Вы находитесь на странице: 1из 676


qxd 1/12/09 3:01 PM Page i

A Companion to

A Companion to Metaphysics, Second Edition Edited by Jaegwon Kim, Ernest Sosa, and Gary S. Rosenkrantz
2009 Blackwell Publishing Ltd. ISBN: 978-1-405-15298-3
9780631199991_1_pre.qxd 1/12/09 3:01 PM Page ii

Blackwell Companions to Philosophy

This outstanding student reference series offers a comprehensive and authoritative survey of philosophy as
a whole. Written by todays leading philosophers, each volume provides lucid and engaging coverage of the
key figures, terms, topics, and problems of the field. Taken together, the volumes provide the ideal basis for
course use, representing an unparalleled work of reference for students and specialists alike.

Already published in the series: 21. A Companion to Genethics

Edited by Justine Burley and John Harris
1. The Blackwell Companion to Philosophy, 22. A Companion to Philosophical Logic
Second Edition Edited by Dale Jacquette
Edited by Nicholas Bunnin and Eric Tsui-James
23. A Companion to Early Modern Philosophy
2. A Companion to Ethics Edited by Steven Nadler
Edited by Peter Singer
24. A Companion to Philosophy in the Middle Ages
3. A Companion to Aesthetics, Second Edition
Edited by Jorge J. E. Gracia and Timothy B. Noone
Edited by Stephen Davies, Kathleen Higgins, Robert
Hopkins, Robert Stecker, and David E. Cooper 25. A Companion to African-American Philosophy
Edited by Tommy L. Lott and John P. Pittman
4. A Companion to Epistemology
Edited by Jonathan Dancy and Ernest Sosa 26. A Companion to Applied Ethics
Edited by R. G. Frey and Christopher Heath Wellman
5. A Companion to Contemporary Political
Philosophy (two-volume set), Second Edition 27. A Companion to the Philosophy of Education
Edited by Robert E. Goodin and Philip Pettit Edited by Randall Curren
6. A Companion to Philosophy of Mind 28. A Companion to African Philosophy
Edited by Samuel Guttenplan Edited by Kwasi Wiredu
7. A Companion to Metaphysics, Second Edition 29. A Companion to Heidegger
Edited by Jaegwon Kim, Ernest Sosa and Gary S. Edited by Hubert L. Dreyfus and Mark A. Wrathall
30. A Companion to Rationalism
8. A Companion to Philosophy of Law and Legal Edited by Alan Nelson
31. A Companion to Ancient Philosophy
Edited by Dennis Patterson
Edited by Mary Louise Gill and Pierre Pellegrin
9. A Companion to Philosophy of Religion
32. A Companion to Pragmatism
Edited by Philip L. Quinn and Charles Taliaferro
Edited by John R. Shook and Joseph Margolis
10. A Companion to the Philosophy of Language
33. A Companion to Nietzsche
Edited by Bob Hale and Crispin Wright
Edited by Keith Ansell Pearson
11. A Companion to World Philosophies
34. A Companion to Socrates
Edited by Eliot Deutsch and Ron Bontekoe
Edited by Sara Ahbel-Rappe and Rachana Kamtekar
12. A Companion to Continental Philosophy
35. A Companion to Phenomenology and
Edited by Simon Critchley and William Schroeder
13. A Companion to Feminist Philosophy Edited by Hubert L. Dreyfus and Mark A. Wrathall
Edited by Alison M. Jaggar and Iris Marion Young
36. A Companion to Kant
14. A Companion to Cognitive Science Edited by Graham Bird
Edited by William Bechtel and George Graham
37. A Companion to Plato
15. A Companion to Bioethics Edited by Hugh H. Benson
Edited by Helga Kuhse and Peter Singer
38. A Companion to Descartes
16. A Companion to the Philosophers Edited by Janet Broughton and John Carriero
Edited by Robert L. Arrington
39. A Companion to the Philosophy of Biology
17. A Companion to Business Ethics Edited by Sahotra Sarkar and Anya Plutynski
Edited by Robert E. Frederick
40. A Companion to Hume
18. A Companion to the Philosophy of Science Edited by Elizabeth S. Radcliffe
Edited by W. H. Newton-Smith
41. A Companion to the Philosophy of History and
19. A Companion to Environmental Philosophy Historiography
Edited by Dale Jamieson Edited by Aviezer Tucker
20. A Companion to Analytic Philosophy 42. A Companion to Aristotle
Edited by A. P. Martinich and David Sosa Edited by Georgios Anagnostopoulos
9780631199991_1_pre.qxd 1/12/09 3:01 PM Page iii

A Companion to
Second Edition

Edited by

A John Wiley & Sons, Ltd., Publication

9780631199991_1_pre.qxd 1/12/09 3:01 PM Page iv

This second edition first published 2009

2009 Blackwell Publishing Ltd
except for editorial material and organization 2009 by Jaegwon Kim, Ernest Sosa and Gary S.

Edition history: Blackwell Publishers Ltd (1e, 1996)

Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwells publishing
program has been merged with Wileys global Scientific, Technical, and Medical business to form

Registered Office
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom

Editorial Offices
350 Main Street, Malden, MA 02148-5020, USA
9600 Garsington Road, Oxford, OX4 2DQ, UK
The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

For details of our global editorial offices, for customer services, and for information about how
to apply for permission to reuse the copyright material in this book please see our website at

The right of Jaegwon Kim, Ernest Sosa and Gary S. Rosenkrantz to be identified as the author of the
editorial material in this work has been asserted in accordance with the Copyright, Designs and Patents
Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise,
except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of
the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may
not be available in electronic books.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand
names and product names used in this book are trade names, service marks, trademarks or registered
trademarks of their respective owners. The publisher is not associated with any product or vendor
mentioned in this book. This publication is designed to provide accurate and authoritative information in
regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in
rendering professional services. If professional advice or other expert assistance is required, the services of
a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data

A companion to metaphysics / edited by Jaegwon Kim, Ernest Sosa, and Gary S. Rosenkrantz. 2nd ed.
p. cm. (Blackwell companions to philosophy)
Includes bibliographical references and index.
ISBN 978-1-4051-5298-3 (hbk. : alk. paper) 1. MetaphysicsDictionaries. I. Kim, Jaegwon.
II. Sosa, Ernest. III. Rosenkrantz, Gary S.
BD111.C626 2009
110.3 dc22

A catalogue record for this book is available from the British Library.

Set in 10/12.5pt Photina by Graphicraft Limited, Hong Kong

Printed in Singapore by Utopia Press Pte Ltd

1 2009
9780631199991_1_pre.qxd 1/12/09 3:02 PM Page v


List of Contributors vii

Introduction xiii

Part I Extended Essays 1

Part II Metaphysics From A to Z 95

Index 637

9780631199991_1_pre.qxd 1/12/09 3:03 PM Page vii


Felicia Ackerman John Bigelow

Brown University Monash University, Australia

Robert Ackermann Akeel Bilgrami

University of Massachusetts, Amherst Columbia University

Marilyn McCord Adams John Biro

Christ Church, Oxford University of Florida

Robert Merrihew Adams Simon Blackburn

Yale University and Mansfield College, Oxford University of Cambridge and the University
of North Carolina at Chapel Hill
Jan A. Aertsen
Thomas-Institut at the University of Cologne, Ned Block
Germany New York University

C. Anthony Anderson
Paul A. Boghossian
University of California, Santa Barbara
New York University
Richard E. Aquila
M.B. Bolton
University of Tennessee
Rutgers University
D.M. Armstrong
Michael E. Bratman
University of Sydney, Australia
Stanford University
Keith Arnold
University of Ottawa, Canada Harold I. Brown
Northern Illinois University
Bruce Aune
University of Massachusetts, Amherst Douglas Browning
University of Texas at Austin
Thomas Baldwin
University of York Panayot Butchvarov
University of Iowa
George Bealer
Yale University Robert E. Butts

Frederick Beiser Alex Byrne

Syracuse University Massachusetts Institute of Technology

9780631199991_1_pre.qxd 1/12/09 3:03 PM Page viii

c on t r i buto r s
Steven M. Cahn Dagfinn Fllesdal
City University of New York, Graduate Stanford University and University
School of Oslo, Norway

Keith Campbell Matthew Davidson

University of Sydney, Australia California State University, Bernardino

Albert Casullo Graeme Forbes

University of Nebraska, Lincoln University of Colarado at Boulder

Peter Caws Richard Fumerton

George Washington University University of Iowa

Richard M. Gale
Arindam Chakrabarti
University of Pittsburgh
University of Hawaii at Manoa
Richard Gallimore
John J. Compton
University of North Carolina at Greensboro
Vanderbilt University, Tennessee
Patrick Gardiner
John Corcoran
State University of New York at Buffalo Brian Garrett
Australian National University
John Cottingham
University of Reading Don Garrett
New York University
Edwin Curley
University of Michigan Rolf George
University of Waterloo, Canada
Helen Daly
University of Arizona Roger F. Gibson
Washington University in Saint Louis
Harry Deutsch
Illinois State University Carl Ginet
Cornell University
Cora Diamond
University of Virginia Peter Godfrey-Smith
Harvard University
Alan Donagan
Thomas A. Goudge
Rolf A. Eberle Jorge J.E. Gracia
University of Rochester, New York State University of New York at Buffalo

Catherine Z. Elgin John Greco

Harvard University Saint Louis University

Fred Feldman Reinhardt Grossmann

University of Massachusetts, Amherst Indiana University

9780631199991_1_pre.qxd 1/12/09 3:03 PM Page ix

cont ri but ors

Rick Grush Paul Horwich
University of California, San Diego New York University

Charles Guignon Paul Hovda

University of South Florida Reed College, Oregon

Bob Hale M.J. Inwood

University of Sheffield Trinity College, Oxford

Michael Hallett Frank Jackson

McGill University, Montreal Australian National University

Chad D. Hansen Janine Jones

University of Hong Kong University of North Carolina at Greensboro

Ross Harrison Lynn S. Joy

Kings College, Cambridge University of Notre Dame

W.D. Hart Robert Kane

University of Illinois, Chicago University of Texas at Austin

John Heil Christopher Kirwan

Washington University in Saint Louis Exeter College, Oxford

Mark Heller Arnold Koslow

Syracuse University Brooklyn College

Risto Hilpinen John Lachs

University of Miami Vanderbilt University, Tennessee

Eli Hirsch Karel Lambert

Brandeis University, Massachusetts University of California, Irvine

Herbert Hochberg Stephen Leeds

University of Texas at Austin University of Wisconsin, Milwaukee

Joshua Hoffman Keith Lehrer

University of North Carolina at Greensboro University of Arizona and University
of Miami
Christopher Hookway
University of Sheffield Ramon M. Lemos

James Hopkins Jarrett Leplin

Kings College, London University of North Carolina at Greensboro

Terence E. Horgan Ernest LePore

University of Arizona Rutgers University

9780631199991_1_pre.qxd 1/12/09 3:03 PM Page x

c on t r i buto r s
John Leslie Trenton Merricks
University of Guelph University of Virginia

Barry Loewer Phillip Mitsis

Rutgers University New York University

J.M. Moravcsik
Lawrence B. Lombard
Stanford University
Wayne State University, Michigan
Alexander P.D. Mourelatos
Douglas C. Long University of Texas at Austin
University of North Carolina at Chapel Hill
Kevin Mulligan
E.J. Lowe University of Geneva, Switzerland
Durham University
Daniel Nolan
William Lyons University of Nottingham
Trinity College, Dublin
Martha C. Nussbaum
Charles J. McCracken University of Chicago
Michigan State University
David S. Oderberg
Neil McKinnon University of Reading
Monash University, Australia
Anthony OHear
University of Buckingham
Brian P. McLaughlin
Rutgers University
Dominic J. OMeara
University of Fribourg, Switzerland
Ernan McMullin
University of Notre Dame Walter R. Ott, Jr.
Virginia Polytechnic Institute and State
Joseph Margolis University
Temple University, Pennsylvania
George S. Pappas
Ned Markosian Ohio State University
Western Washington University, Washington
Terence Parsons
Gareth B. Matthews University of California Los Angeles and
University of Massachusetts, Amherst University of California, Irvine

Anthonie Meijers David Pears

Delft University of Technology, Netherlands Christ Church, Oxford

Alfred R. Mele Terence Penelhum

Florida State University University of Calgary, Canada

Joseph Mendola Alvin Plantinga

University of Nebraska at Lincoln University of Notre Dame

9780631199991_1_pre.qxd 1/12/09 3:03 PM Page xi

cont ri but ors

Ruth Anna Putnam Wesley C. Salmon
Wellesley College, Massachusetts
David H. Sanford
Diana Raffman Duke University
University of Toronto
Geoffrey Sayre-McCord
Peter Railton University of North Carolina at Chapel Hill
University of Michigan
Richard Schacht
Andrew J. Reck University of Illinois at Urbana-Champaign
Tulane University, Louisiana
Frederick F. Schmitt
Nicholas Rescher Indiana University
University of Pittsburgh
Richard Schmitt
Thomas Ricketts Brown University
University of Pittsburgh
George Schumm
Richard Robin Ohio State University
Mount Holyoke College, Massachusetts
Jorge Secada
Gideon Rosen University of Virginia
Princeton University
Charlene Haddock Seigfried
Jay F. Rosenberg Purdue University

Gary S. Rosenkrantz David Shatz

University of North Carolina at Greensboro Yeshiva University, New York

David M. Rosenthal Fadlou Shehadi

City University of New York, Rutgers University
Graduate School
Donald W. Sherburne
David-Hillel Ruben Vanderbilt University
Birbeck College, London and New York
University in London Sydney Shoemaker
Cornell University
Paul Rusnock
University of Ottawa, Canada Peter Simons
University of Leeds
Nils-Eric Sahlin
Lund University, Sweden Lawrence Sklar
University of Michigan
R.M. Sainsbury
University of Texas at Austin and Kings John Skorupski
College, London University of St Andrews

Nathan Salmon Robert C. Sleigh, Jr.

University of California, Santa Barbara University of Massachusetts, Amherst

9780631199991_1_pre.qxd 1/12/09 3:03 PM Page xii

c on t r i buto r s
Barry Smith Michael Tye
State University of New York at Buffalo University of Texas at Austin

Quentin Smith James Van Cleve

Western Michigan University University of Southern California

Elliott Sober J. David Velleman

University of Wisconsin, Madison New York University

Roy A. Sorensen Georgia Warnke

Dartmouth College University of California, Riverside

Robert Stalnaker Joan Weiner

Massachusetts Institute of Technology Indiana University

Howard Stein Nicholas White

University of Chicago University of Utah

S.G. Williams
Guy Stock
Worcester College, Oxford
University of Dundee
Kenneth P. Winkler
Richard Swinburne
Yale University
University of Oxford
Kwasi Wiredu
Paul Teller University of South Florida
University of California, Davis
Allen W. Wood
Neil Tennant Stanford University
Ohio State University
Larry Wright
Amie L. Thomasson University of California, Riverside
University of Miami
Edward N. Zalta
James E. Tomberlin Stanford University

Martin M. Tweedale Dean Zimmerman

University of Alberta, Canada Rutgers University

9780631199991_1_pre.qxd 1/12/09 3:03 PM Page xiii

jaegwon kim, ernest sosa, and gary rosenkrantz

Because it is the most central and general subdivision of philosophy, and because it is among
the oldest and most persistently cultivated parts of the field, metaphysics raises special difficul-
ties of selection for a companion such as this. The difficulties are compounded, moreover,
by two further facts. First, metaphysics is not only particularly old among fields of phi-
losophy; it is also particularly widespread among cultures and regions of the world. And,
second, metaphysics has provoked levels of skepticism unmatched elsewhere in philosophy;
including skepticism as to whether the whole subject is nothing but a welter of pseudo-
questions and pseudo-problems.
In light of this a project such as ours needs to delimit its approach. In accomplishing this,
we had to bear in mind the space limitations established by the series, and also the fact that
other volumes in the series would be sure to cover some questions traditionally viewed as
metaphysical. These considerations led to our including some such questions, which we
thought would be covered more extensively in Samuel Guttenplans A Companion to the
Philosophy of Mind, for example, or in Peter Singers A Companion to Ethics, but which should
be treated in this Companion, if only briefly and for the sake of a more complete and self-
contained Companion to Metaphysics. In addition, we tried to give a good sense of the sorts
of skeptical objections that have been raised to our field as a whole. As for the spread of
metaphysics across cultures, traditions, and regions of the world, we opted again to include
some coverage of the non-western, while at the same time keeping our focus firmly on the
western tradition from the Greeks to the present. What is more, even within the western
tradition we needed to be selective, especially once we came to the present century.
Philosophy in the present century has grown explosively, especially in the so-called ana-
lytic traditions common to North America and the British Commonwealth countries, along
with Scandinavia and some enclaves in the rest of Europe and on other continents. Our
focus has been for the most part on these traditions, although, again, as with non-western
traditions, we have paid some attention to the schools and traditions that have flourished
best in Continental Europe.
We had to be selective also in our treatment of contributors to metaphysics. The account
of the work of a philosopher included in our Companion will most often reflect the contri-
butions of that philosopher to metaphysics. A certain artificiality is therefore inevitable, and
readers should bear this in mind. Other companions in our series will, therefore, provide,
at least sometimes, a helpful supplement to the discussions of individual figures found in
these pages.
In this second edition of the Companion, many of the first edition articles have been updated;
additions include more than thirty new entries on important contemporary contributors to
metaphysics and a new section of ten extended essays on major topics in metaphysics.
Cross-references will be made by the use of small capitals, both in the text and at the end
of each article.
Brown University, Rutgers University, and the University of North Carolina at Greensboro

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 1

Part I

Extended Essays

A Companion to Metaphysics, Second Edition Edited by Jaegwon Kim, Ernest Sosa, and Gary S. Rosenkrantz
2009 Blackwell Publishing Ltd. ISBN: 978-1-405-15298-3
9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 3

Causation Causes and effects are often not contiguous.

A switch on the wall is distant from the
Making something happen, allowing or electric light overhead that it controls.
enabling something to happen, or prevent- Pulling a button on an alarm clock makes
ing something from happening. Mental and it ring six hours later. The New York per-
extra-mental occurrences, of all spatial and formance of three musicians in 1937 con-
temporal dimensions, great and small, have tributes causally to what one hears on the
causes and are causes. Our awareness of Perth radio in 2007. Although intervals
the world and our action within the world of space, time, or spacetime separate the
depends at every stage on causal processes. causes and effects in these examples, spatio-
Although not all explanations are causal, temporally continuous causal paths connect
anything that can be explained in any way them. The path has no spatial or temporal
can be explained causally. Like other meta- gaps or breaks. (A rigorous definition of
physical concepts, the concept of causation continuity requires the notion of a limit
applies very broadly. Yet this fundamental found in calculus textbooks.) The path is
concept continues to elude metaphysical causal because for any two positions, a and
understanding. While there is some general c, on the path, there is an intermediate
philosophic agreement about causation, there position b on the path such that either
is also considerable disagreement. Causal something at a causes something at b that
theories of knowledge, perception, memory, causes something at c, or the causation runs
the mind, action, inference, meaning, refer- in the other direction, cba. An explanation of
ence, time, and identity through time, take what constitutes a causal path that does
a notion as fundamental that philosophers not use the notion of causation would serve
understand only incompletely. as a reductive definition of causation. The
HUME is the dominant philosopher of explanation above, which uses the notion
cause and effect. A running commentary of causation explicitly, serves only to state
on Humes views and arguments, pro and a spatio-temporal necessary condition of
con, could cover most contemporary philo- causation.
sophical concerns with causation (Hume, Causes and effects are events. This is a
1739, esp. Bk. I, Pt. III; Hume, 1748, esp. majority view (see Davidson, 1980). Idiom-
sects. IV, V, VII). According to Hume, it is atic speech often mentions something other
not the experience of an individual causal than a change, or non-change, or occur-
transaction, but experience of other trans- rence, as a cause or effect, as in Richard
actions, relevantly similar, that provides what makes me furious. The question is whether
causation involves in addition to priority an available paraphrase such as Reading
and contiguity. Experiences of regularities or what Richard writes makes me become
constant conjunctions condition our expec- furious brings events back into the picture
tations. We project our conditioned feelings as causes and effects. If both causes and
of inevitability on external objects as a kind effects are always of the same kind, then
of necessity that resides in the objects them- causal paths can continue indefinitely both
selves (see Hume, 1748, sect. VII). Limitations from the past and into the future. On the
of space preclude extensive quotation and dis- other hand, the strategy of reducing all causal
cussion of these and other primary texts. statements by paraphrase to statements
A number of paragraphs in this entry about events does not convince philo-
begin with the statement of a view about sophers who hold that sometimes facts, pro-
causation. The next sentence then classifies perties, or aspects of events are irreducible
the view as prevailing, majority, controversial, relata of causal relations (see Sanford, 1985).
or minority. Some of these classifications may Some philosophers who concentrate on
themselves be controversial. Their purpose questions of agency and freedom entertain
is only to help organize the entry. views of agent causation: in human action
Continuous causal paths connect causes a person is an irreducible cause (see action
with their effects. This is a prevailing view. theory). Although Lucys putting on her

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 4

c a u s a t io n
shoes involves many instances of event from observing its sensible qualities, we
causation, the ultimate cause of Lucys shoes cannot figure out a things causal capacities.
being put on is Lucy herself. And when we do come to believe, from a
Causation is the transfer of something much broader experience, what they are,
from cause to effect. This is a controversial our evidence does not entail our conclu-
view. In one version of this view, causation sion. It is still logically possible that any-
transfers some quantity subject to a conser- thing will happen next. Our beliefs about
vation law of physics. Hans reichenbach the physical properties of belts and pulleys are
propounded and Wesley salmon developed fallible and based on more than an initial
another version in terms the mark trans- visual impression. Still, given the physical
mission of a mark, a modification that properties of the belt and pulley, the spatial
satisfies certain requirements. The trans- relations between them, and the assump-
mission of a mark between processes is a tion that the belt moves in a certain direc-
transmission of structure. There are clear tion, one can figure out which way the
positive instances of this view. One contro- pulley rotates. Although one can draw on
versy involves the generalization of these experience of similar set-ups that involve
instances. Another questions whether the belts and pulleys when closing the final
application of a notion such as mark gap of causal inference, it is unnecessary
requires some prior causal commitment. to do so. Reason can bridge the gap un-
There is no element of genuine a priori aided by additional experience (see Sanford,
reasoning in causal inference. This is a 1994).
majority view. Most philosophers believe By the very nature of causation, effects are
that Hume refuted the rationalists (see never earlier than their causes. This is a
rationalism) before him (such as Spinoza, majority view. Mackie (1974, ch. 7) dis-
Descartes and, on this issue, Hobbes) and the cusses the conceptual possibility of backward
idealists after him (such as McTaggart causation and provides further references.
and Blanshard) who hold that causation There are also serious philosophical discus-
is intrinsically intelligible. Given a determin- sions of the conceptual possibility of time
ate event, according to Hume, anything travel in which in there are closed causal
might happen next, so far as reason and loops (see Lewis, 1986).
logic are concerned. The contrary of every By the very nature of causation, causes
matter of fact is still possible; because it can are always earlier than their effects. This is
never imply a contradiction (Hume, 1748, a controversial view. Other requirements of
p. 25). Cause and effect are distinct exist- causal connection are symmetric in form;
ences, and the mind never perceives any they do not distinguish effects from causes.
real connexion among distinct existences Defining causal priority in terms of tem-
(Hume, 1739, p. 636). Reason by itself can- poral priority thus has theoretical appeal.
not predict what will happen next after one But there is also a theoretical drawback:
billiard ball bumps into another. But from the equally appealing account of temporal
what should one attempt to make such pre- priority by reference to causation will be
dictions, from descriptions of the events in circular if the explanation of causal priority
question? If so, which logical relations do or is to be temporal. Moreover, simultaneous
do not obtain will depend on the nature causation appears not only to be possible, but
of the description. Any event has logically actual. Physics assures us that much of
independent descriptions, and any two events this appearance is illusion. Since nothing
have descriptions that are not logically transmits motion faster than the speed of
independent (see Davidson, 1980, essay 1). light, the motion of ones fingers, that grip
The view that there is at least sometimes an the handle of a teaspoon, does not, strictly
intelligible connection between cause and speaking, cause the simultaneous motion of
effect does not rely on inventing clever the bowl of the spoon. Other cases of appar-
descriptions. Rather, it concedes a lot to ent simultaneous causation, however, do
Hume without conceding everything. Just not involve bridging a spatial gap, as when

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 5

a moving belt turns a pulley with which it causeeffect pairs are really coincident joint
is in direct contact. effects of a common cause, such as God.
We cannot directly perceive causal rela- Current discussions of causation disregard
tions. This is a majority view that Hume such views and take it for granted that
influences greatly with his example of the manipulations are causes.
impact of billiard balls. We can see motions Causation depends on manipulation; a
and changes in motion in the balls. We can correct general account of causation is in
see that one ball touches the other immedi- terms of manipulation. This is a minority
ately before the second begins to move. We view. Just because one might reach this
cannot see that there is a causal relation view by means of the conceptual fallacy dis-
between the two motions. Nor can we tell, cussed above, that does nothing to prove
just by observing the sensible qualities of a it false. When distinguished from a view
thing, what are its causal capacities and about relations between concepts, however,
dispositions. the theory must deal, by appeal to analogy
Our sense of touch and our perceptions of or imagination, with causal instances in
the positions and movements of our limbs which humans do not and sometimes can-
enable our direct perception of causal rela- not actually participate, such as those that
tions (see von Wright, 1971, pp. 6674). involve clusters of galaxies.
This is a minority view. The causal rela- A correct general account of causation is
tions between ones arm movement and the in terms of intervention. This is a contro-
movement of a cue stick one grasps is a versial view, which is currently the center of
more promising candidate for an object of a robust research program (see Woodward,
direct perception than the impact of billiard 2003). This program is careful to dis-
balls that is merely seen. The conceptual fal- tinguish its technical term intervention
lacy (here so named) may be tempt one from the ordinary term manipulation.
here. This is a mistaken inference of the Manipulations are performed by agents.
form that since we cannot conceive of A While agents also intervene, some natural
without having the concept of B, therefore processes that involve no agents, directly or
the existence of A requires the existence indirectly, are also called interventions. On
of B. It views ontological dependence as the other hand, the notion of an intervention
following from conceptual dependence. is explicitly causal. Its descriptions use
Granted the minority view that our concep- the notion of a causal path. Not all of the
tion of causation depends on our conceptions descriptions in the literature are equivalent.
of ourselves as agents who make things Here is one description:
happen in the physical world, and as patients
INT is an intervention between two vari-
affected by occurrences in the physical world,
ables X and Y on the same causal path if
it does not follow that the existence of cau-
and only if INT completely determines
sation requires the occurrence of such
the value of X; every causal path between
INT (or any cause of INT) and Y goes
Manipulations are causes. This is a pre-
through X; and if there is a causal path
vailing view. Many languages have many
between Z and Y that neither includes
verbs for specific manipulations such as
nor is included by the path between X
cook, shake, turn, and hold that we understand
and Y, INT does not affect Z.
as causal relations. The view is not strictly
a truism since it is inconsistent with seriously Adding fertilizer does not affect the amounts
held positions such as the following. (a) of water and light, which are relevant
There really is no physical world; its ap- variables on causal paths that include the
pearance is an illusion; and from this it growth of tomatoes. According to this
follows that there really are no genuine definition of intervention, does the addition
manipulations or physical causal relations. of fertilizer then intervene on the causal
(b) Although there really are physical events, path between nitrogen level and tomato
those we commonly but wrongly take as growth? Weeds complicate the answer to

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 6

c a u s a t io n
this question. When fertilizer stimulates weed All physical laws are causal laws. This is
growth, a better tomato crop requires pulling a majority view. Some philosophers deny
some weeds and bringing the addition of that all laws of nature, for example Newtons
fertilizer under the general description of first law of motion, are causal laws. Consider
intervention may also require it. a body traveling in a straight line, not
Since intervention is a thoroughly causal changing direction or speeding up or slow-
notion, an interventionist account of a ing down. Where is the causation? Opinions
specific causal connection is not reductive divide on the adequacy of responses such
in the sense of using only non-causal con- as Its motion from B to C is caused by its
cepts. This need not render such accounts cir- immediately prior motion from A to B.
cular. The use of the notion of intervention Events related as cause and effect, when
to support the presence of a specific connec- appropriately described, instantiate a phys-
tion, such as between nitrogen and growth ical law. This is a majority view. These
rate, need not assume its presence to begin appropriate descriptions typically use con-
with. This accords with the function of cepts different from the ones we ordinarily
experiments. Experiment is a thoroughly use in describing the causal transaction.
causal notion, yet we use experiments to Causation in the everyday world supervenes
confirm and to disconfirm causal hypotheses. on causal relations that the fundamental
Theorems about interventions have a laws of nature directly cover. If such super-
wide scope in understanding the roles of venience is universal, there are no causal dif-
experiments in various sciences. This is a ferences without differences of fundamental
controversial view. From a precise definition properties and spatio-temporal arrangements.
of intervention and some strong assump- A singular causal statement need not entail
tions about probabilistic relations between a law, but it does entail that there is a law
variables, theorists prove theorems about that covers, probably as described differ-
directed causal graphs. (There is no attempt ently, the events mentioned (see Davidson,
here to summarize these results.) While 1980, essay 7).
the theorems themselves are neither trivial Causal attribution and the acceptance of
nor controversial, there is not a consensus corresponding conditional statements are
about the manner and scope for their useful closely related. This is a prevailing view.
application to actual causal processes. Hume connects causation with conditionals
Some generalizations that have no excep- in this famous passage:
tions, and some statements of conditional Similar objects are always conjoined with
probability, are causal laws. This is a pre- similar. Of this we have experience.
vailing view. Some universal laws are not Suitable to this experience, therefore, we
causal because they are mathematical or may define a cause to be an object, fol-
logical laws. Some universal truths are not lowed by another, and where all the objects
laws because they are mere accidental similar to the first are followed by objects
regularities. If all swimming birds eat fish, similar to the second. Or in other words,
this does not imply that there is a law-like if the first object had not been, the second
connection between birds swimming and never had existed. (Hume, 1748, p. 76)
their eating fish. Finding evidence against
an accidental regularity, whether quite sur- What Hume puts in other words is scarcely
prising, or not at all surprising, does not a restatement of what goes before. It never-
upset our general theories about the world. theless expresses an important and influen-
Providing a general account of the differ- tial claim, that a cause is necessary for its
ence between laws and accidental general- effect.
ization is a major theoretical see law of Kate turned the key, and the engine
nature undertaking. There are many com- started. But if the engine would have
peting theories about the character of phys- started at that very moment anyway,
ical laws, for example, the view that laws are without Kates key turn, then Kates
relations between properties or universals. turning the key did not start the engine.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 7

