the counterfactual scenarios taken above to ground
externalism. An alternative, internalist interpretation would maintain instead
that psychological properties are necessarily preserved across intrinsic physical
duplicates precisely because they are and must be grounded in the discrimina-
tive capacities and transparent epistemic outlook of the individual.14 An indi-
viduals psychological make-up cannot outstrip what that individual can do
and how things seem to her, as it were. S and counterfactual S in each of the
scenarios have the same capacities to discriminate and classify things, and (in
some sense) have the same views about the things they encounter: there is noth-
ing that allows them to distinguish the actual from the counterfactual situation
in each case. Consequently, according to thorough-going internalism, there can
be no psychological dierence between them.
One way of upholding the view is to think of the relevant concepts as descrip-
tive, encapsulating the subjects beliefs (or theories) about the things referred
to. For instance, the concept both S and counterfactual S express by the term
silver might be shiny metal oen used to make jewellery and that needs to be polished
to be kept clean and . . . The concept they express by the term game might be kind
of activity undertaken for enjoyment, involving at least two people, involving rules in
accordance with which you can win or lose; and the concept they express by the
term sofa might be religious artefacts that look as if they may be sat upon but . . .
Note that if the original concepts (here thought of as descriptive) are to be
individualistically individuated, then the concepts used in the descriptions
must of course themselves be individualistically individuated. But this kind of
descriptivism (which many will view as independently problematic) is not
essential to the view. One could instead treat the concepts minimally.15 On this
view, the concept both S and counterfactual S express by the term silver is a
concept that has in its extension silver, twilver and everything else that S cannot
distinguish from them (as it does on the descriptive view). But in order to
express the concept we would need to introduce a new term such as shmilver.
Similarly, the concept they express by the term game is the concept shgame,
which has in its extension games that involve at least two people; and the con-
cept they express by the term sofa is the concept safo, which has in its extension
religious artefacts that look like sofas. The subjects discriminative capacities
and epistemic outlook here serve to individuate her concepts and hence deter-
mine the extensions of those concepts but are not taken up as descriptive elements
of the concepts themselves. On both the descriptive and the non-descriptive
versions of through-going internalism new terms need to be introduced into
our language in order to express with accuracy the concepts had by indi-
viduals whose beliefs dier from the norm (or, more generally, from our own),
140
as illustrated by the use of new terms in the examples just given. And on both
views, S possesses the same concepts as her counterfactual self in virtue of
having the same discriminative capacities and epistemic outlook on the world,
but she has dierent concepts from those in her linguistic community. This
stands in marked contrast to predicative externalism, according to which S has
dierent concepts from her counterfactual self but shares many concepts with
those in her linguistic community despite varying degrees of understanding
and competence which result in a wide variety of discriminative capacities and
epistemic outlooks across individuals within that community.
Two kinds of two-factor internalism
The third and fourth kinds of predicative internalism are more complicated.
They acknowledge that the counterfactual scenarios outlined establish that
S and counterfactual S have dierent thoughts in some sense, but aim none-
theless to retain a sense of content which is preserved across intrinsic physical
duplicates, in order to respect the internalist conception of sameness of epistemic
outlook. Both therefore maintain that a thought has a narrow and a broad
content and are thus kinds of two-factor theory.
According to the first of these, the internal component of a thought (its
narrow content) is a function that determines its external component (its broad,
truth-conditional content) given a context (an environment).16 Thus when S
and counterfactual S u er the sentence Silver jewellery is cheaper than gold
jewellery, the narrow content of their thoughts is the same, but the broad
content of their thoughts diers simply in virtue of their location in dierent
environments: Ss thought concerns silver, and is true if and only if silver
jewellery is cheaper than gold jewellery; whereas counterfactual Ss thought
concerns twilver, and is true if and only if twilver jewellery is cheaper than
gold jewellery.
There are similarities between this two-factor theory of predicative thought
and the two-factor theory of singular thought discussed in the Singular Exter-
nalism section above. On both views the only form of truth-conditional content
is broad. And yet on both views the thoughts of intrinsic physical duplicates
share a kind of content even though they have dierent truth conditions.
However, the similarity does not extend beyond the superficial level, and the
dierences are important. According to the two-factor theory of singular
thought, singular thoughts have contents that are intrinsically representational
independent of context, and can be applied to (or thought of) dierent indi-
viduals in dierent circumstances. The two factors involved in a singular
thought are first, a content (is nutritious, say), and second, (potentially) an
individual of whom the content is thought (a particular apple, for instance).
141
The content of a singular thought is not itself divided into a narrow and a broad
component. According to the two-factor theory of predicative thought now
under consideration, in contrast, the content of a predicative thought is itself
divided into a narrow component and a broad component. Crucially, the
narrow component is not representational: only the broad component is. The
narrow component is a function and can be understood only in terms of its
inputs and outputs: that is, only in terms of the broad, truth-conditional content
it produces once the individual is situated in a particular environment. This
puts pressure on the idea that the narrow component of a predicative thought
is properly conceived as a form of content at all.
The second kind of two-factor theory of predicative thought is also a racted
both by the externalist interpretation of the counterfactual scenarios and by the
internalist conception of sameness of epistemic outlook. However, it aims to
draw a distinction between broad and narrow content consistent with all con-
tent being representational in some sense. On this view, the broad content of a
subjects thought is determined, in line with externalist considerations, in part
by relations she bears to objective properties in her environment. The narrow
content of a thought, on the other hand, is individuated by the epistemic pos-
sibilities it allows and excludes.17 The underlying thought here is that intrinsic
physical duplicates are in the same epistemic position in the sense that they
cannot distinguish between the relevant actual and counterfactual situations
and that the narrow content of a thought encapsulates this fact. For example,
suppose that S and counterfactual S both u er the sentence Silver jewellery
is cheaper than gold jewellery. The thoughts they thereby express have dier-
ent broad, truth-conditional contents: one concerns silver whereas the other
concerns twilver. However, the thoughts are taken to have the same narrow
content because a purely qualitative description of a situation in which silver
jewellery is cheaper than gold jewellery is identical to a purely qualitative
description of a situation in which twilver jewellery is cheaper than gold
jewellery. Given the epistemic position of S and counterfactual S, both situa-
tions verify their thoughts and hence they share a narrow content. The success
of the position clearly depends on the possibility of describing situations in
purely qualitative terms terms not subject to externalist considerations. As
such, the position depends upon the truth of a restricted rather than a general
form of externalism. If all terms were subject to externalist considerations then
there would be no terms available to feature in the qualitative descriptions
required to ground this notion of narrow content. Moreover, there is a question
about whether it makes sense to think of a thought as having two forms of
representational content where only one of these is truth conditional.
I have discussed and argued against all four forms of internalism elsewhere
and will not repeat the arguments here.18 Instead I now turn to some of the
142
primary metaphysical and epistemological considerations that surround the

internalism/externalism debate.
Metaphysical Considerations
According to externalism, psychological properties do not supervene locally on

a subjects intrinsic physical make-up. This throws up two related metaphysical
concerns: first, how to retain a naturalistic theory of the mind; and second, how
to make sense of mental causation. The two concerns are intimately connected
and have provided much of the motivation for internalism.
Since the late 1950s and early 1960s, the question of how psychological prop-
erties relate to lower level properties and ultimately to properties of interest
to the physical sciences has dominated discussions in philosophy of mind.
The requirement that they must be related in some intimate and significant
way has been regarded as crucial to an account of the mind that is scientifically
and naturalistically respectable. Type-physicalism, according to which psycho-
logical properties are identical to physical properties, clearly satisfies the
requirement. However, the postulated identity of psychological properties with
physical properties comes under pressure from arguments to the eect that
psychological properties are multiply realizable that individuals in dierent
physical states could nonetheless be in the same mental state. Such arguments
have motivated forms of token-physicalism, according to which each token
mental state of an individual is identical to or realized by some physical state of
that individual, even if the psychological property of which it is an instance
is not identical to the physical property of which it is an instance. But the
dispute between type-physicalism (of various kinds, including behaviourism)
and token-physicalism (of various kinds, including functionalism) takes place
within a common theoretical framework: physicalism.
Externalism, in contrast, rules out all forms of physicalism.19 The minimal
claim of physicalism is that every token psychological state of an individual
is either identical to or realized by a token physical state of that individual.20
More specifically, physicalism is defined by its commitment to local psycho-
physical supervenience. This makes clear why it is inconsistent with external-
ism. As such, externalism has been thought to sever the psychological from the
physical, and hence to rule out a naturalistic theory of the mind. However,
although externalism is inconsistent with physicalism, it is consistent with
materialism the view that every entity is composed of physical ma er. This
is a weaker doctrine than physicalism, but is strong enough to secure a natural-
istically respectable theory of the mind. Since externalism is consistent with
materialism, it is consistent with naturalism.
143
There is, however, a related worry about anti-individualistically individuated

properties. In particular, it has been thought that only individualistic properties
can be causal properties: that causal powers must be intrinsic. If this is right,
then externalism is commi ed to the claim that psychological properties are
not causal properties, which undermines the intuitive and commonly held idea
that our beliefs and desires cause our actions.21 The worry here is that although
externalism is consistent with naturalism, its understanding of psychological
properties renders them insignificant because psychological properties thus
conceived would have no causal powers and hence make no dierence in the
world. Consequently, even if externalism is naturalistically respectable, it does
not yield an account of the mind that is scientifically respectable.22
However, the assumption that causal properties must be intrinsic is mis-
guided. Indeed, scientific practice demonstrates that many sciences study
pa erns of causation involving entities in their normal environment, and the
properties to which they appeal in causal explanations are individuated in a
way that presupposes such relations between entities and their environment.23
Thus astronomy studies the motions of the planets; geology studies land
masses on the surface of the Earth; physiology studies hearts or optic fibres in
the environment of a larger organism; psychology studies activity involving
intentional states in an environment about which those states carry informa-
tion; the social sciences study pa erns of activity among persons (Burge, 1989,
p. 317). Because such properties are individuated with reference to a normal
environment, the properties are anti-individualistic; and yet such properties
are also individuated with reference to their causal powers, and hence there is
no question that they are causal properties.
The view that emerges is a view according to which psychological properties
are causal and yet fail to supervene on lower level properties. It follows that
individuals who are classified as of the same kind from the perspective of one
science may be classified as of dierent kinds from the perspective of another.
Thus for example two individuals may be exactly similar from the perspective
of neuroscience but significantly dierent from the perspective of psychology
because they instantiate the same neurophysiological properties but dierent
psychological properties. This is the case for S and counterfactual S in each of
the scenarios described in the Predicative Externalism section. This allows us to
identify an error in the internalists reasoning. Internalists o en point out that
S and counterfactual S would exhibit the same behaviour non-intentionally
described (they would follow exactly similar trajectories through space, exhibit
the same speech pa erns, classify things in the same way, and so on), which
of course is true, but they go on to conclude that S and counterfactual S
instantiate the same psychological properties. However, the similarity in beha-
viour is to be explained by the fact that S and counterfactual S instantiate the
same neurophysiological properties and is consistent with their instantiating
144
dierent psychological properties. The former may well be individualistically

individuated even though the la er are not.
Externalism is inconsistent with physicalism. However, it is consistent both
with a naturalistic theory of the mind and with the claim that psychological
states are causally ecacious.24
Epistemological Considerations
Central to the internalism/externalism debate in the philosophy of mind has

been the question of whether externalism is consistent with the intuitive
claim that a subject knows what she is thinking in an epistemically privileged
way. There are two primary areas of concern. The first is the achievement
problem and the second is the consequence problem. I deal with each
in turn.25
The achievement problem
According to externalism, what concepts we possess and hence what thoughts

we can think depends on contingent, empirical relations we bear to objective
properties in our environment. The question then arises, how can we know our
thoughts in a direct, non-empirical, authoritative manner when those thoughts
depend on our relations to the environment? Imagine that S is periodically
switched from the actual situation (in which she is related to silver) to the
counterfactual situation (in which she is related to twilver). Suppose further
that a er each switch she stays long enough to acquire the concept appropriate
to the new environment. Under such a hypothesis S will at certain points in
time think that silver jewellery is cheaper than gold jewellery, and at other
points in time think that twilver jewellery is cheaper than gold jewellery. And
yet there would be no break in the continuity of Ss life because there would
be (by hypothesis) no discernible dierence between the two environments.
The changes in her environment would pass undetected, and so, crucially,
would the changes in her thoughts. On this basis it is argued that S does
not know what she thinks in a direct, non-empirical, authoritative manner.
Rather, S requires empirical knowledge of her environmental relations in
order to know what she thinks. There are various ways one might take the
argument: one might conclude that the mere possibility of such switches under-
mines the direct, non-empirical, authoritative nature of self-knowledge; or one
might conclude that the close epistemic possibility of such switches undermines
the direct, non-empirical, authoritative nature of self-knowledge; or one might
conclude that only actual switching undermines the direct, non-empirical,
145
authoritative nature of self-knowledge.26 However it is taken, externalists have

been broadly uniform in their response, which has two main strands.
First, it is pointed out that the concepts available at the second-order level of
thought (concepts employed to think about ones first-order thoughts) are deter-
mined (in part) by relations to the very same set of environmental conditions
that determine the concepts available at the first order level of thought (con-
cepts employed to think about the world). As such, S could not be in error about
her thoughts simply in virtue of the dependence of those thoughts on her envi-
ronmental relations. S could not, as it were, think she was thinking that silver
jewellery is cheaper than gold jewellery but really be thinking that twilver
jewellery is cheaper than gold jewellery. This kind of error would involve using
the concept twilver at the first-order level while simultaneously using the con-
cept silver at the second-order level. Rather, the same concept (whether it be
silver or twilver) would be employed at all levels of thought. Consequently, the
kind of threat envisaged is ill-conceived.27
Second, it is pointed out that the (partly environmental) conditions that indi-
viduate a thought are presupposed in the thinking of that thought but need not
themselves be known in order for it to be known that that is the thought one
is thinking. Perceptual knowledge presupposes that certain background condi-
tions obtain (that lighting conditions are reasonable, that one is not hallucinat-
ing, and so on), but such background conditions need not be established by the
subject before she can be said to know by looking that there is, for instance, an
apple on the table in front of her. Similarly, it may be that particular instances of
self-knowledge presuppose relations to objective properties in ones environ-
ment, but a subject need not know that such relations obtain in order to know
what she thinks.28 Indeed, if one had to know the individuating conditions of a
thought in order to know one was thinking it, then neither internalism nor
externalism would be consistent with the direct, non-empirical, authoritative
nature of self-knowledge: we do not have such direct, non-empirical, authorita-
tive knowledge about our environmental relations (as would be required if
externalism were true); but we do not have such direct, non-empirical, authori-
tative knowledge about our intrinsic physical make-up either (as would be
required if internalism were true). This merely shows that the demand for
such knowledge of individuating conditions is irrelevant to questions about
self-knowledge.29
But the achievement problem surfaces again in the guise of the argument
from memory.30 Suppose S thinks a second-order thought at t1: I think silver
is shiny. Suppose she is then switched from the actual to the counterfactual
environment where she remains long enough to acquire the concept twilver.
At t2 (some point later), when reflecting on what she thought at t1, she will
think with concepts relevant to the counterfactual environment and hence
think, it is argued: I thought twilver was shiny. The content of her thought at t2 is
146
false, since it does not capture the content of her thought at t1. Consequently, S
does not know at t2 what she was thinking at t1. This is taken to undermine
self-knowledge of externally individuated past thoughts. Moreover, it is argued,
self-knowledge of externally individuated current thoughts is also undermined.
A er all, if S does not know at t2 what she was thinking at t1, and there is no
reason to think she has forgo en anything in the interim, there is reason to
think she never knew at t1 what she was then thinking.
As with the initial argument, there are various ways the argument might be
taken depending on whether one thinks mere possibility, close possibility or
actuality the relevant epistemic factor. But here two dierent responses have
emerged. According to the first, the argument shows that externalism does
undermine the direct, non-empirical, authoritative nature of ones knowledge
of ones past thoughts, but it does not show that externalism undermines the
direct, non-empirical, authoritative nature of ones knowledge of ones current
thoughts.31 This can be made plausible, for instance, by acknowledging a new
way in which one might be said to forget something (namely, by being switched
between subjectively indistinguishable environments), or alternatively, by
maintaining that forge ing is not the only way in which one might fail to
know at t2 what one knew at t1 (since one might instead be switched between
subjectively indistinguishable environments). The general moral here is that
although one may need to rely on empirical considerations to the eect that the
environment has remained broadly stable in order to know what one thought
in the past, the non-empirical warrant for knowledge of ones current thoughts
is not thereby undermined. The disruption, as it were, is confined to knowledge
of past thoughts.
According to the second line of response, the argument does not show that
externalism undermines the direct, non-empirical, authoritative nature of ones
knowledge either of ones current or of ones past thoughts.32 This second line
of response is bolder and can be made plausible, for instance, by showing
how the content of past thoughts can be preserved in memory even across
undetectable switches between diering environments. This has been advo-
cated by Burge, who distinguishes substantive event memory, which refers
back to earlier events, from preservative memory, the function of which is to
hold the contents of thought in place so the subject can determine logical and
epistemic relations between them. Preservative memory does not refer back to
earlier thinkings, but rather holds contents in place for the purposes of, for
instance, critical reasoning.33 While substantive event memory might be under-
mined by externalism, preservative memory will not be, precisely because
preservative memories do not refer back to independent events.34
It is important to note that externalists have not oered a theory of self-
knowledge in response to the achievement problem in either of its guises.
Rather, they have tried to show how the arguments are misguided. Two things
147
are clear. First, there is no widely accepted theory of self-knowledge, either

internalist or externalist, and work remains to be done here. But, second, the
achievement problem does not bring to light any specific diculties for the
externalist. Rather, and perhaps unsurprisingly, externalist theories of self-
knowledge and of memory will look rather dierent from internalist ones.
The consequence problem
The consequence problem emerges once an answer to the achievement problem

has been assumed. The problem arises when one combines a non-empirical
warrant for the claim that one is thinking a particular thought, with a non-
empirical warrant for the claim that thoughts of that kind depend on the envi-
ronments being a particular way, to yield, surprisingly, a non-empirical warrant
for a claim about the nature of ones environment. For example, S might reason
as follows: (P1) I think silver is shiny; (P2) if I think silver is shiny then I must
be related to silver; therefore (C) I am related to silver. (P1) is an instance of self-
knowledge and hence taken to be non-empirically warranted; (P2) is arrived
at through philosophical theorising and hence taken to be non-empirically
warranted; but then it seems that (C) can be warranted non-empirically, which
is generally thought to be wildly implausible. Given the implausibility of
having a non-empirical warrant for claims such as (C), it is argued, externalism
is inconsistent with the direct, non-empirical, authoritative nature of self-
knowledge.35
The consequence problem has generated a vast amount of literature and
three primary externalist responses have emerged. The first concerns the exter-
nalist conditional that connects thoughts with the environmental conditions
they presuppose. According to this position, conditionals such as (P2) are false
because they commit the externalist to a stronger thesis than is either plausible
or established by the counterfactual scenarios that support it. And a true exter-
nalist conditional, which stated a genuine dependency relation between the
thinking of a thought and environmental conditions necessary for it, would,
according to this view, be so weak that non-empirical knowledge of its con-
sequent would not be implausible at all. For instance, it might state that in
order for S to think that silver is shiny, S would have to be related to some basic
kinds of things (but not necessarily to silver).36
The second response abstracts away from questions about the content of the
externalist conditional and focuses instead on what is wrong with Ss reasoning
even if her reasoning is sound. According to this strategy, arguments such as
the one S reasons through are epistemically defective in roughly the same
way that a question-begging argument is epistemically defective: neither a
148
question-begging argument nor an externalist argument of this kind will be

persuasive to a subject who doubts the conclusion. According to this position,
the non-empirical warrant available at (P1) fails to transmit across the known
conditional (P2) to provide a non-empirical warrant for (C), because, roughly,
the non-empirical warrant for (P1) is only available on the assumption that the
environmental conditions stated in (C) obtain.37
The final position argues, in contrast, that there is nothing epistemically
wrong with Ss reasoning. On this view, self-a ributions are warranted non-
empirically and without the need of a prior warrant for the claim that the
environmental conditions that help to individuate the thoughts obtain. More-
over, in the absence of a doubt about whether the environmental conditions
obtain, the non-empirical warrant for (P1) remains undefeated and can legiti-
mately transmit via (P2) to (C). The response has met with incredulity, but is,
I think, reasonable given two facts. First, a claim is equally open to doubt
whether it is warranted empirically or non-empirically. This is important
because it separates the question of whether there is anything epistemically
defective with Ss reasoning from the question of whether Ss conclusion can
be deployed in a straightforward argument against the sceptic. Ss reasoning
is epistemically legitimate, but does not refute external world scepticism.
Second, (although I have not argued this here) externally individuated thoughts
depend not only upon the existence of relations between the thinking subject
and objective properties in her environment, but on the subjects knowledge
of such relations. Consequently, while externalist arguments of this kind can
provide a subject with non-empirical warrants for claims about her environ-
mental relations, the subject will already have empirical warrants for such
claims.38, 39
Conclusion
In this chapter I have tried to provide a relatively comprehensive overview of

the internalism/externalism debate in the philosophy of mind and its implica-
tions. However, it should be clear where my allegiance lies. In the predicative
case I find the considerations that motivate externalism persuasive, the theo-
retical gains of externalism significant, and the considerations against exter-
nalism misguided. On the metaphysical side, externalism is consistent with a
naturalistic theory of the mind and with the claim that mental states are the
causes of our actions. On the epistemological side, externalism is consistent
with the direct, non-empirical, authoritative nature of self-knowledge, and in
addition has the potential to ground an adequate theory of justification.40 In the
singular case, however, a two-factor theory strikes me as theoretically superior
149
for thoughts expressed both by sentences containing demonstratives and by

sentences containing proper names. This yields a theory according to which
names and demonstratives do not function in the same way as predicates,
whether in language or in thought: our fundamental contact with the world
is demonstrative, which itself grounds an externalist theory of the mind.
150
8 The Philosophies of
Cognitive Science1
Margaret A. Boden
Cognitive Science and Pluralism
Theres no such thing as the philosophy of cognitive science. Rather, there are
competing philosophies of and within the field. Thats partly because the
concepts and techniques of artificial intelligence and artificial life have been
changed and enriched since the 1940s. For although psychology is the thematic
heart of cognitive science, its intellectual heart is AI/A-Life or AI, for short.
Psychology (both animal and human) is the thematic heart because cognitive
science studies all aspects of the mind or mind/brain, or, if you prefer, embodied
experience and behaviour. It ranges from low-level vision to enculturated
thought, from infantile development to adult personality, and from individual
behaviour to social phenomena. It investigates not only cognition, but emotion
and motivation too. So the field is badly named: outsiders are o en misled,
assuming that it deals only with cognition.
Cognitive science diers from other forms of psychology in using computa-
tional concepts of various kinds. Very broadly speaking, these fall into two
main types: formalist/symbolic and connectionist/dynamical. Much of the
philosophical interest lies in the dierences between these approaches.
Some computational concepts, in the broad sense intended here, denote
formal computations on symbolic representations. These typify classical AI, or
GOFAI, Good Old-Fashioned AI (Haugeland, 1985, p. 112). Others draw on
cybernetic ideas about embodied and self-organizing systems. These include
situated robotics, wherein the robots rely on direct reflex responses to environ-
mental cues; dynamical systems understood in terms of physical laws; and
self-equilibrating neural networks. And all approaches sometimes include
the sort of computation (mutation and natural selection) thats eected by
evolution.
Cognitive scientists o en express their theories as computer models, because
this is the best way of testing their coherence and implications. (Testing for their
truth, of course, involves comparisons with the actual phenomena.) The com-
putational concepts implemented in such models are substantive theoretical
151
terms. That is, the mind and the brain are theorized as mechanisms that actu-
ally carry out computations of one kind and/or another. Put another way, the
mind is conceptualized as what computer scientists call a virtual machine,
defined in abstract (computational) terms but implemented in the brain.
The varieties of computation mentioned above are best suited for dierent
tasks. All of them (and more) will be needed for an adequate account of the
complex virtual machine which is the human mind (Sloman, 2000; Minsky,
1985, 2006). Some interesting hybrid systems have already been implemented,
in which psychological phenomena are modelled by a combination of GOFAI
and connectionist methods (Norman and Shallice, 1986; Cooper et al., 1995).
However, the various techniques are o en wrongly assumed to be mutually
exclusive (e.g. Dreyfus and Dreyfus, 1988). Thats part of the reason for the
multiplication of philosophies of cognitive science.
Another reason for this pluralism is the fundamental divide between
analytic and continental philosophy. Almost all the early philosophers of
cognitive science came from the analytic community, and were commi ed to
functionalism. Philosophical traditions such as phenomenology were scarcely
mentioned, although Hubert Dreyfus (1967, 1972) was an important exception.
Recently, there have been a empts to combine these two traditions, or even
to forsake one for the other; see the Embodiment, Enactiveness, and Phenome-
nology section below.
The Fate of Functionalism
Functionalism analyses mental states as causal-computational functions: inter-

nal representations and information-processes, which interact with each other
and mediate between input and output. These were defined by Hilary Putnam
(1960, 1967) in terms of Turing-computation, the mind being glossed as the
program implemented in the brain. This implies multiple realizability, because
definitions of abstract functions say nothing about the details of implementa-
tion and because a given program can be realized on many dierent computers.
In other words, philosophers and theoretical psychologists, too can ignore
the brain.
Putnam was the first professional philosopher to recommend functionalism
as the basis for a computational research programme in psychology. But it
had been outlined much earlier in Kenneth Craiks (1943) work on cerebral
models (see the Varieties of Representation section below) and in the seminal
paper of cognitive science (McCulloch and Pi s, 1943). Moreover, Allen Newell
and Herbert Simon had already implemented a theory of problem-solving
which, they insisted, in no way depended on how neural mechanisms realize
information processing in the brain (Newell et al., 1958). A few years later (and
152
independently of Putnam), functionalist philosophies were developed at length

by Newell and Simon (1972, 1976), myself (Boden, 1965, 1970, 1972), and Aaron
Sloman (1978).
The general thesis of functionalism is still widely (though not universally:
see the Embodiment, Enactiveness, and Phenomenology section below) accepted
as the core philosophy of the field. But it has developed into several mutually
disputatious positions.
One is due to Fodor. He remains the philosopher of cognitive science
most faithful to Putnam, or rather, to early Putnam (see the Embodiment,
Enactiveness, and Phenomenology section below). He too describes mental
states in terms of GOFAIs formal-symbolic representations and computation.
He too ignores neuroscience, regarding connectionism as concerned only with
implementation (see the Concepts and Connectionism section below). And he
too takes belief/desire explanation (folk psychology) as the starting-point for
a scientific psychology.
But he has added two highly controversial claims. First, that mental repre-
sentations are composed of atomic items in a language of thought (LOT), whose
primitives cognitive science must discover, and second, that these primitives
are present at birth. Admi edly, babies dont understand natural language. But
Fodor insisted that learning the meaning of airplane is a ma er of identifying
and combining the relevant already-present LOT primitives.
Two further controversial claims were soon added (Fodor, 1983). Namely,
that the mind contains a number of innate, functionally distinct, information-
processing modules; and that these are the only aspects of mentality which can
be scientifically understood. Non-modular computation, Fodor said, does occur
in the higher mental processes. But he argued that there are so many degrees
of freedom here that psychologists cant hope to find laws to predict specific
thoughts, nor even to explain them post hoc. The many discussions of this view-
point (e.g. Karmilo-Smith, 1992; Samuels, 1998) involve both empirical and
philosophical arguments. Thats not surprising, for the philosophy of cognitive
science isnt, and shouldnt be, a ma er for philosophers alone; see Conclusion
below.
Daniel Denne s view on mental representations was very dierent from
Fodors. Whereas Fodor posited representations (formulae in the LOT internal
code) actually present in the mind/brain, Denne described them in a more
ambiguous fashion. The ambiguity took two forms. On the one hand, Denne
didnt restrict himself to GOFAI computation. And on the other, he held that
to ascribe representations beliefs, desires, goals, fears to an organism is to
speak instrumentally, not realistically. Even qualia, he said, are fictions; see the
Consciousness section below.
In his first book (Denne , 1969), Denne spoke of a personal (intentional)
stance for explaining behaviour, and he clarified this idea soon a erwards
153
(Denne , 1971). Now, he distinguished three descriptive/explanatory stances:

the physical stance considers the system as a material thing; the design
stance the proper direction for philosophy and psychology focuses on the
teleological functions for which the system was designed or evolved; and the
intentional stance considers it as a system with beliefs and desires linked by
rational principles.
Even chess-playing computers, he said, have to be described in this folk-
psychological way for their behaviour to be predicted and understood. But
even in the human case, intentional discourse (which he saw as highly impre-
cise, and which assumes rationality rather than explaining it) was a ma er of
interpretation, not discovery. On being accused of instrumentalism (e.g. by
John Searle, 1980), he back-pedalled, comparing intentional states to real
pa erns and abstracta such as centres of gravity (Denne , 1981b, 1987a).
Denne s comments on the multiple imprecision of ascriptions of proposi-
tional a itudes would be endorsed by Paul and Patricia Churchland (Church-
land and Churchland, 1981; Paul Churchland, 1986; Patricia Churchland, 1986).
Their philosophy of eliminative materialism compared folk-psychological terms
to mediaeval talk about witches, a pseudo-factual discourse that would eventu-
ally be wholly eclipsed by science. (That science, on their view, would eschew
GOFAI-based explanations for connectionist ones; see the Concepts and
Connectionism section below.) So they were functionalists, of a sort (as Andy
Clark was too: 1989, 1993). But they had moved a long way from Putnam.
Putnams functionalism was an advance on the previous scientifically ori-
ented philosophy of mind, identity theory (Place, 1956). For it posited only
token-token identity of mental and brain states, not type-type identity, which
last implied that dogs and Martians simply cant feel pain (Lewis, 1980). But
opponents accused it of four major flaws: inability to admit the existence of
qualia (Block, 1978); failure to account for intentionality (Searle, 1980); neglect
of the implications of Godels theorem (Lucas, 1961; Penrose, 1989, pp. 1028);
and a sidelining of neuroscience and the brain in the commitment to multiple
realizability.
That last criticism was sometimes made even by philosophers highly
sympathetic to functionalism. Indeed, it was one of the reasons for the move
towards connectionism, discussed in the next section.
Concepts and Connectionism
Connectionist cognitive science began in the early 1940s (McCulloch and

Pi s, 1943), but was ignored by most philosophers until the late 1980s, when a
particular version of it became prominent: parallel distributed processing
(PDP). This oered a more philosophically plausible account of concepts than
154
GOFAI did. Broadly inspired by the brain, it was less biologically implausible
too. Even so, there are many dierences between PDP systems and real neural
networks; multiple realizability had been diluted only a li le.
In GOFAI, in the psychological research that inspired it (Bruner et al., 1956;
Hunt 1962), and in philosophical writings based on it (Fodor, 1975), concepts
were defined in terms of necessary and sucient conditions. This contributed
to the notorious bri leness of GOFAI systems: in the absence of explicit excep-
tions, just one missing criterion would render a concept inapplicable. Later
psychological work suggested that concepts were less neat and tidy (Rosch
and Mervis, 1975). And some philosophers, of course, had already said so
(Wi genstein, 1953).
In PDP systems, concepts are implemented not by all-or-none activations of
neatly listed defining criteria, but by equilibrium states of the whole network,
involving mini-representations of many dierent, and even partially conflict-
ing, facets. Since each facet is continuously weighted, and the weights can
change according to the contextual evidence, concepts arent accepted/rejected
outright but are given varying degrees of confidence.
Many philosophers were enthusiastic (Clark, 1989, 1990, 1993, 1996; P. M.
Churchland, 1986, 1989b; P. S. Churchland, 1986; Churchland and Sejnowski,
1992; Thagard, 1988, 1989, 1990). They used connectionist ideas not just as a
way of glossing concepts, but as a way of thinking about mind in general
including epistemology, philosophy of science, and ethics.
One major implication of PDP appeared to be that the GOFAI/Chomskian
emphases on explicit rules and on innateness were each mistaken. One PDP
network not only learnt from input examples to form the past tenses of regular
and irregular verbs, but the changes in its performance over time seemed to
match aspects of infants behaviour, which Chomsky had claimed were incon-
trovertible evidence for innate explicit rules (Rumelhart and McClelland,
1986).
The Chomskians fought back. While some focused on criticizing the details
of that specific network (Pinker and Prince, 1988), others argued that PDP
models in general, because of their holistic nature, couldnt satisfy the general-
ity constraint (Evans, 1982), nor capture the compositionality, productivity,
and systematicity of language (Fodor and Pylyshyn, 1988). For Fodor (and for
Newell too: 1980, 1990), connectionism was not a theory of cognition as such,
but of its implementation.
Fodor still insists that a GOFAI-based psychology is the only [theory of
cognition] weve got thats worth the bother of a serious discussion (Fodor,
2000, p. 1). Worth the bother or not, this dispute has triggered a huge literature.
Some philosophers have claimed that PDP helps us to understand the
nature of non-conceptual content, and how it can lead to genuinely conceptual
meaning (Cussins, 1990). Paul Smolensky (1987, 1988) rebu ed Fodors critique
155
of PDP (for Fodors response, see Fodor and McLaughlin, 1990), saying that
GOFAI-explanations of thought and language are approximations to finely
detailed sub-symbolic accounts. Smolenskys fellow connectionist Clark (1989,
Chapter 8; 1991) disagreed. In arguing that that PDP could deliver systematic-
ity, he admi ed that productivity was problematic for PDP systems (more so
than sequential order). But he said that compositionality neednt be built into
the basic architecture, as it is for GOFAI systems; instead, it could emerge from
it. Given the public availability of language, our brain can model individual
words and thereby create a virtual machine with formalist properties. So
Smolensky was wrong: GOFAI accounts of thinking arent mere approxima-
tions of PDP accounts, but are true (or false) descriptions of processes in the
relevant virtual machine.
The PDP group themselves had suggested such a machine: they believed
that, to be able to do logic/mathematics or engage in hierarchical thinking, the
brain must somehow emulate a von Neumann computer (Norman, 1986). But
despite various a empts (e.g. Touretsky and Hinton, 1985, 1988; Hinton, 1990;
Elman, 1990, 1993), no one has yet shown how this is possible. PDP models
still cant match the formalist strengths of GOFAI.
The eorts of many philosophers to justify either GOFAI or connectionism as
the key to the philosophy (and psychology) of mind are ill judged. The mind is
a complex virtual machine that includes both these types of computation and
probably many others (see the What is Computation? section below). More-
over, the fact that the brain consists of interconnected neurones doesnt
justify dismissing connectionism as mere implementation, or GOFAI as
non-biological. For a virtual machine is defined in terms of its computational
functions, not its physical implementation.
It follows that brain anatomy alone cant justify eliminative materialism.
It may be that the concepts of folk psychology arent useful in a scientific
psychology. But thats to say that they may not figure crucially in the brains
virtual machine. Their relevance/irrelevance cant be decided purely on the
grounds that the brain is a mass of interconnected neurones.
Varieties of Representation
Cognitive science typically posits cerebral models (Craik, 1943) that is,
representations in the mind/brain. And in positing such models/representa-
tions, cognitive scientists also oer hypotheses about the computational pro-
cesses that manipulate them (build, compare, transform, combine them).
In GOFAI, in psychological theories inspired by GOFAI, and also in
Putnams functionalism, these are thought of as symbolic representations with
compositional semantics. Another way of pu ing this is to say that the virtual
156
machine which is the mind is supposed by many to be purely symbolic-

computational. However, work in AI and in neuroscience has shown that there
are many other possibilities.
One set of alternatives comprises connectionist representations of various
kinds. These may be localist (each network unit being assigned a specific mean-
ing) or distributed (the representation being spread across the whole network,
wherein individual units may carry dierent meanings according to context).
Yet more types of representation have been posited. The evidence oered is
largely behavioural, but is increasingly backed by research in neuroscience
(Parker et al., 2002). The varieties include: physical and/or spatiotemporal
analogues; feature detectors; deictic (situationist), linguistic, and iconic repre-
sentations; sensorimotor models based on tensor network geometry; hier-
archical conceptual and sensorimotor schemas; culturally relevant scripts;
temporary online representations; and neurophysiological emulators.
Some of these types of internal model are believed to be symbolic-
computational by some philosophers, whereas others are explicitly contrasted
with GOFAI representations. Not all of them are accepted by all workers in the
field. But nearly all cognitive scientists oer theories couched in terms of some
sort/s of representation. (For exceptions, see the Embodiment, Enactiveness,
and Phenomenology section).
From the philosophical point of view, there is another question: namely,
what counts as a representation. This was a query raised long ago by Craik
(1943) who leant heavily on the criteria of use (for survival) and of similarity,
and also by Marvin Minsky (1965) who stressed the importance of use. How-
ever, most empirical cognitive scientists employ the words representation
and model in largely intuitive senses, obscuring important philosophical
distinctions and begging controversial philosophical questions.
Philosophers of cognitive science have tried to clarify the issues here, usually
with some reference to aspects of the scientific literature (e.g. Sloman, 1971,
1975, 1978, Chapter 7; Pylyshyn, 1973; Cussins, 1990; Kirsh, 1991; van Gelder,
1995; Clark and Grush, 1999; Clark and Wheeler, 1999; Grush, 2004). Their
views dier greatly in detail, but three things are now clear. First, there need
not be, although there may be, a similarity between the representation and
the thing represented. Second, if there is, the theorist must explain just how
the similarity is exploited and used. And third, for something to function
(sic) as a representation, it must somehow mediate in the agents thought or
behaviour where that somehow should be explicitly spelt out, preferably in
computational and/or neuroscientific terms.
Philosophers of mind with scientific sympathies, but no special interest
in cognitive science, ignore the specific mechanisms suggested by AI and
neuroscience. But they do oer naturalistic accounts of representation, and/or
of intentionality, explaining sense-making in terms of evolutionary biology
157
(Papineau, 1984, 1987; Millikan, 1984), causal regularities (Dretske, 1984, 1995a),
or biological autopoiesis (Jonas, 1966; Di Paolo, 2009). Since the concept of
representation is o en used in defining cognitive science, disagreements about
what representation is can be reflected in conflicting judgments about the
fields scope and success. Given Fodors formalist sense of representation, con-
nectionism is either a refutation of cognitive science (as Dreyfus claims) or a
mere implementational adjunct to it (as Fodor believes). Given a more catholic
definition of representation, connectionism is an interesting example of cogni-
tive science, and further examples have been mentioned above. Similarly,
dynamical and autopoietic theories and situated robotics (all of which deny
representations) must be excluded from the field unless one can show, as some
critics have argued, that they too involve (non-formalist) representations.
Because of these problems, and despite the huge influence in philosophical
circles of Fodors view, its best not to use the term representation in defining
cognitive science.
Some experimental psychologists outside cognitive science explicitly deny
the existence of mental representations. For instance, James Gibsons (1950,
1966) ecological approach emphasizes direct (computation-free) responses
to environmental cues. There have been many debates between Gibsonians
and the followers of David Marr, who analyzed low-level vision and object-
recognition in terms of a hierarchy of representations (Marr, 1982; Ullman, 1980;
Hinton, 1980; Sloman, 1989; Norman, 2002). Marrians typically complain that
although Gibsonians have reported important empirical data about responses
to perceptual cues, they fail to ask how those responses are possible.
Gibsons theory includes the notion of aordances (Gibson, 1977). He
claimed that perception doesnt only, or even primarily, provide knowledge
about the things in the external world, but rather provides (direct) knowledge
of the possibilities for action that are aorded by those things. A gap, for example,
is perceived as something that can be moved through, and a smile (or other
bodily expression of positive emotion) as aording approach without fear.
Not all aordances have to be learnt: the visual cli experiment suggests
that even very young babies can see (sic) that a sudden drop in the floor is
dangerous, and refuse to crawl over it (Gibson and Walk, 1960).
Fodor has complained that Gibson gave no principled way of deciding
what counts as an aordance (Fodor and Pylyshyn, 1981). For him, some
specific story about computations over representations is needed to explain
perception of any kind. Today, that view is less prominent.
The (non-computational) representational theory of perception has been a
philosophical embarrassment ever since Descartes. Accordingly, ideas about
direct, representation-free perception have also been developed within philo-
sophy by the continental phenomenologists. Some of them, such as Maurice
Merleau-Ponty (1962), borrowed heavily from the Gestalt psychologists; and
158
all would have been more sympathetic to the Gibsonians than to the Marrians.
But none admi ed the possibility of a scientific (naturalistic) explanation of
intentionality, not even in neuroscience, never mind AI. According to them,
what counts as a representation cannot be understood in terms of information-
processing mechanisms of any kind, nor in terms of biological evolution either.
Rather, intentionality is seen as a basic philosophical notion, to be understood
(though not explained) in terms of situatedness, embodiment, in-dwelling, and/
or Dasein (see the Embodiment, Enactiveness, and Phenomenology section).
The Extended Mind
In the early days of cognitive science, the psychologist Jerome Bruner wrote
about how cognitive technologies (motor action, imagery/drawing, and
language) enable us to think (Greenfield and Bruner, 1969; Cole and Bruner,
1971). Similarly, Newell and Simon (1972) stressed the use of external memory
such as wri en sums and memos in problem-solving. And the cognitive anthro-
pologist Anthony Wallace (1965) showed how drivers regulate their journeys
by monitoring the input/feedback from roads, trac lights, landmarks, and road
signs, and from clutch, indicators, and gear shi . These cognitive scientists, and
others a er them, have provided a wealth of evidence that cultural artefacts
in general (and above all, language) enable us to think thoughts, and to raise
and answer questions that would otherwise have been beyond our powers.
However, they didnt go so far as to claim that our minds are partly
constituted by these artefacts. Nor did those early visionaries who predicted
that the availability of certain types of information technology then, only
just conceivable, but now common would deeply change the way we think
(Bush, 1945; Engelbart, 1962). Scientists and engineers in general dont normally
ask constitutive questions.
Philosophers, however, o en do. So the anthropologist Cliord Geertz (1973),
speaking in philosophical mode when recommending a non-psychological
view of culture and anthropology, said that the mind is located outside the
head. And the philosophers Clark and David Chalmers went even further,
locating the mind and the self largely in the external (physical/cultural) world
(Clark and Chalmers, 1998; Clark, 2003a, 2008b).
Clark and Chalmers argued that while there is a clear distinction between
brain (or body: Clark, 1997) and artefacts, considered as material things, there
is no principled distinction between the (abstractly defined) processes of control
that link them in behaviour. If an amnesiac constantly consults a personal
notebook to decide what to do, and even how to do it, the complex feedback
loops and information sources involved cant be neatly assigned to either brain
or world, but only to a closely coupled merger of the two. If, very broadly
159
speaking, the mind is what the brain does (the credo of functionalism), then the
mind itself is partly constituted by things in the world.
Fodor (2009) isnt convinced. Computation and control, he insists, are
grounded in the brain. He insists, also, that the defining feature of mind is con-
tent, or aboutness, and that only brain processes can have content in a non-
derivative sense. If these processes continually consult, aect, and are aected
by aspects of the outside world, it doesnt follow that those aspects notebooks,
i-phones, whatever are actually parts of the mind. One doesnt have to
share Fodors views on content to agree; its not clear that theres any great
advantage, either philosophical or scientific, in making the strongly counterin-
tuitive constitutive claim. Nevertheless, some philosophers are sympathetic:
Clark and Chalmerss article won a Philosophers Annual prize as one of the
years ten best papers in philosophy. And some cognitive scientists, likewise,
were persuaded. In certain circles, extendedness and externalism have
become buzz-words.
This philosophical approach is especially well suited to cognitive scientists
talk of distributed and situated cognition. Much as AI-work on PDP refers
to distributed processing and distributed representations, so AI-work on
autonomous agents refers to distributed cognition (Bond and Gasser, 1988).
The cognition (problem-solving, as well as knowledge) is thought of as being
distributed over a number of interacting agents. And those agents are usually
described as situated, meaning that they arent controlled by a top-down
problem-solving GOFAI program but by bo om-up cues from the environ-
ment, to which they respond directly, in a near-reflex fashion (Agre and
Rosenschein, 1996; cf. Brooks, 1991a, 1991b).
Both these ideas are highlighted by the cognitive anthropologist Edwin
Hutchins (a follower of Wallace) in a study of ship navigation (Hutchins, 1995).
He shows that this skill isnt located inside the head of a single person, or
even several. It emerges from a complex coupling of individual personalities,
social roles and conventions, maps and instruments, ship-design, mariners
knowledge, problem-solving (o en spread across several crew members), and
a variety of bodily skills.
The thesis of the extended mind need not be combined with the concept of
embodiment (although the notion of situatedness is very close to it). But even
before the prize-winning article, Clark had argued for a philosophy of mind
grounded in embodiment.
Embodiment, Enactiveness and Phenomenology
The notion of embodiment was important in the philosophy of mind long

before it entered cognitive science, for it is a key concept of continental
160
phenomenology. Indeed, Dreyfus (1967) early a ack on AI drew on Martin

Heidegger and Merleau-Ponty, and declared that computers will have to
have bodies in order to be intelligent.
By this, he didnt mean that robots would be intelligent. For a body isnt
just a self-moving physical object. For autopoietic philosophers (see below), a
body is a living thing: an autonomous system that is the result of physical self-
organization. For phenomenologists, it is the physical system through which
the intentional agent (actor) is coupled with, or situated within, the world.
The philosophy of embodiment is closely related to enactivism, a position
that stresses the agents bodily activity as a condition of perception and thought
(Varela et al., 1991; No, 2004). The evidence includes intriguing experiments
on change-blindness and ina entional blindness (Simons and Chabris, 1999;
ORegan and No, 2001). Here, what one would expect to be startling environ-
mental information (e.g. one interlocutor being replaced in mid-conversation
by another, or a gorilla cavorting amidst a group of humans) simply does not
register in the perceivers consciousness if they are actively (sic) a ending to
something else.
Within psychology, the Gibsonians had long focused on how the subjects
own bodily movements influence perception. But this notion, sometimes called
animate vision, became prominent in the philosophy of cognitive science
only much later when continental philosophy finally crossed the intellectual
Channel to the Anglo-American world.
When cognitive science began, the phenomenological movement was
generally discounted, even despised, by analytical philosophers. Eventually,
however, some analytically trained philosophers, including John McDowell
(1994b), Michael Morris (1991, 1992), and even Putnam himself (1982, 1988,
1997), drew closer to the continental tradition. They now oered philosophies
of mind wherein representations were explicitly denied. And folk psychology,
beloved of many functionalists, was dismissed as a scientistic myth that is
wrong in every particular (Morris, 1992, p. 111).
Morris was unusual in deigning to address some claims of cognitive science.
While rejecting internal representations that is, non-semantically individuated
objects/events, such as brain states, that have meaning), because no natural-
istically identified event can be inherently meaningful, he oered alternative
accounts of the phenomena explained by cognitive scientists in terms of them
(Morris, 1992, pp. 2830). He even allowed that cognitive scientists may some-
times be justified, for scientific purposes, in stipulating that certain states/
events in the brain have such-and-such meanings. But he believed this would
be appropriate only for low-level cognitive capacities.
A few philosophers sympathetic to the continental tradition take science
(though not functionalist cognitive science) more seriously. Humberto Maturana
and Francisco Varela, for instance, formulated the philosophy of autopoiesis
161
(a concept similar to that of metabolism: Boden, 2000a), which explains biologi-

cal phenomena in terms of spontaneous self-organization (Maturana and
Varela, 1980). Maturana had co-authored the seminal neuroscientific paper
on What the Frogs Eye Tells the Frogs Brain (Le vin et al., 1959), but when
writing on autopoiesis he specifically rejected its functionalist language and
assumptions. Varela and other autopoietic psychologists describe the activity
of the autonomous agent as a system of intimate couplings between organism
and environment (Varela et al., 1991). Indeed, autopoietic enactivism defines
cognition (a.k.a. mind) as the bringing-forth of a world by an autonomous
embodied agent, coherently engaging with its environment (Di Paolo, 2009).
Another representation denier, prepared to take science seriously, is Timothy
van Gelder (1995). He argues that cognition should be understood in terms not
of computation, but of dynamical systems describable by the laws of physics.
However, his arguments fall foul of two issues discussed in the What is
Computation? section. First, computation should not be interpreted only as
formal symbolic information-processing. And second, the mind is best seen
as a many-levelled virtual machine, so that events that are physically imple-
mented in the brain by dynamical systems may be understood as diverse
aspects of the virtual machine, including, among many other things, GOFAI
computations and representations.
Ironically, given the anti-scientific bias of the continental tradition, some
(empirical and philosophical) work in cognitive science has been deeply
influenced by Merleau-Ponty and Heidegger. Thats not to say that these
cognitive scientists take Heideggers philosophy as gospel. Michael Wheeler
(2005), for instance, specifically denies that Dasein is restricted to language-
using human beings, and absent from animals; he also demurs from the rejection
of any internal representations; and he criticizes the anti-realist implications of
phenomenology. But some Heideggerian concepts (skilful coping, readiness/
unreadiness-to-hand, presence-at-hand, background) are used by him to
describe and justify enactive and dynamical cognitive science research.
A new twist on the realism/anti-realism debate has arisen in connection
with the technology of virtual reality (VR), and in particular with respect to the
1999 film The Matrix. The human beings in this story are supposed to be
experiencing a purely virtual world, while their living but inert bodies are
farmed for energy by the machines in charge. The machines are clever enough
to fool the people that they are walking, talking, and eating in a world like our
own. Several leading philosophers of cognitive science have considered the
implications, sometimes with reference to specific scientific findings/theories
in the field. Some of their discussions are in online papers (Dreyfus and
Dreyfus, 2002; Clark, 2003b; Chalmers, 2003b), others (Searle and Denne ) in
online interviews on the same website, and yet others in print (Dreyfus, 2003;
McGinn, 2005). (See also Irwin, 2002; Lawrence, 2005; Grau, 2005.)
162
Consciousness
Much as there is no such thing as the philosophy of cognitive science, so there

is no such thing as the problem of consciousness. The many problems of con-
sciousness encompass various intriguing aspects of the human mind. Most
have been illuminated by cognitive science. Philosophical discussions, here, are
usually informed by empirical evidence drawn from studies of human beings
(and animals), and also from computer modeling.
Crucially, such evidence reports many phenomena that cannot be shoehorned
into a Cartesian vision of error-free introspection and unitary consciousness.
Experimental and clinical psychologists have long challenged common-sense
intuitions on ma ers ranging from a ention (Broadbent, 1958), through percep-
tion (Kolers and Rosner, 1960; Kolers, 1972) and intelligence (Damasio, 1994), to
voluntary action (Libet, 1985a, 1985b, 1999; Libet et al., 1979) and its debilitating
breakdown (Norman and Shallice, 1986; Cooper et al., 1995).
Neuroscientists have provided relevant evidence too. But the methodology
of brain-imaging has not been especially helpful. It has discovered many
mind-body correlations; but Descartes predicted such correlations, so this
hardly counts as a philosophical advance. Usually, it isnt even a scientific
advance, since most work of this type reports a-theoretical fishing expeditions,
which ignore the complexities of the psychological phenomena concerned.
When neuroscience is relevant, thats because it throws light on the neural
mechanisms underlying the computations involved in consciousness. For
example, autistic children unable to master theory of mind do not show
normal brain activations when presented with intentional concepts (Frith and
Frith, 2000).
Much of the past puzzlement about consciousness has rested on the seem-
ingly paradoxical phenomenon of reflexive self-consciousness. However, a
computational philosophy, thanks to the computer scientists concept of recur-
sion, allows for non-paradoxical theories of self-reflection. These include
accounts of meta-cognition (higher-order thought), adduced in explaining
not only conscious reasoning and deliberative self-control but also hypnosis
(Dienes and Perner, 1999, 2007).
Similarly, puzzles about the possibility of dissociative states, such as multiple
personality syndrome, dissolve when the mind is viewed as a complex system
comprised of partly independent teleological structures, controls, and memo-
ries/data-bases (Boden, 1994). Free will can be understood as a feature of
cognitive/motivational systems with sucient complexity to compute the likely
eects of alternative (sequences of) hypothetical actions, and to compare them
with respect to moral and personal preferences of many kinds before deciding
what to do (Boden, 1978; Denne , 1984). The self itself can helpfully be seen
as a narrative construction, involving self-image and self-ideals, which guides
163
and interprets the persons behaviour (Boden, 1972, pp. 23660, 32733;
Denne , 1991a, Chapter 13).
All those puzzling phenomena were addressed at length by Denne (1991a),
who analysed them in broadly computational terms. Other cognitive scientists
have oered general theories of consciousness somewhat similar to his, such
as Bernard Baars (1988) integrative idea of the global work space, Thomas
Metzingers (2003) self-model theory, and Slomans architectural account
described below (see also Crick and Koch, 1990; Frith et al., 1999). But Denne s
is among the most philosophically sophisticated.
Of Denne s many provocative claims, the most controversial is his position
on what Chalmers (1996) calls the hard problem namely, how to explain
(or even admit) the existence of subjective experiences, or qualia. Early critics of
functionalism declared that it couldnt account for qualia, because zombies,
behaving exactly like humans, but with no conscious experiences, are in
principle possible (Block, 1978; Zi, 1959). Denne disagrees: the concept
of zombies, he says, is incoherent, and belief in their possibility ridiculous
(Denne , 1991a, Chapter 10, p. 4; 1995). Sloman (1996c, 1999) argues, similarly,
that nothing could have the same computational architecture as us (necessary
for it to behave exactly like us), yet lack sensation.
On one key point, however, these two computationalist zombie-deniers
disagree with each other. For Denne , qualia are mere fictions (Denne ,
1988; 1991a, Chapter 12). Once everything has been said about the behavioural
aspects (the many subtle discriminative and associative activities) of seeing
blue, or tasting chablis, nothing more remains. For Sloman, by contrast, qualia
do exist. They arent visible as overt behaviour, nor verbally describable to
others, but nor are they events in some essentially mysterious immaterial
world. They are internal (but self-accessible) computational states, whose
generation and functions are possible only in complex computational
systems virtual machines of a particular architectural type (Sloman and
Chrisley, 2003).
Slomans philosophy of mind takes the design stance seriously. Drawing
on a wealth of experience in AI, and knowledge of a wide range of animal
behaviour, Sloman makes relatively specific suggestions about which sorts of
computational structures and processes could and which could not generate
particular types of cognition and control (Sloman, 1978, 1993, 2000). For
example, various types of anxiety, and the complex emotion of grief, are made
possible by distinct types of computational mechanism, some of which can
already be modelled, up to a point, in computers (Wright et al., 1996).
Sloman regards most philosophical discussion of consciousness as fixated
on highly confused concepts (Sloman and Chrisley, 2003; Sloman, 2010).
These include not only long-familiar notions (such as qualia), but also suppos-
edly more precise and more recent terminology (such as phenomenal and access
164
consciousness: Block, 1995). This conceptual confusion partly accounts for the
rise of new mysterian accounts of consciousness (Flanagan, 1992, pp. 8f.),
according to which it is either u erly unintelligible to human minds (McGinn,
1989, 1991) or intelligible only in terms of arcane, even undiscovered, aspects
of quantum mechanics (Penrose, 1994; Hodgson, 1991) and/or of an informa-
tion-imbued universe (Chalmers, 1996).
Searle (1980, 1992) isnt one of the new mysterians, but he might be termed
an old mysterian. In rejecting functionalist/computational accounts of con-
sciousness (and intentionality), he insists that we know that it is a biological
phenomenon, just as digestion and photosynthesis are. This knowledge,
however, is merely the fact that we now have even more evidence than
Descartes did to believe that the brain causes/generates consciousness. What
we want to know is How? At the level of material stu (compare: lactose,
chlorophyll), the emergence of consciousness is intuitively unintelligible. Searle
oers no new ideas, mysterian or not: no micro-tubules, no dual-aspect
information, no special types of computation or computational architecture.
He simply says that the problem of consciousness is a scientific problem, and
leaves the scientists to get on with it.
There are now several conferences, and an international journal, dedicated
to machine consciousness. Sometimes, this phrase is glossed by references to
mechanisms explaining human or animal consciousness. O en, however, it
is used to suggest that computers of a certain sort are, or anyway would be,
genuinely conscious. Some such claims are philosophically (and computation-
ally) thin (e.g. Aleksander and Dunmall, 2003; Aleksander, 2005). A few are
more substantive. For instance, Sloman has outlined a specification for a
machine whose normal functioning could lead it to discover within itself
something akin to qualia, as a result of developing an ontology for describing
its sensory contents (Sloman and Chrisley, 2003; cf. Sloman, 2010, Section 13).
Mind and Life
Many philosophers assume that life is necessary for mind (e.g. Scriven, 1953,
p. 233), and that, because computers arent alive, strong AI is impossible
(Geach, 1980, p. 81). However, the necessity of the life/mind linkage is more
o en taken for granted than explicitly justified and when arguments are given,
they are usually weak.
Cyberneticists in general assume that the same principles of control govern
both life and mind that is, they are strong continuity theorists with respect to
the ontological similarity of life and mind (Godfrey-Smith, 1994b). However,
they pay scant a ention to issues such as self-reflection and reasoning.
Even language, though sometimes mentioned, isnt considered in any detail.
165
So philosophers of biology who construe life as self-organization claim that

life can culminate in cognition, but what they say about cognition applies as
much to oak trees as to humans (Pa ee, 1966, 1989; Jonas, 1966; Maturana and
Varela, 1980; Cariani, 1992; Sober, 1992).
Nor do they normally ask whether mind without life is possible. The
existentialist theologian Hans Jonas did tackle this question (Jonas, 1966,
pp. 6491). But his argument that life is essential for mind rested on the prob-
lematic anti-Cartesian claim that all self-organized (metabolic) ma er is, in
a sense, ensouled. A less provocative way of pu ing this is to say that the
adaptivity of an autonomous (autopoietic) living system enables it to construct
both norms and sense, so that ethics and intentionality are seen as naturalized
(Di Paolo, 2009). Even here, however, very li le is said about specifics, although
there are some intriguing ideas about the origin of sociality (De Jaegher and
Di Paolo, 2007).
If life really is necessary for mind (intentionality), then A-life is philo-
sophically, and perhaps methodologically, prior to AI. In other words, an under-
standing of life as such (which is one aim of A-life) should bu ress, and perhaps
even lead to, the understanding of mind. It would follow, too, that strong A-life
(i.e. virtual life, or life in cyberspace) is necessary for strong AI.
Despite the claims of a few A-life scientists (e.g. Ray, 1992), strong A-life is
impossible. Metabolism is a criterion of life, and computers dont metabolise.
They use energy, but what biologists mean by metabolism is more than mere
energy dependency. It is the self-production and self-maintenance of the phys-
ical organism by energy budgeting that involves self-equilibrating energy
exchanges of some necessary complexity (Boden, 1999). Since theres no univer-
sally agreed definition of life, someone might suggest that metabolism be
omi ed. However, the only gain would be to allow the possibility of virtual
life, which begs the question. And there would be a huge loss namely, the
explanatory power that the concept of metabolism, with the associated laws
of bio-energetics, provides with respect to all living things.
What is Computation?
Philosophers typically believe that we know what computation is because

Alan Turing (1936) told us. They rely unquestioningly on his definition
computation as formal symbol manipulation in expressing their own views
on mind and/or cognitive science (e.g. Putnam, 1960; Fodor, 1975, 1980; Searle,
1980; Pylyshyn, 1980; Haugeland, 1985; Copeland, 1993).
However, ma ers are more complicated. Turings is still the only rigorous
definition. But as the practices of AI and computer science the virtual machines
involved have become increasingly varied, additional senses have arisen
166
(a dozen are distinguished in Smith, 2002a). In general, these treat computation

not as abstract, uninterpreted, symbol manipulation but as actual processes
implemented in computers.
Some logicians have asked just which computers are equivalent to Turing
machines, and which are not (Bringsford, 1994; Calude et al., 1998; Copeland
and Sylvan, 1999; Scheutz, 2002). Sloman (2002) has questioned the historical
relevance of Turing machines to the development of AI. As for their philo-
sophical relevance, he says that No programmer or computer engineer has, to
my knowledge, ever thought of programs in [Searles, i.e. Turings] way, and as
a programmer myself I have never thought of programs that way (p.c., quoted
in Boden, 2006, p. 1415); and he has shown that the notion of Turing equiva-
lence doesnt capture the richness of the concept(s) of computation needed to
understand minds (Sloman, 1996b).
For our purposes, three senses of computation can be distinguished. The
first is Turings. The second is one whose denotation has changed over the
years, and will continue to do so namely, Whatever methods (virtual machines)
are actually used in computer modelling. These include not only the many methods
within GOFAI and connectionism, but also (for instance) approaches that
simulate the computational eects of chemicals diusing through the brain
(Husbands et al., 1998; Smith et al., 2002).
The third sense is the most philosophically interesting. Its also the most
unclear, because its a group of rather dierent senses, all informed by the same
general aim: to present computation as intentional, and meaning as computa-
tional. What computers do is conflated with what minds do, by some account of
intentionality that supposedly applies to both.
One example is Newell and Simons theory of physical symbol systems, seen
as the necessary and sucient means for general intelligent action, whether
in minds or computers (Newell and Simon, 1976; Newell, 1980). A symbol was
defined as a physical pa ern with causal eects. Likewise, causal definitions
were given of intentional terms such as representation, interpretation, designation,
reference, naming, standing for, and aboutness.
Another third-style example is due to Sloman, who holds that we dont yet
understand computation largely because we dont yet understand causation.
He dismisses causal theories of meaning (like Newell and Simons) that assume
some physical relation between a symbol and its referent, because we can refer
to non-existing things (Sloman, 1986). He stresses the virtual causal processes
required for understanding, arguing that a virtual process can properly be said
to cause a physical one, so that qualia (which he analyses in computational
terms; see the Consciousness section above) arent epiphenomenal, but really
do cause changes in the brain (Sloman, 1992, 1996a, 1996c).
The most controversial third-style position is that of Brian Cantwell Smith
(1996). His early analysis of computation as causal, not formal, rested on insights
167
about what computers actually do (Smith, 1985). It has developed into a philo-
sophy of presence that is a new metaphysics: an account of the emergence of
objects, individuation, particularity, subjectivity, and meaning. Mind, or inten-
tionality, is a form of active registration that requires a relatively high degree of
disconnectedness, or autonomy, as well as connectedness. This gives rise to sub-
jectivity and objectivity alike. Smith sees no fundamental distinction between
intentionality in people and computers: computation is inherently participatory,
and computers have intentional capacities ultimately grounded in practice,
analogous to the human practices stressed by Heidegger and Dreyfus (Smith, 1996,
pp. 305, 149). Both physical objects and intentional subjects, he argues, arise
from the participatory engagement of distinguishable regions of the meta-
physically basic dynamic flux. This flux is described by field-theoretic physics,
and having no objects involves neither individuality nor particularity. Objects
emerge, or are constructed, as a result of dynamic participatory relations. So where
others were broadening Dasein from humans to animals (see the Embodiment,
Enactiveness, and Phenomenology section above), Smith broadened construc-
tive dynamical interaction from animals to rocks, and even to atoms.
Smith claims to have retained the major insights of both continental and
empirical-analytic traditions, without any of their problematic ontological
assumptions. He also claims that his metaphysics gives us norms as well as
facts: a way of living right as well as a way of speaking truthfully (Smith, 1996,
p. 108). Some philosophers are deeply impressed: John Haugelands jacket-
blurb says Smith recreates our understanding of objects essentially from scratch
and changes, I think, everything. Others are only partly persuaded, allowing
that the physical implementation of computation is important although reject-
ing Smiths account of it (Searle, 1990b, 1992, p. 209). Yet others are highly scep-
tical, and also repelled by his vividly purple prose. (My own view is that he has
helped himself to the dynamic flux, his version of Kants noumenal world,
without proper licence, despite his claim, in the final 60 pages, to have pulled
this concept up by its own bootstraps.)
As weve seen, Smith isnt the only one to regard the concept of computation
(understood intuitively as what computers do) as problematic. So the core
thesis of cognitive science, like that of physicalism, should be interpreted
transparently, not opaquely (Chrisley, 2000). The claim isnt that mind can be
explained by our current ideas about computation, but that its explicable by
whatever theory turns out to be the best account of what computers do.
Conclusion
Thirty years ago, Sloman predicted that philosophers today would be pro-
fessionally incompetent if they werent well informed about developments in
168
AI (Sloman, 1978, p. xiii). He was right. This chapter has described a host of
examples where the findings of cognitive science, and especially the concepts
developed within AI/A-life, have provided intriguing questions, and some-
times plausible, or even satisfying, answers, for the philosophy of mind. These
questions and answers cannot properly be ignored. Even those who reject the
fundamental assumption that there can be a naturalistic psychology should
(like Morris) at least engage lightly with some of the data/theories of cognitive
science.
Moreover, cognitive science can contribute to longstanding problems
ranging from the philosophy of science (Sloman, 1978; Thagard, 1988, 1989;
Churchland, 1989b; Whitby, 1996) and of religion (Arbib and Hesse, 1986; Boyer,
1994), through metaphysics (Smith, 1996; Sloman, 1996a), to ethics (May et al.,
1996), aesthetics (Turner, 1991; Boden, 2000b, 2007; Boden and Edmonds, 2009),
and philosophical logic (McCarthy and Hayes, 1969; Pearl, 2000). Thats hardly
surprising. Insofar as philosophy is the study of systems of thought, and ways
of knowing, it is concerned with the mind. So a science of the mind is likely to
have connections to, and implications for, most areas of philosophy.
Arguably, the study of metaphysics is dierent. But even there, the concern
is with how we should identify and think about the most basic categories of
being. AI models using natural language, for example, must o en employ
notions of space, time, and cause. The AI definitions of these categories leave
much to be desired. But AI researchers with philosophical expertise have con-
tributed not only to AI modelling but also to philosophical understanding of,
for example, causation (Sloman, 1996a; Pearl, 2000).
As that last example suggests, the potential for dialectical enrichment goes
both ways. When philosophers engage with the field in a serious spirit, they
may come up with views that influence the science itself. They may help it to
advance not only by clarifying current scientific concepts but also by oering
new insights for empirical study.
Examples of this salutary eect include Denne s work on the intentional
stance, which has been used by cognitive ethologists to guide experimentation
and theorizing on the minds of many species (Grin, 1978, 1984). (Phenome-
nologists who refuse intentionality to animals will describe these data in dier-
ent terms; but they must admit that interesting new data have been discovered
by cognitive ethologists, and that the writings of philosophers have played
some part in furthering this.) Another piece by Denne which influenced
empirical research was his (Denne , 1978) snippet on modelling the whole
animal, although some cognitive scientists, to be sure, had already worked out
that message for themselves (Arbib et al., 1974; Arbib, 1982).
Two more examples, concerning the acquisition of conceptual representations,
are due to Clark, writing in cooperation with a developmental psychologist
(Clark and Karmilo-Smith, 1993) and a connectionist AI scientist (Clark and
169
Thornton, 1997). My own analysis of the concept of creativity, and of the com-
putational mechanisms underlying it, has prompted research in both AI and
psychology (Boden, 2004). Yet more instances are due to Sloman, including
his work on temporary representations (which influenced the recent dual
pathways theory of vision: Goodale and Milner, 1992), on virtual machines
and computation, on emotions and mental architecture, and on the space of
possible minds.
So the appellation cognitive scientist covers many people who are primarily
regarded as philosophers. Theres no fundamental distinction between those
who actually do science and those who restrict themselves to commenting on
it. Thats not to say that science can, of itself, answer philosophical questions.
But the days when respectable philosophy, and especially the philosophy of
mind, was thought to exclude any reference to science should by now be well
and truly over.
170
9 Representation
Georges Rey
Introduction
Representation has come to be used in contemporary philosophy and cogni-

tive science as an umbrella term to include not only pictures and maps, but
words, clauses, sentences, ideas, concepts, indeed, virtually anything that is
a vehicle for intentionality (i.e. anything that stands for, means, refers to or
is about something). We will abide by this broad usage, and for the most
part not distinguish among these dierent forms of representation and
intentionality here.
Following many authors, however, we can distinguish original from derived
intentionality and representations. Ordinary maps, drawings, paintings, and
the words and sentences of natural language, are representations that stand for
or about something by virtue of how they are deliberately used for usually social
purposes by people and perhaps some animals: their status as representations
is derived in complex ways from the ideas and intentions of those people and
animals. But what of the ideas and intentions themselves? On pain of regress,
their intentionality is not derived from anyones deliberately using them for
such a purpose, but is original. I shall confine the discussion of representation
here to representations with original intentionality, and presume that these are
all mental ones.1
The issues of original intentionality and representation have become parti-
cularly important in the last several decades due to the increasing interest in
a computational/representational theory of thought (CRTT) in a wide variety
of approaches to psychology from, for example, theories of reasoning and
decision-making, to theories of vision, language acquisition and animal naviga-
tion. In the second section of what follows, Ill briefly review the needs and
prospects of CRTT, turning in the third section to consider a general puzzle
about intentionality and representation, before turning in the remaining
sections (fourth through sixth sections) to the dicult issues of determining
by virtue of what something represents whatever it represents namely, has
the (representational) content it has.
171
CRTT
The program
Since at least the demise of behaviourism, it has been taken for granted that
people and many animals exhibit all sorts of pa erns of behaviour that seem
explicable only on the assumption that they are capable of perceiving, remem-
bering, reasoning, planning, decision making, and, for some, systematically
expressing their thoughts in natural languages. There are two general proper-
ties of such thinking things: the transitions between their states are o en
rational, and they occur o en by virtue of the intentional content of the states.
Thus, people are sensitive to deductive, inductive and abductive relations
among their thoughts, and to decision theoretic pa erns in much of their
planning and behaviour, most of which can only be expressed by reference
to the truth-valuable contents of the constituent states. Its standardly truth-
valuable contents that are at least one of the relata of imply, infer, confirm
or rationalize, as when, for example, the hypothesis that it has rained is con-
firmed by the evidence of the wet streets, and the act of taking an umbrella
is rationalized by a desire that one not get wet.2
Dualists such as Descartes and Brentano have argued that rationality and
intentionality arent assimilable to a general physical theory of the world. How-
ever, many people have thought that recent advances in various formalizations
of reasoning and computation suggest otherwise, and so have pursued a CRTT.
Building on the proposals of Alan Turing regarding the nature of computation,
CRTT postulates that there exists a medium of formal representation (a lan-
guage of thought) and a set of computational processes defined over it that
could account for rational (and many irrational) phenomena.3 Since these
computations and representations could be physically realized in the brain,
it promises an answer at least to Descartes challenge that reason cant be
extracted from the power of ma er.
There is still the intentionality of the representations to be accounted for,
and it might be feared that CRTT depends too heavily on our understanding of
artifactual computers, whose intentionality, like that of a book, is derived from
the original intentionality of its programmers. To be sure, the plausibility of
CRTT is based partly from such an understanding, and frequently specific
CRTT proposals are tested by running them on artifactual computers. How-
ever, this reliance on artifactual machines is inessential. Indeed, CRTT departs
from current research in artificial intelligence (AI) in two important respects,
(a) it is not, as AI routinely is, concerned merely with understanding how to
get a computer to solve some problem in some way or other; rather, it seeks
to understand the actual inner workings (at a computational level) of humans
and animals, and (b) it decidedly does not suppose that human or animal
172
Representation
intentionality is derived from the intentionality of any programmers, but has

some other naturalistic source, as well discuss at length below fourth through
sixth sections).
As things stand, CRTT is not so much a theory as a research program. It
might be compared to the postulation of atoms as constituents of chemical
phenomena, without any clear idea of what the precise character of those atoms
might be. It leads us to ask interesting and o en empirically testable questions
about, for example, the precise character of the medium of representation its
expressive power, the kind of information it needs to represent as well as
about the character of the computations defined over those representations:
whether they are serial, parallel, connectionist or dynamic, and to what extent
they are modularized or encapsulated from one another. Discovering the subtle
principles and algorithms by which we understand the world and adjust our
behaviour to it is not something to be expected in our grandchildrens lifetimes,
if ever. But CRTT does seem to be the only serious framework in which these
issues can be raised.
(Non-)physical (non-)locality
Its important to appreciate both a serious problem that CRTT aims to solve and
a constraint on its solution. An organisms sensitivities begin to present dicul-
ties for any purely physicalistic proposal once one considers the detection of
non-local, non-physical properties by local physical agents. It is a fundamental
fact about the kinds of entities that include people and animals that their causal
interaction with their environment is entirely local and physical. In the case
of human beings this interaction is confined to six or seven sense modalities,
or transducer systems, that convert various physical phenomena into electro-
chemical impulses and motor systems that convert them back to bodily beha-
viour. As animals become more intelligent, the relations between these inputs
and outputs become increasingly complex, and it becomes increasingly hard
to explain how an animal or other device is capable of realizing them.
For example, successful models of animal navigation need to assume that
the animal can represent some of the spatio-temporal structure of that experi-
ence (e.g. how long it has been since it last found food at a given place). Indeed
as Gallistel notes:
[T]he representation of temporal intervals in rats and pigeons appears to be