If-then statements about what would have anyway. Dorothy was standing by, ready to
happened if something else had occurred throw a stone toward the bottle if Lucy did
are called counterfactuals, contrary-to-fact not. Standby causes, over-determination,
or subjunctive conditionals. A conditional prevention, and other examples serve as
of the form If a had not happened, then b counterexamples to simple formulations of
would not have occurred says that a is counterfactual conditional accounts. This
necessary for b: it is impossible for b to occur leads to formulations that are less simple,
without a. If it is impossible for a which in turn stimulates the invention of
to occur without b, then a is sufficient for b. examples of increasing complexity, and so on,
For example, the downward movement back and forth. (See essays in Collins et al.,
of a lever of the first kind is sufficient for 2004.) Opinions are divided about where
the upward movement of its other end. The this process is leading.
necessity of a for b is often separate from the Replacing the notion of counterfactual
sufficiency for a for b; the thesis that a cause dependence with the notion of influence
is both necessary and sufficient for its effect results in a counterfactual account that runs
is quite strong. Events or conditions we sin- more smoothly. This is a minority view.
gle out as causes often are neither necessary One event influences another when each
nor sufficient for their effects. Adding Bobs belongs to a range of similar events and
Super-Grow fertilizer speeded up the growth there is a range of true counterfactuals of
of the tomato plants, but it was not really the form if event c (in the first range) had
necessary. Other brands would have had occurred, then event e (in the other range)
the same effect. Just by itself, moreover, it would have occurred. A mass hanging on a
also was not sufficient; for other factors, spring influences its length, which varies
independent of adding the fertilizer, such systematically with the mass. (Within a cer-
as light, water and the absence of large tain range of values, the relation between
amounts of concentrated sulfuric acid, were mass and length is invariant. Invariance and
also necessary for the quick growth of the intervention both figure in causal graph
plants. We can still use the notions of neces- theory.) Adding acid to a base exemplifies
sity and sufficiency to spell out the causal causal influence. As more acid is added, more
relevance of adding Bobs Super-Grow to base is neutralized. There is, however, a
the plants rapid growth. It is presumably causal relation in this process that seems
an inus condition of the growth; that is, it is not to fit the definition of influence. As more
an insufficient but non-redundant part of an acid is added, it is not until all the base is neu-
unnecessary but sufficient condition of rapid tralized that the next drop of acid causes a
growth (Mackie, 1974, p. 62). Inus condi- sudden, large increase in acidity (decrease in
tions involve somewhat complicated coun- pH). It remains to be seen how the influence
terfactual conditionals. The pair of simpler view accommodates this and similar tipping
conditionals that express necessity and suf- point examples in which a small event pro-
ficiency, If a had not happened, neither duces large effect by upsetting an equilibrium.
would b and If a had happened, then so Questions of causation, inductive support,
would b together express counterfactual laws of nature, and counterfactual condi-
dependence. tionals are bound closely together. This is a
Causation can be defined in terms of prevailing view. The following distinctions
counterfactual dependence (Lewis, 1986, are closely associated, and any one can ex-
essays 17 and 21). This is a controversial plain the others: acceptable vs. unacceptable
view. Counterexamples provide one source counterfactual conditionals; laws of nature
of controversy. Counterexamples to a claim vs. accidental generalizations; a particular
of the form A=B are in general examples of observations inductively confirming vs. not
A that are not B or examples of B that are not confirming a hypothesis. Acceptable coun-
A. Lucy threw a stone that broke a bottle. terfactual conditionals, but not unaccept-
If Lucy had not thrown the stone, however, able ones, fall under laws (as Chisholm and
a stone would have broken the bottle Goodman have argued). On the other hand,

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 8

c a u s a t io n
laws, but not accidental generalizations, sup- follows from the above standard explana-
port acceptable counterfactuals. Laws, unlike tions of necessary for and sufficient for. This
accidental generalizations, are hypotheses view does not entail the stronger view that
that their instances confirm. These inter- a is a necessary condition of b if, and only if, b
connections, although mutually explana- is a sufficient condition of a. Causal examples,
tory, are arranged in a tight circle and thus among others, show that condition of is not
evoke a sense of theoretical uneasiness. a symmetric relation. The presence of light,
Philosophers who aspire to develop a theory for example, is a causally necessary con-
of causation attempt to break out of the dition of the growth of tomatoes, which is
circle by explaining one distinction in the not in turn a causally sufficient condition
family without appeal to additional distinc- for the presence of light. No one attempts
tions in the same family. Different theories to produce light by growing tomatoes. A
attempt to break out in different places and theory of the direction of conditionship can
also differ in their assignments of explanatory help account for the direction of causation
priority. For example, one theory holds that (Sanford, 1975).
a relation between particulars is causal A totality of conditions necessary for an
when it falls under a law, while another occurrence is jointly sufficient for it. This
holds that a generalization is a law when is a controversial view, and not a logical
particular causal relations fall under it. No truth, in the technical sense of sufficient
views prevail about the best way to achieve spelt out above. There is an ordinary sense
equilibrium in these theoretical matters of sufficient, however, namely enough,
concerning causation. lacks nothing. When everything necessary
An adequate theory of causation should for b obtains, the aggregate is collectively
be in terms of Probability. This is a sufficient for bs occurrence, because jointly
controversial view. When an event causes the members of the aggregate are enough
another, the occurrence of the cause often nothing necessary for b is missing (see
increases the probability of the occurrence of Anscombe, 1981, p. 135). It is not a logical
the other. However this is not always so. contradiction to maintain that an event did
Attempts of formulate universal generaliza- not occur even though nothing necessary for
tions connecting probability with causation its occurrence was missing. This contention
run up against examples such as the fol- runs against the grain of the following
lowing (an earlier example with more details): controversial view:
Lucy aims a stone at a bottle. She throws it, Something necessitates every event. This
and the stone breaks the bottle. Whenever is a controversial view. Although what we
they engage in the sport of throwing stones call a cause often falls far short of being
to break bottles, Dorothy throws a stone if sufficient for its effect, it is common to
Lucy doesnt. Although Lucy often misses, assume that every effect has some, usually
Dorothy almost never misses. Lucy didnt more complicated, sufficient cause. The main
miss this time, however. Her throw broke the issue is not whether some occurrences
bottle. The probability that the bottle would are totally without causal antecedents,
break if she did not throw (and dead-eye but whether, in the technical sense of suf-
Dorothy threw instead) is nevertheless higher ficient, every event has a sufficient cause.
than if she did throw. Qualifications of a pro- If every event has a sufficient cause, and
babilistic account can accommodate partic- every cause is an event, then a classic ver-
ular examples such as this one, but then, sion of Determinism is true. Every event
following a pattern of dialectic common in is a link on a branching chain of causal
technical philosophy generally, and speci- necessitation that runs from the beginning
fically with the associated counterfactual to the end of the universe. The occurrence
accounts of causation, new ingenuous of any event is causally consistent with
counterexamples are not far behind. exactly one set of events causally con-
a is necessary for b if, and only if, b is nectible with it, whether these events are
sufficient for a. This is a prevailing view that earlier or later.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 9

Modern physics, for example in its treat- these corresponding forms are true: when
ment of atomic decay, discourages belief C causes E, a suitably situated relation R
in determinism. Definitions that resemble obtains; and when a suitably situated rela-
Mackies definition of an inus condition pro- tions R obtains, C causes E. (This formula-
vide for the possibility of causation without tion is due to L. Paul.) The Euthyphro
sufficiency: a is a suni condition of b, for Question is whether (a) C causes E because
example, if there is something x such that the R obtains or (b) R obtains because C causes
disjunction a or x is a necessary condition of E. A philosophical reductive definition, acc-
b, and x is not a necessary condition of b ount, or analysis of causation should hope
(Sanford, 1984, p. 58). to give an answer of form (a). Some popular
Accounts of specific causal connections accounts appear to favor answers of form (b).
often refer to causal mechanism. This is a pre- Consider a counterfactual statements and a
vailing view. One of the early truly effective corresponding causal statement:
drugs was aspirin. As everyone knows, it
If Kate had not turned the key, the engine
relieves pain. What scientists did not know,
would not have started.
but for years hoped to find, was the mecha-
Kates turning the key caused the engine
nism of aspirins effect. This goal is different
to start.
from discovering a more general or more
fundamental law. Many scientists try to It is more natural to say that the condi-
understand mechanisms rather that find tional is true because turning the key
general laws that cover certain phenom- caused the engine to start rather than that
ena, and this is true not just in medicine, turning the key caused the engine to
biology, and chemistry, but in many other start because the conditional is true. Some
special sciences. conditionals are true because of causal
A general account of causation should connections; causal connections do not
refer to causal mechanisms rather than obtain because conditionals are true (see
to causal laws. This is a minority view. Sanford, 2003, chs. 1114). Similarly,
Although operations of mechanisms, of causal connections explain the effectiveness
whatever size, seem generally to involve of manipulation rather than the other way
three-dimensional motions, a general theory around. Causal connections also explain
of causation as mechanism would want a the effectiveness of interventions, although
more detailed account of what a mecha- interventionist theory does not represent
nism is. Also, some causal connections are itself as reductive. Theories in terms of the
so direct that there seems to be no room transfer of something, or in terms of under-
for a mediating mechanism. Lucy threw a lying mechanism, whatever their difficulties,
rock that hit a tree before it reached the promise to give appropriate answers to the
wall. The tree interrupted the flight path Euthyphro Question.
of the rock. Where should one look for the In Book II of the Physics, Aristotle dis-
mechanism of this causal interaction? cusses four kinds of aitia or causes. The pre-
In Platos dialogue The Euthyphro Socr- sent article deals only with efficient causes.
ates and Euthyphro reach a point where In the Second Analogy of the Critique
they agree that everything all the gods love of Pure Reason (1781), Kant argues that
is pious and that everything pious all the gods all changes conform to the law of cause
love. Socrates goes on to ask whether all the and effect. In Of Induction, Book III of
gods love pious things because they are A System of Logic (1843), J. S. Mill presents
pious, or whether things are pious because experimental methods for establishing causal
all the gods love them. We may call probing relevance. In his 1912 lecture, On the
questions of this form Euthyphro Questions and Notion of Cause, Russell claims that the law
proceed to ask them about treatments of of causation is a relic of a bygone age; but
causation that aspire to provide reductive Russells own theoretical constructions in
accounts. Suppose that some theory is suf- some later writings depend heavily on causal
ficiently refined that both conditionals of notions.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 10

f i c t i onal entities
b i b l i og rap hy Sanford, D.H.: The Direction of Causation
and the Direction of Conditionship, Journal
Anscombe, G.E.M.: Metaphysics and the of Philosophy 73 (1975), 193207.
Philosophy of Mind, Collected Philosophical Sanford, D.H.: The Direction of Causation
Papers, vol. 2 (Minneapolis: University of and the Direction of Time, Midwest Studies
Minnesota Press, 1981). in Philosophy 9 (1984), 5375.
Beauchamp, T. and Rosenberg, A.: Hume Sanford, D.H.: If P, then Q: Conditionals and
and the Problem of Causation (New York and the Foundations of Reasoning, 2nd edn.
Oxford: Oxford University Press, 1981). (London: Routledge, 2003; originally
Collins, J., Paul, L., and Hall, N., ed.: Coun- published 1989).
terfactuals and Causation (Cambridge, MA: Schaffer, J.: The Metaphysics of Causation,
MIT Press, 2004) in the Stanford Encyclopedia of Philosophy.
Davidson, D.: Essays on Actions and Events The Stanford Encyclopedia of Philosophy
(Oxford: Oxford University Press, 980). (http://plato.stanford.edu) is an online
Dowe, P.: Causal Process, in the Stanford resource of substantial entries that typic-
Encyclopedia of Philosophy. ally have helpful bibliographies. Entries
Faye, J.: Backward Causation, in the Stan- undergo periodic revision.
ford Encyclopedia of Philosophy. Strawson, G.: The Secret Connexion: Cau-
Hausman, D.M.: Causal Asymmetries (Cam- sation, Realism and David Hume (Oxford:
bridge: Cambridge University Press, 1998). Oxford University Press, 1989).
Hitchcock, C. Probabilistic Causation, in the Woodward, J.F.: Causation and Manipul-
Stanford Encyclopedia of Philosophy. ability, in the Stanford Encyclopedia of
Hume, D.: Enquiry Concerning Human Under- Philosophy.
standing (London, 1748): ed. L.A. Selby- Woodward, J.F.: Making Things Happen: A
Bigge (Oxford: Oxford University Press, Theory of Causal Explanation (New York:
1894); 3rd edn. rev. P.H. Nidditch (Oxford: Oxford University Press, 2003).
Oxford University Press, 1975). Wright, G.H. von: Explanation and Under-
Hume, D.: A Treatise of Human Nature, Book standing (Ithaca, NY: Cornell University
I (London, 1739); ed. L.A. Selby-Bigge Press, 1971).
(Oxford: Oxford University Press, 1888);
david h. sanford
2nd edn. rev. P.H. Nidditch (Oxford: Oxford
University Press, 1978).
Lewis, D.K.: Philosophical Papers, vol. II
(Oxford: Oxford University Press, 1986). Fictional Entities
Mackie, J.L.: The Cement of the Universe, 2nd The first question to be addressed about
edn. (Oxford: Oxford University Press, fictional entities is: are there any? The usual
1980; originally published 1974). grounds given for accepting or rejecting the
Menzies, P.: Counterfactual Theories of view that there are fictional entities come
Causation, in the Stanford Encyclopedia of from linguistic considerations. We make
Philosophy. many different sorts of claims about fictional
Psillos, S.: Causation and Explanation characters in our literary discussions. How
(Chesham, Bucks.: Acumen; Montreal: can we account for their apparent truth?
McGill-Queens University Press, 2002). Does doing so require that we allow that
Salmon, W.C.: Scientific Explanation and the there are fictional characters we can refer
Causal Structure of the World, (Princeton, to, or can we offer equally good analyses
NJ: Princeton University Press, 1984). while denying that there are any fictional
Sanford, D.H.: Causal Relata, in E. LePore entities?
and B. McLaughlin, ed., Actions and Events While some have argued that we can
(Oxford and New York: Blackwell, 1985), offer a better analysis of fictional discourse if
28293. we accept that there are fictional charac-
Sanford, D.H.: Causation and Intelligibility, ters, others have held that even if thats
Philosophy 69 (1994), 5567. true, we have metaphysical reasons to deny

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 11

fictional ent i t i es
the existence of fictional entities. Some have discourse will have import for whether or not
supposed that accepting such entities would we should accept that there are fictional
involve us in contradictions and so must be entities we sometimes refer to, and if so,
avoided at all costs, while others have held what sorts of thing they are and what is
that, even if contradiction can be averted, we literally true of them.
should refrain from positing fictional entities Given these very different types of fictional
if at all possible since they would be utterly discourse, many different approaches have
mysterious, involve us in positing unex- been developed, some of which accept and
plained differences in kinds of being, or some of which deny that there are fictional
violate reasonable calls to parsimony. entities. Many of the differences among
them may be seen as products of differences
1. Linguistic Considerations in which of the four types of discourse each
takes as its primary case and central motiv-
At least four sorts of fictional discourse may ator though of course all are ultimately
be distinguished: obliged to say how we should understand
(1) Fictionalizing discourse (discourse within each type of discourse.
works of fiction), e.g., [Holmes was] Perhaps the most popular approach to
the most perfect reasoning and observ- fictional discourse has been to deny that
ing machine that the world has seen in there are any fictional entities, and to
A Scandal in Bohemia. handle the linguistic evidence by adopting a
(2) Non-existence claims, e.g., Sherlock pretense theory. It is plausible that authors
Holmes does not exist. in writing works of fiction (and so writing sen-
(3) Internal discourse by readers about the tences of type (1)) are not making genuine
content of works of fiction. This may assertions at all, but rather simply pretend-
be either intra-fictional (reporting the ing to assert things about real people and
content of a single work of fiction, e.g., places (Searle, 1979, p. 65). (Though see
Holmes solved his first mystery in his Martinich and Stroll, 2007, ch. 2, for chal-
college years,) or cross-fictional (com- lenges to this.) Inspired by this observation
paring the contents of two works of about discourse of type (1), full-blown pre-
fiction, e.g., Anna Karenina is smarter tense theories of fictional discourse (such as
than Emma Bovary). that developed by Kendall Walton) treat all
(4) External discourse by readers and critics four forms of fictional discourse as involving
about the characters as fictional char- pretense and so as making no genuine ref-
acters, e.g., Holmes is a fictional erence to fictional entities. Discourse of type
character, Hamlet was created by (3), on these views, involves readers play-
Shakespeare, The Holmes character ing along with the pretense authorized by
was modeled on an actual medical the work of fiction, and so pretending that
doctor Doyle knew, Holmes appears what is stated in works of fiction is true.
in dozens of stories, Holmes is very Claims like Holmes solved his first mystery
famous. in his college years are authorized moves in
the game of pretense licenced by the work,
The puzzles for fictional discourse arise which is why we find them more acceptable
because many of the things we want to say than parallel claims like Holmes drove a
about fictional characters seem in conflict white Plymouth.
with each other: How, for example, could While that extension of the pretense view
Holmes solve a mystery if he doesnt exist? seems plausible enough, more difficulties
How could Hamlet be born to Gertrude if arise for handling external discourse and
he was created by Shakespeare? Any theory non-existence claims. Walton takes exter-
of fiction is obliged to say something about nal claims of type (4) to invoke new ad
how we can understand these four kinds hoc unofficial games of pretense other
of claim in ways that resolve their apparent than those authorized by the story, where,
inconsistencies. And any theory of fictional e.g., we pretend that there are two kinds

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 12

f i c t i onal entities
of people: real people and fictional char- in his college years would be false. Cross-
acters (1990, p. 423), or pretend that fictional statements can be handled simi-
authors are like gods in being capable of larly by taking them to fall in the context
creation, etc. Even apparently straightfor- of an agglomerative story operator that
ward non-existence claims (type 2) are appeals to the total content of the relevant
treated as involving pretense: first invoking stories, taken together, e.g., According to
a pretense that there is such a character to (Anna Karenina and Madame Bovary [taken
refer to (using the name Sherlock Holmes), agglomeratively] ), Anna Karenina was more
and then in the same breath betraying intelligent than Emma Bovary (Sainsbury,
that as mere pretense, with the addition of forthcoming).
doesnt exist (1990, p. 422). The full-blown But like the pretense view, the negative free
pretense approach thus seems to implausibly logic view has more difficulties accounting
take as pretenseful precisely the (type 2 and for the apparent truth of external claims
type 4) talk about fiction that is designed of type (4), since their truth cannot be
to step outside of the pretense and speak accounted for by taking them as implicitly
from the real-world perspective. It also offers reporting what is true according to the
contorted and ad hoc readings of what seem fiction. Various ad hoc ways of interpreting
to be straightforward literal claims (cf. these claims have been tried, e.g., Holmes
Thomasson, 2003). So while pretense theo- is a fictional character, may be read as
ries do well at addressing internal and reporting that, according to some fiction,
fictionalizing discourse, they are much less Holmes exists (Sainsbury, forthcoming). But
plausible adopted as across the board given the variety of external claims that
approaches but if we cant adopt them must be rewritten in different ways, these
across the board, they cant be used to remain the biggest thorn in the side of neg-
avoid positing fictional entities. ative free logic theories.
Various other approaches to fictional dis- On the other side of the debate are those
course have been proposed which dont who argue that we can only or best handle
rely on taking pretense to be ubiquitous in fictional discourse by allowing that there
fictional discourse, yet still avoid accepting are fictional entities and that at least some-
that there are fictional entities. The best times our discourse refers to them. But even
developed of these is Mark Sainsburys among those who accept that there are
(2005) negative free logic approach, which fictional entities there are widespread dis-
takes as its central motivation the truth agreements about what we should consider
of claims of type (2): non-existence claims them to be and what is literally true of them.
involving fictional names. On the negative Some realist views about fiction are
free logic view, fictional names are non- inspired by the apparent truth of internal
referring terms, and all simple sentences claims of type (3), and so take fictional enti-
using non-referring terms are false. Thus ties to be beings that (in some sense) have the
Holmes exists is false (as Holmes doesnt properties the characters of the story are
refer), and so its negation Holmes doesnt said to have, so that claims like Holmes
exist is true (Sainsbury, 2005, p. 195), solved his first mystery in his college years
leaving us with a far simpler and more is true because there is a fictional entity,
plausible account of the truth of non- Holmes, who in some sense has this property.
existence claims than pretense views pro- These views have taken many forms with
vide. Internal discourse by readers can still some taking the fictional entities to be
be held to be true even though it involves possible people, others taking them to be
non-referring names, since these claims are Meinongian non-existent objects, and others
plausibly held to be implicitly prefixed with still taking them to be pure abstract entities
a fiction operator, where According to the such as kinds.
fiction, Holmes solved his first mystery in One natural approach inspired by the
his college years may be true even if the desire to accommodate the truth of type (3)
simple claim Holmes solved his first mystery internal claims is to take fictional characters

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 13

fictional ent i t i es
to be merely possible people described by the character in the relevant work (or works)
the stories. Kripke expressed this idea when of fiction.
he wrote Holmes does not exist, but in The simple version of this approach en-
other states of affairs, he would have existed counters difficulties of the kind that led
(1963/1971, p. 65). But Kripke himself to Russells (1905/1990) criticisms of
later (1972, p. 158) rejected this answer, and Meinong. For the stories ascribe to Holmes
his rejection of it has generally been taken not only properties like being a person and
on board. His grounds for rejecting it come solving mysteries, but also properties like
from considerations about reference: the existing, in conflict with the apparent
name Sherlock Holmes is not a description truth that Holmes doesnt exist. Indeed
(which could be fulfilled by various possible Meinongian theories take non-existence
individuals); instead, if it refers at all, it picks claims of type (2) to be straightforwardly
out the individual to whom the speakers true since, although there are the relevant
use of the name bears a historical connection, fictional entities, they do not exist. So the
and it refers to that very individual across Meinongian is in danger of contradiction by
all possible worlds. So if there happened taking Holmes and the like both to exist
to be someone in the actual world who (since Meinongian objects are supposed
coincidentally was just as Holmes is said to to have all of the properties ascribed to
be in the novels, that would not show that them) and not to exist (since they are non-
he was Holmes. Similarly, if there are indi- existent objects).
viduals in other possible worlds who fulfill The central achievement of neo-
the descriptions in the books, that does not Meinongians such as Terence Parsons
show that any of them is Holmes. Moreover, (1980) and Edward Zalta (1983) has been
since there will be a great many different pos- to show how these contradictions may be
sible individuals who fulfill the descriptions, avoided. Parsons avoids them by distin-
it seems there would be no non-arbitrary guishing two kinds of properties: nuclear
way of saying which of these is Holmes properties (like being a man, being a detec-
(Kripke, 1972, pp. 1578). tive, etc.) and extra-nuclear properties (like
Given the problems with possibilist views, existing, being possible, etc.). He then holds
the most popular realist treatments of that only the nuclear properties ascribed
fictional entities have been not possibilist to the character in the story are actually
but Meinongian and abstractist views. possessed by the corresponding objects, so
Meinong himself was not interested in we do not have to conclude that Holmes
fiction per se, but rather sought to develop a exists. Nonetheless, we do need some way
general theory of the objects of speech and to mark the fact that there may be objects
cognition (1904/1960). If there is knowledge, (arguably, like Macbeths dagger) that dont
Meinong thought, there must be something exist according to the stories, as well as
known, if there is a judgment, there must objects that (like Macbeth) are said to exist.
be something judged, and so on. So, for To mark this, Parsons suggests that there are
example, if we know that the round square watered down nuclear properties corre-
is round, there must be something (the sponding to each extra-nuclear property, so
round square) of which we know that it is that Holmes does not exist (extra-nuclear)
round. Some of these objects of knowledge, but does have watered-down (nuclear)
however (like the round square) do not existence. Zalta (1983), following Ernst
exist. Meinongian views thus take seriously Mally, avoids contradiction by a different
the truth of internal (type (3)) sentences route: distinguishing two modes of predica-
like Holmes solved his first mystery in his tion: encoding and exemplifying. Fictional
college years, and take fictional entities to entities encode all of those properties they
be the non-existent objects truly described are said to have in the stories, but that
in such sentences so on these views a does not mean that they exemplify them. So
fictional entity is the object that (in some Holmes encodes existence but exemplifies
sense) has all of the properties ascribed to non-existence, and contradiction is avoided.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 14

f i c t i onal entities
A third view along similar lines takes for the truth of at least some type (4)
fictional entities to be existing abstract external claims. Whether fictional entities are
objects of some sort rather than to be taken to be unactualized possibilia,
Meinongian non-existent objects. Nicholas non-existent objects, or abstract kinds, it
Wolterstorff develops one such view, acc- seems that in any of these cases the work
ording to which fictional characters are of authors writing stories is completely
not persons of a certain kind, but person- irrelevant to whether or not there are these
kinds which do exist (1980, p. 144). On fictional entities: the relevant possibilia,
this view, authors do not refer to anyone non-existent objects, and abstract kinds
when they write fictional stories; instead, were around just as much before as after
they delineate a certain kind of person by acts of authoring, and so we cant take seri-
describing certain sets of characteristics. ously the idea that authors create fictional
The fictional character Holmes is not a characters on any of these views. The best
person, but a certain kind of person, or these views can do to account for the appar-
person-kind, that has essentially within it ent truth of claims such as Hamlet was
those properties the work attributes to the created by Shakespeare is to say that it
character, e.g., being a man, being clever, is at least true that Shakespeare described
being a detective. . . . As abstracta, of course or selected Hamlet from among all the
kinds cant literally have such properties as available possibilia, non-existent objects, or
being clever or solving mysteries but they abstract kinds and, by writing about that
can be defined by the properties essential object, made it fictional. (Below I will return
within them. So on this view, type (3) to discuss some metaphysical difficulties
claims such as Holmes solved his first mys- these views also face.)
tery in his college years are true just in All of the views canvassed thus far
case the properties expressed by the predicate whether or not they accept that there are
(solving ones first mystery during ones fictional entities face difficulties accounting
college years) are essential within the person- for the apparent truth of certain external
kind Holmes (1980, p. 159). Many (but not (type 4) sentences. This has inspired several
all see below) of the properties attributed recent theorists to begin by taking this sort
to characters in external discourse, e.g., of discourse as the focal case a view that
being famous, appearing in stories, may be requires accepting that there are fictional
properties these abstract person-kinds gen- characters and that these are created by
uinely have rather than properties essential authors in the process of writing works of
within the kind. fiction. Since they take fictional characters
But neither of these strategies helps Wolter- to be products of the creative activities of
storff cope with (type 2) non-existence claims, authors, call these artifactual views of
for existence is ascribed to Holmes in the fiction.
stories, and so is essential to that person-kind, The phenomenologist Roman Ingarden
and the abstract entity that is that person- suggested something like an artifactual view
kind also exists. Wolterstorff suggests two of fiction in his (1931) The Literary Work
alternative ways of understanding non- of Art, where he treats fictional characters
existence claims: either as saying that the (and the literary works in which they appear)
relevant person-kind has never been exem- as purely intentional objects objects owing
plified, or (acknowledging Kripkes point) their existence and essence to consciousness.
that the author was not referring to anyone Saul Kripke (apparently independently)
when he used the name in writing the story suggests that fictional entities are human
(1980, p. 161). creations in his unpublished 1973 John
Despite their differences, possibilist, neo- Locke lectures. He argues that fictional
Meinongian, and abstractist views are alike characters exist in the ordinary concrete
in taking most seriously internal (type 3) world (not another possible world), but
claims about fictional characters, and as a they do not exist automatically as pure
result they face similar difficulties accounting abstracta do. Instead, although they are