a rich one, in the formal sense of rich. Successful computational models of
timing behaviour appear to require decision processes that employ operations
isomorphic to the addition, subtraction, and division of represented intervals.
(Gallistel, 1990, p. 586)
173
But how does an animal keep track of temporal structure in a way that
enables it to perform these computations and so sensitively modify its beha-
viour? What must its brain be like to do this? For starters, it had be er
somehow represent events as having specific temporal properties, and then
somehow store those representations for further use in combination with other
such representations in ways available for executing the vector algebra of dead
reckoning (Gallistel, 1990, 2008). Understanding mechanisms that could be
sensitive in these and other ways is one of the fundamental problems for any
psychology.
Moreover, human sensitivities to stimuli are not confined to non-local
properties. As contemporary psycholinguistic research shows, even very young
children seem to be sensitive to such categories as noun or verb that are in no
plausible sense physical properties of the stimuli (the properties, at any rate, do
not appear in physics or physiology). What kind of mechanism could possibly
produce such sensitivities? Noun phrases patently do not share any transduc-
ible property. It would seem impossible to rig a device to be sensitive to all and
only noun phrases in any way other than by in one way or another building into
the device the principles of grammar, and, with them, at least representations
of the relevant grammatical categories over which those principles are defined.
Similarly for the plethora of other non-physical properties to which people
are obviously sensitive (e.g. timidity, audacity, pomposity; being composed by
Beethoven, a fall in corporate profits, a proposal of marriage, or a declaration
of war). The only plausible way something could be sensitive to these things
is by having a mind, and the hope is that CRTT will explain how this helps
by showing how it emerges from computations over representations.
Turing and Marr
A crucial feature of Turings account of computation is that the transitions

between states turn on local physical properties (e.g. 0s and 1s that could
be individuated by their physical shape (or, in your modern computer, by
electro-magnetic properties). It is this property of Turings proposal that enables
it to avoid the familiar homunculus objections wouldnt one need a mind in
order to read the symbols? that have been raised to CRTT by Wi genstein,
Ryle and most recently, John Searle: the operation of the machine is brute
causal, not requiring any intelligence that it can thereby be enlisted to explain.4
By contrast, ge ing a machine to detect any other categories say, whether a
symbol was beautiful, terrifying, created by an Australian, or a representation
of justice would very likely require a mind, or another computer (i.e. some
procedure for proceeding from the detection of local, physical properties to
the target ones).
174
Representation
The pioneering work of David Marr (1982) on vision provides a paradig-

matic application of Turings idea, with a richness of ideas and degree of suc-
cess that has led it to become a paradigm for much work in the CRTT tradition.
Roughly, Marrs strategy was to explain visual recognition of familiar objects
in terms of computation of intensity values on the retina, proceeding to the
determination of continuous lines, or edges of an apparent object, the location
of this surface of this object in a two-dimensional grid, and then to computation
of conical shapes with which the original input is compared. The important
suggestion here is that, through mathematical computations of certain ideal-
ized figures, co-variant relations may be gradually be established with non-
local or non-physical properties such as being a rabbit or a chair. What promises
to make this account explanatory is the presumption that at every stage a com-
putation is being proposed on a physically specifiable representation whose
content, it is hoped, can also be specified by the theory.
Unfortunately, like virtually all psychologists, Marr focuses almost entirely
on the computational aspects of the processes. There is li le said systematically
about the issue of the content of the representations themselves. Indeed, it is
an odd and in some ways unfortunate sociological fact that by and large the
only cognitive scientists who have tried systematically to address this issue
are philosophers, and even they have paid comparatively li le a ention to
representation per se, as opposed to a itudes and intentionality generally. This is
particularly puzzling given the virtual ubiquity of the notion of representation
in psychology and cognitive science. Even though computations can be specified
and studied without appeal to content, content is arguably presupposed by any
computational theory: in vision, representations of, for example, edges and
objects; in decision making, of options, losses and gains; in parsing, of nouns
and verbs. Indeed, even the computations of Turing machines are standardly
defined over numerals that represent numbers. Yet the specific content of the
representations, and their relations to what they represent, is simply stipulated
or intuitively taken for granted.
Referential Opacity
The philosopher who brought the problems of representation and intentional-

tiy to modern a ention was the Austrian philosopher Franz Brentano who
noted some logical peculiarities exhibited by a very large class of mental
verbs, the so-called propositional a itude verbs, such as think, notice,
believe, desire (verbs that take a sentence complement a that . . . or a to
. . . clause as a direct object).5 These peculiarities are o en conveniently
referred to by the name of referential opacity, to bring out the contrast with
normal, referentially transparent verbs.
175
Most non-mental verbs are referentially transparent insofar as
(a)the terms involve a relation between real things: if x kills or kisses y, then
both x and y had be er exist: you cant kill/kiss something that doesnt
exist; this is contrast with a mental verb like think, since you can think
that x is angry without there being anything of which you think it: the
Greeks thought Zeus was o en angry, but there was no Zeus; and
(b) whatever objects are related by a transparent verb are so related no
ma er how the objects are named or described, so long as the names
of descriptions of them truly apply. Thus, if Oedipus killed Laius, then,
since Laius is, in fact, his father, it follows that Oedipus killed his father.
Or, to put it another way, if the expressions Laius and his father both
refer to the same thing, then they can be substituted for each other in a
normal, referentially transparent verbal contexts such as x killed y.
Again, this contrasts with a mental verb such as think: Oedipus can
think he killed Laius without thinking he killed his father; its the fact
that the one claim doesnt imply the other that gets him into trouble. By
contrast with kills, most propositional a itude terms such as x thinks
that p, are opaque, so to say; the light of reference doesnt shine
directly onto the referent, if any, of expressions in the direct object clause
of the term, but somehow bounces o the expressions themselves.
Now, curiously, representation seems to suer from one of these problems

but not the other. At least on one natural reading, it doesnt exhibit the resis-
tance to substitution: if Oedipus represents Laius then he represents his father,
whether he knows it or not (although see note 8). But, on the other hand,
it seems to exhibit resistance to existential generalization: from the fact that
Oedipus represents Zeus it doesnt follow that there exists something such that
Oedipus represents that thing, there being no Zeus. Like the propositional
a itudes, representation can be about nothing, at any rate nothing real (and
what else is there?).
To be as neutral as possible about the complex issues raised here, we need
to allow that there is a crucial ambiguity in ways of talking about what repre-
sentations represent. On the surface represent would appear to be simply a
two-place relation, as in:
(a)The word cats represents cats.
But this cant be quite right, since
(b)The word Zeus represents Zeus.
176
Representation
would then be false, for lack of Zeus: you cant bear a real relation to some-
thing that doesnt exist. But theres surely a reading of (b) that makes it true,
since, again, Zeus is not meaningless. Its merely empty (which Ill confine to
meaningful expressions).
So what does an empty term like Zeus represent? Well, its an interesting
fact that an almost universal response is that of Quines (1953a) fictitious
philosopher, McX: it represents an idea in your head, as many people
might be inclined to put it. But this is absurd, since (a) whatever else might
be in your head, there are certainly no bearded gods there; and, in any
case, (b) if Zeus is an actual idea in your head, then Zeus would turn out to
exist aer all!
There have been a wide variety of replies to this puzzle.6 Again, to maintain
neutrality between disputes, we should simply note that the word represent
(and, for that ma er, virtually any intentional idiom) seems to suer from a
systematic ambiguity along the following lines:
(REP) (a)If we are talking about a representation, x, of some real thing, y,

then we o en take x to represent that real thing, y; thus Nixon
represents the actual man Nixon.
(b)When there isnt, as in the case of Zeus, then we rely on talk
about the content of the expression y (which I will abbreviate by
placing brackets around an expression, e.g. [Zeus]).
The first usage might be called the existential, the second the (purely)
intentional usage of represent (and other intentional idioms).7 Ive expressed
the second, intentional use with deliberate vagueness. It would be tempting to
say so a purely intentional use of representation of y, for lack of any y, is
really about an intentional content. But this wouldnt be correct, since someone
thinking about Zeus and his philandering ways isnt thinking about the phi-
landering ways of an intentional content. Speaking more carefully, we should
say something such as: when x represents a y that doesnt exist, a person is
standing in the thinking relation to the content [y]; but this doesnt entail she
is thinking about [y]. Even in a purely intentional usage, thinking about y is
one thing; thinking about [y] quite another.
Arguably, psychology is concerned with pure intentional content, even in
cases that might be described existentially. Although its convenient to describe
navigating birds and bees as representing, for example, the azimuth of the sun,
(e.g. see Gallistel, 1990), this seems to carry the misleading suggestion that the
birds and the bees actually represent the sun as the sun, which is doubtful;
presumably they dont really have the concept [sun], and would react just as
well to lights in a planetarium.8
177
Providing an account of intentional content, what animals represent things

as, is, however, no easy ma er. There have been two main strategies, internalist
and externalist.
Internalist Strategies
One of the most natural ideas about meaning is that it is some sort of intro-
spectible idea inside ones mind and/or head, say, an image or a inclination
to make one inference rather than other.
Images and stereotypes
The idea that mental representations are images, or perhaps maps by which
we steer (Lewis, 1994) can be traced back to Aristotle (see Cummins, 1989).
There seems to be a presumption that imagistic representation is somehow
unproblematic: an image, such as a green triangular one, represents what it
resembles. As appealing as this idea has been, it wont get us very far. In the first
place, one doesnt find entities that are actually triangular and green in the
brains of people who think about green triangles. Perhaps talk of resemblance
is just a way of talking about a correspondence between features of neural
events and real world properties, but then the naturalization problem is the
problem of specifying why one correspondence rather than another provides
the correct interpretation: as Wi genstein (1953) noted, there are infinite
numbers of projections from a triangle to arbitrary objects in the world.
Moreover, even if some mental representations can be usefully thought of
as images, it is extremely implausible that all, or even a significant proportion
of representations can be. Images simply dont combine to produce logically
complex images. What image, for example, could represent a negative fact, as
in the thought that there are no green triangles? A green triangle with a black
X superimposed upon it? How is that to be distinguished from a unnegated
image of a triangle with a black X superimposed on it? And the problem, of
course, only gets worse when one considers conditionals, quantified sentences,
modal claims, and all the indefinitely complex thoughts that people are
manifestly capable of thinking (e.g. try forming a distinctive image for If not
every green triangle is inside a square, then either all the figures are illusory
or Im blind). And over all this there hovers the problem of abstract ideas, such
as those of uncle, number and justice; what are the distinctive images for them?
While imagistic representations might well play some role in some cognitive
processes, they simply seem inadequate for any serious logical thought.
Moreover, as we already noted, in a CRTT, computations are defined over local
178
Representation
physical properties, not over non-local properties such as distances between

points in an image. Sentences seem to be the only physically realizable objects
that begin to have anything like the requisite expressive and computationally
tractable properties.
Conceptual roles
One kind of proposal that has been immensely popular in the twentieth century
is that the meaning of a representation is determined by its epistemic (eviden-
tiary and logical), conceptual role in reasoning. Logical operators provide the
most plausible examples; thus, a certain expression # might plausibly mean
[and] in virtue of the fact that thinking P#Q tends to cause thinking P and
thinking Q, which in turn tends to cause thinking P#Q (cf. Peacocke, 1992).
However, extending the account past [and] merely to other logical particles
is dicult. There are substantial disputes about or and the law of excluded
middle, as well as about how to understand conditionals. Are we to take these
dierences to be dierences in the meaning of the logical words, or simply in the
theories that people have about them (cf. Williamson, 2006)? Moving away
from purely logical cases, the problem seems even more daunting. With a li le
imagination, it seems always possible to construct a story whereby someone
(particularly a philosopher) could reasonably deny a standard role for a con-
cept while still seeming to possess it, simply by having a suciently bizarre
theory about the world. Thus, some creationists seem to be denying that humans
are animals; some nominalists, that numbers are abstract; and some idealists
that tables are material objects. There would seem to be no epistemic connec-
tion so secure that, with a li le ancillary theory, someone couldnt break it and
yet still be competent with the relevant concept. So how could some epistemic
connection provide the meaning of a concept?
A further problem, stressed by Fodor and LePore (1992), is that conceptual
roles dont seem compositional in the way that content ought to be. The content
[pet fish] ought, a er all, to be some kind of compositional function of the con-
tent [pet] and the content [fish]. However, its not at all clear that the conceptual
role of [pet fish] is a compositional function of the role of [pet] and the role of
[fish]; there may well be inferences that are peculiar to pet fish that are not
shared by pets or fish alone (e.g. only pet fish live in bowls).
It is tempting to try to limit the relevant roles to certain constitutive infer-
ences. The most famous eort to do this was the verifiability theory of meaning,
which was an a empt to spell out the meaning of a concept in terms of the
sensory evidence that would confirm or disconfirm its application. It is widely
thought that such a theory encountered insuperable problems. For purposes
here the most serious were Quines (1953b) observations about confirmation
179
holism: claims are (dis)confirmed not individually, but only in conjunction with
a great many other claims, making it dicult to see how to isolate certain
epistemic connections as constitutive of the meaning of a specific claim. This
was the heart of Quines famous challenge to any theory of meaning, and to
eorts to distinguish analytic claims, such as that bachelors are unmarried,
whose truth seemed to be due to meaning alone, from claims that were simply
tenaciously believed.9
These observations about belief variability and confirmation holism have
led many advocates of a conceptual role semantics to the desperate measure
of semantic holism: all of a terms epistemic connections are constitutive of its
meaning. But this seems to make a havoc of psychology. It would entail that
any change in any of a persons beliefs would ipso facto be a change in the
content of all of their thoughts. Since by accretion, reasoning or forgetfulness,
our beliefs are constantly changing, no one would ever have the same thoughts
twice; indeed, one wouldnt remember anything with the same content. It would
be a cosmic coincidence if two people ever shared a belief, since to agree on
anything theyd have to agree on absolutely everything. There could be no
serious generalizations about mental states, not even of the well-confirmed
kind regarding visual illusions. This was perhaps no problem for Quine who
was a behaviourist and sceptical of intentional psychology. However, it is
a serious problem for post-behaviourist cognitive approaches such as that
of CRTT.
A general problem for internalism
In addition to Quines scepticism about distinguishing ma ers of meaning from

mere ma ers of belief, a general problem with any purely internalist theory
emerges from considering the observations of Kripke (1972/1980), Putnam
(1975), and Burge (1979) regarding the ways in which the meanings of many
terms depend crucially upon the environment of the agent. Putnam famously
imagines a planet called Twin Earth that is exactly like the earth in every
respect (including history) except for having a novel chemical compound,
XYZ, everywhere that the earth has H2O. He claims that, since in fact water
is H2O, there is no water on Twin Earth. Now consider some Earthling adult,
Sophie, who knows no chemical theory, and her twin, Twin Sophie, on Twin
Earth, and suppose (along the improbable lines of the story) that they are
molecule-for-molecule duplicates of each other. Certainly, there would seem
to be no internal psychological dierences between them. And yet, if XYZ is
not genuine water, then Sophie and Twin Sophie are not referring to the same
substance when they use the word water. (Perhaps a clearer example would
180
Representation
be their use of Aristotle: where Sophies use refers to the Ancient Greek
philosopher, Twin Sophies refers to Twin Aristotle!). If meaning is what
determines extension, then, as Putnam pithily put it, meanings just aint in
the head (Putnam, 1975, p. 227).
Whatever one thinks of Putnams examples, still a further problem can be
raised in terms of a general question for any theory that takes thought to
be formal symbol manipulation. Imagine a computer that used exactly the
same program one day to play chess, the next to fight a war; there would seem
to be no purely internal facts that would distinguish what it was representing
on the one occasion vs. the other. It would seem that something external to the
computer must determine the meaning of at least some of its representations.
Externalist Theories
Interestingly enough, the suggestion that meaning might be something external

seems to be every bit as natural as the idea that its internal. One of the oldest
theories of meaning is what is sometimes called the Fido/Fido theory, according
to which the meaning of a representation is the object for which it stands.
Obviously, so stated, the theory is false; for starters, theres the problem of
empty representations such as Zeus that we mentioned earlier. However, there
are sophisticated versions of it, ones that appeal to the actual causal history
of a representation; ones that appeal to co-variational relations between a rep-
resentation and phenomena in the world; and teleofunctional theories that
appeal to evolutionary selectional processes.
Historical causal theories
Critics of internalist semantics proposed an alternative semantic picture (few

claimed a serious theory here) whereby the reference of a token term is deter-
mined by actual causal chains linking a speakers use of a particular word to
users of that word to dub the thing (object, substance, kind) to which the speaker
thereby refers. Kripke (1972/1980) and Putnam (1975) vividly argued for such
an account of proper names and for natural kind terms. Thus, Kripke argued
that what determines the reference of a name such as Aristotle is a chain of
uses of the name that extends from present uses all the way back to Aristotles
original dubbing with the name, and Putnam suggests a similar story about
water that takes earthlings back to H2O and Twin Earthlings back to XYZ.
Although actual causal histories may well have some role to play in a theory
of reference, it is widely recognized that such histories are not in themselves
181
sucient for a naturalistic theory, since at every stage of such causal chains
there are events (such as ostensions, dubbings, communications, understandings)
that require intentional characterization. For example, they would seem to
require that the original dubber had one thing rather than another in mind on
the occasion of the dubbing. But what determines that it was the infant,
Aristotle, and not, for example, his nose, or the kind, Greek, human being, or
animal, that Aristotles parents had in mind? All these things, a er all, are
equally in the causal path described by the dubbers ostending finger. (This is
the qua problem, discussed at length in Devi and Sterlney, 1987.)
Co-variation locking theories
A natural answer to the question of what a dubber dubbed might be: whatever
kind of thing she would discriminate as that thing with the term; that is, what-
ever she would apply the term to, as opposed to everything she wouldnt.
Along these lines, a number of philosophers have considered ways in which
states involving token expressions might have, in addition to actual causal
histories, certain counterfactual dispositional properties to co-vary with certain
phenomena in the world. Intentional meaning is treated as a species of so-called
natural meaning, the kind of meaning that is said to obtain between dark clouds
and rain, red spots and measles, expansions of mercury in a thermometer and
ambient temperature. One event naturally means another if there is a causal
law connecting them, or, as Dretske (1981) put it, the one event carries informa-
tion about the other. A sentence, on this view, means what it carries information
about: the sentence its raining means that its raining, since it carries the
information (causally co-varies) with the fact that its raining.10
Nevertheless, so stated, the view is open to several immediate objections.
As stated, almost everything would mean something, since almost everything
is reliably caused by (and carries information about) something. So there must
be some further condition on genuine mental meaning. Here, most tokenings
of representations are produced in the absence of the conditions that they never-
theless mean: Thats a horse can be thought on a dark night in the presence of
a cow, or just idly in the presence of anything. In his influential discussions
Fodor (1987, 1991b) calls these la er usages wild; the property whereby
tokens of symbols can mean things that arent on occasion their actual cause
he calls robustness.
The problem for any co-variational theory is to account for robustness. In
doing so it needs to solve what has come to be called the disjunction problem:
given that among the causes of a symbols tokenings there both are meaning-
forming and wild causes, what distinguishes them? In particular, what makes
182
Representation
it true that some symbol F means [horse] and not [horse or cow on a dark
night], or [horse or cow on a dark night or w2 or w3 or . . .] (where each wi is
one of the purportedly wild causes)?11
Several proposals have been advanced for handling the disjunction problem.
They have in common trying to constrain the occasions on which the nomic
connection is meaning-forming.
Ideal co-variation
A natural suggestion is that the meaning-constitutive conditions are those that

obtain under some optimal conditions, conditions that obtain when nothing is
interfering with the belief formation system. One of the first such theories
was that of Stampe (1977), which was taken up by Stalnaker (1984) and inde-
pendently proposed (and then later rejected) by Fodor (1980/91, 1987). The
a raction of such a theory lies in its capturing the idea that two individuals
meaning the same thing by some symbol consists in their agreeing about what
it would apply to were everything else about the world known. Their disagree-
ments are to be explained as due to their limited epistemic positions and
reasoning capacities, which interfere with their being as omniscient as the right
conditions for agreement would require. Insisting on such distinctions is like
insisting on a distinction between guided missiles that end up at a certain
location because thats where they were aimed, from those that, aimed else-
where, ended up there because of an error in navigation. Wherever they
happen to end up, the missiles are locked onto a certain destination. Similarly,
terms have a certain meaning by virtue of being locked onto a certain phenom-
enon in the world.
Such theories in this way suggest a way of isolating semantic stability from
issues of epistemic dierences. In particular, a co-variation theory allows us to
capture what in the world an agent is ge ing at in her use of a symbol, isolating
that from her relative epistemic success or failure in reaching it. It thereby
provides a basis for psychologically important predictions about how an agent
will react to further evidence and argument, distinguishing rational from
merely verbal revisions of thought.
Although such theories might not be straightforwardly false, they do seem
to be subject to a number of diculties, the chief one consisting in the circular-
ity that seems unavoidable in specifying the optimal conditions: its hard to
see how to specify the conditions without employing the very intentional idiom
the theory is supposed to explain. For example, it is dicult to see how to rule
out the interference of other intentional states (e.g. the aforementioned bizarre
theories that are the bane of conceptual role accounts, or the cooperation
183
of certain intentional states, such as a ending to, thinking of, wanting to get
things right).
In order to avoid these problems Fodor (1987, 1990c) went on to propose
another kind of co-variation, what has come to be known as the asymmetric
dependency relation. Although it makes no explicit appeal to ideal epistemic
conditions, much of its motivation can be appreciated by thinking of the ideal
co-variational theory in the background.
Fodors asymmetric dependencies
According to the ideal co-variational theory, under epistemically ideal condi-

tions, tokenings of a predicate co-vary with the property it expresses. But, of
course, tokens of it may also be produced by many properties it doesnt express.
Tokens might be produced by things erroneously taken to have the property,
by things associated with the property, by mere thoughts about the property,
etc. Now, one way to understand the asymmetric dependency theory is first
to notice that, plausibly, all these la er cases depend upon the ideal case but
not vice versa. The wild tokenings depend upon the ideal ones, but the ideal
ones dont depend upon the wild ones (ge ing things wrong depends upon
ge ing things right in a way that ge ing things right doesnt depend upon
ge ing things wrong). Thus, the property horse causes a tokening of some
symbol C because some horses (e.g. those at the far end of the meadow) look
like cows and, under ideal conditions, cow causes C. Milk causes C because
seeing some milk causes one to think milk, and this reminds one of where milk
comes from, which, under ideal conditions, would be the sort of thing that
causes cow tokenings.
So formulated, of course, the account still mentions ideal conditions, and
these Fodor has conceded cannot be specified non-circularly. His further
audacious suggestion is that mention of the ideal conditions here is entirely
inessential: the structure of asymmetric causal dependency alone, abstracted
from any specific conditions or causal chains, will do all the required work!
Specially, all we need do is existentially generalize over the ideal epistemic con-
ditions, thereby avoiding the need to specify them. To a first approximation:
A representation R means [p] only if its a law that, under some conditions,
R is entokened if p, and all other tokenings of x asymmetrically depend upon
this law. Thus, C means cow only if, under some conditions, C is entokened
if theres a cow present, and tokenings of C when cows arent present asym-
metrically depend upon this law.12
An abundance of objections were raised to this theory in the 1990s (e.g. see
Loewer and Rey, 1991; Loewer, 1997) to which Fodor made abundant and
o en ingenious replies (see, e.g., see Fodor, 1987, 1991b). We shall confine our
184
Representation
a ention here to problems that arise for any externalist theories but only a er
se ing out another important class of them.
Teleofunctional theories
Millikan (1984), Papineau (1987), Dretske (1988) and Neander (2004) have been
working on a general account of meaning based upon the role mental states
play in a biological account of the evolution and life of an organism. On this
view, intentional states possess certain information-carrying, evolutionarily
determined functions, even if there are no conditions under which they pres-
ently execute them. These functions fix the states intentional content. Thus, a
frogs tokening F whenever a black speck crosses its retinal field might be
interpreted as meaning [fly], since its flies that are responsible for the survival
of frogs. Such teleological ideas can be combined with co-variational approaches.
Thus, Dretske (1988) proposes treating meaning as the recruitment of a co-
variation between an internal state and a worldly phenomenon on behalf of
some adaptive response. The fact of the co-variation itself (say, between a state
of the frog and the motion of a fly) is an explanatorily significant cause of the
frogs shooting out its tongue at such nutritious prey.
All such teleofunctional approaches face some general problems. A worry
that has been much discussed about selectionist theories in general is that they
are Panglossian in assuming that all useful traits were evolutionarily selected.
Critics of teleosemantics, like Fodor (1987, 1991b), argue that we have no reason
to think that large parts of thought and language, and their meaning, may not
be unselected eects of whatever was selected, and so of no help in determining
semantic content.
Indeed, certain properties of thought and language do seem to outstrip any
pressures from natural selection. Insofar, for example, as many of our concepts
involve commitments to full potentially infinite universal quantifications, it is
dicult to see how they, as opposed to more modest finite cousins, could have
been selected by a finite history (see Peacocke, 1992, p. 131).
But perhaps the most serious worry about such teleofunctional approaches
to content is that, given the vicissitudes of natural selection, the wrong contents
might get assigned to psychological states. Pietroski (1992) imagines a case
in which an animal gets a racted by red flowers on high hills and increases
its selectional advantage by avoiding predators in the low lying valleys. The
teleofunctionalist would appear to be commi ed to claiming that the state of
responding to the red flowers actually has something like the content [avoid
predator], even if the animal couldnt actually recognize a predator when face
to face with it. Standard selectionist explanations of intentional states tend to
presuppose the contents of such states and so would be unable to explain them.
185
General problems with externalism13
A general diculty that arises for any pure externalist account is the dis-
tinctions of the mind seem to outrun the distinctions that the external world
independently provides. There are not only distinctions among concepts (such
as [renate] and [cordate]) that happen in the real world to be co-extensive, but
there are distinctions among necessarily co-instantiated concepts (i.e. concepts
that are instantiated in all the same possible worlds and/or counterfactual
situations).
There are two kinds of such co-instantiated concepts: those expressed by
necessarily co-extensive terms, such as triangle and trilateral, eucalyptus
and gum, circle and point equidistant locus of co-planar points (whatever
satisfies one in each of these pairs necessarily satisfies the other); and what
might be called the necessarily co-divided ones, such as rabbit, vs. unde-
tached rabbit parts, vs. temporal stage of a rabbit. Dierent things satisfy
each of these three expressions, but whenever an agent is presented with some-
thing that satisfies one of them shes presented with something that satisfies
the others (cf. Quine, 1960, Chapter 2; Gates, 1996). Or consider simply the
phenomenon of subception, whereby many animals are able to recognize
groups of things of certain modest cardinality (Gallistel, 1990, Chapter 10).
Do such animals plausibly have the same concept [three] that people have?
Theres this reason to think not: unlike other animals, most people have a
concept controlled by general principles, for example, Peanos axioms for
arithmetic (e.g. zero is a number; every number has a successor), whereby we
can be led by reasonings into understanding a potential infinity of complex
arithmetic truths.
Recalling our early discussion of empty terms and purely intentional uses
of represent, a particularly crucial set of cases of necessarily co-instantiated
concepts are the necessarily uninstantiated ones (e.g. [largest prime], [round
square]). One way to think of dealing with such cases is to try dealing with
them as logically complex, so [largest prime] would actually involve a logically
complex construction about of symbols meaning [large] and [prime] (see Fodor,
1990c). However, there seem to be plenty of non-complex cases. As Plato pointed
out, no one ever perceives a genuine circle: all the figures we could possibly
encounter are only very crude approximations, at best ragged complex poly-
gons.14 Moreover, although laws may arguably relate properties that happen
to be non-instantiated (e.g. some specific mass that nothing ever happens to
possess), if one limits oneself to properties that genuinely figure in scientific
laws, it would be dicult to see how there could be laws about, for example,
unicorns, angels or ghosts. Its not at all clear that these things are even meta-
physically possible, much less suciently possible for there to be scientific laws
relating them to entokenings of representations.15
186
Representation
Still another problem for a pure co-variation theory is presented by concepts

involving varying degrees of response-dependency, such as [shameful], [funny],
[bizarre]. The problem here is that the same response-dependent concept
can cause its possessors to lock onto dierent phenomena. A er all, it is a com-
monplace that dierent people find dierent things funny, shameful, tragic,
bizarre; and this is not in all cases plausibly due to dierences in their epistemic
position. You and I may disagree about whats funny, well, just simplicter,
perhaps as a result of simply very brute dierences between our nervous
systems. Consequently, there is no reason to expect convergence even under
any (ideal) circumstances on which all other uses of such concepts asymmetri-
cally depend, despite our patently sharing them.
For all these reasons it would seem that any externalist theory will need to
be supplemented by some facts about a terms inferential role: square bears
a direct relation to four sides that circle lacks; undetached proper rabbit
part, but not rabbit play dierent roles in mereological inferences; [funny]
is tied to laughter whether or not people ascribe it to the same things. For all
the interest of externalist intuitions and the problems with internalist ones,
there seems to be something semantic in the head. So perhaps the internal and
externalist strategies ought to be combined.
Ecumenical Approaches
There are two ways to combine the two strategies: two factor theories and
appeals to basicality.
Two factor theories: narrow and wide content
Two factor theories posit both an internal and external factor to meaning.
One suggestion might be to allow the representational vehicle itself to be a
component of content. This suggestion seems a natural way of distinguishing
necessarily co-instantiated concepts that have dierent constituent structure.
Thus, the thought that water is wet is distinct from the thought that H2O is wet
by virtue of the fact that the content of the one thought involves a syntactically
complex expression, H2O, and the other doesnt.
However, such a suggestion by itself is unlikely to suce for all the cases;
a er all, there can be logically simple expressions names, simple predicates
that are necessarily co-extensive (Mark Twain/Sam Clemens, Zeus/Jupiter,
eucalyptus/gum), and although they might be associated with dierent con-
tents within the mind of a single person, it doesnt seem as though they should
always express dierent contents across people. Indeed, as the above examples
187
show, dierent people could have the same thought about Sam Clemens, Zeus
or gum trees without the vehicles of thought being actually spelt the same!
Intuitively, all that would seem to ma er is that the role of the vehicles in their
thought be the same.16 Consequently, some two-factor theorists (e.g. Loar, 1981;
Block, 1986a) suggest that the narrow component be identified with a terms
conceptual role, along the lines but also subject to the diculties sketched
above.
Another suggestion is to identify the narrow content with a rule in the
head that along the lines of Grice (1961/65) and Putnam (1975) may leave
some sort of blank space to be filled in by the specialist, a kind of indexical
element that permits a full semantic content to be determined by the context
with which the agent interacts, much as the semantics of indexical terms
such as I,now, this and that do (cf. Kaplan, 1989). White (1982) and Fodor
(1987) develop this strategy, generally identifying the narrow content of a
LOT expression with a function (in the set theoretic sense) that maps a context
onto a broad content. For example, the narrow content of Sophie and Twin
Sophies water is the function that maps Sophies context onto H2O and her
twins context onto XYZ. When Sophie u ers Water is wet, she thereby
expresses the content [H2O is wet], while when Twin Sophie u ers it she
expresses the content [XYZ is wet]. Two symbols have the same narrow con-
tent just in case they serve to compute the same such function: it is this that
is shared by Sophie and her twin. (See Chalmers (forthcoming) for further
development of this strategy.)
Of course, this strategy will be subject to the same Quinean worries we
raised earlier with regard to conceptual role theories: how do we distinguish
those roles that are essential and constitutive of meaning from those that are
mere ma ers of belief. A tentatively promising approach to those worries has
recently emerged in the work of Paul Horwich and Michael Devi .
Basicality?
Along lines strikingly similar to Fodors asymmetric dependency proposal,

Horwich (1998, 2005) proposes to treat meaning as the property of the use
of a word that is explanatorily basic: the one that best explains all the other
use properties of the term (Horwich, 1998, p. 41), and provides a number of
examples (see Horwich, 1998, pp. 45, 129):
The basic property for and is a tendency for x to accept p and q if x accepts
both p and q; for red: a disposition to apply red to an observed red surface; for
one: holding true Peanos axioms; for Aristotle: holding true This is Aristotle,
pointing to Aristotle.
188
Representation
Although one might quarrel with the examples and worry about the defla-
tionary context in which Horwich proposes his view, there seems to me some-
thing right about it which could be applied to CRTT, at a first pass, along the
following lines:
(BAS) The content of a representation is determined by the property of a

meaningful tokening of a term that is explanatorily basic: the one on which
all other tokens with that meaning asymmetrically/explanatorily depend
by virtue of that property.
Note that the explanatorily basic property need not be a purely internal one
but might well involve relations to external phenomena. That is, (BAS) has
Fodors proposals potentially as special cases, cases in which the basic proper-
ties are ones about actual language use, or about how symbols manage to be
locked onto actual phenomena in the world. (BAS) is simply not limited to
such cases.17
Although (BAS) is by no means a reduction of intentionality (intentional
notions are still mentioned in it), it still does some important work. Insofar as
basic properties are suciently local, it permits conceptual stability despite
wide epistemic and other sorts of surface variation in how people use words
and concepts. The issue is not whether people agree in their surface behaviour
but whether their responses are controlled by the same basic properties, an
issue not so easily addressed. And (BAS) allows for empty concepts such as
[unicorn] and [circle], and response-dependent ones such as [funny] and [good],
where the basic properties seem to be mostly in our internal responses, not
in the variable things to which we are responding.
(BAS) also concedes to Quine that there may well be no adequate way to
define theoretical terms such as electron or species, since the basic facts in
these cases may be precisely as theoretically diuse as Quines holistic view of
theoretical confirmation emphasizes. But on the other it may allow for some
local basic facts of the sort that seem to explain the intuitions about meaning
that people have about trivial cases such as bachelor. Thus, it seems to capture
not only internalist and externalist intuitions, but also what was always reason-
ably driving a Quinean scepticism about intentional content. If this is correct,
then it may well be at least the most ecumenical strategy to pursue in trying
to provide an adequate theory of the content of mental representations.
189
10 Mental Causation
Neil Campbell
In a seminal paper on the problem of mental causation Jaegwon Kim helpfully

characterizes the issue in terms of the following question: How is mental
causation possible given X? where X is an assumption we have some indepen-
dent reason to respect which makes mental causation prima facie problematic
(Kim, 1990a, p. 121). For the past forty years or so the formulation of Kims
question that has dominated the philosophical scene is the following: How
is mental causation possible given non-reductive physicalism? The widespread
commitment to non-reductive physicalism is due primarily to considerations
about multiple realization and mental anomalism, which I will treat as
working assumptions for the sake of this discussion. By denying that mental
properties are reducible to physical properties, however, it is unclear how the
mental can play a genuine causal role in the production of human physical
behaviour. For the most part this diculty has been articulated in one of two
ways. The first is driven by considerations about the nomological character
of causation and has been formulated primarily as a challenge to Donald
Davidsons version of non-reductive physicalism, anomalous monism (Davidson,
1970). The second, which is articulated in terms of exclusion pressures, is
broader in scope and is owed principally to Jaegwon Kim (1988, 1989a, 1990a,
1993a, 1994, 1998, 2005). What both versions of the problem share is the
conclusion that a robust account of mental causation seems impossible if
we deny that mental properties are reducible to physical properties. Indeed,
both lines of argument purport to show that non-reductive physicalism leads
to type epiphenomenalism, the causal inecacy of mental properties.
My goal in the sections to follow is to sketch out these two versions of the
problem and to explore some ways of dealing with them. My discussion is
divided into three sections. In the first section I outline anomalous monism and
the objection that it entails epiphenomenalism. In the second section I provide
a sketch of the exclusion argument against non-reductive physicalism. In the
third and final section I show that both arguments against non-reductive phys-
icalism rely on questionable metaphysical assumptions about the nature of
events that render them either misguided or question-begging.
190
Mental Causation
Anomalous Monism and the Problem of Mental Causation
In his paper Mental Events Davidson (1970) sought to reconcile three claims
that appear to be true yet seem to be mutually inconsistent:
1. At least some mental events interact causally with physical events.

2. Events related as cause and eect fall under strict laws.
3. There are no strict psychophysical laws.
The apparent inconsistency is that the truth of (1) and (2) entails the falsity
of (3). If a mental event such as my deciding to close the door causes the door
to close then (2) seems to imply that there ought to be a law connecting my
deciding to close the door and its closing, but this is just what the third claim
denies. Davidsons method of reconciliation involves a particular understand-
ing of the second claim. According to Davidson, when events stand in a causal
relation they have true descriptions that instantiate a strict law; not every true
description of the events is amenable to the formulation of such laws. In fact,
given the holism of the mental and the rational principles that guide mental
ascription Davidson argues that mental vocabulary is unsuitable for the for-
mulation of strict laws. Since only physical predicates are appropriate for the
formulation of strict laws and mental events enter causal relations with phys-
ical events, it follows from (2) that mental events have physical descriptions
and hence, are themselves also physical events. Since there are no strict psy-
chophysical laws mental concepts cannot be reduced to physical concepts,
so we have an ontological reduction of mental to physical events without a
conceptual reduction of mental to physical properties. Since it is individual
events and not properties that are identified, Davidsons anomalous monism is
a token rather than type identity theory.
Davidsons brand of non-reductive physicalism would seem to provide a
simple and elegant account of mental causation. Mental events such as deci-
sions or choices cause other events, including physical events, because they are
themselves physical events. A number of critics (Honderich, 1982, 1983, 1984;
Hess, 1981; Stoutland, 1976, 1980, 1985; Kim, 1989a, 1993a; Antony, 1989), how-
ever, have argued that anomalous monism entails the inecacy of mental prop-
erties and consequently fails to provide an adequate account of mental causation.
Although this argument takes many forms, the basic reasoning is roughly as
follows.1 Davidson faces a dilemma when it comes to the issue of mental causa-
tion. At the heart of the dilemma is the observation that we ordinarily distin-
guish between the properties of an event or object that are causally responsible
for the production of a given eect and those that are irrelevant. For example,
if I throw my glass on the floor and the impact of the glass against the concrete
191
causes the glass to sha er, some of the properties of the cause seem to ma er
and some dont. The fact the glass was blue and that it contained water seem
peripheral to the sha ering, whereas the velocity at which the glass was travelling
when it struck the floor, the angle of impact and the structure of the glass seem
more important. When it comes to identifying the law that connects events like
the first with events like the second it seems only natural to suppose that the la er
rather than the former properties will be implicated. That is, there is more likely
a law connecting the structure, velocity, and angle of impact of the glass with
its breaking than one couched in terms of the colour and contents of the glass.
Since mental events are, according to Davidson, identical to physical events,
it seems that mental events have both mental and physical properties. Given the
example of the sha ering glass it is reasonable to suppose that when there
is causal interaction between a mental event and a physical event we should
be able to identify which properties of the mental event enabled it to play the
causal role it did. This is o en expressed in the form of the question, Is it
the mental event as mental (i.e. in virtue of its mental properties) that causes
behaviour, or is it the mental event as physical that has causal ecacy?
Davidsons claim that the only strict laws there can be are physical laws sug-
gests that it is in virtue of the events physical properties that it caused what
it did. That is, if it is in virtue of the law-engaging physical properties of the
impact of the glass against the floor that caused the glass to sha er, by analogy
it seems reasonable to suppose that it is in virtue of the law-engaging physical
properties of a mental event that the event caused what it did. While this is
consistent with Davidsons three claims his critics think this falls too far short
of a robust account of mental causation. For what this first option means is that
mental events cause behaviour solely in virtue of their physical (i.e. neuro-
biological) properties. That is, when I decide to get up for a drink and then rise
from my chair my rising is caused by the event that was my deciding, but my
behaviour is not caused in virtue of the fact that the cause was a deciding, was
a desire for a drink, or in virtue of any of its mental properties. These are all
as irrelevant to the production of the eect as the colour and contents of the
glass were irrelevant to its sha ering. This hardly seems like mental causation
anymore. As Jerry Fodor once famously said,
If it isnt literally true that my wanting is causally responsible for my reaching,

and my itching is causally responsible for my scratching, and my believing
is causally responsible for saying . . . if none of that is literally true, then
practically everything I believe about anything is false and its the end of the
world (Fodor, 1990b, p. 156).
If mental events cause only in virtue of their physical properties then it

isnt literally true that Fodors wanting is causally responsible for his reaching;
192
Mental Causation
certain physical properties of the event are causally responsible for his reach-
ing. For many of Davidsons critics this is not good enough.
Since the first option is not very appealing what if one argued for the other,
according to which mental events cause in virtue of their mental properties?
This would certainly address Fodors concern, for then his wanting would
literally be causally responsible for his reaching because it would be in virtue
of the fact that his wanting has the mental properties it does that he in fact
reaches. However, this option fares no be er for the Davidsonian because it
entails psychophysical laws. This is because, as we saw earlier, it seems that
events cause in virtue of their law-engaging properties. If a mental event
causes in virtue of its mental properties, then this reintroduces psychophysical
laws, contradicting Davidsons third claim. Worse still, the reintroduction of
psychophysical laws revives the possibility of psychophysical reduction. So
although claiming that mental events cause in virtue of their mental properties
might provide a robust account of mental causation, it does so at the cost of
mental anomalism and, at least potentially, of non-reductive physicalism itself.
Neither of the two options then, is a ractive to the non-reductive physicalist,
in which case it looks like Davidsons brand of non-reductive physicalism
stumbles on the question of mental causation.
Mental Causation and the Exclusion Principle
The second formulation of the objection to non-reductive physicalism is based

on Kims principle of causal-explanatory exclusion and is intended to be more
far-reaching than the argument just examined, for Kim thinks the exclusion
argument is a problem for any version of non-reductive physicalism, not just
Davidsons. At the heart of the argument is Kims principle of explanatory
exclusion, which states that there can be no more than a single complete and
independent explanation of any one event (Kim, 1988, p. 233). Kim considers an
explanation to be a complex of statements that can be divided into explanans
and explanandum propositions, where the explanandum is the proposition in
need of explaining and the explanans the proposition that does the explaining.
However, just because explanations are defined in terms of propositions this
does not mean one should think of them as arguments or as logical derivations
with the explanandum as the conclusion in the way Hempel did (Hempel, 1963,
1965, 1996; Hempel and Oppenheim, 1953). In Kims view Hempels approach
leads to what he calls explanatory internalism or explanatory irrealism
because the focus on logical or derivational relations between propositions
comes at the cost of neglecting the relations between the events in the world
that the propositions are about. Kim prefers a more deeply externalist account
of explanation that is grounded not in relations between items in our epistemic
193
corpus but in events and relations in the world. For this reason he adopts what
he calls explanatory realism, which claims that a proposition C is an explanans
for E in virtue of there being some determinate relation R holding between
events c and e. Kim, then, takes R to be the explanatory relation that grounds
the explanans relation between propositions C and E.
On the realist view, our explanations are correct or true if they depict these
relations correctly, just as our propositions or beliefs are true if they correctly
depict objective facts; and explanations could be more or less accurate
according to how accurately they depict these relations. Thus, that c is related
by explanatory relation R to e is the content of the explanation consisting of
C and E; it is what the explanation says (Kim, 1988, p. 226).
Since the most prominent species of explanation is causal explanation,

Kim claims that the most plausible candidate to fulfil the role of R is the causal
relation itself in such cases. Hence, he thinks that explanatory realism entails
causal realism, the view that causal relations are mind-independent relations
between events in the world2 and that every event has a unique and determinate
causal history whose character is entirely independent of our representation of
it (Kim, 1988, p. 230). Bringing explanatory realism and causal realism together
Kim defines having an explanation in terms of the possession of causal know-
ledge: To have an explanation of event e in terms of event c is to know, or
somehow represent, that c caused e (Kim, 1988, p. 230).
Kims commitment to realism also leads him to locate the individuating
role of explanations in the explanatory relation itself:
Explanatory realism yields a natural way of individuating explanations:

explanations are individuated in terms of the events related by the explan-
atory relation (the causal relation, for explanations of events). For on realism
it is the objective relationship between events that ultimately grounds
explanations and constitutes their objective content. This provides us with a
basis for regarding explanations that appeal to the same events standing in
the same relation as giving, or stating, one explanation, not two, just as two
inequivalent descriptions can represent the same fact (Kim, 1988, p. 233).
This is a fully extensional view of explanation. Logically inequivalent descrip-

tions of the cause or the eect in explanatory claims will state the same explana-
tion since they are grounded in the same metaphysical relation R between the
same events.
Kims explanatory realism plays a central role in his justification for the
principle of explanatory exclusion. He asks us to imagine that we have two
causal explanations for the occurrence of a single event e, one in terms of c1 and
194
Mental Causation
another in terms of c2. If we explore the various ways in which c1 and c2 might
be related, it turns out that the explanations fail to be complete or independent.
Kim identifies six possibilities: (1) c1 is identical to c2, (2) c1 is distinct from c2 but
is reducible to or supervenient on it, (3) c1 and c2 are both partial causes of e,
(4) c1 is a proper part of c2, (5) c1 and c2 are dierent links in the same causal
chain leading to e, and finally, (6) e is causally overdetermined by c1 and c2.
There is no need to discuss all of these options. It is clear that if c1 and c2 are both
partial causes, then neither event is sucient on its own for the eect, and so
according to Kim an explanation that appeals to either cause alone will be
incomplete because it leaves out a central causal factor. Similarly, if c1 and c2 are
sequential links in a causal chain, then the explanation in terms of c2 fails to
be independent of the explanation in terms of c1 in virtue of the dependence of
c2 on c1. Hence, Kim plausibly assumes that if the events referred to in two
explanations are not independent of one another, then the explanations them-
selves also fail to be independent. The only time there can be two complete and
independent explanations of the same event is option (6), when the event is
causally over-determined (i.e. when there are two independent causes, each
of which is sucient for the eect). Kim admits this possibility but claims that
genuine cases of over-determination are suciently rare, in which case the
principle of explanatory exclusion is a plausible general principle.
Kim has used the exclusion principle to place considerable strain on the
concept of mental causation. Suppose that George rises from the couch. On the
one hand we have an explanation for his rising that appeals to the instantiation
of a mental property, such as his desire for a beer; on the other hand, since his
rising is a physical event it seems to have a purely physical explanation in terms
of a neurobiological property. This la er claim implicitly appeals to the causal
closure of the physical domain which states roughly that for the occurrence
of any physical event there is a physical cause which is sucient for it.3
Kim claims we should find something puzzling about having both of these
explanations for why George rises from the couch:
When these two claims are viewed together, we should find the situation
perplexing and somewhat unse ling . . . We want to ask: Which really did
it? Whats the real story? The premises of the two causal explanations are
mutually consistent; however, there is something perplexing and perhaps
even incoherent about accepting both as telling us what caused Georges
behaviour, without an account of how the two accounts are related to each
other. Each explanation specifies a cause of Georges behaviour. But how
are the two supposed causes related to each other? (Kim, 1990a, p. 125)
Kim then surveys the six possibilities mentioned earlier and shows that none of
these is very promising. He claims the possibility that the mental and physical
195
causes of Georges behaviour are each partial causes, distinct links in the same
causal chain, or proper parts of the same cause are highly implausible, and I
agree. Over-determination is not an option either because this would require
an unexplained coincidence of causes that is systematic. Occasional cases of
causal over-determination can be tolerated (e.g. the smouldering cigar and the
lightning strike simultaneously ignited the haystack) because circumstances, as
unusual as they may be, can lead to a coincidental convergence of sucient
causes. Appealing to over-determination to explain mental causation requires
that all intentional actions are systematically over-determined by mental and
physical causes. While this is not impossible, it is not a very a ractive option
since there is something troubling about the idea of systematic coincidences.
The only live options for Kim are identity and supervenience.
Although Kim was at one time optimistic about using supervenience to
account for the relation between intentional and neurobiological explanations
(see Kim, 1984) he points out that such an approach faces some serious
problems. The trouble is that in order to use supervenience to show that one
explanation depends on the other one must oer a characterization of the
supervenience relation between mental and physical properties that captures a
suciently robust notion of dependence, and this has been lacking in standard
accounts. Weak and global supervenience are, according to Kim, too weak,4
and strong supervenience arguably implies reduction, which is precluded by
the non-reductive physicalist. Hence, the only real hope for a solution lies in
the identification of mental with physical properties. Indeed, this is precisely
what Kim argues in his most recent work (Kim, 2005), but this is once again to
give up on non-reductive physicalism and espouse a version of reductionism.
These considerations show that the non-reductive physicalist lacks an appro-
priate account of how the two explanations of Georges behaviour are related,
in which case the neurophysiological explanation excludes the intentional
explanation. This puts the legitimacy of all psychological explanations in jeop-
ardy, for the above line of reasoning generalizes to every case where human
beings seem to act for reasons.
The Metaphysics of Events
So how can the non-reductive physicalist respond to either or both of these

objections?
Some authors (Pereboom, 2002; Pereboom and Kornblith, 1991) have appealed
to the idea that mental properties are constituted out of their physical base
properties and thereby inherit their causal ecacy, though the relevant notion
of constitution is not as clear as one might like.5 Others (Noordhof, 1999b; Block,
196
Mental Causation
2003; Gulick, 1992; Bontly, 2002; Gille , 2001; Burge, 1993b; Menzies, 2003) have
argued that Kims reasoning can be generalized to all irreducible supervenient
properties in which case there is no causation anywhere but at the fundamental
physical level, which is absurd. Another reply is to acknowledge the possibility
that mental properties over-determine their eects, though few seem to take
this idea very seriously (Bontly, 2005; Ezquerro and Vicente, 2000; Vicente, 1999;
Kallestrup, 2006; Sparber, 2005; Raymont, 2003). Finally, some authors (LePore
and Loewer, 1987) place their hopes in the idea that mental properties can be
shown to be causally relevant to physical causation and that this relevance,
though not the same as ecacy, is robust enough to rescue mental causation.
In the remainder of this discussion I would like to explore an alternative
approach. I will suggest that both formulations of the objection share meta-
physical assumptions about the nature of events that render the objections
either incoherent or question-begging, in which case non-reductive physicalism
is not as implausible as the arguments make it seem. Since the form of this
response is slightly dierent depending on the target argument, I will discuss
each in turn, but since both responses rely on the same general idea, the second
will build on the first.
The reply to the argument against anomalous monism was actually provided
by Davidson himself (Davidson, 1993), though he did not make it as perspicu-
ous as one might like. Davidsons main line of response was to claim that given
his view of events and causation the objection raised against him makes no
literal sense. Featuring heavily in this reply is his claim that causation is an
extensional relation between events.
If causality is a relation between events, it holds between them no ma er

how they are described. So there can be descriptions of two events (physical
descriptions) which allow us to deduce from a law that if the first event
occurred the second would occur, and other descriptions (mental descriptions)
of the same events which invite no such inference. We can say, if we please
(though I do not think this is a happy way of pu ing the point), that events
instantiate a law only as described in one way rather than another, but we
cannot say that an event caused another only as described. Re-describing
an event cannot change what it causes, or change the events causal ecacy
(Davidson, 1993, pp. 67).
In Davidsons view, treating causation as an extensional relation means that

causes operate independently of the way we describe or classify them. Thus,
not only is it irrelevant to the causal powers of an event that we can describe it
using mental vocabulary, it is also irrelevant to the causal ecacy of physical
events that they can be described in the physical vocabulary. It is events that
197
have the power to change things, not our various ways of describing them
(Davidson, 1993, p. 12). This means there is no room for the idea of an event
causing as mental or as physical or in virtue of its mental or physical proper-
ties; events, not their properties, cause, and to be a mental or physical event
just is to be described using mental or physical vocabulary. Hence, Davidsons
view of events as concrete particulars and his nominalism about properties
prevent the epiphenomenalist objection from ge ing o the ground. The entire
objection depends on the seemingly innocuous assumption that events have
properties and cause what they do in virtue of some subset of those properties.
But Davidsons metaphysics is incompatible with this assumption because
events are simple entities that dont have properties as constituents; properties
are instead simply ways of describing events, and if events dont care how
we describe them, as his extensionalist thesis claims, it is hard to see how one
could claim that events cause in virtue of either their mental or their physical
properties. Neither option is possible, yet it was by forcing one option or the
other that the epiphenomenalist objection got going in the first place.
So Davidsons response, with which I am sympathetic, is that the objection
is formulated on the basis of a certain view of the metaphysics of events and of
causation that is foreign to his philosophy.6 The alternative view treats events
as property exemplifications, according to which an event is a structured com-
plex of which some property is a part. This is not to say that this alternative
metaphysics is false or should be rejected; indeed, there is something very
a ractive about the account of events and causation assumed by Davidsons
critics.7 The point is that it is illegitimate to import these assumptions into
the argument against Davidsons position, unless, that is, one can show that
Davidson himself accepts these assumptions. Since it is quite clear that he does
not, the objection takes aim at a position that bears only a faint resemblance
to Davidsons own views, and hence misses the mark. That this is the case is,
I think, made quite plain in the following rejoinder from Kim:
The issue has always been the causal ecacy of properties of eventsno
ma er how they, the events or the properties, are described. What the critics have
argued is perfectly consistent with causation itself being a two-termed
extensional relation over concrete events; their point is that such a relation
isnt enough: we also need a way of talking about the causal role of properties,
the role of properties of events in generating, or grounding, these two-termed
causal relations between concrete events. (Kim, 1993a, p. 21)
While it might be true that what many of Davidsons critics want is a way of
talking about the causal role of properties, Kim continues to assume that this
way of speaking about causation and about properties makes sense within a
198
Mental Causation
Davidsonian framework when he suggests in the above passage that properties

themselves can have multiple descriptions. If one thought of properties as real
entities this makes a certain amount of sense, but if one adopts Davidsons
nominalism about properties, according to which a property is just a way of
describing an event, it is hard to see how descriptions themselves can be
redescribed. The dispute about anomalous monism then, is largely a product
of underlying metaphysical views about the nature of events and causation.
Until one can show that anomalous monism leads to epiphenomenalism on
Davidsons own terms (i.e. within his metaphysical views of events and causa-
tion), it seems to me that the critics are wasting their breath. As we shall see,
similar issues also complicate the version of the problem of mental causation
that appeals to the principle of exclusion.
To make my concerns about the argument from exclusion clear I need to
elaborate a li le more on Kims alternative theory of events. Kim (1969, 1973,
1976) has long advocated a property exemplification theory. According to Kim
events are structured complexes with three basic types of constituents. First,
each event involves an object since events usually denote something undergoing
a change or alteration. Second, since events are thought of as occurrences, each
event also has a time at which it occurs or over which it endures. Third, each
event has what Kim calls a constitutive property since in order for the relevant
object to undergo a change there must be a modification of its properties.
According to Kim constitutive properties are among the important properties,
relative to . . . [an explanatory] theory, in terms of which lawful regularities can
be discovered, described, and explained (Kim, 1976, p. 37). The canonical
description of an event then, takes the form [x, P, t] where x is the constitutive
object, P is its constitutive property, and t is the time at which the event occurs.
The constitutive property, which is exemplified by the constitutive substance,
determines the generic event under discussion and is distinguished from other
properties the event (as opposed to the substance) exemplifies. For example,
Kim says that Socrates dying at t is a constitutive property of the event [Socrates,
dying, t] and that occurring in prison is merely exemplified by the event, though
not constitutive of it (Kim, 1973, 12). Consequently, Kim is careful to point out
that not every description of an event tells us about its constitutive elements.
These observations are important because unless we are aware of all of an
events constitutive elements we will be unable to distinguish it from other
events.
In contrast to the Davidsonian model of events, the property exemplification
model is a fine-grained account because each event possesses exactly one
constitutive property only, whereas Davidson does not distinguish between
constitutive properties and the various properties an event exemplifies. On
Kims view then, if two properties F and G are tokened at the same time by the
199
same object, but F G, then the tokening of F and the tokening of G are distinct
events. This is made quite explicit in Kims Identity Condition for events:
[x, P, t] = [y, Q, t] just in case x = y, P = Q, and t = t. (Kim, 35)
This identity condition is behind Kims well known disagreement with

Davidson about whether Brutuss stabbing Caesar is a distinct event from his
killing Caesar. Davidson treats these as alternative descriptions of a single event
whereas Kim distinguishes these as distinct events because the property of
stabbing cannot be identified with the property of killing since some stabbings
are not fatal.
While there are certainly many interesting questions about how one should
distinguish the constitutive properties of an event from other properties the
event exemplifies and about how to individuate events, my interest in Kims
theory of events concerns its implications for his criterion of individuation for
explanations. What I want to suggest is that there is reason to suspect the exclu-
sion principle is an implication of Kims account of events. To see how this is
so imagine that the principle of explanatory exclusion were false, such that
there could be multiple explanations for a single event. One way this could
happen, as several authors have suggested (Marras, 1998; Campbell, 2007,
2008a; Campbell and Moore, 2009; Raymont 2003), is to adopt what is o en
called the dual explanandum reply to the exclusion argument. According to this
reply a single event can generate multiple explananda by tokening more than
one property at the time in question. Hence, there can potentially be as many
explanations of an event as there are facts about it or properties it tokens. So, for
example, if a single event such as Georges rising from the couch simultane-
ously tokens the property of being an intentional action and the property of
being bodily movement of a specific type, then we can have more than one
explanation for the occurrence of that event qua the tokening of one property
or the other. Hence, one explanation might appeal to Georges desire for a beer
while the other might appeal to a neurobiological property. Relative to how
the event is described each explanation is complete within its own domain.
The thing to notice about this possibility, however, is that it is precluded by
Kims identity condition for events. According to Kim we must treat the token-
ing of Georges desire for a beer and the tokening of bodily movement of a
specific type as distinct events unless we can identify the two properties in
question. The trouble is that according to the non-reductive physicalist this
identification is not up for grabs. By Kims identity condition this means that
the events must be distinct and so this cannot be a case of a single event having
more than one explanation. For since Kims causal realism claims that each
event has a unique causal history we have to assume that Georges rising to get
200
Mental Causation
a beer has a distinct cause from his bodily movement. Kims identity condition
renders the dual explananda strategy impossible, for any a empt to ground
multiple explanations of a single event in distinct properties, the event tokens
will run afoul of Kims identity condition for events. Thus the a empt to show
that a single event can have more than one explanation by fragmenting it into
multiple explananda (according to which property is tokened) has no prospect
for success within Kims metaphysics. Such an approach could only succeed on
a Davidsonian course-grained account of event identity.8 But this means there
is an important sense in which Kim has begged the question in his use of the
principle in debates about mental causation.9 The principle obviously holds
for someone who, like Kim, accepts a fine-grained theory of events, but there
are many who prefer a course-grained Davidsonian approach, and there is cer-
tainly much room for doubt about whether or not the exclusion principle holds
under the conditions of this alternative metaphysics.10 Thus, by making it seem
as though the principle of explanatory exclusion holds regardless of ones
metaphysical theory of events, Kim does the issue a serious disservice.
If I am correct that Kim and other critics of non-reductive physicalism have
assumed something like Kims property exemplification view of events, this
goes a long way to discrediting the two arguments surveyed in my discussion.
However, there is an even more serious concern here, at least about Kims use
of exclusion to argue against non-reductive physicalism. Since the argument
relies on Kims theory of events it is entirely question-begging. This is because
Kims version of the property exemplification theory of events already assumes
the falsity of non-reductive physicalism.
As we saw, a central claim of Kims theory of events is the identity condition,
which he uses to individuate events. The identity condition states that two
events are identical if and only if their constitutive elements are identical. That
is, for event [x, P, t] to be identical to event [y, Q, t], x must be identical to y,
so we have here the same constitutive object, t must be identical to t, so the
events occur at the same time and have the same duration, and property P must
be identical to property Q. So if P is a physical property and Q is a mental prop-
erty, on Kims schema the mental event [y, Q, t] can be identical to the physical
event [x, P, t] only if we can identify the mental property Q with the physical
property P. This precludes the very possibility that defines most forms of non-
reductive physicalism, namely, that mental events are physical events but that
mental properties cannot be identified with physical properties. This means
that if an argument against non-reductive physicalism assumes Kims theory of
events the argument begs the question, for the property exemplification theory
already assumes non-reductive physicalism is false. To the extent that Kims
version of the argument from exclusion depends on the property exemplifi-
cation theory then, the exclusion argument begs the question.11
201
The lesson the above observations hold for those concerned about the
problem of mental causation seems to be that the problem cannot be isolated
from metaphysical questions about the nature of events and the role of proper-
ties in their individuation. Since ones assumptions about such ma ers can
have a profound eect on ones treatment of the problem of mental causation,
it seems only prudent to clear such ma ers up first, or at least to be forthright
about them from the start.
202
11 Personal Identity
E. J. Lowe
Why Should Personal Identity Be Philosophically Interesting?
It may be wondered why a chapter on personal identity belongs in a volume on

the philosophy of mind rather than in one on metaphysics. The answer is that
the topic belongs to both branches of philosophical inquiry: to metaphysics
because the notion of identity is a central one in that domain and to the philo-
sophy of mind because persons are prime examples of minded beings. However,
it might be supposed that, since the notion of identity is a universal one, there
can be nothing special to say about personal identity as such, beyond saying that
it involves the application of this notion to minded beings of a certain kind.
Some philosophers would undoubtedly agree with this view. They would urge
that the theory of identity, if indeed it deserves to be dignified by the title
theory, is exhausted by an account of the logical properties of the identity
relation, which reduces to the fact that it is a reflexive relation that is governed
by Leibnizs law or, more precisely, by the principle of the indiscernibility of
identicals. This is just the principle, taken to be a necessary truth that things that
are identical share all their properties, or, rather more cautiously expressed,
in a way that doesnt presuppose the existence of properties: that whatever is
true of something is true of anything identical with that thing. If that view were
correct, then there would be nothing to be said about personal identity beyond
the banality that persons, like anything else, can be said to be identical only if
they are indiscernible from one another.
Thus, for example, by this account, there is nothing more to be said regard-
ing the hypothesis that I am identical with, say, Napoleon than that it is true
only if I dier from Napoleon in no discernible way. Of course, this provides us
only with a logically necessary condition for the truth of that hypothesis, not
a logically sucient one. However, if the converse principle of the identity of
indiscernibles is also accepted, then even this deficiency is remedied and our
hypothesis may be judged to be true if and only if there is no discernible dier-
ence between Napoleon and me. It might be supposed that this is then the
end of the ma er, since it is just obvious isnt it? that there are indeed dis-
cernible dierences between Napoleon and me, such as that he won the ba le
of Austerlitz but I did not. But why should anyone be so confident that I didnt
203
win the ba le of Austerlitz? The reply may be oered: because I obviously

didnt even exist at the time of that ba le. But why should anyone be so confi-
dent of that? It can only be because something is being presupposed about the
nature of persons which constrains the possibilities of identifying one person
with another, such as that I cant be identical with a person none of whose
experiences I can remember having, or with a person whose body was destroyed
before the body that I have now was created. It is presuppositions like these
that make it seem obvious that I cant be identical with Napoleon, but they
have nothing to do with Leibnizs law as such, since they relate specifically to
the presumed nature of persons, as opposed to things of various other kinds.
This shows, then, that much more needs to be said about personal identity
than can be captured simply by applying the logical properties of the identity
relation to the particular case of persons. Specifically, what is needed is a prin-
cipled account of the identity-conditions of persons, or, to use John Lockes
helpful phrase, an account of what their identity consists in. In modern par-
lance, what we must endeavour to establish is a criterion of identity for persons.
And, as Locke himself insisted, this will require us to provide an account of
what persons essentially are. As he succinctly puts it: This being premised to
find wherein personal identity consists, we must consider what Person stands
for (Locke, 1975 [1690], II, XXVII, p. 9). Famously, his own answer to this la er
question is that a person is a thinking intelligent Being, that has reason and
reflection, and can consider it self as it self, the same thinking thing in dierent
times and places (Locke, 1975 [1690]: II, XXVII, p. 9). However, before we can
examine this and other proposals concerning the nature and identity-conditions
of persons, we should step back to take a wider view of the kind of the enter-
prise that we are embarking upon, by looking more closely at the notions of
identity and criteria of identity.
Identity and Identication
It seems evident that the expression is identical with, symbolized in logic

and mathematics by the equality sign, =, is a relational expression and hence
denotes a certain relation in which things can stand to one another. However, if
so, then it is a very peculiar relation, in that it can never literally hold between
one thing and another thing, but only between a thing and itself. Other relations
can, of course, hold between a thing and itself, such as the relation of admiring:
someone can obviously admire him or herself. But this relation can also hold
between dierent things, as when Peter admires Jane. Identity is peculiar as
a relation in that it necessarily holds only between a thing and itself and,
indeed, this has led some philosophers to deny that it is really a relation at all.
However we classify it, though, it can certainly seem strange. Since everything
204
Personal Identity
is identical with itself and with no other thing, one might wonder how facts of
identity can fail to be u erly trivial and uninteresting.
Part of the solution to this conundrum is provided by distinguishing, as we
must do anyway, between identity and identification. Identification is a cognitive
act and a far from trivial or easy one. One and the same object may o en be
identified in dierent ways, even by the same thinker, and it may not be evident
to such a thinker that, indeed, he or she has identified the same object in two
such ways. To be able to identify an object is, typically, to be in possession of
some descriptive information which applies uniquely to that object. But, as
Frege (1960 [1892]) pointed out, a thinker can be in possession of two such
pieces of information without necessarily thereby knowing that they apply to
the same object. To use his famous example, it was an astronomical discovery of
considerable magnitude that the Evening Star (Hesperus) is the Morning Star
(Phosphorus). Similarly, it would be a stunning discovery to find out that the
victor of Austerlitz (Napoleon) is the author of this chapter (Jonathan Lowe).
As we shall see, the role of identity criteria is to impose certain constraints
on what can count as an acceptable answer to such a question concerning iden-
tification. But in order to understand that role, we first need to say a li le bit
more about identity as such.
Identity, as has been remarked, is a reflexive relation a relation which,
of necessity, holds between everything and itself. We can formalize this as
follows:
(x)(x = x)
As was also remarked earlier, identity is subject to Leibnizs law, which for
our purposes may be formalized in this way:
(x)(y)(x = y (F)(Fx Fy))
Here F stands for any condition that may hold true of an object, so that the
above formula eectively arms that, for any things x and y, if x is identical
with y, then anything true of x is also true of y, and vice versa. From the fore-
going two principles, it is easy to derive two other logical properties of the
identity relation: its symmetry and its transitivity, expressible by the following
two formulas:
(x)(y)(x = y y = x)
(x)(y)(z)((x = y & y = z) x = z)
Together, these four formulas exhaust the properties of the identity relation
from a purely logical point of view. They pin that relation down uniquely, as
205
being not only an equivalence relation reflexive, symmetrical and transitive

but also, more specifically, as being the only such relation all of whose equiva-
lence classes are necessarily single-membered, with each such member being
an ordered pair of a thing and itself, of the form x, x. To make this la er point
clearer, each equivalence class of the same height relation is the class of all those
pairs of objects that share a certain height and, clearly, while it might happen
to be the case only one object has a certain height, it is also possible for more
than one object to have the same height. Hence, some of these equivalence
classes may contain ordered pairs of dierent objects, such as Peter, Jane,
Jane, Mary and Peter, Mary, assuming that Peter, Jane and Mary all have the
same height. But the equivalence classes of the identity relation are all boringly
uniform, each having a unique member such as Peter, Peter or Jane, Jane
because, obviously enough, Peter is identical only with Peter, Jane only with
Jane, and so on. These rather austere logical points are not made idly here, since
they will be seen to have a direct bearing on what can qualify as a satisfactory
criterion of identity for things of a given kind. We may sum the situation up
by saying that while an equivalence relation such as the same height relation
may be described as being an exact similarity relation, the identity relation is
necessarily stricter than that, in that it can fail to hold even between objects that
are in every respect exactly similar.
Criteria of Identity
A criterion of identity is a principle which specifies, in a non-trivial way, logi-

cally necessary and sucient conditions for the identity of objects of a given
sort or kind, K. The qualification in a non-trivial way is needed to exclude
principles that are uninformative or circular. Such a principle may take one
or other of two dierent forms and, depending on which it takes, it may be
described as being either a one-level or a two-level identity criterion (see
Lowe, 1997). One-level criteria take the following form:
(x)(y)((Kx & Ky) (x = y RKxy))
Here, RK denotes what we may call the criterial relation for objects of kind
K. And note that such a relation must, of course, be an equivalence relation
reflexive, symmetrical and transitive because identity itself is an equivalence
relation and RK has to hold between Ks just in case they are identical. The
best-known example of such a one-level identity criterion is the axiom of exten-
sionality of set theory, which tells us that if x and y are sets, then x is identical
with y if and only if x and y have the same members, so that in this case having
the same members is the relevant criterial relation. However, Frege, who founded
206
Personal Identity
the formal theory of identity criteria, favoured two-level identity criteria, which
may be wri en in the form:
(x)(y)(fK(x) = fK(y) RKxy)
Here, fK denotes what could aptly be called the K-function. The best way to
illustrate this is by means of Freges own famous example of such an identity
criterion, his criterion of identity for directions (see Frege, 1953 [1884], p. 74).
A direction (in the geometrical sense of the word) is always a direction of
something, namely, a line. And Freges criterion of identity for directions is
just this: the direction of line x is identical with the direction of line y if and only
if x and y are parallel. So, in this case, the K-function is the direction of function
and the criterial relation for directions is parallelism between lines. Observe that
both the relation of having the same members and the relation of parallelism
between lines are, as required, equivalence relations.
It should be easy to see why the two dierent forms of identity criteria
receive their respective names. A two-level criterion specifies the identity-
conditions of things of a kind K in terms of an equivalence relation between
things of another kind; thus, in the case of Freges criterion, it specifies the iden-
tity-conditions of directions in terms of an equivalence relation between lines.
In contrast, a one-level criterion specifies the identity-conditions of things of a
kind K in terms of an equivalence relation between those very things; thus, in
the case of the axiom of extensionality, it specifies the identity-conditions of
sets in terms of an equivalence relation between those sets. We shall see that
this dierence between the two forms of identity criteria is significant in the
context of a search for an adequate criterion of personal identity. For a two-level
criterion of personal identity will be appropriate only if we can think of persons
as being objects of a functional kind, in the sense that directions are.
Something more should now be said about the requirement that a criterion
of identity be non-trivial and, more particularly, non-circular. Clearly, it would
be blatantly circular to allow the criterial relation in a one-level criterion of
identity for Ks simply to be the relation of identity itself. It is true, but just
trivially so, that if x and y are Ks, then x is identical with y if and only if x and
y are identical. But sometimes a putative identity criterion can be circular in a
less obvious way: for example, the putative identity criterion for sets which
states that if x and y are sets, then x is identical with y if and only if x and y
include exactly the same sets. It is indeed logically necessary and sucient for
the identity of sets x and y that x and y include exactly the same sets (bearing
in mind that every set includes itself), but since what we are seeking is an
informative way of specifying the identity-conditions of sets, it is clearly
unsatisfactory to do so by appealing to a criterial relation in this case, the
relation of including the same sets which is itself defined at least partly in terms
207
of sameness (i.e. identity) between sets. Another example of such circularity is

provided by Donald Davidsons well-known proposal regarding the identity-
conditions of events, namely, that events x and y are identical if and only if x
and y have the same causes and eects (Davidson, 1980 [1969]). For, since he takes
causes and eects themselves to be events, this proposal amounts to the circular
claim that events x and y are identical if and only if the same events cause both
x and y and the same events are caused by both x and y (see Lowe, 1989). A
criterion of identity for Ks should never appeal to or rely upon, in its formula-
tion of the criterial relation for Ks, sameness (i.e. identity) between Ks. Unfor-
tunately, circularity of this kind in a putative identity criterion is not always
easy to spot and sometimes needs considerable work to tease out. This is a
problem that aicts certain well-known a empts to formulate an adequate
criterion of personal identity, as we shall see.
One final point needs to be made about identity criteria in general. This is
that they are here being taken to be metaphysical principles, not merely epistemic
or heuristic ones. Thus, for example, while it is true in the case of human persons
that having the same fingerprints provides strong empirical evidence for
identity between such persons, it certainly isnt true that human personal iden-
tity consists in having the same fingerprints for, quite apart from anything
else, a human person can obviously survive the loss of his or her fingerprints
(by losing his or her fingers) and indeed can even, in these days of modern
medicine, acquire someone elses fingerprints (as a result of a hand-transplant).
So it cant be true, quite generally and of necessity, that human persons x and
y are identical if and only if x and y have the same fingerprints.
What is a Person?
Locke, as we noted earlier, very wisely observed that This being premised to
find wherein personal identity consists, we must consider what Person stands
for (Locke, 1975 [1690], II, XXVII, p. 9). We cannot hope to formulate an
adequate criterion of identity for objects of a kind K unless we have a pre y
good idea as to what Ks are. But what exactly are we asking when we ask a
question of the form What are Ks? The short, but, I think, correct answer is
that we are inquiring into the nature or essence of Ks. As for what the word
essence means in this context, we again do well to quote Locke who said that
in the proper original signification of the word it denotes the very being of
anything, whereby it is, what it is (Locke, 1975 [1690]: III, III, p. 15). From this
we may glean that, at the very least, we do not know what a K is unless
we know to what ontological category Ks belong. Unfortunately, in the case
of persons this immediately gives rise to a problem, namely, that dierent
208
Personal Identity
philosophers over the ages and across cultures have had very dierent views as
to what, in this sense, persons are. Some have held that persons are essentially
immaterial substances (spirits or souls), some that they are combinations
of such a substance with a material one (a body), some that they are purely
material substances (such as living animals), some that they are phases of such
substances (rather as caterpillars and bu erflies are dierent phases of the
same kind of insect), some that they are non-substances (such as bundles of
experiences, or functional roles that substances can occupy), some that they
are not even individual entities of any kind but rather universals of a certain
type, some that they are transcendental entities which cannot be identified
with items of any kind that are located in the world of space and time, some
that they are literally non-entities having a purely fictional status.
What is the source of this remarkably wide dierence of opinion concerning
the nature or essence of persons? Perhaps this: the key ingredient in anyones
conception of a person seems to be the conviction that, at least, he or she herself
is a person. Thus, possession of the first-person perspective is at the heart of any-
ones conception of a person, whatever else may also be part of it. A person
is, first and foremost, something that conceives of itself as thinking, feeling or
doing various things (see Lowe, 1996, Chapter 1). Such a conception is one that
requires the deployment of the first-person pronoun, I, or some expression
equivalent to that, for its articulation. But the peculiar feature of this pronoun,
from a semantic point of view, is that its competent use apparently does not
require of the user any very specific conception of what kind of thing it desig-
nates. This is why Descartes (1984 [1641], II) could famously claim to be certain
of the truth of the cogito I think and thereby certain of his own existence, while
still professing uncertainty as to what he was. In the end, of course, he concludes
that he is essentially a thinking thing, a substance whose essence is thinking
(in the broadest sense of that term) and which excludes any other property
(at least, any material property, such as shape or mass).
Locke, as we have already seen, is less prescriptive concerning the nature
or essence of persons, saying only that a person is a thinking intelligent Being,
that has reason and reflection, and can consider it self as it self, the same
thinking thing in dierent times and places (Locke, 1975 [1690], II, XXVII, p. 9).
This definition of personhood certainly builds in the notion that a person is a
self-aware subject of thought and experience, but it is far from clear that we
should take Locke to be implying, by his use of the capitalized word Being,
that persons are substances, much less that they are essentially immaterial
substances (or, indeed, that they are essentially material ones either). In fact,
it would appear that Locke held human persons to be, strictly speaking, non-
substances, with their ontological status being that of modes, or bundles of
modes (mode being Lockes preferred term for an individualized property, or
209
what would nowadays be called by metaphysicians a trope). This is because,

while he believed that thoughts and other mental modes have to be borne by
substances and that these substances are in all probability spiritual rather than
material in nature, he held that you or I, as human persons having such
thoughts, could not be identified with any such substance, since you or I could
in principle survive a change in respect of the substance bearing our thoughts
at dierent times in our lives. This, obviously, is connected with Lockes own
theory of personal identity and his preferred criterion of personal identity, to
which we shall return shortly.
So the problem is that, while practically everyone might agree that, whatever
else a person is, a person is something that is, or at least is capable of being,
aware of itself as having thoughts, this formulation apparently leaves it almost
entirely open as to what kind of thing this something is. In fact, it even seems
to leave open the possibility that there need be no one kind of thing that a per-
son could be. If that is the case, however, then it would appear to be misguided
to search for a criterion of personal identity as such, since persons of dierent
kinds could be expected to comply with the identity criteria, whatever they
might be, associated with the kinds in question. For example, if it is held that
human persons as opposed, say, to android persons of science fiction lore
are animals of a certain kind and thus that I, as a human person, am identical
with such an animal (a biological organism of the species homo sapiens), then
it should be concluded that my identity-conditions are just those of one
such animal that I began to exist when it did and will cease to exist when it
does. This view, known as animalism, is currently fairly popular among meta-
physicians (see Olson, 1997), perhaps on account of its thoroughly naturalistic
flavour and perhaps too because it eectively does away with all the traditional
problems of personal identity of the sort that Lockes account generates.
On the other hand, the idea that persons are not really a single kind of things
and thus that things of radically dierent kinds, with quite dierent identity-
conditions, could all qualify as persons is prima facie counterintuitive and even
rather disturbing in its apparent moral implications. As Locke so aptly put
it, person is a Forensick Term appropriating Actions and their Merit (Locke,
1975 [1690], II, XXVII, p. 26): it is indispensable for our moral and legal practices
of apportioning praise and blame and oering rewards and punishments.
Ones natural presumption is that each person has and should have a moral
concern for his or her own future and, more generally, for the futures of all
other persons. But if there is no unified conception of what would count as
the future of a person as such, because persons of dierent kinds can have
quite dierent identity-conditions, it may be hard to see what exactly could
be the basis of such a universal moral concern. Indeed, if animalism were true
regarding human persons such as you and I, why, a er all, should I have any
210
Personal Identity
moral concern for your or my future as such, given that the animals that you
and I supposedly are have identity-conditions which dont entail that those
futures are ones in which you or I exist as persons at all?
Reflections such as these suggest that it is strongly built into the common-
sense conception of a person that all persons are essentially persons, so that
my ceasing to be a person would entail my ceasing to exist altogether. Lockes
definition of personhood, whatever its defects, is clearly intended by him to
have this consequence and to that extent seems to be more in tune with com-
mon sense than a view like animalism is. This, in any case, is a good point
at which to look more closely at Lockes own proposed criterion of personal
identity, not only because it is interesting in its own right but also because it
is, in eect, the first explicitly formulated criterion of personal identity to be
found and has remained highly influential. This is not to deny that preceding
philosophers were implicitly commi ed to various criteria of personal identity
which can be deduced from their writings. The point is just that Locke has
the distinction of being the first philosopher who explicitly acknowledged
the notion of a criterion of identity although he did not use that term for it
and applied it to the case of persons.
Lockes Criterion of Personal Identity
According to Locke,
[S]ince consciousness always accompanies thinking, and tis that, that makes
every one to be, what he calls self; and thereby distinguishes himself from all
other thinking things, in this alone consists personal Identity, i.e. the sameness
of a rational Being. And as far as this consciousness can be extended
backwards to any past Action or Thought, so far reaches the Identity of that
Person; it is the same self now as it was then; and tis by the same self with
the present one that now reflects on it, that that Action was done. (Locke,
1975 [1690], II, XXVII, p. 9)
It is a ma er of some controversy among Locke scholars how exactly this

passage should be unpacked (see Lowe, 1995, Chapter 5, and Lowe, 2009,
Chapter 7), but most commentators take it to be expressing a memory-based
criterion of personal identity, on the understanding that the kind of memory
that we are here concerned with is what is sometimes called autobiographical
or experiential memory (e.g. remembering seeing a certain film some years
ago), as opposed to the mere memory of impersonal facts (such as remembering
that the Ba le of Hastings was fought in 1066).
211
Here is one way in which one might a empt to frame Lockes proposed
criterion in the form of a one-level identity criterion, as such criteria were
formulated earlier:
(x)(y)((x is a person & y is a person) (x = y (t1)(t2)(e)((x

experiences e at t1 y remembers e at t2) & (y experiences e at t1 x
remembers e at t2)))),
where t1 and t2 are any two times at which both x and y exist (with t1 being
earlier than t2) and e is a variable ranging over individual conscious experi-
ences, such as a conscious experience of having a particular thought or under-
taking a particular action. What the foregoing formula says, in plain English, is
just this: if x and y are persons, then they are the same person if and only if any
conscious experience had by x at any earlier time is remembered by y at any
later time, and vice versa (restricting ourselves here to times at which both x
and y exist, of course, since no person can experience or remember anything at
a time at which he or she doesnt exist). This criterion entails, obviously, that
a person must always remember every conscious experience that he or she
ever formerly had. That, however, is extremely implausible. Indeed, its implau-
sibility was fairly soon exploited by Thomas Reid (1975 [1785]) to construct
a refutation of Lockes proposed criterion by means of his well-known brave
ocer example, as follows.
We can readily imagine there being an elderly general who remembers
saving the regiments standard when in ba le as a young ocer and who, as
a young ocer, remembered stealing apples as a boy. But it also seems quite
conceivable that the elderly general has entirely forgo en the boyhood episode.
Suppose, indeed, as seems prima facie conceivable, that the elderly general remem-
bers every experience of the young ocer and the young ocer remembers
every experience of the boy, but that the elderly general remembers only some
of the experiences of the boy. Then it seems to follow that, by Lockes criterion
(as we have stated it), the elderly general is the same person as the young ocer
and the young ocer is the same person as the boy, but the elderly general is not
the same person as the boy. This, however, blatantly conflicts with the transiti-
vity of identity and implies that Lockes proposed criterial relation for personal
identity remembrance of past experience, as we may call it is not, as required,
an equivalence relation. We might seek to remedy ma ers by relaxing Lockes
criterion so as to require only that a person remember some of the experiences
that he had at any earlier time in his life. (Very possibly, indeed, this is all that
Locke himself really meant to imply.) But then the counterexample can be
modified by having the elderly general remember only some of the young ocers
experiences, who in turn remembers only some of the boys, while the general
remembers none of the boys which again seems perfectly conceivable.
212
Personal Identity
Modications to Lockes Criterion
Fortunately, if Reids objection does indeed expose a fatal flaw in Lockes

proposed criterion, then it is one that is fairly easily rectified. For, given a
certain non-transitive relation, R, it is always easy enough to define in terms
of R another relation, R*, which is guaranteed to be transitive, namely, the
so-called ancestral of R. Consider, for example, the parenthood relation, in which
any parent stands to his or her children, a relation which is evidently not transi-
tive. The ancestral of this relation is (appropriately enough) the relation of
being an ancestor of. This is the relation that holds between x and y if and only
if there is a chain of individuals, beginning with x and ending with y or vice
versa, such that adjacent individuals in the chain stand in the parenthood
relation. Thus, my great-grandmother is an ancestor of mine because she is a
parent of a parent of a parent of mine. And the relation of being an ancestor of
is plainly transitive: necessarily, if x is an ancestor of y and y is an ancestor of z,
then x is an ancestor of z. So, it seems, all that we need to do to save Lockes
criterion of personal identity from Reids objection is to replace the relation of
remembrance of past experience by the ancestral of that relation, call it connected-
ness of remembered past experience. The elderly general does stand in this relation
to the boy, it seems, given that he remembers every experience of the young
ocer who, in turn, remembers every experience of the boy. Certainly, he does
so if at every time between now and when he was a boy, there existed a person
who remembered every experience had by a person existing at the preceding
moment of time, beginning with the elderly general and ending with the boy.
By the modified Lockean criterion, then, even a person whose autobiographical
memory is limited to a span of just a few minutes or seconds and there are in
fact such unfortunate individuals can in principle be identified with a person
living many years ago.
Here it may be wondered what Locke and indeed we should say about
the possibility of persons undergoing periods of complete loss of consciousness,
as appears to happen in deep sleep or a coma. It would be consistent with
the modified Lockean criterion to say that persons simply cease, temporarily,
to exist during such periods. If a neo-Lockean wished to avoid saying this,
however, then it seems that the criterion would have to be modified further,
by replacing appeal to the actual remembrance of past experience by appeal to
a capacity for such remembrance, which can be retained during periods of
complete unconsciousness.
Now, of course, the fact that connectedness of remembered past experience, as
we have decided to call it, is a transitive relation doesnt guarantee that it is,
as required, an equivalence relation, since to have that status it needs also to
be reflexive and symmetrical. That it is reflexive might seem to be relatively
uncontroversial, but that it is symmetrical is certainly not, for the following
213
reason. It seems at least prima facie conceivable that two distinct persons, A
and B, existing at a time t2 should both stand in this relation to a single person,
C, existing at an earlier time t1. But if the relation is both transitive and sym-
metrical, this implies that A and B stand in the relation to each other. Why?
Call the relation R for short. Then, we are given that (1) A is R to C and (2) B is
R to C. But if R is symmetrical, then it follows from (2) that (3) C is R to B.
And given that R is also transitive, it follows from (1) and (3) that (4) A is R to B.
Here it may be suggested that an advocate of the modified Lockean criterion
should just bite the bullet and accept that in such circumstances A and B are
not, a er all, two distinct persons. But this simply isnt sustainable, even by the
standards of the modified Lockean criterion. For the kind of circumstances
that we are now envisaging are ones in which a single person, C, supposedly
undergoes a process of fission, spli ing into two distinct persons who go
on to build up, therea er, quite dierent and unconnected stores of autobio-
graphical memory. This, supposedly, might occur as a result of the bisection of
Cs brain into its two hemispheres, each of which is then transplanted into the
head of a dierent human body (see Nagel, 1979 [1971]). At the later time t2, it
simply will not be true to say that A stands in the R relation to B, because there
will be experiences that B had a er the fission event which are not connected
to any memory that A has at t2. Suppose, for example, that at some time a er the
fission event, B experiences a toothache. Will it be the case that at t2 A remem-
bers the past experience of someone who remembers the past experience of
someone . . . who remembers Bs toothache experience? Surely not: for the
memory-chain in question will take us back to C at the moment of fission, but
not forward from there to Bs toothache experience.
Another thing that we should bear in mind in assessing the merits of the
modified Lockean criterion is this. While it is necessary that a criterial relation
should be an equivalence relation, this is not sucient, since such a relation
is required to hold between objects of a kind K just in case they are identical.
Consequently, it cannot hold between distinct Ks. However, it may readily be
argued that the relation of connectedness of remembered past experience doesnt
meet this demand, if we are prepared to countenance, in addition to cases of
personal fission, cases of personal fusion (for instance, as a result of a reversal
of the kind of brain-bisection and double transplant operation described
earlier). For if a single person, C, existing at a time t2, stands in this relation to
both of two distinct persons, A and B, existing at an earlier time t1, then it
follows since at most one of A and B can identical with C that C stands in this
relation to at least one person who is not identical with C.
Of course, it may be objected that these imagined cases of personal fission
and fusion are purely imaginary and not really possible. But that is much too
big a debate to be entered into here. Suce it to say that such cases present
a prima facie problem for the modified Lockean criterion. A rather dierent
214
Personal Identity
problem that might be raised for it is the following, which leads to an

accusation first made by Joseph Butler (1975 [1736]) that the criterion is
implicitly circular. The criterion appeals to the notion of a person, P, remember-
ing some past experience, e. But isnt it in fact a logically necessary condition
of Ps genuinely remembering e (in the first-personal, autobiographical sense of
remembering) that P him or herself should actually have experienced e? How
could you properly be said to remember having an experience which you
didnt have? Wouldnt that simply be a false memory, that is, a false impres-
sion of remembering something, rather than a genuine memory of anything?
If that is so, then, as Butler urged, memory presupposes personal identity and
hence cannot be what constitutes it.
The standard modern response to this objection is to concede it, but then to
modify the Lockean criterion still further by appealing instead to the notion
of quasi-memory, where this is understood to be a mental state with all the
features of autobiographical memory except that it is not a defining condition
of the state that one can quasi-remember only experiences that one had oneself
(see Parfit, 1984, Chapter 11). It is allowed; that is to say that it is at least logi-
cally possible to quasi-remember the experiences of another person. However,
this appeal to the notion of quasi-memory in defence of a neo-Lockean criterion
of personal identity is a two-edged sword. For although it enables the defender
of such a criterion to avoid the Butlerian charge of circularity, it does so at the
expense of ruling out one kind of reply to the problem raised earlier in cases of
personal fusion. Thus, while it may be objected that C in such a case cannot
genuinely stand in the ancestral of the memory relation to a past experience of
someone distinct from C, it has to be allowed that C can stand in the ancestral
of the quasi-memory relation to such an experience. It must also be acknow-
ledged that the notion of quasi-memory is far from being uncontroversial, since
a good many philosophers doubt whether it really makes sense (see Wiggins,
2001, Chapter 7).
Suppose, however, that we set aside such doubts. What if anything can be
done to further modify a neo-Lockean criterion that appeals to the ancestral
of the quasi-memory relation (i.e. to the relation of connectedness of quasi-
remembered past experience), in order to safeguard it against the threat posed by
putative cases of personal fission and fusion? There is a simple enough answer:
simply build into the criterion a clause excluding any such branching. Then
we can say, in nutshell, that persons x and y, existing at times t1 and t2 respec-
tively, are identical just in case there is a non-branching chain of connected
quasi-memories linking y at t2 to an experience of x at t1. Writing out this
criterion formally, in the style deployed earlier, would be too complicated to
be very useful, but it can certainly be done. However, such a criterion is still
open to objection, even waiving any diculty that one might have with the
notion of quasi-memory. One objection is that it violates what is sometimes
215
called the only x and y principle: this is the principle that, in order to se le a
question about whether an object x is identical with an object y, only facts
about x and y should be deemed relevant, not facts concerning other objects
(see Wiggins, 2001, p. 96). The proposed condition on non-branching violates
this principle, because it amounts to the requirement that person xs identity
with person y is conditional upon there existing no other person, z, in addition to
x and y as a result of some fission or fusion process involving them. However,
it may be questioned why we should regard the only x and y principle as
sacrosanct. Why shouldnt we just concede that identity can sometimes
be extrinsically determined, by being dependent on the existence or non-
existence of other objects in addition to those whose identity is at issue?
Another objection is that the new version of the Lockean criterion is at odds
with our moral convictions concerning the importance of personal survival.
For, if my survival requires the future existence of someone who is identical
with me, then it seems that, by the neo-Lockean criterion, my surviving or
not surviving can turn upon the seemingly irrelevant ma er of whether or
not I will at some point undergo fission or fusion (however far-fetched such
scenarios may seem). To this, however, it may be replied that the real lesson
of this is that my survival, in the sense in which it is or should be something
of importance to me, should be defined not in terms of my identity with some
future person but rather in terms of there being at least one such person who
is linked to me by a connected chain of quasi-memories; if there is more than
one, as in a fission case, then so much the be er, on this view (see Parfit, 1984,
Chapter 12 and Chapter 13).
Another Circularity Objection to the Neo-Lockean Approach
Does this mean, then, that the final, non-branching version of the neo-Lockean
criterion is finally acceptable? Does it satisfy all the requirements of an adequate
criterion of personal identity? Very arguably, it does not, for it still seems vul-
nerable to a charge of implicit circularity, although one of a dierent sort from
Butlers. Recall again that the neo-Lockean criterion appeals to the notion of a
person, P, remembering or, rather, quasi-remembering some past experi-
ence, e. But now we need ask ourselves this: how are memories and experiences
themselves individuated? Such items are mental states or events. But what are
their identity-conditions? We can already rule out the Davidsonian criterion of
identity for events as a way of se ling this question because we found it to be
implicitly circular. It was so because it sought to identify events on the basis of
the sameness of their causes and eects, while also taking these causes and
eects to be events themselves. So it defined sameness of events in terms
appealing to sameness among events, a blatantly circular procedure, leaving us no
216
Personal Identity
clearer as to what the identity of events consists in. But the neo-Lockean cri-
terion likewise appears to be implicitly circular, albeit in a rather more round-
about way. For it is strongly arguable that the only adequate criterion of identity
for mental states and events will be one which makes reference to their subjects,
which, in the case of personal memories and experiences, will be the persons
who have those memories and experiences (see Strawson, 1959, Chapter 3).
Let us focus on the case of experiences, although the same reasoning
will apply equally to memories. On the view now being recommended, part of
what makes an experience of mine numerically distinct from a qualitatively
indistinguishable experience of yours is the very fact that it is mine as opposed
to yours. The only other possible distinguishing feature seems to be the time
at which an experience occurs. In short, the following seems to be a very plau-
sible criterion of identity for personal experiences:
(x)(y)((x is a personal experience and y is a personal experience)

(x = y (x and y are qualitatively indistinguishable & (P1)(P2)(t1)(t2)
(P1 has x at t1 & P2 has y at t2 & P1 = P2 & t1 = t2)))),
where P1 and P2 are variables ranging over persons and t1 and t2 are vari-
ables ranging over times. In plain English, what this formula says is just this: if
x and y are personal experiences, then they are the same personal experience if
and only if x and y are qualitatively indistinguishable experiences had by the
same person at the same time. It is quite clear that the criterial relation invoked
by this criterion is, as required, an equivalence relation. But, equally, it is obvi-
ous that it appeals to the notion of sameness between persons and hence presup-
poses that notion. Accordingly if, as I strongly suspect is the case, this is the
only adequate criterion of identity for personal experiences, then the neo-
Lockean criterion of personal identity is implicitly circular inasmuch as it
will need to rely on the foregoing criterion for a specification of the identity-
conditions of the experiences to which it appeals for the purposes of identifying
persons. Clearly, at any rate, we cannot both individuate persons in terms of
their experiences (as the neo-Lockean criterion a empts to do) and individuate
personal experiences in terms of the persons having them (as the foregoing
criterion does). And to the extent that the foregoing criterion of identity for
personal experiences looks to be in good order, it is the neo-Lockean criterion
that must be rejected as inadequate (see Lowe, 2009, Chapter 7).
Some Loose Ends and a Brief Conclusion
I have focused on the neo-Lockean approach because it is, deservedly, by far

the most prominent one in the modern literature on personal identity, whether
217
it is being endorsed or being a acked. But something should be said now about
some alternative approaches. First of all, so far we have considered only the
prospects for a one-level criterion of personal identity. But on some views of
what persons are, a two-level criterion might seem more appropriate for
instance, if persons are taken to be functional states or roles that objects of
appropriate kinds can occupy. Thus, one such view would be that a persons
body, or a special part of that body, such as its brain, is the object that occupies
the functional role in question. Suppose that being a person is a functional
role of a brain (e.g. it might be taken to be the role of being a producer of first-
person thoughts). Then a criterion of personal identity could be expected to take
something like the following two-level form:
(x)(y)(the person of brain x = the person of brain y (brain x and brain y

are RP-related),
where RP denotes a certain equivalence relation among brains. Indeed,

on one view, this relation would simply be identity itself. There would be no
circularity in the criterion on this account, since it would simply be defining
personal identity in terms of brain-identity, and persons and brains are here
being taken to be items of quite dierent kinds. So this approach is by no means
identifying a person with his or her brain. The brain-identity criterion of per-
sonal identity just implies that a persons identity tracks that of the persons
brain, so that, for example, if a person As brain is transplanted into evacuated
head of another person Bs body, then person A acquires person Bs body; and if
person As brain is switched with person Bs, then we have a body-swap, with
person A acquiring person Bs body and person B acquiring person As body.
The scenario is really very similar to Lockes famous imaginary example of the
prince and the cobbler, who supposedly undergo a body-swap, although what
Locke envisaged was that the soul of the prince entered the cobblers body and
the soul of the cobbler entered the princes body (Locke, 1975 [1690], II, XXVII,
p. 15). However, although Locke thought that this scenario was in principle
possible, he did not, of course, subscribe to a soul-identity criterion of personal
identity because he thought that the same person could in principle have
dierent souls at dierent times and that the same soul could, at dierent times,
be the soul of dierent persons. (For a modern defence of a soul-identity
criterion, see Swinburne, 1986, Chapter 8.)
I have nothing to say in recommendation of a two-level approach such as the
brain-identity criterion, although it will clearly appeal to some philosophers
and psychologists. Such an approach is clearly inappropriate if we regard the
term person as denoting a distinct kind of substantial being, an individual
substance, rather than a certain kind of state or role that such a substance can
occupy. Certainly, common sense and ordinary language strongly suggest the
218
Personal Identity
former view. I feel myself to be some thing, with distinctive properties such as
thought and feeling, rather than my being merely some property or feature of
some other thing, such as my brain. But it must be confessed that a satisfactory
criterion of personal identity that supports this conviction is still very elusive.
On the other hand, we should be open to the possibility that personal identity
is so basic in our ontological scheme that we should not really expect to be able
to formulate such a criterion. For, as we have seen, criteria of identity for objects
of a kind K always appeal to objects of other kinds in specifying a criterial rela-
tion for K-identity. If persons are really fundamental in our ontological scheme,
we should not expect to be able to appeal to such other kinds of objects in their
case. That being so, we should probably conclude that personal identity is
primitive and simple, in the sense that nothing more informative can be said
about the identity of persons than that in some cases it just obtains and in
others not (see Lowe, 2009, Chapter 7).
219
12 Embodied Cognition and
the Extended Mind
Michael Wheeler
The Flight from Cartesianism
There is a seductive image of intelligent action that sometimes gets labelled

Cartesian. According to this image, as I shall present it here, the psychological
understanding of the operating principles by which an agents mind contrib-
utes to the generation of reliable and flexible, perceptually guided intelligent
action remains conceptually and theoretically independent of the details of
that agents physical embodiment. Less formally, one might say that, in the
Cartesian image, the body enjoys no more than a walk-on part in the drama of
intelligent action. Whether or not the Cartesian image is Cartesian in the sense
that it ought to be a ributed to Descartes himself is a ma er that demands
careful exegetical investigation (e.g. see Wheeler, 2005 for an analysis which
concludes that, by and large, it should). In general, positions that are currently
identified as Cartesian may not map directly or completely onto Descartess
own views. This potential mis-match is an example of a widespread phenome-
non and should come as no surprise. Were Karl Marx with us today, he might
well express serious misgivings about some of what has been said and done
in the name of Marxism. In Descartess case, his views have been handed down
to us via a rich intellectual history of contested interpretations and critical
debate. Inevitably, perhaps, some ideas that now bear the stamp Cartesian
will have as much to do with that intervening process as they have to do with
Descartes himself. Anyway, for now, I intend to ignore the question of pro-
venance. What is crucial in the present context is that the two views of intelli-
gent action with which I shall be concerned in this chapter the hypotheses of
embodied cognition and of the extended mind may be understood as dierent
stop-o points in a flight from the image in question.
To bring all this into be er view, we can adapt an example due to Clark
(1997, pp. 634) of some dierent ways in which an intelligent agent might
solve a jigsaw puzzle. Here is a strategy suggested by the Cartesian image.
On the basis of perceptual information about the problem environment (the
unmade jigsaw), the agent solves the entire puzzle in her head, using some
220
combination of mental imagery, judgement, inference, reasoning, and so on.

The solution arrived at in this way is then executed in the world, through a
series of movement instructions that are dispatched from the mind, to the hands
and arms. Things may not always go according to plan, of course, but any fail-
ures experienced during the execution phase act as nothing more than percep-
tual prompts for some newly initiated in-the-head planning. Now, it is quite
obvious that the puzzle-solving mind at the core of this activity needs a body
to execute the movement instructions generated by that mind; and nothing
in the account on oer suggests that there could be minds without brains.
(Substance dualism is not the issue.) Nevertheless, in this Cartesian scenario,
the fact is that the body makes only an impoverished contribution to the intel-
ligence on display. The nature of this impoverishment becomes clear once a
second vision of jigsaw competence is placed on the table. According to this
new vision, certain bodily acts, such as picking up various pieces, rotating those
pieces to help pa ern-match for possible fits, and trying out potential candi-
dates in the target position, are deployed as central aspects of the agents
problem-solving strategy. In the unfolding of this alternative plot, the details of
the thinkers embodiment, in the guise of the specific embodied manipulative
capacities that she deploys, plays an essential supporting role in the story of
intelligent action. This is an example of embodied cognition.1
Notice that problem-solving strategies which essentially involve bodily
acts will o en encompass a richer mode of environmental interaction than
is present in Cartesian contexts. Thus in our Cartesian jigsaw-completing
scenario, the physical environment is arguably no more than a furnisher of
problems for the agent to solve, a source of informational inputs to the mind
(via sensing), and a stage on which sequences of pre-specified actions, choreo-
graphed in advance by prior neural processes, are simply executed. By contrast,
in the alternative, embodied cognition scenario, although the physical environ-
ment remains a furnisher of problems and a source of informational inputs, it
has also been transformed into a readily available external resource which is
exploited by the agent, in an ongoing way, to restructure the piece-finding prob-
lem and thus reduce the information processing load being placed on the inner
mechanisms involved. Indeed, the external factors in play in particular, the
geometric properties of the pieces themselves participate in a kind of ongoing
goal-achieving dialogue with the agents neural processes and her bodily move-
ments. In so doing, those external factors account for some of the distinctive
adaptive richness and flexibility of the problem-solving behaviour. The embod-
ied mind is thus also a mind that is intimately embedded in its environment.
Once one starts to glimpse the kind of environmental contribution to
intelligent action ushered in by embodied solutions, it is but a small step,
although one which is philosophically controversial, to the second of our target
positions, namely the extended mind hypothesis (Clark and Chalmers, 1998).2
221
According to this hypothesis, there are actual (in this world) cases of intelligent
action in which thinking and thoughts (more precisely, the material vehicles
that realize thinking and thoughts) are spatially distributed over brain, body and
world, in such a way that the external (beyond-the-skin) factors concerned are
rightly accorded cognitive status. In other words, actions and loops through non-
biological structure [sometimes count] as genuine aspects of extended cognitive
processes (Clark, 2008b, p. 85). So, if the extended mind hypothesis is true, it is
not merely the case that thinking is sometimes (and perhaps sometimes essen-
tially) causally dependent in complex and intricate ways on the bodily exploita-
tion of external props or scaolds. Indeed, bare causal dependence of thought
on external factors is not sucient for genuine cognitive extension (a point
rightly emphasized by Adams and Aizawa, 2008). Rather, if the extended mind
hypothesis is true, thought must sometimes exhibit a constitutive dependence
on external factors. This is the sort of dependence indicated by talk of beyond-
the-skin factors rightly being accorded cognitive status. Stretching our thesbian
metaphor beyond reasonable limits, this is the twist in the tale of intelligent
action where the scenery and the props get a mention in the cast list.
In one short chapter, I cannot hope to give a comprehensive field guide to
embodied cognition and the extended mind. So my goal will be more modest.
I shall endeavour to cast light on a specific issue which lies at the very heart
of the contemporary debate, namely the character of, and the argument for,
the transition from embodied cognition to cognitive extension (see also, Clark
2008a, 2008b; Wheeler forthcoming a and c; Rowlands, forthcoming). Here,
then, is where I am going. In the second section, I shall present some empirical
research from cognitive science which illuminates the embodied cognition
hypothesis, henceforth EmbC. In the third section, I shall suggest that once one
has accepted the resulting picture of intelligent action, there remains a philo-
sophical choice to be made over how to conceptualize the role of the body in
the action-generation process, a choice between what Clark (2008a) identifies as
a radical body-centrism and a newly interpreted functionalism. In the fourth
section, I shall explore the connection between the second of these options and
the extended mind hypothesis, henceforth ExM. My suggestion will be that
the basic character of one of the central philosophical arguments for ExM, the
argument from parity, makes that functionalist option more a ractive. In the
fi h section, I shall seek to strengthen the emerging picture by showing how
a key element of the argument from parity may be secured.
Body Matters
As I shall use the term, orthodox cognitive science encompasses the bulk of
research in both classical cognitive science (according to which, roughly, the
222
mind recapitulates the abstract structure of human language, in that it is

characterized by a combinatorial syntax and semantics) and mainstream con-
nectionism (according to which, roughly, the mind recapitulates the abstract
structure of the biological brain, in that it is organized as a distributed network
of interconnected simple processing units). Although I shall not give a full
defense of the claim here, it is arguable (e.g. see Wheeler 2005) that the
Cartesian image of an explanatorily disembodied and disembedded mind has
been a core feature of orthodox cognitive science and of the sort of scientifically
oriented philosophy of mind that rides shotgun with that science.
This is not to say that no orthodox cognitive scientist has ever expressed the
view that bodily acts in close interaction with environmental structures might
play a crucial and active part in generating complex behaviour. Simon famously
discussed the path followed by an ant walking on a beach in order to make
precisely this point (Simon, 1969; for discussion, see Boden, 2006, pp. 42930,
and Haugeland 1995/98, pp. 20911). Moreover, the conceptual geography in
this vicinity demands careful mapping. For one thing, orthodox connectionism
takes its basic inspiration from a psychologically crucial part of the organic
body, namely the brain. Indeed, the much recorded ability of orthodox connec-
tionist networks to perform cognitively suggestive feats of graceful degrada-
tion, flexible generalization, fluid default reasoning, and so on, can, in many
ways, be identified as a natural consequence of that nod to embodiment. So
the claim that the disembodied aspect of the Cartesian image has been at work
in this area of orthodox cognitive science needs to be backed by some sort of
evidence (more on that soon). In addition, as we shall see later, the language-
like compositional structures of the classical framework and the distributed
network-style structures of connectionism may be rendered fully compatible
with ExM, so it is not as if those structures must necessarily be associated with
the Cartesian image. Nevertheless, it remains true, I think, that the Cartesian
image has historically held sway as part of the received orthodoxy in cognitive
science.
All that said, things are on the move. Over the past two decades, cognitive-
scientific models generated from the EmbC perspective have become increas-
ingly common. And to the extent that such models provide illuminating,
compelling and fruitful explanations of intelligent action, EmbC as a paradigm
garners empirical support. It is in this context that it will serve our current
purpose to make a brief visit to the sub-discipline of contemporary artificial
intelligence known as situated robotics. Roboticists in this camp shun the clas-
sical cognitive-scientific reliance on detailed internal representations (although
they dont necessarily shun all forms of representation). The case for this scepti-
cism about representational control o en turns on the thought that where the
adaptive problem faced by an agent involves integrating perception and action
in real time so as to generate fast and fluid behaviour, detailed representations
223
are just too computationally expensive to build and maintain. So situated

roboticists favour an alternative model of intelligent action in which the robot
regularly senses its environment (rather than checks an internal world model)
to guide its actions. It is this commitment that marks out a robot as situated
(Brooks, 1991). One of the key lessons from research in this area is that much
of the richness and flexibility of intelligence is down not to centrally located
processes of reasoning and inference, but rather to integrated suites of special-
purpose adaptive couplings that combine neural mechanisms (or their robotic
equivalent), non-neural bodily factors, and environmental elements, as equal
partners in a behaviour-generating strategy. Unsurprisingly, then, the field of
situated robotics is a rich storehouse of examples of embodied cognition.
To illustrate just how explanatorily powerful the appeal to embodiment
may be in cognitive science, consider the following challenge. Clark and
Thornton (1997) claim that there are certain learning problems so-called type-2
problems where the target regularities are inherently relational in nature,
and so are statistically invisible in the raw input data. Type-2 problems are thus
to be contrasted with type-1 problems, which involve non-relational regulari-
ties that are visible in that data. According to Clark and Thornton, this leaves
cognitive science with a serious diculty, because empirical testing suggests
that many of the most widely used, o-the-shelf artificial intelligence learning
algorithms (e.g. connectionist back-propagation and cascade-correlation, plus
others such as ID3 and classifier systems) fail on type-2 problems, when the raw
input data is presented. This fact would, of course, be no more than a nuisance
for cognitive science if such learning problems were rare; but, if Clark and
Thornton are right, type-2 problems are everywhere in relatively simple
behaviours (such as approaching small objects while avoiding large ones), and
in complex domains (such as grammar acquisition). Clark and Thornton pro-
ceed to argue that the solution to this diculty involves the internal presence
of general computational strategies that systematically re-represent the raw
input data so as to produce a non-relational target regularity. This output
re-representation is then exploited by learning in place of the initial input
coding. In eect, the process of re-representation renders the type-2 learning
problem tractable by transforming it into a type-1 problem.
So where do embodiment and situated robotics come in? Scheier and Pfeifer
(1998) demonstrate that a type-2 problem may be solved by a process in which
a mobile agent uses autonomous bodily motion to actively structure input
from its environment. Once again the strategy is to transform an intractable
type-2 problem into a tractable type-1 problem, but this time there is no need
for any computational inner re-representation mechanism. The test case is the
type-2 problem presented by the task of avoiding small cylinders while staying
close to large ones. Scheier and Pfeifer show that this problem may be solved
by some relatively simple, evolved neural network robot controllers. Analysis
224
demonstrated that most of these controllers had evolved a systematic circling

behaviour which, by inducing cyclic regularities into the input data, turned a
hostile type-2 climb into a type-1 walk in the park. In other words, adaptive
success in a type-2 scenario (as initially encountered) was secured not by
inner re-representation, but by an approach in which the agent, by exploiting
its body and through the interaction with the environment . . . can actually gen-
erate . . . correlated data that has the property that it can be easily learned
(Scheier and Pfeifer, 1998, p. 32).
Scheier and Pfeifers canny and frugal solution to Clark and Thorntons
challenge shows how being an embodied agent (of a mobile kind) can yield
dividends in the cognitive realm, and thus how a proper sensitivity to what
we might call gross embodiment has an impact on cognitive science. A dier-
ent, but equally important, perspective on how embodiment may shape our
understanding of cognition comes into view if we switch scale, and concentrate
instead on the detailed corporeal design of biological systems. Once again, as
we shall see, situated robotics provides an experimental context in which an
appeal to embodiment may be developed and tested.
As the flip-side of its claim to biological plausibility, mainstream connectionism
tends to promote a vision of biological brain processes as essentially a ma er
of electrical signals transmi ed between simple processing units (neurons) via
connections (synapses) conceived as roughly analogous to telephone wires.
However, as Turing once remarked, [i]n the nervous system chemical phenom-
ena are at least as important as electrical (Turing, 1950, p. 46). The factoring
out of brain-based chemical dynamics by mainstream connectionist theorizing
thus indicates another dimension along which the embodiment of cognition is
sidelined by orthodox cognitive science. So what happens when such chemical
dynamics are brought into view?
Reaction-diusion (RD) systems are distributed chemical mechanisms
involving constituents that are (a) transformed into each other by local chemical
reactions and (b) spread out in space by diusion. Such systems explain how
unicellular organisms such as bacteria manage to distinguish between dierent
relevant environmental factors, adapt to environmental change, and co-ordinate
collective behaviour. Thus behaviour that researchers in the field of artificial
life o en describe as minimally cognitive may be achieved by RD systems.
Many of the molecular pathways present in unicellular organisms have been
conserved by evolution to play important roles in animal brains, so an under-
standing of the ways in which RD systems may generate minimally cognitive
behaviour will plausibly help us to explain the mechanisms underlying higher-
level natural cognition. Against this background, Dale and Husbands (2010)
show that a simulated RD system (conceived as a one-dimensional ring of
cells within which the concentration of two coupled chemicals changes accord-
ing to dierential equations governing within-cell reactions and between-cell
225
diusion) is capable of intervening between sensory input (from whiskers)

and motor output (wheeled locomotion) to enable a situated robot to achieve
the following minimally cognitive behaviours: (a) tracking a falling circle (thus
demonstrating orientation), (b) fixating on a circle as opposed to a diamond
(thus demonstrating discrimination), (c) switching from circle fixation behav-
iour to circle avoidance behaviour on the presentation of a particular stimulus
(thus demonstrating memory). As Dale and Husbands (2010, p. 17) put it,
a range of robust minimally cognitive behaviours may be exhibited by a
seemingly homogenous blob of chemicals, a revision to our understanding
of how cognition works that is inspired by our taking seriously the details of
biological corporeal design.
In this section I have highlighted two important examples of the way in
which embodiment may have an impact on cognitive theory. In the next section
I shall address a further question: in the light of the examples of corporeal
impact to which I have drawn a ention, how, in general terms, are we concep-
tualize the fundamental contribution of the body to cognitive phenomena?
Two Kinds of Embodiment
Clark (2008a) observes that there are two dierent, although o en tangled,
strands of thinking at work within contemporary accounts that stress embodi-
ment. In the following passage, he unravels those strands for us.
One . . . depicts the body as intrinsically special, and the details of a creatures
embodiment as a major and abiding constraint on the nature of its mind:
a kind of new-wave body-centrism. The other depicts the body as just one
element in a kind of equal-partners dance between brain, body and world,
with the nature of the mind fixed by the overall balance thus achieved: a
kind of extended functionalism (now with an even broader canvas for
multiple realizability than ever before). (Clark, 2008a, pp. 567)
In order to see this division of ideas in its proper light, one needs to say
what is meant by functionalism, as that thesis figures in the debate with which
we are concerned here. The final emphasis is important, because although Clark
does not address the issue, the kind of functionalism plausibly at work in
the transition from EmbC to ExM is not the kind most usually discussed by
philosophers, although I think it is the kind most usually assumed in cognitive
psychology. To bring our target version of functionalism into view, we can
exploit McDowells (1994) distinction between personal-level explanations, which
are those concerned with the identification and clarification of the constitutive
character of agency (roughly, what it is to competently inhabit a world), and
226
sub-personal explanations, which are those concerned with mapping out the
states and mechanisms (the parts of agents, as it were) that causally enable
personal-level phenomena. Functionalism, as I shall understand it here, is a
sub-personal causal-enabling theory. It is not, as it is in its more common
philosophical form, a way of specifying constitutive criteria for what it is to
undergo types of personal-level mental states. Depending on ones account
of the relationship between personal and sub-personalsub-personal levels of
explanation, one might be a sub-personal functionalist while rejecting function-
alism at the personal level. In this paper I shall say nothing more about personal-
level functionalism. My concern is with the sub-personal version of the view,
i.e., with the claim that what ma ers when one is endeavouring to identify
the specific contribution of a sub-personal state or process qua cognitive is not
the material constitution of that state or process, but rather the functional role
which it plays in the generation of personal-level cognitive phenomena by
intervening between systemic inputs, systemic outputs and other functionally
identified, intra-systemic, sub-personal states and processes.
With that clarification in place, lets return to the division of ideas recom-
mended by Clark. In the present context, it will prove useful to re-draw that
division in terms of a closely related distinction between two kinds of material-
ity, namely vital materiality and implementational materiality (Wheeler, forth-
coming c). The claim that the materiality of the body is vital is tantamount to
the first strand of embodied thought identified by Clark, (i.e. that the body
makes a special, non-substitutable contribution to cognition, generating what,
elsewhere, Clark [2008a, p. 50] calls total implementation sensitivity). On
the other hand, if the materiality of the body is merely implementational in
character, then the physical body is relevant only as an explanation of how
mental states and processes are instantiated in the material world. The link
between implementational materiality and functionalism becomes clear when
one notes that, on any form of functionalism, including the sub-personal one
presently on the table, multiple realizability will be at least an in-principle prop-
erty of the target states and processes. Because a function is something that
enjoys a particular kind of independence from its implementing material
substrate, a function must, in principle, be multiply realizable, even if, in this
world, only one kind of material realization happens to exist for that function.
And since the multiple realizability of the mental requires that a single type of
mental state or process may enjoy a range of dierent material instantiations,
the specific material embodiment of a particular instantiation cannot be a major
and abiding constraint on the nature of mind. Put another way, the implemen-
tational materiality of the mental (or something akin to it) is plausibly necessary
for mental states and processes to be multiply realizable. And this remains
true when ones functionalism and thus the level at which the behaviour-
generating causal states and processes qua cognitive are specified is pitched at
227
a sub-personal level. By contrast, where the materiality of the body is vital,

multiple realizability is, if not ruled out altogether, at least severely curtailed
(e.g. see Shapiro, 2004, especially p. 167).
Armed with the conceptual distinction just made, how are we to conceptual-
ize the role of the body in each of our two flagship examples of embodied cogni-
tion; as a case of vital materiality (supporting a new wave body-centrism) or as
a case of implementational materiality (supporting a functionalist picture)?
My immediate answer to this question might come as something of a surprise.
For, as far as I can see, each of our examples might be interpreted according to
either vision of embodiment. Heres why.
To see Scheier and Pfeifers cylinder discriminating robots as an instance
of vital materiality, one might begin with the observation that Clark and
Thorntons appeal to an inner process of re-representation exemplifies a com-
putational information processing approach to solving the problem. One might
then suggest, with some plausibility it seems, that the way in which Scheier
and Pfeifers robots exploit gross bodily movement in their specific circling
behaviour provides us with a radical alternative to computational information
processing as a general problem-solving strategy, an alternative available only
to agents with bodies of a certain kind. To see Dale and Husbands minimally
cognitive RD system as an instance of vital materiality, one might interpret
that system as an example of what Collins calls embrained knowledge. For
Collins, knowledge is embrained just when cognitive abilities have to do with
the physical setup of the brain, where the term physical setup signals not
merely the way neurons are interconnected, but also factors to do with the
brain as a piece of chemistry or a collection of solid shapes (Collins, 2000,
p. 182). Embrained knowledge so defined is an example of total implementa-
tion sensitivity and thus establishes vital materiality. And the evidence from
Dale and Husbands that the spatio-temporal chemical dynamics of RD systems,
as plausibly conserved in the evolutionary transition from unicellular organ-
isms to animal brains, may generate minimally cognitive behaviour surely
provides an example of cognitive abilities being to do with the physical setup of
the brain, that is, of embrained knowledge.
Now lets look at things from a dierent angle. To see Scheier and Pfeifers
robots as providing an instance of implementational materiality, one might
argue that the restructuring of the learning problem achieved by their bodily
movements is functionally equivalent to the restructuring of that problem
eected by Clark and Thorntons inner re-representation strategy. In both cases,
a type-2 learning problem (intractable to standard learning algorithms as it
stands) is transformed into a type-1 problem (and so rendered tractable).
Thus one might think in terms of alternative material realizations of a single
multiply realizable, functionally specified contribution (the transformation of
the statistical structure of the target information), a contribution that may be
228
performed by inner neural mechanisms or by bodily movements. To see Dale

and Husbands RD system as an instance of implementational materiality,
one need note only that the experiments described briefly above are designed
explicitly as (something close to) replications, using an RD system, of experi-
ments in minimally cognitive behaviour carried out originally by Beer (1996,
2003; Slocum et al., 2000) using continuous recurrent neural networks (CNNs).
RD systems thus emerge as one kind of vehicle for functionally specified
mechanisms of orientation, discrimination and memory, mechanisms that could
in principle be realized in other ways, such as by CNNs.
One might worry here that RD systems and CNNs are not alternative real-
izations of certain functionally specified mechanisms, but rather alternative
ways of achieving certain minimally cognitive behaviours without there being
any more specific functional unity in terms of processing architecture. And
indeed, one might well analyze RD systems as examples of Collins embrained
knowledge, and thus of vital materiality (see above), while analyzing CNNs as
a dynamically richer form of connectionism, and thus as a kind of microfunc-
tionalist theorizing (Clark 1989) that demands an implementational notion of
materiality.3 But any such uncertainty in how to interpret the case is arguably
grist to my mill, since it will be an illustration of the very issue of under-
determination that I have set out to highlight.
As things stand, we seem to confront something of an impasse in our a empt
to understand the fundamental contribution of embodiment to cognitive theory.
To escape from this impasse, it seems to me, we have no option right now but
to look beyond the thought that the understanding we seek may be directly
read o from the available science. In the next section I shall present, analyse
and briefly defend one of the central philosophical arguments for ExM, namely
the argument from parity. I shall then explain why that argument forges a link
with the functionalist perspective on embodiment. Given that vital materiality
is inconsistent with functionalism, this suggests a consideration in favour of
the view that the fundamental contribution of the body to cognitive theory is
a ma er of implementational materiality. At the very least, if the argument
from parity is indeed sound, then the implementational view of embodiment
is correct.
From the Parity Principle to Extended Functionalism
According to ExM, there are actual (in this world) cases of intelligent action in
which thinking and thoughts (more precisely, the material vehicles that realize
thinking and thoughts) are spatially distributed over brain, body and world, in
such a way that the external (beyond-the-skin) factors concerned are rightly
accorded cognitive status. To see how one might argue philosophically for this
229
view, we need to make contact with what, in the ExM literature, is called the
parity principle. Here is how that principle is formulated by Clark (drawing
on Clark and Chalmers, 1998, p. 8):
If, as we confront some task, a part of the world functions as a process which,
were it to go on in the head, we would have no hesitation in accepting as part
of the cognitive process, then that part of the world is (for that time) part of
the cognitive process. (Clark, 2008b, p. 77)
The general idea here seems clear enough: if there is functional equality with
respect to governing intelligent behaviour (e.g. in the way stored information
is poised to guide such behaviour), between the causal contribution of certain
internal elements and the causal contribution of certain external elements, and
if the internal elements concerned qualify as the proper parts of a cognitive
system (state, process, mechanism, architecture . . .), then there is no good
reason to deny equivalent status to the relevant external elements. Parity of
causal contribution mandates parity of status with respect to the cognitive. But
if the general idea of the parity principle is clear enough, the details of how
to apply it are not, so we need to pause here to get clear about those details
(for a similar analysis, see Wheeler, forthcoming c).
One interpretation of the parity principle is suggested by the way in which
it is applied by Clark and Chalmers themselves to the near-legendary (in ExM
circles) case of Inga and O o (Clark and Chalmers, 1998). In this imaginary
scenario, Inga is a psychologically normal individual who has commi ed to
her purely organic (neural) memory the address of the New York Museum of
Modern Art (MOMA). If someone asks her the location of MOMA, she deploys
that memory to retrieve the information that the building is on 53rd Street. O o,
on the other hand, suers from a mild form of Alzheimers, but compensates for
this by recording salient facts in a notebook that he carries with him constantly.
If someone asks him the way to MOMA, he automatically and unhesitatingly
pulls out the notebook and looks up the relevant fact, viz. that the museum is
on 53rd Street. Clark and Chalmers claim that there is a functional equivalence
between (a) the behaviour-governing causal role played by O os notebook,
and (b) the behaviour-governing causal role played by the part of Ingas brain
that stores the same item of information as part of her purely organic memory.
By the parity principle, then, O os memory turns out to be extended into the
environment. Moreover, argue Clark and Chalmers, just as, prior to recalling
the information in question, Inga has the non-occurrent dispositional belief
that MOMA is on 53rd Street, so too does O o, although while Ingas belief is
realized in her head, O os is realized in the extended, notebook-including
system.
230
If we reflect on precisely how the parity principle is intended to work in

this particular case, we would be forgiven for thinking that the benchmark for
parity (the set of conditions that the O o-plus-notebook system would need
to meet in order to count as cognitive) is fixed by whatever Ingas brain does.
But although Clark and Chalmerss text sometimes leaves rather too much room
for this reading of the parity principle, it would be a tactical disaster for the
advocates of ExM if that really were what was meant. As Menary (a fan of ExM,
but not of the parity principle), drawing on work by Su on (di o), observes:
[O]nly at the grossest level of functional description can [the claim of

equivalence] be said to be true. O o and his notebook do not really function
in the same kind of way that Inga does when she has immediate recall from
biological memory. There are genuine and important dierences in the way
that memories are stored internally and externally and these dierences
ma er to how the memories are processed. John Su on has pointed out that
biological memories stored in neural [i.e., connectionist] networks are open
to eects such as blending and interference (see Su on [2006] for discussion).
The vehicles in O os notebook, by contrast, are static and do no work in
their dispositional form (Su on, 2006). (Menary, 2007, p. 59)
Other critics of the parity principle have appealed to the psychological data
on various extant inner cognitive capacities, as delivered by cognitive science,
in order to construct similar failure-of-parity arguments (e.g. see Adams and
Aizawa, 2008 on primacy and recency eects in organic memory; for discus-
sion, see Wheeler forthcoming a and c). The general version of the worry, how-
ever, is this: if (a) the relatively fine-grained functional profiles of extant inner
cognitive systems set the benchmark for parity, then (b) any distributed (over
brain, body and world) systems that we might consider as candidates for
extended counterparts of those cognitive systems will standardly fail to exhibit
full functional equivalence, so (c) parity will routinely fail, taking with it the
parity argument for cognitive extension.
Right now things might look a li le bleak for a parity-driven ExM, but
perhaps we have been moving too quickly. Indeed, it seems to me that the kind
of anti-parity argument that we have been considering trades on what is in fact
a misunderstanding of the parity principle. To see this, one needs to think more
carefully about precisely what the parity principle, as stated above, asks us to
do. It encourages us to ask ourselves whether a part of the world is functioning
as a process which, were it to go on in the head, we would have no hesitation
in accepting as part of the cognitive process. So we are encouraged to imagine
that exactly the same functional states and processes which are realized in the
actual world by certain externally located physical elements are in fact realized
231
by certain internally located physical elements. Having done this, if we then

judge that the now-internal but previously external processes count as part of a
genuinely cognitive system, we must conclude that they did so in the extended
case too. A er all, by hypothesis, nothing about the functional contribution of
those processes to intelligent behaviour has changed. All that has been varied
is their spatial location. And if the critic were to claim that that being shi ed
inside the head is alone sucient to result in a transformation in the status of
the external elements in question, from non-cognitive to cognitive, he would,
it seems, be guilty of begging the question against ExM.
To apply this understanding of the parity principle to the case of O o
and Inga, one must start with the functional contribution of O os notebook
in supporting his behaviour, and ask whether, if that functional contribution
were to be made by an inner element, we would count that contribution, and
thus its realizer, as cognitive. If the answer is yes, then we have a case for ExM.
Crucially, at no point in this reasoning have we appealed to Ingas organic
memory (the relevant extant human inner) in order to determine what counts
as cognitive. And while rather more would need to be said about the precise
functional contribution of O os notebook, our reconceived argument from
parity does not succumb to criticisms that turn on any lack of fine-grained func-
tional equivalence between the target distributed system and some extant
example of inner human cognition.
It is, of course, possible to conduct a debate that revolves around the func-
tional contributions of certain elements, without that being an issue that con-
cerns functionalism as such (cf. Chalmers, 2008). So what is the link between
the parity principle and functionalism? The parity principle is based on the
thought that it is possible for the very same type-identified cognitive state or
process to be available in two dierent generic formats one non-extended and
one extended. Thus, in principle at least, that state or process must be realizable
in either a purely organic medium or in one that involves an integrated combi-
nation of organic and non-organic structures. In other words, it must be multi-
ply realizable. So, if we are to argue for cognitive extension by way of parity
considerations, the idea that cognitive states and processes are multiply
realizable must make sense. As we have seen, functionalism provides one well-
established platform for securing multiple realizability. Moreover, although
functionalism has standardly been developed with respect to what is inside the
head (e.g. the brain of some nonhuman entity may be wired up dierently, or
it may be silicon-based rather than carbon-based, without that aecting the
rights of that entity to be judged a cognizer), there isnt really anything in the
le er of functionalism as a generic philosophical outlook that requires such an
internalist focus (Wheeler, forthcoming a and c). According to (sub-personal)
functionalism, when one is endeavouring to identify the specific contribution of
a sub-personal state or process qua cognitive, it is not the material constitution
232
of that state or process that ma ers, but rather the functional role which it
plays in the generation of personal-level cognitive phenomena by intervening
between systemic inputs, systemic outputs and other functionally identified,
intrasystemic, sub-personal states and processes. There is nothing in this schema
that requires multiple realizability to be a between-the-ears phenomenon. So
functionalism allows, in principle, for the existence of cognitive systems whose
boundaries are located partly outside the skin. It is in this way that we arrive at
the position that, following Clark, I shall call extended functionalism (Clark,
2008a, 2008b; Wheeler forthcoming a and c).
We have seen already that there will be functional dierences between
extended cognitive systems (if such things exist) and purely inner cognitive
systems. So, if extended functionalism and the parity principle are to fly
together, what seems to be needed is some kind of theory that tells us which
functional dierences are relevant to judgements of parity and which arent.
To that end, here is a schema for a theory-loaded benchmark by which parity
of causal contribution may be judged (Wheeler forthcoming a, b and c). First
we give a scientifically informed account of what it is to be a proper part of a
cognitive system that is fundamentally independent of where any candidate
element happens to be spatially located. Then we look to see where cognition
falls: in the brain, in the non-neural body, in the environment, or, as ExM pre-
dicts will sometimes be the case, in a system that extends across all of these
aspects of the world. On this account, parity is conceived not as parity with the
inner simpliciter, but rather as parity with the inner with respect to a scientifically
informed, theory-loaded, locationally uncommied account of the cognitive. So the
parity principle now emerges not as the engine room of the extended mind, but
as an heuristic mechanism that helps to ensure equal treatment for dierent
spatially located systems judged against an unbiased and theoretically moti-
vated standard of what counts as cognitive. It is a bulwark against what Clark
(2008b, p. 77) calls biochauvinistic prejudice.
This idea of a scientifically informed, theory-loaded, locationally uncom-
mi ed account of the cognitive is tantamount to what Adams and Aizawa
(e.g. 2008) call a mark of the cognitive. In the interests of expository elegance,
I shall default to Adams and Aizawas term. The most obvious next step in
this dialectic would be for me to specify the or, given the possibility that
the phenomena in question will reward a disjunctive account, a mark of the
cognitive. In the next section I shall make a tentative proposal.4
A Mark of the Cognitive
Newell and Simon, two of the early architects of artificial intelligence, famously
claimed that a suitably organized physical symbol system has the necessary
233
and sucient means for general intelligent action (Newell and Simon, 1976,
p. 116). As anyone familiar with cognitive science will tell you, a physical sym-
bol system is (roughly) a classical computational system instantiated in the
physical world, where a classical computational system is (roughly) a system
in which atomic symbols are combined and manipulated by structure sensitive
processes in accordance with a language-like combinatorial syntax and seman-
tics. I shall take it that the phrase means for general intelligent action points to
a kind of cognitive processing. More specifically it signals the sort of cognitive
processing that underlies the same scope of intelligence as we see in human
action . . . in any real situation behavior appropriate to the ends of the system
and adaptive to the demands of the environment can occur, within some limits
of speed and complexity (Newell and Simon, 1976, p. 116). What we are con-
cerned with, then, is a human-scope cognitive system. Notice that the concept
of a human-scope cognitive system is not a species-chauvinistic notion. What
ma ers is that the system exhibit roughly the same degree of adaptive flexibil-
ity we see in humans, not that it have our particular biological make-up, species
ancestry or developmental enculturation.
Against this background, Newell and Simons physical symbol systems
hypothesis may be unpacked as the dual claims that (a) any human-scope cog-
nitive system will be a physical symbol system, and (b) any physical symbol
system of sucient complexity may be organized so as to be a human-scope
cognitive system. In eect, then, the hypothesis is equivalent to the claim that
being a suitably organized physical symbol system is the mark of the (human-
scope) cognitive. To unpack that claim, the physical symbol systems hypothesis
advances a scientifically informed, theory-loaded account of the (human-scope)
cognitive, one that supports a computational form of functionalist theorizing.
But can it tick all our boxes by being a locationally independent account too?
The answer, it seems, is yes. For while classical cognitive scientists in general
thought of the symbol systems in question as being realized inside the head,
there is nothing in the basic concept of a physical symbol system that rules
out the possibility of extended material implementations. Indeed, as I shall
now argue, the idea of an extended physical symbol system has much to recom-
mend it.
In a series of compelling treatments that combine philosophical reflection
with empirical modelling studies, Bechtel (1994, 1996; see also Bechtel and
Abrahamsen, 1991) develops and defends the view that certain human-scope
cognitive achievements, such as mathematical reasoning, natural language
processing and natural deduction, are the result of sensorimotor-mediated
interactions between internal connectionist networks and external symbol
systems, where the la er feature various forms of combinatorial syntax and
semantics. It is useful to approach Bechtels suggestion (as he does himself)
234
by way of Fodor and Pylyshyns (1988) well-known claim that connectionist

theorizing about the mind is, at best, no more than a good explanation of how
classical states and processes may be implemented in neural systems. Here is a
brief reminder of Fodor and Pylyshyns key argument. It begins with the empir-
ical observation that thought is systematic. In other words, the ability to have
some thoughts (e.g. that Elsie loves Murray) is intrinsically connected to the
ability to have certain other thoughts (e.g. that Murray loves Elsie). If we have
a classical vision of mind, the systematicity of thought is straightforwardly
explained by the combinatorial syntax and semantics of the cognitive represen-
tational system. The intrinsic connectedness of the dierent thoughts in ques-
tion results from the fact that the processing architecture contains a set of atomic
symbols alongside certain syntactic rules for recombining those symbols into
dierent molecular expressions. Now, Fodor and Pylyshyn argue that although
there is a sense in which connectionist networks instantiate structured states
(e.g. distributed connectionist representations have active units as parts), com-
binatorial structure is not an essential or a fundamental property of those states.
This leaves connectionist networks inherently incapable of explaining the sys-
tematicity of thought, and thus of explaining thinking. What such systems
might do, however, is explain how a classical computational architecture may
be implemented in an organic brain.
Bechtel agrees with Fodor and Pylyshyn on two key points: first, that where
systematicity is present, it is to be explained by combinatorially structured
representations, and second, that connectionist networks fail to realize combi-
natorial structure. He does not need to endorse Fodor and Pylyshyns claim that
all thought is systematic, however. For his purposes, all that is required is that
some cognitive activities (e.g. linguistic behaviour, natural deduction, math-
ematical reasoning) exhibit systematicity. One might gloss this by saying that,
for Bechtel, being a physical symbol system is a, not the, mark of the cognitive.
Bechtels distinctive next move is to locate the necessary combinatorial struc-
ture in systems of representations that remain external to the connectionist
network itself. Given the idea that our inner psychology should be conceived
in connectionist terms, this is tantamount to saying that the necessary combina-
torial structure resides not in our internal processing engine, but rather in
public systems of external representations (e.g. wri en or spoken language,
mathematical notations). As Bechtel (1994, p. 436) himself puts it, the property
of systematicity, and the compositional syntax and semantics that underlie that
property, might best be a ributed to natural languages themselves but not
to the mental mechanisms involved in language use. (Notice that, for Bechtel,
the mental is restricted to the inner. This is an issue to which we shall return).
For this interactive solution to work, it must be possible for the natural sen-
sitivity to statistical pa erns that we find in orthodox connectionist networks
235
to be deployed in such a way that some of those networks, when in interaction

with specific external symbol systems, may come to respect the constraints of a
compositional syntax, even though their own inner representations are not so
structured. Bechtels studies suggest that this may be achieved by exploiting
factors such as the capacity of connectionist networks to recognize and general-
ize from pa erns in bodies of training data (e.g. large numbers of correct deri-
vations in sentential arguments), plus the temporal constraints that characterize
real embodied engagements with stretches of external symbol structures (e.g.
dierent parts of the input will be available to the network at dierent times,
due to the restrictions imposed by temporal processing windows). The conclu-
sion is that by dividing the labor between external symbols which must con-
form to syntactical principles and a cognitive system which is sensitive to those
constraints without itself employing syntactically structured representations,
one can perhaps explain the systematicity . . . of cognitive performance
(Bechtel, 1994, p. 438).
How should we interpret the distributed solutions that Bechtel favours:
as examples of embodied cognition or as instances of cognitive extension?
Bechtel himself stops short of the extended option. Thus, as we have just
seen, he tellingly describes systematicity as a feature of cognitive performance
rather than as a property of the cognitive system, and states that the composi-
tional syntax and semantics might best be a ributed to natural languages
themselves but not to the mental mechanisms involved in language use (my empha-
sis). What this indicates is that, for Bechtel, the genuinely cognitive part of
the proposed solution remains skin-side. Lets see what interpretation we get,
however, once we apply the parity principle. If the envisaged system of syntax-
sensitive processes and combinatorially structured symbols were all stued
inside the agents head, we would, I think, have no hesitation in judging the
symbol structures themselves to be bona fide parts of the agents cognitive
architecture. Equality of treatment therefore seems to demand that the external
symbol structures that figure in the functionally equivalent distributed version
of that solution also be granted cognitive status. On the strength of the parity
principle, then, what we have here are models of extended cognition.5
Of course, the foregoing direct appeal to parity considerations takes us only
part of the way toward ExM. As we have seen, parity-based arguments remain
inconclusive until they receive backing from some mark of the cognitive that
sets the benchmark for parity. Its at this point that we see the impact of the
physical symbol systems hypothesis, conceived as specifying a mark of the
cognitive. For, I suggest, both the wholly inner and the environment-involving
versions of the Bechtel-style network-plus-symbol-system architecture are
instantiations of suciently complex and suitably organized physical symbol
systems. Since both exhibit that mark of the cognitive, both are cognitive
systems, and the la er is an extended cognitive system. Given the functionalist
236
character of the physical symbol systems hypothesis, such considerations

strengthen further our reasons for thinking that the fundamental contribution
of the body to cognitive theory is to be conceived in terms of implementational
materiality, not vital materiality. One way to appreciate the plausibility of this
picture is to reflect on the most obvious objection to it.
In response to the view just sketched, many cognitive scientists will want
to complain that the kinds of pa ern-matching and pa ern-completion pro-
cesses realized by connectionist networks are not equivalent to the syntactic
rules present in classical systems, implying that the analysis of the Bechtel
architectures as extended physical symbol systems is suspect. With all due
respect this is, I think, a failure of the imagination. It is of course true that the
network processes concerned are not explicitly rule-driven in a classical sense,
but two considerations strongly indicate that this is not the end of the ma er.
First, the keystone of Bechtels model is the thought that the networks involved
are genuinely sensitive to the constraints of a compositional syntax. Thus, pend-
ing good arguments to the contrary, one might insist that Bechtels networks
implicitly realize the rules in question, at least in the minimal sense that, in
this case (although not in others), classical-style rules will provide a perfectly
reasonable, high-level, idealized description of the networks processing activ-
ity. (The fact that there is idealization here should not concern anyone. For
one thing, idealization is part of scientific explanation. For another, as we have
seen, orthodox connectionist models are themselves abstract idealizations of
real brains.) Secondly, and from a more radical perspective, it may be that the
classical rules are not implicitly realized in the neural network alone. If we
think of those rules as principles that govern the skilled embodied manipula-
tions of certain external material symbols, it might be more accurate to think
in terms of dynamic sub-personal vehicles that include not just neurally-
implemented connectionist elements, but also non-neural bodily factors,
including physical movements. On either analysis of how the rules in question
are realized, the objection under consideration would fail.
A Parting of the Ways
In exploring the relationship between embodiment and cognitive extension,