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 15

fictional ent i t i es
in some sense abstract entities, they are here: denials that Sherlock Holmes exists
contingent and exist only given concrete may be read as denials that there is any
activities of writing or telling stories. John such person (Thomasson, 1999, p. 112),
Searle (1979, pp. 712) similarly claims or any object answering the descriptions
that authors, in writing stories and pre- in the stories (van Inwagen, 2003, p. 146).
tending to refer to people, instead create Alternatively, these non-existence claims
fictional characters to which others can may be read as noting that past users of the
then refer. More recently, artifactual views name mistakenly supposed that the name-
of fiction have been defended by Schiffer use chain led back to a baptism rather
(1996) and Salmon (1998), and developed than a work of fiction (van Inwagen, 2003,
at length by Thomasson (1999, 2003). pp. 1467; cf. Thomasson, 2003). If some
(van Inwagen (1977, 1983, 2003) develops such solution to the problem of non-
a similar view according to which fictional existence claims can be shown to be plaus-
characters are theoretic entities of literary ible and non ad hoc, artifactual theories
criticism, but he is noncommittal about may offer the best overall way to handle
whether or not they are created.) fictional discourse a way which does
Artifactualist theories take external (type require positing fictional entities.
4) claims about fictional characters e.g.,
that Holmes is a fictional character created 2. Metaphysical Considerations
by Arthur Conan Doyle, who modeled
Holmes on a medical doctor to be literally None the less, many think that we have
true. On Thomassons view, fictional char- metaphysical grounds to resist positing fic-
acters are abstract artifacts created by tional entities even if we can offer a somewhat
authors activities in writing or telling stories, better account of language by accepting
and dependent for their ongoing existence that there are such entities and that we
on those stories (and copies or memories of sometimes refer to them. These arguments
them). The status of fictional characters as have run in parallel to the developing
created, dependent, abstracta, she empha- theories of what fictional entities are.
sizes, is like that of many social and cultural As we have seen, Russell originally claimed
entities such as laws of state, symphonies, that Meinongian objects were apt to infringe
and works of literature themselves: none of the law of contradiction (1905/1990, 205);
them may be identified with any concrete an objection that kept fictional entities
entity, none has a definite spatial location, largely undefended for over seventy years.
but all come into existence at a particular time While neo-Meinongians showed how to
given certain types of human activity. avoid contradiction, their views were none
Most artifactualists, like Searle, take the less widely rejected for drawing a dis-
fictional characters to be created by authors tinction between what objects exist and
pretending to refer to real people and places, what objects there are (or over which we
and so take fictionalizing (type 1) discourse may quantify) a distinction many philo-
to involve mere pretended assertions. sophers claim to find incomprehensible
Artifactualists generally do not take (type (van Inwagen, 2003, pp. 138 42).
3) internal discourse to state literal truths Abstractist and possibilist solutions, of
about properties these fictional entities course, are more acceptable to those
have; instead, they (like Sainsbury fictional already inclined to accept abstract objects,
entities) typically read these as shorthand for or possible worlds and the objects in them.
claims about what is true according to the But even if one accepts that there are pla-
fiction or (following Walton) about what is tonistic abstracta or mere possibilia (see the
accepted in games of pretense authorized extended essay on realism and antirealism
by the story. about abstract entities), other problems
The greatest difficulty for artifactual views arise in supposing that fictional characters
arises in handling (type 2) non-existence are among them. As mentioned above,
claims. Various strategies may be used fictional characters are generally thought

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 16

f i c t i onal entities
to be created, contingent features of the character somewhat differently than she did,
actual world, but neither of these is true and allows that a later author may ascribe
of either platonistically conceived abstracta new properties to a preexisting fictional
(which are eternal and necessary) or of character, provided she is familiar with that
mere possibilia (which are not created by character and intends to refer back to it
authors and are merely possible). Moreover, and ascribe it new properties (Thomasson,
some stories are (intentionally or uninten- 1999, pp. 679).
tionally) inconsistent, and so some of their None the less, artifactualist views face
characters cant be treated as possible objects other metaphysical objections. Although the
having all the properties ascribed in the artifactualist treats fictional characters as
story. created entities, they are also clearly abstract
Another metaphysical problem that arises in some sense: though not eternal and
for both possibilist and abstractist views necessary like the Platonists abstracta,
comes from the fact that they (like the they still lack a spatio-temporal location
Meinongian views before them) take the (and are not material) (Thomasson, 1999; see
descriptions in works of fiction to deter- also concrete/abstract). But the very idea
mine which object we are talking about: the that there may be created abstracta strikes
fictional entity is the possible person or some as hard to swallow. As Inwagen puts
abstract entity that has, or has essential it Can there really be abstract things that
within it, all of the properties ascribed to are made? Some might find it implausible
the character in the story. But this leads to to suppose that even God could literally
problems with the identity conditions for create an abstract object (2003, pp. 153
fictional characters (see Thomasson, 1999, 4). Thomasson (1999) addresses these wor-
ch. 5). For these views entail that no ries by noting that those who accept the
fictional character could have had any prop- existence of such ordinary social and cultural
erties other than those they are ascribed. objects as laws, marriages, symphonies, and
If the author made even a minor change in works of literature themselves are appar-
the work, so that the character is ascribed ently already committed to the existence of
so much as one different property (however created abstracta, so that no special problems
trivial), she would have written about a arise in accepting created abstracta to
different possible person, or delineated a dif- account for fictional characters. Of course
ferent person-kind. As a result, these views this companions in guilt argument leaves
must hold that sequels, parodies, and even us with two choices: allow that there are
revised editions must always include entirely abstract artifacts and accept the existence of
different characters from the original texts fictional characters, literary works, laws,
in violation of our standard assumption that etc., or deny the existence of all of these and
an author may change what she says about find some way of paraphrasing talk about
a given character, and that sequels may the latter entities as well as about fictional
describe the further adventures of one and characters. But those who would take the lat-
the same character. (Meinongian theories ter route should note that even accounting
face similar difficulties with handling iden- for fictional discourse itself is much more
tity conditions.) difficult if we cannot make reference to
Artifactualist views avoid metaphysical the stories in which they appear.
dificulties like these by taking fictional char- A final and persistent metaphysical argu-
acters (like works of literature themselves) ment against fictional entities is that, since
to be created by activities of authors and it would be much more parsimonious to
individuated primarily by their historical deny the existence of fictional characters,
origin. The artifactualist typically treats his- we should do so if at all possible. The parsi-
torical continuity not properties ascribed mony argument can be addressed in several
as the primary factor for the identity of ways. First, it is worth noting that even
a fictional character. This leaves open the Occams razor only tells us that it is vain
idea that an author might have described a to do with many what can be done with

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 17

fictional ent i t i es
fewer but if we can provide a better If we deny that there are fictional entities
account of fictional discourse by accepting (and so deny that we ever refer to them), we
fictional entities, the antirealist about fictional must explain how we can have true state-
entities is not really doing the same thing ments involving non-referring terms. If
as the realist, with fewer entities. Second, as we accept that there are fictional entities,
Thomasson (1999) notes, it is not obviously we must explain how we can refer to non-
more parsimonious to do without fictional existent objects (if we take a Meinongian
characters if we must posit abstract artifacts view), merely possible objects, or abstracta
in some other arena, e.g. to make sense of our (whether Platonist or artifactual) a task
talk about novels, symphonies, laws of state, that is especially difficult for causal theories
and the like. of reference, since none of these entities are
The most potentially powerful, though also obviously a part of the actual causal order.
the most controversial, response to parsi- Issues regarding fictional entities also
mony-based arguments comes from a certain have broader relevance for work in meta-
minimalist or pleonastic approach to their physics. If artifactualists like Thomasson
ontology proposed by Stephen Schiffer are correct, then whether or not one accepts
(1996). On Schiffers view, pretenseful uses that there are fictional characters is closely
of a fictional name in works of literature, e.g. connected to the issue of whether one
[Holmes was] the most perfect reasoning accepts other mind-dependent social and
and observing machine that the world has cultural objects such as laws and nations,
seen, automatically license us to introduce stories and symphonies. Moreover, our stance
the singular term the fictional character regarding fictional entities has central rele-
Sherlock Holmes which may then be used vance for issues of ontological commitment
in a hypostatizing way in literary discussions. and quantification: If the Meinongian is right,
Given those prior pretenseful uses, that we can quantify over entities that dont
singular term is guaranteed to refer to a exist, and existence must be distinguished
fictional character. But if all that it takes for from quantification. If the minimalist is
fictional names to be guaranteed to refer to right, then the measure of ontological com-
characters is that these names be used pre- mitment is not whether or not we quantify
tensefully in works of literature, it is not at over the relevant entities for if we accept
all clear that someone who accepts that that there are authors who use fictional
there are pretenseful uses of these names names pretensefully in writing works of
in works of literature but denies that there fiction, we are already tacitly committed to
are fictional characters is genuinely offer- fictional characters regardless of whether
ing a more parsimonious view. Instead, as they explicitly quantify over them.
Thomasson argues (2003), such a person
would be only twisting the ordinary rules of See also the az entry on fictional truth,
use for terms like fictional character by objects, and characters.
artificially inflating the conditions it takes for
there to be such characters not offering
bibl i ography
a genuinely more parsimonious ontology.
Ingarden, R.: The Literary Work of Art,
3. Broader Relevance trans. George Grabowicz (Evanston, IL:
Northwestern University Press, 1931).
The question of whether or not we should Kripke, S.: Naming and Necessity (Oxford:
accept that there are fictional entities and Blackwell, 1972).
if so, what sort of thing they are has been Kripke, S.: Semantical Considerations on
a recurrent topic throughout the history of Modal Logic, in Reference and Modality, ed.
analytic philosophy because of its broader Leonard Linsky (Oxford: Oxford University
relevance for a range of other philosophical Press, 1971; originally published 1963).
issues. First, as we have seen in section 1, Martinich, A.P. and Stroll, A.: Much Ado
it has relevance for our theory of language. about Nonexistence: Fiction and Reference

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 18

f r e e w ill
(Lanham, MD: Rowman and Littlefield, Free Will
Meinong, A.: On the Theory of Objects, in The metaphysical problem of free will has
Realism and the Background of Phenomen- arisen in history whenever humans have
ology ed. Roderick Chisholm (Atascadero, reached a certain stage of self-consciousness
CA: Ridgeview, 1960; originally published about how profoundly the world may
1904). influence their behavior in ways unknown
Parsons, T.: Non-existent Objects (New to them and beyond their control (Kane,
Haven, CT: Yale University Press, 1980). 1996, pp. 956). Various authors describe
Russell, B.: On Denoting, in The Philosophy this stage of self-consciousness as the recog-
of Language, 2nd edn. Ed. A.P. Martinich nition of a conflict between two perspect-
(New York: Oxford University Press, ives we may have on ourselves and our
1990; originally published 1905). place in the universe (e.g., Nagel, 1986).
Sainsbury, M.: Reference Without Referents From a personal or practical standpoint,
(Oxford: Clarendon Press, 2005). we believe we have free will when we view
Sainsbury, M.: Serious Uses of Fictional ourselves as agents capable of influencing
Names (forthcoming). the world in various ways through our
Salmon, N.: Nonexistence, Nos 32:3 choices or decisions. When faced with choices
(1998), 277319. or decisions, open alternatives seem to
Schiffer, S.: Language-Created Language- lie before us a garden of forking paths
Independent Entities, Philosophical Topics into the future, to use a popular image.
24:1 (1996), 14967. We reason and deliberate among these
Searle, J.: Expression and Meaning: Studies in alternatives and choose. We feel (1) it is
the Theory of Speech Acts (Cambridge: up to us what we choose, and hence
Cambridge University Press, 1979). how we act; and this means we could have
Thomasson, A. L.: Fiction and Metaphysics chosen to act otherwise. As Aristotle
(Cambridge: Cambridge University Press, said, when acting is up to us, so is not
1999). acting. This up-to-us-ness also suggests
Thomasson, A. L.: Speaking of Fictional that (2) the ultimate sources of our choices,
Characters, Dialectica 57:2 (2003), 207 and hence of our actions, lie in us and
26. not outside us in factors beyond our
van Inwagen, P.: Creatures of Fiction, control.
American Philosophical Quarterly 14:4 Because of these features, free will is often
(1977), 299308. associated with other valued notions such
van Inwagen, P.: Existence, Ontological as moral responsibility, autonomy, genuine
Commitment, and Fictional Entities, in The creativity, self-control, personal worth or
Oxford Handbook of Metaphysics, ed. Michael dignity, and genuine desert for deeds or
Loux and Dean Zimmerman (Oxford: accomplishments (Kane, 1996, ch. 6). These
Oxford University Press, 2003). two features of free will also lie behind
van Inwagen, P.: Fiction and Metaphysics, various reactive attitudes that we naturally
Philosophy and Literature 7 (1983), 67 take toward the behavior of ourselves
77. and others from a personal standpoint (P.
Walton, K.: Mimesis as Make-Believe (Cam- Strawson, 1963). Gratitude, resentment,
bridge, MA: Harvard University Press, admiration, indignation, and other such
1990). reactive attitudes seem to depend upon
Wolterstorff, N.: Works and Worlds of Art the assumption that the acts for which we
(Oxford: Clarendon Press, 1980). feel grateful, resentful, or admiring had
Zalta, E.: Abstract Objects (Dordrecht: Reidel, their origins in the persons to whom these
1983). attitudes are directed. We feel that it was
up to them whether they performed those
amie l. thomasson acts or not.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 19

f ree will
d e t e r m i nism and co mp atib ilism not exist in a completely determined world
are incompatibilists about free will. The
But something happens to this familiar pic- opposite view is taken by compatibilists, who
ture of ourselves and other persons when hold that, despite appearances to the con-
we view ourselves from various impersonal, trary, determinism poses no threat to free will,
objective or theoretical perspectives (Nagel, or at least to any free will worth wanting
1986, p. 110). From such perspectives it (Dennett, 1984).
may appear that our choices or decisions Compatibilists characteristically argue that
are not really up to us, but are deter- all the freedoms we recognize and desire in
mined or necessitated by factors unknown ordinary life e.g., freedoms from coercion
to us and beyond our control. The advent of or compulsion, from physical restraint, from
doctrines of determinism in the history of addictions and political oppression are
philosophy is an indication that this worry really compatible with determinism. Even
has arisen. Doctrines of determinism have if the world should be deterministic, they
taken many forms. People have wondered argue, there would still be an important
whether their actions might be determined difference between persons who are free
by Fate or by God, by the laws of physics from constraints on their freedom of action
or the laws of logic, by heredity or environ- (such as coercion, compulsion, addiction
ment, by unconscious motives or hidden and oppression) and persons who are not free
controllers, psychological or social condi- from such constraints; and we would prefer
tioning, and so on. There is a core idea run- to be free from such constraints rather than
ning through all these historical doctrines not, even in a determined world. Com-
of determinism that shows why they are a patibilism was espoused by some ancient
threat to free will. All doctrines of determin- philosophers, such as the Stoics, and also
ism whether logical, theological, physical, by Aristotle, according to some scholars.
biological, psychological or social imply But it became especially influential in the
that at any time, given the past and the modern era, defended in one form or
laws of nature (see law of nature) and of another by philosophers such as Hobbes,
logic, there is only one possible future. Locke, Hume, and Mill, who saw com-
Whatever happens is therefore inevitable patibilism as a way of reconciling ordinary
or necessary (it cannot but occur), given experience of being free with modern
the past and the laws. science. Compatibilism remains popular
Doctrines of determinism thus seem to among philosophers and scientists today
threaten both features of free will men- for similar reasons. By contrast, incompati-
tioned earlier. If determinism is true, it bilists of the modern era, such as James,
seems that it would not be (1) up to regard compatibilism as a quagmire of
agents what they chose from an array of evasion or a wretched subterfuge, as
alternative possibilities, since only one Kant called the compatibilism of Hobbes
alternative future would be possible, given and Hume.
the past and laws. It also seems that, if Compatibilists also characteristically warn
determinism were true, the (2) sources or against confusing determinism with fatal-
origins of choices and actions would not ism, the view that whatever is going to
be in the agents themselves, but in some- happen, is going to happen, no matter what
thing outside their control that determined we do. Compatibilists, such as Mill, argue
their choices and actions (such as the that what we decide and what we do will
decrees of fate, the foreordaining acts of make a difference to how things turn out,
God, heredity and upbringing or social even if determinism should be true. And
conditioning). since we do not know the future, we have to
Those who believe, for these or other rea- deliberate and try to decide upon the best
sons, that free will and determinism are not course of action, whether determinism is
compatible and hence that free will could true or not.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 20

f r e e w ill
t h e c onseq uence ar gument Mill) who defend hypothetical or conditional
a n d i n c o mp atib ilism analyses of can and can do otherwise.
According to such analyses, to say we can
The Compatibility Question (Is free will
do otherwise means that we would do
compatible or incompatible with determin-
otherwise, if we chose or wanted to do other-
ism?) has thus been central to modern
wise. If such hypothetical analyses are
debates about free will. And the popularity
correct, the conclusion of the Consequence
of compatibilism in the modern era has
Argument (if determinism is true, no one can
placed the burden of proof on incompatibil-
do otherwise would fail. For, being able to
ists to show why free will must be incom-
do otherwise would merely entail that one
patible with determinism. Incompatibilists
would have done otherwise, if (contrary to
have tried to meet this challenge in various
fact) one had chosen or wanted to do other-
ways. The most widely discussed of their
wise, or if the past had been different in
arguments for the incompatibility of free
some way; and such a claim would be con-
will and determinism in modern philosophy
sistent with saying that ones present action
is called the Consequence Argument. It is
was determined by the actual past and
stated informally by one of its defenders
laws. Much debate about the compatibility
(van Inwagen, 1983, p. 16) as follows:
of free will and determinism has thus con-
If determinism is true, then our acts are the cerned such hypothetical analyses of can
consequences of the laws of nature and and could have done otherwise favored
events in the remote past. But it is not up by classical compatibilists. Incompatibilists
to us what went on before we were born; reject hypothetical analyses; and powerful
and neither is it up to us what the laws of objections have been made against them by
nature are. Therefore the consequences J.L. Austin, R.M. Chisholm and K. Lehrer,
of these things (including our own acts) are among others. Yet hypothetical analyses
not up to us. continue to be defended by many compat-
ibilists, e.g., Davidson and Lewis. (For an
To say it is not up to us what went on overview of the debates, see Berofsky, in Kane,
before we were born, or what the laws of 2002.)
nature are, is to say that there is nothing
we can now do to change the past or alter
al t ernative possibi l i t i es and
the laws of nature (it is beyond our control).
moral responsibi l i t y
If such things are beyond our control,
and our present actions are necessary con- A more radical compatibilist challenge to
sequences of the past and laws of nature the Consequence Argument consists in deny-
(as determinism entails), then altering the ing altogether the assumption that free
fact that our present actions occur would will requires the power to do otherwise, or
appear to be beyond our control as well. In alternative possibilities. Call this assumption
short, if determinism is true, no one can do AP (for alternative possibilities). If AP is
otherwise than he or she actually does; and false if free will does not in fact require
if free will requires the power to do otherwise, the power to do otherwise then the Con-
or alternative possibilities, then no one sequence Argument, it is argued, would
would have free will. also fail to show that free will and deter-
This argument has generated a large crit- minism are not compatible. But on what
ical literature. Each premise and step has grounds could one deny that free will
been questioned. (For useful summaries of requires the power to do otherwise? The
the issues, see van Inwagen, 1983; Fischer, answer lies in the connection between free
1994; Ekstrom, 2000; Kapitan, in Kane, will and moral responsibility. Freedom of will
2002). Compatibilists have usually chal- is not just any kind of freedom of action or
lenged the argument in either of two ways. freedom to do what you want. Freedom of will
The first challenge comes from classical has a special relationship to responsibility
compatibilists (such as Hobbes, Hume, and or accountability for ones actions. Indeed,

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 21

f ree will
many philosophers actually define free will though he could not have done otherwise:
as that kind of freedom (whatever it may and PAP is false.
be) that is necessary to confer true moral As with the Consequence Argument, an
responsibility (and hence genuine praise- enormous literature has developed around
worthiness and blameworthiness) on agents. these Frankfurt-style examples. (Overviews
The connection between free will and of the literature can be found in Fischer,
moral responsibility is important because a 1994 and Widerker and McKenna, 2003.)
number of new compatibilists, including Of many objections that have been made
Frankfurt (1969), Dennett (1983), Fischer against the use of such examples to under-
(1994), and Wallace (1994), have denied mine PAP, the most discussed objection is
that moral responsibility requires the power one originally made by Kane (1985) and
to do otherwise, or alternative possibilities. developed independently by Widerker (1995)
They reject a principle that Frankfurt has and Ginet, among others. The objection
called the Principle of Alternative Possibilities insists that if some morally responsible (free
(PAP): Persons are morally responsible for will) choices are undetermined up to the
what they have done, only if they could moment they occur, as incompatibilists
have done otherwise. If free will is the kind require, then a Frankfurt controller like
of freedom required for moral responsibility Black could not know in advance which
and PAP is false, then AP would be false as choice the agent Jones was going to make
well: Free will (in the sense required for until the choice was actually made. In that
moral responsibility) would also not require case, if the controller did not intervene, the
the power to do otherwise or alternative agent would have alternative possibilities;
possibilities. and if the controller did intervene, he would
Two kinds of examples have been offered have to do so in advance to make the agent
by new compatibilists to show the falsity choose as he wished. But in that case,
of PAP. The most widely discussed of these the controller would be responsible for the
two kinds of examples are called Frankfurt- choice, not the agent. To meet this objection,
style examples, after Harry Frankfurt, who a host of new, more sophisticated, Frankfurt-
introduced the first such example in 1969. style examples have been developed in the
Frankfurt posited a controller named Black, past decade by David Hunt, Eleonore Stump,
whom we might suppose is a neuro- Alfred Mele and David Robb, Derk Pereboom,
surgeon with direct control over the brain of and others. The jury is still out on the
an agent Jones. Jones faces a choice between efficacy of these new Frankfurt-style ex-
doing A (say, voting for a Democrat) and B amples. (For a discussion of this literature,
(voting for a Republican). Black wants see Widerker and McKenna, 2003.)
Jones to do A, but he does not want to inter-
vene unless he has to. So if Black sees that
hi erarc hical theories and
Jones is going to choose A on his own,
other new compat i bilist vi ews
Black will not intervene. Only if Black sees
that Jones is going to choose B will he inter- New compatibilists, such as Frankfurt, have
vene in Joness brain, making Jones choose also put forward novel (compatibilist) ac-
A. Frankfurt asks us to consider the case counts of free will, according to which free
where Jones chooses A on his own and will does not require the power to do other-
Black does not intervene. In such a case, wise. In a seminal essay, Frankfurt (1971)
Frankfurt argues, Jones could well have argues that persons, unlike other animals,
been morally responsible for his choice, have the capacity for reflective self-evalua-
since Jones acted on his own and Black tion that is manifested in the formation of sec-
did not intervene. Yet Jones could not have ond-order desires (p. 7) desires to have or
done otherwise, since, if he had given any not to have various first-order desires. Free
indication of choosing otherwise, Black would will and responsibility require that we
not have let him. So it seems that Jones can assess our first-order desires or motives and
be morally responsible for his choice even form second-order volitions about which

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 22

f r e e w ill
of our first-order desires should move us to of P. Strawson (1962) and Wallace (1994),
action. Our wills the first-order desires emphasize the role of reactive attitudes,
that move us to action are free, according such as gratitude, resentment and indigna-
to Frankfurt, when they are in conformity tion, in our understanding of freedom and
with our second-order volitions, so that we responsibility. (For critical surveys of many
have the will (first-order desires) we want of these new compatibilist theories, see the
(second-order desires) to have and in that essays by Haji and Russell, in Kane, 2002).
sense we identify with our will. Another significant new compatibilist
Such a theory of free will is called hier- approach to free will is semi-compatibilism,
archical for obvious reasons. Classical whose chief defender is Fischer (1994; see also
compatibilism is deficient, according to Fischer and Ravizza, 1998). Fischer is con-
hierarchical theorists such as Frankfurt, vinced by Frankfurt-style examples and
because it gives us only a theory of freedom other considerations that moral responsibility
of action (being able to do what we will) does not require alternative possibilities. But
without a theory of freedom of will in terms he also argues that freedom does require
of the conformity of first-order motives forking paths into the future, and hence
to higher-order motives (being able, so to alternative possibilities; and he is convinced
speak, to will what we will). Hierarchical by the Consequence Argument that deter-
theories remain compatibilist, however, since minism rules out alternative possibilities.
they define free will in terms of a conformity The result of these competing considerations
(or mesh) between desires at different is semi-compatibilism: moral responsi-
levels without requiring that desires at any bility is compatible with determinism, but
level be undetermined. freedom (in the sense that requires alterna-
Other new compatibilist accounts of free tive possibilities) is not compatible with
will, such as those of Watson (1975) and determinism.
Wolf (1990), are also mesh theories, but
they reject Frankfurts hierarchical view.
hard determini sm and hard
For Watson, the relevant mesh required
inc ompatibi l i sm
for free agency is not between higher and
lower-order desires, but between an agents Incompatibilists have also put forth new
valuational system (beliefs about what accounts of free will in modern philo-
is good or ought to be done), which has sophy and new defenses of its incom-
its source in the agents reason, and the patibility with determinism. Incompatibilism,
motivational system (desires and other however, may take two opposing forms:
motives), which has its source in appetite. Incompatibilists who affirm the existence of
Watson thus revives the ancient Platonic free will and hence deny the truth of deter-
opposition between reason and desire minism are called libertarians in modern
arguing that freedom consists in a certain free will debates. By contrast, incompati-
conformity of desire to reason. Wolfs rea- bilists who affirm determinism, and thus
son view takes this approach in another deny the existence of free will, have tradi-
direction that also has ancient roots. She tionally been called hard determinists. Hard
argues that freedom consists in being able determinism will be considered here first
to do the right thing for the right reasons, and then libertarianism.
which requires in turn the ability to appreci- Classical hard determinism (as held by
ate the True and the Good. Wolfs theory dHolbach, for example) consists of three
thus has a stronger normative component theses: (i) free will (in the strong sense
than other compatibilist theories. required for ultimate responsibility and
Other new compatibilist approaches to desert) is not compatible with determinism;
freedom with a normative component in- (ii) there is no free will in this strong sense
clude those of Michael Slote, Paul Benson, because (iii) all events are determined by
and Philip Pettit and Michael Smith. Still natural causes (i.e., determinism is true).
other new compatibilist theories, e.g., those Modern skeptics about free will who are

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 23

f ree will
sympathetic to hard determinism, such as One of these modern skeptics about free
Honderich (1988), Pereboom (2001), and G. will, Pereboom calls this successor view
Strawson (1986) tend to accept theses (i) and to hard determinism, hard incompatibilism,
(ii), but remain non-committal about (iii) which is a useful designation for those who
whether universal determinism is true. hold theses (i) and (ii), but remain non-
These modern skeptics about free will committal about thesis (iii).
are aware that, with the advent of quantum Why do hard incompatibilists continue to
physics in the twentieth century, it is far believe that incompatibilist or libertarian
less clear that the physical world is the free will does not exist, if they are unsure of
deterministic system imagined by classical the truth of universal determinism? There are
Newtonian physics. In the eighteenth cen- several reasons. First, while hard incompat-
tury, with Newtonian physics in mind, ibilists remain non-committal about inde-
LaPlace famously imagined that a superin- terminism in physics generally, they tend to
telligence, knowing all the forces of nature believe that human behavior is regular and
and the exact positions and momenta of determined for the most part. If indeter-
particles at any one time, could predict minism does exist in the microphysical
with certainty every future event in the world, in the behavior of elementary particles,
minutest detail. its macroscopic effects on human behavior,
Today it is customary to distinguish pre- they argue, would be negligible and of no
dictability or this sort from determinism. For significance for free will. Second, hard incom-
it is known that even in some classical patibilists are convinced by developments
physical systems (such as those that exhibit in sciences other than physics in biology
chaotic behavior), future behavior may (greater knowledge of genetic influences),
not be predictable, even though such sys- neuroscience, psychology, psychiatry, social
tems continue to be deterministic. Modern and behavior sciences that more of our
quantum physics complicates this classical behavior than previously believed is deter-
picture even further (at least on standard mined by causes unknown to us. For example,
interpretations of it), by insisting that no controversial neuroscientific experiments of
superintelligence could know the exact Libet (2002) and others have led psycho-
positions and momenta of all particles at logists, such as Wegner (2002), to argue
any moment because the particles do not that our familiar experiences of conscious
have both exact positions and momenta at willing may be an illusion.
the same time (the Heisenberg uncertainty New research in the neurosciences in
principle); and hence their future behavior general has had an increasing impact on
is not predictable or determined. Yet free will debates. (For discussions of this
issues of determinism and indeterminism in impact, see Walter, 2001; Dennett, 2003;
physics remain unsettled because there is Wegner, 2002; and the essays in Libet et al.,
continuing controversy about the interpre- 1999. For discussions of the implications of
tation of quantum physics and about its quantum physics and other developments
metaphysical implications. in the physical and behavioral sciences
As a consequence, modern skeptics about for free will, see the essays of Hodgson and
free will usually remain non-committal about Bishop, in Kane, 2002; Earman, 1986; and
the truth of universal determinism (thesis iii), the essays in Atmanspacher and Bishop,
preferring to leave that debate to the phy- 2004.)
sicists. But these modern skeptics about There is a third reason why hard incom-
free will continue to hold the first two the- patibilists are skeptical of an indeterminist
ses of classical hard determinism, namely or libertarian free will. They insist that if
that (i) free will in the true responsibility- quantum indeterminism at the micro-level did
entailing sense is incompatible with sometimes have macroscopic effects on the
determinism and that (ii) there is, and can human brain or behavior, such indetermin-
be, no incompatibilist (or libertarian) free ism would be of no help to believers in liber-
will of this true responsibility-entailing kind. tarian free will, since such indeterminism

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 24

f r e e w ill
would not enhance, but would only dimin- free will and political libertarianism share
ish, freedom and responsibility. Suppose a a name from the Latin liber, meaning free
choice was the result of a quantum jump and they share an interest in freedom. But
or other undetermined event in a persons libertarians about free will are not neces-
brain, they argue. Such undetermined effects sarily committed to political libertarianism
in the brain or body would be unpredictable and may (and many do) hold differing
and occur by chance, like the sudden occur- political views.
rence of a thought or the spasmodic jerking To defend libertarianism about free will,
of an arm quite the opposite of what we take one has to do more than merely argue for
free and responsible actions to be. Undeter- the incompatibility of free will and deter-
mined events in the brain and body would minism. One must also show that we can
therefore undermine our freedom rather actually have a free will that is incompat-
than enhance it. ible with determinism. Many philosophers,
Hard incompatibilists have also been including both hard determinists and com-
concerned with the impact their denial of free patibilists, have argued that an incompat-
will would have for morality and the mean- ibilist free will of the kind that libertarians
ing of life. Some of them, such as Honderich affirm is not even possible or intelligible and
and Pereboom, argue that we can still live that it has no place in the modern scientific
meaningful lives without the illusion of picture of the world. Critics of libertarianism
libertarian free will, though some import- note that libertarians have often invoked
ant life-hopes and attitudes would have obscure and mysterious forms of agency or
to change. For example, we could no longer causation to defend their view.
believe that criminal punishment was ultim- In order to explain how free actions can
ately deserved. Yet, we could still incarcer- escape the clutches of physical causes and
ate criminals to deter them and others from laws of nature, libertarians have sometimes
committing future crimes or to reform them. posited a disembodied mind or soul in the
But other philosophers, such as Smilansky manner of Descartes, which is outside
(2000), who also believe libertarian free of the physical realm and not governed
will is impossible, argue that the effects on by physical laws, yet capable of influencing
society and moral life would be dire if most physical events. Other libertarians, such as
people became convinced that we do not Kant, have appealed to a noumenal self,
have an incompatibilist or libertarian free outside space and time, not subject to sci-
will. Smilansky provocatively suggests that entific causes and explanations. Still other
while we do not have such an incompatibilist libertarians, such as Reid, appeal to a spe-
free will, we must continue to foster the illu- cial kind of agent- or immanent causation
sion which most ordinary persons share that is irreducible to ordinary forms of
that we do have such a free will for the sake causation (see the extended essay) in terms
of morality and social order. of events common to the sciences. Appeals
such as these, and other appeals by liber-
tarians to uncaused causes or unmoved
l i b e r t ar ian views o f fr ee will
movers, have invited charges of obscurity
Libertarianism is the name usually given to or mystery against libertarian views of free
those who hold that (i) free will and deter- will by their opponents. Even some of the
minism are incompatible, (ii) free will (in greatest defenders of libertarianism, such
this incompatibilist sense) exists and so (iii) as Kant, have argued that we need to
determinism is false. Libertarianism about free believe in libertarian freedom to make sense
will in this sense should not be confused of morality and true responsibility. But we
with the political doctrine of libertarianism, cannot completely understand such a free-
the view that governments should be limited dom in theoretical and scientific terms.
to protecting the liberties of individuals so The problem that usually provokes skep-
long as the individuals do not interfere with ticism about libertarian free will has to do
the liberties of others. Libertarianism about with an ancient dilemma: If free will is not

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 25

f ree will
compatible with determinism, it does not be uncaused by events, they are not uncaused
seem to be compatible with indeterminism by anything. Free actions are caused by the
either. Events that are undetermined, such agents themselves in a sui generis way that is
as quantum jumps in atoms, happen merely not reducible to causation by states or
by chance. So if free actions must be un- events of any kinds involving the agent,
determined, as libertarians claim, it seems physical or mental. Other agent-causalists,
that they too would happen by chance. But such as Clarke (2003), allow that reasons
how can chance events be free and respons- for action (such as desires and beliefs) can
ible actions? To defend their view, liber- causally influence choices and actions. But
tarians must not only show that free will is he also postulates a special non-event cau-
incompatible with determinism, they must sation by agents to explain what tips the
also show how free will can be compatible with balance between reasons for one choice or
indeterminism. action rather than another. Critics of agent-
Libertarian accounts of free will have causal theories, such as Watson, argue that
taken a number of different forms in the appeals to a special kind of non-event cau-
attempt to address this problem. It is now sation by substances are no less mysterious
customary to distinguish three main types than Kantian appeals to noumenal selves or
of libertarian theories: (1) non-causalist Cartesian appeals to disembodied minds to
(or simple indeterminist) views; (2) causal explain undetermined free choices. Agent-
indeterminist or event-causal views; and causalists have attempted to rebut such
(3) agent-causal views. charges in various ways. (For an overview
Non-causalist or simple indeterminist lib- of the debates see the essays of OConnor
ertarian views rely on a distinction between et al., in Kane, 2002; Clarke, 2003, ch. 10).
two ways of explaining events, explanations Causal indeterminist or event-causal views
in terms of reasons and purposes (desires, (the third kind of libertarian theory) are of
beliefs and intentions) and explanations more recent origin. Such views were first
in terms of causes. Non-causalists, such as suggested, though not developed in detail, in
Ginet (1990) and McCann (1998), argue the 1970s by David Wiggins and Robert
that free actions can be explained in terms Nozick as an alternative to non-causalist and
of the agents reasons for action (desires, agent-causal views. The first fully developed
beliefs, etc.), without being caused or deter- causal indeterminist view was that of Kane
mined, because explanations in terms of (1985, 1996). Causal indeterminists attempt
reasons are not causal explanations. Non- to explain undetermined choices without
causalist views raise important questions appealing to claims that reasons cannot
of action theory about the nature of be causes of actions and without appealing
action, about the distinction between actions to extra factors such as noumenal
and other events (see event theory), about selves, disembodied minds, or non-event
whether reasons for action can be causes of agent causes to explain free actions. Causal
action, and so on. Critics of non-causalist indeterminists allow that free actions
or simple indeterminist views note that, may be caused by reasons, intentions and
for non-causalists, free actions are literally other states and processes of the agent,
uncaused events, and the critics raise questions but not deterministically caused. Undeter-
about how events that are uncaused can be mined, they point out, need not mean
under the control of agents. uncaused: Reasons can cause actions non-
Agent-causalist libertarians follow Reid in deterministically or probabilistically, so that,
postulating a special kind of causation by an while libertarian freedom must be indeter-
agent or substance that does not consist in minist, it need not be contra-causal.
causation by events or states of affairs, as is Causal indeterminist or event-causal
common for forms of causation studied by the libertarian views come in two varieties.
sciences. Agent-causalists, such as Chisholm Deliberative views (first suggested by
and OConnor (2000), insist against simple Dennett and later developed by Mele, 2006
indeterminists that, while free actions may and Ekstrom, 2000) place the indeterminism

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 26

f r e e w ill
early in the deliberative processes of agents, Consequence Argument, have appealed to the
in the undetermined coming to mind of requirement of alternative possibilities or
thoughts, memories and other considera- AP, or branching paths into the future. But
tions that influence subsequent choice. By a number of modern incompatibilists about
contrast, so called centered causal inde- free will, have argued that another require-
terminist views (such as that of Kane) insist ment of free will, a requirement of ultimate
that indeterminism can in some cases persist responsibility or UR, is as important as AP,
right up to the moment of choice itself. perhaps even more important, to debates
An important criticism of causal indeter- about the incompatibility of free will and
minist views of the centered variety is the determinism. The basic idea of UR is this:
so-called luck objection, an objection that To be ultimately responsible for an action,
has been used against other libertarian views an agent must be responsible for anything
as well, agent-causal and non-causalist. (See that is a sufficient cause or motive for the
Mele, 2006 and Haji, 2003 for extended actions occurring. If, for example, a choice
discussions of this objection.) Mele puts the issues from and can be sufficiently explained
luck objection this way: Suppose John fails by an agents character and motives (together
to resist the temptation to tell a lie. If his with background conditions), then to be
choice to lie is a free act in the libertarian ultimately responsible for the choice, the
sense then it must have been undetermined agent must be at least in part responsible by
up to the moment it was made. This means virtue of choices or actions voluntarily per-
John could have done otherwise (could formed in the past for having the character
have chosen not to lie), given exactly the and motives he or she now has. Compare
same past up to that moment (since inde- Aristotles claim that if a man is responsible
terminism implies same past, different pos- for the wicked acts that flow from his char-
sible outcomes). Thus we can imagine a acter, he must be responsible for forming
counterpart of John, John*, in an alternative the wicked character from which these
possible world with exactly the same past who acts flowed.
did resist temptation and chose not to lie. Mele The importance of this UR condition was
argues that, since there is nothing about first noted in recent free will debates inde-
the powers, capacities, states of mind, moral pendently by Kane (1985) and G. Strawson
character and so on that is different in (1986). Both agreed that UR could not be
the pasts of John and John* right up to the satisfied in a deterministic world, so it pro-
moment they chose that could explain the vided a further argument for the incompat-
difference in their choices, then the differ- ibility of free will and determinism that did
ence in their choices would have been not appeal to AP. But Kane and Strawson
merely a matter of luck. John* got lucky in disagreed about whether UR was an intelli-
attempting to resist temptation, while John gible or satisfiable condition. Kane, a liber-
did not; and it would not be fair to reward tarian, attempted to show that UR could
one and punish the other for what was be satisfied. While Strawson, a hard incom-
merely a matter of luck. A considerable lit- patibilist, argued that UR was an unsat-
erature has been generated by this luck isfiable condition since it would either
objection. Causal indeterminists and other require an impossible infinite regress of past
libertarians have tried to answer it in vari- voluntary actions by which we formed our
ous ways, but many believe it cannot be later characters or it would require some
answered. initial character-forming acts that were
not determined by prior character. Such
initial acts would either be determined by
u l t i m a te r esp o nsib ility
something external to the agent or would
One final topic concerning incompatibilist occur merely by chance. This regress argu-
and libertarian views of free will deserves ment, which Strawson called the Basic
mention. Most arguments for the incompat- Argument, poses a significant challenge
ibility of free will and determinism, like the to libertarian accounts of free will; and

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 27

f ree will
attempts to answer it by libertarians have also Ginet, C.: On Action (Cambridge: Cambridge
been an important part of current free will University Press, 1990).
debates. Haji, I.: Deontic Morality and Control
The requirement of ultimate responsibility (Cambridge: Cambridge University Press,
or UR has played another role in free will 2002).
debates. Some incompatibilists, now called Honderich, T.: A Theory of Determinism,
source incompatibilists (including some 2 vols. (Oxford: Clarendon Press, 1988).
hard incompatibilists, such as Pereboom, Kane, R., ed. Free Will and Values (Albany,
and some libertarians, such as Eleonore NY: State University of New York Press,
Stump and Linda Zagzebski) argue that UR 1985).
is the primary condition required for an Kane, R., ed.: The Oxford Handbook of Free Will
incompatibilist free will and that alternat- (Oxford and New York: Oxford University
ive possibilities (AP) are not required for Press, 2002).
free will at all. Source incompatibilists of Kane, R., ed.: The Significance of Free Will
this sort are now often distinguished from (Oxford: Oxford University Press, 1996).
leeway incompatibilists, who hold the Libet, B., Freeman, A., and Sutherland, K.,
more traditional view that AP is the prim- ed.: The Volitional Brain: Towards a Neuro-
ary reason why free will and determinism are science of Free Will (Thorverten, Devon:
incompatible. Disputes between these two Imprint Academic, 1999).
views about the comparative importance McCann, H.: The Works of Agency: On
of UR and AP for free will have thus also Human Action, Will and Freedom. (Ithaca,
become a significant part of modern debates NY: Cornell University Press, 1998).
about free will. Mele, A.: Free Will and Luck (Oxford: Oxford
University Press, 2006).
Nagel, T.: The View from Nowhere (New York:
b i b l i og rap hy Oxford University Press, 1986).
OConnor, T.: Persons and Causes: The Meta-
Atmanspacher, H. and Bishop, R., ed.: Be- physics of Free Will (New York: Oxford
tween Chance and Choice: Inter-disciplinary University Press, 2000).
Perspectives on Determinism (Thorverton, Pereboom, D.: Living Without Free Will
Devon: Imprint Academic, 2002). (Cambridge: Cambridge University Press,
Clarke, R.: Libertarian Accounts of Free 2001).
Will (Oxford: Oxford University Press, Smilansky, S.: Free Will and Illusion (Oxford:
2003). Clarendon Press, 2000).
Dennett, D.: Elbow Room (Cambridge, MA: Strawson, G.: Freedom and Belief (Oxford:
MIT Press, 1984). Oxford University Press, 1986).
Dennett, D.: Freedom Evolves (New York: Strawson, P. F.: Freedom and Resent-
Vintage Books, 2003). ment, Proceedings of the British Academy
Earman, J.: A Primer on Determinism 48 (1962), 125.
(Dordrecht: Reidel, 1986). van Inwagen, P.: An Essay on Free Will
Ekstrom, L. W.: Free Will: A Philosophical (Oxford: Clarendon Press, 1983).
Study (Boulder, CO: Westview, 2000). Wallace, R. J.: Responsibility and the Moral
Fischer, J. M.: The Metaphysics of Free Will: Sentiments (Cambridge, MA: Harvard
A Study of Control (Oxford: Blackwell, University Press, 1994).
1994). Walter, H.: Neurophilosophy of Free Will
Fischer, J. M. and Ravizza, M.: Responsibility (Cambridge MA: MIT Press, 2001).
and Control: A Theory of Moral Responsib- Watson, G. Free Agency, Journal of Phil-
ility (Cambridge: Cambridge University osophy 72 (1975), 20520.
Press, 1998). Widerker, D.: Libertarianism and Frank-
Frankfurt, H.: Freedom of the Will and the furts Attack on the Principle of Alternative
Concept of a Person, Journal of Philo- Possibilities, The Philosophical Review 104
sophy 68 (1971): 520. (1975), 247 61.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 28

i n d i v i duatio n
Widerker, D. and McKenna, M., ed.: Moral must, of course, be distinct from identity
Responsibility and Alternative Possibilities itself in order for the criterion in question
(Aldershot, Hants: Ashgate Publishers, to be informative and non-circular. (An
2003). equivalence relation is one that is reflexive,
Wegner, D.: The Illusion of Conscious Will symmetrical, and transitive.) An adjectival
(Cambridge MA: MIT Press, 2002). term lacks a criterion of identity because
Wolf, S.: Freedom Within Reason (Oxford: there is no single condition that things
Oxford University Press, 1990). to which it applies must satisfy in order to
be identical (other than, trivially, identity
robert kane
itself). Thus, there is no such condition that
any red thing must satisfy in order to be
identical with another red thing: whether or
not one red thing is identical with another
For reasons which will become clear, it depends at least in part on what sort or kind
is appropriate to begin a general account of of red things they are and then the relevant
individuation with some discussion of sortal criterion of identity will be that supplied
terms and concepts. The expression sortal by the relevant sortal term, be it, say, cat,
is a coinage of John Lockes (Locke, 1975, apple, or star.
III, III, p. 15). He held a sortal name to Sortal concepts are what sortal terms
signify the complex general idea of a certain express or convey although, of course, we
sort of things (Locke, 1975, III, VI, p. 1). shouldnt assume that for every sortal con-
Prime examples of sortal terms, sometimes cept there exists a sortal term (much less
also called substantival general terms, are a sortal term in every natural language)
cat, apple, mountain, and star. Sortal which expresses or conveys it. Another name
terms may be contrasted with adjectival terms, for sortal concepts is individuative concepts,
such as red, round, and heavy. It for reasons that will become plain when I
is commonly said that the key distinction come, in a moment, to introduce the notion
between sortal and adjectival terms is that or, rather, the notions of individuation.
while both possess criteria of application, But what, quite generally, are concepts sup-
only the former possess criteria of IDENTITY posed to be? Of course, this in itself is
(Dummett, 1981, pp. 5478). A criterion a highly contentious question. Here I shall
of application for a general term tells us simply state one widely held view of the
what it applies to. In other words, it determines matter, which is that a concept is a way of
the extension of the term: the SET of entities thinking of some thing or things (Lowe, 2006,
all and only the members of which are pp. 856). Since thought is a mental process,
correctly described by the term, such as the this means, in effect, that concepts are mental
set of cats in the case of the sortal term cat properties of a certain kind. For properties
and the set of red things in the case of the or qualities, quite generally, are ways of
adjectival term red. A criterion of identity being ways things are (Lowe, 2006, pp. 14,
for a sortal term tells us what determines 901). For example, roundness is a way of
whether or not one thing that the term being shaped and redness is a way of being
applies to is the same as, or numerically colored. By the same token, concepts, being
identical with, another thing that the term ways of thinking of things, are ways of
applies to: whether or not, for instance, the being and hence properties and, more
cat that is now sitting on the mat is the specifically, mental properties, since thought
same cat as the cat that was formerly sleep- is a mental process. So much for the onto-
ing on the sofa. Where K is a sortal term, logy of concepts. But we speak of thinkers
the general form of a criterion of identity grasping or failing to grasp concepts. We may
will be this: If x and y are Ks, then x is iden- take this simply to be a matter of their being
tical with y if and only if x is RK-related to y able, or not being able, to think of things
(Lowe, 1989b). Here RK denotes a certain in certain ways. Someone who grasps the
equivalence relation on Ks a relation which concept of a cat is able to think of certain

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 29

i ndivi duat i on
things in this case, certain living organisms applicable: so that the sortal term cat, for
in a certain way. What way is that? Well, instance, denotes the cats all of them
of course, such a person is able to think of that exist (or, perhaps, all that do, did, or will
certain living organisms as being cats. And exist). Another view and more popular
what does this involve? Well, among other view is that a sortal term that has denota-
things, it involves being able to think of tion denotes a sort or kind of things con-
these organisms as possessing certain char- ceived as a type of UNIVERSAL , which has as
acteristic properties, such as furriness and its particular instances all of the particular
warmbloodedness, and most importantly things to which it is applicable. According
for present purposes as satisfying a certain to this view, the sortal term cat denotes a
criterion of identity. We neednt suppose, substantial universal or kind, whose par-
however, that a person who grasps the con- ticular instances are all the individual cats
cept of a cat must be able to articulate such that do, did, or will exist (Lowe, 1989a,
a criterion in an explicit form, in line with pp. 157 63).
the general form of a criterion of identity Now let us turn to another key notion
stated earlier. Indeed, it is notoriously difficult that needs to be clarified for present pur-
even for philosophers to formulate clear poses: that of an object. This is a philosophical
and uncontroversial criteria of identity for term of art, which admits of various different
many kinds of things, even when we seem interpretations, some narrower than others.
to have a good implicit grasp of such criteria In its very broadest use, object is inter-
that is manifested in our ability to make changeable with the very general all-purpose
confident identity-judgments concerning term entity. In this sense, anything what-
things of those kinds. ever that does or could exist is an object,
So far, we have discussed sortal terms including numbers, properties, propositions,
and sortal concepts. In addition, however, events, surfaces, waves, holes, and places, as
there are sorts or kinds, which sortal terms well as common-or-garden material objects,
and concepts purportedly designate. I say such as apples, tables, and rocks. However,
purportedly for this reason if for no other: I propose to use the term object more
even granted that a sortal term or concept narrowly to mean an entity that does at
may designate a really existing sort of things, least possess determinate identity conditions
we can hardly insist that it must do so. The and the kind of unity that makes it something
point is exactly parallel to one that may that is, at least in principle, countable (Lowe,
be made concerning adjectival terms and 1998, pp. 5861, Lowe, 2006, pp. 75 6).
concepts or, more generally, predicates and Some of the items listed earlier do not indis-
predicative concepts: that they may, but putably qualify as objects by this criterion:
need not, designate anything. For example, for example, waves do not. In what follows
it is natural to suppose that red denotes we shall restrict our attention for the
a certain color property or quality, redness. most part to material objects. However, it is
But, for familiar reasons, it may disputed important to recognize that the notion of a
whether there really are any color properties material object is still a very broad notion
at all. It would, of course, be quite extrava- indeed. Crucially, material objects do not
gant to suppose that cats dont exist, but collectively constitute a sort or kind in
the history of human thought is replete the sense discussed earlier. In other words,
with examples of sortal terms that failed material object is not a sortal term and does
to designate anything, such as mermaid, not express or convey a sortal concept. The
dragon, unicorn, and centaur. What, reason is simple enough: it is simply not the
however, should we say concerning the sortal case that all material objects are governed
terms that do designate or denote some- by the same criterion of identity. Thus,
thing: what, exactly, do they denote? Various for example, both cats and mountains are
answers are possible, one being that they material objects, but they do not share the
denote, in plural fashion, all of the various same criterion of identity. All it takes for
particular things to which they are something to qualify as a material object

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 30

i n d i v i duatio n
is that it (a) be an object, in our narrower convey the same criterion of identity, as must
sense, and (b) be composed of matter. Both the sortal terms dog and animal, on pain
cats and mountains qualify by this stand- of incoherence. For it is not even metaphys-
ard, as do many other material objects ically possible that objects of kinds governed
governed by yet other criteria of identity, by different criteria of identity should be
such as tables and stars. identical (Lowe, 1989a, ch. 4). Hence if
The next important thing to notice is the sortal terms cat and animal, say,
this. Although two different sortal terms, conveyed different criteria of identity, no
each designating a different sort or kind individual cat could be identified with any
of things, may convey different criteria of individual animal, which is plainly absurd.
identity for the individual objects to which But if cat and dog must, for the forego-
they apply, this is not necessarily the case ing reason, both convey the same criterion
and, indeed, is very often not the case. Very of identity as animal does, then they must
often, two such sortal terms convey exactly clearly convey the same criterion of identity
the same criterion of identity. This is the as each other and the same applies in the
case, for instance, with the sortal terms cat case of all other sortal terms denoting
and dog and, indeed, with all sortal terms animal kinds. This, then, is why I maintain
denoting kinds of living organism (Lowe, that all animal kinds share the same crite-
1998, p. 45). Particulars of all these kinds rion of identity.
share the same criterion of identity, which The foregoing discussion, if it is along
is that of living organisms in general. So it the right lines, reveals that general names
is likewise with all kinds of material artefact, as Locke would have called them fall into
for instance, such as tables and computers: at least three distinct classes. First, there are
they all share the same criterion of identity, non-adjectival general terms like material
which differs from that governing living object which are certainly not sortal terms,
organisms. But why, it may be asked, must because they do not convey any criterion
we suppose that all living organisms, say of identity whatever. Second, there are
and certainly all animals share the same cri- regular sortal terms, such as cat, dog,
terion of identity? For the following reason. mountain, star, and table, which not
Animal unlike, for instance, material only convey a criterion of identity but also
object does at least appear to be a sortal purportedly denote certain distinct sorts or
term in good standing, conveying a criterion kinds. Intermediate in generality between
of identity for the objects to which it applies. these two classes of general names are
After all, we can always intelligibly ask non-adjectival general terms like living
whether an individual animal encountered organism and material artefact, which
on one occasion is or is not identical with do convey a criterion of identity but are too
another individual animal encountered on general to qualify as regular sortal terms.
another occasion and in order to determine What these terms designate are not, properly
the answer to such a question, we do not speaking, specific sorts or kinds but, rather,
necessarily need to know what sort or sorts certain ontological CATEGORIES or, more
of animal these individuals are. Indeed, we precisely, certain categories of object (com-
may well be uncertain, at least at an early pare Dummett, 1981, p. 583). What cats
stage of our inquiries, whether we are con- and dogs and all other such sorts or kinds
fronted with just one sort of animal or two, have in common is that they are all kinds
because the individual animals encoun- of living organism. The individual members of
tered on the two occasions may exhibit very all these kinds all belong to the same onto-
considerable morphological differences, as logical category, the hallmark of this fact
in the case of a tadpole and a mature frog. precisely being that they are all governed by
However, cats and dogs, say, clearly are the same criterion of identity. In effect,
both sorts of animal. (Indeed, they are clearly then, we can identify those general terms that
different sorts of animal.) But in that case denote ontological categories categorial
the sortal terms cat and animal must terms, as we may aptly call them as being

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 31

i ndivi duat i on
the most general terms that still convey decomposition.) Material objects of some
criteria of identity for the objects to which other kinds, however, are apparently not
they apply. And categorial terms fall, in individuated by their material parts (at any
respect of their degree of generality, in level of decomposition). Living organisms
between regular sortal terms and transcat- seem to be a case in point, for they can
egorial terms, such as material object. I undergo a change of any of the material
must emphasize that criteria of identity, on parts that they possess at any time during
this view, are not empirically discoverable their careers. It is not even clear that as
principles, but are, rather, a priori ontolo- some philosophers suggest (for instance,
gical principles which delimit what is and is Kripke 1980) living organisms are indi-
not metaphysically possible for the objects viduated by the material parts that they
governed by them (Lowe, 1998, ch. 8). possess at their moment of origin, since it
With this stage-setting in place, we can seems that these too could always have
now at last introduce the term individua- been different (Lowe, 1998, pp. 165 6).
tion itself. This term has two senses (Lowe, It should be clear from these examples
2003). In one sense which we may call the that metaphysical principles of individu-
cognitive sense individuation is a cognitive ation are closely related to criteria of identity.
achievement, consisting in the singling out But they should not be confused with them.
of an object in thought (compare Wiggins, A metaphysical principle of individuation
2001, pp. 67). In this sense, it is we, or tells us what determines the identity of an
thinkers quite generally, who individuate object, in the sense that it tells us what
objects, whenever we single them out in determines which object it is. A criterion of
thought. But in a quite different sense identity, by contrast, tells us what deter-
which we may call the ontological sense indi- mines whether an object belonging to a
viduation has nothing to do with cognition given ontological category is or is not ident-
or thinkers, but is simply a certain kind of ical with another such object. In the latter
metaphysical determination relation between case, we are concerned with identity con-
entities. In this sense, an object is individu- ceived as a relation, whereas in the former
ated by one or more other entities, its case we are concerned with identity in the
individuator or individuators. An objects sense of individual ESSENCE (to use a traditional
individuators, in this ontological sense, are term). Identity in this sense, or individual
the entities which determine which object essence, is as John Locke aptly put it
it is. A simple example drawn from the the very being of any thing, whereby it is,
domain of abstract objects will serve for what it is, this being, according to Locke, the
illustrative purposes. A set, then, is individu- proper original signification of the word
ated, in the ontological sense, by the enti- essence (Locke, 1975, III, III, p. 15).
ties that are its members, at least in all cases Plausibly, every object is individuated in the
in which it has members (not, thus, in the ontological sense. In every case, something
case of the empty set). If a set has members, some entity or entities individuates the
its members, and these entities alone, deter- object in question, in the sense of determin-
mine which set it is. Turning to the case of ing which object it is. To suppose that there
material objects, we can see that material are unindividuated objects seems incoherent.
objects of some kinds are individuated by For an unindividuated object would be an
their material parts (at least at some level object concerning which there was no fact of
of decomposition): for example, a heap of the matter as to which object it was, and it is
stones is individuated by the stones that very hard to see how this could be the case.
make it up, because which heap it is is deter- Now, clearly, objects of different ontological
mined by which stones make it up. (Note, categories are individuated, in the onto-
however, that a heap of stones is not indi- logical sense, in very different ways. For
viduated by the subatomic particles that example, mountains and islands are indi-
make it up at any given time, which is why viduated, at least partly, by their geograph-
it is important to specify the relevant level of ical locations. But living organisms and

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 32

i n d i v i duatio n
material artefacts are plainly not. However, already been noted, we cannot uncritically
the claim that every object is individuated assume that every predicate denotes a pro-
might raise in the minds of some critics the perty, if by a property we mean some really
worry that an infinite regress is thereby existing entity, be it a universal or a so-called
threatened. The thought would be that if trope. There is no obvious reason to suppose
every object is individuated and, moreover, that the predicate now in question denotes
is individuated by one or more objects, then a property in this sense. That being so, it is
there is no end to individuation and so, per- hard to see what we can say about the
haps, no object at all really gets individuated. empty set other than that it is self-
However, there are two ways, at least, to individuating. It alone is the only entity that
counter this worry. One is to point out that determines which set it is, since nothing
it was implied earlier only that every object else does.
is individuated by some entity or entities, However, it may seem that, because the
but not that the entities in question must empty set is an abstract object, we can draw
always themselves be objects. Indeed, I said few lessons from its case when considering
that mountains and islands are partly the individuation of material objects. But that
individuated by their geographical locations, conclusion would be too hasty. For what
but geographical locations are doubtfully makes it plausible to say that the empty
objects at all and are certainly not material set is self-individuating is the fact that it
objects. Another point to bear in mind is is an object that does not appear to depend
that nothing said so far implies that objects for its identity on anything other than itself
may never be self-individuating. In fact, it (on this notion of identity-dependence, see
is plausible to claim that what we may Lowe, 1998, pp. 1479). But this also seems
call material SUBSTANCES are indeed self- to be a characteristic of what we are calling
individuating, including living organisms. material substances, including living organ-
According to this view, for example, what isms. We may take it to be an essential fea-
determines which animal a given animal is ture of such substances that, even though
is nothing other than that very animal. they are composed of matter, they are
The idea that some objects are self- capable of changing their material parts
individuating is certainly far from being and, indeed, could have been made up, at any
absurd (Lowe, 2003). Indeed, in some cases given time, of material parts numerically
it seems extremely compelling: for instance, distinct from those that actually make them
in the case of the empty set. For, given that up at that time. This, if true, is why they
every set is individuated and that sets which do not depend for their identity upon such
have members are individuated purely by parts, in the way that something like a heap
those members, we seem to have little option or pile of stones does. But, given that they
but to say that the empty set individuates do not depend for their identity upon their
itself, for it has no members to individuate material parts, it is not clear what else they
it in the only way that other sets are indi- could depend on for their identity, other
viduated. In opposition to this view, it than simply themselves.
might be suggested that the empty set is in Perhaps, in the end, saying that material
fact individuated by a certain property that substances are self-individuating is not so
it alone possesses and possesses necessarily very different from saying, as some meta-
the property of being the only set that has physicians do, that they are individuated
no members and that since this property by their so-called HAECCEITIES or thisnesses
is an entity that is distinct from the empty (Rosenkrantz, 1993). According to this view,
set itself, that set is not self-individuating. what determines which animal a given
However, this assumes that the predicate animal, a, is is that animals haecceity its
is the only set that has no members, property of being that animal or, in other
which undoubtedly applies uniquely to the words, its property of being identical with
empty set, does indeed denote a certain a. But, it may be asked, are there really
property which that set possesses. But, as has such properties as the property of being