I have presided over a parting of the ways between, on the one hand, ExM,
understood as involving an extended functionalist commitment to a kind of
open-ended multiple realizability, and, on the other, a particular strain of
EmbC that depicts the organic body as, in some way, intrinsically special in
the generation of cognitive phenomena. At root this fork in the theoretical road
may be traced to a fundamental disagreement over how philosophy and
cognitive science should conceive of the materiality of the body as just one
237
implementing substrate among possible others, or as a vital and irreplaceable

determinant of cognitive life. I have presented a case for thinking that we
should follow the ExM path to implementational materiality. But in this fast
moving and complex debate, wrong turns and dead ends will abound. Under
such circumstances, drawing up a road map will always be a hazardous task,
and I expect there to be many moments of disorientation and puzzlement
along the way, before we arrive at a detailed theory of the embodied and
extended mind.
238
13 Current Issues in the
Philosophy of Mind
Paul Noordhof
In the broadest terms, the issues which lie at the heart of discussions in
philosophy of mind have not changed and are unlikely to change in the near
future, or even, I hazard a guess, the quite distant future. People will still seek
to understand the nature of consciousness in its various forms; to understand
the nature of intentionality or, indeed, other ways in which our mental life
may concern the world around us; to describe, and account for, the special
access each of us has to our own mental lives; to scrutinize the basis of self-
consciousness; to worry about whether mental phenomena have an appropri-
ate causal explanatory impact, and so on. Nor do new philosophical theories
in these areas arrive thick and fast. Instead, at dierent times, dierent theories
and their motivations receive particular development and emphasis. These
facts about philosophical discussion give rise to unfortunate moments when
placed on the spot by inquiring vice chancellors and sceptical governments
or other, prospective, suppliers of research funding. Nevertheless, since we are
among friends here, we can acknowledge it without embarrassment.
Other awkward moments o en occur when some of the folk just mentioned,
or the others professing to have an interest in the role of philosophy in intel-
lectual life at large, remark that while the themes upon which it focuses are
big (with the slight suggestion, about some of them, that philosophers are on
a fools errand), the current contributions of philosophers are technical and
specialized, as if, given that we can all recognize the themes, we should all be
able to, without too much study, appreciate what progress has been made.
However, precisely because philosophical progress is one of developing our
understanding of particular types of theories by se ing them out in more detail
and/or developing the motivation for them, it is entirely unsurprising that the
statement of these developments will seem more specialized and technical
than is desirable. Philosophers should make every eort to help with this but
not to apologize for it.
Once you get down to the requisite level of detail for the progress and details
to show up, identification of the main themes currently in play and suggestions
about how these will develop, will be controversial (I guess it is no news that
239
philosophers love an argument) and more than likely receive dierent spins
by dierent philosophers working in the field. That might explain why my
overriding sense is one of trepidation. Nevertheless, I shall suppress this in
what follows and state as clearly as I can what seem to me the main points of
emphasis in a selection of the fields detailed above.
Specifically, I shall discuss how development in our understanding of phys-
icalism has more or less stabilized around an approach which reveals some-
thing rather interesting: first, some of the reason for believing it is undermined
while, second, it is more entrenched as a starting point because working within
this framework brings useful explanatory rigor. I shall discuss how the debate
about mental causation is slowly turning into one which focuses on the onto-
logical commitments of thought and talk about the mental, which purports
to be about physical properties which are not to be identified with those of
physics. Then I turn to a empts to dismiss dualist intuitions regarding the
explanatory gap between neural properties of the brain and phenomenal con-
sciousness independent of an admission of ignorance, or development of a
theory of phenomenal consciousness. I explain why it is increasingly recog-
nized that such a empts fail. Recent theories of phenomenal consciousness
have sought to explain it by appeal to representational properties and a separate
theory of subjective awareness (roughly, a theory of how we are conscious of
the phenomenal content determined by representational properties). I explain
how the appeal to representational properties has deepened our understanding
of the motivation for invoking qualia and issues in the philosophy of perception
and intentionality. I then outline how apparently distinct theories of subjective
awareness have converged. In the final section, I examine relatively new
approaches to the understanding of intentionality which a empt to come to
terms with the diculties that proponents of reductive theories of intention-
ality faced. In this context, I briefly touch on the putative normativity of the
mental.
I mourn not being able to discuss the many interesting developments in
our understanding of self-consciousness, self-deception, mental illness and
introspection as well as some of the fascinating texture of more specific mental
states such as that of imagination, auditory perception and so on. It proved
impossible to do so in a piece of acceptable length and some kind of structure.
Physicalism
The characterization of physicalism
At the risk of losing a few readers at the beginning, it seems to me that there is
more general agreement now as to how we should go about characterizing
240
physicalism. Identify some basic properties in terms of the subject ma er of

physics, specifically, a development of todays physics which resembles it.
Admi edly, talk of resemblance is vague but this is entirely as it should be.
It is quite conceivable that there will be developments of physics in which we
wonder whether the properties so characterized really count as physical in
any sense continuous with talk of the physical previously and conclude that it
is just not clear either way. Appeal to resemblance provides a way of steering
between the o -cited dilemma that, if we appeal to current physics to char-
acterize the physical, we may soon find that there are no physical things, when
current physics is superseded, and if we appeal to the ideal future physics,
then physicalism becomes trivially true.
The point is not that, if we simply appeal to current physics, we would
have no reason to believe in the truth of physicalism since, so understood, it is
likely to be false. As Andrew Melnyk points out, we might accept scientific
theories even though the likelihood of their being true is low in the absence
of rivals with higher probability of being true (Melnyk, 2003, pp. 2257). Nor
is this first element part of an a empt to provide a conceptual analysis of
physicalism. It need be no part of somebodys grasp of the concept of physical-
ism that they know that, in the circumstances, envisaged we may not know
whether or not it would be correct to conclude that it was false. Instead, it
is a considered judgement as to what is plausible to say about the truth of
physicalism in the circumstances, given conceptual knowledge and everyday
observations about our usage of this notion.
With this foundation for the characterization of physicalism in place,
a ention then turns to the connections between these properties (narrowly
physical properties) and other properties that, more broadly, we are inclined to
view as physical. It is here that controversy breaks out and its resolution will
have significant consequences for the legitimate means to evaluate whether
physicalism is true and the constraint it places upon understanding the nature
of the mental.
One way in which broadly physical properties could be related to narrowly
physical properties is, once again, resemblance. Properties as disparate as being
a mountain, being a table, being a cell and being an anteater, would all count
as physical in the broadest sense because they resemble narrowly physical
properties. However, the claim has only to be stated for its unsatisfactory
character to be felt. There are such huge dierences between these properties,
and huge dierences between any of them and narrowly physical properties,
that it is far from obvious exactly how these properties all resemble the
narrowly physical properties suciently to make them broadly physical. This
thought together with the idea that somehow physics is, at least in part,
concerned with the nature of things from which all the rest are composed,
suggests an alternative.
241
Just as broadly physical objects are composed from, and thereby nothing
over and above, arrangements of narrowly physical objects, so broadly physical
properties are those which are constituted from, and, thereby, nothing over and
above arrangements of, narrowly physical properties. The idea is relatively
unproblematic in the case of objects. Integrated spatial arrangements of small
scale physical objects seem naturally to result in large scale objects. In many
cases, physics will concern these small scale objects and other sciences focus
on the larger scale. Opponents of physicalism o en remark that physics con-
cerns the large scale too (Crane and Mellor, 1990). Wise physicalists will not
deny this. They will just insist that any object not recognized by physics will
be composed from objects which are so recognized.
Unfortunately, talk of constitution is not so easily taken across to properties.
If we take properties to be universals so that, for example, there is only one
property of being a hydrogen atom then it is hard to see how the property of
being CH4 methane can be composed of the properties of being a carbon
atom and being a hydrogen atom (Lewis, 1986). There are just two of these
la er properties. So it cannot be the case that the property of being methane
is made up of, at least, five elements: one property of being a carbon atom,
and four properties of being a methane atom.
If we take the relevant notion of property to be property instances, then this
case becomes less problematic. There is no diculty in thinking of the property
instance of being methane as composed, in part, from one property instance
of being a carbon atom, and four property instances of being a hydrogen
atom. Instead, the diculty is that the connection between narrowly physical
property instances and broadly physical ones doesnt seem invariably to fit
this model. In the case of CH4, the arrangement of the properties of atoms
identified is the way of being methane. There are many dierent ways in which
arrangements of instances of narrowly physical properties may be arranged to
make up an instance of a property of being a mountain, or a cell, or an ant-eater.
What is the connection between a particular way in which some narrowly
physical properties are arranged and the broadly physical property which is
said to result in these cases? This is the well known phenomenon of variable
realization applied to non-mental cases.
One kind of minimal answer appeals to supervenience. Supervenience
has been characterized in a number of dierent ways and ge ing clear on the
precise connections between all the various dierent alternatives is, perhaps,
not the most engrossing way of spending a Sunday a ernoon. A currently
popular one holds that physicalism is true of our world if a minimal physical
duplicate of it, as far as arrangements of narrowly physical properties are
concerned, is a duplicate simpliciter (Jackson, 1998, p. 12). For some word, w,
physical properties broadly conceived are just any which meet the following
condition. Given P is instantiated in w, then it is instantiated in all minimal
242
physical duplicates of w. The connection between arrangements of basic physi-

cal properties and broadly physical properties holds of metaphysical necessity.
If it held of merely nomological necessity, then any property nomologically
related to arrangements of physical properties and this might include ghostly
properties, properties of ectoplasm, properties of Descartess immaterial mind
and so on could count as physical. All that would be ruled out is that these
properties could not vary with arrangements of narrowly physical properties
unless the physico-psychological laws changed.
The formulations focus is upon what makes a world one in which physical-
ism is true because physicalism is doctrine primarily concerning the nature of
worlds. Concern with individual objects and properties is relevant only in so
far as their occurrence in a world is a way of physicalism about that world
being falsified. It also avoids thorny questions about exactly which arrange-
ment or arrangements of narrowly physical properties are responsible for the
instantiation of broadly physical properties. Nevertheless, as already remarked,
in the background is the thought that, for each broadly physical property,
there is one or more arrangements of narrowly physical properties which
metaphysically necessitate it. If broadly physical properties are nothing more
than arrangements of narrowly physical properties then it should not be the
case that the arrangement in question is present and yet the broadly physical
property is not. Constitution by instances is thus seen as just one example of a
general relationship between families of properties which reveals something
about the nature of the supervening properties and explains how they involve
nothing over and above the properties of the supervenience-base: those arrange-
ments of narrowly physical properties which metaphysically necessitate them.
But now the problems come thick and fast.
Appeal to minimal physical duplicates of our world addresses two dicul-
ties. The first is that physicalists do not want to deny the possibility of immate-
rial minds, only assert the possibility of material ones. By explicitly talking of
physical duplicates, we exclude worlds with immaterial minds from consider-
ation. The second is that we wouldnt want to conclude that physicalism is
false in our world if there is a world which matches our world in arrangements
of narrowly physical properties but which diers in mental properties due to,
in that world, the additional occurrence of non-physical ectoplasm. So long as
none of this stu is present in our world, physicalism would be true. Talk of
minimal physical duplication sets such worlds aside as well because, while
these worlds may duplicate the arrangements of narrowly physical properties,
they dont stop right there. They have ectoplasmic properties in addition whose
instantiation does not come along with the instantiation of arrangements of
narrowly physical properties.
The appeal to minimal physical duplication gives rise to our first diculty.
Suppose that there is a non-physical property (by this I mean neither narrowly
243
nor broadly physical) that breaks the connection between a particular arrange-
ment of narrowly physical properties and the mental property for which they
are putatively sucient: a blocker. Then, intuitively, the connection between
the arrangement of narrowly physical properties and the mental property is
not tight enough. The possibility of the connection being blocked seems to
imply that the connection is merely nomological and, hence, no dierent from
that which holds between arrangements of narrowly physical properties
and the non-physical mental properties characteristic of emergent dualism.
Nevertheless, appeal to the idea of minimal physical duplicates sets aside such
worlds because the relationship is broken by non-physical properties and,
hence, the possibility does not reveal that physicalism is false (Hawthorne,
2002, pp. 1046).
Some argue that, as a result, physicalism should be understood in terms
of suciency in the absence of blockers and, hence, reject the intuitive verdict
that physicalism is false when there is the possibility of a non-physical blocker
(e.g. Leuenberger, 2008, pp. 14860). Prima facie, this is unacceptable for the
reason identified above. The possibility of blocking suggests that the connec-
tion is no be er than emergent dualists endorse. Moreover, if the connection
between arrangements of narrowly physical properties and mental properties
is loose enough to be blocked by non-physical properties, then the connection
between the arrangements of narrowly physical properties and mental proper-
ties holds of nomological necessity, in which case, there is no reason the con-
nection could not also be blocked, if the physico-psychological laws were
dierent, by the presence of some of the same narrowly physical properties
which, in our world (lets say), were responsible for the instantiation of mental
properties. So there is no worry about excluding the worlds with non-physical
blockers by talking of minimal physical duplicates. The looser connection
would show up in the minimal physical duplicate worlds too. (I discuss one
line of objection to this point under the second diculty that this formulation
of physicalism faces below.)
Blockers dont have to be causal blockers, although talk of stu which one
shouldnt add when preparing food, or algoplasm, which makes phenomenal
properties disappear, suggests that this is standardly how they are conceived
(Hawthorne, 2002; Leuenberger, 2008; also Montero in this volume). Instead,
blockers may disrupt a relationship in which at least one of the relata is, partly,
relational. For example, I might have the property of being alone in a room
which I (or my counterpart) do not possess in a minimal narrowly physical
duplicate of my world because, in that world, there is a non-physical poltergeist
in it with me. In these circumstances, there is no reason to expect that, in a mini-
mal physical duplicate world with dierent laws, the property of loneliness
may fail to be instantiated because one of the physical properties has become a
blocker. However, all this shows is that we should refine our understanding of
244
minimal physical duplicate. Minimal physical duplicates arent those which

simply stop right there as far as the arrangements of narrowly physical proper-
ties are concerned, they also keep all the truths concerning properties which
arent instantiated compatible with that arrangement the same. This does not
trivialize the characterization of supervenience because we are not including,
in the characterization of a minimal physical duplicate, that it is a duplicate in
all other positive properties.
With this recent line of concern set aside, two main lines of worry have been
expressed about the proposed characterization of physicalism. The first is
that it would fail to characterize physicalism if a powers ontology were true
(OConnor, 1994; Wilson, 2005). According to this ontology (which may, actu-
ally, be the case), properties potential to stand in causal relations (their causal
profile) are an internal fact about the property. They are not fixed by indepen-
dently holding laws. The problem arises if a powers ontology is taken to include
the thesis that properties causal profile are essential to them. In which case, it is
not possible for a certain property to be instantiated in certain circumstances
and the relevant causal relations not hold. Suppose that emergent dualism is
true and, hence, that some arrangement of narrowly physical properties cause,
but are distinct from, a non-physical mental property. Then it will be essential
to these properties that they cause the non-physical property. So, it will be
true that any minimal physical duplicate world will be a world in which this
non-physical property occurs.
There are two points to make about this concern. The first is that a powers
ontology need not be commi ed to taking the causal powers which, in part,
or in whole, characterize properties as essential. The distinction between
intrinsic and essential applies as much to properties as it does to individuals.
Just as Socrates shape is not essential but intrinsic to him, so a propertys causal
potential need not be essential to it in all respects even if intrinsic to it.
Specifically if emergent dualism is true and there are fundamental physico-
psychological laws, that part of a propertys causal profile which concerns
its generation of non-physical mental properties might be absent and yet the
narrowly physical property arrangement still be present (see Noordhof, 2010).
Second, as we shall see, the principal reason for being an emergent dualist
stems from the nature of phenomenal properties (those properties that deter-
mine what it is like to undergo a conscious mental life). Taking even these
properties to be functional properties that is properties characterized in terms
of causal profile has been a way of avoiding supposing that phenomenal
properties are non-physical. In which case, we do not need to provide a defini-
tion of physicalism which excludes emergent dualist worlds in which a powers
ontology is true. There are no such worlds.
By far the most familiar line of resistance to the characterization of physical-
ism in terms of supervenience alone concerns whether it is enough or whether
245
we need to appeal, in addition, to the idea of there being an explanatory con-

nection between the arrangement of narrowly physical properties and the
mental property in order for the la er to be appropriately characterized as
broadly physical. The easiest way to express the worry is to consider what we
might say if there were a capricious God who made sure that, in all possible
worlds, if certain arrangements of narrowly physical properties were instanti-
ated, then non-physical mental properties would be. In that case, wouldnt
the favoured account pronounce that the properties in question are physical
when it shouldnt? One answer is to insist that just as God cant make the
putative necessary truths of mathematics false (say), so likewise he cannot
make the following putative necessary truth false concerning non-physical
properties: It is possible that a certain arrangement of narrowly physical prop-
erties is instantiated and the non-physical mental property is not. If he really
could fix it that all possible worlds contain the non-physical mental property
co-instantiated with the arrangement of narrowly physical properties, this
would be a refutation of the possible worlds analysis of possibility rather
than an objection to this analysis of physicalism since, clearly, they could have
come apart if God hadnt held them together. A second answer is just to dier-
entiate between God-enforced co-instantiation and other sorts with no further
appeal to explanatoriness. More generally, we might hold that the crucial dier-
ence is between internal and external metaphysical necessitation. Externally
sourced metaphysical necessitation for instance, in Gods will reveals noth-
ing about the character of the properties which supervene upon the narrowly
physical properties. Internally sourced metaphysical necessitation does. This
response does not appeal to any fuller account of explanation of the kinds
canvassed below.
Unfortunately, the response doesnt help us to capture the dierence
between emergent dualism and ethical non-naturalism in ethics, which is the
other reason folk cite to show that supervenience needs supplementation (e.g.
Horgan, 1993, pp. 55760). It is argued that moral properties are metaphysically
necessitated by arrangements of natural properties. If physicalism is commi ed
to mental properties being natural, then they must have some feature which
distinguishes them from non-natural moral properties.
I have addressed this argument elsewhere. In brief, the way in which pro-
ponents of ethical non-naturalism took them to be non-natural is not a way
which need distinguish them from mental properties. G. E. Moore, the most
famous non-naturalist, cites as his considered opinion that, while there are
many dierent intrinsic natural properties, each of which is ought-implying,
there is no common property (apart from the disjunction) entailed by them all
and also ought-implying. That is, there is no separate ought property even
though there are many dierent ought-implying properties (see Moore, 1942,
p. 605; Noordhof, 2003a, p. 96). Thus the argument is that there is variable
246
realization with no common property. What Moore is not taking as the basis
of his ethical non-naturalism, is that there is an intrinsic action-guiding or
ought property. Yet, it is the la er which would threaten the characterization
of physicalism. In those circumstances, we would have Moore claiming that,
since there are intrinsic ought properties, non-naturalism is true while, at the
same time, we would be asserting that the fact that other properties stand in
the very same way to arrangements of narrowly physical properties shows
that these other properties are natural.
Roll forward to the most prominent current defender of ethical non-
naturalism: Russ Shafer-Landau. He describes his position as taking the same
approach to moral properties as non-reductive physicalists do to the mental
(Shafer-Landau, 2003, pp. 728). He accepts that moral facts are intrinsically
normative but seems to take this as no more problematic than the properties
of phenomenal consciousness. It is not that he supposes that this intrinsic
normativity is what makes moral facts non-natural and, hence, a problem for
those who seek to characterize physicalism by appeal to the same kind of
relationship of supervenience. So the proposal does not need defence against
those who appeal to the very same relationship to characterize ethical non-
naturalism.
Even if we dont need to appeal to some kind of explanatory relationship, it
is reasonable to ask whether there is some kind of explanation of the relation-
ship. I do not want to rule this out, indeed, far from it. The point of defending
the supervenience-only approach is simply to keep our explanatory options
open. For example, consider currently the most popular explanation on oer:
the subset view. According to it, a property F is realized by another property G
if and only if the causal powers of an instance of the property F are a subset of
the causal powers of an instance of the property G (This is a rough preliminary
formulation, see Shoemaker, 2007, pp. 1213, 2231 for details). It is question-
able whether this is, in fact, true of any variably realized property (e.g. see
Noordhof, 1997, p. 246; Noordhof, 1999b, pp. 11314). However, for the sake
of argument, suppose that it is. Taking it as a general requirement places restric-
tions on the kind of explanation we should look for in every case in which we
suppose that the mental supervenes upon the narrowly physical.
Endorsement of it is behind Jaegwons Kims conclusion that there is a
non-physical residue to the mental: qualia (a particular kind of phenomenal
property, see below). He takes it that reduction of the mental to the physical
requires that it is functionalizable and accepts that, in the case of qualia, exhaus-
tively specifying their nature in functional terms is not going to be plausible
because of the possibility of spectrum inversion (Kim, 2005, pp. 16573). The
conclusion, of course, does not follow if other kinds of explanation are allowed.
If physicalism only requires that mental properties are metaphysically necessi-
tated by physical properties, then there are various ways in which this might
247
hold: functionalization is only one. Another might be macro-micro reduction

and who knows what other options there are!
As the examples illustrate, the appeal to explanation in the characterization
of physicalism does not imply that the envisaged relation between arrange-
ments of narrowly physical properties and broadly physical properties is
epistemic (for another view, see Montero in this volume). Instead, it serves to
characterize a certain ground for the relationship of metaphysical necessitation
to hold, a ground which then may be cited in explanation. There is nothing
epistemic about macro-micro reduction or functionalization of a property and
accounting for its instantiation in terms of arrangements of narrowly physical
properties together with, if required, laws.
Thus there are two considerations in favour of the proposal. First, it is the
minimum required to capture the idea that there is nothing over and above
arrangements of the physical without becoming commi ed to particular kinds
of explanation of this. Second, it allows for the possibility that we are ignorant,
or even cognitively closed, to the exact way in which mental properties may be
no more than an arrangements of physical properties and this is desirable if
we have no substantial intellectual grounds for being more restrictive for rea-
sons that will become apparent when we consider appeals to the explanatory
gap in favour of dualism below.
Much a ention has been paid to the proper formulation of physicalism.
It might reasonably be wondered why we should care. Suppose a more restric-
tive account of physicalism is oered as a result of which it is concluded that
physicalism is not true. What is the harm in that? To an extent, I strongly sym-
pathize with the line of thought behind this objection. Nevertheless, I think
there is merit in the role that physicalism plays even though it may not ma er
precisely what we call the doctrine that plays the role.
One motivation in favour of physicalism is undermined by my characteriza-
tion of it. The a raction of physicalism is o en allied with being responsive to
scientific development and the providing a naturalistic account of the nature
of the mental. Let there be no mystery mongering! Let us seek to integrate
our philosophical reflections with the findings of science. However, insisting
that physicalism will only be true if the resulting physics in some way resem-
bles our own allows for the possibility that there may be scientific develop-
ments which falsify physicalism. Under those circumstances, wouldnt our
situation be just as good? Wouldnt a scientific theory of consciousness which
took advantage of these new developments be just as substantial as one which
was compatible with the truth of physicalism. In which case, what is the
merit of physicalism as opposed to robbing it of its negative connotations
scientism? Taking mental properties either to be scientifically recognized
properties or supervening upon scientifically recognized properties?
248
Physicalism, as a sub-category of scientism so understood, is unlikely to

have any more virtues than scientism so, in that sense, lacks additional support.
Nevertheless, it plays a significant role in another respect. Physicalism places
some constraints upon how we should, currently, seek to explain our mental
lives. The problem is that, without such a constraint, it is hard to see how one
could oer something substantial in the area. It is all too easy for the philo-
sopher to postulate as a non-physical substance or property whatever it
is which has the target property that it is allegedly hard to explain in physical-
istically acceptable terms, thus, making it a case of lo, and in one bound, Jack
was free. The diculties are particularly apparent in the case of those who
seek to do the intellectually decent thing and give an account of how talk of
non-physical properties may be put to explanatory work by formulating a
viable theory of panpsychism or relating it to developments in quantum
mechanics (for discussion of this, see Seager, 1999, pp. 21652).
Thus, it is not really a desire for simplicity which characterizes the physicalist.
It is rather that, when dualists provide support for their doctrine, they justify
it in terms of there being things which resist integration into the sciences, prop-
erties such as consciousness or being free that cannot be explained in terms of
other properties (e.g. see Mawson in this volume. Li le positive is said about
the nature of the explanation. It is rather that Dualism is a response to the inabil-
ity to explain in other terms. Now I dont want to rule out the possibility that
no physicalistic explanation of consciousness and the rest is available. It is
simply that, since it is hard to prove a negative, it is reasonable to continue
with the physicalist perspective until such time as the dualist explains how
their proprietary form of explanation works.
This point is not quite Philip Pe its point that the physicalism should be
presumed as a worst case scenario (Pe it, 1993). Will an account of conscious-
ness and the rest be available if physicalism is true? It is rather the methodogi-
cal point that dualist accounts of the kind of explanation they oer have, for the
most part, been conspicuous by their absence unless you count as an explana-
tion of property P, the existence of the property P. I dont think we should go out
of the way to make ourselves intellectually suer (as Pe it seems to suggest).
However, we need to be sure that what we are oering as an explanation is
subject to constraints at least as rigorous in its own terms as those imposed by
the physicalist. In the absence of this, physicalism commands our a ention. As
a result, both sides in the debate will, and should, have an interest in exploring
further the kinds of explanations that they respectively oer. Dualists, because
only then will they push home their advantage in accounting for aspects of our
mental life; physicalists, because any development of the range of explanations
on oer promises to provide relief from their assailants. An example of this
will be discussed when we come to phenomenal consciousness.
249
Mental causation and physicalism
It is well known that the main argument in favour of physicalism is the

so-called over-determination argument (set out in Ravenscro in this volume).
It has probably received its best articulation only recently and it is contentious
whether it is sound (Papineau (1993a), Chapter 1). Those who question the
plausibility of physicalism either question whether it is unacceptable to allow
that there is systematic over-determination or question the evidence that the
world is causally closed (Mellor, 1995, pp. 1035; Lowe, 2000a or 2000b). The
la er claim, in particular, it seems to me is in need of further scrutiny. What
may be agreed to on all sides is that a particular arrangement of narrowly
physical property instances is sucient to fix the probability of any eect.
Nevertheless, some of these arrangements bring with them, according to
the emergent dualist, non-physical mental properties. It is an open question
whether these arrangements fix the probability of the eect partly in virtue of the
non-physical mental properties, for whose instantiation they are responsible.
Thus, there is a slide between two claims:
(C1)Every target eect has its probability fixed by arrangements of nar-

rowly physical property instances alone.
(C2)Every target eect has its probability fixed by arrangements of nar-
rowly physical property instances, perhaps partly in virtue of the
instantiation of non-physical properties.
Either of (C1) or (C2) is will be compatible with the evidence that the
physical world is causally closed. However, only the first would establish that
non-physical mental properties, if they exist, are epiphenomenal.
A lot of energy has been focused, though, on whether a particular form
of physicalism, non-reductive physicalism, is compatible with the ecacy of
mental properties. If it is not, then non-reductive physicalism has no advantage
over property dualism, given the success of the over-determination argument.
Suppose that mental properties supervene upon physical properties in the
way identified above. Suppose that, in particular, an instance of a physical
property P1, p1, necessitates an instance of a mental property M1, m1 and that,
for any candidate eect of m1 we accept (from considerations of the causal
closure of the physical) that it is caused by p1. (As others do, I simplify. p1
would obviously be an arrangement of narrowly physical properties). To fix
ideas, consider m1 to be an extended period of pain, which eventually becomes
unbearable, from leaving ones hand on a hot plate, p1, a firing of neurones
in the brain specifically related to pain and the target eect: the behaviour of
withdrawing ones hand from the hot plate.
250
The challenge can come in two forms. The first concerns whether the instances
of mental properties can be ecacious if the instances of physical properties
are. As the question is o en put, is there any further work for the mental prop-
erties to do? (For details of this kind of argument, see Ravenscro , in this
volume). The second, and more pressing, challenge is whether, even if these
instances of mental properties are ecacious in virtue of their relationship
with instances of narrowly physical properties, they are ecacious in virtue of
being mental properties.
Responses to the first challenge emphasize that the instances of mental
properties are either identical to instances of narrowly physical properties
or realized by arrangements of them (identity: Robb, 1997; Macdonald and
Macdonald, 1986, 1995; Macdonald, 1992; realized: Noordhof, 1999b, 2010;
Shoemaker, 2007). Appeal to identity is a ractive because it reproduces the
solution to how mental events and states can be ecacious by being token
identical to physical events and states for property instances. They are men-
tioned in a footnote by Campbell but assimilated to the event response he
favours on behalf of Davidson (Campbell, in this volume). However, the dis-
tinct virtue of the property instances approach is that it allows for finer grained
identifications of relevance than allowed by Campbell, given his rejection of
the property exemplification view of events. The main problem is to justify
the claim that mental property instances are identical to physical property
instances, given that they are distinct properties.
Appeal to realization avoids this diculty, and is independently a ractive
because it seems implausible to say that instances of mental properties are
identical to narrowly physical properties as opposed to realized by them. Such
approaches, though, are said to come unstuck when they try to a ribute to
the mental properties ecacy that they concede belongs to arrangements of
narrowly physical properties in virtue of which they are realized.
There are various a empts to deal with this relying, for example, on the
claim that the causal profile of mental properties is a subset of the causal profile
of its realizer or that they share ecacious property instances (Shoemaker,
2007; Paul, 2007). Opponents of the first have asserted that instances of mental
properties are redundant because they recapitulate a subset of the causal role
that the physical property instance already brings (e.g. OConnor and Churchill,
2010, pp. 513). This seems unfair. If mental properties have a causal profile
which is a subset of the causal profile of the narrowly physical properties, then
it is reasonable to hold that anything which is a manifestation of this profile
reveals the ecacy of mental property instances. It is in virtue of this part of
the profile that a narrowly physical property is ecacious. Opponents of the
second have questioned whether instances of mental properties do anything if
their ecacy is the result of, for example, sharing an instance of mass with an
251
arrangement of narrowly physical properties. The property instance they share

seems straightforwardly narrowly physical (Ney, 2007, pp. 5012).
Both these kind of replies and objections to them share a common assump-
tion that I think needs to be questioned. They seem to start from a position
in which they grant, for the sake of argument perhaps, the truth of non-
reductive physicalism and, given that, worry that, while there may be instances
of broadly physical properties, it is questionable whether there are instances
of broadly physical causation. The absence of the la er is taken to undermine
the case for the existence of mental properties as some kind of independent
argument.
The true position to me seems closer to this. If you allow that there are
instances of broadly physical properties, then there is no reason to come to a
dierent verdict with regard to one particular kind of broadly physical prop-
erty, the relation of broadly physical causation (on the assumption that causa-
tion is a relation; for those who deny it is, then that particular non-relational
broadly physical property which is causation). Similarly, if you are inclined
to say that instances of mental properties are arrangements of instances of
physical properties and, in that sense, realized by them, then mental causation
can be understood in the same way (e.g. Noordhof, 1999a, 2010).
The assumption is devastating to the prospects of non-reductive physical-
ism. Its friends try to identify something which counts as a piece of narrowly
physical causation for example, a subset of the powers or the causal con-
sequences of shared narrowly physical property instances which its enemies
then deny is the work of the mental. Instead, we should understand that
arrangements of narrowly physical properties will result in an arrangement
of causal relations which, if the former is the supervenience-base for a mental
property, the la er is the supervenience-base for mental causation.
There may be an issue about the existence of mental properties. For instance,
suppose, as some argue, they might be taken to be determinable physical
properties, where the arrangements of narrowly physical properties might
be taken to be the determinates just as colour, say, is the determinable of a
particular shade of red (the determinate) (e.g. Yablo, 1992). Then if determin-
able properties dont exist, mental properties dont exist either. Instead, we have
sentences involving mental predications such as having a pain in his hand
which are made true by the determinates: arrangements of narrowly physical
properties. What I claim is puzzling is to think that you can derive an issue
for the existence of mental properties from some reflections about causation as
if it isnt a property to which the same considerations apply (see Noordhof,
2010, p. 85).
As a result, I suspect that the main issue that is going to confront non-
reductive physicalism in the near future is not concerned with ecacy per se
but which derives from reflections upon truth-making. If the truth-makers of
252
mental statements by the lights of the non-reductive physicalist are arrange-

ments of physical properties which metaphysically necessitate the truth of
these statements, then there is no reason to postulate the existence of mental
properties as truth makers. There is simply mental predication and mental
statements (for a coherent statement of this view, see Heil, 2003).
I can imagine two kinds of response to this issue. The first is to refine our
understanding of the truth-making relation. The standard approach to truth-
making is to take the truth-maker to involve metaphysical necessitation of the
truth of some target statement. Instead, it might be thought that truth-making
involved appeal to the in virtue of relation which cuts finer (Rodriguez-Pereyra,
2006). The second is to recognize some additional dimension of assessment
which validates allowing for the existence of supervening properties. That is
my own favoured approach and brings me to the second challenge concerning
mental causation for non-reductive physicalism.
The question of whether it is in virtue of being the realization of, or instan-
tiation of a mental property, that an arrangement of physical properties has a
certain eect implicitly involves a level of generality. The precise statement of
it is technical (see Noordhof, 1999a), but intuitively the thought is that if a
certain type of target eect is caused as a result of a number of dierent
realizations of that mental property indeed all such realizations in conducive
circumstances then the explanation for this is that it is in virtue of the mental
property that that type of eect is caused.
Truth-making may well be one important element in the proper understand-
ing of the ontology to which our statements commit us. But as important,
it seems to me, is providing rational support for inferences we are inclined to
make. To put it crudely, and joining in the myth that bulls care about red and
are not colour-blind, the motivation for recognizing the existence of the deter-
minable red, as well as the determinates corresponding to dierent shades of
red, is that assigning one determinate or another to a cloak will not provide
us with any grounds for taking other shades of red cloak also to enrage bulls.
For the la er, we need to appeal to the property of red. Suppose that, for the
sake of argument, the causal powers of property of being scarlet have, as a
subset, those causal powers which are a ributed to the property of red, then
rather than suppose that means that there is no need to postulate the existence
of red because they are conveyed by the property of scarlet, we should conclude
that scarlet conveys them by being, among other things, an instance of the
property of being red and, thereby, have causal powers beyond its determinacy
(for more discussion see Gille and Rives, 2005; Noordhof, 2010, pp. 868).
Therefore issues about mental causation raise, and can only be appropriately
resolved, if we consider more fully, the dierent ways in which our ontological
commitments may manifest themselves and avoid the mistake of only focusing
on a certain issue, viz. truth-making.
253
Phenomenal consciousness
It has become standard to distinguish between what Ned Block dubbed access
consciousness and phenomenal consciousness. A state is access conscious if and
only if it is poised to play a certain role in our practical deliberations or cogni-
tions (e.g. see Block, 1995, p. 231). Obviously quite a lot is packed into talk of
poised here. Intuitively, some state may be poised to play a role, although
we are not conscious of its presence, because the state has an impact on the
workings of our minds at the relevant points. So less illuminatingly we may say
that access consciousness just involves a being aware that we have a certain
state, or being aware of it. I dont want to presume here that access conscious-
ness is awareness of facts rather than objects. Nothing more is added to this
awareness. By contrast, a state is phenomenally conscious if, and only if, there
is something it is like to be in that state.
Many reductive theories of phenomenal consciousness break down phe-
nomenal consciousness into two components. The first component concerns
the content of phenomenal consciousness, what we may dub the phenomenal
content (some call it phenomenal character). The phenomenal content is just a
characterization of anything it is like for a subject to undergo the mental life
for example, his or her pain, the taste of banana cake, the sound of the waves
breaking on the shore, perceiving that there is a desk lamp lit on the desk, the
sense of yearning to go outside. All of these count. To introduce a stipulation
now, which I will discuss in more detail later, let phenomenal properties
be those properties which determine the phenomenal content of our mental
states. Thus, there is a phenomenal property of my experience of the waves
breaking on the shore which determines that the phenomenal content of my
experience is the waves breaking on the shore. Let manifest objects and proper-
ties be those things which our experience concerns, for example, the waves
(object) breaking (property).
The second component derives from the fact that it seems at least possible,
according to some theories of phenomenal content, that states should have
it and yet not be conscious because the subject is not aware of the state instan-
tiating the relevant content. So there is a story to tell about what makes the
subject aware of the phenomenal content of a particular state he or she is in.
This second component is o en called subjective awareness. I will only discuss
access consciousness in the context of this la er notion since, arguably, a theory
of access consciousness is a theory of subjective awareness.
Discussion of phenomenal consciousness in recent years has centred on
two dierent issues. The first concerns the proper explanation of why it
seems that phenomenal consciousness resists explanation in a way which is
acceptable to physicalists. The second concerns specific theories of phenomenal
254
consciousness, in particular those which seek to explain it in terms of the

representational properties of mental states.
Phenomenal properties and the explanatory gap
In discussing whether phenomenal properties are compatible with physical-

ism, o en the main focus is Frank Jacksons knowledge argument (Jackson,
1982, 1986). The thought experiment is discussed briefly in Hu o (in this
volume) and, in detail, by Ravenscro (in this volume). However, just as much
as Thomas Nagels discussion concerning what it is like to be a bat, Jacksons
argument contains elements which obscure the ma er it is meant to illuminate.
It seeks to provide an answer to the question about whether phenomenal prop-
erties are, at least, broadly physical by consulting us on the circumstances
in which a subject, usually Mary, would obtain knowledge of what it is like
to experience the colour red (Nagel, 1974). It is reasonable to suppose that a
necessary condition of such knowledge is that one either has knowledge by
acquaintance of the colour in question or is able to imagine it (Nemirow, 1990;
Lewis, 1990; Conee, 1994; Noordhof, 2003c). It is quite clear that a subject need
not be able to have either through reading about the nature of red experiences
in books of neuroscience. So the thought experiment cannot be taken to show
anything about whether there is a non-physical fact concerning what it is like
to experience something red rather than something about the nature of know-
ledge of what something is like. It also potentially fails to distinguish between
two options: Mary doesnt know what it is like because she is ignorant of some
non-physical fact or she fails to possess a concept of what it is like to experience
red (never having done so).
Discussion focusing on the explanatory gap is more direct (Levine, 1983).
Here the emphasis is on the fact that we lack, and appear to need, an explana-
tion of how, if physicalism is true, an experience of red is realized by one
arrangement of narrowly physical properties and an experience of green is
realized by another arrangement of narrowly physical properties. If the discus-
sion in The Characterization of Physicalism section above is correct, then no
assumption need be made about the kind of explanation required; nor should
we conclude that, if no explanation is forthcoming, then physicalism is false.
Nevertheless, acknowledging that there is a gap, lets the dualist in to propose
an account of what that is, so that needs to be taken seriously.
My defence of a characterization of physicalism in terms of, implicitly,
metaphysical necessitation alone may suggest that an even more direct route
to the question of whether phenomenal properties are physical is by consider-
ing zombie arguments: those which rest on the claim that it is possible for a
255
physical duplicate of me to lack phenomenal consciousness (see Chalmers,

1996). However, such arguments invite the response Why should we take
seriously our views about what may be possible in such a complex ma er?
This forces us into a debate about the degree to which our modal intuitions
are rational grounds for arriving at a view about the fundamental nature of
reality.
Focus on the explanatory gap provides a dierent kind of answer. Talk
of metaphysical necessitation has given us some indication of the kind of
explanation we are looking for that is, not merely a causal/nomological expla-
nation that could also be oered by the emergent dualist. Nevertheless, the
restricted nature of this assumption means that the following options are
kept open. The explanatory gap reveals an ontological gap (Chalmers, 1996), or
it is a gap in our knowledge, a conceptual gap, or, as I shall characterize it, an
experiential gap.
Until perhaps recently, the most popular approach has been to take it to be
a conceptual gap. Features of our concepts of phenomenal properties, it is
argued, account for why we take there to be an explanatory gap when, in fact,
there is no ontological gap. Although the approaches dier in emphasis, there
is surprising agreement on the general feature which does the work. All agree
that our concepts of these properties phenomenal concepts directly refer to
the properties they concern that is, they refer without the aid of descriptions
true of the properties in question (Sturgeon, 1994; Papineau, 1998a; Tye, 1999).
Some add to this the thesis that states with phenomenal properties such as
our experience of a particular pain are, at the same time, tokens of the
phenomenal concepts of these states (Balog, 1999, p. 525). Although the precise
formulation of the agreed-upon point diers, the desired upshot is that there
are no descriptions that we can know a priori apply to a particular phenomenal
property. There is simply the fact that we have a concept which does so apply.
The next step in the argument is to suggest that the standard form of
theoretical identification which would bridge an explanatory gap is:
(1)M has causal role R (an a priori truth)

(2)A(P) occupies causal role R (an a posterior truth)
Therefore,
(3)M = A(P).
Here M stands for some phenomenal property, and A(P) for an arrange-
ment of narrowly physical properties which are supposed to realize the
phenomenal property. Because there are no descriptions we can know a priori
are true of M, we cannot know a priori that M has a causal role, R. Therefore,
256
such theoretical identifications are unavailable for phenomenal properties and

there is no way in which the explanatory gap can be bridged. We have an expla-
nation for why there is an explanatory gap without conceding that there is an
ontological one. Note that even if you took explanatory identifications to appeal
to dierent descriptions about M and A(P) than their causal role, the point
about direct reference promises to deliver an account of the explanatory gap.
Unfortunately, the account has seemed more and more unsatisfactory.
Quite a few have begun wondering whether there is any substantial account
of the workings of phenomenal concepts to be given (e.g. Tye, 2009, Chapter 3).
Ill allow, for the sake of argument, that there is. Here are two points which are
increasingly familiar even granting this. First, since we are seeking to relate
all the various Ms ultimately to arrangements of narrowly physical properties,
R must be characterized in terms of the la er. But it is very unlikely that we
know a priori any R characterized in terms of narrowly physical properties
for any non-narrowly physical properties and not just phenomenal ones. The
standard illustration is water. Moreover, there is no explanatory gap regarding
whether water is H20, banished when we formulate the appropriate R by what-
ever means, and use it to identify H20 (Block and Stalnaker, 1999, pp. 1323).
The impression that we havent identified the root of the explanatory gap is
reinforced by the second point. Indexical or demonstrative concepts simply
dont have their reference fixed by description. Yet, we dont get explanatory
gaps arising to the same degree here. For example, consider John Perrys
example of the messy shopper. John Perry looks for the messy shopper who
is leaving a trail of flour in the local supermarket only to discover that he
himself is the messy shopper. When he thinks to himself I am the messy
shopper, he does not feel plagued by an explanatory gap. How could it be
that I am the person who is a messy shopper? Yet, presumably that is what
the proponent of the kind of account of the explanatory gap I have detailed
above should expect, given their theory as to its source.
To my knowledge, no satisfactory replies have been given to these concerns.
They throw into considerable doubt the supposition that the explanatory gap
is a conceptual one. Of course, in one sense, if physicalism is true, it must be.
If the explanatory gap does not correspond to an ontological one, then state-
ments about phenomenal mental states must express dierent concepts from
those which occur in statements about the brain. The point is rather that, in
recognizing this fact, we still have no explanation as to why we feel that there
is an explanatory gap.
An alternative approach to the explanatory gap is to account for it in terms
of our ignorance of some crucial facts. One kind of ignorance would involve
being ignorant of the narrowly physical properties which would explain the
gap. Another would be that there is a way in which we can explain how
one type of property supervenes upon, or is realized by, another which we
257
have yet to understand. The proposal is perhaps most famously associated

with the name of Colin McGinn, although versions of it have now been
expounded by Joseph Levine and Daniel Stoljar, and it was even present in
Nagels original article on what it is like to be a bat (McGinn, 1989; Levine,
2001; Stoljar, 2006; Nagel, 1974).
The principal dierence between McGinn and Stoljar concerns whether
we will forever be ignorant given the very structure of our perception and
cognition, or whether we are just ignorant at the moment. Stoljar is quite right
to observe that we dont need to make the former stronger claim in order to
have a potential account of the explanatory gap. This is a good thing because
McGinns argument that we cannot even arrive at a theoretical understanding
of the property which accounts for the link between the brain and phenomenal
consciousness is questionable unless the property in question is supposed, in
some way, to be implied by perception and thereby have a character very close
to that which we experience (Stoljar, 2006, pp. 923).
Nevertheless, there are problems that the ignorance approach must face.
The first is that physics just provides us with properties identified in terms of
structure and dynamics. It has no use for intrinsic qualitative properties. If
something like the la er is required to explain the nature of consciousness, then
it is not that we are ignorant of some narrowly physical property, it is rather
that panpsychism is true. We are ignorant of some qualitative element which
escapes physics but is omnipresent in the natural world. Now it is true that
the claim that the missing element must be qualitative has not been made out.
Nevertheless, in pu ing forward an ignorance account of narrowly physical
properties which ignorance has been made manifest to you by the fact that
you cant explain phenomenal consciousness at least raises the possibility
that, whatever you are ignorant of, is not a narrowly physical property, or
arrangements of the same.
Second, as Levine emphasizes, part of the problem is not that we are
ignorant of the narrowly physical properties, or the means by which they are
arranged, which would explain phenomenal properties, it is rather that we
have no idea of what kind of explanation should be oered (Levine, 2008), or, at
least, proponents of the ignorance approach do not make clear the kind of
explanation that they think that we should provide, of which we are ignorant
of the constituents. Because of this, it is hard to be certain whether we are
missing the vital ingredients.
This brings me to the third kind of approach to the explanatory gap. The
basic idea is that what we fundamentally need to explain is why our experience
of arrangements of narrowly physical properties are very dierent from the
experiences which are supposed to be these arrangements of narrowly physical
properties. Once we are in a position to account for this dierence, we will be
able to bridge the explanatory gap.
258
The basic form of the answer seems obvious and remarking upon it is
quite familiar. It is that, obviously, the reason why these things seem to be very
dierent is that when we experience the arrangements of narrowly physical
properties that, allegedly, realize an experience, we have a dierent object of
experience to what that experience concerns, for example, the waves breaking
upon the shore. Yet it is the la er that characterizes what it is like to have the
experience.
By itself, this comment goes very li le way to resolving the situation. Sup-
pose the current object of my experience is a certain arrangement of physical
properties which, let us presume for the sake of argument, realizes our experi-
ence of a red wall. Then what we want to know is why our experience of
this arrangement of physical properties in no way reveals the fact that the
arrangement is an experience of a red wall, or if it does, then in what way it
does and how can this be squared with what it is like to be in that experience.
We cannot just trade upon the fact that there are dierent objects of experience,
we want to understand how one object of experience should manifest itself
in the other object of experience.
The point can be driven home if we consider the special case in which we
are experiencing an arrangement of physical properties which either is exactly
like the arrangement of physical properties which is a realization of this kind of
experience in another or one which is very like the arrangement if, for example,
we are having an introspective experience of one of our own experiences. In
the former case, we want to know why our experience of the arrangement of
physical properties in another does not reveal that it is an experience of the
same type as we are currently undergoing. In the la er case, we would be having
an introspective experience of e* (an experience very like our introspective
experience of it) and asking why it in no way reveals that it is very like the
arrangement of physical properties which realize the introspective experience.
Looked at this way, an a empt to bridge the explanatory gap has two
components. First, we need a theory of what makes a particular experience
of a property P. Second, we need an account of what makes P show up with
the character it has in our experience. In the next section, I will be considering
various accounts of the first, including the currently fashionable one which
I favour: representationalism. The second component will not receive a ention
though it is important. As many have noted, if we a empt to bridge the
explanatory gap by focusing on what our experiences concern, then the
explanatory gap is reintroduced regarding the properties the experiences
concern. For example, if colour is a profile of surface reflectance properties,
then why does it look the way it does in experience? However, the room for
manoeuvre here is far greater. The world is not blurred, but my blurred per-
ception due to myopia has a perfectly satisfactory explanation to bridge the
non-blurred-blurred gap.
259
Obviously, then, in an extended sense the recommended response is a

version of the ignorance approach. However, it is distinctive in that a char-
acterization of the kind of explanation required has been provided and it is
suggested that, when we possess an explanation of that kind, the explanatory
gap will no longer be present. Indeed, it is compatible with the proposal that, in
fact, we are not ignorant of the arrangements of narrowly physical properties
which account for phenomenal consciousness (or need not be) for the explana-
tory gap to arise. We just dont appreciate their significance until we appreciate
what makes an experience of the manifest objects and properties they concern.
This brings us to the question of the character of phenomenal properties.
Theories of phenomenal consciousness
It is quite possible to challenge the division of labour regarding theories of

phenomenal consciousness observed below into theories of phenomenal con-
tent and theories of what it is to be aware of that content (e.g. Neander, 1998).
However, the reasons for this are best understood by considering the two
elements separately.
Phenomenal content
There are, broadly speaking, three approaches to phenomenal content, more

precisely, the phenomenal properties which determine it, currently favoured in
the literature: those which take qualia, representational properties and brute
non-representational relations of awareness respectively as that which deter-
mines the phenomenal content of mental states. That these three approaches
constitute three distinct options may not be immediately obvious. For example,
it may be said that qualia se le the phenomenal content of our mental lives by
our being aware of them. Moreover, rightly, any reader of the word determines
will want to know what that involves.
Ma ers can be clarified if we consider the case of veridical perception
where the three approaches most obviously come into immediate contact.
Although this might not be historically accurate, a productive way of thinking
about the case of veridical perception is to see appeals to qualia, and to brute
non-representational relations of awareness, as a response to the perceived
inadequacies of the representational approach.
According to representationalism, the representational properties of experi-
ence determine the phenomenal content of that experience (see Harman, 1990;
Dretske, 1995; Tye, 1995, 2000). Representationalists take the kind of properties
that make beliefs have truth conditions, desires have satisfaction conditions
260
(e.g. that I have a fig now) and, outside the mental arena, sentences have truth
conditions, and suppose that they will also be the basis for an account of the
phenomenal content of an experience. For example, some of the representa-
tional properties of a particular perception may determine that the content of
the experience is that there is a grey elephant at such and such a location.
Although, in one sense of determine, what determines the phenomenal content
of the experience is that the elephant and the property of being grey are con-
stituents. In another sense, the one to which the representationalist appeals,
the representational properties determine the phenomenal content by certain
objects and other properties being the manifest objects and properties of the
experience. Representationalists o en, and increasingly controversially,
motivate their account by appealing to the fact that, in characterizing what
an experience is like, we dont notice properties of the experience so much as
go right through to what the experience concerns the manifest objects and
properties suggesting that representation is at work to determine that content.
It is reasonably clear how the representationalist approach may seek to
make a contribution to the question of the explanatory gap mentioned at the
end of the last section. If we had an account of the representational properties
of an experience, we could see what elements of an arrangement of physical
properties helped to explain how that experience was an experience of a red
wall. To illustrate, rather than because the account is independently plausible,
if what made an experience of a red wall was the fact that it was causally cor-
related with red walls in optimal circumstances, then experiencing these causal
facts would be to experience what made that experience have the character
it does.
A acks to the representationalist picture come from, broadly, two angles.
On the one hand, critics seek to identify experiences which have phenomenal
dierences but where it is plausible that their representational content is the
same (Peacocke , 1983; Block, 1990, 1996; Shoemaker, 1990, 1991). On the other
hand, critics emphasize the obvious dierences between two very dierent
kinds of mental states: beliefs and perceptions (e.g. Martin, 2002).
The general form of the debate under the first heading has been that, for
every phenomenal dierence identified by the critics of representationalism
which dierence is supposed to encourage the introduction of qualia as addi-
tional phenomenal properties the representationlist identifies a representational
dierence a er all. Thus qualia are usually thought to make a dierence to
what an experience is like by being possessed by the experience without
being representational properties and, for some, without being the objects of
awareness. Others are inclined to accept that they are objects of awareness and,
hence, in one respect, the way in which they determine the phenomenal content
of our mental lives is by being manifest properties se led by the third account
of phenomenal content appealing to brute awareness.
261
For example, it has been remarked by representationalisms critics that our

visual and gustatory experience of wine is very dierent and yet may well be
of the same chemical properties which give rise to distinct aects on the light
we experience and on the taste buds (Shoemaker, 1990). Representationalists
response has been to emphasize that, even if there are similar properties
experienced, they will be experienced in dierent ways because of the other
bundles of properties with which they are associated. Of course, representa-
tionalists cannot prove that this accounts for the phenomenal dierence but
the force of the reply is no worse than the hypothesis that the same properties
are experienced and yet there is a phenomenal dierence: stand-o.
Some apparent diculties for the representationalist have arisen because of
the account of representational properties they have endorsed. For example,
one of its foremost proponents, Michael Tye, takes an experience to represent
a blue object, say if the experience is of a type where it is optimally causally cor-
related with blue objects, and so on and so forth for the other colours. Consider
now the alleged possibility of inverted spectrum. It is hard to make sense of this
because the basic idea is that the two inverted subjects see the colours of objects
in dierent ways while otherwise having their experiences caused in a the
same way and giving rise to the same responses. If representation of colour is
by causal correlation, this should not be possible if you accept the claim that,
although the subjects see the colours in dierent ways (one sees red the way the
other sees green, and so on) it would be a mistake to suppose that one subject is
systematically misrepresenting the world where the other subject gets it right.
There are a number of dierent possible replies that the representationalist
might make. One is to challenge the claim that spectrum inversion is a possibil-
ity. If there is independent reason to believe that the account of representational
properties is correct, then this would be one ground for resistance. Unfortu-
nately, this is far from clear. As we shall note later, causal correlation in optimal
circumstances faces problems as an account of representation. More promising,
with regard to dealing with the alleged possibility of spectrum version, is to
note that our conviction that it is possible may mistake the intrinsic character
of what is represented for the intrinsic character of what is doing the represent-
ing. Thus, if colour is an intrinsic property of the surfaces of objects at least
to the extent that the colour of an object does not seem to depend upon the
colour of other objects then we are inclined to judge that we could have see
that object with another colour instead or indeed all the colours swapped
round when, in fact, this is not possible. To be fully developed, this approach
will have to explain how we are convinced of this possibility when colour
illusions show that how we experience a colour is o en a function of the other
colours experienced along with it.
A third option is to take the possibility of spectrum inversion as providing
additional grounds for supposing that the account of representational properties
262
in terms of causal correlation in optimal conditions is a mistake. Some other

account of representational properties is appropriate. This is an approach
I favour. Rather than take an analysis of representational properties to dictate
what is represented in experience, we revise our account in response to what
seems to be represented in experience.
Other phenomenal states seem to present diculties for the representation-
alist because we come with certain assumptions about what must be repre-
sented. These assumptions o en stem from a view about the correct analysis
of representational properties. The clearest examples of this are the alleged
non-presentational states such as that of free-floating anxiety or general depres-
sion. These states may involve a number of dierent presentations of the world
in the way distinctive of the mental set of the depressed or anxious person,
but there is an element of these states it is alleged which involves no
presentation at all.
Representationalist treatments of these kind of cases either tend to claim that
this element is captured by the various ways in which a depressed or anxious
person represents the world to him or herself, for example, as lacking things of
value to do, or to urge, as a result of your favoured theory of representational
properties, that, in fact, something is represented a er all (Crane, 1998, seems
to suggest the former strategy). As an illustration of the la er approach, Tye
has argued that the general phenomenal character of depression relates to
low states of certain chemicals in the body which is said to lie at the root of
depression (Tye, 1995).
Neither move is satisfactory without further defence. When we experience
the world as lacking value, it is doubtful whether what is represented is that the
world lacks value. It is rather that our a itude to what is in the world is that
we find no value in it. So there seems to be a retreat from taking phenomenal
content to be exhausted by what is represented to taking mental states essen-
tially to involve some form of representation even if the phenomenal content is
not exhausted by it. On the other hand, working out that something is repre-
sented a er all, according to a preferred account of representational properties,
still gives rise to a puzzle. We do not have to work out what is represented in
experience when we take it to represent something square or red. These proper-
ties are manifest in experience. If we conclude that the general character of
depression is that it presents low lithium levels in the body, then the obvious
question is why wasnt this obvious in experience.
Both these critics of representationalism and the responses overlook another
possibility. They all seem to assume that if a state, in a certain respect, involves
nothing but a certain character it is like to be in that state, then that state couldnt,
in that respect, actually be presenting that character. But this can be challenged.
If a certain property seems to characterize a state but does not involve the
presentation of anything else, it is reasonable to consider the hypothesis that
263
the state in question in fact, partly, presents that character. Thus, alleged non-
presentational states in fact involve higher order states which are presentations
of some of these lower order states properties (for a suggestion of this kind of
response, see Byrne, 2001; Noordhof, 2003b).
A number of issues are raised by this possibility. The first is that there are
cases of phenomenal dierence where it is part of our understanding of this
phenomenal dierence that it does not involve a presentation of this dierence.
The classic example is blurred vision. When we blurriedly experience some-
thing, our experience precisely doesnt seem to be an experience of something
which is blurred (Crane, 2006, pp. 1301). A second is that if we do take puta-
tive non-presentational states as involving higher order presentations, are
we, in eect, undermining the a raction of representationalism by, as it might
seem, allowing that what is represented in this case are the very phenomenal
properties that representationalists were seeking to do away with, viz. qualia?
I raise the second issue only to set it aside. I guess your view about that will
depend upon what you pack into the idea of qualia. In principle, representa-
tionalists should not be concerned about what counts as the manifest properties
of experience so long as it is representational properties which determine that
they are the manifest properties.
The idea that there might be phenomenal dierences which correspond to
representational dierences and yet which dont involve presentations of
these properties the blurred vision case may, fruitfully, be related to the
idea that there might be dierent degrees of phenomenal presence. Here are
two examples. Suppose that you are currently experiencing a red tomato.
Then it seems plausible to say that our experience is of something which has a
backside. Indeed, one might think, the phenomenal content of our experience
would be rather dierent if we experienced the tomato as not having a backside
rather than as experiencing it as having a backside. How should we capture this
dierence? Second, many feel happy about saying that our visual experiences
present colours and shape to us but they balk at saying it presents water as
opposed to something which flows, is transparent, etc. But just as with the
case of the backside of objects, it seems undeniable that there is a phenomenal
dierence at some level between experiencing water and experiencing this
more cautiously described kind of thing.
It is tempting for the representationalist to deal with these cases by
suggesting that the phenomenal content is the result of the judgements we are
inclined to make on the basis of the experience. Thus, it is because we dont
take our blurred experience of the world to be an experience of a blurred world
that the blurriness is not presented as a property of the world. But the imple-
mentation of this strategy is not straightforward. For example, suppose that
you are convinced that the forest you are currently seeing is an elaborate holo-
gram. Then you wont be inclined to take the experience to be of trees and so
264
forth. Nevertheless, it seems that what you experience is still appropriately

characterized as having a phenomenal content characterized in terms of trees
and not some feature shared by tree and hologram presentations (Siegel, 2006,
pp. 4945).
Some have been tempted to see an argument for the enactive account of
perception to be derived from the claim that the backside of objects is phenom-
enally present to us in our experience in some sense (e.g. as absence). According
to, perhaps, its leading philosophical proponent, Alva No, perceptions involve
the utilization of sensori-motor knowledge to have the content that they do
and what is responsible for our experience of a tomato with a backside is that
we have sensori-motor expectations concerning what would happen if we
changed our location relative to the tomato (No, 2004, pp. 779).
Although highly suggestive, the proposal has diculties. We have many
sensori-motor expectations. Exactly how are some recorded in the phenomenal
content of experience and others not in a way which accords with the phenom-
ena to which we have just adverted? For example, I have certain expectations
about what would happen if I moved into the next room regarding what
I would see. Yet, it is stretching it to say that the next room has a grade of phe-
nomenal presence even as absence in my experience of the current room in
which I am located. Perhaps what is crucial is sensori-motor skills which are
in some way integrated with our perception of particular objects or kinds. But
it is unclear why we need to adopt such an approach rather than argue that
there are conditions that must be met for an object or property to be a manifest
object or property of experience, which together with a representation of only
some of its properties will mean that other of its properties are present by
their absence.
A third issue raised by the tendency to be liberal about what is represented,
for example, by allowing representation of properties of mental states, is
what we might dub the treatment of putatively non-representative sensory
residue. For example, No suggests that without the sensori-motor skills in
play, we will have sensuous material without anything being represented
(No, 2004, p. 10). Similarly, A. D. Smith has argued that our perceptual
experience may be intrinsically but not essentially representational. When our
sensations dont make us aware of normal physical objects, we simply have
sensuous ma er without representation. Although, later, he concedes that
sensations might, in those circumstances, provide us with presentation of
non-objective states of our body, he takes this to be distinct from perception
of independent objects characteristic, as he sees it, of the intentionality of
perception (Smith, 2002, pp. 1237).
Is there a substantial issue here or can all the facts be captured whether one
decides to build more or less into representation? One consideration in favour
of being more concessive over what can be represented is that the theoretical
265
cost of explaining how there might be more substantial representation derived

out of more primitive cases of representation may be less than seeking to pro-
vide an entirely new characterization of what characterizes the phenomenal
content of our mental lives and, most importantly, how it relates to, and con-
stitutes, the classic cases of intentionality in perception.
The debate between representationalists and those who appeal to brute
relations of awareness, that is, proponents of relational views of perception
(relationists), arise from the la ers diagnosis of the former as a empting to
assimilate perceptual experience to belief or judgement. Yet, there is an obvious
dierence in phenomenal content. In the case of perceptual experience, objects
and properties in the world seem presented to us. In the case of belief or
judgement, this is not the case. My belief that snow is white as opposed to
my perception that snow is white involves nary a hint of the delectable, chilly,
white, fluy stu.
Broadly speaking, the debate has had two areas of focus. One has been
those conscious states that seem to be phenomenally similar to perceptual expe-
rience and yet involve objects that dont, or at least neednt, exist. Indeed it is
not even that the objects and properties neednt exist as that we seem to have
an experience of them without standing in the kind of relationship which is
distinctive of perceptual experience. The obvious examples of these are imagi-
nation and hallucination. If representationalists can make good the charge that
these states have the same phenomenal content as (or appropriately similar
phenomenal content to) perceptual experiences, then their opponents are wrong
to characterize perceptual experience in such a way that the explanation of
its phenomenal content is very dierent to that of these states and appeals to
relations.
Typically, then, proponents of relational views of perception are keen to
deny that there is this similarity in phenomenal content. In the case of imagina-
tion, they are inclined to say that the reason why we suppose that there is a
similarity of phenomenal content is that when we sensuously imagine some-
thing, we imagine having a perceptual experience of that thing (dubbed the
dependency thesis). Imagining having a perceptual experience gives us an
explanation for why we feel that the phenomenal content is the same when
it isnt.
It is debatable whether this can be the basis for a successful response. It is
worth nothing that we can also think that we are having a perceptual experi-
ence of black dog with a wet nose (you are having that thought now) or
believe that we are. Nevertheless, the phenomenal content is very dierent
(see Martin, 2002; Noordhof, 2002).
In the case of hallucination, a disjunctive approach is adopted. First, it is
asserted that perception and hallucination are two dierent kinds of states with
no common content. Second, a non-ontological account is given of phenomenal
266
similarity. Thus it is said that when we hold that perception and hallucination
are phenomenally similar, that just means that they are indiscriminable to
introspection alone. The account of why they are indiscriminable may vary.
All that is denied is that phenomenal similarity should be understood simply
in terms of similarity in phenomenal content (Martin, 2006, p. 369). In the case
of perceptual experience, the determinant of its phenomenal content is that it
involves a certain relation to an object and its properties. In the case of halluci-
nation, no account is given of its phenomenology. Some hold that that is because it
has no phenomenology (Fish, 2009). Others may allow that it has a phenomenal
content but just not like that of perceptual experience.
Giving an account of the various ways in which hallucination may be
indiscriminable from the corresponding perceptual experience and resolving
the question of whether hallucination does have a phenomenal content are
pressing issues for proponents of the relational view. I think it is fair to say that
it is upon their treatments of these issues that the success of their approach
turns. Nevertheless, as we remarked, representationalists have a diculty of
their own: the apparent substantial phenomenal dissimilarity between percep-
tual experience and belief or judgement.
The standard strategies are to hold that perceptual experience involves a
richer, perhaps non-conceptual, content. Thus the phenomenal dierence is to
be accounted for by supposing that there are more representational properties
at work in the case of perception and that what is represented is organized non-
propositionally (e.g. spatially and/or without requiring that the subject who
has the states possesses the relevant concepts; see, for example, Tye, 1995).
While these are plausible dierences from belief, it is questionable whether
they will do. Representionalists hold that representational properties have
phenomenal significance. Thus even in the case of belief, one would expect its
representational properties to have some phenomenal significance. The belief
that there is a dog barking should have a whi of a dog about it. The belief
that it is brown should give us a flash of brown. But none of this is the case.
For this reason, representationalists are going to have to appeal to an analy-
sis of the second component of phenomenal consciousness, namely subjective
awareness. They need to do this anyway to explain why there is no phenome-
nology in the case of unconscious mental states which, we may presume, have
representational properties. But perhaps, in addition, such an appeal will help
dierentiate between conscious judgement or belief and conscious experience.
Subjective awareness
Theories of subjective awareness divide broadly into those which take the
consciousness of a certain state to be an intrinsic fact and those who take it to
267
be an extrinsic fact. However, as we shall shortly see, it is hard to retain

this division and there is a distinct danger that theories of the former sort
collapse into theories of the la er sort when suitable refinements are made.
These theories become part of a full theory of phenomenal consciousness
because they appeal to materials discussed under the previous section to dif-
ferentiate between what it is like to experience various kinds of mental states
those with dierent types of content. Theories of subjective awareness explain
the circumstances in which we are aware of this content. An exception to these
remarks is the provocative theory of consciousness oered by Ted Honderich.
He takes consciousness to be constituted by the existence of a mind-dependent
world of objects and properties. Thus, for him, what we might dub the phenom-
enal content is the account of our awareness of it (Honderich, 2004, Chapters
710; for discussion, see Noordhof, 2006b).
Among theorists who take the consciousness of a certain state to be an
extrinsic fact, there is a further division between those who adopt a perceptual
model and those who appeal to the idea of higher order thought instead. The
former hold that something like the following is the case.
A conscious mental state is one which is scanned by the introspective

perceptual mechanism, an unconscious one is one which is not. (Lycan, 1996,
pp. 1415)
This theory raises in particularly stark form many of the problems which have
plagued all extrinsic accounts of consciousness, so I will rehearse them here
first. Possibly the most familiar objection is that it makes our experience of
our own conscious states highly fallible. The internal scanning mechanism
may misfire so that either we introspect that we are in a dierent state to the
one we are, in fact, in, or we think we are having a certain mental state when
we are not. You could, for example, perceive that you are in serious pain when,
in fact, thats just a false case of introspection.
In the case of pain, it is possible to construct an evolutionary argument
in favour of radical fallibility being extremely unlikely. Our pain system is
important for us to recognize when our bodies are being damaged and crea-
tures which fail to appreciate when this is happening are unlikely to survive
long enough to pass on their genes. On the assumption that the capacity to
introspect is coded into our genes as one might expect, given it is a sensory
mechanism we can expect considerable accuracy (Lycan, 1996, p. 18). This
is much less obvious in the case of other mental states. It would be productive
to consider this argument on a case by case basis and consider whether our
relative confidence concerning our judgements about whether we are in those
states correspond to what we might predict by the evolutionary story.
268
Nevertheless, rather than focus on this, I want to discuss more immediate and
challenging diculties that follow on from this initial observation.
If what makes a mental state conscious is that we perceive it, then what
about the objects in the world? They are the objects of perception. Do we
thereby make them conscious? For example, gaze at a rock. Does your gazing
make this a conscious rock? You might reply that, yes, the rock is an object of
consciousness but it is not, itself, conscious. However, in making this response,
you are presuming that the conscious state is the product of the perceptual/
introspective mechanism. That is not the position of those who adopt this
theory of consciousness. Their claim is that the conscious state is the one
which is the input into the process the object of the perceptual-introspective
mechanism.
So, instead, the reply that is that rocks arent conscious because we do not
call them mental states (Lycan, 1996, pp. 234). This reply is either weak or
under-described. If the denial reflects a stipulation regarding our use of the
term conscious, then the reply seems weak. It concedes that, in terms of the
reality, there is considerable similarity between the state the rock is in, as an
object of perception, and the state one of our mental states is in, as the object
of mental perception. It just denies that we apply conscious to the former. The
situation would be no dierent to remarking that sparkling white wine pro-
duced by the champagne method could not be champagne if it did not come
from the champagne region. This might be an important point regarding mar-
keting but, unless there is something else to be said on the taste front, entirely
uninteresting to the drinker. On the other hand, if it is suggested that we limit
applying conscious to mental states because, in addition, they are states of a
cognitive/aective system and this aects the nature of consciousness, then the
reply is certainly under-described. It is also questionable whether the result is,
strictly speaking, a perceptual model of consciousness since now, the a ribu-
tion of consciousness to something is not just a ma er of it being perceived but,
in addition, these further facts Ive mentioned. This is probably why, as I can
tell, the foremost proponent of the perceptual model William Lycan seems
to have in mind a response of the former sort, with its a endant weakness.
The third challenge that the perceptual model faces concerns what we should
say if the perceptual-introspective mechanism misfires and proclaims that
we are undergoing a certain state when we are not. What are mental halluci-
nations like? If you claim that, when a subject hallucinates that she, say, is
imagining a green frog, then it is exactly for her as if she were imagining this.
Then it seems that we have to allow the possibility that there are some intrinsi-
cally conscious states, our hallucinations. If we allow intrinsic consciousness
here, then why not everywhere? (cf. Byrne, 1997, pp. 1212, with regard to
higher order thought theory).
269
For this reason, I suspect that proponents of extrinsic accounts of conscious-

ness must insist that there is nothing it is like to hallucinate a mental state since,
in those circumstances, there would be no state of which we could be conscious.
However, there are costs of making this move. First, suppose that a subject
introspects that she is imagining a green frog and has a still higher-order mental
state that she is introspecting that she is imagining a green frog. Then there
must be some conscious state the subject is in she must be conscious of a hal-
lucinatory introspective state by the terms of the theory. But whats that
like? Arent we forced to say that it is exactly like thinking that one is con-
sciously imagining a green frog? Second, if we allow the object of introspection
to determine the content of introspection, then arent we forced to conclude that
it doesnt ma er what we introspect ourselves to be undergoing because it does
not determine the content of our conscious states? The la er is just se led by
whatever state we are in being the object of an introspective state.
To deal with the first point, it seems we must abandon the idea that a state
is conscious simply by being the object of another state. There must be a first-
order state at the end of the chain. Otherwise, we will have phenomenal
similarity between conscious hallucinatory states and non-hallucinatory states
without the phenomenal content that the la er is meant to supply. In response
to the second point, it may be suggested that ge ing the state we are in roughly
right is a necessary condition of the state being conscious as a result of being
perceived.
Why do some philosophers appeal to higher order thought rather than
perception in seeking to provide an extrinsic account of consciousness? Proba-
bly the following considerations have weighed with them. First, introspection
seems to involve no distinctive sensory qualities for itself. If this is taken to be
a necessary condition for perception, then introspection fails to count as a case
of perception. Second, there is the problem of the rock mentioned earlier. It
might be thought less likely that thought would stand in the same relationship
to a rock as higher-order thought to the mental state of which it is supposed
to make us conscious. However the last point is not at all obvious. Rocks seem
just as likely, and in similar ways, to trigger thoughts about them as mental
states trigger higher-order thoughts about them. With regard to the former
point about distinctive sensory qualities, it is hard to know what significance to
a ach to it. If the higher-order states are not supposed to the contribute to the
content of conscious states, but just make us aware of the lower-order state,
then these sensory qualities cannot be provided by the states of perception
themselves but must rather lie in their object. If mental states are representa-
tional, one might expect that they provide no sensory qualities of their own
but convey the sensory qualities possessed by what they concern.
There are also disadvantages appealing to higher order thought rather than
perception namely, that it makes consciousness a feature of creatures with
270
conceptual capacity the capacities needed to have thoughts and also relates
it to self-consciousness because the higher-order thought is characterized as
that I am in such and such a mental state. Those who wish to a ribute phenom-
enal consciousness to other animals and children at an early stage of develop-
ment, o en balk at this commitment. The only way this can be resolved,
I suspect, is if we have a much clearer idea of what is involved in the basic
conceptual capacities at work. Naturally, higher-order thought theories down-
play their sophistication, but the question is whether they can do so while, at
the same time, making it legitimate to think that the states they postulate
are genuine thoughts rather than non-sensory higher-order representations
of some non-thought-like kind (e.g. Rosenthal, 1986, p. 344; Rosenthal, 1991a,
pp. 323; Rosenthal, 1993).
It is standard to appeal to the presence of higher-order perception or thought,
rather than the disposition to have these things to account for consciousness.
The reason oered is that being conscious of something is an occurrent state
rather than a dispositional one (Rosenthal, 1993, pp. 2089). However, this
seems to be a confusion. Once it is recognized that the conscious state is the
state which the disposition to have an introspection or thought concerns, then
we have an explanation of in what sense being conscious of something is
occurrent namely, that there is an occurrent state which is the object of a
disposition.
One thing that either type of theory emphasizes is that the higher-order
perception or thought itself is not conscious. What would make it conscious,
in turn, is an even further higher-order thought or perception about it. This
is usually thought to be an unsatisfactory feature by those who propose a
reflexive intrinsic theory of consciousness (e.g. Kriegel, 2009, Chapters 1 and 4).
Such theorists begin by pu ing forward something like the following.
S is a conscious state if and only if S represents itself in the right way.
Thus suppose that I am having a conscious perceptual experience of a desk

lamp, then I am in a state with two contents. The first of these is of the desk
lamp. The second concerns the state which is of the desk lamp. Proponents of
this approach struggle to ensure that this doesnt re-introduce an extrinsic
account of consciousness.
Obviously if it is said there is a mental state which has two components the
first of which has the content that there is a desk lamp, the second of which has
the content that this state is of a desk lamp then we have a dierence in name
only from an extrinsic account of consciousness. The only dierence is that
reflexive accounts of consciousness have made the decision to call the complex
of states to which extrinsic theories of consciousness appeal a state itself.
Obviously no proponent of a reflexive account of consciousness is going to
271
accept this description of their position. So the interesting question is how they
avoid it.
One suggestion has been that the conscious state has the causal roles both
for the experience that there is a desk lamp there and the awareness of this
experience. Co-instantiation of the causal roles explains the reflexivity of the
state (Shoemaker, 1994, pp. 2425). Obviously co-instantiation of roles by
itself wont do. There are any number of roles which may be coin-instantiated.
What is of particular significance is that one of the roles is that of an awareness
of the experience. At the minimum, this means that one of the roles must be of
a state which represents the experience in question. This makes the proposal
a version of representationalism in which the correct account of the repre-
sentational properties of awareness of the experience is to be given in terms of
causal role.
Some have argued that appeal to coincidence of roles a dispositional
ma er is once more ill-suited to explain the occurrent nature of consciousness
(Kriegel, 2005, pp. 378). If what I have argued before is correct, then this is
a mistake. The occupant of these coinciding roles is occurrent. In any event, it
has influenced the currently most prominent exponent of an intrinsic theory of
consciousness to provide a weakened version of this approach to take this
into account (Kriegel, 2005, pp. 4451). According to Uriah Kriegel,
M is conscious if there are M* and M** such that both M* and M** are proper
parts of M, M is a complex (not merely a sum but an integrated whole
like a molecule) of M* and M** and M* represents M by (indirectly)
representing M**. (Kriegel, 2009, pp. 2268)
The idea is that M contains proper parts, M* and M** such that M** causes and
hence is represented by M* and M* indirectly represents M by representing
M** in roughly the way that a painting can represent a house by representing a
portion of it, the rest being obscured by trees etc.
Obviously a lot of work is being done by talk of an integrated whole and the
idea of indirect representation. Since there seems nothing to rule out integrated
wholes in which representation is only of a part, talk of representation of a part
of the overall state, which thereby represents a whole, requires a particular
kind of, as yet unspecified, integration. To remark it is unspecified is not to say
that, over time, such an account cannot be provided. The question is whether
indirect representation is what we need. In the pictorial case, although the
whole house is represented, it is plausible that only part of the house the part
not occluded is phenomenally present in the strongest sense. The remainder
may have a weaker grade of phenomenal presence, depending upon the out-
come of the discussion mentioned earlier. However, it is questionable whether
this is what we want in the case of our consciousness of a mental state. Is there
272
some part of our conscious state of which we are not conscious when we
are conscious of it? A natural candidate would be the reflexive element.
However, proponents of the reflexive account of consciousness usually empha-
size that the reflexive character is part of what we are aware of (e.g. Kriegel,
2009, p. 117).
Nevertheless, perhaps this can be turned into a virtue. It has o en been
remarked that we are not conscious of the nature of our mental states but just
conscious of what they are of in being conscious of them. It might be argued
that a lower grade of phenomenal presence implied by indirect representation
nicely explains this.
One of the motivations for the moving to an intrinsic, reflexive account of
consciousness was the conviction that unconscious states were not the kind
of states to make the states they concern conscious and yet, if it was insisted
that the higher order states were conscious, a vicious regress would ensue.
Obviously, one might ask the same question of M*. Are we conscious of it and,
if so, what makes us so. Kriegels answer is that it is part of a globally conscious
state. However, it seems we may have replaced an infinite regress with an
explanatory circle. We are conscious of M (and M**) by having M* representing
M** and conscious of M* by having M* as part of M. If that move is allowed to
the proponent of reflexive consciousness, why isnt an equivalent available to
the proponent of an extrinsic account of consciousness.
Nor does it seem that this approach to consciousness avoids the possibility
that we could have M** without M* and hence the same questions arise as to
whether we would, in those circumstances, have a hallucinatory mental state
or no phenomenal state at all. Reflexive accounts of consciousness are likely to
have to draw upon the same materials as extrinsic accounts of consciousness
here.
Although we began this section with two theories in play, it appears that
they have converged upon something remarkably similar. A conscious state
involves, at least, two parts. The first is the state of which we are conscious; the
second is the element which makes us conscious of it. The second element,
therefore, cannot be present and provide us with consciousness independent of
the first element. There is nothing it is like just to be in a state with the second
element. The dierences lie in the account of indirect representation and the
claim that the second element is part of the state of which we are conscious.
The plausibility of these claims relate back, I have suggested, to the diering
degrees of phenomenal presence we may recognize in our conscious states.
If this convergence is to be evaluated further, work is required on the content
of the second element. I have suggested that an appeal to a causal relationship
between it and the first isnt mandatory to explain its content. Perhaps a suc-
cessful explanation of the content may restore some of the dierence between
the two approaches once more. It also seemed that the second element couldnt
273
convey consciousness by any old misrepresentation of the first element. It had