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 33

i ndivi duat i on
identical with a? Is the property of being single out an object in thought before hav-
identical with a really an entity that is dis- ing any conception of what sort of object it
tinct from a itself? Some metaphysicians is. For example, in thinking about a particu-
may find it hard to believe so. But if haec- lar animal lets call it Tom a thinker
ceities are not genuine entities in their own surely need not be able to single out Tom as
right, it is difficult to see what it can mean being, say, a cat, as opposed to a dog, or a pig.
to say that animal a is individuated by its Maybe so. But can a thinker successfully
property of being identical with a, other single out in thought a particular animal,
than simply to say that a individuates itself such as Tom, without even grasping that
that a itself is the only entity that determines Tom is an animal, or at least a living organ-
which animal a is. And this is, I think, a per- ism? Is it possible, for example, for a thinker
fectly coherent thing to say. Nor should it be successfully to single out in thought a par-
supposed that once we say this about some ticular animal, Tom, while grasping only
objects, we shall be obliged to say it about all that Tom is a material object? It is hard to see
objects. For we have already seen that there how this can be possible. For, it seems, one
are plenty of objects, such as piles of stones cannot successfully single out an object in
and sets that have members, which are thought without grasping which object it is
plainly not self-individuating. Furthermore, that one has thus singled out. However,
it seems very reasonable to say that at least this is the point at which the cognitive and
some entities must be self-individuating, the metaphysical notions of individuation
on pain of the sort of infinite regress that come together in a crucial way. Which object
was mooted earlier. (Thus, if there are haec- a given object is is something that is deter-
ceities, must not they be self-individuating?) mined by that objects individuator or
So why not say this about material sub- individuators and, as we have seen, objects
stances, together with, perhaps, other objects of different types have different types of indi-
such as the empty set? Anyway, let us adopt viduator. Material objects as such have no sin-
it as a working assumption in what follows gle type of individuator, because material
that material substances, including animals objects as such do not constitute an onto-
and other living organisms, are indeed self- logical category but, rather, fall into many
individuating in the ontological sense. diverse ontological categories, such as living
So far, however, I have said very little organisms, material artefacts, and geological
about individuation in the cognitive sense, but formations.
this notion too raises important metaphysical Turning aside, for a moment, from the case
issues, concerning the nature of thought. of material objects, consider the following
What I did say is that individuation in this question: can we intelligibly suppose that
sense is a cognitive achievement, consisting a thinker could successfully single out in
in the singling out of an object in thought by thought an abstract object, such as a set for
a thinker, that is, by a person. A sortalist, in example, the set of prime numbers smaller
this connection, is a theorist who maintains than 10, {2, 3, 5, 7}, or the set of planets
that a thinker can successfully single out an closer to the sun than Jupiter, {Mercury,
object in thought on a given occasion only Venus, Earth, Mars} without grasping that
as an object of some specific sort, that is, as the object in question is indeed a set and
falling under or satisfying some specific sor- thereby grasping its criterion of identity and
tal concept a concept that the thinker in principle of individuation? For, assuming
question must therefore grasp and apply in as we now are that the set in question is
individuating that object on that occasion, non-empty and therefore individuated by
in the cognitive sense of individuate. An its members, how could a thinker know
anti-sortalist, correspondingly, is a theorist which object this set is without grasping
who denies the foregoing claim. On the face what it is that determines which object it is,
of it, the sortalist thesis as I have just for- namely, its members something that, it
mulated it is clearly too strong. For, it may seems, requires the thinker to grasp that
be urged, a thinker can surely successfully what this object is is a set. But if a thinker

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 34

i n d i v i duatio n
does not know which object it is that he is requires us to suppose that thinkers cannot
thinking about, how can he really be said to perceive objects which do not fall under
have singled out that object in thought? sortal concepts grasped by them. It is then
To single out an object in thought is, at the pointed out that very frequently we find our-
very least, to think something about that selves perceiving some object while simply
very object. But how can a thinkers thoughts having no idea at all what sort of object it is
be said to fasten upon a certain object in par- that we are perceiving. This may happen
ticular, as opposed to some other object, if that when, for example, an archaic artefact
thinker cannot even be said to know which of unknown purpose is dug up and we
object it is that he is thinking about? ask ourselves, What on earth is this a
It doesnt appear that matters are funda- drinking vessel, perhaps, or an oil lamp, or
mentally different in the case of thoughts something designed to be used in a religious
about material, as opposed to abstract, rite? (compare Campbell, 2002, pp. 701).
objects. Accordingly, it is hard to see how We undoubtedly see and feel the object,
a thinker could successfully single out a however, and can talk about it intelligibly. So
material object in thought while conceiving is this not a case in which we have managed
of it as nothing more specific than a material to single out the object in thought but without
object that is, as an object composed of having a sortal conception of it, quite contrary
matter. For conceiving of an object in this to the sortalist thesis? The first thing that
way would leave entirely open the question must be said about this type of example
of what determined which object it is and is that, of course, we have already conceded
yet, without his having a grasp of what a cor- that the sortalist thesis is too strong. The most
rect answer to that question would be it is that we should say is that we cannot single
hard to see how a thinker could be said out an object in thought without having, at
to know which object he was, supposedly, least implicitly, a categorial conception of
thinking about. Thus, while we should be it, and thereby having at least an implicit
happy to allow that a thinker can success- grasp of the criterion of identity that the
fully single out a material object in thought object satisfies. We could call this the cat-
without conceiving of it as belonging to egorialist thesis, as opposed to the stronger sor-
some quite specific sort or kind, such as the talist thesis. The latter is stronger, because it
kind cat, or the kind table, or the kind moun- implies the former, but the reverse is not the
tain, we should insist that he must grasp, at case. Now, in the foregoing archaeological
least implicitly, to what ontological category example, no challenge to the categorialist
the object in question belongs such as thesis was even threatened, since we were
living organism, or material artefact, or geo- supposing the discoverers of the mysterious
logical formation. This is not at all to imply, object in question to be convinced, at least,
of course, that the thinker need be able to that what they had found was a material
have a linguistic command of such categor- artefact of some kind and material artefacts
ial terms as these, only that he must have constitute an ontological category.
at least an implicit grasp of the relevant cri- In this context, it is vitally important to
teria of identity and principles of individuation. distinguish between thought and perception.
For without such a grasp the thinker cannot The categorialist thesis is the claim that a
really be said to know what it is that he thinker cannot successfully single out an
is, supposedly, thinking about. And without object in thought without conceiving of that
knowing that, he cannot really be said to have object as falling under a certain ontological
singled out an object in thought. category and thereby grasping a correspond-
However, it is unlikely that this claim ing criterion of identity that he conceives
will go entirely unchallenged. One kind of it to satisfy. But perceiving is not thinking and
challenge that is likely to be raised against there is no reason at all why the cat-
it focuses on the perceptual capacities of egorialist should not accept that a person
thinkers. Against sortalists, it is sometimes can perceive an object without having any
complained that their position improbably conception whatever as to what ontological

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 35

i ndivi duat i on
category it falls under. Indeed, there are makes me acquainted with as individuator.
compelling reasons to accept precisely this. But if one grasps an objects principle of
For it is evident that many non-human individuation and is also acquainted with
animals perceive objects in their immediate the entities which, according to that prin-
environment, even though it would be utterly ciple, are its individuators, then one knows
extravagant to suppose that those animals which object it is. For example, if I grasp
are capable of categorizing those objects the principle of individuation for sets and
ontologically or grasping the relevant cri- am acquainted with the prime numbers
teria of identity for those objects. A dog, smaller than 10, then I know which set, and
for instance, can surely see its feeding bowl, hence which object, the set of prime numbers
without recognizing that what it sees is a smaller than 10 is. What is special about
material artefact. But, conceding this, let us material substances together, maybe,
then ask: can the dog successfully single out with some other objects, such as the empty
that object in thought? Can the dog think set is that a thinker does not need to be
about its feeding bowl that very object, acquainted with anything else in order to be
as distinct from any other? There seems acquainted with such a substances individ-
to be no compelling reason to suppose that uator, so that a grasp of such a substances
it can. general essence together with perceptual
We may conclude now with a final ques- acquaintance with that substance provides
tion: what cognitive significance, if any, is a thinker with a grasp of that substances
there in the fact assuming that it is a individual essence and thereby an ability to
fact that material substances are, in the single out that substance in thought, that
ontological sense of individuation, self- is, to individuate it in the cognitive sense.
individuating? There seems to be consider- But whether this is the only way in which a
able cognitive significance in this fact. For thinker can acquire a grasp of the individual
what it apparently implies is that it is suf- essence of a material substance is another
ficient for a thinker to be able to single out and difficult question.
such a substance in thought that that thinker
should have perceived that substance at See also the az entry on individuation.
some time, knowing on that occasion that
what he was perceiving was an instance of
bibl i ography
a certain category of material substance,
and have retained a memory of this experi- Campbell, J.: Reference and Consciousness
ence. For instance, having seen a certain (Oxford: Clarendon Press, 2002).
animal, a, knowing that what I was then Dummett, M.A.E.: Frege: Philosophy of
seeing was an animal, and remembering Language, 2nd edn. (London: Duckworth,
this perceptual encounter with a, I can sub- 1981).
sequently have singular thoughts about a Kripke, S.A.: Naming and Necessity (Oxford:
that is to say, I can continue to single out Blackwell, 1980).
a in thought. In other words, I continue to Locke, J.: An Essay Concerning Human
know which animal a is, even if I never again Understanding, ed. P. H. Nidditch (Oxford:
have perceptual contact with a. To put this Clarendon Press, 1975 [1690]).
another way, I continue to grasp as indi- Lowe, E.J.: The Four-Category Ontology: A
vidual essence. For, we are supposing, what Metaphysical Foundation for Natural Science
individuates a, in the ontological sense, is (Oxford: Clarendon Press, 2006).
just a itself it is just a itself that deter- Lowe, E.J.: Individuation, in M.J. Loux and
mines which object a is. Hence, my percep- D.W. Zimmerman, ed., The Oxford Hand-
tual encounter with a, provided that it is book of Metaphysics (Oxford: Oxford
informed by a grasp of the category of object University Press, 2003), 7595.
to which a belongs and thus a grasp of Lowe, E.J.: Kinds of Being: A Study of Indi-
as general essence, which it shares with viduation, Identity and the Logic of Sortal
all other members of the same category Terms (Oxford: Blackwell, 1989a).

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 36

t h e m i nd / b o d y p r o b lem
Lowe, E.J.: The Possibility of Metaphysics: are irreconcilable (bodies are extended in
Substance, Identity, and Time (Oxford: space, minds are non-spatial), so modes
Clarendon Press, 1998). of thought and modes of extension are
Lowe, E.J.: What Is a Criterion of Identity?, incommensurable. Now we are faced with a
The Philosophical Quarterly 39 (1989b), problem: how could mental goings-on have
121. material effects; how could material occur-
Rosenkrantz, G.S.: Haecceity: An Ontological rences affect the mind? This is Descartess
Essay (Dordrecht: Kluwer, 1993). mindbody problem.
Wiggins, D.: Sameness and Substance Renewed In fact there are two problems here.
(Cambridge: Cambridge University Press, The first arises from the difficulty of under-
2001). standing how spatial and non-spatial entities
could engage causally. The difficulty is especi-
e.j. lowe
ally pressing for Descartes who regards
mental and physical substances as operating
on very different laws or principles.
The Mind/Body Problem A second difficulty arises from our concep-
Reinhardt Grossmann calls the mind the tion of the physical world as a self-contained
great garbage bin of ontology (1983, closed system. Physical events have, we sup-
p. 256). What seems real but lacks physical pose, purely physical causes. If non-physical
respectability we consign to the mind. A minds affect the physical world, it looks
long tradition places secondary qualities as though they would have to initiate or
(colors, tastes, sounds, odors) in the mind. intervene in physical processes. Were that so,
These are thought, not to be out there, but the physical world would not be a closed
to be subjective transitory occurrences in system governed by physical law a daunt-
the minds of observers (see quality, primary/ ing prospect that threatens the garbage-bin
secondary). Hume regarded causation status of the mental.
(see the extended essay) as a psychological
The self-contained nature of the physical
projection, and it seems natural to distin-
world could be expressed in terms of a
guish the world as experienced from the
conservation principle. Descartes, writing
world as it is. The idea that minds incor-
before Newton, imagined that what was
porate non-worldly, non-physical elements,
conserved was motion. Minds could not
however, evidently places minds outside the
initiate or inhibit motion in the physical
physical realm. What science casts asunder,
world. Minds could, however, have phy-
philosophers must piece together. Hence the
sical effects without violating physical
mindbody problem.
closure by altering the direction taken
Although he did not invent the mind
by material particles. This solution un-
body problem, Descartes (15961650)
raveled with Newtons introduction of
is responsible for its modern formulation
force, which moved physics from Car-
(see Matson, 1966). Immediately after prov-
tesian kinematics to a dynamical system.
ing his existence by noting that the thought
Nowadays we think that what is con-
expressed by I exist must be true if I
served is massenergy. In either case
can so much as consider whether it is true
Descartess account of mindbody inter-
(Meditation 2), Descartes asks, What am I?
action is no longer viable.
He answers, a thing that thinks, a think-
ing substance. Descartes regards planets Malebranche (16381715), a Cartesian,
and trees, not as substances, but as modes, sees the problem and rejects interaction.
ways extended matter is organized. On the According to Malebranche (and there are
one hand we have extended substance and suggestions of such a view in Descartes),
its modes: material bodies (see matter). On not only is there no mentalphysical causa-
the other hand, we have thinking substances, tion, there is no purely physical causation.
minds, and their modes: thoughts, images, Whatever happens is the result of Gods
feeling (see soul). Just as minds and bodies making it the case that mental and physical

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 37

t he mi nd / body probl em
substances are as they are at every moment. Many of Descartess contemporaries and
The world resembles a succession of images most of his successors rejected this picture.
on a movie screen. In a movie, events on A mental substance might be a substance
the screen succeed each other. Their cause, with mental properties; a physical substance,
a projector, is not a member of the sequence, one with physical properties. This leaves
however, but something entirely outside it. open another possibility: some substances
For Malebranche, God does not cause, but might have both mental and physical prop-
occasions events in the world. God does erties, a dualism of properties, not sub-
this, not by intervening in worldly processes, stances. Perhaps mental properties are just
but making it the case at every instant that distinctive properties of certain complex
a world exists containing those processes physical systems.
(see occasion, occasionalism). We should Property dualism can be developed in vari-
not be shocked by the thought that mental ous ways. According to Epiphenomenalists
events are causally impotent: physical events T. H. Huxley (18251895), for instance
are in the same boat! mental occurrences are by-products of brain
Leibniz (16461716) depicts a world processes. They resemble squeaks made by
comprising an infinity of independent sub- a complex machine that play no role in the
stances each reflecting the world from a machines operation. When you bark your
unique point of view. On this conception shin, you feel a pain. This feeling is a result
the physical world amounts to a virtual of a chain of events in your nervous system
world made up of these points of view. leading from your shin to a region of your
Events unfold in each substance independ- brain. In the simplest case, the neurological
ently but in perfect harmony with events event that gives rise to your painful sen-
in every other substance. Causal interac- sation also produces bodily motions that
tion is a harmless illusion. might otherwise be thought to be caused by
Both Malebranche and Leibniz skirt the the sensation. Conscious states and bodily
mindbody problem by rejecting mental motions are correlated, not because con-
physical interaction altogether. If there is sciousness is causally efficacious, but because
no mindbody interaction, there is no mind conscious states and bodily motions have
body problem. Such maneuvers, however, common causes.
exact a heavy price. Can we reasonably On the one hand, epiphenomenalism
abandon the idea that physical events are enables us to sidestep worries about mental
causally connected? Could we ever be satisfied goings-on intervening in the physical world
with an account of the world according thereby violating closure. On the other hand,
to which mental occurrences perceptual we are left with two significant worries. First,
experiences, for instance are not brought as in the case of Malebranche and Leibniz, we
about by physical occurrences, and thoughts will need to abandon the idea that men-
and decisions never give rise to actions tality makes a difference in what we do.
and utterances? Must we settle for the idea You might worry about this, not merely
that mindbody interaction is illusory? because it seems on the face of it implaus-
For Descartes, mental and physical sub- ible, but because it is hard to see how con-
stances are, God aside, mutually exclusive and sciousness could possibly bestow any sort of
exhaustive. Each kind of substance has a dis- evolutionary advantage on creatures pos-
tinctive attribute: mental substances think, sessing it. True, consciousness could be an
but are not extended; physical substances evitable by-product of evolutionarily adaptive
are extended, but do not think. Mental and physical processes, but it is hard to believe
physical properties are modes of these attri- that consciousness itself is evolutionarily
butes, determinate ways be being extended irrelevant (Nichols and Grantham, 2000).
or thinking. Being spherical and being red are A second worry concerns the production
ways of being extended. An experience of of conscious experiences. These are caused
a spherical red object, in contrast is a mode by physical processes in the brain, but how
of thought, a way of being conscious. is this supposed to work? What exactly is

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 38

t h e m i nd / b o d y p r o b lem
involved in the production of a non-physical states and properties. Second, you might
event? simply deny that there are minds or mental
Epiphenomenalists tell us that conscious- states or properties (Churchland, 1981; Stich,
ness arises from the brain, but what is 1996). To see the difference, consider the
this arising from relation? Mental events discovery of DNA and its consequences for
presumably involve mental properties, but genetics. We now think we can map genes
where are these properties? They seem not onto complex molecular structures, thereby
to be among those we discover when we reducing genes to DNA (see reduction,
probe the brain. Are they invisible? Are they reductionism). Compare reduction of this
somehow outside space or spacetime? The kind to the demise of phlogiston. Seventeenth-
Cartesian problem concerned how extended century chemists explained combustion by
and non-extended things could interact. supposing that flammable materials con-
The problem arises anew for epiphenom- tained phlogiston, a fluid driven out when the
enalism in relation to the production of materials were heated. Advances in chemistry
mental properties or events. The situation rendered phlogiston superfluous. Phlogiston
appears bleak. We have a robust conviction was not reduced to more fundamental
that, although mental and physical pro- goings-on, but stricken from the scientific
perties are utterly different, interaction inventory. Eliminativists believe a similar
between minds and bodies is commonplace. fate lies in store for the mind.
The difficulty is to square this with closure, According to eliminativists, talk of mental
our conviction that the physical world as states and properties belongs to an out-
a whole is causally closed, massenergy is moded folk theory of human and animal
conserved. (For a dissenting view, see Lowe, behavior. At one time we explained natural
1996.) occurrences by supposing objects were
One elegant solution is to deny the exist- animated by spirits. Such explanations
ence of minds and mental properties alto- were gradually supplanted by explanations
gether. If there are no minds, no mental adverting exclusively to physical processes.
properties, there is no mindbody problem. Nevertheless, we persist in regarding human
Hobbes (15881679) argued that we are bodies (and the bodies of most animals)
nothing more than elaborate machines. In as animated by spirits. We comprehend the
a way, Hobbes is just extending Descartess behavior of intelligent creatures by suppos-
official view. Descartes held that most human ing they are conscious of their surround-
behavior and all behavior of non-human ings and do what they believe will subserve
creatures could be explained mechanically. their interests. Advances in neuroscience,
Only in the case of behavior resulting from however, promise to undermine folk psy-
rational mental processes (most notably chology and its posits just as chemical
linguistic behavior), do we need to posit discoveries undermined phlogiston.
mental causes. If ratiocination, however, You might worry that this way of fram-
were just a matter of calculation (think of a ing the issues stacks the deck. Consider
computing machine to get a feel for what ordinary beliefs about ordinary objects:
Hobbes has in mind) we would have no tables, trees, volcanoes. Physics and chem-
need to imagine that our bodies are con- istry assure us that these things are at bot-
trolled by minds with distinctive mental tom just clouds of particles. We can explain
properties. the behavior of these particles without
A conception of this kind, materialism, can positing the ordinary entities, and there is no
be developed in two ways (see physicalism/ prospect of smoothly reducing the ordin-
materialism). First, you might think, as ary things to respectable physical chemical
Hobbes does, that mental states and proper- kinds. Should we eliminate tables, trees,
ties are reducible to, that is identifiable volcanoes? Mightnt it be better to see talk of
with, physical states and processes. On this tables, trees, and volcanoes as reflecting an
view, minds turn out to be brains, mental inventory of genuine objects that happen to
states and properties turn out to be physical be of no interest to the physicist or chemist?

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 39

t he mi nd / body probl em
Physics and chemistry provide us with the are conditioned by our involvement with
deep story about the world, a world that the world to do as we do. The mechanisms
includes the fundamental things and includes are simple but, in combination, yield complex
as well tables, trees, and volcanoes. These are responses.
not add-ons any more that the forest is Meanwhile, philosophers, inspired by
something in addition to the trees. Wittgensteins (1953, 38) insistence that
Whether or not you are moved by such philosophical problems arise when lan-
considerations, even tough-minded philo- guage goes on holiday, were crafting
sophers have found eliminativism hard to a philosophical version of behaviorism.
swallow. We can explain away elimin- Gilbert Ryle campaigned against the
ate, consign to the garbage bin ghosts Cartesian myth, the conception of minds
by supposing that they are illusions, but it is as ghosts in the machine. The mistake,
hard to see how this could work with states thought Ryle, was to regard mental events
of consciousness. Illusions seem ineluctably as private, inwardly observable goings-on
mental. An illusory feeling of pain is still a that, while not quite physical, had physical
feeling. Conceiving of mental phenomena causes (incoming stimuli) and effects (bodily
as only in the mind is scarcely a recipe for motions). Ryle thought this picture stemmed
their elimination. The problem of reconciling from a category mistake (see categories):
illusions with the physical world is just the representing the facts of mental life as if they
mindbody problem all over again. belonged to one logical type or category . . .
Materialism dissolves the mind-body when they actually belong to another
problem by subtracting the mental as a dis- (1949, p. 16). A child, watching a parade,
tinct category (see physicalism/material- is told that a regiment is marching past.
ism). Others, idealists, move in the opposite Puzzled, the child remarks, I see soldiers, but
direction: all that exists are minds and their where is the regiment? The child thinks
contents. The physical world is, as George a regiment is something alongside or over
Berkeley (16851753) would put it, a and above the soldiers, a peculiar sort
mere appearance. (For a more recent of object. So it is with us and the mind.
variant, see Foster, 1982.) One advantage of Scrutinizing the body, we fail to observe
idealism is that it is not hard to see how the mind and conclude that minds must
physical objects could turn out to be be organs like the brain but invisible to out-
illusory. A disadvantage is that idealism side observers. Rather, Ryle thinks, talk of
addresses the world in a way deeply at odds minds and states of mind is a way of indi-
with tenor, if not the substance, of modern cating what intelligent agents do or would
science. The sense is that idealism works, do under various circumstances. Thoughts
but only by tossing out the baby with the and feelings are not inner states. Your
bath water. thinking of Vienna is just a matter of your
The urge for scientific respectability under- being disposed to remark on Vienna or
lies the advent of psychological behaviorism respond with Vienna when queried.
during the first half of the twentieth century. Neither Wittgenstein nor Ryle denied
Behaviorists were intent upon distancing that there were inner states, only that states
themselves from reliance on introspective of mind were identifiable with such states.
techniques to study states of consciousness Their aim was to challenge the picture of
prominent in the nineteenth century. By mental goings on as being causally related
their lights this meant providing tough- to physical goings on. Your forming an
minded operational characterizations of intention to stroll does not cause your sub-
important concepts and shunning anything sequent strolling. Puzzling over mindbody
that might prove objectively unverifiable interaction in such cases manifests a category
(Skinner, 1963). The result was psychology confusion. Your intention illuminates or
minus the mental trappings. Behavior was makes sense of your subsequent action.
to be explained by contingencies of rein- Actions, which presumably have purely
forcement and operant conditioning. We physical causes, are understood in light of

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 40

t h e m i nd / b o d y p r o b lem
thoughts and desires. The philosophical of mind, its what-its-like-ness, its sub-
mistake is to see these states as ghostly jectivity (Nagel, 1974), seem utterly unlike
internal causes of behavior. any physical properties we might hope to
Despite attempts to move us away from the discover in the brain. Smart noted that this
Cartesian model of minds as inner control was so with lightning and electrical dis-
centers, philosophical behaviorism came charges, water and H2O. In both cases prop-
under fire from philosophers who found erties encountered in experience differed
behaviorist analyses of mental states im- from those we discover via scientific invest-
plausible. Such analyses seek to reduce talk igation, yet this does not prevent us from
of mental states to talk of behavior or identifying lightning and water with elec-
behavioral dispositions (see disposition). trical discharges and H2O, respectively. In
If you believe the ice is thin, you will avoid the case of water and lightning, however,
skating on it, or at least be disposed to avoid we compare properties of the appearance of
skating on it but only assuming that you water or lightning with properties of the
want not to fall through. Your wanting not stuff that gives rise to the appearance. In
to fall through could be analyzed behavi- the case of minds and brains, the roles are
orally, but only by mentioning still further reversed. What we are trying to explain are
states of mind. What we do or would do the appearances. It would be futile to suggest
depends, it would seem, on interrelations that we are aware only of the appearances
among beliefs and desires, and this resists of states of mind.
reductive analysis. Philosophical behaviorism succumbed to
Whatever states of mind are, they do seem pressure from the identity theory and trans-
to affect behavior causally and to be caus- formed itself into functionalism. The stum-
ally responsive to perceptual inputs from bling block for psychological behaviorism
the environment. In the 1950s, U.T. Place came with the advent of the computing
(1956) and J.J.C. Smart (1959), colleagues at machine and the Chomskeyean revolution
the University of Adelaide, put forward a in linguistics. Chomsky (1966) argued that
mindbrain identity thesis. Mental states, behaviorist categories were hopelessly in-
although not analyzable in physical terms, adequate to account for human linguistic
might nevertheless be identified with states capacities. At the same time, computing
of the brain: sensations are brain-processes. machines were coming to be seen as af-
This is not something that could be worked fording explanatorily tractable models of
out solely by attending introspectively to intelligent behavior. Alan Turing (1950),
ones own states of mind, any more than one echoing Hobbes, argued that intelligence
could work out that lightning is an elec- could be understood as computation. It
trical discharge or that water is H2O, merely would be possible in principle to build a
by reflecting on familiar properties of light- mind by programming a machine that
ning and water. Identities of this kind are dis- would process symbols so as to mimic an
coverable only after careful scientific study. intelligent human being.
When we investigate the brain, we discover Turing proposed a test for intelligence,
that it has the kind of administrative stand- the imitation game. Start with two people,
ing in the processing of incoming stimulation A and B, a man and a woman, commun-
and the production of behavior we associate icating via teletype with a third person, the
with the mind. The simplest explanation interrogator. The interrogator queries A
for this coincidence of roles is that the brain and B in an effort to determine which is the
is the mind, mental states are states of the woman. The woman must answer truth-
brain. fully, but the man can prevaricate. A wins
Plenty of scientists and non-philosophers the game when he convinces the interrog-
have thought this for a long time, why not ator that he is B. Now, imagine a cleverly
philosophers? Philosophers see the task of programmed digital computer replacing A.
reconciling mental and physical properties If the machine succeeds in fooling the inter-
as fraught with difficulty. The feel of a state rogator about as often as a person would,

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 41

t he mi nd / body probl em
we should, Turing contends, count it as traced to Aristotle, who, at times, seemed to
intelligent. be thinking along these lines (De Anima
Despite important advances in techno- Book II, 13). One difficulty for any such view
logy, events have not born out Turings is that it seems possible to imagine systems
optimistic prediction that machines would that preserve the same patterns of internal
pass his test by the turn of the century. Still, relations as minds, but are not minds. Ned
work in artificial intelligence (AI) has pro- Block (1978) imagines the population of
gressed on several, less adventurous fronts. China organized in the way an intelligent
Although attacks on AI (most famously by system might be organized. Although the
Hubert Dreyfus, 1972 and John Searle, Chinese nation is a functional duplicate of a
1980) have been inconclusive, philosophical conscious agent, it is hard to think that the
enthusiasm for the thesis that the nature nation, as opposed to the individuals who
of the mind can be captured by a computer make it up, constitutes a conscious mind.
program has waned. One question is whether The functionalist picture is one of higher-
consciousness might supply some needed level mental properties realized by, but dis-
spark, and this brings us back to the funda- tinct from lower-level physical realizers.
mental mindbody problem. The result is non-reductive physicalism:
The advent of the digital computer minds and their properties are grounded
encouraged philosophers to separate what in the physical world, but not reducible to
could be called hardware questions from their physical grounds. A similar picture
questions about software. Perhaps we has been inspired by Donald Davidsons
should view the mind, not as a physical anomalous monism. Davidson (1970)
machine, but more abstractly, as a pro- describes the mental as supervening on
gram running on a physical machine, the the physical. Davidson borrows the notion of
brain. What is important is not the minds supervenience from R.M. Hare, who had
physical implementation, but networks borrowed it from G.E. Moore. Both Hare
of internal relationships that mediate inputs and Moore were concerned with issues in
and outputs. So long as this pattern is ethics. Both, though for different reasons,
preserved, whatever the nature of the held that, although moral assertions could
underlying hardware, we have a mind. not be translated into non-moral, natural
This is one way of thinking about func- assertions, moral differences required non-
tionalism (Fodor, 1968). Functionalists moral differences. If St. Frances is good, an
note that we are comfortable ascribing states agent indistinguishable from St. Frances in
of mind to very different kinds of physical relevant non-moral respects a molecular
system. A human being, an octopus, and duplicate of St. Frances must be good
a Martian could all be said to feel pain, as well. Davidson applied this idea to the
although physical states that might be relation between mental and physical
thought to realize pain in each could be descriptions: agents alike physically (agents
very different. This thought led to the thesis answering to all the same physical descrip-
that states of mind are multiply realizable. tions) must be alike mentally (must answer
A property the pain property, for instance to the very same mental descriptions).
that has different physical realizers cannot Reduction fails in both ethics and psy-
be identified with any of those realizers. chology because agents could be alike
This sounds like old-fashioned dualism. But morally or mentally, yet differ physically.
realized properties are realized physically. Supervenience fits nicely with multiple
In this regard they are shaped by, and realization, so nicely that some philosophers
dependent on, physical goings-on. began to think of supervenience as provid-
Functionalists focus on structure. What ing an account of the realizing relation.
matters to a mind is not the medium in Considerable effort was expended on refining
which it is embodied (flesh and blood, silicon the supervenience concept. The result was a
and metal, ectoplasm), but its organization. proliferation of kinds and grades of super-
Thus construed, functionalism is sometimes venience and much discussion as to which

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 42

t h e m i nd / b o d y p r o b lem
best reflected the relation between mental and mental properties are distinct, higher-level
physical properties (Kim, 1990). Superveni- properties, how are they supposed to figure
ence, however, is a purely formal, modal in causal relations involving lower-level
notion. If you know that the As supervene physical goings-on? So long as we embrace
on the Bs (moral truths supervene on nat- closure, it appears that physical events
ural truths, mental truths supervene on bodily motions, for instance must have
physical truths), you know that the Bs in wholly physical causes. The prospect of
some fashion necessitate the As. But what mental properties making a causal dif-
is responsible for this necessitation? What ference in the physical world is evidently
is it about the Bs that necessitates the As? inconsistent with mental properties being
There are a number of possibilities: (1) irreducible to physical properties and the
the As are the Bs; (2) the As are made up physical worlds being causally closed. We
of the Bs; (3) the Bs include the As as parts; must choose, it seems, between epiphenom-
(4) the As are caused by the Bs; (5) the As enalism mental properties, although real, are
and the Bs have a common cause. None of physically impotent and systematic over-
these fit what proponents of supervenience determination some events have mental
or multiple realizability appear to have in causes as well as physically sufficient causes.
mind, however. Sydney Shoemaker (1980) Kim argues that over-determination is a false
has suggested that causal powers option. We thus face a choice between
bestowed by mental properties are a sub- epiphenomenalism, on the one hand, and, on
set of powers bestowed by a variety of the other hand, the abandonment of the
physical realizing properties. When one of non-reductivist hypothesis. Mental properties
these physical properties is on the scene, the are either reducible to physical properties
mental property is thereby on the scene, or epiphenomenal. Perhaps, Kim suggests,
option (3) above. Derk Pereboom (2002), most mental properties are reducible. Those
invoking the idea that a statue, although that are not, qualitative properties of con-
constituted by a particular lump of scious experiences, for instance, the qualia,
bronze, is not identical with the lump, must be epiphenomenal: real, but causally
argues that instances of mental properties are impotent.
wholly constituted by, but not identifiable This is close to the line advanced by David
with their physical realizers, option (2). Chalmers (1996) in a ringing defense of the
These accounts of the realization relation irreducible nature of qualia. Chalmers divides
locate mental properties within the physical mental attributes into those characterizable
causal nexus. It is hard to see, however, in information processing terms and those
how any such account could preserve the that are essentially conscious. The former
thought that mental properties are really logically supervene on fundamental physical
distinct from their realizers while mingling features of organisms: a system with the
their causal powers with powers of the right sort of functional organization will be
realizers. Powers comprising a subset of a intelligent and, in general, psychologically
things physical powers would seem to be explicable. consciousness, on the other
physical powers; and powers of a statue are hand, although determined by the physical
hard to distinguish from powers of the facts, is not reducible.
bronze that constitutes the statue. To facilitate the distinction he has in
Non-reductive physicalism has proved mind, Chalmers imagines zombies, creatures
popular because it promises to preserve the resembling us but altogether lacking in
distinctiveness and autonomy of the mental, conscious experiences (Kirk, 1974). Such
while anchoring it firmly in the physical creatures are impossible in our world, that
world. However, non-reductive physicalism is, given actual laws of nature. The con-
has come under fire from Jaegwon Kim ceivability of zombies, however, suggests
(2005) and others for failing adequately that laws governing the production of
to accommodate mental causation, the conscious qualities are fundamental in
centerpiece of the mindbody problem. If the sense that they are additions to laws

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 43

t he mi nd / body probl em
governing fundamental physical processes. Suppose, however, we distinguish proper-
Think of such laws as analogous to Euclidian ties of things experienced from properties of
axioms. Laws governing consciousness experiences. The sunset is red, the breeze
resemble the parallel postulate in being balmy, the sand warm, and the waves mur-
independent of the rest. Their presence or mur softly. Colors sounds, odors, and the
absence has no effect on physical goings-on. like are not properties of our experiences
Outwardly, a zombie world is indistinguish- of such things, but properties of things we
able from ours. experience, or at any rate properties we rep-
Both Kim and Chalmers render conscious resent such things as possessing. The point
qualities qualia epiphenomenal, perfectly was made by J.J.C. Smart (1959) in his
real, but physically irrelevant. The result original discussion of mindbrain identity,
is what Kim calls modest physicalism and, more recently, others have sought
physicalism plus a mental residue a to demystify qualia by arguing that what
conception reminiscent of Descartess idea have been regarded as irreducible qualities
that much human behavior is explicable on of conscious experiences are, in reality, only
mechanical principles alone. The difference qualities we represent things as having
is that, whereas Descartes embraced inter- (Harman, 1990 and Lycan, 1996). Were
actionism mental properties are causally that so, there would be no insurmountable
potent Kim and Chalmers regard con- gulf between mental properties, including
sciousness as qualitatively remarkable but properties of conscious experiences, and
causally inert. unexceptional physical properties. Much of
Other philosophers with physicalist lean- the mystery of consciousness might be due
ings are not so ready to throw in the towel. to confusion over what experiential proper-
What exactly are mental qualities, the so- ties could be (see experience).
called qualia? Describe a dramatic sensory Here we have representation playing the
scene: a sunset viewed from a tropical beach. garbage-bin role: embarrassing or incon-
Your description will invoke a panoply of venient features of the world are consigned
vivid qualities: colors, odors, sounds. Were to representations of the world. Still, it is
we to look inside your head, however, we difficult to shake the idea that representings
would observe none of this. Colin McGinn are themselves permeated with irreducibly
asks how Technicolor phenomenology mental qualities. Your being in pain might
could arise from grey soggy matter (1989, involve your representing a bodily state as
p. 349). As C.D. Broad reminds us, prop- painful, but this representing is, or certainly
erties of brains seem utterly different from seems to be, qualitatively loaded.
properties of our conscious experiences. What we might hope to learn from all
Let us suppose, for the sake of argument, this? The mindbody problem takes hold
that whenever it is true to say that I have only when we respect the integrity of both
a sensation of a red patch it is also true to the physical and the mental. More often
say that a molecular movement of a certain than not this has meant accommodating
specific kind is going on in a certain part of the mental to the physical, thereby privil-
my brain. There is one sense in which it is eging the physical. The ideal solution would
plainly nonsensical to attempt to reduce the involve finding a niche for the mental
one to the other. There is something which within the physical realm, but that seems
has the characteristic of being an aware- hopeless, no more promising than reduc-
ness of a red patch. There is something which tion or elimination. Perhaps we are deluding
has the characteristic of being a molecular ourselves. Perhaps we have erred in letting
movement. It would surely be obvious even Descartes set the agenda and assuming at the
to the most advanced thinker who ever outset that the mental and the physical
worked in a physiological laboratory that, are mutually exclusive. Suppose, instead,
whether these somethings are the same or it turned out that the mental/physical
different, there are two different character- distinction were not metaphysically deep.
istics (Broad, 1925, p. 622). In that case, we would have no mystery as

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 44

t h e m i nd / b o d y p r o b lem
to how mental (in the sense of non-physical) of being mental or physical. Spinozas
properties could have physical (in the sense attributes differ from Descartess, however, in
of non-mental) causes or effects. being attributes of a single substance and
Consider Davidsons anomalous mon- in being, at a deeper level, unified. In singl-
ism. Davidson is commonly read as ing out attributes, we are abstracting in
holding that mental properties depend on, Lockes sense, engaging in partial con-
but are not reducible to physical properties. sideration of a substance. Abstraction is
A mental event is an event with a mental a mental act, but what is abstracted is in
property; a physical event is an event with no way mind-dependent.
a physical property. This leaves open the These are deep metaphysical waters, but
possibility of token identity without type the mindbody problem cries out for a deep
identity: one and the same event could be solution. Perhaps it is time to abandon
both mental and physical by virtue of pos- the Cartesian presumption that the mental
sessing a mental property and a (distinct) and the physical differ in a fundamental
physical property. The problem of mental way, along with all the many attempts at
causation arises because we think that events reconciliation beholden to the Cartesian
have the effects they have solely in virtue of presumption. As noted, such attempts have
their physical properties. Mental properties tended to privilege the physical. The mental
piggyback on physical properties, but is seen as reducible to or dependent on the
appear causally inefficacious. physical in some way. For Davidson and
Although this picture is widely attributed Spinoza, the physical is in no regard privi-
to Davidson, it is pretty clearly not what leged. We have one world, variously pro-
Davidson has in mind. Davidson speaks of pertied, describable in various ways, with
descriptions and predicates, not properties. An various degrees of specificity. To imagine
event is mental, he holds, if it answers to that dramatic differences in our modes of
(satisfies) a mental description; it is phys- classification must reflect fundamental
ical if it satisfies a physical predicate. One metaphysical discontinuities is to mistake
and the same event, including the events features of our representations of the world
causally efficacious constituent properties, for features of the world.
could answer to both a mental and a phys- Or so Spinoza and Davidson think.
ical description. For Davidson, the mental Whether a move to monism represents
physical distinction is classificatory, not progress or merely one more philosophical
metaphysical. Everything in the world could byway leading nowhere remains to be seen.
be given a physical description and so counts Meanwhile, philosophers will continue to
as physical. Some portions of the world till familiar soil in familiar ways in hopes of
could also be described using mental terms. bringing forth some new solution to the
truthmakers for applications of mental mindbody problem.
predicates will be fully describable using
a physical vocabulary. This is so despite
bi bliography
the fact that, owing to very different appli-
cation conditions, there is no prospect Block, N.J.: Troubles with Functionalism,
of analyzing mental predicates in physical in C.W. Savage, ed., Perception and Cogni-
terms. tion: Issues in the Foundations of Psychology,
A view of this kind treats mental and Minnesota Studies in the Philosophy of
physical as classificatory designations, not Science 9 (Minneapolis: University of
fundamental metaphysical categories. In Minnesota Press, 1978), 261325.
this regard it resembles Spinozas neutral Broad, C.D.: The Mind and Its Place in Nature
monism. Spinoza (16321677) held that (Paterson, NJ: Littlefield Adams, 1960;
there is but a single substance possessing originally published 1925).
multiple attributes, including the mental Chalmers, D.: The Conscious Mind: In Search
and the physical. Finite physical or mental of a Fundamental Theory (New York:
entities are modes of these attributes, ways Oxford University Press, 1996).

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 45

modality and possibl e worl ds

Chomsky, N.: Cartesian Linguistics: A Chapter Place, U.T.: Is Consciousness a Brain Pro-
in the History of Rationalist Thought (New cess?, The British Journal of Psychology
York: Harper and Row, 1966). 47 (1956), 4450.
Churchland, P.S.: Eliminative Materialism Ryle, G.: The Concept of Mind (London:
and the Propositional Attitudes, Journal Hutchinson, 1949).
of Philosophy 78 (1981), 6790. Searle, J.R.: Minds, Brains, and Programs,
Dreyfus, H.L.: What Computers Cant Do: A Behavioral and Brain Sciences 3 (1980),
Critique of Artificial Reason (New York: 41724.
Harper and Row, 1972). Shoemaker, S.: Causality and Properties, in
Feyerabend, P.K. and Maxwell, G., ed.: Mind, Peter van Inwagen, ed., Time and Cause
Matter and Method: Essays in Philosophy (Dordrecht: Reidel, 1980), 109 35.
and Science in Honor of Herbert Feigl (Min- Skinner, B.F.: Behaviorism at Fifty, Science
neapolis: University of Minnesota Press, 140 (1963), 9518.
1966). Smart, J.J.C.: Sensations and Brain Pro-
Fodor, J.A.: Psychological Explanation: An cesses, Philosophical Review 68 (1959),
Introduction to the Philosophy of Psycho- 14156.
logy (New York: Random House, 1968). Stich, S.P.: Deconstructing the Mind (New
Foster, J.: The Case for Idealism (London: York: Oxford University Press, 1996).
Routledge and Kegan Paul, 1982). Turing, A.M.: Computing Machinery and
Grossmann, R.: The Categorial Structure of Intelligence, Mind 59 (1950), 434 60.
the World (Bloomington: Indiana Univer- Wittgenstein, L.: Philosophical Investiga-
sity Press, 1983). tions, trans. G.E.M. Anscombe (Oxford:
Harman, G.: The Intrinsic Quality of Experi- Basil Blackwell, 1968; originally published
ence, Philosophical Perspectives 4 (1990), 1953).
Kim, J.: Physicalism, or Something Near Enough john heil
(Princeton, NJ: Princeton University Press,
Kim, J.: Supervenience as a Philosophical Modality and Possible Worlds
Concept, Metaphilosophy 12 (1990), 1
27. Propositions are evaluated not only as true
Kirk, R.: Zombies versus Materialists, or false, but as necessarily or contingently
Proceedings of the Aristotelian Society, true or false. That seven plus five equals 12
suppl. vol. 48 (1974), 13552. is necessary; that George W. Bush was
Lowe, E.J.: Subjects of Experience (Cambridge: the President of the United States in 2008
Cambridge University Press, 1996). is contingently true, and that Saul Kripke
Lycan, W.G.: Consciousness and Experience has seven sons is merely possible. What
(Cambridge, MA: MIT Press, 1996.) sort of fact makes it true that these pro-
McGinn, C.: Can We Solve the MindBody positions have the modal status that they
Problem?, Mind 98 (1989), 34966. have? The problem is sometimes put in epis-
Matson, W.I.: Why Isnt the MindBody temological terms: empiricists, for example,
Problem Ancient?, in Feyerabend and ask how experience could give us reason
Maxwell (1966), 92102. to believe that a proposition is not just true,
Nagel, T.: What Is it Like To Be a Bat?, but necessary. But the real problem behind
Philosophical Review 83 (1974), 43550. this question is not epistemological, and
Nichols, S. and Grantham, T.: Adaptive not dependent on any thesis about the
Complexity and Phenomenal Conscious- sources of our knowledge. Even if an oracle
ness, Philosophy of Science 67 (2000), gave us unlimited access to matters of
64870. fact about the world, we would still face the
Pereboom, D.: Robust Nonreductive Materi- question, what could make it the case
alism, Journal of Philosophy 99 (2002), that some fact was not just true, but had
499531. to be true?

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 46

m od a l i ty and p o ssib le wo r ld s
According to one traditional response to Quine distinguished three grades of modal
this problem, modal propositions are made involvement (Quine, 1953). (He was skeptical
true by relations of ideas or linguistic con- even of the first, but saw them as increasingly
ventions: not by the way the world is, but by problematic.) The first grade was a necessity
the way we conceive or describe it. But (on predicate of sentences: being logically true,
this view) what is necessary is not that we or perhaps being analytic. The second grade
conceive or describe the world as we do. If it was a move from a predicate of sentences
is necessary that all uncles are male, it is to an operator on sentences from uncles
not because it is necessary that we should are male is necessary to necessarily, uncles
have adopted certain conventions to use are male. Quine argued that the move
the worlds uncle and male in certain involved a use-mention confusion, since
ways. What is said to be a matter of con- operators are to be interpreted in terms of
vention that a certain sentence be used to functions whose arguments are the values of
say something that is true no matter what expressions, and not the expressions them-
the facts are is different from what is said selves. To stipulate that a sentence of the form
to be necessary, which is the proposition necessarily p shall be true whenever the
itself that this sentence is convention- sentence that is in the place of p satisfies
ally used to express. So how can linguistic the necessity predicate constrains the inter-
conventions, or facts about the way we pretation of the operator, but does not
conceive of things, explain necessity and determine it. Quine argued that the move
contingency? In any case, it is hard to see how from the first to the second grade of modal
some statements widely thought to be nec- involvement, while based on a use-mention
essary could be true by convention. How mistake, was in itself relatively harmless,
could the way we talk or think make it true until one made the further move to the
and necessary that something (a number, third grade, which was to allow the opera-
or God, for example) should exist, or that a tor to operate on open as well as closed sen-
particular thing (Hillary Clinton, say) should tences that is, to allow quantification into
be a member of a particular kind (human modal contexts. The first move disguised
being)? the fact that modal contexts were really
The way we have put the problem is quotational, and so that quantification into
already contentious, since it assumes that the modal context was, implicitly, quantifica-
things that are said to be true or false, and tion into a quotation. One could repair the
necessary or contingent, are propositions damage and avoid incoherence, he argued,
(see proposition, state of affairs). An only by making metaphysical commitments
adequate theory of modality must give that he and the empiricist developers of
some account of what propositions are, or modal logic that he was criticizing would
of whatever the bearers of truth and neces- agree are unacceptable.
sity are taken to be. One way to begin that is It is true that modern modal logic began
motivated by the empiricists idea that neces- (with C.I. Lewis) as a project of analyzing
sity has its source in relations of ideas or logical necessity, and deducibility, so Quines
in the meanings of words is with a predicate, analysis is appropriate as an ad hominem
not of propositions, but of the sentences of argument against his intended targets
some language (see empiricism). Paradigms (C.I. Lewis and Rudolf Carnap). But modal
of necessary truths, according to this ap- concepts in general have much wider appli-
proach, are statements that are logical truths, cation. We may be concerned with what
or truths in virtue of meaning. The signi- must or might happen, in various senses, and
ficance of this alternative starting point can with what would or might have happened
be illustrated by looking at W.V. Quines under various conditions that did not, in
criticisms of modal logic which began with fact, obtain, with the dependence and inde-
the assumption that our most basic modal pendence of facts on other facts, and these
concepts are applied to linguistic expres- concerns arise in our attempts to under-
sions, rather than to what they express. stand and act on the empirical world, and not

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 47

modality and possibl e worl ds

just in logic and semantics. To understand agree that they have truth conditions, and
modal concepts more generally, it seems that their truth conditions are essential to
appropriate to begin with something like them. Any theory of propositions will say
facts, states of affairs, or propositions as the that the class of propositions determines a
things to which modal predicates are applied. structure that can be characterized by some
If we begin with a predicate of propositions, familiar interdefinable relations: entailment,
rather than sentences, then Quines three incompatibility, consistency, etc. If we start
grades of modal involvement look quite dif- with a notion of consistency or compatibil-
ferent. Suppose we assume, about proposi- ity, as a property of sets of propositions, we
tions, only that if we ascribe a well-defined can define the other relations that are
predicate to a determinate entity that is required in terms of it. We assume that con-
within the range of the predicate, we will sistency will satisfy the following property:
have expressed a proposition. Then the if a set of propositions is consistent, then
move from the first to the second grade of so is any subset of it. It will be assumed, in
modal involvement looks unproblematic: a minimal theory of propositions, that every
the proposition expressed by a sentence of the proposition has a contradictory, where the
form necessarily p will be as well defined notion of a contradictory is definable in
as the predicate of propositions with which terms of the consistency relation as follows:
we began. And the move to the third grade proposition x is a contradictory of proposition
to an operator on open sentences into y, if and only if, first, the set {x,y} is incon-
which one may quantify looks unprob- sistent, and second, any consistent set of
lematic as well, for the following reason: if propositions is either consistent with x, or
an operator on propositions is well-defined, consistent with y. A set of propositions
then so is a corresponding operator on entails a proposition x if and only if the set
propositional functions (functions from indi- {y} is inconsistent, where y is a con-
viduals to propositions). Suppose the neces- tradictory of x. Two propositions will be
sity operator,  is interpreted with a equivalent if and only if they are mutually
function that takes (for example) the proposi- entailing. A minimal theory might identify
tion that Socrates is human to the proposi- equivalent propositions. Even if finer dis-
tion that it is necessary that Socrates is tinctions between propositions are required
human. Suppose that the open sentence for some purposes, we can go some way
x is human expresses a function from toward a theory of modality while ignoring
individuals to propositions. Then it is nec- such distinctions.
essary that x is human will express the It is clear from the requirements of a min-
propositional function whose value, for any imal theory of propositions that the most
individual a, is the proposition that the pro- basic modal properties are not something
position which is the value of x is human added onto a minimal theory of proposi-
for argument a is necessary. tions, but are constitutive of it. Intuitively,
But even if the move through the grades the consistency of a set is the possibility that
of involvement is unproblematic, given the the members of the set all be true together,
assumption that what we start with is a and a necessary truth is a proposition that
predicate of propositions, a clear account of is entailed by every set. This is possibility in
modality still need an account of propositions the widest sense; more restrictive notions of
(about which Quine was famously skeptical). possibility and other modal properties and
There are many conceptions of proposition, relations might be defined with additions
and lots of controversies about how this no- to the basic structure.
tion is best understood, but fortunately we Necessity, according to a familiar slogan
can go some way toward an account of mod- going back at least to Leibniz, is truth in all
ality while making only minimal assumptions possible worlds, and the notion of a possible
about exactly what propositions are. world has played a prominent role in
Whatever propositions are, all who are contemporary treatments of modality, both
willing to talk at all about such things will in formal semantic models, and in the informal

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 48

m od a l i ty and p o ssib le wo r ld s
characterization of philosophical problems. claim about ontology, but an attempt to
(See Kripke, 1963 for an exposition of the get clearer about what such claims come to.
model theory, and Kripke, 1980 for an in- I think the thesis that necessity is truth in
fluential treatments of philosophical problems all possible worlds should be understood
in metaphysics and the philosophy of lan- in a similar spirit. The paraphrase of modal
guage that uses the possible worlds frame- claims and questions into the language of
work.) The notion is a controversial one, possible worlds solves some of the more
and there are substantive disagreements superficial puzzles about referential opacity
about how it should be understood, and and merely possible individuals by diagnos-
about whether any notion of possible world ing scope ambiguities and by separating
should play a central role in an account of questions about names and words from
modality. But at least a minimal concept of questions about the individuals, kinds and
possible world can be defined within the properties that the names and words are
basic minimal theory of proposition. Within used to designate. And it brings to the sur-
that theory, we can define maximal con- face and gives new form to the underlying
sistent classes of propositions: classes that are metaphysical questions about the nature
consistent, and that for every proposition of modality.
contain either that proposition or its con- In the context of this simple framework,
tradictory. One might identify a possible I will consider a number of interrelated
(state of the) world with these maximal sets metaphysical problems about modality.
of propositions. Or in an alternative formu- First, if possible worlds are to be taken as basic
lation, one might take a set of possible worlds entities in our ontology, what kind of thing
as the primitive basis of ones theory, and are they? What is it that makes it true that
define the propositions as sets of them. there are the possible worlds that there are?
Whichever primitive notion one begins with, Different philosophers who take possible
there will be, in a minimal theory, a one-one worlds to be fundamental to an explanation
correspondence between sets of possible of modality give radically different answers
worlds and coarse-grained propositions. to these questions. David Lewis argued
(See Adams, 1974 for an analysis of pos- that we should take other possible worlds
sible worlds in terms of propositions, and literally as concrete particular universes,
Stalnaker, 2003, ch. 1, for a discussion of the spatio-temporally disconnected from our
relation between propositions and possible own (Lewis, 1986). Most other philosophers
worlds.) who take possible worlds seriously explain
The point of spelling out this minimal them as possible states of the world, ways
theory of propositions, possible worlds and the world might be. (See Kripke, 1980;
basic modal properties and relations is to set Plantinga, 2003; Stalnaker, 2003 for
up a framework in which the substantive actualist accounts. See Divers, 2002 for a
metaphysical questions about modality can survey of a range of accounts of possible
be sharpened and clarified. We will consider worlds.) These contrasting answers take
questions about the nature of possible worlds on different explanatory burdens and give
and their role in a metaphysical account different response to various more specific
of modality, but it is useful first to see the problems about modality. Second, can we
minimal framework as an attempt to provides give an account of modality that is reductive
only a paraphrase of problematic modal in some sense, and if so, what is being
claims in a language in which ambiguit- reduced to what? David Lewis argued that his
ies and equivocations are more easily avoided, realist analysis of modality in terms of pos-
and in which the structure of modal claims sible worlds was a reduction of modal to
and questions are more perspicuously dis- non-modal notions, but others have dis-
played. The thesis that necessity is truth puted this. Alternatively, one might try to
in all possible worlds is like Quines thesis reduce the notion of a possible world to some-
that to be is to be the value of a bound vari- thing more basic. Is a reduction of modal
able. The Quinean thesis is not a substantive to non-modal notions something we should

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 49

modality and possibl e worl ds

seek, or take to be a benefit of a theory if it merely possible worlds, which seems to mean
succeeds, in its own terms, in giving one? that it is committed to the existence of
Third, might there have been things that things that do not exist.
do not in fact exist? If so, what can be said Any response to this challenge that seeks
about what is merely possible? What onto- to defend the coherence of the account
logical commitments are required to make must distinguish a sense in which merely
sense of the possibility of things that do not possible worlds exist from a sense in which
actually exist? The Lewisian modal realist has they do not, and there are two very different
no problem here (at least no new problem), strategies for making this distinction. The
but the actualist needs either to explain modal realist answers the question by
what we are really talking about when we distinguishing two different ranges for the
seem to be talking about things that might, quantifier one unrestricted and one
but dont exist, or else to reject the thesis restricted. When we talk about absolutely
that there might have been things other everything that exists, we include a plural-
than those there are. Fourth, whether or ity of possible worlds (as well as merely pos-
not there might have been things that do sible donkeys, people and things). But we
not actually exist, it seems obvious that the most often use the quantifiers so that they
things there are might have been different in range over a restricted domain: everyone
various ways from the ways they in fact might, for example, mean all the people
are. Does this imply that the same things exist invited to the party. Even when we are
in many possible worlds? Is there a problem making very general claims, we are often
about the identification of individuals across (according to the modal realist response to
possible worlds, and if so what is it? There are this challenge) restricting our quantifiers to
different theoretical accounts of the rela- things in our vicinity, broadly construed.
tions between the individuals that exist in dif- Our vicinity, on this construal, includes that
ferent possible worlds, and of the relations part of reality that is spatio-temporally con-
between particular things and the properties nected with us. In this broad but restricted
and relations that they exemplify. sense, there is only one possible universe
1. Modal realism vs. actualism. The that exists: the one we are in. But the other
basic contrast between possibilist, or modal universes, like the actual people who were
realist accounts of modality on the one not invited to the party, are equally real.
hand and actualist accounts on the other is For the actualist, the distinction is of a
central to many of the more specific issues different kind. According to this theory, the
in the metaphysics of modality. According only things that exists, in the most absolute
to the modal realist, there are literally many and unrestricted sense, are actual things.
universes, individuated by the spatial and The relevant distinction is not in the range
temporal relations between things in them. of the quantifier, but in the kind of thing
Two things count as worldmates denizens that one is talking about. Possible worlds,
of the same possible world if and only if properly construed as things that there are
they are spatio-temporal relations between many of, are more accurately labeled pos-
them. But for the actualist, everything that sible states of the world, and states are the
is real is actually real. Possible worlds are pos- kind of thing that may be instantiated or
sible ways that a world might have been. The exemplified. (For a given state, there may or
difference between the two kinds of theory may not be something that is in that state).
comes out in the contrasting answers that We might distinguish a notion of possible
they give to the following general challenge world meaning a thing that exemplifies
to the coherence of the idea of a merely a given possible state of the world from a
possible world: notion of possible world as the state itself
A merely possible world is a world that is something that is perhaps not exempli-
not actual, which is to say a world that does fied. Using possible world in the first sense,
not exist. But the possible worlds analysis of there is only one of them (within the domain
modality is committed to the existence of of absolutely everything), while using it in

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 50

m od a l i ty and p o ssib le wo r ld s
the second, there are (in this same domain) Possible states of the world, or ways a world
many, only one of which is exemplified. might be, are not representations of a world,
The modal realist doctrine can be separated but properties that a world might have.
into two theses, one metaphysical and one While properties allow for the distinction
semantic. The metaphysical thesis is that between existing and being exemplified that
there is a rich plurality of spatio-temporally the actualist needs to distinguish the sense
disconnected universes, rich enough to in which merely possible worlds exist from
obey certain principles of recombination. the sense in which they do not, properties
(Roughly, for any two things in different differ crucially from representations in the
universes, there will be a universe that following way: representations, whether
contains intrinsic duplicates of both.) The pictorial, linguistic, mental, or of some other
semantic thesis is that statements about form, face a problem of intentionality;
what is necessary and possible are properly it makes sense to ask, of a representation,
interpreted by quantifiers that range over what is it that explains why the representa-
these universes (A sentence of the form tion has the representational content that
Possibly P is true if and only if P is true in it has? There is no analogous question about
one of these universes.) Judged separately, properties. One cannot intelligibly ask, of a
both theses seem highly implausible. What property, what makes it that particular pro-
reason do we have to believe in this extra- perty, rather than some other one? I think
vagant ontology? And even if we did, what Lewiss critique trades on the fact that the
does it have to do with what is necessary actualists do not have an answer to a ques-
and possible in the actual world? But while tion like this about the things they are
Lewis granted the prima facie implausibility calling possible worlds.
of his doctrine (he took what he called the In the context of Lewiss overall meta-
incredulous stare to be the most serious physical picture, the thesis that possible
challenge to his metaphysical view), he states of the world are a kind of property
argued that the two parts of the doctrine does not provide an alternative to modal
must be judged together, and that together realism, since on Lewiss account, properties
they provide an indispensable foundation are classes, individuated by their extensions,
for a rich family of modal concepts. Despite and so a total way a world might be will be
its initial implausibility, the fruitfulness of a unit set, with the world that is that way
the doctrine that provides this foundation as its member (see class, collection, set).
is sufficient reason to believe that it is true. On this account of properties, there will be
Crucial to this defense of modal realism is many possible total states of the world only
the thesis that the rich family of modal con- if there are many things that are in those
cepts cannot rest on a more modest founda- states. It is an irony of Lewiss modal realism
tion. To use Lewiss rhetoric, the claim is that that the metaphysically extravagant doc-
we cannot have the paradise that this trine is grounded in Quinean ontological
family of concepts brings on the cheap. In austerity a rejection of any notion of prop-
this context, Lewis criticizes several ver- erty or attribute that cannot be identified
sions of the actualist alternative to modal real- with its extension. The actualist is com-
ism, arguing that none of them is up to the mitted to a more robust notion of property,
job. All of the actualist accounts that Lewis and so needs an explanation of what pro-
considers assume that possible worlds must perties are.
be representations of a world: either linguis- 2. Reduction. Possible worlds, construed
tic representations, something like scale as concrete universes, are the fundamental
models, or perhaps just simple and primitive primitive elements of the modal realist
representations. I think Lewis is right that a theory, and are clearly prior, in the order
notion of possible world as representation of explanation, to propositions, which are
cannot provide an adequate foundation identified with sets of possible worlds (see
for our modal concepts, but there are other concrete/abstract). Necessity and possib-
alternatives that Lewis does not consider. ility and the other modal notions are all