to be roughly right. Indeed, since expression of the content of the second
element seems to be behind our self reports about our conscious states, and we
are generally recognized to have a special first-person authority concerning
them, we might have to be more than roughly right. Yet, specifying the right
kind of content to convey consciousness is no easy ma er. It is this which
has led some to wonder whether the division of our theorizing concerning
phenomenal consciousness into these two elements is a mistake.
I mentioned at the end of my discussion of phenomenal content that the
dierence between conscious judgement/belief and perception may lie in the
theory of subjective awareness. Here is how it might work. The content that
the higher order must have to convey consciousness upon the state which is
its object (Ive stated this in terms that proponents of the intrinsic as well as
extrinsic theories of consciousness can accept) varies depending upon the kind
of state involved. In the case of perception, the way in which the second ele-
ment makes that perception conscious is by correctly representing, or being
disposed to represent, the content of that perception. Whereas, in the case of
conscious judgement, the way in which the second element makes that judge-
ment conscious is by correctly representing that one has a judgement that p
(where p is the content of the judgement). This might account for the fact that,
in the case of perception, what we are aware of is the content of perception:
the objects and properties in the world. Whereas, what we are aware of, in the
case of judgement, is the representational properties of the judgement, the
meanings that make up its expression.
If something like this is along the right lines, then there is, as yet, no decisive
reason to reject representationalism about phenomenal consciousness, pending
successful treatment of an account of subjective awareness. Whether such an
approach will be compatible with the truth of physicalism will turn on arriving
at a proper understanding of the nature of representational properties. In the
final section of this chapter, I briefly touch on some issues that have been
considered under this heading.
Intentionality and Normativity
Just as the evaluation of representationalism has engendered an increased

sophistication in the discussion of phenomenal content, there has also been an
interesting discussion of the nature of intentionality. According to the standard
picture, intentionality involved direction upon a content which was determined
by the representational properties of a state. Belief is, perhaps, the paradigm
example. Various features were a ributed to intentionality so characterized, of
which, perhaps, the most significant is that there could be directedness upon
274
objects and properties which do not exist. Allowing that this is so is compatible
with also allowing, as externalists insist (see Sawyer in this volume for a good
discussion of externalism), that what contents a subject entertains may be set-
tled by the environment. Even if you allow that a subject thinks about water
rather than twater because they have water in their environment, it does not
follow from that that there can be no circumstances in which the subject thinks
about water without there being any in the environment. However, when we
turn our a ention to the case of perceptual experience, it has seemed to a grow-
ing number that it does not have these features. We cannot perceive a dog
unless the dog is present and we stand in a relation to it.
This has given rise to an interesting range of research questions. Can there be
non-representational intentionality as well as representational intentionality?
Are there both propositional and non-propositional forms of intentionality?
Can intentionality come in both conceptual and non-conceptual forms? What
are the relationships between the answers to these questions?
Such questions have given impetus to a field which appeared to have been
going through a bit of an arid spell. If we compare the flurry of activity in the
late 1970s and 1980s with regard to providing a reductive analysis of inten-
tionality (i.e. in terms compatible with the truth of physicalism) with the cur-
rent state of aairs, it seems clear that the emphasis has changed (Field, 1978;
Dretske, 1981, 1988); Fodor, 1984, 1987); Millikan, 1984, 1986, 1989); Papineau,
1984, 1987, 1993a); Stalnaker, 1984; Block, 1986c; Cummins, 1989; Whyte, 1990,
1991). To some extent, the reductive programme has stalled. As far as causal-
informational accounts have concerned, there has been li le further develop-
ment since Fodors A Theory of Content back in 1990. The most significant
competitor Millikan and Papineaus appeal to biological function still faces
controversy over whether it successfully has dealt with the criticisms concern-
ing indeterminacy and error which it was introduced to supply (Millikan, 1984;
Papineau, 1987, 1998b; see Ravenscro and Rey in this volume, for a description
of the disjunction problem which lies at the heart of this debate). A disappoint-
ing feature of the current lack of further developments has been the tendency
to use toy reductive accounts of intentionality in the development of represen-
tationalist accounts of consciousness and the like both to resolve (in the case
of bodily sensations) and present diculties (in the case of inverted earth)
as if the toy accounts in question were close to being right as opposed to
having been shown to be, more than likely, mistaken.
One consequence of this situation is that there has been renewed interest
in the nature of conscious intentionality with the suggestion that the solution
to the diculties which bedevil earlier reductive accounts will be found here.
For example, Galen Strawson takes experience (conscious experience) to be the
primary form of intentionality because he believes that only cognitive experien-
tial qualities will bring sucient determinateness to our intentional states to
275
make misrepresentation possible. Suppose that I am currently experiencing a

desk lamp. Then what makes my experience of a desk lamp rather than any
item on the causal chain between my eyes and the desk lamp (or indeed any
item further back on the causal chain a er the desk lamp) is the cognitive
experiential quality of my experience which involves a taking it to be of a desk
lamp rather than my retinal stimulation (say). Put me in a dierent environ-
ment in which my experiences are caused dierently by mammals which tend
to sit around on desks in various unnatural poses and my experiences will be
mistakenly about desk lamps rather than them because I take my experience
in that way (Strawson, 2008, pp. 295302).
Of course the problem with appeal to taking something consciously a certain
way to resolve these diculties to which Strawson is sensitive is that its
resolution can seem very like magic. It is also potentially threatening to the
hopes of representationalists to provide an account of phenomenal conscious-
ness by appealing to the representational properties of our mental states. Never-
theless, the fact that, in consciousness, there seems no problem with these issues,
where accounts which do not appeal to it struggle, promises to make the dis-
cussion of consciousness more productive. We have a putative explanatory
role for consciousness to play and a way of considering what it brings where
other accounts fall short, if, in the end, they do. By the same token, the develop-
ment of a way of dealing with these issues without explicit appeal to conscious-
ness provides an insight into the nature of consciousness because they are
accounts of something which seems eortless at the level of consciousness.
Perhaps this is also part of what lies behind the recent a raction to enactive
theories of perceptual consciousness. The way in which our sensori-motor skills
are integrated with the causal process from our object of perception to states of
the brain se le that the causal process supports perception of that particular
object rather than events later or earlier in the chain (e.g. see Nos remarks on
prosthetic perception and non-standard causal chains in No, 2003, pp. 97100).
Another aspect of the case in its favour is the idea that perceptual content is best
specified in terms of what is immediately available to a perceiving subject to
explore in the world and not what is represented (No, 2004; 2006, pp. 4268).
This has led some to consider whether perceptual consciousness following on
from cognitive processes (see Wheeler in this volume) is constituted in part
by extra-cranial processes (see Hu o in this volume). We couldnt have the
phenomenal consciousness we do without the environment. Proper defence
of this view is going to require systematic exploration of the kind of relation-
ships which take place in the brain, upon which consciousness is taken to
supervene as a working hypothesis, to consider whether they are reproduced
in our interaction in the environment.
Others have appealed to biological function to provide a proper account
of the objects of perception, (e.g . Davies, 1983). Whether such appeals supply
276
the required materials will turn on a detailed examination of the grounds for
a ributing these functions, and a resolution of the issue of indeterminacy.
At the moment it looks like dierent a ributions of content may be a ributed
to an organism compatible with oering an explanation of their cognitive and
conative processes in terms of biological functions. However, perhaps this is a
failure to consider these ma ers at the right level of detail.
Further consideration of biological function also has utility in discussing the
second strain of recent investigation into the nature of intentionality to which
I wish to draw a ention. Problems with developing a naturalistic account
of intentionality have led some philosophers to consider the extent to which
intentionality, and the states which possess it, have normative properties. There
seems to be a sense in which it is legitimate to hold that if you perceive that
a white rabbit is nibbling grass to the le of the burrow, and you consider the
ma er, you ought to have the corresponding belief. If you grasp the concepts of
2 + 2 and 4, then you ought to arm that 2 + 2 = 4 (if you consider the ma er).
If you possess the concept dog, and are in the business of making dog judge-
ments, then you ought to apply the concept to dogs you come across and so on.
These oughts appear to be normative oughts and are related to the nature of
concepts and the meaning of words.
Apart from the intrinsic interest in understanding these normative claims,
the focus on normativity has promised to provide traction on the issue of
coming to a final evaluation of the possibility of a reductive account of inten-
tionality. Insisting upon the normative nature of content provides one way of
articulating why reductive accounts of intentionality arent possible. Everybody
knows you cant naturalize morality; now, the claim continues, we can see why
we cant naturalize intentionality either. Taming this normativity, on the other
hand, provides a way of keeping the option open.
A distinction is o en drawn between two ways in which these oughts can be
taken: as related to a norm or standard or as action-guiding (e.g. Ha iangadi,
2007, pp. 378). It is pre y much uncontroversial that they can be understood
in the first sense and, so understood, present no more diculty for naturalism
than the diculty it already faces with regard to intentionality in general. The
question is whether they should be understood in the second. Do they provide
guidance to action independent of our desires to follow these standards?
The action-guiding sense of ought in this context is compared with the case of
morality in which it is supposed that, when it is said we ought to behave
or ought not to behave in a certain way, this is independent of our desires to
behave, or not behave, that way.
It seems unlikely that the oughts relating to content have any action-guiding
role independent of the aims of the mental states in which they figure. Never-
theless, it is less clear whether we may take these aims to be rooted in our
interests. For example, is the aim of belief to be accounted for in terms of our
277
desire to act on the basis of what we take to be true (see Noordhof, 2001, for a
suggested positive answer)? Or is believing the truth desirable in a moral sense
(see Horwich, 2006, for a suggestion along these lines)? If the aims of mental
states are not rooted in our interests, then there may be scope to recognize a
class of categorical content-related oughts in the context of the states through
which these contents are possessed. However, I am not convinced that a posi-
tive answer to this last question threatens the reductive programme any more
than if they are conditional upon our interests. Categorial oughts may be more
controversial, in that it is questioned whether we really ought to do such and
such if we are uninterested in doing so, but the threat to the reductive pro-
gramme if any does not stem from where the normative properties are a rib-
uted but that they are at all. It is just as bad to suppose that if you have certain
desires, you ought to do such and such. What exactly does it mean to a ribute
such a property to the world (if thats what you are doing)?
Biological function provides a potential source of these aims which would
enable an understanding of them in physicalistically acceptable terms. How-
ever, the application of such an appeal to the understanding of the aims of
mental states is not straightforward. Consider the case of the aim of belief. It is
o en remarked that the aim of belief to be true explains why we cannot
consciously believe at will. How should we interpret the cannot? On one
interpretation, the cannot is as strong as metaphysical necessity. It is literally
not possible for a creature consciously to be able to do this. Appeal to biological
function has no capacity to explain this unless it is not possible for there to be
non-evolved creatures with beliefs.
Perhaps mental states can only have aims if the creature whose states they
are is biologically evolved. Then we would have to conclude that, either no
non-evolved organisms can have beliefs or that beliefs should be understood
by their causal profile independent of the aim. For ease of discussion, I shall
adopt the la er option (which is independently defensible (see Noordhof, 2001;
for a view to the contrary, see Velleman, 2000). It does not ma er for the issue
raised below since it can be simply re-cast for a state with the causal profile
alone. In any event, it is plausible that there may be creatures with states
with that causal profile which are non-evolved. So there is a limitation to the
biological explanation of the impossibility of consciously believing at will
right here.
However, things are worse than that. Appeal to biological function seems
in no be er shape to explain why creatures cant consciously believe at will
even if they are biologically evolved and the cannot is simply nomological. It is
a familiar fact that things with a certain biological function can malfunction
(e.g. the human heart can fail to pump blood). While we might accept that crea-
tures that can consciously will that they are in a state with the causal profile of
belief may die out quickly, there should be anomalies just as there are in all
278
other cases. Creatures are born with hearts that dont work; why arent they
born with consciously willable beliefs? In my view, the proper answer to this
question reveals something about the nature of consciousness, viz. that it
makes manifest the a ractiveness of being disposed to act on what you take
to be true (Noordhof, 2001).
Working out the connection between functions derived from our biological
heritage, features of consciousness, cognitive aims and so forth provides a
rich domain of study. It enables us to think about the role of various kinds of
normativity in the proper understanding of our mental life.
Concluding Remarks
Debate in the philosophy of mind in 1950s to 1970s seemed primarily focused

on which of the main physicalist theories of mind, if any, were true. Everything
was seen through the prism of counterexamples to the type-type identify
theories, problems for functionalism and the like. Focus in the late 1970s to the
early 1990s was on three issues which intertwined. First, there as the possible
truth of eliminativism about the mental; second, there was the debate about
whether externalism was true about mental content; third, there as the anxiety
about mental causation even if physicalism was true. It is only a slight exag-
geration to say that issues which came up under the second and third items
o en were pressed into service as a possible argument for the first: eliminativ-
ism. Although I have touched upon some of these issues in the current chapter,
I have tried to reflect my sense that, while the work from the 1950s was essential
to arrive at our current position, in many respects, the debate has moved in a
healthier direction.
Now there is far more focus on the detail of conscious states and the kind
of materials to which we should appeal to understand their character. It is
true that I have given the impression that the new centre of gravity is over
whether phenomenal content should be understood as the representationalist
recommends. However, the result of this discussion has been a much richer
understanding of both the nature of phenomenal consciousness, the nature
of intentionality and how they relate to each other or, if not richer under-
standing, at least richer appreciation of the issues at stake.
279
Glossary
action. Actions have been central to modern philosophy of mind ever since
Descartess contemporaries first criticized the account of mental causation sug-
gested by his substance dualism. Since behaviour involves physical movements
it was deemed that its causes must also be physical on pain of having to explain
how, when and where dualistic causation takes place. Descartes located it
in the pineal gland, a view rejected by modern substance dualists. Other posi-
tions in the philosophy of mind, such as those of property dualism, anomalous
monism, behaviourism, identity theory and eliminativism have all been moti-
vated by concerns related to the causation of behaviour.
While some philosophers use the terms action and behaviour interchange-
ably, others reserve the former for behaviour that is intentional and/or volun-
tary (at least under some description). Voluntary mental acts, such as the act of
calculating inside ones head, pose a problem for any species of behaviourism
which treats all behaviour as (necessarily) being publicly observable.
In the philosophy of mind actions are typically identified with events and/or
processes, however there is much dispute over which events actions are to be
identified with. Some insist that actions are identical to bodily movements. For
example, Davidson (who maintained that all actions are intentional under some
description) identified actions with movements of our bodies. Yet as Hornsby
has rightly cautioned, we must not conflate my moving my body with the
(mere) bodily movement I bring about when I move my body (for the term
bodily movement may be used in both a transitive and an intransitive sense).
This motivates the competing view that actions are the causes of bodily move-
ments (according to Hornsby, for example, all actions are tryings which may or
may not cause our bodies to move). We may wish to further follow Hornsby in
also distinguishing between the thing one did and (the event of) ones doing it
(thus mirroring the Fregean distinction between the thing one believes and
ones believing it). A third view, put forth by von Wright, identifies action with
the causing of an event. While this does not immediately rule out the possibility
that actions are events, it is perhaps misguided to always seek a precise location
of Xs causing of Y, as evidenced in the literature by various counterintuitive
claims concerning the spatio-temporal location of killings involving slow deaths.
Such issues regarding the ontology and individuation of actions are of
particular relevance to the question of whether reasons are causes which in
turn relates to questions concerning causation, agency, control and (ultimately)
free will.
280
Glossary
C. S.
Danto, A. (1973), Analytical Philosophy of Action, Cambridge: Cambridge University
Press.
Hornsby, J. (1980), Actions, London: Routledge.
Moya, C. (1990), The Philosophy of Action, Cambridge: Polity Press.
Stout, R. (2005), Action, Teddington: Acumen.
akrasia. See will.
animal minds. There is much debate over the extent to which some or all
non-human animals can be said to have minds. The arguments typically
revolve around what it is to have a mind, and in particular what it is to have a
so-called mental state such as a belief or desire. For example, some philo-
sophers maintain, contra behaviourism, that a dog cannot believe that the cat
is at the top of the oak tree unless it has the concept of an oak tree, where a
concept is a linguistic representation of some kind. The question of animal
minds is thus also closely related to questions about language and/or concept
acquisition. A dog may now desire to go for a walk immediately, but it cannot
now desire to go for a walk next Tuesday. Frankish here contrasts a behaviour-
based concept of mind with the language-involving concept of supermind.
C. S.
Beko, M., and D. Jamieson (eds) (1996), Readings in Animal Psychology, Cambridge:
MIT Press.
Glock, H. J. (2000), Animals, Thoughts and Concepts, Synthese, 123, 3564.
Searle, J. (1994), Animal Minds, Midwest Studies in Philosophy, 19, 20619.
anomalous monism. A metaphysical thesis, argued for by Davidson, that insists

on the identity of particular mental and physical events. It maintains that any
given causally ecacious mental event described in a mentalistic idiom must
be logically identical to some physical event. As such, when we speak about
causally ecacious mental happenings or properties we are in fact just speak-
ing about physical happenings or properties using dierent vocabulary. There
is just one thing present, not two hence the monism. This is the case even
though these particular happenings admit of radically dierent and irreducibly
incommensurable descriptions in the specialized vocabularies of folk (or every-
day) psychology and physics. The dierent ways that we describe such hap-
penings ma ers to our capacity to make systematic predictions and explanations
of what follows from such happenings. Thus the allegedly normative character
of our everyday mentalistic scheme for a ributing propositional a itudes, such
as beliefs, precludes the possibility of developing our mentalistic idiom into
a science of the mind that incorporates strict laws; hence, the anomalous
nature of the mental. By contrast, Davidson assumes that an ideal physics one
281
Glossary
capable of predicting and explaining the true causes of any given happening
only trades in strict causal laws. These assumptions frame his famous argument
that if mental events do in fact cause physical events then they must be logically
identical to some physical event or other, since only an explanation couched at
the level of an ideal physics can possibly get at the true cause of any physical
happening.
D. H.
Davidson, D. (1980), Mental Events, in Essays on Actions and Events, Oxford:
Clarendon Press.
(1987), Problems in the Explanation of Action, in Metaphysics and Morality,
Oxford: Blackwell.
Macdonald, C. (1989), Mind-Body Identity Theories, London: Routledge.
artificial intelligence. The project of programming computers or, more gener-

ally, designing machines, to engage in behaviour that would be counted as
intelligent if executed by a human. Artificial intelligence was the dominant
component of cognitive science in the 1960s and 1970s. During that period the
focus tended to be on high level cognitive skills such as game playing, logical
reasoning and natural language understanding. Such capacities were studied
in isolation from other cognitive, perceptual and motor abilities and were
modelled by processes involving the rule governed manipulation of represen-
tations. In the 1990s a new approach developed that focussed on designing
machines that do not heavily rely upon internal representations but can
successfully navigate real-world environments and so exhibit the intelligence
of animals much simpler than humans.
M. C.
Denne , D. (1998), Brainchildren: Essays on Designing Minds, Cambridge, MA: MIT
Press.
Haugeland, J. (ed.), Mind Design II, Cambridge, MA: MIT Press.
behaviourism. In philosophy the term behaviourism is typically associated

with a range of views according to which mental phenomena are to be analysed
(if not quite analysed away) by reference to behaviour. According to the ana-
lytic or logical behaviourism of Carnap and Hempel, for example, the verifi-
cation of any psychological statement will involve behavioural observations.
This view should be distinguished from the methodological behaviourism of
Watson (which states that psychology may legitimately only concern itself with
observable behaviour) and the even more radical psychological behaviourism
of Skinner (according to which all human an non-human behaviour is to be
explained without reference to the subjects inner life). More contentiously,
282
Glossary
behaviourism has also been associated with Ryle and Wi genstein, who both
emphasized the conceptual relations between behaviour and mental ascrip-
tions, though they arguably fell short of confirming hypotheses about psycho-
logical events in terms of behavioural criteria (to use Sellars characterization
of what makes someone a behaviourist). A recent counterexample to some
(though by no means all) strands of behaviourism is that of Galen Strawsons
Weather Watchers, viz. beings who are hypothesized to have a mental life
despite being constitutionally incapable of any sort of behaviour, as this is
ordinarily understood.
C. S.
Smith, L. (1986), Behaviorism and Logical Positivism, Stanford: Stanford University
Press.
Stout, R. (2006), The Inner Life of a Rational Agent, Edinburgh: Edinburgh University
Press.
belief. While Russell and McTaggart both wrote about states of mind as early
as 1921, and Turing talks of the internal states of machines in his influential
1950 Mind article Computing Machinery and Intelligence, it was not until the
rise of functionalism in the 1960s (and in particular Hilary Putnams 1964
paper Minds and Machines) that philosophers began to refer to beliefs as
mental states, alongside desires and other mental phenomena. Yet belief does
not generally appear to be a state of mind in the ordinary (emotional) sense of
the term, besides which one can only be in so many states of mind at any given
moment. Some theorists hold that there is an episodic occurrent form of belief
in which thoughts are actively brought to mind. It is true that we could not
have many of these at any given time, but this is not because they are states we
might be in. One may be in a state of nervousness or anxiety but it ordinarily
makes no sense to say that a subject, mind, or brain is in a state of belief (though
one may find oneself in a state of disbelief).
Beliefs so construed are held to have representational contents which relate
to (possible or actual) sates of aairs, much as the line on a gramophone record
relates to the notes played by the recorded orchestra. Representationalist
thought dates at least as far back as Hume, if not Plato. Wi genstein a acked
this characterization in his Philosophical Investigations which, among other
things, sought to repudiate his own, earlier, picture theory of the mind. To believe
that p, he came to think, is not to form any kind of representation of p, but sim-
ply to take p to be the case, thus explaining the impossibility of meaningfully
stating p, but I dont believe p. Moores paradox was founded on the worry
that the statement in question could nonetheless be true (since p, but Moore
doesnt believe that p is unproblematic). But there is no real paradox here, for
on the case-stating view (but not on the state-reporting one) to say that one
283
Glossary
believes that p is not to explicitly say anything about oneself, though one may
be disclosing much through conversational implicature.
Frankish helpfully distinguishes between two strands of belief, associated
with two distinct kinds of mental processing and, more generally, two concep-
tions of mind. The first (basic belief) is typically non-conscious, passive, non-
occurrent and a ributable on purely behavioural grounds (see animal minds).
The second (superbelief) may be held consciously, typically requires linguistic
conceptualization, and is frequently occurrent. Frankish argues that basic
beliefs and superbeliefs may conflict. For example, I superbelieve that the
indicator in the new car is on the le , yet on each turn I move my hand to the
right, as in the old car, suggesting that my basic belief (which I may become
aware of) contradicts it. This oers an alternative approach to Moores paradox,
on which the asymmetry between first and third person disappears. If p and
I dont believe p refer to a itudes of dierent types, then they may both be
assertable, even if both (or neither) of them are self-descriptions.
The ontology of (both kinds of) belief also crosses paths with theories of
truth. Inspired by Freges distinction between believing and the thing believed,
White has suggested that the term belief is simply ambiguous: it can either
refer to a proposition (which may be true or false) or to the believing of such a
proposition. However we do not believe our beliefs any more than we desire or
desires or fear our fears, and it is at best awkward to talk of believing, desiring,
or fearing propositions. In contrast to what we believe, a belief may be imagina-
tive, and this need not coincide with ones believing being imaginative (for I may
unimaginatively just latch onto your imaginative belief). Following Gilbert
Ryle, it is arguably a category mistake to think that we have beliefs in the same
sense in which we can be said to have pencils, mortgages, or family in Tanzania.
To claim, instead, that we paradigmatically ascribe beliefs to beings when
they behave (or when it is assumed they would behave) as if something were
the case is not to deny we may legitimately talk of beliefs that no one has
ever had; on the contrary there are numerous things which one could take
to be the case even if nobody has ever actually done so. Indeed, two people
may have exactly the same belief, for the beliefs we have are not particular
tokens of universal types, though ones having a belief is an instance of a
general case.
C. S.
Collins, A. (1987), The Nature of Mental Things, Notre Dame: Notre Dame Press.
Frankish, K. (2004), Mind and Supermind, Cambridge: Cambridge University
Press.
Steward, H. (1997), The Ontology of Mind, Oxford: Oxford University Press.
White, A. (1972),What We Believe, in N. Rescher (ed.), Studies in the Philosophy
of Mind, APQ monograph series no. 6, Oxford: Blackwell.
284
Glossary
causal closure. Causal closure is the claim that all physical events have a phys-
ical cause. If one accepts causal closure one denies the causal ecacy of the
non-physical. Ghosts cannot drag iron chains, and God cannot intervene in
the physical world. Non-physical minds cannot interact with physical bodies
and thus causal closure is thus at odds with mind-body dualism.
D. O.
Papineau, D. (2002), Thinking about Consciousness, Oxford: Oxford University
Press.
Chinese room argument. An argument devised by Searle to undermine strong

artificial intelligence, the doctrine that an appropriately programmed computer
would have a mind. Searle, who does not understand Chinese, imagines him-
self locked in a room that contains English instructions telling him how to
correlate some Chinese symbols with others. The instructions only mention
the syntactic properties of the symbols. Sheets of Chinese symbols are fed
into the room. Searle responds to this by following the English instructions, cor-
relating the new symbols with the symbols wri en on a batch of sheets. These
symbols are then copied onto blank sheets which are in turn posted out of
the room. It turns out that the input symbols are questions wri en in Chinese
and the output symbols are sensible answers to those questions.
Searles symbol-manipulating behaviour mimics that of a competent speaker
of Chinese, but he does not understand a word of it. The crucial point is that
Searle does exactly what a computer does symbol manipulation according
to syntactic but not semantic properties. Searle does not understand Chinese,
just as a computer does not understand any of the symbols it manipulates.
Thus, he claims, no computer, however it is programmed, is capable of under-
standing Chinese or any other language. He generalizes this result to all cogni-
tive capacities.
M. C.
Preston, J., and M. Bishop (2002), Views into the Chinese Room: New Essays on Searle
and Artificial Intelligence, Oxford: Oxford University Press.
Searle, J. (1980), Mind, Brains and Programs, Behavioural and Brain Sciences, 3,
41724.
computationalism. The view that the mind is a computer or ensemble of

computers embodied in the brain and that thinking is a kind of computation.
Interpreted narrowly, a computer is a mechanical device that manipulates
syntactically structured symbols by means of the application of symbol mani-
pulation rules. Although the device is sensitive only to the syntactic properties
of such symbols, the symbols do have meaning so that the device processes
285
Glossary
information when it computes. Such a view of the mind is central to what has
become known as classical cognitive science and is associated with Putnam,
Fodor and others. Interpreted more broadly, neural networks of the kind postu-
lated by connectionists are computers. So, to some, connectionism is a form of
computationalism.
M. C.
Fodor, J. (2008), LOT 2: The Language of Thought Revisited, Oxford: Oxford University
Press.
Haugeland, J. (ed.) (1997), Mind Design II, Cambridge, MA: MIT Press.
conceivability argument. See zombies.
concepts. Concepts are the constituents of thought. To possess a concept is to

be able to think about a certain aspect of the world. I can think of the stu in
my cup as coee, as bi er, as brown and as hot; and to think in these ways
I require the concepts coee, bi er, brown and hot. The concept coee is necessary
for various kinds of thoughts, for example, the concept coee enables me to
believe that coee is my favourite beverage, to desire a particular Italian rich
roast blend, and to hope that my cafetiere is not empty.
D. O.
Margolis, E. (ed.) (1999), Concepts: Core Reading, Cambridge: MIT Press.
connectionism. Connectionism is an approach in cognitive science according

to which the mind is a neural network or ensemble of such networks. Connec-
tionism came to prominence in the 1980s and is contrasted with what has
become known as classical cognitive science, as connectionist networks typi-
cally do not manipulate syntactically structured symbols or store information
by means of such symbols. Champions of connectionism are impressed by
several features of connectionist networks including their similarity to the
brain, their capacity to learn, their possession of a content addressable memory,
and their ability to deal with noisy input data and internal damage. One
particularly prominent debate surrounding connectionism relates to Fodors
charge that connectionism cannot account for the fact that it is a psychological
law that thought is systematic in the respect that anyone capable of thinking
that x stands in relation R to y (e.g. that John loves Jane) is also capable of
thinking that y stands in relation R to y (e.g. that Jane loves John).
M. C.
Bechtel, W., and A. Abrahamsen (2002), Connectionism and the Mind, second edition,
Oxford: Blackwell.
286
Glossary
consciousness. Consciousness signifies a number of dierent phenomena.

Sometimes the subject of the a ribution of consciousness is the creature per se.
We speak of a person or creature as being conscious or not. But there are
cases in which the subject of the a ribution is a particular mental state of the
creature. According to most philosophers, there are dierent kinds of con-
sciousness. Following Block (1995), they distinguish between phenomenal
(P-consciousness) and access (A-consciousness). A-conscious states are con-
scious propositional a itudes, for example, beliefs, judgements and desires.
Although P-conscious states can be A-conscious too, they can occur indepen-
dently of A-conscious states. The mark of A-conscious states is that they are
available for use in reasoning and rationally guiding speech and action. The
notion of A-consciousness is dispositional, not access, but accessibility is
required (also called global accessibility). On the other hand, the distinguishing
mark of P-consciousness, namely what makes a mental state P-conscious, is
that there is something it is like to be in that state. There are also higher-order
conceptions of consciousness, o en called reflexivity or monitoring conscious-
ness, which involve representation of ones mental states. Finally, there is the
notion of self-consciousness, which involves what psychologists call a theory
of mind, namely the ability to a ribute mental states in everyday life and to
reflect upon our mental lives and the mental lives of others.
D. P.
Armstrong, D. M. (1981), What is Consciousness?, in D. Rosenthal (ed.), The Nature
of Mind, Ithaca, NY: Cornell University Press, pp. 5567.
Block, N. (1995), On a Confusion about a Function of Consciousness, Behavioral
and Brain Sciences, 18, 22747.
Rosenthal, D. (2002), How Many Kinds of Consciousness?, Consciousness and
Cognition, 11 (4), 65365.
content. The term content is used equivocally in the philosophical literature.

It can refer either to what is thought about a thoughts subject ma er (e.g. a
state of aairs or object; whether real or imaginary) or the specific manner or
way in which that subject ma er is thought about. Content of the first type is
extensional. That of the second type is intensional (with an s). An important
dierence between these two kinds of content is that the former but not the
la er is transparent to substitution of coextensive terms. It is normally taken to
be a defining feature of contentful mental states that they are truth-evaluable
(i.e. such mental states have the property of being either true or false). Content-
ful mental states come in conceptual and non-conceptual varieties. Those of the
la er sort are allegedly possessed by thinkers who are unable to specify for
287
Glossary
themselves by means of concepts, verbally or otherwise, how it is that they

think about the world.
D. H.
Evans, G. (1982), The Varieties of Reference, Oxford: Oxford University Press.
desire. A psychological inclination towards an object or aim. Hume famously

claimed that desires do not have any representative quality, which renders
[them] a copy of any other existence or modification (Hume, 1978, p. 415). This
has recently been interpreted as suggesting that, unlike beliefs, desires do not
aim to represent any other existence. Their role is not to copy or represent the
world but to change it. Following this line of thought, in contemporary philo-
sophy of mind desires are identified primarily by their role in combining with
beliefs (about how the world is and can be changed) to generate actions which
will result in the world being changed to fit their content. Both beliefs and
desires are thought of as propositional a itudes, but as a itudes with distinct
aims or directions of fit. Beliefs aim to fit the world while desires aim to get the
world to fit them. What exactly is to be understood by this is not however
always agreed upon, nor is it always agreed that the role of desires in combin-
ing with beliefs to give rise to action can truly capture the full and varied nature
of desire.
I. M.
Hume, D. (1978), Treatise of Human Nature, ed. L. A. Selby-Bigge, Oxford: Clarendon
Press.
Sobel, D., and D. Copp (2001), Against Direction of Fit Accounts of Belief and
Desire, Analysis 61 (1), 4453.
dispositions. A disposition is best described as a propensity or liability of an

object, animal, or person to behave in certain ways under certain circumstances
(in which the disposition in question is triggered by the environment). Philo-
sophers of mind distinguish between the mental and physical dispositional
properties of any given subject, though o en they a empt a reduction of the
former to the la er. The current trend of describing dispositions as mental and/
or physical states which may act as mental or behavioural causes is at odds
with certain forms of behaviourism and various Wi gensteinian insights that
are (sometimes erroneously) associated with it.
C. S.
Mumford, S. (1998), Dispositions, Oxford: Oxford University Press.
dualism. Any view which posits the existence of just two kinds of something,
as opposed to just one kind of something (monism) is dualistic in nature. In the
288
Glossary
philosophy of mind there are substance dualists, most famously Descartes,

who maintain that ultimately there are just two kinds of stu in the universe:
physical stu and mental stu. There are also property dualists who argue
that there are at least two kinds of properties: physical and mental properties.
Dualists of whatever stripe hold that mental stu or mental properties are
neither reducable to nor explainable entirely in terms of the physical. The men-
tal really is its own kind of something, and it belongs on the final inventory of
the universe, right alongside quarks or mass.
Arguments for dualism come from many quarters: reflection on what Mary
did or didnt know, zombies, personal identity, language use and modality or
bare conceivability. Perhaps more famous are the diculties for dualism: its
lack of explanatory power, the problem of interaction, maybe just a general
failure to fit into a conceptually tidier monism. Whats interesting, too, is the
grip dualism has on us despite the answers we give to the arguments in its
favour and the troubles we find in thinking it true.
R. D.
Chalmers, D. (1996), The Conscious Mind, Oxford: Oxford University Press.
Descartes, R. (1642/1984), Meditations on First Philosophy, in J. Co ingham,
R. Stootho and D. Murdoch (trans. and eds), The Philosophical Writings of
Descartes, vol. 2, Cambridge: Cambridge University Press.
eliminative materialism. Eliminative materialists claim that our common sense

description of the mind our folk psychology is false. We do not have beliefs,
desires, hopes and fears. In a complete, true theory of the mind these categories
will not be reduced to physical categories; rather, they will be eliminated in
favour of the categories of a materialist theory that explains human cognition in
physical, probably neurophysiological, terms. Some eliminative materialists, such
as Quine, accept that folk psychology is indispensable to our everyday dealings:
we shall thus continue to talk of beliefs and desires even though, strictly speak-
ing, we do not have them. Others, notably Paul and Patricia Churchland, claim
that we should strive to je ison such false ways of speaking. In the future
we shall come to speak of each other, and see ourselves, not as believers and
creatures of desire, but rather, in terms of the categories of neuroscience.
D. O.
Churchland, P. (1979), Ma er and Consciousness, Cambridge: MIT Press.
emotions. Mental states such as fear, happiness, anger and (on some views)
desire, commonly associated with distinctive pa erns of sensation, thought
and behaviour. What constitutes the underlying nature of such states is a ma er
of much controversy as is the issue of whether they form a unified ontological
category at all.
289
Glossary
The etymology of the term emotion points to the most obvious theory of
emotions as visceral movements. Theories which emphasize this visceral or
sensation aspect of emotions have been widely criticized however for leaving
out the whole intentional dimension of emotions and for being unable to
account for either the rationality of emotions or for their complex role in ratio-
nalizing other states and actions. Such concerns underlie a number of more
recent theories of emotions cognitivist and perceptual theories in particular
according to which emotions are just special kinds of judgements or experi-
ences of the world as having emotion-specific evaluative properties such as
of being frightening, dangerous, worth avoiding or worth obtaining. These
theories retain however, in common with sensation theories, the view that
emotions are essentially conscious episodes, thus leaving them open to the
charge that they cannot make room for the existence of unconscious or non-
conscious emotions (i.e. ones not currently occupying ones a ention). As against
such theories, it is thus sometimes argued that emotions are best thought of as
mental dispositions dispositions to behave in various ways, to have ones
experiences of the world transformed or coloured in distinctive ways, and to
have a wide range of other mental states, both episodic and dispositional,
including thoughts, beliefs, desires and even other emotions.
How each of these approaches might be refined to accommodate the varied
aspects and wide ranging roles of emotions in our mental lives is the focus of
much current research. Questions being actively addressed include ones about
the rationality of emotions, about their role in motivating action, about the
nature of emotional experience and emotional expression as well as of course
fundamental questions about the ontological nature of emotions and about
our ability to know them.
I. M.
James, W. (1884), What is an Emotion? Mind, 9, 188205.
Sartre, J. (1962), Sketch for a Theory of the Emotions, London: Methuen.
Wollheim, R. (1999), On the Emotions, New Haven: Yale University Press.
epiphenomenalism. Mental events seem to have a causal eect on each other

and on our bodies. A pain can cause me to be depressed and to rub my sprained
ankle. Epiphenomenalism, though, is the thesis that such causal relations are
an illusion. Mental events play no causal role. They are, rather, side eects of
physical events in our brains and bodies. Physical event F causes me to rub my
ankle, and also has the side eect of causing me to feel pain. Certain pains, then,
tend to be followed by such behaviour, but this is because these pains and
ankle-rubbings are the product of a common cause, F, and not because pains
cause this behaviour.
290
Glossary
D. O.
Jackson, F. (1982), Epiphenomenal Qualia, The Philosophical Quarterly, 32, 12736.
explanatory gap. Contemporary neuroscience is teaching us that our mental

states correlate with neural processes in the brain. However, although we
know that consciousness arises from a physical basis, we dont have a good
explanation of why and how it so arises. Pain experiences for instance, correlate
with C-fibre firing, but even if we know that the feeling of pain correlates with
C-fibre stimulation and that say, the existence of such pain states is contingent
on the occurrence of such neural events, we still want to know why it doesnt
correlate with a neural state of another kind or why it is pain rather than the
feeling of elation or an itch that correlates with that particular kind of neural
state. This leads us to the more general question of why is it that there holds
such a correlation at all. Trying to answer such questions raises the problem of
the explanatory gap. The general idea is that physical properties the subject
ma er of physics can be exhaustively explained in objective-scientific terms,
that is, in terms of function and structure, but it appears that phenomenal prop-
erties cannot be explained in those terms. It is further claimed that phenomenal
properties cannot be explained in terms of cognitive abilities either in that the
la er can be given at least in principle a functional characterization.
Philosophers divide into five groups: (1) there is no explanatory gap or
there is one that is easily closable; (2) there is a deep explanatory gap for now,
but an answer might be forthcoming or we might someday close it; (3) there
is a permanent explanatory gap (we cannot close it in principle because we
suer from cognitive closure), but there is no ontological gap; (4) there is a
permanent explanatory gap, but we will never know whether there is an onto-
logical gap; (5) there is a permanent explanatory gap and a corresponding
ontological gap.
D. P.
Chalmers, D., and F. Jackson (2001), Conceptual Analysis and Reductive Explana-
tion, Philosophical Review, 110, 31561.
Levine, J. (1983), Materialism and Qualia: The Explanatory Gap, Pacific Philosophical
Quarterly, 64, 35461.
McGinn, C. (1989), Can We Solve the Mind-Body Problem? Mind, 98, 34966.
extended mind. The extended mind thesis holds that at least some cognitive
processes essential for enabling the completion of specific acts of cognition are
not wholly within the boundaries of the skin or the skull. Focusing on cases of
belief formation, Clark and Chalmers argue that sometimes successful cogni-
tion unavoidably depends on the use of environmental resources (e.g. appeal to
291
Glossary
information contained in notebooks or other devices). Moreover, they argue

that if the manipulation and use of such resources to support cognition were to
occur entirely within the bounds of the subjects head we would not hesitate
to class them as cognitive. On these grounds, it is argued that we have no rea-
son to deny that the machinery of the mind can, at least when certain specified
criteria are met, extend into the wider environment.
D. H.
Clark, A. (2009), Supersizing the Mind: Embodiment, Action and Cognitive Extension,
Oxford: Oxford University Press.
Clark, A., and D. Chalmers (1998), The Extended Mind, Analysis, 58, 719.
first-person authority. A subjects position as ultimate authority or expert

regarding the contents of his or her own mind. The literature on self-knowledge
is divided both on whether this authority exists and on whether, assuming it
does exist, it is epistemic (the result of our enjoying some form of privileged
access to our own mental states) or merely the consequence of some special
feature of our self-ascriptive judgements of the form I believe that p or I am in
pain. Sceptics about first-person authority note that failures to know what we
believe, desire, fear, etc. are both possible and frequent, as testified by common
cases of self-deception. Yet, despite the existence of such cases, it seems undeni-
able that at least sometimes our authority is unrivalled. If I now come to the
conclusion that space is not Euclidean, I seem to be far be er placed to know
that I have reached this view than anyone else, no ma er how a entive they
might be to my behaviour. The central task of any theory of self-knowledge is
to explain the authoritative position we at least appear to stand in with respect
to a wide range (if not all) of our current mental states.
Philosophers of a non-epistemic persuasion try to point to various distinc-
tive features of our authoritative statements (e.g. to their being self-verifying, to
our act of u ering them itself making us count as having them, or to their being
mere expressions of mental states rather than judgements about them). At the
other end of the spectrum, defenders of epistemic approaches to first-person
authority point to the non-authoritative status of some of our mental self-
ascriptions (about our unconscious minds) to argue that the authoritative
character of other of our self-ascriptions (about our conscious minds) cannot
be a feature of our self-ascriptive judgements considered merely as such that
is, merely as judgements with a particular form (the self-ascriptive form) or
subject ma er (our mental states). If a statement of the form I believe that p can
in some cases be authoritative and in other cases not be authoritative, the
authoritative status of those instances of self-ascription which are authoritative
cannot arise out of their form and content alone, but must arise instead out
of the particular way in which they were reached. Which line of approach,
292
Glossary
epistemic or non-epistemic, is the closest to the truth is a ma er of ongoing

debate, a debate further complicated by an independent threat thought to be
posed by externalism about mental content to the authority of even those of our
mental self-ascriptions which seem to be least open to challenge by others.
I. M.
Moran, R. (2001), Authority and Estrangement: An Essay on Self-Knowledge, Princeton,
NJ: Princeton University Press.
first-person/third-person perspective. The first-person perspective on mental

states comprises our subjective apprehension of our own mental states: the
knowledge we have, for example, of what we believe and desire, and of what it
feels like to have pains and tickles. The third-person perspective comprises an
objective description of mental states. Science, for example, a empts to provide
such a description.
This distinction is related to two fundamental issues in the philosophy of
mind. First, I seem to have a special kind of privileged access to my own mental
states. The Cartesian view is that we are infallible with respect to the contents
of our minds. I seem to be able to know with certainty what I am thinking and
feeling. You, however, can be in error with respect to my thoughts and feelings:
you can think that I am happy when I am not. Others, however, have construed
such first person authority in a weaker sense. I may have a distinct kind of
non-inferential access to my own mental states, but I am not infallible. There
are times, for example, when I may not know that I am jealous, or it may be
obvious from my behaviour that I believe that a certain climbing route is dan-
gerous, even though I am not aware that I have such a belief. Nevertheless the
default assumption is that I am in the best position to know what I am thinking
and feeling.
Second, there would appear to be aspects of the first-person perspective that
cannot be accounted for by a third person theory. Various kinds of dualists
argue that subjective features of the mind cannot be given an objective descrip-
tion: the way a ripe peach tastes to me cannot, for example, be explained in
objective, scientific terms.
D. O.
Alston, P. (1971), Varieties of Privileged Access, American Philosophical Quarterly,
8, 22341.
Nagel, T. (1979), What is It Like to Be a Bat? in Mortal Questions, Cambridge:
Cambridge University Press.
folk psychology. Folk psychology is the name given by philosophers to the

everyday practice of explaining, predicting and understanding intentional
293
Glossary
actions (i.e. our own and those of others) in terms of reasons. Engaging success-
fully in this practice requires being able to answer a particular sort of why
question by competently deploying the idiom of mental predicates (beliefs,
desires, hopes, fears, etc.) and a ributing these mental state terms appropri-
ately. Sometimes folk psychology is used to refer, more restrictively, to the
complete set of propositions and generalizations (or at least a perspicuous
presentation of the core body of these) that its practitioners are implicitly
commi ed to when using mental state terms in order to make sense of actions.
Used in this way, folk psychology is o en imagined to denote the theory that
would be obtained by systematically describing all of the relevant folk com-
mitments in a systematic way. There are several accounts of the nature of folk
psychology and how best to explain the relevant abilities associated with it
for example, it has been variously characterized as being, in essence: a kind
of theory; a practice involving modelling or simulation of mental states; and a
narrative practice. Among some of those who regard it as a theory its elimi-
nation has been called for by stressing its folk status and highlighting its lack
of fit with growing modern science.
D. H.
Carruthers, P., and P. Smith (eds) (1996), Theories of Theories of Mind, Cambridge:
Goldman, A. (2006), Simulating Minds: The Philosophy, Psychology and Neuroscience
of Mindreading, Oxford: Oxford University Press.
Hu o, D. D. (2008), Folk Psychological Narratives: The Socio-Cultural Basis of
Understanding Reasons, Cambridge, MA: MIT Press.
frame problem. The problem of understanding how a computational system

could accurately revise a complex network of representations of the world
to maintain its accuracy when one element of that network has been changed.
Following Denne , the problem can be illustrated by considering a robot driven
by a computer. The robot carries out an action, changing the world in a limited
way. As the world is a complex causal system, this change can be expected to
bring about many other changes in the world, some of them quite distant, while
leaving many other aspects of the world unchanged. The problem the robot
faces is that of eecting the appropriate changes without having to examine
every individual element of its complex representation of the world for doing
that would threaten cognitive breakdown if the network was of any great mag-
nitude. While many artificial intelligence workers have addressed the problem
head on, Fodor, a long-time advocate of computationalism, regards the problem as
so grave as to suggest that central cognition might not be a form of computation
a er all.
294
Glossary
M.C.
Denne , D. (1998), Brainchildren: Essays on Designing Minds, Cambridge, MA: MIT Press.
Fodor, J. A. (2000), The Mind Doesnt Work That Way, Cambridge, MA: MIT Press.
functionalism. Tables are not defined in terms of their physical structure since
they can be made out of all kinds of stu, including metal, wood, and plastic.
They are defined according to their function: a table is (roughly) something
which we use to put things on. Functionalists have a related view of the mind.
I can feel pain and so perhaps can a tuna fish, yet our brains are structured
dierently. It is also conceivable that creatures on other planets can feel pain,
and perhaps future robots, but such beings will have very dierent brains to us
and to tuna. Mental states are not therefore defined in terms of their physical
structure; they are, rather, defined by their causal relations. Pains are the kind
of state that are caused by bodily damage and that lead to avoidance behaviour
and depression. If an alien creature is in the kind of state that bears those
relations with its behaviour and other mental states, then it is in pain. As Hilary
Putnam (1975, p. 291) claims: we could all be made of Swiss cheese and it
wouldnt ma er.
A key problem for functionalism lies in accounting for the subjective feel
of mental states, or what is called their phenomenology or qualitative nature.
Pain may have the causal relations that functionalists say it does, but pain
also feels a certain way it hurts and it is not clear how functionalists can
account for this fact.
D. O.
Putnam, H. (1975), Philosophy and Our Mental Life, in Philosophical Papers, vol. 2,
Cambridge: Cambridge University Press.
H20/XYZ. XYZ is a fictional chemical invented by Hilary Putnam in his Twin

Earth thought experiment. It is superficially identical to what we on Earth call
water: it is colourless, tasteless, odourless and thirst-quenching; it rains from
the sky and flows down rivers into the sea. It is, however, physically dierent
from H20 (water on Earth). It is not comprised of hydrogen and oxygen; its
chemical formula is very long and complicated and is abbreviated to XYZ. It is
dierent stu. Descriptivists claim that both XYZ and H2O are water since (at
normal temperature and pressure) they both satisfy our common sense descrip-
tion of water: it is the stu that flows down rivers into the sea, is colourless,
etc. Essentialists, however, claim that chemists have discovered an essential
property of water, that it is H2O, and nothing can be water that does not have
this chemical structure. Conclusions concerning this are relevant to debates
concerning semantic and cognitive externalism.
295
Glossary
D. O.
Putnam, H. (1975), The Meaning of Meaning, in Philosophical Papers, vol. 2,
higher-order thought. A thought or belief about ones own mental states. In

the philosophy of mind the term most o en comes up in discussions of so
called higher-order thought theories of consciousness, according to which
the presence of a higher-order thought or belief is what makes the lower-order
state it is about conscious. Numerous objections have been raised against such
views, generally taking the form of counterexamples aimed at showing that,
and how, a states being conscious can come apart from its being accompanied
by a second-order belief about it. As a result of such objections, higher-order
thought theories have been revised and refined in numerous ways and have
taken on a number of distinct forms. Whether any such theories can however
ultimately stand up to scrutiny while retaining their essence as theories of
what constitutes consciousness remains open to question.
I. M.
Rosenthal, D. (1991), Two Concepts of Consciousness, in D. Rosenthal (ed.), The
Nature of Mind, Oxford: Oxford University Press.
intentionality. Intentionality is the aboutness or directedness of thought

towards an object. A thought intends an object, establishes an intentional rela-
tion with it as the thing that the thought is about, which is then also and equiva-
lently said to be the object intended by the thought or the thoughts intended
object. Intentionality is the property of thought or the expression of thought
whereby it is about, refers to, or is directed on something.
The intentionality of thought is recognized by Aristotle and emphasized by
certain of the medieval philosophers. It was revived for modern philosophy
in the nineteenth century by Franz Brentano in his 1874 book, Psychologie vom
empirischen Standpunkt [Psychology from an Empirical Standpoint]. Brentanos
intentionality thesis combines two parts. Brentano maintains both that: (1) every
psychological (psychic) phenomenon intends an object, and (2) intended objects
belong to and are contained in the psychological phenomena by which they
are intended. Many of Brentanos later followers accepted (1) but denied (2),
and most if not all intentionalists in the philosophy of mind today are of the
same inclination, accepting Brentanos intentionality thesis in the general form
of Brentanos thesis (1), sometimes with modifications, while rejecting the spe-
cific immanence or in-existence thesis that Brentano accepted in (2). Without (2)
it is possible to say in closer conformity to common sense that at least some
intended objects exist outside the mind, so that we can perceive and want,
hope for, believe about, doubt, love, hate, and so on objects that do not merely
296
Glossary
exist in but transcend the acts of perception, wanting, hoping, etc. Thus, at least
some intended objects are intended as existing in the external world outside of
thought if they do not simply belong to the thinkers imagination.
Brentano, F. (1973), Psychology from an Empirical Standpoint, ed. L. McAlister, trans.

A. Rancurello, D. Terrell and L. McAlister, London: Routledge and Kegan Paul.
Denne , D. (1987), The Intentional Stance, Cambridge: MIT Press.
Searle, J. (1983), Intentionality: An Essay in the Philosophy of Mind, Cambridge:
internalism/externalism. Internalism maintains that the defining properties of

mental states are non-relational or intrinsic. Such properties are either identical
to or supervene on the internal, microstructural properties of sentient and sapi-
ent beings (e.g. their brain or bodily states). Externalism maintains that at least
some mental state properties are extrinsic and hence necessarily relational. The
dierence between intrinsic and extrinsic properties can be illustrated by the
dierence between mass and weight. The mass of a body (an intrinsic property)
is a measure of how much ma er it contains, whereas the weight of the same
body (an extrinsic property) depends on situational factors such as the force
of a raction between it and other bodies. If internalism is true mental states
can be individuated wholly by appeal to internal properties that reside inside
the skin or head of an agent; if externalism is true mental state individuation
will necessarily require an additional appeal to external factors as well.
D. H.
Fodor, J. A. (1994), The Elm and the Expert, Cambridge, MA: MIT Press.
Putnam, H. (1975), The Meaning of Meaning, in Mind, Language and Reality:
Philosophical Papers, vol. 2, Cambridge: Cambridge University Press.
(1988), Representation and Reality, Cambridge, MA: MIT Press.
introspection. The special way each person has of knowing the contents of his
or her own mind, although what this way is, whether it is in fact special and
whether it is a way of knowing at all, are all ma ers of some controversy. The
etymology of the term introspection suggests that introspection is essentially
a form of inner perception and that to introspect is in some sense to look
inside. But, one might ask, in what sense? The verbs to perceive and to see
are pre-theoretically used in a variety of ways, in particular to mean quite gen-
erally to know or to understand as when speaking of seeing what someone
means. Similarly, to look is o en used to mean to think about or to investi-
gate as when promising to look into some ma er. In its modern theoretical
sense however perception refers fundamentally to external sense perception.
Speaking of introspection in the theoretical context of philosophy can thus be
seen to establish an analogy between so called inner perception and external
297
Glossary
sense perception, or between our way of knowing our own mind on the one
hand and our way of knowing that which lies outside it, through sense percep-
tion, on the other. Few current philosophers however believe this analogy to be
legitimate.
Introspection is argued instead to be either a (particularly quick and well
informed) process of inference from observation of our behaviour (and so not
a special way of knowing, distinct from our way of knowing the mental states
of others), or, some essentially non-epistemic process of avowal of our current
mental states (such as that of mere expression of our beliefs and desires and
so not a way of knowing at all). Beyond standard epistemic and non-epistemic
accounts of introspection, the recent literature has seen a number of further,
o en more subtle, views of introspection being espoused on both sides. The
central task for any such theory remains that of providing an account of at least
the appearance of our having a special way of knowing our own minds which
is unlike our way of knowing the minds of others and which displays certain
key characteristics (e.g. of immediacy, first-person authority and immunity to
certain types of error).
I. M.
Armstrong, D. M. (1968), A Materialist Theory of Mind, London: Routledge and
Kegan Paul.
Heal, J. (1994), Moores paradox: a Wi gensteinian approach, Mind (January).
Shoemaker, S. (1996), Self-Knowledge and Inner Sense, in S. Shoemaker, The
First Person Perspective and Other Essays, Cambridge: Cambridge University Press.
language of thought. A language that the mind employs in thinking, as postu-

lated by Fodor. For Fodor, the language of thought is shared by all members of
the human species regardless of what natural language they speak. Like a natu-
ral language, it consists of a vocabulary of a finite number of simple symbols
along with a finite number of formation rules for combining these simple sym-
bols into more complex symbols. In addition to their syntactic properties, sen-
tences of the language of thought have semantic properties, and the meaning or
content of such a sentence is a product of the meaning of its component simple
symbols and its syntactic structure. For Fodor, whenever an individual has a
particular propositional a itude (e.g. a belief that dogs chase postmen) there
will be a sentence of the language of thought bearing the appropriate meaning
or content located in his or her mind. Sentences of the language of thought are
physically embodied by states of the brain, but they are multiply realizable at
the physical level.
M. C.
Cain, M. J. (2002), Fodor: Language, Mind and Philosophy, Cambridge: Polity.
298
Glossary
Fodor, J. A. (2008), LOT 2: The Language of Thought Revisited, Oxford: Oxford

University Press.
logical behaviourism. See behaviourism.
mental causation. It is widely assumed that mental properties, such as being

angry, feeling uneasy, or noticing that the bridge is unsafe, can and typically
do cause action and behaviour. Such mental causes are o en thought to be
productive (i.e. mental events or properties generate other events, just as the
collision of one billiard ball with another is thought of as forcefully producing
motion). However it is also possible to think of mental causation in the weaker
terms of counterfactual dependence where the occurrence of one event depends
upon the occurrence of another, as in the case in which my washing up the
dishes depends upon my remembering a promise I made to my wife (such that
had I not remembered, the washing up would not have been done by me).
Those philosophers who insist that mental causation must be of the produc-
tive variety and who also endorse some form of non-reductive physicalism
encounter a famous problem: the exclusion or pre-emption problem. It is the
worry that any causal contribution that a mental property might make to the
occurrence of another event (e.g. an action) will be systematically usurped by
its neural or other physical realizers. This is due to the assumed relation of
ontological dependence holding (vertically) between mental states and their
realizers. As the problem is entirely metaphysical it holds even though, for
practical purposes, the mentalistic and, say, neuroscientific explanations are
quite dierent.
D. H.
Davidson, D. (1980), Actions, Reasons and Causes, in Essays on Actions and Events,
Oxford: Clarendon Press.
Heil, J., and A. Mele (eds) (1993), Mental Causation, Oxford: Oxford University Press.
Kim, J. (2008), Mind in a Physical World: An Essay on the Mind-Body Problem and
Mental Causation, Cambridge, MA: MIT Press.
mentalese. See language of thought.
methodological behaviourism. See behaviourism.
mind-body problem. The mind-body problem is the central and most dicult
problem in the philosophy of mind. It is the challenge of explaining the meta-
physics of the relation between the physical body, especially the brain and
nervous system, and the events of consciousness experienced by a thinking
psychological subject.
299
Glossary
Eorts to solve the mind-body problem can be divided into two main cate-
gories, those that (1) try to reduce or eliminate all meaningful reference to the
mind to or in favor of purely physical material entities, properties and events,
and (2) those who argue that no such reductions or eliminations could possibly
be adequate to the relevant data we find not only in reflecting on the content of
our own subjective psychological lives and those we a ribute to others, that
we learn about from their expressions of thought, but also in the objective
external behavior of other psychological subjects. The main theories in category
(1) include eliminative or reductive (a) behaviorism, (b) materialism, and (c)
functionalism, the la er subsuming (d) computationalism as a special type
that is otherwise simply identified with functionalism. These theories maintain
either that there are no such things as thoughts, states of consciousness, or the
mind, or that whatever can truly be said of mental entities and events can be
interpreted in a vocabulary consisting entirely of terms for purely physical
material substances, entities, properties and events. The main theories in
category (2) include: (a) what is alternatively called, a er Descartes, Cartesian,
substance, or ontic dualism, and (b) property dualism. Property dualism in
category (2b) is in turn sub-divisible into (i) intentionalist- and (ii) qualia-based
philosophies of mind that emphasize either intentionality or the existence and
nature of qualia as explanatorily ineliminable and physically or materially
irreducible. There is also no reason why a (2b) property dualism could not
accept both (i) and (ii) in opposition to mind-body eliminativism or reductivism
in trying to solve or at least clarify the mind-body problem, and then alterna-
tively ordering (i) and (ii) in terms of explanatory or other priority, with either
(i) taking precedence over (ii) or the reverse, or of treating (i) and (ii) as distinct
but explanatorily equally significant and important grounds for denying the
truth of any mind-body solution in category (1). Since it is absurd to suppose as
eliminativism does that, despite appearances, thoughts and the mind do not
exist, appearances themselves being states of mind, and since Cartesian, sub-
stance or ontic dualism is widely believed to be indefensible, it is possible
to speak in practical terms of the mind-body problem as a contest between
some form of physical reductivism, behavioral, material, or functional (compu-
tational) on the one hand, and, on the other, some form of property dualism
that emphasizes intentionality over qualia or qualia over intentionality, or
gives equal importance to both intentionality and qualia in understanding the
nature of mind.
D. J.
Descartes, R. (1641), The Meditations on First Philosophy, in which the Existence of God
and the Distinction of Mind and Body are Demonstrated, trans. and ed. J. Co ingham,
Jacque e, D. (2009), Philosophy of Mind: The Metaphysics of Consciousness, London:
Continuum Books.
300
Glossary
McGinn, C. (2000), The Mysterious Flame: Conscious Minds in a Material World, New
York: Basic Books.
moral psychology. The branch of ethics that deals with human and animal
psychology, particularly as it relates to moral judgement and motivation. Its
primary aim is to investigate the relation between vice and virtue and our
(general and individual) cognitive and conation abilities including those of
belief, desire, impulse, intention and volition. Various views within moral psy-
chology debate the extent to which ethical norms and reasons are relative to
agential character traits or dispositions (e.g. (a) whether or not moral judge-
ments necessarily motivate and (b) whether one can have a normative reason
for doing something even if there was nothing in their motivational set that
could ever (either directly or indirectly) move them to do it).
C. S.
Blackburn, S. (1998), Ruling Passions, Oxford: Oxford University Press.
Co ingham, J. (1998), Philosophy and the Good Life, Cambridge: Cambridge University
Press.
Smith, M. (1994), The Moral Problem, Oxford: Blackwell.
multiple realizability. Computational devices can be made out of various

kinds of physical systems: metal cogs (as was the first computer, Babbages dif-
ference engine), valves, silicon circuit boards, or perhaps, in the future, lasers.
Computers are therefore multiply realizable. Similarly, according to the func-
tionalist, human physiology is not the only kind of physical system in which
minds can be realized. There may perhaps be alien species that have mental
states akin to ours they might feel pain and desire food but these aliens
could be made out of dierent physical stu. Mental states are therefore
multiply realizable.
D. O.
Kim, J. (1992), Multiple Realizability and the Metaphysics of Reduction, Philosophy
and Phenomenological Research, 52, 126.
neural network. A network of simple processing units modelled on the human

brain. Such networks come in a variety of forms, but a standard network con-
sists of units arranged into three layers: an input, an intermediate and an out-
put layer. At any point in time each unit will be in a state of activation and
require a certain amount of stimulation in order to become active (this is its
threshold value). Each unit in the input layer is linked by a number of connec-
tions to many units in the intermediate layer and the intermediate units are
similarly linked to units in the output layer. Impulses are transmi ed along
these connections.
301
Glossary
The connections have weights so that they can amplify or dampen the
strength of an impulse they carry and can be either excitatory or inhibitory.
When units in the input layer are stimulated by the outside world, impulses
pass along connections to the intermediate layer so stimulating activity there.
Impulses are then passed to the output layer resulting in pa erns of activity
at that level. Consequently, the system transforms pa erns of activation at
the input layer into pa erns of activation at the output layer. The systems
input-output behaviour is determined by the nature and weight of the con-
nections and the threshold values of the units. Adjusting the connection
weights will alter the systems input-output behaviour. As the pa erns of
activation can have semantic significance, the network can serve as an infor-
mation processor.
M. C.
Bechtel, W., and A. Abrahamsen (2002) Connectionism and the Mind, second edition,
Oxford: Blackwell.
other minds. I know that I have a mind, but how can I be sure that others do?
I see people writing shopping lists, running for the bus and talking to each
other, but it is possible that all these people are just mindless automata, their
actions akin to the behaviour of non-sentient robots. I might be the only mind
in existence! This is called the problem of other minds.
Various solutions to this problem have been oered. I could come to know
that others have a mind by analogy. I know that my behaviour is caused by my
mental states, and since others behave in a similar way to me, I can infer that
their behaviour is caused by their mental states. In contrast behaviourists claim
that behaviour is not merely the surface eect of underlying mental causes.
As Ryle puts it in The Concept of Mind, Overt intelligent performances are not
clues to the workings of minds; they are those workings. Boswell described
Johnsons mind when he described how he wrote, talked, ate, fidgeted and
fumed. For the behaviourist, then, it is my perceptual experience of the beha-
viour of others that justifies my belief in other minds. Last, and most popular,
is a theoretical account of the mind. I am justified in believing that others have
minds in the same way that I am justified in believing that stars are giant nuclear
reactions. The physics of nuclear reactions can be used to predict and explain
the behaviour of stars, and folk psychological categories can be used to predict
and explain the actions of people. The reasoning applied here is inference to
the best explanation. If there is a theory that explains the occurrence of certain
phenomena be er than any alternative theory, then we are justified in believing
that theory. I am therefore justified in believing in folk psychology and the
existence of other minds.
302
Glossary
D.O.
Avramides, A. (2001), Other Minds, London: Routledge.
perception. Perception can be generally characterized as a process in virtue

of which we select, organize and interpret sensory stimulation and sensation
into a coherent experience of the world. We can categorize perception as
inner and outer: inner perception bodily awareness involves awareness
or apprehension of the goings on in our bodies (proprioception); and outer or
sensory perception involves awareness or apprehension of the goings on in
the external world outside our bodies. The la er involves the use of our senses
of sight, hearing, touch, smell and taste. The traditional philosophical problem
of perception involves the question of whether we perceive physical objects
and properties directly or indirectly in virtue of being directly aware of
some sensory or mental items that represent the physical objects and their
properties.
D. P.
Maund, B. (2003), Perception, Chesham: Acumen.
Robinson, H. (1994), Perception, New York: Routledge.
perceptual content. Perceptual content refers to the things and their properties
that feature in ones perceptual experience, or in other words, it is what is con-
veyed to one by ones perceptual experience via the five sense-modalities and
proprioception. According to some philosophers, perceptual content is repre-
sentational since it seems that our perceptual experiences are by their nature
such that they present the world as being a certain way. Perceptual experiences
seem to have accuracy conditions; they are accurate in certain circumstances
and inaccurate in others and therefore are assessable for accuracy.
If one thinks that experiences have representational content then one thinks
of them as belief-like in some respects: Believing for instance, that there is a cup
on the table is being in a state with representational content. But if one claims
that an experience has representational content, that does not commit one to
identifying experiences with beliefs, since (1) experience may not be the same
a itude as belief, and (2) if (1) is false, they may be both a itudes to a dierent
kind of content. Regarding the la er, some philosophers believe that perceptual
states can represent the world without the subject of those states possessing the
concepts required to specify their content. On this view, ones experience of the
world is not constrained by ones conceptual capacities.
We can further ask, in virtue of what do perceptual experiences have the
content they have and represent the state of aairs they represent? Externalism
about perceptual content (also called phenomenal externalism) holds that the
contents of experience are not determined by the internal states of the brain but
303
Glossary
rather by external facts. For example, my Twin Earth counterpart (a molecular

duplicate of me in a dierent external environment) and I, being in exactly the
same brain state, have dierent experiences. On this view, what makes it the
case that a particular experience has the content is has depends on (external)
relations outside the subjects body, such as social and causal relations to things
in the environment. According to internalism, the contents of experience are
determined by the internal states of the brain, not by external facts; my Twin
Earth counterpart and I, being in exactly the same brain state, cannot have
dierent experiences.
D. P.
Crane, T. (2003), The Intentional Structure of Consciousness, in Q. Smith and
A. Jokic (eds), Consciousness: New Philosophical Perspectives, Oxford: Oxford
University Press.
Gunther, Y. (2003), Essays on Non-Conceptual Content, Cambridge, MA: MIT Press.
Lycan, W. (2001), The Case for Phenomenal Externalism, Philosophical Perspectives,
15, 1735.
phenomenal concepts. Phenomenal concepts are our concepts of conscious

states. Proponents of the phenomenal concept strategy (Stoljar, 2005) claim that
these concepts have a special nature and given that nature, it is not surprising
that we find an explanatory gap between physical processes and phenomenal
properties: the former are conceived under physical concepts and the la er
under phenomenal concepts. The gap then involves the relationship between
these concepts and our possession of phenomenal concepts can be explained
in physical terms.
D. P.
Chalmers, D. (2007), Phenomenal Concepts and the Explanatory Gap, in T. Alter
and S. Walter (eds), Phenomenal Concepts and Phenomenal Knowledge: New Essays on
Consciousness and Physicalism, Oxford: Oxford University Press, pp. 16795.
Stoljar, D. (2005), Physicalism and Phenomenal Concepts, Mind and Language, 20,
46994.
phenomenology. The study of phenomena or the appearances of things as

they appear in our experience. The properties in virtue of which there is some-
thing it is like for one to be in a mental state are called phenomenal properties
and constitute the ways in which our experiences or phenomenally conscious
states dier; there is for instance something it is like for one to feel a sharp pain
or an itch in ones finger, as there is also something it is like for one to smell
coee brewing or to see the vivid colours of a sunset. Mental states with such
properties include perceptual experiences, bodily sensations and felt emotions
or moods but may also include conscious thoughts and propositional a itudes
304
Glossary
(i.e. experiences in which no qualitative property, for example, sensory or aec-

tive quality) is somehow involved. An example might be the phenomenon of
understanding experience (Strawson, 1994). The la er refers roughly to there
being something it is like for one to understand a spoken sentence over and
above the stream of sound or of any sensory qualities that may be somehow
involved.
Phenomenology in the Continental tradition refers to the study of struc-
tures of consciousness as experienced from the first-person point of view;
the study of experiences themselves and their interrelations, not the search
for laws or causal explanations. Phenomenology is not concerned so much
with the world as such but rather the world as appearing to consciousness.
It goes beyond the study of the phenomenal character of our experiences,
addressing the meaning things have in our experience, that is, the significance
of objects and events, the flow of time, the structure of mental content, tempo-
ral awareness, bodily awareness, memory, imagination, embodied action and
the self.
D. P.
Brentano, F. (1995/1874), Psychology from an Empirical Standpoint, ed. L. McAlister,
trans. A. Rancurello, D. Terrell and L. McAlister, London and New York:
Routledge.
Carruthers, P. (2006), Conscious Experience versus Conscious Thought, in
U. Kriegel and K. Williford (eds), Consciousness and Self-Reference, Cambridge,
MA: MIT Press.
Smith, D., and A. Thomasson (eds) (2005), Phenomenology and Philosophy of Mind,
Oxford and New York: Oxford University Press.
Strawson, G. (1994), Mental Reality, Cambridge, MA: MIT Press.
physicalism. According to substance dualism there are two kinds of substance

in the world, mental and material (or physical). Descartes claimed further that
each of us is a union, made up of a material substance the body and a mental
substance the mind. By substance we normally mean something that can
exist independently, have properties, and enter into relationships with other
substances. If something is a substance then it can exist in such a way as to
stand in need of nothing beyond itself in order to exist. Physicalists deny that
there are two kinds of substance. According to them, there is only one kind of
substance, namely physical substance, and all that exists in the world are bits
of ma er in space-time.
There are, in the main, two versions of physicalism: reductive and non-
reductive physicalism (or property dualism). According to the former, there is
one kind of substance, physical, and mental properties are reducible to, and
reductively identifiable with, physical properties. According to the la er there
is one kind of substance, physical, but two dierent kinds of properties, mental
305
Glossary
and physical properties, and the former are distinct from and irreducible to
the la er.
D. P.
Chalmers, D. (2002), Consciousness and Its Place in Nature, in id. (ed.), Philosophy
of Mind: Classical and Contemporary Readings, New York: Oxford University
Press, pp. 24772.
Kim, J. (2005), Physicalism, or Something Near Enough, Princeton: Princeton University
Press.
privileged access. The special epistemic position each person stands in with
respect to the contents of his or her own mind. It is o en noted in the literature
on self-knowledge that we seem to be be er placed than anyone else to say
what mental states we are in. Whether we are be er placed by virtue of having
a special way of accessing our own mental states (e.g. some form of inner sense)
or merely by virtue of tending to have more evidence available (due to our
greater proximity with ourselves) is a ma er of some controversy. Equally con-
troversial is the wider issue of whether the authoritative status of our judge-
ments about our own mind is truly the result of a cognitive achievement on
our part (and so the result of a form of access), or merely the consequence of
some special feature of our self-ascriptive judgements of the form I believe
that p or I am happy or indeed I am in pain.
I. M.
Alston, W. (1971), Varieties of Privileged Access, American Philosophical Quarterly, 8.
propositional aitudes. Propositional a itudes are psychological a itudes of

various types that relate a thinker to specific propositions. The content of such
propositions is paradigmatically expressed in natural language by sentential
that-clauses such as, London is twenty miles away. For example, a thinker, X,
can believe, desire, hope, fear, or recognize that p; where the mental state verb
denotes Xs a itude and p denotes the content to which Xs a itude relates. The
propositions in question may be true or false (e.g. things may or may not be as
the thinker takes them to be) and as such propositional a itudes are thought to
be states of mind that possess or relate to truth-evaluable contents. Depending
on ones view of the nature of propositions, propositional a itudes will be
regarded as simple or more complex states of mind. They either relate a thinker
directly to some actual or possible state of aairs or indirectly to some actual or
possible state of aairs via a specific mode of presentation or representation.
D. H.
Russell, B. (1918), The Philosophy of Logical Atomism, The Monist; reprinted in R. C.
Marsh (ed.), Logic and Knowledge: Essays 19011950, London: Unwin Hyman, 1956.
306
Glossary
Frege, G. (1988), Thoughts, in N. Salmon and S. Soames (eds), Propositions and