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 51

modality and possibl e worl ds

definable in terms of the properties of and rela- might happen does happen, somewhere
tions between propositions. Is Lewis right to and sometime, perhaps in a parallel uni-
claim that this theory provides a reductive verse, it would still be wrong to say that
account of modality an explanation of the this is what possibility consists in.
modal in terms of the non-modal? This is a I also agree with Lewis that no actualist
delicate question, since it is debatable what attempt to explain possible worlds in non-
basic concepts count as modal, but I think modal terms (for example, as linguistic
Lewiss claim is a reasonable one. The rea- representations) can succeed. But most
son is that the metaphysical component of versions of modal actualism are not
the theory (the hypothesis of a plurality of attempts to explain the modal in terms of the
parallel universes) is intelligible independ- non-modal, since the basic notions of this kind
ently of the semantic analysis of modal con- of theory whether they are propositions or
cepts in terms of it. The parallel universes are total ways a world might be are charac-
individuated by spatio-temporal relations, terized in terms that presuppose modal
and if it is fair to claim that the notion of a notions. In fact, I think the notion of a prop-
spatio-temporal relation is a non-modal erty, which is used to say what kind of thing
notion, than he theory seems to offer a a possible state of the world is, is itself a
metaphysical characterization of the struc- modal notion: one grasps what property
ture of reality in terms of concepts that one is talking about to the extent that one
modal skeptics should be willing to accept has a sense for what it would be for that prop-
(even if they reject the substantive meta- erty to be exemplified, which is to understand
physical claims made with those concepts). a certain possibility.
So I would concede Lewiss claim that he 3. Merely possible things. It seems at
offers a reduction, but maintain that it is least prima facie reasonable to believe that
debatable whether this is a cost or a benefit there might have existed things that do not
of the overall account. It is not just the in fact exist. For example, Saul Kripke
metaphysical commitments of the theory might have had seven sons, and if he had,
that elicit the incredulous stare; the semantic then seven people who do not in fact exist
analysis of modal notions in terms of it also would have existed (assuming that Saul
seems implausible, since it defines modal Kripke actually has no sons). In the possible
concepts in terms of things that, intuitively, worlds framework, the general thesis is
seem to have nothing to do with modality, modeled by the claim that the domains of
even if one were to accept the metaphysics. some possible worlds contain individuals
The intuitive resistance to the semantic that are not in the domain of the actual
component of the doctrine may derive world. The modal realist has no problem
from the judgment that modal notions are with this thesis, since the actual world is
fundamental, and not properly reduced to just one place among others. Non-actual
something more basic. Compare the way things are just things that are located in
one might react to a project of giving a one of the other places. But for the actual-
reductive analysis of truth to something ist, the domain of the actual world includes
more basic (warranted assertability, per- everything that exists at all, in any sense, so
haps, or what will be believed at the end it seems that actualism is at least prima
of inquiry). Even if such a story could be facie committed to the thesis that every-
spelled out in noncircular terms, one might thing that might exist does exist. This is the
judge that the analysis mistakenly categorizes most serious challenge to the actualist con-
a substantive claim as a definition. Even if it ception. I will describe four strategies that dif-
were correct that, at the end of inquiry all and ferent actualists use to respond to it:
only truths would be believed, this would not The first actualist response begins by
give us an account of what truth is (see noting that to understand talk of possible
theories of truth). Similarly, I think it is individuals, we need a distinction that par-
reasonable to think that even if a principle allels the distinction between two senses of
of plenitude were true, so that everything that the term possible world: just as actualists

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 52

m od a l i ty and p o ssib le wo r ld s
must distinguish a way a world might be from he had seven sons properties whose refer-
a world that is that way, so they must dis- ence could be fixed, in such a world, with a
tinguish individuating properties that an predicate of the form being identical to this
individual might have from the individuals individual, where this individual refers
that has those properties. While actualists are to one of the seven sons. Plantinga grants that
committed to the thesis that there are no we may not have the resources, even in
things that might exist, but do not, they can principle, to refer to particular uninstantiated
allow that there are (and necessarily are) haecceities, but we can talk about them in
properties that are necessary and sufficient to general terms, and that, he argues, is good
determine a unique individual, but that are enough.
in fact uninstantiated. More precisely, the The second response (defended in
view is that there are properties X that meet Williamson, 2002 and in Linsky and
the following condition: it is necessary that Zalta, 1994) is to reject the intuition that
if there exists something that instantiates X, gives rise to the problem that there might
then that thing is necessarily identical to have existed things that do not in fact
anything that instantiates X. The domains exist (as well as the intuition that there are
of the different possible worlds are to be some things that exist only contingently).
understood, according to this response, not This response avoids the problem, and as
literally as domains of individuals, but as its defenders emphasize, it also allows for a
domains of properties of this kind individ- modal logic that is much simpler than what
ual essences, or haecceities (see haeccity). is required when the domains vary from
Alvin Plantinga, who develops and defends possible world to possible world. But of
this response, calls the domains essential course it takes on the burden of explaining
domains (Plantinga, 2003). The basic struc- the divergence between the theory and con-
ture of the orthodox Kripke semantics for flicting intuitions about modal truths that
quantified modal logic, with variable do- seem compelling. How can it be made plaus-
mains, is unchanged by this move; the dif- ible that apparently temporary and contin-
ference is in the interpretation of the formal gent beings such as ourselves exist eternally
models. This response to the problem is and necessarily? How can we accept that
simple, formally conservative, and successful, there actually are things that might have
on its own terms, in reconciling actualism been Saul Kripkes seven sons? The defenders
with intuitions about what might have of this strategy respond to the challenge
been true. But it requires what some regard by acknowledging that people and ordin-
as a metaphysical extravagance: a belief in ary physical objects are only temporarily
a special kind of property that carries with and contingently concrete things, with a
it the particularity of an individual, but that spatio-temporal location. In possible worlds
is also conceptually separable from the indi- and at times when one is inclined to say
vidual. We may have no problem under- that the people and things do not exist, we
standing the property of being identical to should instead say that they exist, but lack
Socrates, but one might reasonably think the features that we are inclined to think
that this is an object-dependent property are essential to being a person or a physical
a property that would exist only if Socrates object. They are in no place, at those worlds
did. But the haecceitist response to the and times, and are neither concrete things,
problem holds that while we use the person nor abstract objects, but particular things that
Socrates to fix the reference of the property have the potentiality to be concrete things.
of being identical to Socrates (or a property This may seem a gratuitously extravagant
that is necessarily equivalent to it), the metaphysics, but Williamson argues that
property itself would exist even if he did not. it is entailed by principles that it is difficult
And furthermore, there actually exist, on to reject. The most controversial of the
this account, properties of this kind that premises of Williamsons argument is the
would be instantiated by Saul Kripkes thesis that singular propositions (and iden-
seven sons in the possible worlds in which tity properties, such as being identical to

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 53

modality and possibl e worl ds

Socrates) depend for their existence on that they entail every (actual) proposition or
the things they are about. As we have seen, its contradictory may fail to be maximal in
Plantinga rejects this thesis (which he another sense: they may entail existential
labels existentialism), but I suggested that propositions without entailing any singular
this is a serious cost of his account. But as propositions that witness the existential
Williamson shows, if we accept it, and also claim. That is, this response claims that
accept that a proposition is true only if there may be cases where an existential
it exists, then we must conclude that the proposition (Such as the proposition that
singular proposition that Socrates does not Saul Kripke had a seventh son) is possibly
exist could not be true, and this seems to true even though there is no singular pro-
imply that Socrates must exist. position (no proposition of the form x is
One might try to avoid the uncomfort- Saul Kripkes seventh son) that is possibly
able choice between Plantingas haecceities true. I think this is right, but making sense
and Williamsons objects of pure potential- of it requires a more radical reinterpretation
ity by rejecting a presupposition that both of the standard semantic models than do
positions apparently share. The need either the theories of Plantinga or of Williamson,
for primitive individual essences or for object Zalta and Linsky. And this response must
dependent propositions would be avoided accept the consequence that there are pro-
if we were able to reduce individuals to their positions (Such as the proposition that
properties. So a third actualist response to our Socrates never existed) which are true with
problem is to adopt some kind of bundle respect to some possible worlds in which the
theory of individuals. If this kind of account proposition itself does not exist. (Different
of individuals were defensible, then we could versions of the fourth response have been
characterize possibilia in terms of ordinary defended in Fine (2005), and Adams (1981).
universal properties and relations, rather 4. Modal properties. Even if we ignore
than in terms of primitive haecceities, and no merely possible individuals, there are prob-
propositions would be dependent on particu- lems with attributions of modal properties
lar individuals. But this kind of metaphys- to actual individuals. De re modal claims,
ical doctrine has a problem accounting for claims about what could or could not have
the potentialities and counterfactual pro- been true of some particular things seem
perties of particular individuals; we will say especially problematic since it is not clear
more about this problem below. how they could be true by convention, or by
There is a fourth response that accepts virtue of the relation of ideas. The possible
the irreducibility of individuals to their worlds picture seems to offer a straight-
properties and relations, and the object- forward paraphrase of such claims: to say that
dependence of singular propositions. It takes David Lewis might have been a plumber,
at face value the intuition that there might but could not have been a fried egg, is to
have been things other than those there say that there is a possible world in which
are, and it avoids a commitment to indi- David Lewis was a plumber, but no possible
vidual essences. I think this is the best world in which he was a fried egg. But is the
response to the problem, though it has its own plumber in the other possible world really the
counterintuitive consequences. The prob- same person as our own David Lewis? What
lems for this strategy come from an imme- is it about him explains his metaphysical
diate consequence of the combination of the incapacity to be a fried egg? Modal realists
object-dependence of singular propositions and actualists answer these questions in
with the contingent existence of individuals: different ways.
that some propositions themselves are things If possible worlds are ways the world might
that exist only contingently. If possible worlds have been then there is no implausibility
are identified with maximal propositions, or in accepting the straightforward assumption
maximal sets of propositions, then possible that Lewis himself inhabits other worlds,
worlds themselves will be contingent objects. since this is only to say that among the
Propositions that are maximal in the sense ways the world might have been but was not

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 54

m od a l i ty and p o ssib le wo r ld s
are ways that David Lewis might have been. bundles of qualities, and reduce individual
Kripke (1980) attempts to demystify coun- essences to such bundles and counterpart
terfactual suppositions about particular relations between them. An actualist might
individuals, arguing that nothing prevents also adopt a counterpart framework, with a
us from simply stipulating, in specifying primitive counterpart relation, for meth-
the counterfactual situation we are talking odological reasons: the aim would be simply
about, that it is a situation in which David to provide a framework that is neutral on
Lewis is a plumber. But Kripke acknowledged controversial theses about essential and
that we might also specify a counterfactual accidental properties, a framework in which
situation in a way that does not explicitly puzzles about identity across times and
identify a particular individual in terms worlds can be formulated in a perspicuous
of the qualitative characteristics, origin, or way. (Actualist counterpart theory is dis-
constitutive parts of the individual, and that cussed in some papers in Stalnaker, 2003.)
in such a case, we might then ask whether Whether one is an actualist or a modal
the individual we have specified is or might realist, and however one explains the appar-
be some particular actual individual. It ent possibility of things that do not in fact
remains puzzling exactly what determines exist, and the relation between particular
the answers to such questions. individuals and the properties and relations
If possible worlds are understood as other that they exemplify, and might exemplify,
places, as the modal realist understands them, there will remain a general puzzle about
then it is no longer plausible to think that the nature and source of modal truth. If
the inhabitants of the actual world will also necessity is true in all possible worlds, what
be found in other possible worlds. The explains why there are just the possible
Lewisian modal realist explains modal worlds that there are? Both actualists and
properties of individuals their capacities modal realists resist the idea that we can
and dispositions, essential and accidental explain modal facts as conventional or
characteristics in terms of the existence, semantic facts: conventions may determine
in other possible worlds, of counterparts of that our words express certain propositions,
the individual individuals in other possible but the propositions themselves are necessary
worlds who are similar, in relevant respects, or contingent, independently of the words
to the given individual. According to coun- that are used to express them. But actualists
terpart theory, David Lewis himself existed and modal realists also agree that to express
only in the actual world, but he might have substantive propositions is to distinguish
been a plumber in virtue of the fact that between the possibilities to locate the actual
there is a possible worlds in which a person world in the space of all possible worlds and
who is like him in certain specific respects this seems to imply that it is not possible to
was a plumber. give a substantive characterization of what
Actualists may also use counterpart theory, is common to all possible worlds. We wont
but for them there is no conflict between the have a clear grasp of the concept of meta-
counterpart analysis and the thesis that the physical possibility until we see a way to
plumber in the other possible world really resolve this tension.
is our own David Lewis. An actualist coun-
terpart theorist may say, as Alvin Plantinga
bibl i ography
does, that the domains of other possible
worlds should be thought of, not as sets of Adams, R.: Actualism and Thisness, Syn-
individuals, but as some kind of property these 49 (1981), 3 41.
that would have been instantiated by an Adams, R.: Theories of Actuality, Nos 8
individual, were the state of the world to have (1974), 21131.
been realized. For the haecceitist, the relevant Divers, J.: Possible Worlds (London and
properties are individual essences, but an New York: Routledge, 2002).
anti-haecceitist actualist might take the Fine, K.: Modality and Tense: Philosophical Papers
relevant individuating properties to be (Oxford: Oxford University Press, 2005).

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 55

persist enc e
Kripke, S.: Naming and Necessity (Cambridge, persistence) yet things persist even when
MA: Harvard University Press, 1980). undergoing no macroscopic change such as
Kripke, S.: Semantical Considerations on that of color. What is it for a material object
Modal Logic, Acta Philosophial Fennica to persist pure and simple? Is this a miscon-
15 (1963), 8394. ceived question because persistence is too
Lewis, D.: On the Plurality of Worlds (Oxford: basic a phenomenon to yield to analysis?
Blackwell, 1986). Or can the metaphysician say something
Linsky, B. and Zalta, E.: In Defense of the informative about what it involves? Is there
Simplest Quantified Modal Logic, Philo- something inherently strange, or even para-
sophical Perspectives 8: Logic and Language, doxical, about the concept of persistence
ed. J. Tomberlin (Atascadero, CA: such that we ought to deny that anything
Ridgeview, 1994), 43158. really persists? Or can we retain the idea of
Plantinga, A.: Essays in the Metaphysics of persistence and instead deny that anything
Modality (Oxford: Oxford University Press, changes since it is change, rather than per-
2003). sistence itself, that raises insoluble problems?
Stalnaker, R.: Ways a World Might Be:
Metaphysical and Anti-metaphysical Essays Spatio-temporal continuity
(Oxford: Oxford University Press, 2003).
Williamson, T.: Necessary Existents, in The standard approach to analyzing persist-
A. OHear, Logic, Thought, and Language ence is in terms of spatio-temporal continuity
(Cambridge: Cambridge University Press, (Coburn, 1971; Swinburne, 1968/1981).
2002), 23351. The idea is that an object F persists through
a temporal interval if and only if it traces a
robert stalnaker
spatio-temporally continuous path through
that interval. Tracing a spatio-temporally
continuous path is then defined in terms
Persistence of overlap between every pair of adjacent
spatio-temporal regions enclosing F during
the interval. The approach is intuitively
Things change. This much looks like a plausible inasmuch as we do tend to think
metaphysical and observational datum. By of persistence in terms of some kind of
the proposition that things change we typic- continuity, or perhaps continuous history,
ally mean that things survive change not involving the persisting object. We tend to
all changes, but most. In other words, we live associate diachronic distinctness (distinct-
in a world in which there is both change and ness over time), not just synchronic distinct-
sameness. My car was red; I have given it a ness (distinctness at a time), with breaks
coat of green paint; now it is green. One of in continuity, for example between my car
the cars qualities has changed, and to that and my house: there is no single continuous
extent the car itself has changed. But we path traced by both of them; their spatio-
would all accept that it is still the same car. temporal histories are discontinuous.
The standard way of putting this philosoph- It turns out, however, that it is far more
ically though not a way we often describe difficult to spell out an adequate continuity
it is to say that the car has persisted criterion of identity over time than it seems,
through a change, in this case of color. since it has to rule out obvious counter-
Yet it is not just change that compels examples. For instance, take a single-celled
metaphysicians to wonder about persistence. organism such as an amoeba, which repro-
It may be that, as Aristotle held, time duces by binary fission. It produces daughter
is the measure of change, and so without organisms neither of which are, it seems,
change there could be no time and hence identical to the original; yet one can trace
no persistence since persistence occurs a continuous path between the pre-fission
in time (extra-temporal existence, such as amoeba and each of its descendants. Or
Gods, would not on this view be a kind of consider a marine flatworm, cut into two

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 56

p e r s i s t ence
segments that then grow into new worms. a single, persisting sphere segment. Such
Perhaps the continuity in these cases is too cases are easily multiplied and provide a
weak, since we could specify a strong form formidable challenge to continuity theories
of continuity such that the difference in of identity. Perhaps the causal relations and
overlap between any adjacent regions was counterfactual dependence between the
indefinitely small, which does not seem to segments is relevantly different from those
obtain in an instantaneous division of the between the single segment at one time in its
sort just mentioned. But strong continuity history and at another, but spelling this out
also has counterexamples both to necessity is no easy matter.
and sufficiency for diachronic identity. As to Ought we to take a different approach
necessity, consider an instantaneous loss or to analyzing persistence? We might take a
gain of parts: a tree has a branch lopped off, cue from Butlers famous criticism of the
yet it still persists, though the continuity is memory criterion of personal identity over
only weak. As to sufficiency, consider the time (Butler, 1975, p. 100), that conscious-
infinite series beginning with a tree, all of the ness of personal identity presupposes, and
other members of which are decreasingly therefore cannot constitute, personal iden-
smaller parts of the tree the lump of plant tity. We might argue (Oderberg, 1993;
matter minus a millimetre of wood, the lump Merricks, 1998) that spatio-temporal con-
minus two millimetres, and so on, where the tinuity gives us evidence of persistence but does
series fully constituted is a real continuum not constitute it. For continuity always pre-
of spatio-temporal parts measured along supposes identity, inasmuch as the objects
some dimension, such as length or width. related by continuity (the car at t1, the car
We can specify a strongly continuous at t2) are themselves persisting objects so
path, but we do not want to say that the how can continuity be used to analyze
tree, although strongly continuous with its persistence if persistence is always part of
parts, is identical to any of them. what is described in describing a case of
The obvious move, at least to counter continuity? If continuity can only have evi-
wholepart tracing confusions, is to place dential force being a symptom of persistence
some sort of sortal restriction on what must but not a criterion, as it is sometimes put
be in the path: since parts of trees are not then the evidence will be defeasible, and
trees, a series such as that just given would in some cases easily so. Where it is absent,
not constitute a single persisting object. But moreover, we may still have good grounds
there are other examples that make difficulty for believing identity to obtain: imagine
even here (Shoemaker, 1979; Forbes, the radical disassembly and reassembly of
1985, pp. 1529; the examples go back in an object, or its vanishing and reappearing.
some respects to Kripkes unpublished (Is the latter a metaphysical impossibility?
lectures on identity over time from 1978). It is certainly conceivable.) Absent any other
Consider a homogeneous rotating sphere. viable analyses of persistence, we might
Take one of its segments, i.e., one of its take it to be a brute fact, an unanalyzable
physical parts that moves with the sphere. phenomenon. We usually know it when
This is clearly a single, persisting object, we see it, though we do make mistakes of
namely a sphere segment. Now imagine a reidentification.
light that constantly illuminates one single
region with the same surface area as the Temporal parts
segment. The segment passes through the
illuminated region one time for every com- A defender of continuity, however, will not
plete rotation. But during the period of one be content with the circularity objection to
rotation, an infinite series of distinct sphere the proposed analysis. We do not have to
segments also pass through that single, think of the relata of the continuity relation
illuminated region. They occupy a strongly as themselves persisting objects: they are,
continuous path, and they all fall under the rather, temporal parts of persisting objects,
same sortal sphere segment, yet they are not terminating ultimately in instantaneous

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 57

persist enc e
temporal parts. Each continuous path is it tracks our best intuitive, pre-theoretical
occupied by temporal parts, and each tem- judgments about what persisting things
poral part is itself analyzable in terms of there are; in other words, temporal part
parts of shorter duration, also on a contin- theory should leave our reidentification
uous path. The circularity is only apparent, practices undisturbed. Ordinary persistence
since the termination of the analysis lies in aside, moreover, if we think that a certain
parts that do not themselves persist we object might, say, vanish and reappear after
might think of these as points in space and an interval, we could count the series of
time, occupied by certain qualities. its temporal parts before disappearance and
Temporal part theory, often known as after reappearance as constituting a single
four-dimensionalism or (misusing an old persistent. This view of persistence could be
word) the theory of perdurance, has many supplemented, or to some extent modified,
defenders. (See, for a small sample: Quine, by a mixture of ontology and evolutionary
1950; Lewis, 1986; Forbes, 1987; Heller, theory along these lines: every materially
1990; Armstrong, 1997.) The general occupied portion of spacetime, no matter
idea is that just as persisting objects have how heterogeneous, constitutes an object.
spatial parts (the wheels on my car, the We humans, for survival purposes, gerry-
branches of the tree) so they also have mander certain portions of spacetime and
temporal parts (that part or stage of my car call these particular, privileged worms
from t1 to t2, the part or stage of the tree from the persisting objects. (See, for example,
t3 to t4). The temporal parts are usually des- Quine, 1981.)
ignated by a kind of hyphenated singular The debate about the existence of tempo-
term: my car-from-Monday-to-Wednesday, ral parts and their putative explanatory role
the tree-from-Thursday-to-Saturday, and as regards persistence continues unabated.
so on for any persistent and for any times The intuitive appeal of four-dimensionalism
however specified. has captured the imagination of many
The idea has intuitive appeal, since we metaphysicians, but it has to face some
know that objects have spatial parts, and serious objections. (For some of the critics,
space and time are in many ways similar. see Geach, 1972; Chisholm, 1976: Appen-
Moreover, it seems that contemporary space dix A; Thomson, 1983; Oderberg, 1993;
time physics, with the theory of relativity at Lowe, 1999, pp. 11418.) For instance,
its core, is at least congenial to temporal the thought that just because an object has
parts if not committed to them. There is, it spatial parts so it must have temporal parts
might be claimed, no space and no time is specious. For in order to generate a suffici-
only spacetime. Spacetime is an onto- ently convincing analogy between space and
logical unity, with objects spread out across time to motivate the thought, it turns out that
both the three spatial and one temporal one has to presuppose the existence of tem-
dimensions, all of which are features of a poral parts in the first place (as can be seen
single block, with objects being (on the in Taylor, 1955; discussed in Oderberg,
favorite metaphor) something like worms 1993, pp. 97103; see also Meiland, 1966).
stretched out across the block, divisible Second, is it true that spacetime physics
into segments. These segments, speaking commits us to temporal parts? To say that
accurately, are supposed to be spatio-temporal Minkowskian spacetime geometry has
parts: there are no purely spatial parts, and shown, as Minkowski himself thought, that
no purely temporal parts, but these spatio- space and time are mere shadows of an
temporal parts are what are called temporal underlying unified spacetime (Minkowski,
parts on the four-dimensionalist way of 1952, p. 75) could be seen as a metaphysical
looking at the universe. step too far since the spatial and temporal
How can the four-dimensionalist get per- dimensions are given differing mathema-
sistence from temporal parts? He might tical treatments in relativity theory. One
simply say that a series of temporal parts should, in addition, be careful about
constitutes a persisting thing if and only if drawing metaphysical conclusions from

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 58

p e r s i s t ence
physicists use of terms such as world-line, instance, that part of its history involving
spacetime worm, and the like, since its being parked by the kerbside or its being
where these terms appear in their work per- owned by me; we include its being green,
sistence is nearly always presupposed as a curved on top, and having five windows.
more fundamental concept rather than Now this intrinsic history is a genuine part
explained or analyzed in those terms. It is in of the cars total history. If you wanted
fact very difficult to motivate a metaphysic to, and you knew my car well enough, you
of temporal parts from spacetime physics could write a rather boring narrative of
(Rea, 1998). its history from dawn on Monday until dusk
It might further be argued, independently on Tuesday. Now, says the skeptic, what is
of considerations from the physics of space to distinguish this history of the car from
time, that the very concept of a temporal part Monday to Tuesday from the supposed car-
of a persisting object is of dubious coher- from-Monday-to-Tuesday? Could God, let
ence. To be sure, the temporal part skeptic alone we, tell them apart? But, comes the
does not deny that some things have tem- reply, the temporal part of the car is a phys-
poral parts: events paradigmatically have ical object, whereas the temporal part of its
them (the first half of the battle, the last five history is not. It is the physical temporal
minutes of the opera), as do processes (the part that makes the history true. If the cars
first hour of a compounds dissolution in history involves its being green on Monday
water) and histories (the medieval history of and receiving a coat of red paint on Tuesday,
Portugal; the first half of my life). What the the Monday-Tuesday temporal part will have
skeptic denies is that persisting objects have a green sub-part existing on Monday and
such parts, and while events and processes a red sub-part existing on Tuesday. Yet the
involve objects that persist, they themselves skeptic will insist that there just is no onto-
do not persist. So what sense can be made logical room for such objects. What makes
of the very idea that a persisting object the cars history what it is from Monday to
could have temporal parts? The hyphenated Tuesday is just the car itself and what is
singular terms mentioned above are a philo- true or false of it: it is green on Monday,
sophers invention; not all such inventions red on Tuesday. This is what makes it the
are bad, of course, but we should not infer case that it has a history with the following
from their existence that what they purport temporal parts the Monday history, in
to refer to exists as well. For how could which it is green, and the Tuesday history,
we how even could God distinguish in which it is red (and, of course, the tem-
between a putative temporal part of an poral part overlapping these in which it is
object and a temporal part of that objects his- changed from red to green). What room is
tory (or career, as it is sometimes called) there for temporal parts of the car itself ?
with exactly the same temporal boundaries? Among various other objections, a couple
The critic needs to be more precise, more are worth raising. Remember that to
though. The history of my car from Monday avoid circularity in the analysis of persis-
to Tuesday involves more than just the car tence, temporal parts will have to terminate
itself: there are all of its relations to other in instantaneous entities of which all the rest
objects that need to be included in that (those with duration) are composed. Yet what
history. The putative temporal part of the car sense can be made of an instantaneous
itself from Monday to Tuesday, however, temporal part? Calling it a spacetime point
is supposed to involve only what is within (with or without qualities associated with
the cars spatial boundaries. The critic, how- or true of it) does not clarify matters. If
ever, can reply as follows. Call that part (not instantaneous stages are really that dura-
temporal, not exactly spatial lets think of tionless then how can they constitute an
it as quasi-spatial) of my cars Monday- object with duration any more than dimen-
through-Tuesday history that involves only sionless points can constitute a region with
the car itself and its intrinsic features its dimension? Wasnt Zeno right all along? One
intrinsic history. Hence we factor out, for reply is to appeal to the Aristotelian notion