A itudes, Oxford: Oxford University Press; alternative translation: The Thought:
A Logical Inquiry, in P. F. Strawson (ed.), Philosophical Logic: Oxford Readings in
Philosophy, Oxford: Oxford University Press, 1976.
Quine, W. V. (1956), Quantifiers and Propositional A itudes, Journal of Philosophy,
53; reprinted in W. V. Quine, The Ways of Paradox, Cambridge, MA: Harvard
University Press, 1966.
psychoanalysis. A psychological method for investigating the mind, originally

with the further clinical goal of treating mental illness. While the method was
founded by Breuer and Freud, it was later developed (in ways that they would
not have necessarily approved of) by Jung, Melanie Klein and Freuds daughter
Anna. Non-clinical psychoanalytic theories have also been put forth by critical
thinkers such as Lacan. Despite numerous important theoretical dierences
between all these, they remain united by a focus on the analysed persons
descriptions of their own thoughts, emotions, defences, dreams, fantasies and
free associations. Psychoanalysis is o en mistakenly conflated with depth
psychology, viz. any psychological approach that focuses on the unconscious.
C. S.
Freud, S. (1964), New Introductory Lectures on Psychoanalysis, trans. J. Strachey,
London: Hogarth Press.
psychological behaviourism. See behaviourism.
psychology. The study of how human and animal minds function, with the
primary aim of providing explanatory theories of how our knowledge of the
mind helps to explain or manipulate behaviour. Much psychological theory
thus o en falls between philosophy and science (in particular psychiatry);
indeed the godfather of modern psychology is o en said to be Nietzsche who
put forth his concept of a will to power as an explanation of the behavioural
drives of all living things. Not unlike the philosophy of mind, psychology
divides itself into numerous specialities such as cognitive, motivational, social,
neural, educational, perceptual, or cultural psychology. Such theories may then
be variously applied in fields as diverse as advertising, military strategy and
emotional counselling.
C. S.
Smith, E. et al. (eds) (2002), Atkinson and Hilgards Introduction to Psychology,
fourteenth edition, Belmont, CA: Wadsworth Publishing.
qualia. The word qualia is used in stronger and weaker ways in the philo-
sophical literature. Sometimes it is used to designate the distinctive quality,
307
Glossary
characteristic feel or phenomenal character of token experiences (perceptions,

sensations, feelings, moods) (i.e. what-it-is-like for a subject to undergo such
experiences). In this broad use it is o en equated with having an idiosyncratic
first-personal point of view or perspective of a certain experiential character.
In its more restrictive use, the term denotes inner, intrinsic and introspectable
mental particulars or properties particulars or properties that are entirely
inaccessible and invisible to third-personal analysis and with which such sub-
jects have a privileged and private acquaintance. These properties are allegedly
unique and are u erly distinct from purely intentional, representational or
functional properties. So construed, some believe that qualia literally constitute
part of the content of logically private, privileged first-person reports of our
inner states of mind and that a special first-personal science of consciousness
would need to be developed if we are to study such properties.
D. H.
Chalmers, D. (1996), The Conscious Mind, Oxford: Oxford University Press.
Denne , D. C. (1991), Consciousness Explained, New York: Penguin Books.
Flanagan, O. (1993), Consciousness Reconsidered, Cambridge, MA: MIT Press.
reasons. Agents are said to act for or in the light of reasons when they act with
some aim or purpose in mind. Many philosophers hold that all intentional
action is performed for a reason while almost all hold the converse view that all
actions performed for reasons are intentional. There is much ontological debate
over whether reasons should be conceived of as facts, states of aairs, proposi-
tions, mental states, or some disjunctive combination. Reasons why we do things
need not be reasons for which we do them, nor are the la er always normative
reasons for acting (though it had be er be possible to act for a good reason).
C. S.
Dancy, J. (2000), Practical Reality, Oxford: Oxford University Press.
Parfit, D. (1984), Reasons and Persons, Oxford: Oxford University Press.
Sandis, C. (ed.) (2009), New Essays on the Explanation of Action, London: Palgrave
Macmillan.
reduction. In its most general terms, reduction is the translation of a theory,

including a single proposition in the limiting case, into another theory that is
ontically commi ed to fewer or fewer kinds of entities or properties, or that
requires fewer explanatory principles than the original theory. When such a
translation is correct, it is said to have been reduced either to the explanatorily
simpler or ontically more economical theory, or, where applicable, to both. One
measure of the eectiveness of a reduction is to compare the complexity of
explanations and the domain of existent objects belonging to both the original
theory and its proposed reduction, a ma er that can sometimes be decided by
308
Glossary
comparing explanations for relative diculty and counting the terms needed
in the vocabularies of target theories and their putative reductions. Where
philosophy of mind is concerned, mind-body reduction in particular is the
reduction of the mind or mental properties to exclusively non-mental, purely
physical behavioural, material, or functional (computational) properties.
Whether such a reduction can possibly succeed in the case of all truths about
consciousness, mental or psychological phenomena, the mind and its thoughts,
is one way of formulating the mind-body problem.
D. J.
Horst, S. (2007), Beyond Reduction: Philosophy of Mind and Post-Reductionist Philosophy
of Science, Oxford: Oxford University Press.
Searle, J. (1992), The Rediscovery of the Mind, Cambridge, MA: MIT Press.
Stich, S. (1983), From Folk Psychology to Cognitive Science: The Case Against Belief,
Cambridge, MA: MIT Press.
representation. Generally, the notion of representation entails one things

standing for something else or itself (self-representation). Representing is not
instantiating. Universals or objective properties (e.g. colours) are instantiated
by particulars (or in the environment). In general, the mind does not instantiate
the properties it represents. An objects shape, for example, is instantiated in the
environment and it is represented in ones perceptual experience. For a be er
illustration, we might contrast what we might call the representational with
the instantiation view of phenomenal character. According to some represen-
tationalists (e.g. Jackson, 1977) the phenomenal character of our sensory experi-
ences, that is, the apparent objects and properties of those experiences, are
merely representational, namely they comprise or contain the content of those
experiences without that content thereby being actually instantiated in the
mind. Contrariwise, according to some philosophers (Russell, 1998) when I
experience a red tomato for example, the content of that experience involves
the instantiation of an oval red object or the properties of this object in the mind.
On this view, the object or the property of looking red is not representational.
It is an intrinsic property of the mind.
The idea of representation has been central in discussions of intentionality
for many years. It is o en assumed that to have intentionality is to have content.
Mental content is otherwise described as representational/intentional or infor-
mational content, and intentionality is seen as the way of bearing or carrying
information. Now, if we say what the intentional content of a state of mind is
we thereby determine the conditions that must be met if this content is to be
satisfied (i.e. the conditions of its truth). Thus if I believe that Gordon Brown is
the British Prime Minister my belief has a certain content thereby describing
the conditions under which it is true. Hence representational states have cor-
rectness conditions partly determined by their contents. A belief for example is
309
Glossary
correct just in case its content is correct, and a proposition which gives that
content is correct just in case it is true.
D. P.
Crane, T. (1992), The Non-Conceptual Content of Experience, in id. (ed.) The
Contents of Experience, Cambridge: Cambridge University Press, pp. 13657.
Jackson, F. (1977), Perception: A Representative Theory, Cambridge: Cambridge
University Press.
Macpherson, F., and D. Platchias (forthcoming), Representationalism, Cambridge,
MA: MIT Press.
Tye, M. (1995), Ten Problems of Consciousness, Cambridge, MA: MIT Press.
self. The I who experiences, thinks, believes and desires. This conception of
the self as the subject of psychological states has long been argued to be respon-
sible for creating the Cartesian fiction of the self as a purely mental entity
which persists unchanged through time and in which psychological states
inhere. Hume famously rejected this view, arguing that we have no impression
of such a self but only of a sequence of constantly changing perceptions
(i.e. mental states) and hence no idea of the self except as a bundle of such
perceptions. Most current philosophers also reject, though o en for indepen-
dent reasons, the Cartesian view of the self as a mental entity, yet o en retain
Descartess assumption that the I of psychological self-a ribution refers to an
entity, namely to the human being which thinks, believes, desires and which
is located in space at the point of origin of our spatial experiences or at the
point of reference of our spatial thoughts of the form I am here or that is
over there.
I. M.
Descartes, R. (1912 [1637]), A Discourse on Method, Meditations and Principles, Toronto:
Dent.
Hume, D. (1978 [1740]), Treatise of Human Nature, ed. L. A. Selby-Bigge, Oxford:
Clarendon Press.
Valberg, J. (2007), Dream, Death and the Self, Princeton: Princeton University Press.
self-consciousness. Awareness of the self, or, more commonly in the philo-

sophy of mind, awareness of ones mental states, and in particular awareness of
them as ones own, with or without additional awareness of any substantive
self. Although sometimes used interchangeably with the term self-knowledge,
self-consciousness is commonly reserved to refer to less explicit, yet (on some
views) more fundamental forms of self-awareness. In the phenomenological
literature for instance, self-consciousness is thought of as something which
does not arise merely upon a entive reflection, but which is already present in,
and forms part of, world-directed conscious thought and experience itself.
310
Glossary
Whether any such form of self-consciousness can be shown to exist, what it

should be taken to amount to, and finally how (if at all and in what sense) it
might be argued to be a prerequisite for fully explicit, reflective self-knowledge
are complex issues touched upon from the philosophy of Kant to the phenom-
enological literature all the way to current debates in analytical philosophy
of mind.
I. M.
Bermudez, J. (1998), The Paradox of Self-Consciousness, Cambridge, MA: MIT Press.
Kant, I. (1929), Critique of Pure Reason, London: Macmillan Press.
Sartre, J. (1969), Being and Nothingness, London: Routledge.
self-knowledge. Typically used in the philosophy of mind to refer to the special

knowledge each person has of the contents of his or her own mind. We seem
to be able to know a wide range of our own thoughts, beliefs, desires and emo-
tions in a special immediate, authoritative way in which we are not able to
know the mental states of others. How is this possible? What is this special way
we have of knowing a certain class of our own mental states? The recent litera-
ture divides the theoretical options as follows: either (a) we know our own
minds inferentially (though perhaps particularly quickly) from observing our
own behaviour, or (b) we do so observationally through a form of inner sense,
or (c) we do so not in any way or on any particular epistemic basis, but rather in
virtue of the holding of some essentially constitutive link between our first-order
conscious states and our second-order self-ascriptive judgements. Which of
these options is the closest to the truth and whether they are in fact jointly
exhaustive of the theoretical options available remain ma ers of lively debate.
I. M.
Boghossian, P. (1998), Content and Self-Knowledge, Philosophical Topics, 17, 526.
Moran, R. (2001), Authority and Estrangement: An Essay on Self-Knowledge Princeton,
NJ: Princeton University Press.
semantics. Semantics is the study of meaning. As natural language symbols

have meaning, much of semantics falls within the domains of linguistics and
the philosophy of language. However, some semantic issues belong to the
philosophy of mind because mental states such as propositional a itudes have
semantic properties. They can refer to and represent particular things, classes
of things, and states of aairs and have truth or satisfaction conditions. For
example, my belief that aardvarks eat termites is about aardvarks, represents
them as being termite eaters, and is true if and only if aardvarks eat termites.
One prominent semantic issue in the philosophy of mind involves the
explanation of how mental states get the semantic properties that they have in
311
Glossary
naturalistic terms (i.e. in terms of lower level properties recognized by the

natural sciences). Another class of issues relates to the role that the mind has is
determining the meaning of words and sentences on the lips of an individual.
In this context, semantic externalists such as Putnam argue that linguistic mean-
ing isnt solely determined by the nature of the individuals mind considered in
isolation from the external world; rather, the nature of the extra-cranial world
at the physical level plays a key meaning determining role.
M. C.
Fodor, J. (1990), A Theory of Content and Essays, Cambridge, MA: MIT Press.
Putnam, H. (1975), The Meaning of Meaning, in id., Mind, Language and Reality:
Philosophical Papers, vol. 2, Cambridge: Cambridge University Press.
supervenience. Supervenience is the (ontic) dependence of the existence of a

token or type of an entity or property on the existence of a token or type of
entity or property. In standard definitions of supervenience originally owing to
Jaegwon Kim, in several analyses of distinct but related concepts of super-
venience, an entity or property can be said to supervene on itself, and on other
entities in which a two-way ontic dependence relation obtains. In his classic
paper, Concepts of Supervenience (Kim, 1993), and later investigations of
related topics, Kim distinguishes between weak and strong supervenience.
Kims definitions have gained widespread discussion, but not universal
adoption. Alternatively, there are also intuitive considerations that support
making supervenience into a one-way relation in which if a token or type of
entity or property X supervenes on a token or type of entity or property Y, then
Y does not also supervene on X. Such a provision also precludes anything
supervening on itself, since if X = Y and X supervenes on Y, it would otherwise
follow logically that Y supervenes on X. Defining supervenience as an asym-
metrical relation from the outset has the further advantage of making intuitive
sense of the term supervenience in which it is suggested that something, the
supervenient entity or property, is superior to, in some way above, the entity or
property on which it supervenes, also known as the supervenience foundation
or base. The instantiation of particular properties can be said in all of these
senses to supervene on the objects that possess the properties, since if the
objects did not exist then neither would the instantiations of their properties.
On Kims distinction, weak supervenience holds that there are no logically pos-
sible worlds within which there are Y-indiscernible but X-discernible properties
or entities. Strong supervenience in contrast implies that there are no Y-indis-
cernible but X-discernible properties within the same or dierent logically
possible worlds. Strong supervenience entails weak supervenience when the
logically possible worlds under consideration coincide, although weak super-
venience generally does not entail strong supervenience.
312
Glossary
An important application of the concept of supervenience is in trying to

understand the metaphysical relation between body and mind. The mind can
be said to supervene on the body, in the sense that mental states, properties, and
events, tokens and types, supervene on physical states, properties, and events,
tokens and types, just in case the mind and its properties would not exist with-
out the body and its properties, although the body could exist without the
mind. An interesting possibility is that in which a given human being functions
as a zombie, living and behaving verbally and in other ways exactly like a con-
scious person, while lacking any conscious mental states. We should expect
that if the mind supervenes on the body, then if the exact same conditions of
body are duplicated, the exact same mental states are also duplicated. If not,
then the failure is due to a lack of law-like regularity in the relations whereby
the existence of mind and mental occurrences depend on the existence of a
living body, especially a more or less normally functioning brain and neural
system, so that no positive type-type correlation occurs between the properties
of mind and body, even if a positive token-token correlation sometimes but
only irregularly and in that sense accidentally obtains. Since we expect that the
natural laws of the physical world governing the body on which consciousness
supervenes are regular if not simply or conditionally necessary, we can also
have confidence in the proposition that if X supervenes on Y, then if Ys pro-
perties as supervenient are duplicated, there will also most probably be a
corresponding duplication of Xs properties.
D. J.
Kim, J. (1993), Supervenience and Mind: Selected Philosophical Essays, Cambridge:
Rowlands, M. (1995), Supervenience and Materialism, Avebury: Ashgate Publishing.
Tooley, M. (1999), Laws of Nature, Causation, and Supervenience, London: Routledge.
teleofunctionalism. Teleofunctionalists explain mental state properties (e.g.

representational content) by focusing on the purpose these properties serve in
answering to the needs of complex systems or creatures. For example, certain
properties of an inner state (e.g. a natural sign or internal indicator) will be said
to have the teleofunction of representing Xs if those properties are supposed to
track or indicate the presence of Xs or if by responding appropriately to them
the organism would, in historically normal conditions, perform as it ought. This
emphasis on purpose that such properties or the responses to them fulfil diers
from standard functionalist theories that focus exclusively on the causal profile
or systemic role of mental states (i.e. those that are understood solely in terms
of actual or counterfactual relations to characteristic inputs, other mental states
and characteristic outputs, for example, being produced by specific kinds of
perceptual stimuli, generating specific kinds of behaviour, etc.). Teleofunctionalists
313
Glossary
lay stress on what a mental state is supposed to do as opposed to what it actu-

ally does or is disposed to do. By comparison, hearts can be said to have the
teleofunction to pump blood whether or not any particular or token heart is in
fact capable of doing so under any actual or nomologically possible conditions.
Despite introducing a normative dimension into their account of mental states,
teleofunctions are ultimately meant to be explained in wholly naturalistic terms,
by appeal to standards set, for example, by evolutionary processes, such as
natural selection, and individual learning and training.
D. H.
Dretske, F. (1988), Explaining Behaviour: Reasons in a World of Causes, Cambridge,
MA: MIT Press.
Millikan, R. (1984), Language, Thought and Other Biological Categories, Cambridge,
MA: MIT Press.
Third-person perspective. See first-person/third-person perspective.
threshold value. See neural network.
transparency. In the literature on self-knowledge, transparency typically refers

to a datum, pointed out by (among others) Evans drawing on a remark by
Wi genstein, that when asked about what we believe, we tend to turn our a en-
tion not to ourselves but to the world, and to consider evidence not explicitly
about our beliefs but about how the world is. In order to answer the question
of whether I believe that it is raining for example I will look not at myself but
out the window, and consider not what mental states I am in but whether it is
raining. How this transparency is best to be explained is a ma er of ongoing
debate. The term transparency is frequently used however also in another
context, namely in the philosophy of perception, where it refers not to a view
about the evidence appealed to in introspection but about what introspection
reveals about our perceptual experiences. According to this transparency view,
introspection reveals only the objects of our experiences out in the world, not
any additional phenomenal properties of our perceptual states themselves.
I. M.
Evans, G. (1982), The Varieties of Reference, Oxford: Clarendon Press.
Martin, M. (2002), The Transparency of Experience, Mind and Language, 17, 376425.
Turing machine. A simple and abstract computing device invented by the

British logician Alan Turing. A Turing machine consists of an infinitely long
tape divided into squares (each of which can either have a 1 or a 0 wri en on
it or be blank), and a read-write head which scans the squares one at a time.
Whenever it scans a square the head will be in one of a finite number of possible
314
Glossary
states. How the machine responds to what it scans will depend upon its state.
Its response will have several elements involving: (1) either leaving the square
unchanged or writing a new symbol on it (i.e. a 1, 0 or a blank); (2) moving
one square to the le or one square to the right or halting; and (3) moving into
some other state or remaining in the same state. The machines response
to any possible symbol for each of the states that it can be in is specified by a
machine table.
Turing proved that for any computable mathematical function there is a
Turing machine that can compute it. A universal Turing machine can be pro-
grammed (by means of strings of symbols printed on its tape) to mimic any
possible Turing machine and so is capable of computing any computable math-
ematical function. Putnam developed an early version of functionalism that
compares mental states with Turing Machine states as the la er are defined by
the machine table in terms of their relations to inputs, outputs and other states
rather than their material constitution.
M. C.
Hodges, A. (1983), Turing: The Enigma, New York: Simon and Schuster.
Putnam, H. (1975), Mind, Language and Reality: Philosophical Papers, vol. 2, Cambridge:
Turing test. A test proposed by Turing as a precise alternative for the meaning-
less question of whether a machine is capable of thought. The test is based upon
the imitation game where the tester presents both a machine and a human
with a series of questions via a teletypewriter. On the basis of the answers
received, the tester tries to work out which respondent is the human and which
is the machine. The machine passes the test if the tester fails to identity it as the
machine.
M. C.
Millikan, P., and A. Clark (1996), Machines and Thought: The Legacy of Alan Turing,
vol. 1, Oxford: Clarendon Press.
Turing, A. (1950), Computing Machinery and Intelligence, Mind, 59, 43360.
Twin Earth. Twin Earth is a fictional planet described in a thought experiment

that originally appeared in Hilary Putnams paper The Meaning of Mean-
ing. Twin Earth is superficially identical to Earth. There is someone there who
looks like you your twin reading a book that is identical in appearance
to the Continuum Companion to the Mind. There is, however, a dierence between
Earth and Twin Earth. The glass of water on your desk, and all the water on
our planet, contains the chemical H20. The glass of liquid on your twins table,
and all such liquid on his planet, contains XYZ. H20 and XYZ appear the same:
they are colourless, tasteless, odourless, thirst-quenching liquids, that rain from
315
Glossary
the sky and flow down rivers into the sea. They are, though, dierent liquids
since they are made of dierent stu. Philosophers of mind discussing this
thought experiment o en refer to the twin on Earth as Oscar, and the one on
Twin Earth as Toscar, and the liquids on their planets as water and twater
respectively.
Putnam took his thought experiment to show that the meanings of natural
kind terms such as water could not be wholly determined by items that are
literally inside the head of a thinker, items such as mental images, or brain or
computational states. Everything inside the heads of Oscar and Toscar is the
same when they are talking about what they both call water, yet Oscars word
refers to water and Toscars to twater. Their words therefore have dierent
meanings even though everything in their heads, and their behaviour, is the
same. As Putnam puts it, Meaning just aint in the head.
Others, including John McDowell and Greg McCulloch, take Putnams
thought experiment to entail a stronger conclusion, one concerning not just
the meanings of words, but the contents of thoughts. Oscar has thoughts with
the content water; Toscar has thoughts with the content twater. Even though
everything in their heads is the same, their thoughts are dierent. The content
of thought is partly determined by our relation to the world: the mind aint in
the head (McCulloch, The Life of the Mind, p. 41). This is cognitive externalism.
Putnam himself has now adopted this position.
D. O.
Pessin, A., and S. Goldberg (1996), The Twin Earth Chronicles: Twenty Years of
Reflection on Hilary Putnams The Meaning of Meaning, New York: Armonk.
what its like. Nagel famously wrote that the fact that an organism has con-
scious experience at all means, basically, that there is something it is like to be
that organism . . . fundamentally an organism has conscious states if and only
if there is something it is like to be that organism something it is like for the
organism the facts of experience [are] facts about what it is like for the experi-
encing organism. (Nagel, 1974, pp. 435, 439, emphasis in the original). This
quote points directly at what-it-is-likeness, the salient but dicult to describe
feature of a conscious state. If theres something it is like for one to be in a men-
tal state then the state is experiential. If there is nothing its like for one to be
in that state, its not. However, to say that there is something its like for one to
be in an experiential state is not merely to mean that there is something that
an experience is like. That there is something that an experience is like is a
mere truism in that it is plain that there is nothing such that it is not like
something. We can say for instance, that there is something that a rock or a
table is like. What-it-is-likeness in the Nagelian sense concerns the individual.
If there is something it is like for the individual to be in a particular mental state
316
Glossary
then that state is experiential. What its like to be in an experiential state is in

the relevant sense what its like for one to be in that state.
D. P.
Jackson, F. (1982), Epiphenomenal qualia, Philosophical Quarterly, 32, 12736.
Nagel, T. (1974), What is It Like to be a Bat?, Philosophical Review, 83 (4), 43550.
Tye, M. (1995), What What Its Like is Really Like, in Ten Problems of Consciousness,
Cambridge, MA: MIT Press, 13355.
will. The faculty of the will grounds our abilities to do things intentionally,
voluntarily, and at will, where this typically (but perhaps not necessarily)
involves acting for reasons. Contrary to an influential British tradition dating at
least as far back as Hobbes, to act voluntarily is not to perform an act of volition
but rather to act (intentionally or otherwise) according to ones own will (or
desire) and not under coercion or duress. This need not involve performing an
act of will, the la er requiring a high degree of motivational strength, courage
and/or eort. A persons will is said to be weak if they are too easily prone
to change their mind under the influence of others. In philosophy, however,
weakness of will has become a technical term for the phenomenon of acting
against ones be er judgement, commonly also referred to in the literature as
akrasia or (even more misleadingly) incontinence.
C. S.
Hacker, P. M. S. (2000), Willing and the Nature of Voluntary Action, in id.,
Wi genstein Mind and Will Part I: Essays, Oxford: Blackwell.
O Shaughnessy, B. (1980), The Will A Dual Aspect Theory, vols. 1 and 2, Oxford:
Oxford University Press.
XYZ See H20/XYZ.
zombies. The philosophical zombie is a living organism which is functionally

and behaviourally indistinguishable from a conscious one (with which they
may share the same environment and have same causal histories), but there is
nothing it is like to be that creature (i.e. the creature is non-conscious). Accord-
ing to some philosophers, it is not only conceivable that this creature can exist
but it can also possibly exist. The explanatory gap argument revolves around
this issue that is, whether such creatures are conceivable and further whether
they are possible. Let P be the proposition that everything physical is as it
actually is and Q the proposition that there are phenomenal or experiential
properties. According to most philosophers, it is conceivable that (P and Q).
Since it is conceivable that (P and Q) then one can ask, why, given that P is the
case, is Q the case? Hence the explanatory gap, namely there is no entailment
from P to Q; one cannot deduce Q from P. According to some philosophers
317
Glossary
(Chalmers, 1996; Jackson, 2001), from this epistemic gap one can infer an onto-
logical gap: if we cannot deduce Q from P then we cannot explain phenomenal
consciousness in terms of physical processes, and if we cannot explain it in
terms of physical processes then phenomenal consciousness is not a physical
process.
D. P.
Brueckner, A. (2001), Chalmers Conceivability Argument for Dualism, Analysis,
61, 18793.
Chalmers, D. (1996), Can Consciousness Be Reductively Explained? in The
Conscious Mind, Oxford: Oxford University Press.
Nagel, T. (1998), Conceiving the Impossible and the Mind-Body Problem,
Philosophy, 73, 33752.
Shoemaker, S. (1999), On David Chalmerss The Conscious Mind, Philosophy and
Phenomenological Research, 59 (2), 43944.
318
Chronology
This is necessarily idiosyncratic, but the hope is to provide a useful desk refer-
ence for philosophers of mind who require speedy access to a date or title as
well as a bit of context for it. There are a few lines of explanation for almost
every entry. I tried to encapsulate in a single sentence as much of what ma ers
as possible. Obviously this almost never works, but a narrative structure really
does help one get a grip on the major events in the history of our thinking about
the mind. The narrative ends in 1949 with Ryles Concept of Mind. My excuse for
stopping there is that its dicult to say what the impact or meaning of more
recent books and papers might be, because we are too close in time to know.
Forgive omissions; it is very hard to tell which books and events should be
included early on or during unfamiliar centuries, and it gets extremely dicult
as we approach the present.
800 BCE
Homeric poems taking shape between the late ninth and early eighth century; they
characterize the soul thinly, as something lost at death, something which then
howls o to Hades.
600 BCE
Thales (fl. 600) might view psyche as a mover, force, or impetus, something which
initiates the movement of moving things, from animals and people to magnets.
Anaximenes (c. 585c. 528) possibly believes that psyche holds a living thing
together and rules or controls it.
Pythagoras (fl. 530) accepts metempsychosis; possibly first to locate the soul in
the head.
500 BCE
Anaxagoras (c. 500c. 428) seems to argue for a materialist world actuated by a
cosmic intelligence, mind or nous.
Heraclitus (fl. 500) might believe that psyche is fire, somehow responsible for the
changes a ending waking, sleeping and death.
Parmenides (early to mid-fi h century) distinguishes between false appearances
and reality as revealed by reason; might flirt with idealism.
400 BCE
Empedocles (c. 495c. 435) probably formulates the first theory of perception; his
talk of the cosmic psychological principles, love and strife, suggest panpsychism
to some.
319
Chronology
Socrates (c. 469c. 399), the man not the mouthpiece; might conceive of soul as the
bearer of moral qualities.
Democritus (c. 460c. 370) elaborates the atomism of Leucippus, including
materialist conceptions of perception and the soul; might be first to tie soul to
intelligence.
Plato (c. 427c. 347) distinguishes soul from body, argues for immortality of the soul,
ties soul to reason, Phaedo; divides the soul into three parts: reason, spirit and
appetite, Republic).
Aristotle (c. 384c. 322) oers an extended, systematic discussion of psychological
phenomena, De Anima and Parva Naturalia, (c. 350); soul characterized as the
form of a living thing.
Epicurus (c. 341c. 271) argues for a radical materialism and for the impossibility
of the soul surviving death.
300 BCE
Zeno of Citium (c. 335c. 263) founds Stoic School, active until c. 520, which
perpetuates a variety of materialist notions of soul, typically conceived as a
breath-like substance diused throughout the body.
200 BCE
The Septuagint produced between the third and first centuries BCE; conceptions
of the soul and mental phenomena as depicted in the Hebrew Bible translated
into Greek.
100 BCE
Lucretius (c. 98c. 51) propounds and expands the philosophy of Epicurus, pro-
ducing the first philosophical treatment of mind in Latin, De Rerum Natura.
BCE/CE
Philo (c. 20 BCEc. 50 CE) blends Greek philosophy and Hebrew thought about
the soul.
100
The Church Fathers (end of the first century to as late as 749 CE) subordinate
philosophical accounts of mind to scriptural ones, raise religious questions, and
shape the intellectual agenda accordingly.
Tertullian (c. 160c. 225) advocates traducianism; argues that soul must be somehow
corporeal if it can be tormented in hell, On the Soul.
200
Origen (c. 185c. 254) holds that souls were created by God for contemplation
but, falling away in distraction, became enveloped in bodies, On First Principles.
Plotinus (204/5270/1) founds Neoplatonism, articulates a conception of soul as part
divine and part entwined with body, as well as an intricate theory of perception,
The Six Enneads.
320
Chronology
300
Augustine (354430) oers a detailed description of and reflection on introspected
mental life, Confessions; has thoughts on action theory, On Free Will; might argue
by analogy for other minds, anticipate the cogito, and influence the Cartesian
conception of mind, On the Trinity and City of God.
400
Boethius (c. 480c. 524) translates Aristotle and Plato into Latin; emphasizes rational
nature of the soul, Contra Eutychen.
900
Avicenna (c. 9801037) integrates Islamic philosophy and Greek thought about
mind and soul, formulates floating man thought experiment, On the Soul.
1100
Averroes (11261198) Latin translations of his commentaries on Aristotle bring
Greek views on mind, through Islamic lenses, back to the West; also develops
his own complex psychology and metaphysics of the soul, Long Commentary on De
Anima.
Vespasian Homilies (c. 1150) contain possibly the first use in English of a variation
on the word soul (sawle), meaning life or life force.
1200
William of Moerbeke (c. 12151286) undertakes a complete translation of Aristotle
into Latin (c. 1250).
Aquinas (c. 12241274) reinterprets Aristotle in the light of Christian teaching,
articulates full-blooded conceptions of mind, soul, intellect, memory, appetite,
self-knowledge, imagination, perception, etc., Summa Theologiae.
1300
William of Shoreham (fl. 1330) writes religious poems containing a forerunner of
mind (mende), which might be the first use in English tied to cognition.
1400
Marsilio Ficiono (14331499) is the first to translate all of Plato into Latin, and the
Platonic conceptions of soul and mind are rekindled, Theologia Platonica de
Immortalitate Animae, (1474).
Pomponazzi (14621525) might anticipate property dualism, On the Immortality
of the Soul.
1500
Shakespeare (15641616) writes Hamlet (c. 1600); some detect Cartesian pre-
suppositions in certain soliloquies; others hear Hamlets repressed desires.
321
Chronology
1600
Hobbes (15881679) defends a causal, empiricist, mechanistic and materialist
conception of mental phenomena, Leviathan (1651).
Descartes (15961650) articulates Cartesian dualism; disentangles new thoughts on
mind from Aristotelian, Platonic and Scholastic thinking; thereby ushers in modern
philosophical reflection on the mental, Meditations on First Philosophy (1641).
Geulincx (16241669) (and Graud de Cordemoy [162684]) follows Descartes;
argues for pre-established harmony before Leibniz, Opera Philosophica (c. 1668).
Spinoza (16321677) rejects Cartesian dualism in favour of dual-aspect monism:
there is one substance, God, Ethics (1677).
Locke (16321704) formulates modern conception of self; raises questions about
personal identity; claims that experience is the source of ideas; sets out limits to
understanding, Essay Concerning Human Understanding (1690).
Malebranche (16381715) largely follows Descartes, but argues for occasionalism,
Dialogues on Metaphysics and on Religion (1688).
1700
Leibniz (16461716) argues for pre-established harmony, Discourse on Metaphysics
(1686).
Berkeley (16851753) writes an account of perception, Essay Towards A New Theory
of Vision (1709) and argues that to be is to be perceived, Principles of Human
Knowledge (1710).
Hartley (17051747) founds associationist school of psychology, Observations on
Man, His Frame, His Duty, and His Expectations (1749).
Reid (17101796) brings common sense to an account of sensation, conception, and
perception; uses memory to inform a notion of self, An Inquiry into the Human
Mind on the Principles of Common Sense (1764), Essays on the Intellectual Powers
of Man (1785).
Hume (17111776) brings the experimental method to bear on mind, follows the
sceptical implications of empiricism through, propounds the bundle theory of self,
Treatise of Human Nature (1739), Enquiry Concerning Human Understanding (1748).
Adam Smith (17231790) considers the nature of sympathy, The Theory of Moral
Sentiments (1759).
Kant (17241804) argues that the structuring activity of the mind makes possible
a world of experience; gives an account of reason, perception, judgement, the
understanding, imagination, etc., a Copernican Revolution in the conception
of mind, Critique of Pure Reason (1781), Critique of Practical Reason (1788), Critique
of Judgement (1790).
Bentham (17481832) articulates modern psychological hedonism, Introduction to
the Principles of Morals and Legislation (1789).
1800
Hegel (17701831) gives an account of the evolution of consciousness as it plays
out in human history, The Phenomenology of Spirit (1807).
Schopenhauer (17881860) sees blind craving, will, at the depressing centre of
human action; our inner experience of it points to the hidden nature of all things,
The World as Will and Representation (1819).
322
Chronology
J. S. Mill (18061873) elaborates on the connection between right and wrong and
pleasure and pain; connects social and political reform to psychology, A System
of Logic (1843).
Kierkegaard (18131855) claims that subjectivity is truth, Concluding Unscientific
Postscript to Philosophical Fragments (1846).
T. H. Huxley (18251895) memorably couches a version of epiphenomenalism
in terms of whistles and steam engines, On the hypothesis that animals are
automata, and its history (1874).
Wundt (18321920) investigates the self-examination of experience, Principles
of Physiological Psychology (1873/4), establishes a laboratory of experimental
psychology in 1879.
Brentano (18381917) reintroduces the Scholastic conception of intentionality as the
mark of the mental, and his elevation of introspection paves the way for the
phenomenological movement, Psychology from an Empirical Standpoint (1874).
Peirce (18391914) raises objections to Cartesian methods and suggests panpsy-
chism, along with further thoughts on signs and representation, The Fixation of
Belief (1877), The Monist series (18911893).
James (18421910) largely sets the agenda for both the philosophy of mind and
psychology by advancing influential accounts of the brain, the mind-body
relation, the stream of consciousness, memory, sensation, imagination, will, and
emotions all peppered with compelling introspective reports, The Principles of
Psychology (1890).
Nietzsche (18441900) calls the subject a grammatical fiction, On the Genealogy
of Morals (1887).
Bradley (18461924) leads the turn towards idealism in the English-speaking
world, rejects empiricist psychology, The Principles of Logic (1883), Appearance and
Reality (1893).
Husserl (18591938) rejects psychologism and formulates the phenomenological
method, Logical Investigations (1900/1); the method of epoch and transcendental
phenomenology itself appear, Ideas (1913).
Bergson (18591941) oers an alternative to phenomenology, finds multiplicity
in consciousness, regards intuition as method, Time and Free Will (1889), Maer
and Memory (1896).
1900
Freud (18561939), father of psychoanalysis, formulates such concepts as
repression, psychosexual motivation, unconscious desire, as well as the id, ego
and super ego, The Interpretation of Dreams (1900), The Ego and the Id (1923),
Introductory Lectures on Psychoanalysis.
Dewey (18591952) brings pragmatism to bear on mind, rejects dualisms in favour
of naturalism and evolution; mind emerges socially; founds the functional
approach to psychology, The Reflex Arc Concept in Psychology (1896), Experi-
ence and Nature (1925).
Whitehead (18611947) rejects materialism for the view that nature is a structure of
evolving processes, Process and Reality (1929).
Russell (18721970) champions analytic method, moves from reflection on sense
data to neutral monism, rejects idealism and psychologism, Knowledge by
323
Chronology
Acquaintance and Knowledge by Description, (1910), The Analysis of Maer

(1927), The Analysis of Mind (1929).
Moore (18731958) brings commonsense realism to metaphysics and epistemology,
Refutation of Idealism (1903), Proof of an External World (1939).
Watson (18781958) gives the boot to consciousness in general and introspection
in particular, Psychology as a Behaviorist Views It (1913).
Broad (18871971) argues for emergent vitalism, considers the possibility of
survival a er death, The Mind and Its Place in Nature (1925).
Wi genstein (18891951) early picture theory of meaning gives way to therapeutic
treatments of problems associated with mental phenomena; private language
argument makes trouble for Cartesian reflection and solipsism, Tractatus Logico-
Philosophicus (1922), Philosophical Investigations (1953), The Blue and Brown Books
(1958), On Certainty (1969).
Heidegger (18891976) urges reflection on Dasein, instead of a misunderstood
conception of Being, reorients numerous mental concepts, Being and Time (1927).
Carnap (18911970) ties meaning to phenomenalistic language, argues that
metaphysics is meaningless, The Logical Structure of the World (1928), Pseudoprob-
lems in Philosophy (1928).
Price (18991984) reflects on perceptual consciousness, sense data, and the role
of concepts in thought, Perception (1932), Thinking and Experience (1953).
Ryle (19001976) ushers in contemporary philosophy of mind, arguing against
Descartes ghost in the machine and for logical behaviourism, The Concept of
Mind (1949).
Feigl (19021988) The Mental and the Physical (1958, as a book with
Postscript and Preface, 1967).
Sartre (19051980) the father of existentialism distinguishes between being-in-
itself and being-for-itself; were both, The Psychology of Imagination (1940), Being
and Nothingness (1943), Critique of Dialectical Reason (1960) Sketch for a Theory of
the Emotions.
Merleau-Ponty (19081961) perception takes centre stage, phenomenology meets
scientific psychology, The Structure of Behavior (1942), Phenomenology of Perception
(1945), The Visible and the Invisible, (1964).
Quine (19082000) Two Dogmas of Empiricism (1951), Quantifiers and Proposi-
tional A itudes, (1956), Epistemology Naturalized, (1969), Word and Object (1960).
Ayer (19101989) applies the verification principle to claims about the mind,
oers an analysis of sense data, Language, Truth and Logic (1936), The Problem of
Knowledge (1956), The Concept of a Person and Other Essays (1963).
Austin (19111960) Sense and Sensibilia (1959).
Malcolm (19111990) Our Knowledge of Other Minds (1958).
Turing (19121954) Computing Machinery and Intelligence (1950).
Sellars (19121989) Empiricism and the Philosophy of Mind (1956), Science and
Metaphysics (1968), Meaning as Functional Classification (1974).
Chisholm (19161999) Perceiving (1957), Person and Object (1976), The First Person
(1981), Brentano and Intrinsic Value (1986).
Geach (1916) Mental Acts (1957).
Davidson (19172003) Actions, Reasons and Causes (1963), Mental Events (1970),
Essays on Actions and Events (1980).
324
Chronology
Anscombe (19192001) Intention (1957), The First Person (1975), Metaphysics and
the Philosophy of Mind (1981).
Strawson (19192006) Individuals (1959), Freedom and Resentment and Other Essays
(1974) Scepticism and Naturalism (1985).
Smart (1920) Sensations and Brain Processes (1959), Philosophy and Scientific Realism
(1963).
Place (19242000) Is Consciousness a Brian Process? (1956).
OShaughnessy (19252010) The Will (1980), Consciousness and the World (2000).
Armstrong (1926) Perception and the Physical World (1961), Bodily Sensations (1962),
A Materialist Theory of the Mind (1968), The Nature of Mind and Other Essays (1980),
Consciousness and Causality (1984), The Mind-Body Problem (1999).
Putnam (1926) The Nature of Mental States (1967), Mind, Language and Reality
(1975) The Meaning of Meaning (1975).
Chomsky (1928) Syntactic Structures (1957), Aspects of the Theory of Syntax
(1965).
Williams (19292003) Problems of the Self (1973).
Rorty (19312007) Philosophy and the Mirror of Nature, (1979).
Shoemaker (1931) Self-Knowledge and Self-Identity (1963), Identity, Cause and Mind:
Philosophical Essays (1984), The First-Person Perspective, and other Essays (1996).
Dretske (1932) Seeing and Knowing (1969), Knowledge and the Flow of Information
(1981), Naturalising the Mind (1995), Perception, Knowledge and Belief (2000).
Searle (1932) Minds, Brains and Programs (1980), Intentionality (1983), Minds, Brains
and Science (1984), The Rediscovery of the Mind (1992).
Fodor (1935) The Language of Thought (1975), Propositional Aitudes (1978), Representa-
tions (1979), The Modularity of Mind (1983), Psychosemantics (1987), A Theory of
Content (1990).
Nagel (1937) What is it Like to be a Bat? (1974), Mortal Questions (1979), View from
Nowhere (1986).
Honderich (1933) A Theory of Determinism (1988), On Consciousness (2004).
Millikan (1933) Language, Thought, and Other Biological Categories (1984), White Queen
Psychology and Other Essays for Alice (1993).
Kim (1934) Supervenience and Mind (1993).
Rosenthal (1939) Two Concepts of Consciousness (1986), Consciousness and Mind
(2005), Consciousness and Its Function (2008).
Kripke (1940) Naming and Necessity (1972), Wigenstein on Rules and Private Language
(1982).
Lewis (19412001) An Argument for Identity Theory (1966), Psychophysical
and Theoretical Identifications (1972), Mad Pain and Martian Pain (1980),
Philosophical Papers, Volume II (1986).
McDowell (1941) Mind and World (1994), Mind, Value and Reality (1998).
Jackson (1943) Perception (1977), Epiphenomenal Qualia (1982), What Mary
Didnt Know (1986).
Block (1942) Psychologism and Behaviorism (1981), On a Confusion about the
Function of Consciousness (1995), Consciousness, Function and Representation (2007).
Denne (1942) Brainstorms (1981), Content and Consciousness (1986), The Intentional
Stance (1989), Consciousness Explained (1992), Kinds of Minds (1996), Brainchildren
(1998), Sweet Dreams (2005), Neuroscience and Philosophy (2007).
325
Chronology
Parfit (1942) Reasons and Persons (1984).

Paul Churchland (1942) Scientific Realism and the Plasticity of Mind (1979), Maer and
Consciousness (1988), A Neurocomputational Perspective (1989).
Jackson (1943) Epiphenomenal Qualia (1982), What Mary Did Not Know (1986).
Patricia Churchland (1943) Neurophilosophy (1986), The Computational Brain (1992),
Brain-Wise (2002).
Burge (1946) Individualism and the Mental (1979), Foundations of Mind (2007).
McGinn (1950) The Character of Mind (1982), Mental Content (1989), The Problem of
Consciousness (1991), The Mysterious Flame (1999), Consciousness and Its Objects
(2004).
Papineau (1947) Philosophical Naturalism (1993), Thinking About Consciousness (2002).
Peacocke (1950) Sense and Content (1983), A Study of Concepts (1992), Truly Understood
(2008).
Tye (1950) Ten Problems of Consciousness (1995), Consciousness, Color and Content
(2000), Consciousness and Persons (2003).
McCulloch (19512001) The Mind and Its World (1995), The Life of the Mind (2003).
Clark (1957) Being There (1997), Supersizing the Mind (2008).
Chalmers (1966) The Conscious Mind (1996).
326
Research Resources
The following is a list of journals, websites and centres devoted to subjects

of interest to philosophers of mind. Many of the journals listed below have
companion websites with at least some free content. This list is, of course, not
exhaustive, but it does contain some useful starting points. The awe-inspiring
Mind Papers (hp://consc.net/mindpapers), compiled by David Chalmers
(Editor) and David Bourget (Assistant Editor), is probably the most useful
starting point of all.
Journals
AI & Society
AI Communications
Annals of Mathematics and Artificial Intelligence
Artificial Intelligence
Artificial Intelligence and Law
Artificial Intelligence Review
Behavior and Philosophy
Behavioural and Brain Sciences
Brain and Cognition
Brain and Language
Brain and Mind
Cognition and Emotion
Cognitive Linguistics
Cognitive Psychology
Computational Intelligence
Consciousness and Cognition
Consciousness and Emotion
Cybernetics and Human Knowing
International Journal of Approximate Reasoning
Journal of Artificial Intelligence Research
Journal of Cognitive Systems Research
Journal of Consciousness Studies
Journal of Culture and Cognition
Journal of Experimental and Theoretical Artificial Intelligence
Journal of Intelligent Systems
327
Research Resources
Journal of Mind and Behaviour

Journal of Theoretical and Philosophical Psychology
Mind and Language
Minds and Machines
Neuroethics
Phenomenology and the Cognitive Sciences
Philosophy, Psychiatry, & Psychology
Psyche Theory and Psychology
Thinking and Reasoning
Trends in Cognitive Science
Websites
Cogprints
hp://cogprints.org/view/subjects/phil-mind.html
Consciousness and the Brain: Annotated Bibliography
www.consciousness-brain.org/
Dictionary of the Philosophy of Mind
hp://philosophy.uwaterloo.ca/MindDict/
Episteme Links
www.epistemelinks.com/
A Field Guide to the Philosophy of Mind
hp://host.uniroma3.it/progei/kant/field/
Internet Encyclopedia of Philosophy
www.iep.utm.edu/
KLI Theory Lab
www.kli.ac.at/theorylab/index.html
Mind Papers
hp://consc.net/mindpapers
Stanford Encyclopedia of Philosophy
hp://plato.stanford.edu/
The Turing Archive for the History of Computing
www.alanturing.net/
Centres and Societies
Association for the Advancement of Artificial Intelligence

www.aaai.org/home.html
Association for the Scientific Study of Consciousness
www.theassc.org/
328
Research Resources
Centre for Cognition and Culture

www.case.edu/artsci/cogs/CenterforCognitionandCulture.html
Centre for Consciousness
hp://consciousness.anu.edu.au/
Center for Consciousness Studies
www.consciousness.arizona.edu/
Centre for Research in Cognitive Science
www.sussex.ac.uk/cogs/index.php
Centre for Research into Embodied Subjectivity
www2.hull.ac.uk/fass/humanities/philosophy/research/centre-for-research-
into-embod.aspx
Centre for Research on Concepts and Cognition
www.cogsci.indiana.edu/
Centre for the Study of Perceptual Experience
www.gla.ac.uk/Acad/Philosophy/cspe/
Cognitive Science Society
hp://cognitivesciencesociety.org/index.html
Consciousness and Self-Consciousness Research Centre
www2.warwick.ac.uk/fac/soc/philosophy/research/conandselfcon/
European Society for Philosophy and Psychology
www.eurospp.org/
International Association for Computing and Philosophy
www.ia-cap.org/
Mind, Meaning and Rationality Research Group
www.open.ac.uk/Arts/philosophy/mmr/index.shtml
Society for Philosophy and Psychology
www.socphilpsych.org/
Southern Society for Philosophy and Psychology
hp://southernsociety.org/
329
Notes
Chapter 1
1 One problem with Armstrongs definition of substance is that objects like the Sun
cannot exist alone for they are essentially spatio-temporal objects and so depend for
their existence on the existence of space and time. I will not a empt to resolve this
issue here.
2 In order to avoid concerns raised by quantum indeterminacy, physical closure
is sometimes expressed by saying that every physical event has a physical cause
sucient to determine its objective probability. See Yablo (1992); for a helpful discussion
see Section 2.3 of Robb et al. (2008).
3 The example is not to be taken too seriously because the neurobiology of pain is
very complex, and cannot be captured by the slogan pain = c-fibre firing. I will use
c-fibre firing as a convenient label for the complex neurobiological process which
are involved in the human pain response.
4 We can distinguish between the property of having a property which occupies the
causal role characteristic of mental state M, and the property which occupies
the causal role characteristic of mental state M. Functionalists dier as to which of
these properties is to be identified with M.
5 Strictly speaking, Davidson only denied the possibility of strict laws linking pro-
positional a itudes with physical events. I introduce propositional a itudes in the
Mental Representation section.
6 For alternative eliminativist strategies, see Stich (1983).
7 Notice the similarity between this proposal and Denne s intentional stance (see
the Eliminativism, Instrumentalism and the Intentional Stance subsection).
8 Traditionally, the key external relations were held to be those of similarity or resem-
blance, but this idea is fraught with diculties (for a quick overview see Ravenscro ,
2005, pp. 1267).
9 Putnam originally used his example to develop a point about the reference of natural
kind terms in ordinary human languages. However, his example has been widely
used to discuss parallel considerations in the philosophy of mind.
10 What Chalmers calls the hard problem is the problem of accounting for phenomenal
consciousness, and what he calls the easy problem includes the problem of accounting
for access consciousness (see Chalmers, 2003a, p. 103).
11 Strictly speaking, qualia is the plural; the singular is quale. However, in line with
most contemporary usage, I will use qualia as both the plural and the singular.
Chapter 2
1 Philosophers in the phenomenological tradition engage in the descriptive project of

revealing the structures of experience in detail. For a useful introduction to their
work see Gallagher and Zahavi (2007).
330
Notes
2 Several prominent critics question the phenomenal concepts strategy on the grounds
that it is questionable that concepts of the requisite first-personal sort might exist
(see Prinz 2007; Tye 2009). Nevertheless, it is arguable that ordinary, public concepts
of experience could do the same work that phenomenal concepts are meant to do
equally well (or badly).
3 For a catalogue of other problems and worries concerning higher order theories
and their relatives see Block (forthcoming).
Chapter 3
1 We hasten to add that we think it is in principle possible to build machines that

could think out of non-biological materials.
2 See Adams and Aizawa (2008) for the pros and cons.
3 In a longer version of this paper (available from the authors) we discuss the views of
the eliminativists and the mysterians about the mental.
4 Antony (1997), Block (1995, 1996, 1998) McGinn (1982) and Searle (1983) are other
well-known critics of the view that sensory or qualitative states are representational
(intentional) states.
5 We take this to mean that Rorty does not take non-occurrent states to be mental even
though others may call them so. Thus, we are taking Rorty to hold the single property
view, not the property cluster view or single system view of the mark of the mental,
though he perhaps could be interpreted as taking the single system view.
6 Rorty (1970a) considered other possible properties to distinguish the mental from
the physical, such as intentionality, introspectability, purposiveness, non-spatiality
and privacy, but rejected all of these. We wont go through his reasons here.
7 Surprisingly Rorty (1970a) says, if science had only beliefs and desires to deal with
(not reports of mental events), science would never have invented the concept of
mind or the mental. (Rorty, 1970a, p. 408)
8 Even Rorty (1972, p. 217) says that robots would require language use similar to that
of humans for their reports to be incorrigible. We take this to require, at the very
least, communicative intentions.
9 Rorty (1970a, p. 405) says: It seems clear that it is the notion held by Cartesian
philosophers that we must explicate if we are to make sense of materialism. This
la er notion must contain properties incompatible with properties of physical
entities. He also says: (Rorty, 1970a, p. 402) It is part of the sense of mental that
being mental is incompatible with being physical, and no explication of this sense
which denies this incompatibility can be satisfactory and again (p. 405) Material
and physical would be vacuous notions without the contrast with mental.
Immaterial and nonphysical are notions that have sense only if the mental is
given as an instance of them.
10 Crane (1998) argues for what he calls the weaker view that all mental states, even
qualitative states, have intentionality even if qualia themselves are not intentional.
Whereas, we read Tye and Dretske as a empting to explain the qualitative character
of mental states in virtue of their being representational states, and thereby inten-
tional. Crane also at least flirts with the view that an intentionalist about sensations,
say pain, may hold that the intentional object presented in a pain sensation is an
internal mental object. Where Tye and Dretske would say the thing represented in a
pain sensation (say, foot pain) is the damage to the foot, not an object in the mind.
Much of Cranes view (1998 is devoted to responding to Searles view that emotions
331
Notes
are non-intentional. Crane gives good reasons to think that this is not true, but we
cannot go into the details here.
11 See Adams and Dietrich (2004) for an account that emphasizes the dierences.
12 Of course there can be hallucinations or artificially caused experiences, but the quali-
tative character of the experience would derive from past representational episodes.
13 All of these are examples of natural signs. Natural signs (such as smoke being a sign
of fire; footprints being a sign of a passerby) have a kind of informational aboutness.
These seem to be the wrong kind of intentionality, at least in part because they need
not produce a phenomenology, cannot be falsely tokened, and have not risen to the
level of semantic meaning. Smoke naturally indicates fire, but smoke means smoke
(and its tokening need not indicate fire).
14 Famously, Descartes believed that non-human animals not only could not think
(were not intentional systems), but could not feel because they were not intentional
systems. Descartes believed that to feel pain, for instance, one must be able to think
the thought Im in pain. So, only intentional systems were able truly to have
phenomenological states. When Fodor says paramecia are only sensory systems, he
does not say whether they may have a phenomenology. He also does not say whether
a purely sensory system may be credited with mental states.
15 Actually, Fitch says conflicting things. At times he seems to say that the reasons
computers cant think is that they dont have nano-intentionality. At others times he
seems to say that hes only talking about vertebrates.
16 We dont know whether Searle would think it is possible to have a purely sensory
conscious being, but we dont find anything he has said that rules it out.
17 See discussion in Ma hen, 2005, Chapter 12.
18 See Adams and Aizawa, 2008.
19 See Dretske, 1981, for the interpretation of the mathematical model for use by
cognitive science that we are following.
20 We accept that even biological structures that are selected to be dedicated informa-
tion processors have a physical-chemical make-up, but physics and chemistry alone
wont explain why these structures are where they are and are doing what they are
doing. For that you need to appeal to selectional history (Dretske, 1995).
21 There is a whole literature on this feeling of magic. See Levine, 2001, and this may
be part of why McGinn adopts the mysterian view.
22 We do not say only biological systems can have minds (as Fitch comes close to
saying and Searle is taken to say). To think, computers would need concepts the
non-derived meanings of which were meaningful to them, not only to us or to their
designers.
23 Acknowledgements: James Garvey, Ken Aizawa, John Barker, Fred Dretske, William
Tecumseh Fitch, Annie Steadman, and to University of Delaware Oce of Under-
graduate Research.
Chapter 4
1 However, substance dualism is not without contemporary defenders. The most

prominent is R. Swinburne, for example, in The Evolution of the Soul (second edition,
Oxford: Oxford University Press, 1997), but see also J. Foster, A Defence of
Dualism, in J. Smythies and J. Belo (eds), The Case for Dualism (Charlo esville, VA:
University of Virginia Press, 1989) and The Immaterial Self (London: Routledge, 1996);
W. D. Hart, Dualism, in S. Gu enplan (ed.), A Companion to the Philosophy of Mind
(Oxford: Blackwell, 1994); and H. Robinson, Ma er and Sense (Cambridge: Cambridge
332
Notes
University Press, 1982) and Dualism, in S. Stich and T. Warfield (eds), The Blackwell
Guide to Philosophy of Mind, (Oxford: Blackwell, 2003).
2 Substance dualism is perhaps not commi ed to hylomorphism; see Swinburne,
The Evolution of the Soul, pp. 3302. For general discussion of the position see
H. Robinson, Aristotelian dualism, Oxford Studies in Ancient Philosophy 1, (Oxford:
Oxford University Press, 1983), pp. 12344, and D. Oderberg, Hylemorphic
dualism, Social Philosophy and Policy, 22, (2005), 7099.
3 For articulation of the nature of the classic Cartesian view see M. Rozemond,
Descartess Dualism (Cambridge, MA: Harvard University Press, 2002) and
J. Hawthorne, Cartesian dualism, in P. van Inwagen and D. Zimmerman (eds),
Persons Human and Divine (Oxford: Oxford University Press, 2007). For discussion
of Aquinass view, see A. Kenny, Aquinas on Mind (London: Routledge, 1994).
4 See J. Kim, Lonely Souls: Causality and Substance Dualism, in T. OConnor and
D. Robb (eds), Philosophy of Mind:Contemporary Readings (London: Routledge, 2003).
See also R. Larmer, Mind-body interactionism and the conservation of energy,
International Philosophical Quarterly, 26, (1986), 27785; E. Mills, Interaction and over-
determination, American Philosophical Quarterly, 33, (1996), 10515, and Inter-
actionism and physicality, Ratio, 10, (1997) 16983; and E. J. Lowe, The problem of
psychophysical causation, Australasian Journal of Philosophy, 70, (1992), 26376. Lowe
himself endorses a form of dualism, albeit not the one argued for here in E. J. Lowe,
The causal autonomy of the mental Mind, 102, (1993), 62944; Subjects of Experience
(Cambridge: Cambridge University Press, 1996), and Non-Cartesian substance
dualism and the problem of mental causation, Erkenntnis, 65,1, (2006), 523.
5 Rene Descartes, in J. Co ingham, R. Stootho and D. Murdoch (trans. and eds), The
Philosophical Writings of Descartes (Cambridge: Cambridge University Press, 1994),
vol. II, p. 275.
6 For a longer discussion of this argument, see my Belief in God (Oxford: Oxford
University Press, 2005), pp. 8799.
7 For articulation of this view, see J. Eccles and K. Popper, The Self and its Brain (New
York: Springer, 1977).
8 I qualify this conclusion somewhat in my forthcoming book Free Will (Continuum).
9 The original may be found in F. Jackson,Epiphenomenal qualia, Philosophical
Quarterly, (1982), 12736 and, for a good contemporary discussion, see the papers
collected by P. Ludlow, Y. Nagasawa and D. Stoljar (eds), Theres Something About
Mary: Essays on Phenomenal Consciousness and Frank Jacksons Knowledge Argument
(Cambridge, MA: MIT Press, 2004).
10 On the issue of the problem of consciousness, Denne , for example, arguably
re-conceptualizes it until it becomes tractable to the natural sciences in his Conscious-
ness Explained (Boston: Li le Brown, 1991), but perhaps thereby merely fails to
address the real issue; see David Chalmers, for example, in The Conscious Mind (New
York: Oxford University Press, 1996 ). See also D. Ross, Denne s Conceptual Reform
in Behaviour and Philosophy, 22 (1994), 4152 and N. Latham, Chalmers on the addition
of consciousness to the physical world, Philosophical Studies, 98, (2000), 6793.
11 I am grateful for the comments of Richard Swinburne on an early dra of this
chapter.
Chapter 5
1 What does it mean to be fundamental? As I understand it, the basic or fundamental

properties (1) may determine each other, as in, for example, F=MA, (2) may determine
333
Notes
other properties, as how the properties of physics determine (many think) the
chemical properties (i.e. once the properties of physics are in place, the chemical
properties are also in place), and (3) are not determined by anything that they do not
themselves determine. See Montero (2006) for a suggestion on how to formulate
physicalism if there is no fundamental level.
2 Indeed, a world that is entirely understandable to a human mind seems, if anything,
to point a non-physicalistic view, as it would seem to hint at a creator that made the
world intelligible to humans. (This isnt to say that any anti-physicalistic view implies
the existence of a creator, but just that the existence of a creator would seem to imply
an anti-physicalistic position.)
3 See Wilson (2010).
4 Terence E. Horgan (1993).
5 For an explanation of this stance see McLaughlin (2001).
6 See Melynk (2003). I argue against his view in Montero (1999).
Chapter 6
1 Jaegwon Kim has pressed this objection repeatedly. He oers the analogy of killing
someone by firing a gun. Had the gun had a silencer, the shot would still have killed
the victim. Thus, the noise of the shot was causally irrelevant. However, by analogy
had the shot been fired with by a gun with a silencer, the shooting would have been
a dierent event, and so would the death. Kim also presses the epiphenomenal charge
by claiming that it is not the mental property but only the physical property of an
event that is causally ecacious. However, for Davidson, causality is a relation
between events not properties, nor does it hold in virtue of properties of events. And
besides, Davidson doesnt believe in properties, and so Kims objections are beside
the point (see Davidson, 1993).
2 Davidson (1986a) A Coherence Theory of Truth and Knowledge, p. 150.
3 Davison rejects the locution, knowing what I mean, preferring, meaning what
I say, and he points out that if we didnt mean what we say we would be
uninterpretable.
4 Note that Denne notes that this objection is based on too limited a view of causation.
See Denne , 1991b. However, he doesnt elaborate.
5 See Fodors 1974 paper Special Sciences: The disunity of science as working
hypothesis.
Chapter 7
1 It also concerns the individuation conditions of psychological contents, states and

concepts. It does not, however, concern phenomenal properties unless they are
conceived as being representational.
2 In order to make sense of the debate between internalists and externalists the neces-
sity involved in both psycho-physical supervenience theses should be understood as
nomological necessity: necessity consistent with the laws of nature.
3 See Putnam (1975) for the original Twin Earth thought experiment. Although Putnam
introduced the Twin Earth thought experiment his argument concerned linguistic
properties rather than psychological properties: indeed Putnam was clearly an inter-
nalist at the time. The argument was adapted by Burge to establish externalism.
334
Notes
See Burge (1979). Putnam himself accepts the adaptation as can be seen by remarks
in his (1996).
4 Putnams original example was of water on Earth and twater on Twin Earth, where
twater is a substance superficially indistinguishable from water but with a dierent
chemical composition. I have not used his example here because it has an irrelevant
complication, namely that S on Earth could not be an intrinsic physical duplicate of
S* on Twin Earth if S were partly composed of water and S* were partly composed of
twater.
5 The view of natural kinds that is taken to underwrite this form of externalism can be
found in Kripke (1972) and Putnam (1975). However, the view remains controversial.
For a thorough overview of recent positions on natural kinds and natural kind terms
see Wilkerson (1998).
6 See Burge (1982). The original example involves an incomplete understanding of
the term arthritis.
7 See Burge (1986a) and (1986b).
8 The example is from Burge (1986b).
9 For reasons of space I have not been able to discuss Davidsons views, which in some
sense straddle the internalism/externalism divide. According to Davidson, roughly
speaking, the content of a thought is determined by the way in which a subject is
best interpreted in the context of a shared world. Thus although the meaning of a
subjects words and the contents of her thoughts are and must be grounded in the
discriminative capacities and epistemic outlook of the individual, they are externally
individuated nonetheless because in order to interpret an individual one must
make essential reference to objective properties in the environment to which she
(and you) are jointly related. See Davidson (1984b) especially Chapters 9 and 10, and
(2001), especially Chapters 1, 2 and 9.
10 See Evans (1982) and McDowell (1977, 1984). See also McDowell (1986).
11 See, for example, Salmon (1986) and Soames (2002).
12 See Burge (1977) for more detail. See also Segal (1989). It should be noted that the
view is consistent with both internalism and externalism about the representational
content of singular thoughts thus conceived. It counts as a form of singular inter-
nalism simply because the representational content of a singular thought is not
dependent for its individuation on the object the thought concerns. For all this,
the predicative element of the thought may depend for its individuation on the
properties to which the thinker is related, for reasons akin to those given in the
Predicative Externalism section above.
13 This view of proper names is advocated by Burge in his (1973). For recent defences
of the view see Elugardo (2002) and Sawyer (2009).
14 Epistemic outlook can itself be understood individualistically or anti-individual-
istically. This makes it dicult to oer a neutral characterization of the internalist
position conceived generally.
15 See Segal (2000) for a thorough defence of this kind of view.
16 See Fodor (1980), McGinn (1982b) and Stich (1983).
17 For this view see Chalmers (2003c).
18 See Sawyer (2007).
19 See Burge (1979), Section IV.
20 Both type- and token-physicalism are commi ed to this claim, although type-
physicalism is commi ed to the stronger claim about property identity as well.
21 That psychological states are the causes of actions (re)gained prominence following
Davidsons (1980), Chapter 1. For an alternative view see Morris (1992).
22 For arguments to this eect see Fodor (1987) and (1991a).
23 See Burge (1989) and (1993b) and Wilson (1995).
335
Notes
24 For recent discussion see Noordhof (2006a) and Kezer and Schouten (2007).
25 In my discussion of both problems I focus on predicative externalism, but the
problems and responses in each case are analogous for singular externalism.
26 For arguments concerning the strength of the argument see Ludlow (1995) and (1997)
and Warfield (1997). See also McLaughlin and Tye (1998) and Brown (2004).
27 See, for example, Burge (1988), Heil (1988) and Peacocke (1999), Chapter 5. Burge
has gone further in identifying a set of what he calls cogito-like judgements that
provide a limiting case of direct, non-empirical, authoritative self-knowledge that
withstand any amount of disruption to presupposed background conditions for
self-knowledge. Cogito-like judgements are self-referential and self-verifying in
virtue of being so. Consequently, they cannot but be true even if their contents are
externally individuated. Examples include: I am now thinking that writing requires
concentration, and I hereby judge that examples need elaboration. See Sawyer (2002) for
a defence of the view.
28 See, in particular, Burge (1988).
29 Davidson oers a dierent response that follows from his particular form of externalism
referred to in Note 9 above. For his response see Davidson (2001), Chapters 1 and 2.
30 See Boghossian (1989) for the original presentation of the argument.
31 See Ludlow (1998).
32 See Burge (1998) and (1993a).
33 In addition, and following discussion of cogito-like judgements about current
thoughts in Note 26, Burge extends the realm of cogito-like judgements to include
thoughts about past thoughts. Thus the was and thereby in judgements such as I was
thereby thinking that p relate to elements in the original thought preservatively rather
than referentially. See Burge (1998).
34 There are related issues concerning the implications of externalism for reasoning.
In addition to Burge (1993a) and (1998), see Boghossian (1992) and Schier (1992).
See Goldberg (2007) for an account of the implications of externalism for content
preservation and discursive justification.
35 McKinsey sparked the controversy with his (1991), although he rejects the inter-
pretation of his argument that led to the controversy. See also Brown (1995).
36 See Brueckner (1992) and Goldberg (2003).
37 For this strategy, see Davies (2000), for example, and Wright (2004), for example.
There are dierences between the positions of Davies and Wright but the general
strategy is the same. The strategy is criticized in Sawyer (2006).
38 See Sawyer (1998) and (2006).
39 Externalism has further epistemological implications that I have not had the space to
discuss here. See, for example, Majors and Sawyer (2005) in which it is argued that
externalism (and only externalism) grounds a reliabilist theory of justification. See
also Majors and Sawyer (2007) for the ramifications of externalism in epistemology
and meta-ethics.
40 I have not had the space to explain this fully here. See Majors and Sawyer (2005).
Chapter 8
1 This chapter is based on the much fuller discussions in my book Mind as Machine:
A History of Cognitive Science (Boden, 2006). No specific references are given below.
But the most directly relevant parts are Chapters 4 and 16 entire, and Sections 1.iii
and iii.bd; 6.iii.c and iv.c; 7.i.eh, iii, and vi.dh; 9.vii and x; 11.iiiii; 12.x; 13.vii;
14.ii and viiixi; and 15.i.
336
Notes
Chapter 9
1 That is, we will be concerned with issues of a psychosemantics, or issues about

the nature and meaning of representations in the mind, not, for example, a linguo-
semantics, as might be provided for the expressions in natural language, which may
or may not be used by a mind to express its thoughts.
2 Note that, pace Denne (1987a) and others who stress the role of norms in intentional
a ribution, even much irrational thought and behaviour is understood intentionally,
as in, for example, the gambler fallacy, disregard of base rates, and discounting of
future satisfactions. Thus, the gamblers fallacy is likely due to errors in reasoning
about probability, not to some non-intentional, purely mechanical breakdown. See
the exchange between Wedgwood (2007) and Rey (2007) for recent discussion.
3 See Fodor (1975), Rey (1997) and Harnish (2002) for extensive discussion of the
program, and the several volumes of Osherson (1995/98) for representative discus-
sions of empirical work within it.
4 Turings famous proposal of Turing machines, and his thesis that they can compute
anything that can be computed should not be confused with the far less plausible
test for intelligence named a er him, according to which a machine would count as
intelligent if a (normal?) human being couldnt distinguish its teletype responses to
questions from those of an intelligent human being. Pace Searles notorious Chinese
Room Argument, CRTT is commi ed to no such behavioural test, but to quite detailed
stories about how behaviour is produced (see Rey, 1997, 2002 for further discussion).
5 Resurrecting (unfortunate) medieval terminology, Brentano thought these peculiari-
ties were the mark of intentionality. For those new to these discussions, intentionality
in this sense means directedness or aboutness, as when we say that a thought is
directed upon or about its object, in the way that a thought about Julius Caesar
is directed upon, or about Caesar. Intentional in the sense of deliberate (as in he
coughed intentionally) is only one of a very large class of states that are intentional
in Brentanos sense.
6 See Quine (1953a), Cartwright (1960/1987), Evere and Hofweber (2000) and
Priest (2005) for rich discussions. See note 15 below for rejections of empty
representations.
7 The distinction is, of course, close to the much-discussed distinctions between
transparent/de re and opaque/de dicto readings of propositional a itudes and/or
their ascriptions (e.g. see Kaplan, 1969). I dont want to assimilate my distinction
immediately to those, both because they are the objects of enough controversy
on their own, and because a usual strategy for understanding them wont work for
represent (e.g. a transparent reading of John thinks of Sam Clemens that hes
funny may well involve John being related to a representation, Mark Twain
is funny, that involves a representation, Mark that in fact represents Sam. But
this understanding obviously cant be available for the term represent itself).
8 This, of course might be a reason to treat representation as also opaque to substitution
in a serious psychology.
9 Quines challenges have recently been vigorously pressed by Fodor (1998) as
arguments against any epistemic account of meaning.
10 A cautionary note: information is used quite freely in cognitive science, o en with
a tacit presumption that its use is sanctioned by the technical notion of information
(as roughly negative entropy) introduced by Shannon and Weaver. Perhaps some
uses are sanctioned in this way, but it is far more likely that the uses in psychology
are simply disguised references to intentional content. Dretske himself introduces a
specifically co-variational notion, in terms of strict co-variance, which is the notion
being recruited here to explain intentional content.
337
Notes
11 This problem is of a piece with the problem Kripke (1982) a ributes to Wi genstein
(1953) about how to distinguish someone who has added 57 and 63 and obtained 5 as
an error from someone who is computing a dierent function, quaddition, which
is identical to addition except for the case of 57 and 63.
12 Its important to note that Fodor doesnt intend his proposal as a sucient condition
on intentionality tout court, but only as a sucient condition for meeting disjunction
objections; see his 1990c, pp. 12731.
13 In Rey (forthcoming) I argue that the representations of geometrical figures (geons)
in early vision, as well of the standard entities of linguistics (words, phonemes) are
empty along these lines.
14 One strategy is to appeal to real, but simply uninstantiated properties, such as unicorn-
hood, which (arguably) can exist even without unicorns (see Fodor, 1987, 1991b).
Another is to claim that at least syntactically simple expressions are not genuine
representations (see Millikan, 2000). See Rey (forthcoming) for discussion.
15 Note that the Zeus/Jupiter example shows one cant rely here on co-reference alone.
16 This is a condensation of a longer discussion in my (2009) in which I try to distill what
seems to me common and correct in Fodors and Horwichs proposals, while rejecting
what seems to me mistaken in each (viz., Fodors externalism and Horwichs defla-
tionism). Devi s (1996) proposal also looks to explanatory roles, but is not so focused
on the asymmetric basicality condition, which seems to me crucial to replying to the
Quinean challenge. Note that (BAS) is deliberately neutral between Horwichs
(explanatory) and Fodors (asymmetric) ways of expressing what seems to me the
common important idea.
Chapter 10
1 My formulation will follow Honderichs (1982) classic statement of the problem.

2 This does not preclude the possibility that other types of relations could play the role
of R. In fact, in his later work on explanation Kim (1994) claims that explanations
track dependence relations, so it is certainly possible that other kinds of dependence
relations (such as mereological dependence) can serve as the explanatory relation in
other explanatory contexts.
3 Closure is motivated by the longstanding problems with Cartesian interaction.
Most recently Kim defines the causal closure of the physical domain as follows: If
a physical event has a cause at t, it has a physical cause that occurs at t (Kim, 2005,
p. 43). He also considers a stronger version of closure: Any cause of a physical event
is itself a physical event, that is, no nonphysical event can be a cause of a physical
event (ibid., p. 50) but claims that this is too strong since it rules out mind-body
causation all by itself and would allow one to dispense of the exclusion principle
altogether in the argument against non-reductive physicalism.
4 The problem with weak supervenience is that it lacks the modal force required to
generate a relation of dependence between properties by virtue of the fact that they
need not co-vary across all possible worlds. The problem for global supervenience
is it is not suciently restrictive because it says nothing about how mental and
physical properties are distributed within any possible world. For a more detailed
discussion see (Kim, 1987, 1990b, 1993b; Campbell 2000).
5 See Kims assessment of this option (Kim, 2005, p. 61). See also (Ney, 2007) for a
helpful critical discussion of the notion of constitution.
6 For defences of Davidson in terms of these kinds of considerations see (Campbell,
1997, 2003, 2008; Gibb, 2006), though Gibb does go on to argue that this reply leads
to an intolerable account of causal relata.
338
Notes
7 Indeed, the consensus in the literature seems to be that the Davidsonian account of
events should be avoided because it renders causation an u erly mysterious relation.
The virtue of a property exemplification view of events is that it permits an account
of why a cause produces its eect in terms of an appeal to its causally relevant
properties (Gibb, 2006).
8 For further discussion see (Campbell and Moore, 2009).
9 Kim thinks that he has not begged any questions since he claims that if it is
aspects of events, rather than events simpliciter, that are explained, then explanatory
exclusion would apply to these event aspects (Kim, 1989a, p. 96) and elsewhere, We
have so far spoken indierently of both events simpliciter and events being a certain
kind or having certain properties as causes and eects. This makes no dierence:
however we individuate causes and eects, we face the same problem [about
explanatory exclusion] (Kim, 1990a, pp. 401). I think we have seen ample reason
to disagree with these claims. Interestingly, Marras and Yli-Vakkuri (2008) have
recently observed that Kims improved version of the exclusion argument (the
supervenience argument) also presupposes a fine-grained account of event identity
and also begs the question against non-reductive physicalism.
10 Indeed, Marras (1998) ably shows it does not.
11 This observation has also been made in a slightly dierent context by Marras
(Marras, 2007; Marras and Yli-Vakkuri, 2008). It is overly simplistic simply to suggest
that the choice we face is between Davidsons and Kims conception of events, though
that is o en the sense one gains from the literature. While theirs are the most promi-
nent views there are others worth considering, such as Chisholms (1970, 1976) and
Lombards (1986), and of course there is Horgans (1978) claim that we can and
should make do without introducing events into our ontology at all. Macdonald and
Macdonald (2006) adopt a modified version of Lombards account and argue that it
allows for the causal ecacy of mental property instances. However, their approach
strikes me as a return to the Davidsonian position since events are identified with
property instances. It would be interesting to explore whether or not the exclusion
argument can be maintained on an alternative model of events. If, like Lombard,
one adopts the position that events can have more than one constitutive property
it seems an exclusion argument formulated in terms of such events would be
even more vulnerable to the dual explanandum reply. Unfortunately, I do not have
sucient space to explore this question here.
Chapter 12
1 More precisely, its an example of the version of embodied cognition with which
we shall be concerned here. There are other versions of the view. For instance,
some embodied cognition theorists concern themselves with the way in which
embodiment has an impact on our understanding of perceptual experience (e.g.
ORegan and No, 2001, No, 2004). Others argue that our embodiment structures
our concepts (Lako and Johnson, 1980, 1999). This sample is not exhaustive.
2 The classic presentation of what I am calling the extended mind hypothesis is by
Clark and Chalmers (1998). See Menary (forthcoming) for a recent collection of
papers. Rather confusingly, the view has always traded under a number of dierent
names, including close variants of the original moniker, such as the hypothesis of
extended cognition (Rupert, 2004) and the extended cognition hypothesis (Wheeler, forth-
coming a), but also active externalism (Clark and Chalmers, 1998), vehicle externalism
(Hurley, 1998; Rowlands, 2003), environmentalism (Rowlands, 1999), and locational
externalism (Wilson, 2004).
339
Notes
3 It has o en been noted that connectionist networks may be analysed in terms of

cognitively relevant functions which need to be specified at a finer level of grain
than those performed by classical computational systems (e.g. using mathematical
relations between units that do not respect the boundaries of linguistic or conceptual
thought), hence Clarks term micro-functionalism.
4 Two comments: First, although this is not the place to launch into a critique of the
details of Adams and Aizawas position, my view is that while they are right that
ExM needs a mark of the cognitive, they are wrong about what that mark might be.
Secondly, my appeal to a scientifically informed, theory-loaded mark of the cognitive
will, in some quarters, be controversial. For example, Clark (2008b) suggests that the
domain of the cognitive should be determined by our intuitive folk-judgements of
what counts as cognitive. His supporting argument is (roughly) that our intuitive
understanding of the cognitive is essentially locationally uncommi ed, while the
range of mechanisms identified by cognitive science is in truth too much of a motley
to be a scientific kind, and so will thwart any a empt to provide a scientifically
driven, theory-loaded account of the cognitive, locationally uncommi ed or other-
wise. I disagree with this assessment. I hold out for a locationally uncommi ed
account of the cognitive that is scientifically driven and theory-loaded on the grounds
(roughly) that our intuitive picture of the cognitive has a deep-seated inner bias,
while Clarks argument for the claim that there is a fundamental mechanistic disunity
in cognitive science is far from compelling (Wheeler, forthcoming b).
5 In previous ExM treatments of Bechtels logical reasoning studies, Rowlands
(1999, pp. 16871) and Menary (2007, also pp. 16871) rely at root not on parity
considerations to justify the claim of cognitive extension, but rather on the integra-
tion of inner connectionist processing with external symbol systems in order to com-
plete a cognitive task that could not ordinarily be achieved by the inner networks
alone. My own view is that the mere fact that an external resource is necessary to
complete a cognitive task is not sucient to establish cognitive extension, as opposed
to a compelling case of embodied-embedded cognition.
340
Bibliography
Adams, F. (1991), Causal Contents, in B. McLaughlin (ed.), Dretske and His critics
(Oxford: Blackwell), 13156.
(2003), Thoughts and Their Contents: Naturalized Semantics, in S. Stich and
T. Warfield (eds), The Blackwell Guide to the Philosophy of Mind (Oxford: Blackwell),
14371.
Adams, F., and Aizawa, K. (2008), The Bounds of Cognition (Oxford: Blackwell).
(1994), Fodorian Semantics, in Stich and Warfield (1994), 22342.
Adams, F., and Dietrich, L. (2004), Swampmans Revenge: Squabbles among the
Representationalists, Philosophical Psychology, 17: 32340.
Agre, P. E., and Rosenschein, S. J. (eds) (1996), Computational Theories of Interaction
and Agency (Cambridge, MA: MIT Press).
Aikns, K. (1993), A Bat without Qualities, in M. Davies and G. Humphreys (eds),
Consciousness (Oxford: Blackwell).
(1996), Lost the Plot? Reconstructing Daniel Denne s Multiple Dra s Theory
of Consciousness, Mind and Language, 2: 143.
Aizawa, K. (2003), The Systematicity Arguments (Dordrecht: Kluwer).
Aleksander, I. (2005), The World in My Mind, My Mind in the World: Key
Mechanisms of Consciousness in Humans, Animals, and Machines (Exeter: Imprint
Academic).
Aleksander, I., and Dunmall, B. (2003), Axioms and Tests for the Presence of
Minimal Consciousness in Agents, in O. Holland (ed.), Machine Consciousness
(Exeter: Imprint Academic), 718. Special issue of the Journal of Consciousness
Studies, 10 (45).
Alexander, S. (1920), Space, Time, and Deity, 2 vols. (London: Macmillan).
Anscombe, G., and Geach, P. (1954), Descartes: Philosophical Writings (Indianapolis:
Bobbs-Merrill).
Anscombe, G. E. M. (1957), Intention (Oxford: Blackwell).
(1965), The Intentionality of Sensation: A Grammatical Feature, in A. No and
E. Thompson (eds) (2002), Vision and Mind, Selected Readings in the Philosophy of
Perception (Cambridge, MA: MIT Press).
Antony, L. (1989), Anomalous Monism and the Problem of Explanatory Force,
Philosophical Review, 98: 15387.
(1997), What Its Like to Smell a Gardenia, Times Literary Supplement 4897
(7 February).
Arbib, M. A. (1982), Modelling Neural Mechanisms of Visuomotor Coordination
in Frog and Toad, in S. Amari and M. A. Arbib (eds), Competition and Cooperation
in Neural Nets, Lecture Notes in Biomathematics, 45 (Berlin: Springer-Verlag),
34270.
Arbib, M. A., and Hesse, M. B. (1986), The Construction of Reality (Cambridge:
Cambridge University Press).
341
Bibliography
Arbib, M. A., Boylls, C. C., and Dev, P. (1974), Neural Models of Spatial Perception
and the Control of Movement, in W. D. Keidel, W. Handler and M. Spreng (eds),
Cybernetics and Bionics (Munich: Oldenbourg), 21631.
Armstrong, D. (1968), A Materialist Theory of Mind (London: Routledge & Kegan
Paul).
(1981), What is Consciousness?, in The Nature of Mind (Ithaca, NY: Cornell
University Press).
Ayers, M. (1990), Locke, vol. 2 (London: Routledge).
Baker, L. R. (2000), Persons and Bodies: A Constitution View (Cambridge: Cambridge
University Press).
Balog, K. (1999), Conceivability, Possibility and the Mind-Body Problem, The
Bechtel, W. (1994), Natural Deduction in Connectionist Systems, Synthese, 101:
43363.
(1996), What Knowledge Must Be in the Head in Order to Acquire Language,
in B. Velichkovsky and D. M. Rumbaugh (eds), Communicating Meaning: The
Evolution and Development of Language (Hillsdale, NJ: Lawrence Erlbaum
Associates).
Bechtel, W., and Abrahamsen, A. (1991), Connectionism and the Mind: An Introduction
to Parallel Processing in Networks (Oxford: Basil Blackwell).
Bedau, M. (1999), Supple Laws in Psychology and Biology, in V. Hardcastle (ed.),
Where Biology Meets Psychology (Cambridge, MA: Bradford Books, MIT Press),
25171.
Beer, R. D. (1996), Toward the Evolution of Dynamical Neural Networks for
Minimally Cognitive Behavior, in P. Maes, M. Mataric, J. Meyer, J. Pollack and
S. Wilson (eds), From Animals to Animats 4: Proceedings of the Fourth International
Conference on Simulation of Adaptive Behavior (Cambridge, MA: MIT Press), 4219.
(2003), The Dynamics of Active Categorical Perception in an Evolved Model
Agent, Adaptive Behavior, 11 (4): 20943.
Behan, D. (1979), Locke on Persons and Personal Identity, Canadian Journal of
Philosophy, 9: 5375.
Block, N. (1978), Troubles with Functionalism, in C. W. Savage (ed.), Perception
and Cognition: Issues in the Foundations of Psychology, Minnesota Studies in the
Philosophy of Science, vol. 9 (Minneapolis: University of Minnesota Press),
261325.
(1986a), Advertisement for a Conceptual Role Semantics, in P. French, T. Uehling
and H. We stein (eds), Studies in the Philosophy of Mind (Minneapolis: University
of Minnesota Press).
(1986b), Advertisement for a Semantics for Psychology, in P. French, T. Uehling
and H. We stein (eds), Midwest Studies in Philosophy, vol. 10 (Minneapolis:
University of Minnesota Press), 61578; reprinted in Stich and Warfield (2003),
81135.
(1986c), Functional Role and Truth Conditions, Proceedings of the Aristotelian
Society, Supplementary Volume 60: 157181.
(1990), Inverted Earth, in J. E. Tomberlin (ed.), Philosophical Perspectives, vol. 4
(Atascadero, CA: Ridgeview), 5379.
(1994), Consciousness, in Gu enplan (1994), 21019.
342
Bibliography
(1995), On a Confusion about the Function of Consciousness, Behavioural and

Brain Sciences, 18: 22747.
(1996), Mental Paint and Mental Latex, in Enrique Villenueva (ed.), Perception,
Philosophical Issues, 7 (Northridge, CA: Ridgeview), 1948.
(1998), Is Experience Just Representing?, Philosophy and Phenomenological
Research, 58: 66370.
(2003), Do Causal Powers Drain Away?, Philosophy and Phenomenological
Research, 67 (1): 13350.
(2007), Consciousness, Accessibility, and the Mesh between Psychology and
Neuroscience, Behavioral and Brain Sciences, 30: 481548.
(forthcoming), Comparing the Major Theories of Consciousness, in M. Gazzangia
(ed.), The Cognitive Neurosciences, 4th edn (Cambridge MA: MIT Press).
Block, N., and R. Stalnaker (1999), Conceptual Analysis, Dualism and the
Explanatory Gap, The Philosophical Review, 108: 146.
Boden, M. A. (1965), McDougall Revisited, Journal of Personality, 33.
(1970), Intentionality and Physical Systems, Philosophy of Science, 37: 20014.
(1972), Purposive Explanation in Psychology (Cambridge, MA: Harvard University
Press).
(1978), Human Values in a Mechanistic Universe, in G. Vesey (ed.), Human
Values: Royal Institute of Philosophy Lectures 197677 (Brighton: Harvester Press),
13571.
(1994), Multiple Personality and Computational Models, in A. Phillips-Griths
(ed.), Philosophy, Psychology, and Psychiatry (Cambridge: Cambridge University
Press), 10314.
(1999), Is Metabolism Necessary?, British Journal for the Philosophy of Science,
50: 23148.
(2000a), Autopoiesis and Life, Cognitive Science Quarterly, 1: 129.
(2000b), Cra s, Perception, and the Possibilities of the Body, British Journal of
Aesthetics, 40: 289301.
(2004), The Creative Mind: Myths and Mechanisms, 2nd edn, expanded/revised
(London: Routledge).
(2006), Mind as Machine: A History of Cognitive Science, 2 vols. (Oxford: Oxford
University Press).
(2007), Creativity and Conceptual Art, in P. Goldie and E. Schellekens (eds),
Philosophy and Conceptual Art (Oxford: Clarendon Press), 21637.
Boden, M. A., and Edmonds, E. A. (2009) What is Generative Art?, Digital
Creativity, 20 (12): 2146.
Boghossian, P. (1989), Content and Self-Knowledge, Philosophical Topics, 17: 526.
(1992), Externalism and Inference, Philosophical Issues, 2: 1128.
Bond, A. H., and Gasser, L. (eds) (1988), Readings in Distributed Artificial Intelligence
(San Francisco: Morgan Kaufmann).
Bontly, T. (2002), The Supervenience Argument Generalizes, Philosophical Studies
109 (1): 7596.
(2005), Exclusion, Overdetermination, and the Nature of Causation, Journal of
Philosophical Research, 30: 26182.
Boyer, P. (1994), The Naturalness of Religious Ideas: A Cognitive Theory of Religion
(London: University of California Press).
343
Bibliography
Braddon-Mitchell, D., and Jackson, F. (1996), The Philosophy of Mind and Cognition
(Oxford: Blackwell).
Brentano, F. (1874), Psychology from an Empirical Standpoint (London: Routledge).
Bringsford, S. (1994), Computation, Among Other Things, is Beneath Us, Minds
and Mahines, 4 (4): 46988.
Broad, C. D. (1925), The Mind and Its Place in Nature (London: Routledge & Kegan
Paul).
Broadbent, D. E. (1958), Perception and Communication (Oxford: Pergamon Press).
Brooks, R. A. (1991a), Intelligence without Reason, in Proceedings of 12th
International Joint Conference on Artificial Intelligence (San Mateo, CA: Morgan
Kauman), 56995; reprinted in Brooks, Cambrian Intelligence: The Early History
of the New AI (Cambridge, MA: MIT Press), 13386.
(1991b), Intelligence without Representation, Artificial Intelligence, 47: 13959.
Brown, J. (1995), The Incompatibility of Anti-Individualism and Privileged
Access, Analysis, 53: 14956.
(2004), Anti-Individualism and Knowledge (Cambridge, MA: MIT Press).
Brueckner, A. (1992), What an Anti-Individualist Knows A Priori, Analysis, 52: 11118.
Bruner, J. S., Goodnow, J., and Austin, G. (1956), A Study of Thinking (New York:
Wiley).
Burge, T. (1973), Reference and Proper Names, Journal of Philosophy, 70: 42539.
(1977), Belief De Re, Journal of Philosophy, 74: 33862.
(1979), Individualism and the Mental, in P. French, T. Uehling, Jr., and
H. We stein (eds), Midwest Studies in Philosophy, vol. 4 (Minnesota: University
of Minnesota Press), 73121.
(1982), Other Bodies, in A. Woodfield (ed.), Thought and Object: Essays on
Intentionality (Oxford: Clarendon Press).
(1986a), Individualism and Psychology, Philosophical Review, 95 (1): 345.
(1986b), Intellectual Norms and Foundations of Mind, Journal of Philosophy,
83 (12): 697720.
(1988), Individualism and Self-Knowledge, Journal of Philosophy, 85: 64963.
(1989), Individuation and Causation in Psychology, Pacific Philosophical
Quarterly, 70: 30322.
(1993a), Content Preservation, Philosophical Review, 102: 45788.
(1993b), Mind-Body Causation and Explanatory Practice, in J. Heil and A. Mele
(eds), Mental Causation (Oxford: Oxford University Press).
(1998), Memory and Self-Knowledge, in P. Ludlow and N. Martin (eds),
Externalism and Self-Knowledge (Stanford: CSLI).
Bush, V. (1945), As We May Think, Atlantic Monthly, 176 (July): 1018; reprinted
in R. Packer and K. Jordan (eds), Multimedia: From Wagner to Virtual Reality
(London: W. W. Norton, 2001), 13553.
Butler, J. (1975) [1736], Of Personal Identity, in J. Perry (ed.), Personal Identity
(Berkeley and Los Angeles: University of California Press).
Byrne, A. (1997), Some Like It Hot: Consciousness and Higher-Order Thoughts,
Philosophical Studies, 86: 10329.
(2001), Intentionalism Defended, The Philosophical Review, 110: 199240.
Calude, C., Casti, J. L., and Dinneen, M. (eds) (1998), Unconventional Models of
Computation (London: Springer).
344
Bibliography
Campbell, J. (1994), Past, Space, and Self (Cambridge, MA: MIT Press).
Campbell, N. (1997), The Standard Objection to Anomalous Monism, Australasian
Journal of Philosophy, 73 (3): 37382.
(2000), Supervenience and Psycho-Physical Dependence, Dialogue: Canadian
(2003), Causes and Causal Explanations: Davidson and His Critics, Philosophia:
Philosophical Quarterly of Israel, 31 (12): 14957.
(2007), Explanatory Pluralism, International Journal of the Humanities, 5 (3):
2529.
(2008a), Explanatory Exclusion and the Individuation of Explanations, Facta
Philosophica, 10.
(2008b), Mental Causation: A Nonreductive Approach (New York: Peter Lang).
Campbell, N., and Moore, D. (2009), On Kims Exclusion Principle, Synthese, 169
(1): 7590.
Cariani, P. (1992), Emergence and Artificial Life, in C. G. Langton, C. Taylor,
J. D. Farmer and S. Rasmussen (eds), Artificial Life II (Redwood City, CA:
Addison-Wesley), 77597.
Carruthers, P. (1996), Language, Thought and Consciousness (Cambridge: Cambridge
University Press).
(2000), Phenomenal Consciousness: A Naturalistic Theory (Cambridge: Cambridge
University Press).
Cartwright, R. T. (1960/87), Negative Existentials, in his Philosophical Essays
(Cambridge, MA: MIT Press).
Chalmers, D. (1995), Facing Up to the Problem of Consciousness, Journal of
Consciousness Studies, 2: 20019.
(1996), The Conscious Mind: In Search of a Fundamental Theory (Oxford: Oxford
University Press).
(1999a), First-Person Methods in the Science of Consciousness, Bulletin from
the Center for Consciousness Studies.
(1999b), Materialism and the Metaphysics of Modality, Philosophy and Phenome-
nological Research, 59: 47393.
(2002a), Consciousness and Its Place in Nature, in Chalmers (2002b), 24772.
(2002b), Philosophy of Mind: Classical and Contemporary and Readings (New York:
Oxford University Press).
(2003a), Consciousness and Its Place in Nature, in Stich and Warfield (2003),
10242.
(2003b), The Matrix as Metaphysics, March 20th. Available at h p://
whatisthematrix.warnerbros.com.
(2003c), The Nature of Narrow Content, Philosophical Issues, 13: 4666.
(2008), Foreword to Andy Clarks Supersizing the Mind, in Clark (2008b), ixxvi.
(forthcoming), Two Dimensional Semantics, in E. Lepore and B. Smith (eds),
Oxford Handbook to the Philosophy of Language (Oxford: Oxford University Press).
Chisholm, R. M. (1956), Perceiving: A Philosophical Study, in D. Rosenthal (ed.),
The Nature of Mind (Oxford: Oxford University Press, 1990), chapter 11.
(1970), Events and Propositions, Nous, 4 (1): 1524.
(1976), Person and Object: A Metaphysical Study (London: G. Allen & Unwin and
La Salle, IL: Open Court).
345
Bibliography
Chomsky, N. (1994), Chomsky, Noam, in Gu enplan (1994), 15367.

Chrisley, R. L. (2000), Transparent Computationalism, in M. Scheutz (ed.), New
Computationalism (Berlin: Academia Verlag, 2000), 10520.
Churchland, P. M. (1981), Eliminative Materialism and the Propositional A itudes,
Journal of Philosophy, 78: 6790.
(1986), Some Reductive Strategies in Cognitive Neurobiology, Mind, 95:
279309.
(1989a), Knowing Qualia: A Reply to Jackson, in his A Neurocomputational
Perspective (Cambridge MA: MIT Press).
(1989b), A Neurocomputational Perspective: The Nature of Mind and the Structure
of Science (Cambridge, MA: MIT Press).
(1995), The Engine of Reason and Seat of the Soul (Cambridge, MA: MIT Press).
Churchland, P. M., and Churchland, P. S. (1981), Functionalism, Qualia, and
Intentionality, Philosophical Topics, 12: 12145.
Churchland, P. S. (1986), Neurophilosophy: Towards a Unified Theory of the Mind-Brain
Churchland, P. S., and Sejnowski, T. J. (1992), The Computational Brain (Cambridge,
MA: MIT Press).
Clark, A. J. (1989), Microcognition: Philosophy, Cognitive Science, and Parallel Distributed
Processing (Cambridge, MA: MIT Press).
(1990), Connectionist Minds, Proceedings of the Aristotelian Society, 90: 83102.
(1991), Systematicity, Structured Representations and Cognitive Architecture: A
Reply to Fodor and Pylyshyn, in T. Horgan and J. Tienson (eds), Connectionism
and the Philosophy of Mind (Dordrecht: Kluwer), 198218.
(1993), Associative Engines: Connectionism, Concepts, and Representational Change
(1996), Connectionism, Moral Cognition, and Collaborative Problem-Solving, in
L. May, M. Friedman and A. Clark (eds), Mind and Morals: Essays on Ethics and
Cognitive Science (Cambridge, MA: MIT Press), 10927.
(1997), Being There: Puing Brain, Body, and World Together Again (Cambridge,
MA: MIT Press).
(2003a), Natural-Born Cyborgs: Why Minds and Technologies are Made to Merge
(Oxford: Oxford University Press).
(2003b), The Twisted Matrix: Dream, Simulation or Hybrid?, 19 December.
Available at h p://whatisthematrix.warnerbros.com.
(2008a), Pressing the Flesh: A Tension in the Study of the Embodied, Embedded
Mind?, Philosophy and Phenomenological Research, 76 (1): 3759.
(2008b), Supersizing the Mind: Embodiment, Action, and Cognitive Extension (New
York: Oxford University Press).
Clark, A. J., and Chalmers, D. J. (1998), The Extended Mind, Analysis, 58: 719.
Clark, A. J., and Grush, R. (1999), Towards a Cognitive Robotics, Adaptive Behavior,
7: 516.
Clark, A. J., and Karmilo-Smith, A. (1993), The Cognizers Innards: A Psycholo-
gical and Philosophical Perspective on the Development of Thought, Mind and
Language, 8: 487519.
Clark, A. J., and Thornton, C. (1997), Trading Spaces: Computation, Representation,
and the Limits of Uninformed Learning, Behavioral and Brain Sciences, 20: 5790.
346
Bibliography
Clark, A. J., and Wheeler, M. W. (1999), Genic Representation: Reconciling Content

and Causal Complexity, British Journal for the Philosophy of Science, 50: 10335.
Cole, M., and Bruner, J. S. (1971), Cultural Dierences and Inferences about
Psychological Processes, American Psychologist, 26: 86776.
Collins, H. (2000), Four Kinds of Knowledge, Two (or Maybe Three) Kinds of
Embodiment, and the Question of Artificial Intelligence, in M. Wrathall and
J. Malpas (eds), Heidegger, Coping and Cognitive Science: Essays in Honor of Hubert
L. Dreyfus, vol. 2 (Cambridge, MA: MIT Press), 17995.
Collins, S. (1982), Selfless Persons: Imagery and Thought in Theravada Buddhism
(Cambridge: Cambridge University Press).
Conee, E. (1994), Phenomenal Knowledge, Australasian Journal of Philosophy, 72:
13650.
Cooper, R., Shallice, T., and Farringdon, J. (1995), Symbolic and Continuous
Processes in the Automatic Selection of Actions, in J. Hallam (ed.), Hybrid
Problems, Hybrid Solutions (Oxford: IOS Press), 2737.
Copeland, B. J. (1993), Artificial Intelligence: A Philosophical Introduction (Oxford:
Blackwell).
Copeland, B. J., and Sylvan, R. (1999), Beyond the Universal Turing Machine,
Australasian Journal of Philosophy, 77: 4667.
Craik, K. J. W. (1943), The Nature of Explanation (Cambridge: Cambridge University
Press).
Crane, T. (1998), Intentionality as the Mark of the Mental, in Anthony OHear
(ed.), Contemporary Issues in the Philosophy of Mind (Cambridge: Cambridge
University Press), 22951.
(2006), Is There a Perceptual Relation?, T. S. Gendler and J. Hawthorne (eds),
Perceptual Experience (Oxford: Oxford University Press).
Crane, T., and Mellor, D. H., (1990), There is No Question of Physicalism, Mind,
99: 185206.
Crick, F. (1994), The Astonishing Hypothesis: The Scientific Search for the Soul
(New York: Charles Scribners Sons).
Crick, F. H. C., and Koch, C. (1990), Towards a Neurobiological Theory of
Consciousness, Seminars in Neuroscience, 2: 26375.
Cummins, R. (1989), Meaning and Mental Representation (Cambridge, MA: MIT Press).
Cussins, A. (1990), The Connectionist Construction of Concepts, in M. A. Boden
(ed.), The Philosophy of Artificial Intelligence (Oxford: Oxford University Press),
368440.
Dale, K., and Husbands, P. (2010), The Evolution of Reaction-Diusion Controllers
for Minimally Cognitive Agents, Artificial Life, 16 (1): 119.
Daly, C. (1997), What are Physical Properties?, Pacific Philosophical Quarterly, 79 (3):
196217.
Damasio, A. R. (1994), Descartes Error: Emotion, Reason and the Human Brain
(New York: Putnam).
Davidson, D. (1970), Mental Events, in L. Foster and J. Swanson (eds), Experience
and Theory (London: Duckworth), 79101; reprinted in Davidson (1980), 20725.
Page numbers in this chapter refer to the reprint.
(1974), Special Sciences or The disunity of Science as a Working Hypothesis,
in Synthese, 28: 97115; reprinted in Fodor (1981).
347
Bibliography
(1975), The Language of Thought (Cambridge, MA: Harvard University Press).

(1980), Essays on Actions and Events (Oxford: Oxford University Press).
(1980) [1969], On the Individuation of Events, in his Essays on Actions and Events
(Oxford: Clarendon Press).
(1981), RePresentations: Philosophical Essays on the Foundations of Cognitive Science
(Cambridge, MA: Bradford Books, MIT Press).
(1982), Rational Animals, in Dialectica, 36: 31727.
(1984a), First-Person Authority, in Dialectica, 38: 10111; reprinted in Subjective,
Intersubjective, Objective (2001).
(1984b), Inquiries into Truth and Interpretation (Oxford: Oxford University Press).
(1985), Fodors Guide to Mental Representation: The Intelligent Aunties
Vade-Mecum, Mind, 94: 76100.
(1986a), A Coherence Theory of Truth and Knowledge in Lepore (1986).
(1986b), The Myth of the Subjective, in M. Krausz (ed.), Relativism: Interpretation
and Confrontation (Indiana: Notre Dame); reprinted in Davidson (2001).
(1987a), Knowing Ones Own Mind, in Proceedings from the American Philosophical
Association, 61: 44158; reprinted in Q. Cassam (ed.), Self-Knowledge (Oxford:
Oxford University Press); and in Davidson (2001).
(1987b), Psychosemantics: The Problem of Meaning in the Philosophy of Mind
(1990), A Theory of Content and Other Essays (Cambridge, MA: Bradford Books,
MIT Press).
(1991), A Modal Argument for Narrow Content, Journal of Philosophy, 88: 526.
(1993), Thinking Causes, in J. Heil and A. Mele (eds), Mental Causation (Oxford:
Oxford University Press), 317.
(1994), The Elm and the Expert (Cambridge, MA: MIT Press).
(2001), Subjectivity, Intersubjectivity, Objectivity (Oxford: Oxford University Press).
Davies, M. (1983), Function in Perception, Australasian Journal of Philosophy, 61:
40926.
(2000), Externalism and Armchair Knowledge, in P. Boghossian and C. Peacocke
(eds), New Essays on the A Priori (Oxford: Oxford University Press).
Davies, M., and Humphreys, G. (1993), Consciousness: Psychological and Philosophical
Essays (Oxford: Blackwell).
Davies, M., and Stone, T. (1995), Mental Simulation: Evaluations and Applications
De Jaegher, H., and Di Paolo, E. A. (2007), Participatory Sense-Making: An Enactive
Approach to Social Cognition, Phenomenology and Cognitive Science, 6 (4): 485507.
Denne , D. C. (1969), Content and Consciousness: An Analysis of Mental Phenomena
(London: Routledge & Kegan Paul).
(1971), Intentional Systems, Journal of Philosophy, 68: 87106.
(1978), Why Not the Whole Iguana?, Behavioral and Brain Sciences, 1: 1034.
(1981a), Brainstorms (Brighton: Harvester).
(1981b), True Believers: The Intentional Strategy and Why It Works, in
A. F. Heath (ed.), Scientific Explanation: Papers Based on Herbert Spencer Lectures
Given in the University of Oxford (Oxford: Oxford University Press), 5375.
(1984), Elbow Room: The Varieties of Free Will Worth Wanting (Cambridge, MA:
MIT Press).
348
Bibliography
(1987a), The Intentional Stance (Cambridge, MA: MIT Press).

(1987b), True Believers: The Intentional Strategy and Why It Works, in his The
Intentional Stance (Cambridge MA: MIT Press), 1335; reprinted in Rosenthal
(1991), 33950.
(1988), Quining Qualia, in A. Marcel and E. Bisiach (eds), Consciousness in
Contemporary Science (Oxford: Oxford University Press), 4277.
(1991a), Consciousness Explained (New York: Penguin Books).
(1991b), Real Pa erns, Journal of Philosophy, 88: 2751.
(1994), Denne , Daniel, C., in Gu enplan (1994).
(1995), The Unimagined Preposterousness of Zombies: Commentary on Moody,
Flanagan, and Polger, Journal of Consciousness Studies, 2: 3226.
(1996), The Case for Rorts, in R. Brandom (ed.), Rorty and His Critics (Oxford:
Blackwell), 91101.
(1997), Kinds of Minds: Towards an Understanding of Consciousness (New York:
Basic Books).
(2001), Consciousness Explained (New York: Li le Brown).
(2003), Whos On First? Heterophenomenology Explained, Journal of Conscious-
ness Studies, 10: 1930.
(2006), Sweet Dreams: Philosophical Obstacles to a Science of Consciousness
Descartes, R. (1637/1985), Discourse on the Method, in J. Co ingham, R. Stootho
and D. Murdoch (trans. and eds), The Philosophical Writings of Descartes, vol. 1
(Cambridge: Cambridge University Press), 11151.
(1642/1984), Meditations on First Philosophy, in J. Co ingham, R. Stootho
and D. Murdoch (trans. and eds), The Philosophical Writings of Descartes, vol. 2
(Cambridge: Cambridge University Press), 1262.
(1994), in J. Co ingham, R. Stootho and D. Murdoch (trans. and eds), The
Philosophical Writings of Descartes, vol. 2 (Cambridge: Cambridge University
Press), 275.
Devi , M. (1996), Coming to Our Senses (Cambridge: Cambridge University Press).
Devi , M., and Sterelny, K. (1987/99), Language and Reality (Cambridge, MA: MIT
Press).
Di Paolo, E. A. (2009), Extended Life, Topoi, 28: 921.
Dienes, Z., and Perner, J. (1999), A Theory of Implicit and Explicit Knowledge,
Behavioral and Brain Sciences, 22: 735808.
(2007), The Cold Control Theory of Hypnosis, in G. Jamieson (ed.), Hypnosis and
Conscious States: The Cognitive Neuroscience Perspective (Oxford: Oxford University
Press), 293314.
Dksterhuis, E. J. (1961), The Mechanization of the World-Picture (Oxford: Clarendon).
Dowell, J. L. (2006), Formulating the Thesis of Physicalism, Philosophical Studies,
131 (1): 123.
(2006), Physical: Empirical not Metaphysical, Philosophical Studies, 131 (1): 2560.
Dretske, F. (1980), The Intentionality of Cognitive States, in D. Rosenthal (ed.), The
Nature of Mind (Oxford: Oxford University Press, 1990).
(1981), Knowledge and the Flow of Information (Cambridge, MA: Bradford Books,
MIT Press).
(1985a), Constraints and Meaning, Linguistics and Philosophy, 8 (1): 912.
349
Bibliography
(1985b), Machines and the Mental, Proceedings and Addresses of the American
Philosophical Association, 59: 2333.
(1986), Misrepresentation, in R. Bogdan (ed.), Belief (Oxford: Oxford University
Press).
(1988), Explaining Behavior: Reasons in a World of Causes (Cambridge, MA: MIT Press).
(1993), Conscious Experience, Mind, 102: 26383.
(1994), Dierences That Make No Dierence, Philosophical Topics, 22 (12): 4158.
(1995), Naturalizing the Mind (Cambridge: MA: MIT Press).
Dreyfus, H. L. (1967), Why Computers Must Have Bodies in Order to be
Intelligent, Review of Metaphysics, 21: 1332.
(1972), What Computers Cant Do: A Critique of Artificial Reason (New York:
Harper & Row).
(2003), Existentialist Phenomenology and the Brave New World of The
Matrix, The Harvard Review of Philosophy, 11 (Fall): 1831. Amended version of
Dreyfus and Dreyfus (2002).
Dreyfus, H. L., and Dreyfus, Stephen (2002), The Brave New World of the
Matrix, 20 November. Available at h p://whatisthematrix.warnerbros.com. For
an amended version see Dreyfus (2003).
Dreyfus, H. L., and Dreyfus, Stuart E. (1988), Making a Mind Versus Modelling
the Brain: Artificial Intelligence Back at a Branch Point, in S. Graubard (ed.),
The Artificial Intelligence Debate: False Starts, Real Foundations (Cambridge, MA:
MIT Press), 1543.
Eccles, J. (1987): Brain and Mind: Two or One?, in C. Blakemore and S. Greenfields
(eds), Mindwaves (Oxford: Blackwell).
Eccles, J., and Popper, K. (1977), The Self and Its Brain (New York: Springer).
Efron, A. (1992): Residual Assymetric Dualism: A Theory of Mind-Body Relations,
Journal of Mind and Behaviour, 13: 11336.
Elman, J. L. (1990), Finding Structure in Time, Cognitive Science, 14: 179212.
(1993), Learning and Development in Neural Networks: The Importance of
Starting Small, Cognition, 48: 7199.
Elster, J. (1999), Alchemies of the Mind: Rationality and the Emotions (Cambridge:
Elugardo, R. (2002), The Predicate View of Proper Names, in G. Preyer and
G. Peter (eds), Logical Form and Language (Oxford: Clarendon Press).
Enc, B. (1982), Intentional States of Mechanical Devices, Mind, 91: 16182.
Engelbart, D. C. (1962), Augmenting Human Intellect: A Conceptual Framework. Report
No. AFOSR 3233, prepared for the Air Force Oce of Scientific Research (Menlo
Park, CA: Stanford Research Institute); reprinted, abridged, in R. Packer and
K. Jordan (eds), Multimedia: From Wagner to Virtual Reality (London: W. W. Norton,
2001), 6490.
Evans, G. (1982), The Varieties of Reference, J. McDowell (ed.) (Oxford: Oxford
University Press).
Evere , A., and Hofweber, T., Empty Names, Fiction and the Puzzles of Non-existence
(Stanford: CSLI).
Ezquerro, J., and Vicente, A. (2000) Explanatory Exclusion, Over-Determination,
and the Mind-Body Problem. Paper read at the Twentieth World Congress of
Philosophy at Boston.
350
Bibliography
Feigl, H. (1958), The Mental and the Physical, in H. Feigl, G. Maxwell and
M. Scriven (eds), Concepts, Theories and the Mind-Body Problem: Minnesota Studies
in the Philosophy of Science, vol. 2 (Minneapolis: University of Minnesota Press),
370497; republished in 1967 with a new postscript, preface to the postscript and
additional bibliography.
(1978), Mental Representation, Erkenntnis, 13; reprinted with postscripts in
N. Block (ed.), Readings in Philosophy of Psychology, vol. 2 (London: Methuen).
(1992), Physicalism, in J. Earman (ed.), Inference, Explanation and Other
Frustrations (Berkeley: University of California Press), 27192.
Fish, W. (2009), Perception, Hallucination, and Illusion (Oxford: Oxford University
Press).
Fitch, T. (2007), Nano-Intentionality, Biology & Philosophy, 23: 15777.
Flanagan, O. (1991), The Science of the Mind (Cambridge, MA: MIT Press).
(1992), Consciousness Reconsidered (Cambridge MA: MIT Press).
Fodor, J. A. (1968), Psychological Explanation: An Introduction to the Philosophy of
Psychology (New York: Random House).
(1975), The Language of Thought (New York: Thomas Crowell).
(1980), Methodological Solipsism Considered as a Research Strategy in Cognitive
Psychology, Behavioral and Brain Sciences, 3: 6373.
(1983), The Modularity of Mind: An Essay in Faculty Psychology (Cambridge, MA:
MIT Press).
(1984), Semantics, Wisconsin Style, Synthese, 59: 231350; and in his A Theory
of Content (Cambridge, MA: MIT Press, 1990), 3149.
(1986a), Banish Dis-Content, in J. Bu erfield (ed.), Language, Mind and Logic
(1986b), Why Paramecia Dont Have Mental Representations, in P. French,
T. Uehling, Jr., and H. We stein (eds), Midwest Studies in Philosophy, vol. 10
(Minneapolis: University of Minnesota Press), 323.
(1987), Psychosemantics: The Problem of Meaning in the Philosophy of Mind
(1988), Connectionism and Cognitive Architecture: A Critical Analysis,
Cognition, 28: 371.
(1990a), Fodors Guide to Mental Representation, in Fodor (1990c), 329.
(1990b), Making Mind Ma er More, in Fodor (1990c), 13759.
(1990c), A Theory of Content and Other Essays (Cambridge, MA: MIT Press).
(1990d), A Theory of Content, I: The Problem, in Fodor (1990c), 5187.
(1990e), A Theory of Content, II: The Theory, in Fodor (1990c), 89136.
(1991a), A Modal Argument for Narrow Content, Journal of Philosophy, 88: 526.
(1991b), Replies, in B. Loewer and G. Rey (eds), Meaning in Mind: Fodor and His
Critics (Oxford: Blackwell), 255319.
(1998), Concepts: Where Cognitive Science Went Wrong (Cambridge, MA: MIT
Press).
(2000), The Mind Doesnt Work That Way: The Scope and Limits of Computational
Psychology (Cambridge, MA: MIT Press).
(2009), Where is My Mind?, London Review of Books, 31 (3), 12 February.
Fodor, J. A., and Lepore, E. (1992), Holism: A Shoppers Guide (Cambridge, MA: MIT
Press).
351
Bibliography
Fodor, J. A., and McLaughlin, B. P. (1990), Connectionism and the Problem of

Systematicity: Why Smolenskys Solution Doesnt Work, Cognition, 35: 183204.
Fodor, J. A., and Pylyshyn, Z. W. (1981), How Direct is Visual Perception?: Some
Reflections on Gibsons Ecological Approach, Cognition, 9: 13996.
Foster, J. (1989), A Defence of Dualism, in J. Smythies and J. Belo (eds), The Case
for Dualism (Charlo esville: University of Virginia Press).
(1991), The Immaterial Self: A Defence of the Cartesian Dualist Conception of Mind
(London: Routledge).
(1996), The Immaterial Self (London: Routledge).
Frege, G. (1953) [1884], The Foundations of Arithmetic, trans. J. L. Austin (Oxford:
Blackwell).
(1960) [1892], On Sense and Reference, in P. T. Geach and M. Black (eds),
Translations from the Philosophical Writings of Golob Frege, 2nd edn (Oxford:
Blackwell).
Freud, S. (1917/1991), Introductory Lectures on Psychoanalysis (Harmondsworth:
Penguin).
Frith, C. D. (2007), Making Up the Mind: How the Brain Creates Our Mental World
Frith, C. D., and Frith, U. (2000), The Physiological Basis of Theory of Mind:
Functional Neuroimaging Studies, in S. Baron-Cohen, H. Tager-Flusberg and
D. J. Cohen (eds), Understanding Other Minds: Perspectives from Developmental
Cognitive Neuroscience, 2nd edn (Oxford: Oxford University Press), 33456.
Frith, C. D., Perry, R., and Lumer, E. (1999), The Neural Correlates of Conscious
Experience: An Experimental Framework, Trends in Cognitive Science, 3: 10514.
Gallagher, S. (2008), Direct Perception in the Intersubjective Context, Consciousness
and Cognition, 17: 53543.
Gallagher, S., and Zahavi, D. (2007), The Phenomenological Mind (London:
Routledge).
Gallese, V. (2006), Intentional A unement: A Neurophysiological Perspective on
Social Cognition and Its Disruption in Autism, Brain Research, 1079: 1524.
Gallistel, C. (1990), The Organization of Learning (Cambridge, MA: MIT Press).
(2008), Learning and Representation, in R. Menzel (ed.), Learning Theory and
Behavior, vol. 1 of Learning and Memory: A Comprehensive Reference, 4 vols. (J. Byrne,
ed.) (Oxford: Elsevier), 22742.
Garre , B. (1998), Personal Identity and Self-Consciousness (London: Routledge).
Gates, G. (1996), The Price of Information, Synthese, 107 (3): 32547.
Geach, P. T. (1980), Some Remarks on Representations, Behavioral and Brain
Sciences, 3: 801.
Geertz, C. (1973), Thick Description: Toward an Interpretive Theory of Culture,
in C. Geertz (ed.), The Interpretation of Cultures: Selected Essays (New York: Basic
Books), 332.
Gibb, S. (2006), Why Davidson is Not a Property Epiphenomenalist, International
Journal of Philosophical Studies 14 (3): 40722.
Gibson, E. J., and Walk, R. D. (1960), The Visual Cli , Scientific American, 202
(April): 6471.
Gibson, J. J. (1950), The Perception of the Visual World (Cambridge, MA: Riverside
Press).
352
Bibliography
(1966), The Senses Considered as Perceptual Systems (Westport, CT: Greenwood Press).
(1977), The Theory of Aordances, in R. Shaw and J. Bransford (eds), Perceiving,
Acting, and Knowing: Toward an Ecological Psychology (Hillsdale, NJ: Lawrence
Erlbaum), 6782.
Gille , C. (2001), Does the Argument from Realization Generalize? Responses
to Kim, Southern Journal of Philosophy, 39 (1): 7998.
Gillet, C., and Loewer, B. (2001), Physicalism and Its Discontents (Cambridge:
Gille , C., and Rives, B. (2005), The Non-Existence of Determinables: Or a World
of Absolute Determinates as a Default Hypothesis, Nos, 39: 483504.
Glenberg, A., and Adams, F. (1978), Type I Rehearsal and Recognition, Journal
of Verbal Learning and Verbal Behavior, 17: 45563.
Godfrey-Smith, P. (1994a), A Continuum of Semantic Optimism, in Stich and
Warfield (1994), 25977.
(1994b), Spencer and Dewey on Life and Mind, in R. A. Brooks and P. Maes
(eds), Artificial Life IV (Cambridge, MA: MIT Press), 809.
Gois, I. (2007), On a Misconception about Consciousness, unpublished Ph.D.
thesis, University of London.
Goldberg, S. (2003), On Our Alleged A Priori Knowledge That Water Exists,
Analysis, 63: 3841.
(2007), Anti-Individualism, Content Preservation, and Discursive Justification,
Nos, 41: 178203.
Goodale, M. A., and Milner, A. D. (1992), Separate Visual Pathways for Perception
and Action, Trends in Neuroscience, 13: 2023.
Grau, C. (ed.) (2005), Philosophical Essays on the Matrix (New York: Oxford University
Press).
Green, C. (2003), The Lost Cause (Oxford: Forum).
Greenfield, P. M., and Bruner, J. S. (1969), Culture and Cognitive Growth, in
D. A. Goslin (ed.), Handbook of Socialization Theory and Research (Chicago: Rand
McNally), 63354.
Grice, H. P. (1961/1965), The Causal Theory of Perception, in R. Swartz (ed.),
Perceiving, Sensing, and Knowing (New York: Doubleday).
Grin, D. R. (1978), Prospects for a Cognitive Ethology, Behavioral and Brain
Sciences, 4: 52738.
(1984), Animal Thinking (Cambridge, MA: Harvard University Press).
Griths, P. (1997), What Emotions Really are: The Problem of Psychological Categories
(Chicago: Chicago University Press).
Grush, R. (2004), The Emulation Theory of Representation: Motor Control, Imagery,
and Perception, Behavioral and Brain Sciences, 27: 377442.
Gulick, R. V. (1992), Three Bad Arguments for Intentional Property Epipheno-
menalism, Erkenntnis 36 (3): 31132.
Gu enplan, S. (1994), A Companion to the Philosophy of Mind (Oxford: Blackwell).
Harman, G. (1990), The Intrinsic Quality of Experience, in J. E. Tomberlin
(ed.), Philosophical Perspectives, 4: 3152; reprinted in N. Block, O. Flanagan and
G. Gzeldere (eds) (1997), The Nature of Consciousness, 66375.
Harnish, M. (2002), Minds, Brains, Computers: An Historical Introduction to the
Foundations of Cognitive Science (Oxford: Blackwell).
353
Bibliography
Hart, W. D. (1994), Dualism, in S. Gu enplan (ed.), A Companion to the Philosophy

of Mind (Oxford: Blackwell).
Ha iangadi, A. (2007), Oughts and Thoughts (Oxford: Oxford University Press).
Haugeland, J. (1983), Weak Supervenience, American Philosophical Quarterly, 19:
93103.
(1985), Artificial Intelligence: The Very Idea (Cambridge, MA: MIT Press).
(1995/1998), Mind Embodied and Embedded, in Having Thought: Essays in
the Metaphysics of Mind (Cambridge, MA: Harvard University Press), Chapter 9,
20737.
Hawthorne, J. (2002), Blocking Definitions of Materialism, Philosophical Studies,
110 (2): 10313.
(2003), From an Ontological Point of View (Oxford: Oxford University Press).
(2007), Cartesian Dualism, in P. van Inwagen and D. Zimmerman (eds), Persons
Human and Divine (Oxford: Oxford University Press).
Heil, J. (1988), Privileged Access, Mind, 97: 23851.
(2003), From an Ontological Point of View (Oxford: Oxford University Press).
(2008), Anomalous Monism, in H. Dyke (ed.), From Truth to Reality: New Essays
in Metaphysics (London: Routledge), 8598.
Heller, M. (1990), The Ontology of Physical Objects: Four-Dimensional Hunks of
Maer (Cambridge: Cambridge University Press).
Hempel, C. (1963), Reasons and Covering Laws in Historical Explanation, in
S. Hook (ed.), Philosophy and History: A Symposium (New York: New York
University Press).
(1965), Aspects of Scientific Explanation (New York: Free Press).
(1966), Philosophy of Natural Science, Prentice-Hall Foundations of Philosophy
series (Englewood Clis, NJ: Prentice-Hall).
(1996), Laws and Their Role in Scientific Explanation, in C. Hempel (ed.),
Philosophy of Natural Science (Englewood Clis: Prentice Hall).
Hempel, C., and Oppenheim, P. (1953), The Logic of Explanation, in H. Feigl and
M. Brodbek (eds), Readings in the Philosophy of Science (New York: Appleton).
Herbert, R. T. (1998), Dualism/Materialism, Philosophical Quarterly, 48: 15975.
Hess, P. (1981), Actions, Reasons, and Humean Causes, Analysis, 40: 7781.
Himma, K. E. (2005), When a Problem for All is a Problem for None: Substance
Dualism, Physicalism and the Mind-Body Problem, Australian Journal of Philo-
sophy, 42 (2): 8192.
Hinton, G. E. (1980), Inferring the Meaning of Direct Perception, Behavioral and
Brain Sciences, 3: 3878.
(1990), Representing Part-Whole Hierarchies in Connectionist Networks,
Artificial Intelligence, 46: 4775. Special issue on Connectionist Symbol Processing.
Hirsch, E. (1982), The Concept of Identity (Oxford: Oxford University Press).
Hodgson, D. (1991), The Mind Maers (Oxford: Oxford University Press).
Honderich, T. (1981): Psychophysical Law-Like Connections and Their Problems,
Inquiry, 24: 277303.
(1982), The Argument for Anomalous Monism, Analysis, 42: 5964.
(1983), Anomalous Monism: Reply to Smith, Analysis, 43: 1479.
(1984), Smith and the Champion of Mauve, Analysis, 44: 869.
(2004), On Consciousness (Edinburgh: Edinburgh University Press).
354
Bibliography
Horgan, T. (1978), The Case Against Events, The Philosophical Review, 87 (1): 2847.
(1983), Supervenience and Microphysics, Pacific Philosophical Quarterly, 63:
2943.
(1993), From Supervenience to Superdupervenience: Meeting the Demands
of a Material World, Mind, 102 (408): 55586.
Horgan, T., and Woodward, J. (1985), Folk Psychology is Here to Stay, Philosophical
Review, 94: 197225.
Hornsby, J. (1980), Actions (London: Routledge).
Horwich, P. (1998), Meaning (Oxford: Oxford University Press).
(2005), Reflections on Meaning (Oxford: Oxford University Press).
(2006), The Value of Truth, Nos, 40: 34760.
Hudson, H. (2001), A Materialist Metaphysics of the Human Person (Ithaca, NY:
Cornell University Press).
(2007), I Am Not an Animal!, in P. van Inwagen and D. Zimmerman (eds),
Persons: Human and Divine (Oxford: Clarendon Press).
Humphreys, N. (1982), Consciousness Regained (Oxford: Oxford University Press).
Hunt, E. B. (1962), Concept Learning: An Information Processing Problem (New York:
Wiley).
Hurley, S. L. (1998), Consciousness in Action (Cambridge, MA: Harvard University
Press).
(forthcoming), Varieties of Externalism, in R. Menary (ed.), The Extended Mind
(Aldershot: Ashgate).
Husbands, P., Smith, T., Jakobi, N., and OShea, M. (1998), Be er Living through
Chemistry: Evolving Gas Nets for Robot Control, Connection Science, 10:
185210.
Hutchins, E. (1995), Cognition in the Wild (Cambridge, MA: MIT Press).
Hu o D. D. (2009), Mental Representation and Consciousness, in W. Banks
(ed.), Encyclopedia of Consciousness, vol. 2 (London: Elsevier), 1932.
Hu o, D. D., and Myin, E. (forthcoming), Radicalizing Enactivism (Cambridge, MA:
MIT Press).
Irwin, W. (ed.) (2002), The Matrix and Philosophy: Welcome to the Desert of the Real
(Chicago: Open Court).
Jackendo, R. (1987), Consciousness and the Computational Mind (Cambridge, MA:
MIT Press).
Jackson, F. (1982), Epiphenomenal Qualia, Philosophical Quarterly, 32: 12736.
(1986), What Mary Didnt Know, Journal of Philosophy, 83: 2915; reprinted in
his Mind, Method and Conditionals (London and New York: Routledge, 1998),
705; and in Rosenthal (1991), 3924.
(1993), Armchair Metaphysics, in J. Hawthorne and M. Michael (eds), Philosophy
in Mind (Amsterdam: Kluwer).
(1998), From Metaphysics to Ethics (Oxford: Oxford University Press).
(2003), Mind and Illusion, in A. OHear (ed.), Minds and Persons (Royal Institute
of Philosophy Supplement 53), 5171.
(2006), On Ensuring That Physicalism is Not a Dual A ribute Theory in Sheeps
Clothing, Philosophical Studies, 131 (1): 22749.
Jackson, F., and Pe it, P. (1990), Program Explanation: A General Perspective,
Analysis, 50: 10717.
355
Bibliography
Johnston, M. (1987), Human Beings, Journal of Philosophy, 84: 5983.

Jonas, H. (1966), The Phenomenon of Life: Toward a Philosophical Biology (New York:
Harper Collins); reprinted, Evanston, IL: Northwestern University Press (2001).
Kallestrup, J. (2006), The Causal Exclusion Argument, Philosophical Studies, 131 (2):
45985.
Kaplan, D. (1969), Quantifying in, in D. Davidson and J. Hintikka (eds), Words and
Objections (Dordrecht: Reidel).
(1989), Demonstratives, in J. Almog, J. Perry and H. We stein (eds), Themes
from Kaplan (Oxford: Oxford University Press).
Karmilo-Smith, A. (1992), Beyond Modularity: A Developmental Perspective on
Cognitive Science (London: MIT Press).
Kezer, F., and Schouten, M. (2007), Embedded Cognition and Mental Causation:
Se ing Empirical Boundaries on Metaphysics, Synthese, 158: 10925.
Kenny, A. (1989), The Metaphysics of Mind (Oxford: Clarendon Press).
(1994), Aquinas on Mind (London: Routledge).
Kim, J. (1969), Events and Their Descriptions: Some Considerations, in N. Rescher
(ed.), Essays in Honor of Carl G. Hempel (Dordrecht: Reidel).
(1973), Causation, Nomic Subsumption and the Concept of Event, Journal of
(1976), Events as Property Exemplifications, in M. Brand and D. Walton (eds),
Action Theory (Dordrecht: Reidel).
(1984), Epiphenomenal and Supervenient Causation, Midwest Studies in Philo-
sophy, 9: 25770.
(1987), Strong and Global Supervenience Revisited, Philosophy and
Phenomenological Research, 48: 31526.
(1988), Explanatory Realism, Causal Realism, and Explanatory Exclusion,
Midwest Studies in Philosophy, 12: 22540.
(1989a), Mechanism, Purpose, and Explanatory Exclusion, Philosophical Perspec-
tives, 3 (Philosophy of Mind and Action Theory).
(1989b), The Myth of Non-Reductive Materialism, Proceedings of the American
Philosophical Association, 63: 127.
(1990a), Explanatory Exclusion and the Problem of Mental Causation, in
E. Villanueva (ed.), Information, Semantics, and Epistemology (Oxford: Blackwell);
reprinted in MacDonald (ed.), Philosophy of Psychology: Debates on Psychological
Explanation (Oxford: Blackwell, 1995).
(1990b), Supervenience as a Philosophical Concept, Metaphilosophy, 21: 127.
(1993a), Can Supervenience and Non-Strict Laws Save Anomalous Monism?,
in J. Heil and A. Mele (eds), Mental Causation (Oxford: Clarendon).
(1993b), Concepts of Supervenience, in J. Kim (ed.), Supervenience and Mind
(1993c), Supervenience and Mind: Selected Philosophical Essays (Cambridge:
(1994), Explanatory Knowledge and Metaphysical Dependence, Philosophical
Issues, 5: 5169.
(1995), Mental Causation: What? Me Worry?, Philosophical Issues, 6: 123151.
(1998), Mind in a Physical World: An Essay on the Mind-Body Problem and Mental
Causation (Cambridge MA: MIT Press).
356
Bibliography
(2003), Lonely Souls: Causality and Substance Dualism, in T. OConnor and

D. Robb (eds), Philosophy of Mind: Contemporary Readings (London: Routledge).
(2005), Physicalism, Or Something Near Enough (Princeton: Princeton University
Press).
(2006), The Philosophy of Mind (Boulder: Westview Press).
Kirsh, D. (1991), Today the Earwig, Tomorrow Man?, Artificial Intelligence, 47: 16184.
Kitcher, P. (1984), In Defense of Intentional Psychology, Journal of Philosophy,
71: 89106.
Kolers, P. A. (1972), Aspects of Motion Perception (London: Pergamon Press).
Kolers, P. A., and Rosner, B. S. (1960), On Visual Masking (Metacontrast); Dichoptic
Observation, American Journal of Psychology, 73: 221.
Kriegel, U. (2005), Naturalizing Subjective Character, Philosophy and Phenomeno-
logical Research, 71: 2357.
(2009), Subjective Consciousness (Oxford: Oxford University Press).
Kripke, S. (1972), Naming and Necessity (Oxford: Blackwell).
(1972/80), Naming and Necessity (Cambridge: Harvard University Press).
(1982), Wigenstein on Rules and Private Language (Cambridge, MA: Harvard
University Press).
Ladyman, J., Ross, D., Spurre , D., and Collier, J. (2007), Everything Must Go:
Metaphysics Naturalized (Oxford: Oxford University Press).
Lahav, R., and Shanks, N. (1982): How to Be a Scientifically Respectable Property
Dualist, Journal of Mind and Behaviour, 13: 21132.
Lako, G., and Johnson, M. (1980), Metaphors We Live By (Chicago: University
of Chicago Press).
(1999), Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western
Thought (New York: Basic Books).
Larmer, R. (1986), Mind-Body Interactionism and the Conservation of Energy,
International Philosophical Quarterly, 26: 27785.
Latham, N. (2000), Chalmers on the Addition of Consciousness to the Physical
World, Philosophical Studies, 98: 6793.
Lawrence, M. (2005), Like a Splinter in Your Mind: The Philosophy Behind the Matrix
Trilogy (Oxford: Blackwell).
Lepore, E. (ed.) (1986), Truth and Interpretation: Perspectives on the Philosophy of
Donald Davidson (Oxford: Blackwell).
Lepore, E., and Loewer, B. (1987), Mind Ma ers, Journal of Philosophy, 84 (11):
63042.
Lepore E., and McLaughlin, B. (eds) (1985), Actions and Events: Perspectives on the
Philosophy of Donald Davidson (Oxford: Blackwell Publishers)
Le vin, J. Y., Maturana, H. R., McCulloch, W. S., and Pi s, W. H. (1959), What the
Frogs Eye Tells the Frogs Brain, Proceedings of the Institute of Radio Engineers, 47
(11): 194059.
Leuenberger, S. (2008), Ceteris Absentibus Physicalism, in D. W. Zimmerman (ed.),
Oxford Studies in Metaphysics, vol. 4 (Oxford: Oxford University Press), 14570.
Levine, J. (1983), Materialism and Qualia: The Explanatory Gap, Pacific Philosophi-
cal Quarterly, 64: 35461.
(1993), On leaving Out What Its Like, in M. Davies and G. Humphreys (eds),
Consciousness: Psychological and Philosophical Essays (Oxford: Blackwell).
357
Bibliography
(1998), Conceivability and the Metaphysics of Mind, Nos, 32 (4): 44980.

(2001), Purple Haze: The Puzzle of Conscious Experience (Oxford: Oxford University
Press; Cambridge, MA: MIT Press).
(2008), Review of Ignorance and Imagination by Daniel Stoljar, Mind, 117:
22831.
Lewis, D. (1980), Mad Pain and Martian Pain, in N. Block (ed.), Readings in
Philosophy of Psychology, vol. 1 (London: Methuen), 21622.
(1983b), Postscript to Mad Pain and Martian Pain, in his Philosophical
Papers, vol. 1 (Oxford: Oxford University Press), 12232.
(1986), Against Structural Universals, Australasian Journal of Philosophy, 64,
2546.
(1990), What Experience Teaches, in W. Lycan (ed.), Mind and Cognition
(Oxford: Blackwell), 499519; reprinted in Ned Block, Owen Flanagan and
Gven Gzeldere (eds) (1997), The Nature of Consciousness, 57995.
(1994), David Lewis, in S. Gu enplan (ed.), A Companion to the Philosophy of
Mind (Oxford: Blackwell).
Libet, B. (1985a), Subjective Antedating of a Sensory Experience and Mind-Brain
Theories, Journal of Theoretical Biology, 114: 56370.
(1985b), Unconscious Cerebral Initiative and the Role of Conscious Will in
Voluntary Action, Behavioral and Brain Sciences, 8: 52966.
(ed.) (1999), The Volitional Brain: Towards a Neuroscience of Free Will. Special issue
of the Journal of Consciousness Studies, 6 (89) (AugustSeptember).
Libet, B., Wright, E. W., Feinstein, B., and Pearl, D. K. (1979), Subjective Referral of
the Timing for a Conscious Sensory Experience, Brain, 102: 193224.
Loar, B. (1981), Mind and Meaning (Cambridge: Cambridge University Press).
(1987), Social Content and Psychological Content, in R. Grimm and D. Merrill
(eds), Contents of Thought: Proceedings of the 1985 Oberlin Colloquium in Philosophy
(Tucson: University of Arizona).
(1990), Phenomenal States, Philosophical Perspectives, 4: 81108.
Locke, J. (1975) [1690], An Essay Concerning Human Understanding, P. H. Nidditch
(ed.), (Oxford: Clarendon Press).
Lockwood, M. (1989), Mind, Brain, and the Quantum (Oxford: Oxford University
Press).
Loewer, B. (1997), A Guide to Naturalizing Semantics, in B. Hale and C. Wright
(eds.), A Companion to the Philosophy of Language (Malden, MA: Blackwell),
10826.
Loewer, B., and Rey, G. (1991), Meaning in Mind: Fodor and His Critics (Oxford:
Blackwell).
Lombard, L. B. (1986), Events: A Metaphysical Study (London and Boston:
Routledge & Kegan Paul).
Lowe, E. J. (1989), Impredicative Identity Criteria and Davidsons Criterion of
Event Identity, Analysis, 49: 17881.
(1992), The Problem of Psychophysical Causation, Australasian Journal of
(1993), The Causal Autonomy of the Mental, Mind, 102: 62944.
(1995), Locke on Human Understanding (London and New York: Routledge).
(1996), Subjects of Experience (Cambridge: Cambridge University Press).
358
Bibliography
(1997), Objects and Criteria of Identity, in R. Hale and C. Wright (eds),

A Companion to the Philosophy of Language (Oxford: Blackwell).
(2000a), Causal Closure Principles and Emergentism, Philosophy, 75: 57185.
(2000b), An Introduction to the Philosophy of Mind (Cambridge: Cambridge
University Press).
(2006), Non-Cartesian Substance Dualism and the Problem of Mental Causation,
Erkenntnis, 65 (1): 523.
(2009), More Kinds of Being: A Further Study of Individuation, Identity and the Logic
of Sortal Terms (Malden, MA and Oxford: Wiley-Blackwell).
Lucas, J. R. (1961), Minds, Machines, and Godel, Philosophy, 36: 11227.
Ludlow, P. (1995), Externalism, Self-Knowledge, and the Prevalence of Slow-
Switching, Analysis, 55: 459.
(1997), On the Relevance of Slow Switching, Analysis, 57: 2856.
(1998), Social Externalism and Memory: A Problem?, in P. Ludlow and N. Martin
(eds), Externalism and Self-Knowledge (Stanford: CSLI).
Ludlow, P., Nagasawa, Y., and Stoljar, D. (2004), Theres something about Mary: Essays
on Phenomenal Consciousness and Frank Jacksons Knowledge Argument (Cambridge,
MA: MIT Press).
Lutz, A., and Thompson, E. (2003), Neurophenomenology: Integrating Subjective
Experience and Brain Dynamics in the Neuroscience of Consciousness, Journal
of Consciousness Studies, 10: 910, 3152.
Lycan, W. G. (1987), Consciousness (Cambridge, MA: MIT Press).
(1996), Consciousness and Experience (Cambridge, MA: MIT Press).
McCarthy, J., and Hayes, P. J. (1969), Some Philosophical Problems from the
Standpoint of Artificial Intelligence, in B. Meltzer and D. M. Michie (eds), Machine
Intelligence, vol. 4 (Edinburgh: Edinburgh University Press), 463502.
McCulloch, W. S., and Pi s, W. H. (1943), A Logical Calculus of the Ideas Immanent
in Nervous Activity, Bulletin of Mathematical Biophysics, 5: 11533.
Macdonald, C., and Macdonald G. (1986), Mental Causes and the Explanation of
Action, Philosophical Quarterly, 36: 14558; reprinted in L. Stevenson, R. Squires
and J. Haldane (eds), Mind, Causation and Action (Oxford: Basil Blackwell, 1986).
Page references in text to la er.
(1995), How to Be Psychologically Relevant, in C. Macdonald and G. Macdonald
(eds), Philosophy of Psychology (Oxford: Basil Blackwell).
(2006), The Metaphysics of Mental Causation, Journal of Philosophy, 103 (11):
53976.
Macdonald, G. (1992), The Nature of Naturalism, Proceedings of the Aristotelian
Society, Supplementary Volume 66: 22544.
McDowell, J. (1984), De Re Senses, in Wright, C. (ed.), Frege: Tradition and Influence
(1986), Singular Thought and the Extent of Inner Space, in P. Pe it and
J. McDowell (eds), Subject, Thought and Context (Oxford: Clarendon Press).
(1994), The Content of Perceptual Experience, The Philosophical Quarterly, 44
(175): 190205.
(1994b), Mind and World (Cambridge, MA: Harvard University Press).
(1977), On the Sense and Reference of a Proper Name, Mind, 86: 15985.
McGinn, C. (1982a), The Character of Mind (Oxford: Oxford university Press).
359
Bibliography
(1982b), The Structure of Content, in A. Woodfield (ed.), Thought and Object:

Essays on Intentionality (Oxford: Oxford University Press).
(1989), Can We Solve the Mind-Body Problem?, Mind, 98 (July): 34966, reprinted
in his (1991), The Problem of Consciousness (Oxford: Blackwell), 122; also reprinted
in N. Block, O. Flanagan and G. Gzeldere (eds), The Nature of Consciousness
(Cambridge, MA: MIT Press, 1997), 52942.
(1991), The Problem of Consciousness: Essay Towards a Resolution (Oxford:
Blackwell).
(1993), Consciousness and Cosmology: Hyperdualism Ventilated, in M. Davies
and G. Humphreys (eds), Consciousness: Psychological and Philosophical Essays
(2005), The Matrix of Dreams, in C. Grau (ed.), Philosophical Essays on the
Matrix (New York: Oxford University Press), 6270.
Mackie, D. (1999), Personal Identity and Dead People, Philosophical Studies, 95:
21942.
McKinsey, M. (1991), Anti-Individualism and Privileged Access, Analysis, 51: 916.
McLaughlin, B. (1981), Anomalous Monism and the Irreducibility of the Mental,
in Lepore and McLaughlin (1985).
(1992), The Rise and Fall of British Emergentism, in A. Beckermann, H. Flohr
and J. Kim (eds), Emergence or Reduction? (Berlin: De Gruyter).
(2001), Physicalism and Alternatives, in N. J. Smelslser and P. B. Baltes (eds),
International Encyclopedia of the Social and Behavioral Sciences (Oxford: Pergamon),
114227.
McLaughlin, B., and Tye, M. (1998), Is Content-Externalism Compatible with
Privileged Access?, Philosophical Review, 107: 34980.
Majors, B., and Sawyer, S. (2005), The Epistemological Argument for Content
Externalism, Philosophical Perspectives, 19: 25780.
(2007), Internal Accessibility and the Opacity of Mental Content, in S. Goldberg
(ed.), Internalism and Externalism in Semantics and Epistemology (Oxford: Oxford
University Press).
Marr, D. C. (1982), Vision: A Computational Investigation into the Human Representation
and Processing of Visual Information (San Francisco: Freeman).
Marras, A. (1998), Kims Principle of Explanatory Exclusion, Australasian Journal
of Philosophy, 76 (3): 43951.
(2007), Kims Supervenience Argument and Nonreductive Physicalism,
Erkenntnis, 66 (3): 30527.
Marras, A., and Juhani Yli-Vakkuri (2008), The Supervenience Argument:
Kims Challenge to Nonreductive Physicalism, in S. Gozzano and F. Orilia (eds),
Tropes, Universals and the Philosophy of Mind: Essays at the Boundary of Ontology
and Philosophical Psychology (Frankfurt: Ontos Verlag).
Martin, M. G. F. (2002), Transparency of Experience, Mind and Language, 17:
376425.
(2006), On Being Alienated, in T. S. Gendler and J. Hawthorne (eds), Perceptual
Experience (Oxford: Oxford University Press), 354410.
Martin, R., and Barresi, J. (eds) (2003), Personal Identity (Oxford: Blackwell).
Ma hen, M. (2005), Seeing, Doing, and Knowing: A Philosophical Theory of Sense
Perception (Oxford: Oxford University Press).
360
Bibliography
Maturana, H. R., and Varela, F. J. (1980), Autopoiesis and Cognition: The Realization
of the Living (Boston: Reidel).
Mawson, T. (2005), Belief in God (Oxford: Oxford University Press).
May, L., Friedman, M., and Clark, A. J. (eds) (1996), Minds and Morals: Essays on
Cognitive Science and Ethics (Cambridge, MA: MIT Press).
Mellor, D. H. (1995), The Facts of Causation (London and New York: Routledge).
Melnyk A. (1997), How To Keep The Physical in Physicalism, Journal of
(2003), A Physicalist Manifesto: Thoroughly Modern Materialism (Cambridge:
Menary, R. (2007), Cognitive Integration: Mind and Cognition Unbounded (Basingstoke:
Palgrave Macmillan).
(ed.) (forthcoming), The Extended Mind (Cambridge, MA: MIT Press).
Menzies, P. (2003), The Causal Ecacy of Mental States, in S. Walter and H.-D.
Heckmann (eds), Physicalism and Mental Causation (Exeter: Imprint Academic).
Merleau-Ponty, M. (1962), Phenomenology of Perception (London: Routledge & Kegan
Paul).
Metzinger, T. (2003) Being No-One: The Self-Model Theory of Subjectivity (Cambridge,
MA: MIT Press).
(ed.) (1995), Conscious Experience (Paderborn: Ferdinand Schningh).
Millikan, R. (1984), Language, Thought and Other Biological Categories: New Foundations
for Realism (Cambridge MA: MIT Press).
(1986), Thoughts without Laws, The Philosophical Review, 95: 4780; and in her
White Queen Psychology and Other Essays for Alice (Cambridge, MA: MIT Press,
1993), 5182.
(1989), Biosemantics, The Journal of Philosophy, 86: 28197; and in her White
Queen Psychology and Other Essays for Alice (Cambridge, MA: MIT Press, 1993),
83101.
(1990), Truth Rules, Hoverflies, and the Kripke-Wi genstein Paradox, The
Philosophical Review, 99: 32353; and in her White Queen Psychology and Other
Essays for Alice (Cambridge, MA: MIT Press, 1993), 21139.
(2000), On Clear and Confused Ideas (Cambridge University Press).
Mills, E. (1996), Interaction and Overdetermination, American Philosophical
(1997), Interactionism and Physicality, Ratio, 10: 16983.
Milner, A., and Goodale, M. (1995), The Visual Brain in Action (Oxford: Oxford
University Press).
Minsky, M. L. (1965), Ma er, Mind, and Models, Proceedings of the International
Federation of Information Processing Congress, 1: 459 (Washington, DC: Spartan).
(1985), The Society of Mind (New York: Simon & Schuster).
(2006), The Emotion Machine (New York: Pantheon).
Montero, B. (1999), The Body Problem, Nos, 33 (2): 183200.
(2006), Physicalism in an Infinitely Decomposable World, Erkentnis, 64 (2): 177191.
Montero, B., and Papineau, D. (2005), A Defense of the Via Negativa Argument
for Physicalism, Analysis, 65 (3): 2337.
Moore, G. E. (1942), A Reply to My Critics, in P. A. Schilpp (ed.), The Philosophy
of G. E. Moore (La Salle, IL: Open Court), 535677.
361
Bibliography
Morris, M. (1991), Why There are No Mental Representations, Minds and Machines,
1: 130.
(1992), The Good and the True (Oxford: Clarendon Press).
Moya, C. (1990), The Philosophy of Action (Cambridge: Polity Press).
Nagel, T. (1974), What is It Like to Be a Bat?, Philosophical Review, 83 (4): 43550;
reprinted in Rosenthal (1991), 4228.
(1979) [1971], Brain Bisection and the Unity of Consciousness, in his Mortal
Questions (Cambridge: Cambridge University Press).
Neander, K. (1998), The Division of Phenomenal Labour: A Problem for Represen-
tational Theories of Consciousness, in J. E. Tomberlin (ed.), Philosophical Perspec-
tives, 12 (Boston, MA, and Oxford: Blackwell), 41134.
(2004), Teleological Theories of Mental Content, in Stanford Encyclopedia of
Philosophy. Available at h p://plato.stanford.edu/.
Nemirow, L. (1980), Review of Nagels Mortal Questions, Philosophical Review,
89: 4756.
(1990), Physicalism and the Cognitive Role of Acquaintance, in W. Lycan
(ed.), Mind and Cognition (Oxford: Blackwell), 4909.
Neurath, O. (1931), Physicalism: The Philosophy of the Vienna Circle, in
R. S. Cohen and M. Neurath (eds), Philosophical Papers 19131946 (Dordrecht:
Reidel, 1983), 4851.
Newell, A. (1980), Physical Symbol Systems, Cognitive Science, 4: 13583.
(1990), Unified Theories of Cognition. (Cambridge, MA: Harvard University Press).
Newell, A., and Simon, H. A. (1972), Human Problem Solving (Englewood Clis, NJ:
Prentice-Hall).
(1976), Computer Science as Empirical Enquiry: Symbols and Search, Com-
munications of the Association for Computing Machinery, 19 (3): 11326; reprinted
in M. A. Boden (ed.), The Philosophy of Artificial Intelligence (Oxford: Oxford
University Press, 1990), 10532.
Newell, A., Shaw, J. C., and Simon, H. A. (1958), Elements of a Theory of Human
Problem-Solving, Psychological Review, 65: 15166.
Ney, A. (2007), Can an Appeal to Constitution Save the Exclusion Problem?,
Pacific Philosophical Quarterly, 88 (4): 486506.
No A. (2003), Causation and Perception: The Puzzle Unravelled, Analysis, 63:
93100.
(2004), Action in Perception (Cambridge, MA: MIT Press).
(2006), Experience without the Head, in T. Szabo Gendler and J. Hawthorne
(eds), Perceptual Experience (Oxford: Oxford University Press), 41133.
Noonan, H. (1998), Animalism versus Lockeanism: A Current Controversy,
Philosophical Quarterly, 48: 30218.
(2003), Personal Identity, 2nd edn (London: Routledge).
Noordhof, P. (1997), Making the Change: The Functionalists Way, British Journal
for the Philosophy of Science, 48: 23350.
(1999a), Causation by Content? Mind and Language, 14: 291320.
(1999b), Micro-Based Properties and the Supervenience Argument: A Response
to Kim, Proceedings of the Aristotelian Society, 99 (1): 109114.
(2001), Believe What You Want, Proceedings of the Aristotelian Society, 2001: 101,
24765.
362
Bibliography
(2002), Imagining Objects and Imagining Experiences, Mind and Language, 17:
42655.
(2003a), Not Old . . . But Not That New Either: Explicability, Emergence and
the Characterisation of Materialism, in S. Walter and H.-D. Heckmann (eds),
Physicalism and Mental Causation (Exeter: Imprint Academic), 85108.
(2003b), Review of Consciousness, Color and Content by Michael Tye, Mind
and Language, 18: 53845.
(2003c), Something Like Ability, Australasian Journal of Philosophy, 81: 2140.
(2006a), Environment-Dependent Content and the Virtues of Causal Explanation,
Synthese, 149: 55175.
(2006b), The Success of Consciousness, in A. Freeman (ed.), Radical Externalism
(Exeter: Imprint Academic), 10927.
(2010), Emergent Causation and Property Causation, in C. Macdonald and
G. Macdonald (eds), Emergence in Mind (Oxford: Oxford University Press), 6999.
Norman, D. A. (1986), Reflections on Cognition and Parallel Distributed Processing,
in J. L. McClelland, D. E. Rumelhart and the PDP Research Group, Parallel
Distributed Processing: Explorations in the Microstructure of Cognition, vol. 1,
Foundations (Cambridge, MA: MIT Press), 53146.
Norman, D. A., and Shallice, T. (1986), A ention to Action: Willed and Automatic
Control of Behavior, in R. Davidson, G. Schwartz and D. Shapiro (eds), Con-
sciousness and Self Regulation: Advances in Research and Theory, vol. 4 (New York:
Plenum), 118.
Norman, J. (2002), Two Visual Systems and Two Theories of Perception: An A empt
to Reconcile the Constructivist and Ecological Approaches, Behavioral and Brain
Sciences, 25: 73144.
Nussbaum, M. C. (1984): Aristotelian Dualism, Oxford Studies in Ancient Philosophy,
2: 197207.
OConnor, T. (1994), Emergent Properties, American Philosophical Quarterly, 31:
91104.
OConnor, T., and Churchill, J. R. (2010), Is Non-Reductive Physicalism Viable
within a Causal Powers Metaphysic?, in C. Macdonald and G. Macdonald (eds),
Emergence in Mind (Oxford: Oxford University Press), 4360.
OLeary-Hawthorne, J., and McDonough, J. K. (1998), Numbers, Minds and Bodies:
A Fresh Look at Mind-Body Dualism, Philosophical Perspectives, 12: 34971.
ORegan, J. K., and No, A. (2001), A Sensorimotor Approach to Vision and
Visual Consciousness, Behavioral and Brain Sciences, 24 (5): 883975, 93973.
Oderberg, D. (2005), Hylomorphic Dualism, Social Philosophy and Policy, 22: 7099.
Olson, E. (1997), The Human Animal: Personal Identity without Psychology (New York:
Oxford University Press).
(2007), What are We? A Study in Personal Ontology (Oxford: Oxford University
Press).
Osherson, D. (1995/98), An Invitation to Cognitive Science, 4 vols. (Cambridge, MA:
MIT Press).
Papineau, D. (1984), Representation and Explanation, Philosophy of Science, 51:
5573.
(1986), Semantic Reductionism and Reference, in J. Bu erfield (ed.), Language,
Mind and Logic (Cambridge: Cambridge University Press).
363
Bibliography
(1987), Reality and Representation (Oxford: Basil Blackwell).

(1993a), Philosophical Naturalism (Oxford: Blackwell).
(1993b), Physicalism, Consciousness and the Antipathetic Fallacy, Australasian
Journal of Philosophy, 71 (2): 16983.
(1998a), Mind the Gap, in J. E. Tomberlin (ed.), Philosophical Perspectives, 12:
37388.
(1998b), Teleosemantics and Indeterminacy, Australasian Journal of Philosophy,
76: 114.
(2002), Thinking about Consciousness (New York: Oxford University Press).
Parfit, D. (1971), Personal Identity, Philosophical Review, 80: 327; reprinted in Perry
(1975).
(1976), Lewis, Perry, and What Ma ers, in A. Rorty (ed.), The Identities of
Persons (Berkeley: University of California Press).
(1984), Reasons and Persons (Oxford: Clarendon Press).
Parker, A., Derrington, A., and Blakemore, C. (eds) (2002), The Physiology of
Cognitive Processes. Special Issue of Philosophical Transactions of the Royal Society:
B, 357, 9571146 (London: Royal Society).
Parsons, T. (1980), Nonexistent Objects (New Haven: Yale University Press).
Pa ee, H. H. (1966), Physical Theories, Automata, and the Origin of Life, in
H. H. Pa ee, E. A., Edelsack, L. Fein and A. B. Callahan (eds), Natural Automata
and Useful Simulations: Proceedings of a Symposium on Fundamental Biological Models
(Washington: Spartan Books), 73106.
(1989), Simulations, Realizations, and Theories of Life, in C. G. Langton (ed.),
Artificial Life (Redwood City, CA: Addison-Wesley), 6377.
Paul, L. (2007), Constitutive Overdetermination, in J. K. Campbell (ed.), Causation
and Explanation, Topics in Contemporary Philosophy series (Cambridge, MA:
MIT Press).
Paull, C., and Sider, T. (1992), In Defense of Global Supervenience, Philosophical
and Phenomenological Research, 52: 83354.
Peacocke, C. (1983), Sense and Content (Oxford: Oxford University Press).
(1992), A Study of Concepts (Cambridge, MA: MIT Press).
(1999), Being Known (Oxford: Oxford University Press).
Pearl, J. (2000), Causality: Models, Reasoning, and Inference (Cambridge: Cambridge
University Press).
Penelhum, T. (1970), Survival and Disembodied Existence (London: Routledge).
Penrose, R. (1989), The Emperors New Mind: Concerning Computers, Minds, and the
Laws of Physics (Oxford: Oxford University Press).
(1994), Shadows of the Mind: A Search for the Missing Science of Consciousness
Pereboom, D. (2002), Robust Nonreductive Materialism, Journal of Philosophy,
99 (10): 499531.
Pereboom, D., and Kornblith, H. (1991), The Metaphyscs of Irreducibility,
Philosophical Studies, 63 (2): 12545.
Perry, J. (1993), The Essential Indexical and Other Essays (Oxford: Oxford University
Press).
(1994), Intentionality (2), in S. Gu enplan (ed.), A Companion Volume to the
Philosophy of Mind (Oxford: Blackwell).
364
Bibliography
(2001), Knowledge, Possibility, and Consciousness (Cambridge, MA: MIT Press).

Pe it, P. (1993), A Definition of Physicalism, Analysis, 53: 21323.
(2009), Consciousness and the Frustrations of Physicalism, in Ravenscro
(2009), 16387.
Pietroski, P. (1992), Intentionality and Teleological Error, Pacific Philosophical
(1994): Mental Causation for Dualists, Mind and Language, 9: 33666.
Pinker, S., and Prince, A. (1988), On Language and Connectionism: Analysis of a
Parallel Distributed Model of Language Acquisition, Cognition, 28: 73193.
Place, U. T. (1956), Is Consciousness a Brain Process?, British Journal of Psychology,
47 (Part 1): 4450.
Poland, J. (2001), Physicalism, The Philosophical Foundations (Oxford: Oxford University
Press).
Polger, T., and Flanagan, O. (1999), Natural Answers to Natural Questions,
in V. Hardcastle (ed.), Where Biology Meets Psychology: Philosophical Essays
(Cambridge, MA: Bradford Books, MIT Press), 22147.
Popper, K. R. (1953), Language and the Mind-Body Problem: A Restatement of
Interactionism, in Proceedings of the 11th International Congress of Philosophy;
reprinted in Conjectures and Refutations (Basic Books, 1962).
(1955): A Note on the Mind-Body Problem, Analysis, 15: 1315.
Popper, K., and Eccles, J. (1977): The Self and Its Brain (New York: Springer).
Premack, D., and Woodru, G. (1978), Does the Chimpanzee Have a Theory of
Mind?, Behaviour and Brain Science, 1: 51526.
Priest, G. (2005), Towards Non-Being: The Logic and Metaphysics of Intentionality
Prinz, J. (2004), Gut Reactions: A Perceptual Theory of Emotion (Oxford: Oxford
University Press).
(2007), Mental Pointing: Phenomenal Knowledge without Concepts, Journal
of Consciousness Studies, 14: 910, 184211.
Pucce i, R. (1973), Brain Bisection and Personal Identity, British Journal for the
Philosophy of Science, 24: 33955.
Putnam, H. (1960), Minds and Machines, in S. Hook (ed.), Dimensions of Mind:
A Symposium (New York: New York University Press), 14879.
(1967), The Nature of Mental States. First published as Psychological Predi-
cates, in W. H. Capitan and D. Merrill (eds), Art, Mind, and Religion (Pi sburgh:
University of Pi sburgh Press), 3748; reprinted in H. Putnam, Mind, Language,
and Reality: Philosophical Papers, vol. 2 (Cambridge: Cambridge University Press,
1975), 42940, and in Rosenthal (1991), 197203.
(1975), The meaning of meaning, reprinted in his Mind, Language, and Reality:
Philosophical Papers Vol. II (Cambridge: Cambridge University Press).
(1982), Why There isnt a Ready-Made World, Synthese, 51: 14167.
(1986) Information and the Mental, in E. Lepore (ed.), Truth and Interpretation,
Perspectives on the Philosophy of Donald Davidson (Oxford: Blackwell).
(1988), Representation and Reality (Cambridge, MA: MIT Press).
(1996), Introduction to A. Pessin and S. Goldberg (eds), The Twin Earth Chronicles:
Twenty Years of Reflection on Hilary Putnams The Meaning of Meaning (New
York: M. E. Sharpe).
365
Bibliography
(1997), Functionalism: Cognitive Science or Science Fiction?, in D. M. Johnson

and C. E. Erneling (eds), The Future of the Cognitive Revolution (Oxford: Oxford
University Press), 3244.
Pylyshyn, Z. W. (1973), What the Minds Eye Tells the Minds Brain: A Critique
of Mental Imagery, Psychological Bulletin, 80: 124.
(1980), Computation and Cognition: Issues in the Foundations of Cognitive
Science, Behavioral and Brain Sciences, 3: 11132.
Quine, W. (1953a), On What There is, in his From a Logical Point of View (New York:
Harper and Row), 119.
(1953b), Two Dogmas of Empiricism, in his From a Logical Point of View
(New York: Harper and Row), 2046.
(1960), Word and Object (Cambridge, MA: MIT Press).
Ravenscro , I. (2005), Philosophy of Mind: A Beginners Guide (Oxford: Oxford
University Press).
(2009), Minds, Ethics, and Conditionals: Themes From the Philosophy of Frank
Jackson (Oxford: Oxford University Press).
(2010), Folk Psychology, as a Theory, in E. Zalta (ed.), Stanford Encylopedia of
Philosophy. Available at h p://plato.stanford.edu/.
Ray, T. S. (1992), An Approach to the Synthesis of Life, in C. G. Langton, C. Taylor,
J. D. Farmer and S. Rasmussen (eds), Artificial Life II (Redwood City, CA:
Addison-Wesley), 371408.
Raymont, P. (2003), Kim on Overdetermination, Exclusion and Nonreductive
Physicalism, in S. Walter and H.-D. Heckmann (eds), Physicalism and Mental
Causation: The Metaphysics of Mind and Action (Exeter: Imprint Academic).
Reid, T. (1975) [1785], Of Mr. Lockes Account of Personal Identity, in J. Perry (ed.),
Personal Identity (Berkeley and Los Angeles: University of California Press).
Rey, G. (1997), Contemporary Philosophy of Mind: A Contentiously Classical Approach
(2002), Searles Misunderstandings of Functionalism and Strong AI, in J. Preston
and M. Bishop (eds), Views into the Chinese Room (Oxford: Oxford University
Press), 20125.
(2007), Resisting Normativism in Psychology, in J. Cohen and B. McLaughlin
(eds), Blackwell Debates in Philosophy of Mind (Oxford: Blackwell), 6984.
(2009), Concepts, Defaults, and Internal Asymmetric Dependencies: Distillations
of Fodor and Horwich, in N. Kompa, C. Nimtz and C. Suhm (eds), The A Priori
and Its Role in Philosophy (Paderborn: Mentis).
(forthcoming), Externalism and Inexistence in Early Content, in R. Schantz (ed.),
Prospects for Meaning (New York: De Gruyter).
Richardson, R. C. (1982), The Scandal of Cartesian Dualism, Mind, 91: 2037.
Robb, D. (1997), The Properties of Mental Causation, The Philosophical Quarterly,
47: 17894.
Robb, D., and Heil, J. (2008), Mental Causation, in E. Zalta (ed.), Stanford
Encyclopedia of Philosophy. Available at h p://plato.stanford.edu/.
Robinson, H. (1982), Maer and Sense (Cambridge: Cambridge University Press).
(1983), Aristotelian Dualism, Oxford Studies in Ancient Philosophy, 1: 12344.
(2003), Dualism, in Stich and Warfield (2003), 85101.
Rock, I. (1983), The Logic of Perception (Cambridge MA: MIT Press).
366
Bibliography
Rodriguez-Pereya, G. (2006), Truthmaking, Entailment, and the Conjunction

Thesis, Mind, 115: 95782.
Root, M. (1986), Davidson and Social Science, in Lepore (ed.), Truth and Inter-
pretation: Perspectives on the Philosophy of Donald Davidson (Oxford: Blackwell).
Rorty, R. (1970a), Incorrigibility as the Mark of the Mental, Journal of Philosophy,
68: 399424.
(1970b), In Defense of Eliminative Materialism, Review of Metaphysics, 24:
11221.
(1972), Functionalism, Machines, and Incorrigibility, Journal of Philosophy, 69:
41758.
Rosch, E. H., and Mervis, C. B. (1975), Family Resemblances: Studies in the Internal
Structure of Categories, Cognitive Psychology, 7: 573605.
Rosenburg, J. F. (1988), On Not Knowing Who or What One is: Reflections on the
Intelligibility of Dualism, Topoi, 7: 5763.
Rosenthal, D. M. (1986), Two Concepts of Consciousness, Philosophical Studies, 49:
32959.
(1991a), The Independence of Consciousness and Sensory Qualities, in
E. Villaneuva (ed.), Consciousness, Philosophical Issues, no. 1 (Atascadero, CA:
Ridgeview), 1536.
(1991b), The Nature of Mind (New York: Oxford University Press).
(1993), Thinking That One Thinks, in M. Davies and G. W. Humphreys (eds),
Consciousness (Oxford: Blackwell), 197223.
(1997), A Theory of Consciousness, in N. Block, O. Flanagan and G. Guzeldere
(eds), The Nature of Consciousness (Cambridge, MA: MIT Press).
(2000), Consciousness, Content and Metacognitive Judgments, Consciousness
and Cognition, 9 (2): 20314.
(2005), Consciousness and Mind (New York: Oxford University Press).
Ross, D. (1994), Denne s Conceptual Reform, Behaviour and Philosophy, 22: 4152.
Rowlands, M. (1999), The Body in Mind (Cambridge: Cambridge University
Press).
(2001), The Nature of Consciousness (Cambridge: Cambridge University Press).
(2003), Externalism: Puing Mind and World Back Together Again (Chesham, Bucks.:
Acumen).
(forthcoming), The New Science of the Mind: From Extended Mind to Embodied
Phenomenology (Cambridge, MA: MIT Press).
Rozemond, M. (2002), Descartess Dualism (Cambridge, MA: Harvard University
Press).
Rumelhart, D. E., and McClelland, J. L. (1986), On Learning the Past Tenses of
English Verbs, in J. L. McClelland, D. E. Rumelhart and the PDP Research
Group, Parallel Distributed Processing: Explorations in the Microstructure of
Cognition, vol. 1, Foundations (Cambridge, MA: MIT Press), 21671.
Rupert, R. (2004), Challenges to the Hypothesis of Extended Cognition, Journal
of Philosophy, 101 (8): 389428.
Russell, B. (1927), The Analysis of Maer (London: Kegan Paul).
Salmon, N. (1986), Freges Puzzle (Cambridge, MA: MIT Press).
Samuels, R. (1998), Evolutionary Psychology and the Massive Modularity
Hypothesis, British Journal for the Philosophy of Science, 49: 575602.
367
Bibliography
Sawyer, S. (1998), Privileged Access to the World, Australasian Journal of Philosophy,

76: 52333.
(2002), In Defence of Burges Thesis, Philosophical Studies, 107: 10928.
(2006), Externalism, Apriority and Transmission of Warrant, in T. Marvarn
(ed.), What Determines Content? The Internalism/Externalism Dispute (Cambridge:
Cambridge Scholars Press).
(2007), There is No Viable Notion of Narrow Content, in B. McLaughlin and
J. Cohen (eds), Contemporary Debates in the Philosophy of Mind (Oxford: Blackwell).
(2009), The Modified Predicate Theory of Proper Names, in S. Sawyer (ed.),
New Waves in Philosophy of Language (London: Palgrave MacMillan).
Scheier, C., and Pfeifer, R. (1998), Exploiting Embodiment for Category Learning,
in R. Pfeifer, B. Blumberg, J.-A Meyer and S. W. Wilson (eds), From Animals to
Animats 5: Proceedings of the Fih International Conference on Simulation of Adaptive
Behavior (Cambridge, MA: MIT Press), 327.
Scheutz, M. (ed.) (2002), Computationalism: New Directions (Cambridge, MA: MIT
Press).
Schier, S. (1992), Boghossian on Externalism and Inference, Philosophical Issues,
2: 2937.
Scriven, M. (1953), The Mechanical Concept of Mind, Mind, 62: 23040.
Seager, W. (1999), Theories of Consciousness (London and New York: Routledge)
Searle, J. R. (1980), Minds, Brains and Programs, The Behavioral and Brain Sciences,
3 (3): 41757. Includes peer commentaries and reply.
(1983), Intentionality (Cambridge: Cambridge University Press).
(1990a), Consciousness, Explanatory Inversion, and Cognitive Science,
Behavioral and Brain Sciences, 13: 585642.
(1990b), Is the Brains Mind a Computer Program?, Scientific American (January):
2025.
(1992), The Rediscovery of the Mind (Cambridge, MA: Bradford Books, MIT Press).
Segal, G. (1989), The Return of the Individual, Mind, 98: 3957.
(2000), A Slim Book about Narrow Content (Cambridge, MA: MIT Press).
Sellars, W. (1954), A Note on Poppers Argument for Dualism, Analysis, 15: 234.
Shafer-Landau, R. (2003), Moral Realism (Oxford: Oxford University Press).
Shannon, C. E., and Weaver, W. (1963), The Mathematical Theory of Communication
(Illinois: University of Illinois Press, 1998).
Shapiro, L. (2004), The Mind Incarnate (Cambridge, MA: MIT Press).
Shoemaker, S. (1963), Self-Knowledge and Self-Identity (Ithaca: Cornell University
Press).
(1970), Persons and Their Pasts, American Philosophical Quarterly, 7: 26985.
(1984), Personal Identity: A Materialists Account, in Shoemaker and Swinburne,
Personal Identity (Oxford: Blackwell).
(1990), Qualities and Qualia: Whats in the mind? Philosophy and Phenomenological
Research, 50: 109131; and in his The First Person Perspective and Other Essays
(Cambridge: Cambridge University Press, 1996), 97120.
(1991), Qualia and Consciousness, Mind, 100: 50724; and in his The First Person
Perspective and Other Essays (Cambridge: Cambridge University Press, 1996), 12140.
(1994), Self-Knowledge and Inner Sense: Lecture 2, the Broad Perceptual
Model, Philosophy and Phenomenological Research, 54; and in The First Person
368
Bibliography
Perspective and Other Essays (Cambridge: Cambridge University Press, 1996),

12140.
(1997), Self and Substance, in J. Tomberlin (ed.), Philosophical Perspectives,
11 (Atascadero, CA: Ridgeview), 283319.
(1999), Self, Body, and Coincidence, Proceedings of the Aristotelian Society,
Supplementary Volume 73: 287306.
(2004), Functionalism and Personal Identity: A Reply, Nos, 38: 52533.
(2007), Physical Realization (Oxford: Oxford University Press).
Siegel, S. (2006), Which Properties are Represented in Perception?, in T. Szab
Gendler and J. Hawthorne (eds), Perceptual Experience (Oxford: Oxford University
Press), 480503.
Simon, H. (1969), The Sciences of the Artificial (Cambridge, MA: MIT Press).
Simons, D. J., and Chabris, C. F. (1999), Gorillas in Our Midst: Sustained Ina en-
tional Blindness for Dynamic Events, Perception, 28 (9): 105974.
Slocum, A. C., Downey, D. C., and Beer, R. D. (2000), Further Experiments in the
Evolution of Minimally Cognitive Behavior: From Perceiving Aordances
to Selective A ention, in J. Meyer, A. Berthoz, D. Floreano, H. Roitblat and
S. Wilson (eds), From Animals to Animats 6: Proceedings of the Sixth International
Conference on Simulation of Adaptive Behavior (Cambridge, MA: MIT Press), 43039.
Sloman, A. (1971), Interactions between Philosophy and Artificial Intelligence:
The Role of Intuition and Non-Logical Reasoning in Intelligence, Artificial
Intelligence, 2: 20925.
(1975), A erthoughts on Analogical Representation, in R. C. Schank and
B. L. Nash-Webber (eds), Theoretical Issues in Natural Language Processing: An
Interdisciplinary Workshop in Computational Linguistics, Psychology, Linguistics, and
Artificial Intelligence, Cambridge, MA, 1013 June (Arlington, VA: Association for
Computational Linguistics), 1648.
(1978), The Computer Revolution in Philosophy: Philosophy, Science, and Models of
Mind (Brighton: Harvester Press). Out of print, but available and continually
updated online at www.cs.bham.ac.uk/research/coga/crp/.
(1986), Reference without Causal Links, in B. du Boulay and L. Steels (eds),
Seventh European Conference on Artificial Intelligence (Amsterdam: North-Holland),
36981.
(1989), On Designing a Visual System: Towards a Gibsonian Computational
Model of Vision, Journal of Experimental and Theoretical AI, 1: 289337.
(1993), The Mind as a Control System, in C. Hookway and D. Peterson (eds),
Philosophy and the Cognitive Sciences (Cambridge: Cambridge University Press),
69110.
(1996a), Actual Possibilities, in L. C. Aiello and S. C. Shapiro (eds), Principles
of Knowledge Representation and Reasoning: Proceedings of the Fih International
Conference (KR 96) (San Francisco: Morgan Kaufmann), 62738.
(1996b), Beyond Turing Equivalence, in P. J. R. Millican and A. J. Clark (eds),
Machines and Thought: The Legacy of Alan Turing, vol. 1 (Oxford: Oxford University
Press), 179220.
(1996c), Towards a General Theory of Representations, in D. M. Peterson (ed.),
Forms of Representation: An Interdisciplinary Theme for Cognitive Science (Exeter:
Intellect Books), 11840.
369
Bibliography
(1992), The Emperors Real Mind, Review of Roger Penroses The Emperors New
Mind: Concerning Computers Minds and the Laws of Physics, Artificial Intelligence,
56, 35596.
(1999), Review of R. Picards Aective Computing, AI Magazine, 20: 1 (March),
12733.
(2000), Architectural Requirements for Human-Like Agents Both Natural
and Artificial. (What Sorts of Machines Can Love?), in K. Dautenhahn (ed.),
Human Cognition and Social Agent Technology: Advances in Consciousness Research
(Amsterdam: John Benjamins), 16395.
(2002), The Irrelevance of Turing Machines to Artificial Intelligence, in M. Scheutz
(ed.), Computationalism: New Directions (Cambridge, MA: MIT Press), 87127.
(2010), An Alternative to Working on Machine Consciousness, International
Journal of Machine Consciousness, 2 (1): 118.
Sloman, A., and Chrisley, R. L. (2003), Virtual Machines and Consciousness, in
O. Holland (ed.), Machine Consciousness (Exeter: Imprint Academic), 13372.
Special issue of the Journal of Consciousness Studies, 10 (45).
Smart, J. J. C. (1959), Sensations and Brain Processes, Philosophical Review, 68:
14156; reprinted in Rosenthal (1991), 16976.
(1978), The Content of Physicalism, Philosophical Quarterly, 28: 23941.
Smith, A. D. (2002), The Problem of Perception (Cambridge, MA: Harvard University
Press).
Smith, B. C. (1985), Prologue to Reflection and Semantics in a Procedural Language,
in R. J. Brachman and H. J. Levesque (eds), Readings in Knowledge Representation
(Los Altos, CA: Morgan Kauman), 3140.
(1996), On the Origin of Objects (Cambridge, MA: MIT Press).
(1998), On Knowing Ones Own Language, in C. Wright, B. C. Smith and
C. Macdonald (eds), Knowing Our Own Minds (Oxford: Oxford University Press).
(2002a), The Foundations of Computing, in M. Scheutz (ed.), Computationalism:
New Directions (Cambridge, MA: MIT Press), 2358.
(2002b), Keeping Emotions in Mind, in P. Goldie (ed.), Understanding Emotions:
Mind and Morals. Ashgate Epistemology and Mind series (London: Ashgate).
Smith, T., Husbands, P., and OShea, M. (2002), Neuronal Plasticity and Temporal
Adaptivity: GasNet Robot Control Networks, Adaptive Behavior, 10: 16183.
Smolensky, P. (1987), The Constituent Structure of Mental States: A Reply to
Fodor and Pylyshyn, Southern Journal of Philosophy, 26: 13760.
(1988), On the Proper Treatment of Connectionism, Behavioral and Brain Sciences,
11: 174.
Smythies, J. R., and Belo, J. (eds) (1989), The Case for Dualism (Charlo esville:
University of Virginia Press).
Snowdon, P. (1990), Persons, Animals, and Ourselves, in C. Gill (ed.), The Person
and the Human Mind (Oxford: Clarendon Press).
(1996), Persons and Personal Identity, in S. Lovibond and S. G. Williams (eds),
Essays for David Wiggins: Identity, Truth and Value (Oxford: Blackwell).
Soames, S. (2002), Beyond Rigidity: The Unfinished Semantic Agenda of Naming and
Necessity (Oxford: Oxford University Press).
Sober, E. (1992), Learning From Functionalism: Prospects for Strong Artificial
Life, in C. G. Langton, C. Taylor, J. D. Farmer and S. Rasmussen (eds), Artificial
Life II (Redwood City, CA: Addison-Wesley), 74966.
370
Bibliography
Sparber, G. (2005), Counterfactual Overdetermination vs. the Causal Exclusion

Problem, History and Philosophy of the Life Sciences 27 (34): 47990.
Stalnaker, R. (1984), Inquiry (Cambridge, MA: MIT Press).
Stampe, D. (1977), Towards a Causal Theory of Linguistic Representation, in
Midwest Studies in Philosophy, vol. 2 (Minneapolis: University of Minnesota
Press), 4263.
Sterelny K. (1990), The Representational Theory of Mind (Oxford: Blackwell).
(2003), Thought in a Hostile World: The Evolution of Cognition (Malden, MA:
Blackwell).
Steward, H. (1996), The Ontology of Mind (Oxford: Clarendon).
Stich, S. (1983), From Folk Psychology to Cognitive Science: A Case Against Belief.
(Cambridge MA: MIT Press).
(1990), The Fragmentation of Reason: Preface to a Pragmatic Theory of Cognitive
Evaluation (Cambridge MA: MIT Press).
(1996), Deconstructing the Mind (Oxford: Oxford University Press).
Stich, S., and Warfield, T. (1994), Mental Representation: A Reader (Oxford: Blackwell).
(2003), The Blackwell Guide to Philosophy of Mind (Malden, MA: Blackwell).
Stoljar, D. (1996), Nominalism and Intentionality, Nos, 30 (2): 26181.
(2000) Physicalism and the Necessary A Posteriori, Journal of Philosophy, 97 (1):
3354.
(2001a), The Conceivability Argument and Two Conceptions of the Physical,
Philosophical Perspectives, 15: 393413.
(2001b), Two Conceptions of the Physical, Philosophy and Phenomenological
Research, 62: 25381.
(2006), Ignorance and Imagination (Oxford: Oxford University Press).
Stout, R. (2005), Action (Teddington: Acumen).
Stoutland, F. (1976), The Causation of Behaviour, in J. Hintikka (ed.), Essays on
Wigenstein in Honour of G. H. von Wright, Acta Philosophica Fennica (Amsterdam:
North-Holland).
(1980), Oblique Causation and Reasons for Action, Synthese, 43: 35167.
(1985), Davidson on Intentional Behavior, in E. Lepore and B. McLaughlin
(eds.), Actions and Events: Perspectives on the Philosophy of Donald Davidson (New
York: Blackwell).
Strawson, G. (2008), Real Intentionality 3: Why Intentionality Entails Conscious-
ness, in Real Materialism (Oxford: Oxford University Press), 281305.
Strawson, P. F. (1959), Individuals: An Essay in Descriptive Metaphysics (London:
Methuen).
Stroud, B. (1986), The Physical World, Proceedings of the Aristotelian Society, 87:
26377.
Sturgeon, S. (1994), The Epistemic View of Subjectivity, The Journal of Philosophy,
91: 22135.
Sussman, A. (1981): Reflections on the Chances for a Scientific Dualism, Journal
of Philosophy, 78: 95118.
Su on, J. (2006), Distributed Cognition: Domains and Dimensions, Pragmatics
and Cognition, 14 (2): 23547.
Swinburne, R. (1984), Personal Identity: The Dualist Theory, in Shoemaker and
Swinburne, Personal Identity (Oxford: Blackwell).
(1986), The Evolution of the Soul (Oxford: Clarendon Press).
371
Bibliography
Thagard, P. (1988), Computational Philosophy of Science (Cambridge, MA: MIT Press).

(1989), Explanatory Coherence, Behavioral and Brain Sciences, 12: 435502.
(1990), Concepts and Conceptual Change, Synthese, 82: 25574.
Thomasson, A. (1998), A Nonreductivist Solution to Mental Causation, Philosophical
Studies, 89: 18195.
Thompson E. (2007), Mind in Life: Biology, Phenomenology, and the Sciences of
Mind (Cambridge, MA: Harvard University Press).
Touretsky. D. S., and Hinton, G. E. (1985), Symbols among the Neurons: Details
of a Connectionist Inference Architecture, Proceedings of the Fourth International
Conference on Artificial Intelligence (Los Angeles, CA), 23843.
(1988), A Distributed Connectionist Production System, Cognitive Science, 12:
42366.
Turing, A. M. (1936), On Computable Numbers with an Application to the
Entscheidungsproblem, Proceedings of the London Mathematical Society, Series 2
(423) (30 November), 23040, and (424) (23 December), 24165.
(1950), Computing Machinery and Intelligence, Mind, 59: 43360; reprinted
in M. A. Boden (ed.), The Philosophy of Artificial Intelligence (Oxford: Oxford
University Press, 1990), 4066. Page numbers in the text refer to the reprinted
version.
Turner, M. (1991), Reading Minds: The Study of Literature in an Age of Cognitive
Science (Oxford: Princeton University Press).
Tye, M. (1995/1996), Ten Problems of Consciousness: A Representational Theory of
the Phenomenal Mind (Cambridge, MA: MIT Press).
(1999), Phenomenal Consciousness: The Explanatory Gap as a Cognitive
Illusion, Mind, 108: 70525.
(2000), Color, Content and Consciousness (Cambridge MA: MIT Press).
(2009), Consciousness Revisited: Materialism without Phenomenal Concepts (Cambridge,
MA: MIT Press).
Ullman, S. (1980), Against Direct Perception, Behavioral and Brain Sciences, 3:
373415.
Unger, P. (1979), I Do Not Exist, in G. F. MacDonald (ed.), Perception and Identity
(London: Macmillan); reprinted in Rea (1997).
Van Cleve, J. (1990), Supervenience and Closure, Philosophical Studies, 58: 22583.
Van Gelder, T. J. (1995), What Might Cognition Be, If Not Computation?, Journal
of Philosophy, 92: 34581.
Van Gulick, R. (2000), Closing the Gap?, Journal of Consciousness Studies 7 (4):
937.
(2003), Maps, Gaps and Traps, in A. Jokic and Q. Smith (eds), Consciousness:
New Philosophical Perspectives (Oxford: Oxford University Press).
(2004), Higher-Order Global States HOGS: An Alternative Higher-Order Model
of Consciousness, in Gennaro, R. (ed.), Higher-Order Theories of Consciousness
(Amsterdam and Philadelphia: John Benjamins).
(2009), Jacksons Change of Mind: Representationalism, A Priorism and the
Knowledge Argument, in Ravenscro (2009), 189218.
Van Inwagen, P. (1990), Material Beings (Ithaca: Cornell University Press).
Varela, F. J., Thompson, E., and Rosch, E. (1991), The Embodied Mind: Cognitive
Science and Human Experience (Cambridge, MA: MIT Press).
372
Bibliography
Velleman, J. D. (2000), On the Aim of Belief, in his The Possibility of Practical

Reason (Cambridge: Cambridge University Press, 2000), 24481.
Vendler, Z. (1972), Res Cogitans (Ithaca, NY: Cornell University Press).
(1984), The Maer of Minds (Oxford: Clarendon Press).
Vicente, A. (1999), Mind-Body Causal Overdetermination, Theoria, 14 (36): 51124.
Von Eckardt, B. (1995), Folk Psychology (1), in Gu enplan (1994), 3007.
Wallace, A. F. C. (1965), Driving to Work, in M. E. Spiro (ed.), Context and
Meaning in Cultural Anthropology (London: Collier-Macmillan), 27796.
Warfield, T. A. (1997), Externalism, Self-knowledge and the Irrelevance of Slow-
Switching, Analysis, 57: 2824.
Wedgwood, R. (2005), Normativism Defended, in J. Cohen and B. McLaughlin
(eds), Blackwell Debates in Philosophy of Mind (Oxford: Blackwell), 85101.
Weiskrantz, L. (1997), Consciousness Lost and Found: A Neuropsychological Exploration
Wheeler, M. (2005), Reconstructing the Cognitive World: The Next Step (Cambridge,
MA: MIT Press).
(forthcoming a), In Defense of Extended Functionalism, in Menary (forthcoming).
(forthcoming b), In Search of Clarity about Parity, in Philosophical Studies, book
symposium on A. Clark, Supersizing the Mind (Clark [2008b]).
(forthcoming c), Minds, Things, and Materiality, in C. Renfrew and
L. Malafouris (eds), The Cognitive Life of Things: Recasting the Boundaries of the
Mind (Cambridge: McDonald Institute for Archaeological Research Publications).
Whitby, B. (1996), Reflections on Artificial Intelligence: The Legal, Moral, and Ethical
Dimensions (Exeter: Intellect Books).
White, S. (1982), Partial Character and the Language of Thought, Pacific Philo-
sophical Quarterly, 63: 34765.
Whyte, J. T. (1990), Success Semantics, Analysis, 50: 14957.
(1991), The Normal Rewards of Success, Analysis, 51: 6574.
Wiggins, D. (2001), Sameness and Substance Renewed (Cambridge: Cambridge
University Press).
Wilkerson, T. E. (1998), Recent Work on Natural Kinds, Philosophical Books, 39:
22533.
Wilkes, K. (1988), Real People (Oxford: Clarendon Press).
Williams, B. (19561957), Personal Identity and Individuation, Proceedings of the
Aristotelian Society, 57; reprinted in his Problems of the Self (Cambridge: Cambridge
University Press, 1973).
(1970), The Self and the Future, Philosophical Review, 59; reprinted in his
Problems of the Self (Cambridge: Cambridge University Press, 1973).
Williamson, T. (2006), Conceptual Truth, The Aristotelian Society, Supplementary
Volume 80: 141.
Wilson, J. (1999), How Superduper Does a Physicalist Supervenience Need to
Be?, Philosophical Quarterly, 50 (194): 3352.
(2005), Supervenience-Based Formulations of Physicalism, Nos, 39: 42659.
(2010), What is Humes Dictum, and Why Believe It?, Philosophy and Phenom-
enological Research, 80: 595637.
Wilson, R. (1995), Cartesian Psychology and Physical Minds (Cambridge: Cambridge
University Press).
373
Bibliography
Witmer, G. (2006), How to Be a (Sort of) A Priori Physicalist, Philosophical Studies,

131 (1): 185225.
Wi genstein, L. (1953), Philosophical Investigations (New York: Macmillan and
Oxford: Blackwell).
Wollheim, R. (2005), The Emotions and Their Philosophy of Mind, in A. Hatzimoysis
(ed.), Philosophy and the Emotions (Cambridge: Cambridge University Press).
Wright, C. (2004), On Epistemic Entitlement: Warrant for Nothing (and Foundations
for Free?), Proceedings of the Aristotelian Society, Supplementary Volume 78:
167212.
Wright, I. P., Sloman, A., and Beaudoin, L. P. (1996), Towards a Design-Based
Analysis of Emotional Episodes, Philosophy, Psychiatry, and Psychology, 3: 10137.
Yablo, S. (1992), Mental Causation, The Philosophical Review, 101: 24580.
Yolton, R. (1983), Thinking Maer (Minneapolis: University of Minnesota Press).
Zahavi D. (2005), Subjectivity and Selood: Investigating the First-Person Perspective
Zi, P. (1959), The Feelings of Robots, Analysis, 19: 648.
Zimmerman, D. W. (2004): Should a Christian Be a Mind-Body Dualist?, in
Contemporary Debates in Philosophy of Religion (Malden, MA: Blackwell).
374
Index
aboutness see intentionality autobiographical memory211, 212, 215

Abrahamsen, A.234 autopoiesis158, 1612, 166
access consciousness254, 287
achievement problem and Baars, B.164
externalism1458 Balog, K.256
action, significance of280 basicality and ecumenical
Adams, F.17, 54, 61, 222, 231, 233, 331n2, approach1889
332nn11, 18 basic belief284
aordances158 Bechtel, W.2346
agents12, 1516, 19, 55, 88, 1024, 10910, Beer, R. D.229
11214, 121, 125, 1602, 183, 188, behaviourism50
2201, 225, 236, 308 significance of2823
Agre, P. E.160 see also individual entries
Aizawa, K.55, 222, 231, 233, 331n2, Beighley, S.54
332n18 belief3, 1216, 43, 589, 68, 109, 121, 129,
Akins, K.27 136, 267, 274, 2834
akrasia see will desire framework and behaviour
Aleksander, I.165 1046, 11113, 11722
Alexander, S.27 and meaning11415
Allen, C.17 see also desire
animal minds281 biological function17, 63, 69, 275,
animate vision161 2768
anomalousness biological memories231
and holism of mental11112, 191 blindsight66, 6970
of mental910, 20 Block, N.15, 24, 40, 51, 69, 154, 164, 188,
monism and7, 110, 190, 1913, 197, 196, 254, 257, 261, 275, 287, 331n3
199, 2812 (Ch 2), 331n4
Anscombe, G.4 blockers and physicalism244
anti-individualism see externalism Boden, M. A.151, 153, 162, 1634, 166,
anti-physicalism267 167, 169, 170, 223, 336n1
Antony, L.191, 331n4 body2226
a priori principles, of kinds of2269
interpretation11214 and mind1, 23
Aquinas79, 88, 89, 90 swapping79
Arbib, M. A.169 see also individual entries
Armstrong, D.4, 57, 330n1 Boghossian, P.336nn30, 34
artificial intelligence and artificial life Bond, A. H.160
(AI/A-Life)151, 166, 167, 168, 169, Bontly, T.197
172, 282 Boswell, J.302
aspectual shape and connection Boyer, P.169
principle656 Braddon-Mitchell, D.15, 25, 29, 30
asymmetric dependencies1845 brain38, 11, 257, 40, 46, 49, 523, 57,
asymmetry, between first-and 616, 77, 824, 89, 130, 15160, 218,
third-person perspective107, 11617 2256, 228, 276, 304
authority, first-person116 brainoscopes82
375
Index
Brentano, F.60, 172, 175, 337n5 Churchland, P. S.154, 155, 289

Psychologie vom empirischen Standpunkt circularity2078
(Psychology from an Empirical objection, to neo-Lockean
Standpoint)296 approach21619
Breuer, J.307 Clark, A. J.54, 154, 155, 156, 157, 159,
Bringsford, S.167 160, 162, 169, 220, 2212, 224, 226,
Broad, C. D. 227, 228, 229, 230, 233, 291, 339n2,
The Mind and its Place in Nature33 340n4
Broadbent, D. E.163 coarse-grained information
Brooks, R. A.160, 224 processing68
Brown, J.336nn26, 35 cogito-like judgements336nn27, 33
Brueckner, A.336n36 cognitive illusion27, 45
Bruner, J.155, 159 cognitive psychology6, 108, 151
Burge, T.18, 144, 147, 197, 335nn3, 68, cognitive science2, 2223
1213, 19, 23, 336nn278, 324 computation and1668
Bush, V.159 connectionism and1546
Butler, J.215 consciousness and1635
Byrne, A.264, 269 embodiment and1602
enactiveness and161, 162
Calude, C.167 extended mind and15960
Campbell, N.190, 200, 251, 338nn4, 6, functionalism and1524
339n8 mind and life and1656
Cariani, P.166 phenomenology and161, 162
Carnap, R.282 and pluralism1512
Carruthers, P.38, 50, 58 representation varieties1569
Cartesianism2201 cognitive slips9
Cartwright, R. T.337n6 cog-objection58
causality111, 144, 167, 197, 334n1 co-instantiation186, 246, 272
and closure78, 195, 250, 285, 338n3 Cole, M.159
and connectionism29 Collins, H.228
and explanations of behaviour105 commonsense psychology see folk
and interaction89 psychology
nomological character of9 communal use and linguistic
power and21, 94, 253 meaning137
realism and194, 2001 compositional syntax and semantics156,
and relevance22 2356
and responsibility889 computational/representational theory of
cerebroscope57 thought (CRTT)171, 337n4
c-fibre firing78, 1011, 291, 330n3 Marr on175
Chabris, C. F.161 (non-) physical (non-) locality1734
Chalmers, D.24, 26, 33, 43, 44, 159, 160, program1723
162, 164, 165, 221, 230, 232, 256, 291, Turing on174
318, 330n10, 335n17, 339n2 computationalism, significance of2856
The Conscious Mind333n10 computational psychology124
charity112, 11314, 115 conceivability argument see zombies
Chinese room argument285 concepts, significance of286
Chisholm, R. M.339n11 Conee, E.255
Chomsky, N.6, 155 confirmation holism180
Chrisley, R. L.164, 165, 168 connectedness
Churchill, J. R.251 of quasi-remembered past
Churchland, P. M.12, 31, 38, 40, 41, 106, experience215
154, 155, 169, 289 of remembered past experience213
376
Index
connectionism3, 645, 103, 109, 113, Dale, K.2256, 228

1223, 1546, 158, 167, 223, 225, Damasio, A. R.163
286 Davidson, D.810, 11, 20, 102, 11214,
causality and29 119, 129, 190, 191, 1978, 208, 251,
and networks2356, 237 280, 281, 330n5, 334nn13, 335nn9, 21,
consciousness67, 23, 358, 211, 268, 336n29, 338n6
26974, 287, 333n10 and causal requirement111
anti-physicalism267 philosophy of mind of10811
cognitive science and1635 against scientific psychology117
folk psychology and12930 on third-person epistemology of
knowledge and2833 mind11517
loss of213 Davies, M.128, 276, 336n37
metaphors of mind and4753 De Jaegher, H.166
non-reductive naturalism, for and Denne , D.2, 1213, 25, 27, 38, 42, 4851,
against427 56, 58, 102, 124, 130, 1534, 162,
physicalism and246 1634, 169, 294, 333n10, 334n4, 337n1
reductively naturalistic on design stance118
frameworks3841 on intentional stance118
strong eliminativism about27 original position of119
substance dualism and8990 revised view of11922
system view of6472 Descartes, R.2, 6, 7, 42, 58, 73, 75, 76, 79,
weak eliminativism about278 845, 88, 89, 90, 158, 163, 165, 172,
consequence problem and 209, 220, 243, 280, 300, 305, 332n14,
externalism1489 333n5
constitutive property199 descriptivism and internalism140
content1314, 27, 49, 56, 108, 115, 1246, design stance118, 154, 164
147, 160, 179, 2701, 278, 287 desire6, 9, 1213, 1516, 21, 68, 97, 1027,
associated133 11013, 11822, 1257, 2778, 288
intentional123, 1245, 130, 172, 1778, determination
185, 189 over-5, 6, 20, 23, 196, 197
narrow versus wide1819, 1412, upward967
1878 Devi , M.182, 338n16
phenomenal254, 2608, 274 Dienes, Z.163
propositional62, 123 Dietrich. L.332n11
representational21, 512, 70, 1359, Di Paolo, E. A.158, 162, 166
142, 171, 175, 189 direct reference theory138
semantic52, 67, 701, 185, 188 disjunction problem1617, 40, 70, 124,
theories1517 1823
continuous recurrent neural networks disposition3, 39, 68, 131, 182, 2712, 288
(CNNs)229 dissociative states163
Cooper, R.152, 163 Dretske, F.16, 25, 38, 60, 61, 678, 69, 70,
Copeland, B. J.166, 167 71, 158, 182, 185, 260, 275, 331n10,
co-variation locking theories and 332nn1920, 337n10
externalism1823 Dreyfus, H. L.152, 158, 161, 162, 168
Craik, K. J. W.152, 156, 157 Dreyfus, S.162
Crane, T.601, 62, 67, 242, 263, 264, Dreyfus, S. E.152
3312n10 dualism4, 249, 2889, 300
Crick, F.38, 164 emergent245
criterial relation and identity2067, 212 interactive property56
Cummins, R.178, 275 interactive substance see interactive
Cussins, A.155, 157 substance dualism
cyberneticists165 substance see substance dualism
377
Index
Dunmall, B.165 explanatory irrealism see explanatory

dynamic flux168 internalism
explanatory realism194
Eccles, J. extended functionalism233
The Self and its Brain333n7 extended mind hypothesis2212, 223,
ecumenical approaches 226, 229, 236, 2912, 339n2
basicality and1889 and cognitive science15960
narrow and wide content and1878 functionalism and2323
Edmonds, E. A.169 parity principle and22932
eliminativism1112, 55 externalism18, 133, 297, 336n39
materialism and106, 154, 156, 289 achievement problem and1458
strong27 asymmetric dependencies and1845
and threat to folk psychology1056 consequence problem and1489
weak278 co-variation locking theories and1823
Elman, J. L.156 historical causal theories and1812
Elster, J.130 ideal co-variation1834
Elugardo, R.335n13 and metaphysical considerations
embodied cognition1602, 2202 1434
embrained knowledge2289 phenomenal3034
emergentism267 predicative1347
and dualism245 problems with1867
emotions27, 1302, 28990 singular1379
see also individual entries teleofunctional theories185
enactiveness and cognitive science external memory159
161, 162 Ezquerro, J.197
enactivists, on consciousness523
Enc, B.62 false memory215
Engelbart, D. C.159 Feigl, H.7
epiphenomenalism6, 20, 22, 190, 198, Fido theory181
199, 290 Field275
about qualia26, 28, 29 first-person authority116, 274, 2923
epistemology1, 24, 32, 44, 46, 54, 77, first-person/third person perspective43,
7981, 95, 107, 11517, 1403, 1459, 45, 48, 49, 51,
17980, 1834, 248, 292, 306 569, 81, 82, 86, 1078, 11517,
equivalence relation and identity206, 119, 12831, 209, 293
207, 212, 213 Fitch, T.634, 67, 68, 69, 332n15
ethical non-naturalism2467 Flanagan, O.25, 41, 165
Evans, G.689, 138, 155, 314, 335n10 Fodor, J. A.12, 14, 17, 19, 21, 623,
events 67, 102, 128, 130, 153, 155, 158,
and behaviour10910 160, 166, 179, 182, 183, 184, 185, 186,
metaphysics of196202 188, 192, 235, 286, 332n14, 334n5,
Evere , A.337n6 335nn16, 22, 337nn3, 9, 338nn12, 14
exclusion problem201, 1936, 200 A Theory of Content275
responses to213 intentional realism of1227
experiential awareness36 folk psychology1112, 1024, 108,
experiential memory see autobiographical 121, 1235, 1312, 156, 161, 289,
memory 2934, 302
explanatorily basic property1889 anomalousness and mental holism
explanatory exclusion principle2001 and11112
explanatory gap457, 53, 240, 248, a priori principles of interpretation
25560, 291, 317 and11214
explanatory internalism193 belief and meaning and11415
378
Index
causal requirement and111 Godfrey-Smith, P.17, 165

consciousness and12930 Gois, I.28
design stance and118 Goldberg, S.336nn34, 36
eliminativist threat to1056 Goodale, M.66, 170
emotions and1302 Good Old-Fashioned AI (GOFAI)151,
idealization and abstraction of120 153, 1556, 160, 162, 167
intentional network and104 granny psychology125
intentional realism and1227 Grau, C.162
intentional stance and118 Greenfield, P. M.159
mentalizing abilities and1289 Grice, H. P.70, 188
and mental life127 Grin, D. R.169
philosophy of mind and10811 Griths, P.278
physical stance and119 Grush, R.157
propositional a itudes and intentional Gulick, R. V.197
actions and1034
realism about1067 Harman, G.25, 32, 260
scientific psychology and1078, Harnish, M.337n3
11718 Hart, W. D.332n1
and self-knowledge107 Ha iangadi, A.277
third-person epistemology of mind Haugeland, J.151, 166, 168, 223
and11517 Hawthorne, J.244, 333n3
Foster, J.332n1 Hayes, P. J.169
frame problem294 Heidegger, M.161, 162, 168
Frankish, K.284 Heil, J.20, 253, 336n27
free will163 Hempel, C.193, 282
Frege, G.205, 2067, 284 Hess, P.191
Freud, S.6, 307 Hesse, M. B.169
Frith, C. D.163, 164 heterophenomenology489, 130
functionalism7, 8, 39, 401, 50, 156, 222, higher-order thought2701, 296
2267, 295 Hinton, G. E.156, 158
cognitive science and1524 historical causal theories and
and extended mind2323 externalism1812
and representation33 Hobbes, T.317
functionalization248 Hofweber, T.337n6
Honderich, T.191, 268, 338n1
Gallagher, S.129, 330n1 (Ch 2) Horgan, T.12, 246, 334n4, 339n11
Gallese, V.129 Hornsby, J.280
Gallistel, C.1734, 177, 186 Horwich, P.188, 278
gamblers fallacy337n2 H2O XYZ18, 21, 180, 181, 188, 295
Garvey, J.336n25, 338n16, 340nn45 human-scope cognitive system2345
Belief in God333n6 Hume, D.97, 288
Free Will333n8 Hunt, E. B.155
Gasser, L.160 Hurley, S. L.55, 339n2
Gates, G.186 Husbands, P.167, 2256, 228
Geach, P. T.165 Hutchins, E.160
Geertz, C.159 Hu o, D. D.35, 52, 255, 276
Gibb, S.338n6, 339n7
Gibson, E. J.158 ideal co-variation externalism1834
Gibson, J. J.158 identification problem and substance
Gille , C.197, 253 dualism7883
God101 implementational materiality and
upward determination and967 body2279
379
Index
incomplete linguistic externalism1367 Johnson, M.339n1

incorrigibility569 Jonas, H.158, 166
indeterminacy112, 330n2 Joyce, J.
and substance dualism834, 86 Ulysses49
indiscernibles identity203 Joycean machine50
individualism see internalism
infants-and-animals objection, and Kallestrup, J.197
mental states58 Kaplan, D.188, 337n7
inferential promiscuity23 KarmiloSmith, A.153, 169
instrumentalism12, 154 Kezer, F.336n24
intentionality21, 55, 604, 67, 112, 1201, Kenny, A.333n3
159, 168, 171, 172, 2967, 332n13 K-function207
and action110 Kim, J.10, 19, 201, 22, 556, 50, 190,
and content123, 1245, 130, 172, 191, 1936, 198201, 247, 312, 333n4,
1778, 185, 189 334n1, 338nn25, 339n9
nano634, 68, 69 Kirsh, D.157
network, and beliefs and desires104 Kitcher, P.12
and normativity2749 knowledge and consciousness2833
orders of678 Koch, C.164
primitive62 Kolers, P. A.163
and psychology1235 Kornblith, H.196
and realism1227 Kriegel, U.271, 272, 273
and sensory states556, 61, 65, 68 Kripke, S.181, 335n5, 338n11
unconscious state and65
intentional stance1213, 1534 Ladyman, J.94
Denne on118 Lako, G.339n1
interactive property dualism56 language of thought1247, 153, 172, 298
interactive substance dualism Larmer, R.333n4
(Cartesian dualism)4, 26 Latham, N.333n10
challenges to46 Lawrence, M.162
non-reductive physicalism and710 Leibnizs law205
reductive physicalism and7, 8 LePore, E.179, 197
see also substance dualism Le vin, J. Y.162
internalism133, 297 Leuenberger, S.244
conceptual roles and17981 Levine, J.45, 95, 255, 258, 332n21
images and stereotypes and1789 Lewis, D.30, 31, 154, 242, 255
and metaphysical considerations143 Libet, B.163
phenomenal304 linguistic meaning1367
singular139 linguosemantics337n1
thorough-going1401 Loar, B.95, 119, 188
two-factor1413 Mind and Meaning106
interpretation theories Locke, J.204, 208, 20916, 218
and a priori principles11214 Lockwood, M.26
linguistic and non-linguistic11415 Loewer, B.184, 197
introspection67, 45, 48, 163, 178, 240, logical behaviourism see behaviourism
259, 26770, 2978 Lombard, L. B.339n11
irrealism, about psychological talk106 Lowe, E. J.203, 206, 208, 209, 211, 217,
Irwin, W.162 219, 250, 333n4
Lucas, J. R.154
Jackendo, R.51 Ludlow, P.333n9, 336nn26, 31
Jackson, F.22, 25, 26, 28, 29, 30, 31, 32, 43, Lutz, A.43
95, 101, 242, 255, 309, 318, 333n9 Lycan, W. G.38, 42, 50, 268, 269
380
Index
Macdonald, C.251, 339n11 metabolism166

Macdonald, G.251, 339n11 meta-cognition163
machine consciousness165 methodological behaviourism
Majors, B.336nn3940 see behaviourism
map theory15 methodological issues2
Marr, D. C.6, 158, 175 Metzinger, T.164
Marras, A.200, 339nn911 microphysics98
Martin, M. G. F.261, 266, 267 Millikan, R.158, 185, 275, 338n14
Marx, K.220 Mills, E.333n4
material stu and consciousness165 Milner, A.66, 170
Matrix, The162 mind54
Ma hen, M.332n17 and body problem1, 23, 299300
Maturana, H.161, 166 lack of unified conception of55
Mawson, T.73, 249 property cluster view of56
May, L.169 single property view of5664
McCarthy, J.169 system view of6472
McClelland, J. L.155 see also individual entries
McCulloch, G.152, 154 minimal physical duplication and
The Life of the Mind316 physicalism2435
McDowell, J.138, 161, 226, 316, 335n10 Minsky, M. L.152
McGinn, C.6, 23, 25, 45, 46, 95, 162, 165, mirror-neurons1289
258, 331n4, 332n21 misrepresentation67
Mckinsey, M.336n35 modest physicalism301, 32
McLaughlin, B. P.27, 156, 334n5, 336n26 Montero, B.92, 244, 332n1, 334nn1, 6
McTaggart, J.283 Moore, D.200, 283, 339n8
Mellor, D. H.242, 250 Moore, G. E.91, 2467, 283
Melnyk, A.93, 94, 241, 334n6 moral psychology301
Menary, R.231, 339n2, 340n5 moral responsibility1, 889
mental causation19, 21, 190, 299 Morris, M.161, 335n21
anomalous monism and1913 mountainmolecule relations and
and anomalousness of mental20 physicalism958
dual explanandum solutions223 multiple dra s model49, 130
exclusion problem201, 1936 multiple instantiation867
metaphysics of events and196202 multiple personality syndrome163
and physicalism2503 multiple realizability2278, 232, 233,
program explanation and22 301
reductive physicalism and212 Myin, E.52
and representational content21 mysterianism55
mentalese see language of thought
mental representation13, 32, 64, 1236, Nagasawa, Y.333n9
153, 178, 189 Nagel, T.23, 25, 37, 214, 255, 258, 316
conceptual role approach1516 nano-intentionality634, 68, 69
dual explanandum solutions and223 narrow versus wide content1819
information-theoretic approaches naturalism144
1617 non-reductive427
narrow versus wide content1819 natural kind externalism1356
representational theory of mind natural meaning70, 182
and14 natural selection17, 185
mental substances4, 26, 305 Neander. K.185, 260
Menzies, P.197 Nemirow, L.30, 32, 255
Merleau-Ponty, M.158, 161, 162 neo-Lockean approach, circularity
Mervis, C. B.155 objection to21619
381
Index
neural network151, 224, 229, 237, 286, Pearl, J.169

3012 Penrose, R.38, 154, 165
Newell, A.152, 153, 155, 159, 167, 2334 perception14, 45, 50, 51, 129, 134, 137,
new mysterianism25 258, 2601, 26571, 274, 276
Ney, A.252, 338n5 conscious689
No, A.52, 161, 265, 276, 339n1 inner297, 303
nominalism199, 199 and non-computational
non-computational representational representational theory158
theory, of perception158 outer303
non-modular computation153 perceptual content51, 276, 3034
non-naturalism2467 Pereboom, D.196
non-nomicity623, 67 Perner, J.163
non-reductive naturalism427 person, significance of204
non-reductive physicalism710, 20, 110, personal identity203
190, 191, 197, 200, 201, 250, 2523, criteria of2068
3056 and identification2046
non-representationalist view, of Lockes criterion of21116
experience52 neo-Lockean approach, circularity
Noordhof, P.196, 239, 245, 246, 252, 253, objection to21619
255, 264, 266, 268, 278, 279, 336n24 person, significance of20811
Norman, D. A.152, 156, 163 and substance dualism857
Norman, J.158 pessimistic physicalism25
Pe it, P.22, 25, 32, 249
OConnor, T.245, 251 Pfeifer, R.2245, 228
occurrent states56 phenomenal blueness51
Oderberg, D.333n2 phenomenal concepts44, 45, 2567, 304,
Olson, E.210 331n2 (Ch 2)
ontology1112, 49, 73, 77, 79, 82, 86, 92, phenomenal consciousness see
956, 100, 165, 191, 219, 240, 245, 253, consciousness
2567, 280, 284, 289, 308 phenomenal content254, 2608
Oppenheim, P.193 phenomenal fundamentalism27
optimistic physicalism25 phenomenal representation33
ORegan, J. K.161, 339n1 phenomenology52, 601, 152, 169,
Osherson, D.337n3 267, 295, 3045
other minds7983, 107, 116, 121, 129, 302 and cognitive science161, 162
over-determination5, 6, 20, 23, 196, philosophical zombies see zombies
197, 250 philosophy of presence168
phlogiston theory12
Papineau, D.45, 467, 158, 185, 250, 256, phredness32
275 physical closure56, 330n2
parallel distributed processing physicalism74, 88
(PDP)1546 anti-267
parallelism, between lines207 characterization of2409, 255
parenthood relation and identity213 and consciousness246
Parfit, D.215, 216 domain of928
parity principle22932 intentionality and normativity
Parker, A.157 and2749
partial integration75 mental causation and2503
participatory engagement168 non-reductive710, 20, 110, 190, 191,
Pa ee, H. H.166 197, 200, 201, 250, 2523, 3056
Paul, L.251 phenomenal consciousness and2545,
Peacocke, C.179, 185, 261, 336n27 26074
382
Index
phenomenal properties and psychophysical nexus and

explanatory gap and25560 consciousness456
physical and98101 psychosemantics337n1
reductive7, 8, 212, 305 putative identity criterion2078
physicalist structuralism94 Putnam, H.7, 8, 18, 38, 152, 153, 154, 161,
physical reports57 166, 181, 188, 283, 286, 295, 312, 315,
physical stance119, 154 316, 330n9, 334n3 (Ch 7), 335nn45
physical substances4 Pylyshyn, Z. W.155, 157, 158, 166, 235
physical symbol systems234, 2367
physics258 qualia24, 26, 2833, 61, 62, 69, 153,
folk104 164, 165, 167, 247, 260, 261, 3078,
of nuclear reactions302 330n11
and physicalism98100, 241 quantum indeterminacy and substance
Pietroski, P.185 dualism834
Pinker, S.155 quasi-memory21516
Pi s, W. H.152, 154 Quine, W. V.177, 17980, 186, 189,
Place, U. T.7, 154 337nn6, 9
Plato73, 186
Poland, J.93 radical body-centrism222
Popper, K. randomness84
The Self and its Brain333n7 rationality9, 13, 105, 108, 10910, 11214,
powers ontology245 120, 121, 172
predicative externalism1347 Ravenscro , I.1, 11, 250, 251, 275, 330n8
predicative thought, two-factor theory Ray, T. S.166
of141, 142 Raymont, P.197, 200
Premack, D.104 reaction-diusion (RD) systems225,
preservative memory147 228, 229
Priest, G.337n6 reasons, significance of308
primitive intentionality62 recursion163
Prince, A.155 reduction, significance of468, 191, 193,
Prinz, J.331n2 (Ch 2) 196, 2478, 3089
privileged access292, 293, 306 reductive naturalism3841
property cluster view, of mind56, 66 reductive physicalism (identity theory)7,
property exemplification theory199 212, 305
propositional a itudes1314, 67, 117, challenge to8
122, 125, 131, 154, 176, 281, 287, 288, referential opacity and1758
298, 306 reflexive consciousness163, 271, 273, 287
beliefs and desires and11112, reflexive relation and identity205, 206,
11516, 123 213
and intentional actions1034 Reid, T.212, 213
propositional content62, 123 representation23, 14, 523, 62, 30910,
psychoanalysis6, 307 2235, 228, 2356, 240, 25967,
psychological behaviourism 27177, 283, 294, 303
see behaviourism cognitive science and1569
psychology computational/representational theory
cognitive6, 108, 151 of thought (CRTT)1725
computational124 content and21, 512, 70, 1359, 142,
folk see folk psychology 171, 175, 189
granny125 ecumenical approaches1879
intentional1235 externalist theories and1817
moral301 and functionalism33
scientific1078, 11719, 121, 124 internalist strategies and17881
383
Index
representation (Contd) semantics1415, 123, 126, 139, 180,

mental causation and21 181, 234, 31112
phenomenal33, 26072 and compositional syntax156, 2356
referential opacity and1758 content and52, 67, 701, 185, 188
see also mental representation promiscuity16, 17
Rey, G.171, 184, 275, 337nn24, sensation129, 131
338nn1314 sensitivities174
Rives, B.253 sensori-motor knowledge and
Robb, D.251 expectations265, 276
Robinson, H.332n1, 333n2 sensory states and intentional states
Rock, I.6 dichotomy556, 61, 65, 68
Rodriguez-Pereyra, G.253 Shafer-Landau, R.247
Root, M.116 Shallice, T.152, 163
Rorty, R.568, 331nn59 Shannon, C. E.67, 337n10
Rosch, E. H.155 Shapiro, L.228
Rosenschein, S. J.160 Shoemaker, S.251, 261, 262, 272
Rosenthal, D. M.38, 50, 271 Siegel, S.265
Rosner, B. S.163 Simon, H.152, 153, 159, 167, 223, 2334
Ross, D.333n10 Simons, D. J.161
Rowlands, M.222, 339n2, 340n5 simulation theory128
Rozemond, M.333n3 single property view, of mind
Rumelhart, D. E.155 incorrigibility569
Rupert, R.339n2 intentionality604
Russell, B.27, 283, 309 singular externalism1379
Ryle, G.174, 283, 284 singular internalism139
The Concept of Mind302 Slocum, A. C.229
Sloman, A.152, 153, 157, 158, 164, 165,
Salmon, N.335n11 167, 168, 169, 170
Samuels, R.153 Smart, J. J. C2, 32
Sawyer, S.133, 275, 335nn13, 18, Smith, A. D.265
336nn27, 3740 Smith, B. C.117, 131, 167, 168, 169
Scheier, C.2245, 228 Smith, Barry C.102
Scheutz, M.167 Smith, T.167
Schier, S.336n34 Smolensky, P.155, 156
Schouten, M.336n24 Soames, S.335n11
scientific psychology1078, 11719, Sober, E.155
121, 124 social externalism1367
Davidson against117 Socrates73
Scriven, M.165 souls
Seager, W.249 and identity criterion218
Searle, J. R.646, 69, 154, 162, 165, 166, and physicalism989
168, 174, 285, 331n4, 332n16 and substance dualism79, 81, 834
seeing-as129 Sparber, G.197
Segal, G.19, 335nn12, 15 special sciences1237
Sejnowski, T, J.155 speech acts49
self, significance of310 Stalnaker, R.183, 257, 275
self-a ributions149 Stampe, D.183
self-awareness37, 209, 310 standing states56
self-consciousness37, 271, 287, 31011 Sterelny, K.2, 14, 182
reflexive163 Stich, S.13, 15, 330n6, 335n16
self-knowledge107, 134, 1457, 292, Stoljar, D.26, 258, 333n9
306, 311, 314 Stone, T.128
384
Index
Stoutland, F.191 transcendental realism25

Strawson, G.2756, 283, 305 transitivity relation and identity205,
Strawson, P. F.217 206, 212
strong eliminativism27 transparency38, 140, 168, 1756, 287,
strongly optimistic physicalism25 314
strongly pessimistic physicalism25 truth-making2523
Sturgeon, S.256 Turing, A. M.166, 172, 174, 225, 283,
subjective awareness254, 26774 314
substance dualism73, 77, 332n1, 333n2 Turing-computation152
consciousness and8990 Turing machines167, 175, 31415, 337n4
freedom and879 Turing test315
identification problem7883 Turner, M.169
interaction problems and835 Twin Earth18, 124, 180, 295, 31516,
personal identity and857 334n3 (Ch 7), 335n4
significance of747 two-factor theory
see also interactive substance dualism of predicative thought, of141, 142
substantive event memory147 singular thought1412
superbelief284 Tye, M.32, 38, 51, 60, 61, 62, 68, 69, 255,
supervenience1011, 52, 967, 111, 257, 260, 262, 263, 267, 331n2 (Ch 2),
133, 196, 242, 245, 247, 31213, 331n10, 336n26
338n4 type identities7, 10
Su on, J.231 type-physicalism143, 335n20
Swampman11, 18 type-type identity40
Swinburne, R.218, 332n1, 333nn2, 11
Sylvan, R.167 Ullman, S.158
symbol167 uncommi ed physicalism256, 30
symmetry relation and identity205, unconscious intentional state65
206, 213 upward determination967
systematicity, of cognitive
performance2356 van Gelder, T. J.157, 162
system view56, 6672 Van Gulick, R.29, 323
consciousness and646 Varela, F. J.161, 162, 166
variable realization242
teleofunctionalism185, 31314 Velleman, J. D.278
teleological theory of content17 via negativa physicalism101
Thagard, P.155, 169 Vicente, A.197
theory of mind104 virtual machines50, 152, 156, 166,
theory-theory128 167, 170
third-person perspective see first-person/ vision for action66
third-person perspective Von Eckardt, B.12
thisness79 von Wright, G.280
Thomasson, A.23
Thompson, E.43, 52 Walk, R. D.158
Thornton, C.170, 224, 228 Wallace, A. F. C.159
thorough-going internalism1401 Warfield, T. A.336n26
thought experiment79, 89, 255 weak eliminativism278
threshold value see neural network weakly optimistic physicalism25
token184 weakly pessimistic physicalism25
-token identity154 Weather Watchers283
token-physicalism143, 335n20 Weaver, W.337n10
total implementation sensitivity227 Wedgwood, R.337n2
Touretsky, D. S.156 Weiskrantz, L.66
385
Index
what-it-is-likeness31, 32, 378, 49, 51, Wi genstein, L.155, 174, 178, 314, 338n11
8990, 255, 2589, 268, 308, 31617 Philosophical Investigations283
Wheeler, M.157, 162, 220, 222, 223, Wollheim, R.131
232, 233, 276, 339n2 Woodru, G.104
Whitby, B.169 Wright, C.336n37
White, S.188
Whyte, J. T.275 XYZ see H2O XYZ
Wiggins, D.215, 216
Wigner, E.100 Yablo, S.23, 252, 330n2
Wilkerson, T. E.335n5 Yli-Vakkuri, J.339nn9, 11
will, significance of317
Williamson, T.179 Zahavi, D.330n1 (Ch 2)
Wilson, J.245, 334n3, 335n23, 339n2 Zi, P.164
Witmer, G.95 zombies24, 34, 434, 97, 164, 31718
386

GARVEY, James. Companion - Philosophy - Mind - Continuum PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

GARVEY, James. Companion - Philosophy - Mind - Continuum PDF

Transféré par

Droits d'auteur :

Formats disponibles

The Continuum Companion to Philosophy of Mind

The Continuum Companions series is a major series of single volume companions

The Continuum Companion to Continental Philosophy, edited by John Mullarkey

James Garvey and Contributors, 2011

All rights reserved. No part of this publication may be reproduced or transmi ed

British Library Cataloguing-in-Publication Data

Library of Congress Cataloging-in-Publication Data

Typeset by Newgen Imaging Systems Pvt Ltd, Chennai, India

1Problems, Questions and Concepts in the Philosophy of Mind 1

13Current Issues in the Philosophy of Mind 239

Fred Adams is Professor of Philosophy, and Chair and Professor of Linguistics

Steve Beighley is on the Neurophilosophy Track at Georgia State University.

Margaret A. Boden is a Fellow of The British Academy, and Research Professor

Mark Cain is Assistant Head of the Department of Religion and Philosophy,

Neil Campbell is Professor of Philosophy at Wilfrid Laurier University. He

Daniel D. Huo is Professor of Philosophical Psychology at the University of

Dale Jacquee is Lehrstuhl ordentlicher Professur fr Philosophie, Schwerpunkt

E. J. Lowe is Professor of Philosophy at Durham University, UK. His books

T. J. Mawson is Fellow and Tutor in Philosophy at St Peters College, Oxford.

Barbara Montero is Associate Professor of Philosophy at the City University of

Paul Noordhof is Anniversary Professor of Philosophy at the University of

Dan OBrien is Research Fellow at Oxford Brookes University, Honorary

Dimitris Platchias is a lecturer at the University of Glasgow. He is the author of

Ian Ravenscro is Associate Professor of Philosophy at Flinders University.

Georges Rey is Professor of Philosophy at the University of Maryland at

Constantine Sandis is a Senior Lecturer in Philosophy at Oxford Brookes

Sarah Sawyer is Senior Lecturer at the University of Sussex. Her research

Barry C. Smith is Professor of Philosophy and Director of the Institute of

Michael Wheeler is Professor of Philosophy at the University of Stirling. Prior

Problems, Questions and Concepts in the Philosophy of Mind

This overview by Ian Ravensco contains a mix of introductory and advanced

Current Issues in the Philosophy of Mind

is not soothsaying. Instead, Noordhof provides an overview and explanation

The bibliography, compiled by Adam Ferner, includes details of all the

You might acclimatize yourself by having a look at the Introduction in

The introduction to nearly every companion volume to a philosophical topic

It turns out that the philosophy of mind really is undergoing a remarkably

an eort to find something be er than dualism. The philosophy of mind has

Philosophers of mind, however, are as likely to march forward with thought

A Useful, Argumentative Companion

Overview of the Contributions

Problems, Questions, and Concepts in the Philosophy of Mind

The Mark of the Mental

Fred Adams and Steve Beighley

Folk Psychology and Scientic Psychology

Internalism and Externalism in Mind

She goes on to outline metaphysical considerations having to do with

The Philosophies of Cognitive Science

representational theory of thought to make sense of such states within a

Embodied Cognition and the Extended Mind

Current Issues in the Philosophy of Mind

physicalism how we now characterize it and how this characterization aects

3. Behavioural and Cognitive Sciences: In addition to the vigorous development

developing and manipulating representations. The third section briefly surveys

Mind and Body

Many problems in the philosophy of mind emerge because we are deeply

The great French philosopher, mathematician and scientist, Ren Descartes,

Elizabeth challenged Descartes to account for the causal interactions

Position 1 Modern science is wrong: the world is not physically closed.

Position 2 Event P is over-determined. That is, event M is a sucient cause of

Position 1 is una ractive because it amounts to endorsing substance dualism

physical brain property P*. Consequently, interactive property dualism is

Position 1 Modern science is wrong: the world is not physically closed.