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 59

persist enc e
of potential infinity: the instantaneous stages interpretations, even if persistence or some-
are no more than limits of a process of thing approximating it is maintained, can the
potential division, but it is not as though equally basic phenomenon of change can be
such things have any actuality. The notion accounted for? We will return to this shortly.
might be a good one, but can the four- First, let us briefly consider a couple of other
dimensionalist appeal to it given that accounts that can be given of persistence.
he has non-circularly to analyze persist-
ence in terms of stages? No one who Stage theory
accepts the distinction between actual and
potential infinity would want to analyze a A view quite similar to standard four-
line in terms of dimensionless points. But the dimensionalism/temporal parts theory/
friend of temporal parts needs just such an perdurance is sometimes called stage theory
analysis if he is to avoid being left with or exdurance (Haslanger, 2003). Accord-
an unanalyzed remainder of persisting, i.e., ing to this theory (Sider, 2001; Hawley,
non-instantaneous, temporal parts. 2001), there are indeed temporal parts
Another point concerns whether four- of things other than events, processes, and
dimensionalism denies the phenomenon it histories, and there are space-time worms
seeks to explain. If persistents are just sums consisting of series of such parts. The basic
of stages, does anything really persist in four-dimensional framework is accepted.
the first place? Rather than genuine persist- What the stage theorist denies, though, is that
ence, doesnt the friend of temporal parts any of these worms are identical to what we
offer us no more than a series of creations identify as ordinary persisting things. When
and annihilations, with new matter literally we talk about persistents we are not talking
springing into existence ex nihilo all the about worms but about the temporal parts
time (Thomson, 1983)? The temporal parts themselves. What we think of as persistents
theorist might bite the bullet here, taking his are no more than stages.
account to be eliminative; though he would Of the various motivations for this
be committed to implausible claims about approach, an important one is to avoid
creation and annihilation. Or he might say what is seen as a problem about spatio-
that this interpretation is true only if he is a temporal coincidence. So, assuming (perhaps
presentist about time, according to which rashly) that personal fission is possible via a
only the present moment and what hap- split-brain operation and transplant, sup-
pens in it are real. More congenial to his pose that I undergo this procedure and two
position, though, is eternalism, according to people each get half of my brain. When they
which all moments of time and what happens awake, each is psychologically continuous
in them are equally real. Matter, on the with me. Call the new persons Bill and Ben.
latter view, does not keep vanishing and Bill is to be tortured, Ben is to live in pleas-
springing into existence; rather, the sum of ure. Should I be worried about what will
stages making up a persistent is, as it were, happen to me after the operation, or not
given all at once not simultaneously, concerned, or both, or neither? On the
but with equal reality. There is no temporal standard four-dimensionalist model, the
becoming: the space-time worm just exists most likely interpretation of events is that two
with its spatio-temporal dimensions. Wher- worms exist before, during, and after the
ever one of the segments is in space-time, so operation the one including the temporal
is the persisting object present, just as my car parts of Bill post-fission and me pre-fission,
is present wherever one of its spatial parts is. these being connected by psychological
The skeptic still worries that an eternalist view continuity; and the one including the con-
also denies persistence: there is no persistence tinuous stages of me and Ben. But this
where there is just creation and annihilation, means that pre-fission, there are two space
but equally no persistence where the object time worms overlapping the me-Bill worm,
is viewed simply as a block in spacetime. and the me-Ben worm. But if persistents are
Moreover, on both presentist and eternalist worms, and persistents include persons, then

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 60

p e r s i s t ence
it looks like there are two persons overlap- theorist allows is that at t1 there is Bill, who
ping coinciding in space and time before was me, and Ben, who was me, and these
the fission occurs. How, though, can two relations are explained in terms of the duplic-
persons be in the very same place at the able relation of continuity. It is hard to see
same time? And what thoughts am I likely how genuine identity and hence persistence
to have a single ambiguous thought in anything like a recognizable form gets
about future pain and future pleasure for into the picture. If, before fission, I really am
me, or two thoughts? Wouldnt the pro- a certain stage, then how, without violating
noun I be ambiguous pre-fission? Could two the necessity of identity, can it be the case
persons share a single thought, ambiguous that there is any stage after fission that is
or not? Surely I wouldnt notice any ambi- me if the pre-fission stage that I am now no
guity when thinking about my fate. longer exists then? So isnt it the case, on the
This interpretation of fission has too stage view (assuming necessity of identity)
many problems, according to the stage the- that I simply cease to exist at fission? In
orist. What we need to say is that there is a which case I will not be Bill and I will not
single person pre-fission, and that person is be Ben, in any recognizable sense of those
a stage, and that stage will be both contin- propositions.
uous with Bill-stages and continuous with The stage theorist will accept that what he
Ben-stages. In this sense it is true to say, for posits is not genuine identity; as one such the-
the stage theorist, that I will be Bill and I will orist puts it, claims of identity between
be Ben, but there is only one of me prior to things at different times make sense, even
fission. To say that I will be Bill (and will be though they are false (Hawley, 2001,
Ben) is akin to what a counterpart theorist p. 156). Yet the stage theorist wants to pre-
such as Lewis says about modal statements. serve the commonsense belief that persons
The counterpart theorist interprets a state- and other persistents do exist. So on the
ment such as I could have been smarter as theory, persons (for example) do exist but no
meaning that in some possible world there identity statements about them are literally
is a counterpart of me who is smarter than true! Well, it takes time to have a thought
I am (in the actual world). Similarly, for the even the simplest and most fleeting of
stage theorist to say that I will be Bill is to thoughts but it cannot then be literally true
mean, properly interpreted, that the stage that it is I who has any thoughts. For per-
that I am is continuous with some future sons are stages, and the only real stages are
Bill-stage. Since the relation is duplicable, instantaneous ones. One can call a three-hour
it can hold simultaneously of the stage that stage of me a stage, but it only has the title
I am and both future Bill-stages and future honorifically, or we might say derivatively.
Ben-stages. It literally has no thoughts, rather it is made
To critics, the stage view fares little better up of instantaneous stages with mental prop-
than standard four-dimensionalism. Pre- erties suitably related in some unspecified
fission (at t), I can truly say that after fission (and arguably unspecifiable) way. But what
(at t1) I will be Bill and that after fission I possible mental properties could an instan-
will be Ben (but not, on a necessary revision taneous stage have, if even the briefest of
of standard reasoning about tense, that thoughts takes time? Presumably, moreover,
I will at t1 be Bill and Ben: see Sider 2001, there is only one of me. But if I am a stage,
pp. 2012). Yet if I will be Bill at t1, is it not which stage? It seems that stage theory
the case that there must at t1 be a stage that retains all of the vices of standard four-
is me, and that that stage is a Bill-stage? Yet dimensionalism but loses any virtues, for at
the same is true for myself and Ben: if I will least the standard worm theorist hold that
be him, then at t1 there must be a Ben-stage there is precisely one of me, and that this
that is me. Yet I cannot be both of them single person is a four-dimensional sum of
since they are distinct! And by the hypo- stages. On the stage view, it looks as though
thesis of equal continuity there is no good eliminativism about persons and other per-
reason to say that I am either. All the stage sistents is the unavoidable consequence: not

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 61

persist enc e
a happy result for a position that wishes properties are had only because one or
to help itself to the commonsense belief that more of its parts has those properties: it is
persons, cars, trees, and the other familiar warm because parts of it are warm, it is
objects of our universe do indeed exist. brick because it has parts made of brick,
and so on. Now whether the worm theorist
Endurance can account for all the properties of a thing
in terms of properties of its temporal parts
Theorists of persistence usually speak as well is highly questionable (see Zimmerman,
of endurantism, taken as the view of vir- 1998), but even restricting ourselves to
tually all those who deny that objects have those that she can so account for, why
(or are) temporal parts, so rejecting four- shouldnt we say that my house changes
dimensionalism of any stripe. There are across space as well, since it has distinct
many such metaphysicians, but whether spatial parts with different properties? My
there is a theory around which they all house does in one sense vary across space,
rally is dubious. The believer in endurance but the endurantist holds that not all vari-
rejects four-dimensionalism for the reasons ation is change, and that something crucial
already given and more. Stage theory, as is lost when change is defined (as it so often
we have seen, denies literal persistence. is in metaphysics texts) as the mere having
Standard worm theory takes there to be of a property by an object at one time and
persistents sums of continuous stages its lacking it at another. For if that is all
but, according to the critic, it denies the there is to change, then objects change
reality of change (see, for example, across space as well, since why on this
Lombard, 1994, replying to Heller, 1992; see view should having a property at a time
also Oderberg, 2004). All there is, on four- be significantly different from having it at
dimensionalism, is replacement of one tem- a place? But change is a fundamentally
poral part by another, or addition of one dynamic phenomenon, involving a real
temporal part to another but neither transition of a thing itself from one state to
replacement nor addition are genuine another. Mere addition, replacement, and/
change. When my red car is painted green, or distinctness of parts do not capture this
a red car-stage (or series of such stages) is phenomenon.
replaced or added to by a green car-stage (or The endurantist is often asserted to hold
series of such stages). As Peirce put it, as a theoretical commitment that a persistent
Phillip is drunk and Phillip is sober would is wholly present at different times (Lewis,
be absurd, did not time make the Phillip of 1986). If there is any theory here, it is the
this morning another Phillip than the denial of four-dimensionalism. But it is not
Phillip of last night (Peirce, 1931, 1.494). the assertion of a recondite metaphysical
The endurance theorist wants to retain state, merely the belief that one and the
both the commonsense belief that there is same object literally exists through time,
literal persistence and the commonsense itself having properties at some times that
belief that there is genuine change through- it loses at others. There are, though, the-
out that persistence. In other words, one oretical consequences of this commonsense
and the same object literally has a property view. For example, endurance rules out any
and loses it. No four-dimensionalist theory approach to fission cases that posits coin-
can hold on to both of these beliefs. True, ciding pre-fission objects, or violation of the
for the worm theorist my car does exist at standard logic of identity (the idea that
every time at which any of its temporal I will be both Bill and Ben is a non-starter),
parts do, but the properties it gains and or any relation weaker than strict identity
loses are gained and lost only in virtue of as capturing what matters as between me
there being distinct stages that have and do and my post-fission descendants. Since
not have those properties respectively. My identity cannot hold between myself and
house also exists at every place at which it both Bill and Ben, the only option for the
its spatial parts do, and many of its intrinsic endurantist is to deny that I continue to

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 62

p e r s i s t ence
exist after fission: I am neither Bill nor Ben. The locus classicus for the law, followed
Since the fission case has psychological con- by virtually all philosophers ever since, is
tinuity built into it, this means denying that Aristotle, who affirms: [T]he same attribute
I really will be psychologically continuous cannot at the same time belong and not
with both Bill and Ben and hence deny- belong to the same subject and in the same
ing that genuine fission is possible, on the respect (Metaphysics, Book Gamma, sect. 3,
assumption that psychological continuity is 1005b19; Ross, 1928; emphasis added).
present wherever personal identity is and When my green car is painted red it certainly
vice versa. (The endurantist could hold the does not, at any one time, possess incom-
weaker position that psychological continu- patible properties the change itself
ity is defeasible evidence of personal identity, ensures that the law is not violated, nor
and simply claim that in the case of myself, could it be. Hence it looks as though the
Bill, and Ben, the evidence is defeated.) problem of temporary intrinsics is spurious:
why would anyone want to affirm the third
Change: metaphysics and semantics proposition unless they had not thought
carefully about the Law of Non-contradiction
What, then, of the phenomenon of change? in the first place, or they wanted something
In contemporary discussion, following Lewis to puzzle about for the sake of it?
(1986), theorists set up what is called the Similarly, though less obviously, change
problem of temporary intrinsics. The idea does not involve any violation of Leibnizs
is that the following propositions are incom- Law. This law (more precisely that half of the
patible: (1) that objects (such as my car or law called the Indiscernibility of Identicals)
me) persist through change; (2) that, across states that if x and y are identical then
the same dimension of change, the intrinsic they share all their properties. Some writers
properties involved in an objects change (e.g., Heller, 1992) argue that if an object
are incompatible (being red and green, or such as my car is red at t and green at t1, then
red and non-red, round and square, round my car at t is discernible from my car at t1,
and non-round, etc.); (3) no object can pos- the first being red and the second non-red.
sess incompatible properties. The problem is But this cannot be, so the properties must be
how, in analyzing change, all of these very possessed by numerically distinct temporal
plausible claims can be held true. Kant, parts (united into a single four-dimensional
for one, gave voice to a worry about how worm). Yet Leibnizs Law is entailed by the
there could be a combination of contradic- Law of Non-contradiction: no object can
torily opposed predicates in one and the both possess a property and lack it at the
same object (Kemp Smith, 1933, A32/ same time and in the same respect. So if x
B48, p. 76). and y are identical, i.e., the same object,
A concern about the way in which this that object x (y) cannot be F and not-F at the
problem has been tackled is that two distinct same time and in the same respect. So it
issues have tended to be conflated: the must be that if x is F, then it (y) is F at the
semantic one of how to represent sentences same time and in the same respect, and vice
describing change in such a way that they versa. So if change violates Leibnizs Law,
do not state a contradiction, and the meta- it violates the Law of Non-Contradiction,
physical one of how change should be which it cannot do. For if change meant
understood in such a way that no contra- that x and y really did not share all their prop-
diction is assumed or implied. Semantics erties, though they were one and the same
and metaphysics are not the same thing, object, a contradiction would result. To say
and so the identity theorist needs to be care- that Leibnizs Law still allows discernibility
ful to separate these issues. Taking the at different times, so is not entailed by The Law
metaphysical one first, the obvious target of Non-contradiction, is to miss the point. For
is the third proposition. Does the Law of if y has a property at t1 incompatible with a
Non-contradiction state that nothing can property that it (x) has at t, there will be a
have incompatible properties? Not at all. contradiction if x does not also have the

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 63

persist enc e
property that y has at t1; it just wont have fast, therefore Fred runs), why cant we
it at t. In other words, Leibnizs Law does not drop temporal adverbs? But if we do so we
mean that objects cannot change: if such end up with a contradiction again my car
change is just what we mean by discernib- is red and non-red so must the adverbial-
ility at different times, that is harmless. But ist say the temporal adverbs cannot be
if we intend something more by the expres- dropped simply to avoid contradiction? If
sion, e.g., that the property y has at t1 is a so, the justification for the solution looks
property that x (=y) lacks at t1, even given that circular: introduce temporal adverbs to block
x exists at t, we end up in contradiction. So: contradiction, but then exclude the stand-
my car at t is red at t, and my car at t is green ard semantic rule for dropping adverbs
at t1; my car at t1 is green at t1, and my car because otherwise there would be a con-
at t1 is red at t. My car does not at any time tradiction. (See further Oderberg, 2004,
possess incompatible properties, though it and also Merricks, 1994 for criticism of
does so at different times. But having incom- adverbialism.)
patible properties at different times does not A third approach is called sentential-
mean that x has any property that y lacks, ism (Oderberg, 2004; see also Myro, 1986,
and conversely. who uses sentential temporal operators
This is where the semantic problem rears but for a different purpose). Taking the
its head, though for all its interest we can separation of semantics from metaphysics
only consider it briefly. The problem is how, seriously, the sententiaist holds that the
semantically, to represent my cars change temporal operators in sentences describing
in such a way that no contradiction is stated change are affixed to atemporal predica-
or implied. What do these expressions such tions. At t, my car is red and at t1, my car
as at t, at t1, and so on, mean? Does it is non-red. The temporal operators create
matter where they are placed in a sentence something like an opaque context: not
stating property possession? There are at strictly, since the context is still extensional
least three alternative proposals for dealing (it doesnt matter what co-referring subject
with this. The four-dimensionalist (including term I use for the sentence to be true),
the stage theorist) applies the metaphysics to but the operators cannot be dropped. Why
the semantics: the temporal qualifiers are not? The main reason is that atemporal
affixed to the subject terms so as to block predications for changeable objects are
a contradiction. My car-at-t is red and my incomplete they do not state facts about
car-at-t1 is green. The subject terms denote the objects, only incomplete information
temporal parts of my car, so no one thing that needs supplementation to make sense.
literally possesses incompatible properties So no inference can sensibly be made for
at any time. For reasons already given, such objects from a temporal to an atem-
the temporal part skeptic will reject this poral predication. This is reinforced by the
approach. Another is called adverbialism semantic fact that when we make predica-
(Johnston, 1987; Hanslanger, 1989): the tions that are not explicitly temporal My
temporal qualifiers are attached to the car is red there is always taken to be
copula. Hence my car is-in-the-t-way red an implicit reference to the present, since
and my car is-in-the-t1-way green. The otherwise the statement would be radically
same object possesses incompatible proper- incomplete and truth unevaluable. Hence
ties but in different ways, and these different semantics is on the side of the sententialist,
ways of property possession remove the whereas the adverbialist has the standard
contradiction. It is difficult to get a grip on rule in favor of adverb dropping to con-
whether adverbialism has any metaphy- tend with.
sical implications, and if so what they are.
One semantic criticism is that the adver- See also the az entries on being and becom-
bialist has to give an account of temporal ing; change; continuant; continuity; identity;
adverb dropping. Since we can usually persons and personal identity; space and
drop adverbs and preserve truth (Fred runs time; temporal parts, stages.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 64

p e r s i s t ence
b i b l i og rap hy Philosophy and Phenomenological Research
54 (1994), 36573.
Armstrong, D.M.: A World of States of Affairs Lowe, E.J.: The Possibility of Metaphysics:
(Cambridge: Cambridge University Press, Substance, Identity, and Time (Oxford:
1997). Clarendon Press, 1999).
Butler, J.: Of Personal Identity, Appendix Meiland, J.W.: Temporal Parts and Spatio-
1 of The Analogy of Religion (1736); repr. temporal Analogies, American Philoso-
in Personal Identity, ed. J. Perry (Berkeley: phical Quarterly 3 (1966), 6470.
University of California Press, 1975). Merricks, T.: Endurance and Indiscerni-
Chisholm, R.: Person and Object (La Salle, IL: bility, The Journal of Philosophy 91 (1994),
Open Court, 1976). 16584.
Coburn, R.: Identity and Spatiotemporal Merricks, T.: There Are No Criteria of
Continuity, in M. Munitz, ed., Identity Identity Over Time, Nos 32 (1998),
and Individuation (New York: New York 10624.
University Press, 1971), 51101. Minkowski, H.: Space and Time, in H.A.
Forbes, G.: Is There a Problem about Lorentz, A. Einstein, H. Minkowski, and H.
Persistence?, Proceedings of the Aristotelian Weyl, The Principle of Relativity (New York:
Society, suppl. vol. 61 (1987), 13755; Dover, 1952); address originally given in
repr. in Haslanger and Kurtz (2006). 1908.
Forbes, G.: The Metaphysics of Modality Myro, G.: Identity and Time, in R. Grandy
(Oxford: Clarendon Press, 1985). and R. Warner, ed., Philosophical Grounds
Geach, P.: Some Problems about Time, in of Rationality (Oxford: Clarendon Press,
his Logic Matters (Oxford: Blackwell, 1972). 1986), 383409.
Haslanger, S.: Endurance and Temporary Oderberg, D.S.: The Metaphysics of Identity
Intrinsics, Analysis 49 (1989), 11925. Over Time (New York: St. Martins Press,
Haslanger, S.: Persistence through Time, 1993).
in M.J. Loux and D.W. Zimmerman, Oderberg, D.S.: Temporal Parts and the
ed., The Oxford Handbook of Metaphysics Possibility of Change, Philosophy and
(Oxford: Oxford University Press, 2003), Phenomenological Research 69 (2004),
31554. 686708.
Haslanger, S. and Kurtz, R.M., ed.: Persis- Peirce, C.S.: Collected Papers, vol. I (Cam-
tence: Contemporary Readings (Cambridge, bridge, MA: Harvard University Press,
MA: Bradford Books/MIT Press, 2006). 1931).
Hawley, K.: How Things Persist (Oxford: Quine, W.V.: Identity, Ostension, and
Clarendon Press, 2001). Hypostasis, The Journal of Philosophy 47
Heller, M.: The Ontology of Physical Objects (1950), 62133; repr. in his From a
(Cambridge: Cambridge University Press, Logical Point of View (New York: Harper and
1990). Row, 1961), and in Haslanger and Kurtz
Heller, M.: Things Change, Philosophy and (2006).
Phenomenological Research 52 (1992), 695 Quine, W.V.: Worlds Away, in his Theories
704. and Things (Cambridge, MA: Harvard
Johnston, M.: Is There a Problem about University Press, 1981).
Persistence?, Proceedings of the Aristotelian Rea, M.C.: Temporal Parts Unmotivated,
Society, suppl. vol. 61 (1987), 10735. The Philosophical Review 107 (1998),
Kemp Smith, N. (trans.): The Critique of Pure 22560.
Reason, by Immanuel Kant (London: Ross, W.D. (trans. and ed.): Aristotles Met-
Macmillan, 1933; originally published aphysics, vol. VIII of The Works of Aristotle
1781 (A), 1787 (B)). Translated into English, 2nd edn. (Oxford:
Lewis, D.: On the Plurality of Worlds (Oxford: Clarendon Press, 1928).
Blackwell, 1986). Shoemaker, S.: Identity, Properties, and
Lombard, L.: The Doctrine of Temporal Causality, in P. French, T. Uehling, and
Parts and the No-Change Objection, H. Wettstein, ed., Midwest Studies in

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 65

r ealism and antirealism about abstrac t entities

Philosophy 4 (Minnesota: University of to accept the existence of propositions and
Minnesota Press, 1979), 32142; also in properties or attributes. (See Quine, 1960, chs.
his Identity, Cause, and Mind (New York: 6, 7 and 1970, ch. 1).
Oxford University Press, 2003). (2) Realism and antirealism, though
Sider, T.: Four-Dimensionalism: An Ontology mutually exclusive, need not exhaust the
of Persistence and Time (New York: Oxford possibilities. One may just be agnostic about
University Press, 2001). whether or not there are abstracta (of some
Swinburne, R.: Space and Time (New York: St given kind). More interestingly, unwillingness
Martins Press, 1968; 2nd edn., 1981). to assert or deny the existence of abstracta
Taylor, R.: Spatial and Temporal Analogies might stem from a conviction that the issue
and the Concept of Identity, The Journal is either hopelessly unclear or confused.
of Philosophy 52 (1955), 599612. Carnaps view that philosophers questions
Thomson, J.J.: Parthood and Identity Across about the existence of numbers, proposi-
Time, The Journal of Philosophy 80 (1983), tions, etc., are not genuinely factual or
20120; repr. in Haslanger and Kurtz theoretical questions at all, but misleadingly
(2006). formulated practical questions calling
Zimmerman, D.W.: Temporal Parts and for a decision whether or not to adopt a cer-
Humean Supervenience: The Incompati- tain linguistic framework rather than an
bility of Two Humean Doctrines, answer assessable as true or false can be
Australasian Journal of Philosophy 76 seen as exemplifying the latter position. (See
(1998), 26588. Carnap, 1950.)
(3) Abstract as opposed to concrete
david s. oderberg entities are commonly taken to be those, if
any, which occupy neither space nor time.
Thus they contrast both with physical enti-
Realism and Antirealism about ties, which occupy both space and time
Abstract Entities (e.g., tables, tennis matches, vapor trails,
and more exotic entities like sub-atomic
1. What Is Realism? particles and force-fields), and with those
entities, if any, which occupy time but not
Realism about abstract entities, in its most space (e.g., mental events, processes and
general form, asserts and antirealism states, on some dualist views), or space but
denies that there are such things. This not time (possible examples: the Greenwich
simple formulation calls for some explanatory Meridian, the North Pole, and spatial points
comment. and regions generally). This explanation
(1) Neither realism, nor its denial, need is not unproblematic. While numbers and
be an all or nothing affair. Realists have sets, and many other standard examples of
asserted, and their opponents have denied, the abstract, have neither spatial nor tem-
the existence abstract entities of several dif- poral location. It makes no sense to ask
ferent kinds universals, mathematical where the number 17 is, or when it came into
entities such as numbers and sets, pro- existence, or how long it will last. But it
positions, and various others (see number; is not clear that this holds for all abstract
class, collection, set; proposition, state entities one might, for example, argue
of affairs). Realists may be selective about that literary and musical works (as distinct
the kinds of abstract entities whose exis- from copies or performances) are abstract
tence they assert, and antirealists may like- entities, but that they have not always
wise be selective denying the existence of existed rather, they came into existence
one kind of abstract entities, while remain- when first composed, and so are not wholly
ing agnostic about, or even accepting, the atemporal. We shall, however, assume its
existence of another. Quine, for example, approximate correctness here. (See con-
was a realist for most of his career about crete/abstract. For skepticism about the
sets and numbers, while steadfastly refusing distinction, see Lewis, 1986, pp. 81 6; for

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 66

r e a l i s m and antir ealism ab o ut abstrac t entities

discussion of difficulties in drawing it, see so as to apply only to what occupies space
Dummett, 1973, ch. 14; Noonan, 1976; and time, or perhaps only to what is
Hale, 1987, ch. 3). capable of causal interaction. If this were
(4) Realism is standardly taken to involve so, the disagreement between his Realism
the further claim that the existence of and realism as characterized here would be
abstracta is objective, this being understood largely if not entirely verbal, at least as far
in terms of mind-independence (see objectivity). as the ontological status of abstract entities
This is both natural and plausible, but not is concerned. One might be similarly tempted
unproblematic. Indeed, the same examples to think that the disagreement between
illustrate the difficulty novels and sym- realists and antirealists in our sense is like-
phonies are (complex) abstract objects, but wise a merely verbal one, in which anti-
ones which would not have existed without realists are simply evincing a prejudice in
a good deal of mental activity on the part favor of restricting application of the word
of their makers. Perhaps the simplest way exists to what is concrete. But while
around this difficulty is to distinguish a some antirealist polemics encourage such
strong form of realism which asserts that a view, the temptation should, in this case,
there are mind-independent abstract enti- be resisted. There is a genuine issue, and it
ties, and weaker forms which assert the concerns knowledge. If his position is to be
existence of abstracta, but not their mind- taken seriously, the realist must claim that
independence. The strong realist thesis is we can have at least some knowledge about
itself open to more and less demanding some abstract entities. But then, given that
interpretations, which differ over how such entities lack spatio-temporal location,
mind-independence is understood. A mini- and so must be incapable of standing in
mum condition for the existence of certain any causal or other natural relations to us,
entities to be mind-independent is that they however remote, he faces a challenge to
would exist even if there were no minds, explain how knowledge about them is pos-
and so exist independently of our actual sible. We shall return to this issue.
knowledge or beliefs about them. But we
may distinguish an extreme form of real- 2. Some Realist Views
ism, according to which the existence of One may, as noted, be a realist about one kind
abstracta is entirely independent, even in of abstract entities but not about others. We
principle, of the possibility of our knowing of illustrate with three examples.
it, from more moderate forms which main- Universals One of the earliest and most
tain that there are abstract entities which famous realist doctrines is Platos Theory
would exist even if there were no thinkers, of Forms, which asserts the existence such
but which accept an epistemological con- things as the Beautiful and the Just in them-
straint to the effect that their existence selves, over and above particular beautiful
must be detectable, at least in principle. objects and just acts which, in Platos view,
(5) Alexius Meinong (1904) (see non- more or less imperfectly exemplify them.
existent objects) denies existence (German: Although Platos usual term for the Forms
Existenz) to abstract entities, but maintains () is often translated as Idea, it is
that they have a different kind of being, clear that he takes them to be abstract enti-
sometimes called subsistence (German: ties existing independently both of our
Bestand). A closely related, but subtly dif- mental activity and of their instantiation
ferent view sometimes called noneism in sensible particulars (see plato). In support
is defended in Routley (1980) and Priest of this view, it may be argued that there
(2005). Meinongs doctrine is standardly is something which different just acts, for
classed as a kind of realism, but in our example, have in common, in virtue of
terms, Meinongian Realism counts, some- which they are all rightly said to be just, and
what paradoxically, as a form of antirealism. that what they have in common does not
It is tempting to suppose that Meinong is depend for its existence upon any of those
using the word exists in a restricted way, particular acts being performed. Each just act

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 67

r ealism and antirealism about abstrac t entities

occurs at a particular time in a particular For every set X there exists a set Y whose
place, but what they have in common has members are exactly the subsets of X
itself no spatio-temporal location. The detailed
If these and similar sentences, taken at face-
interpretation of Platos theory and his
value, are true, then there must be numbers,
arguments for it remain matters of scholarly
sets, etc., to which they refer or over which
controversy, but there is no doubt that his
they quantify. But such sentences are
promulgation of the theory initiated a dispute
widely accepted as true, and are accepted
over the nature and existence of universals
as they stand, without benefit of some re-
often conceived, in opposition to particu-
interpretation which dispels the appearance
lars, as entities such as general properties
of reference to or quantification over num-
which may be wholly present at different
bers, sets, etc. Here we have the premises of
times and places, or instantiated by many dis-
an argument which makes at least a prima
tinct particular objects which has been
facie case for the existence of numbers, sets,
actively pursued in much subsequent philo-
and other mathematical entities abstract
sophy (see universals and particulars).
entities, surely, if any are and hence a case
Propositions Much as realists about uni-
for realism (or Platonism, as it is often called
versals argue for them by appealing to the
see Platonism) about mathematics.
existence of something common to different
particular objects or events which all satisfy
3. Antirealism
some general description or predicate (e.g.,
blue, square, just, etc.), so some philo- Realisms traditional opponents have been
sophers have argued that when different nominalists (see nominalism). Thus in the
speakers or thinkers say or think, say, that medieval dispute over universals, the nom-
172+1 is even, or that Julius Caesar was inalists insisted that there exist only par-
assassinated, there is something common ticular entities, and that the application of
to their distinct linguistic performances or the same general term (or name hence
psychological acts or states. What they the label nominalism) to many distinct
share is a common content what is said particulars does not require the existence of
or thought, as distinct from the saying or a common non-linguistic entity which is
thinking of it. In other words, they all somehow present in each of them, but is
assert, or assent in thought to, the same sufficiently explained by reference to simi-
proposition. Propositions in themselves, in larities between them. Likewise, in the mod-
contrast with the linguistic performances or ern dispute over the existence of abstract
psychological acts or states in which they entities in mathematics, nominalists argue
expressed or encoded, have no spatial or that the acceptance of mathematical theories
temporal location, and hence are abstract involves no unavoidable commitment to
objects (A classic statement of realism about the existence of numbers, functions, sets,
propositions is Bolzano, 1972). or any other ostensibly abstract entities.
Numbers, Sets and other Mathematical Before we consider some of the main strat-
Entities The sentences of pure mathem- egies by which nominalists have sought to
atics almost invariably involve expressions avoid such commitment, we shall briefly
(simple or complex singular terms) whose review their reasons for thinking it is neces-
ostensible role is to make reference to sary or desirable to avoid it.
numbers of some kind, or sets, or other Nominalists have often recommended
mathematical entities, along with quanti- their rejection of abstracta on grounds of
fiers binding variables understood as rang- ontological economy, invoking the meth-
ing over such entities. Simple examples are: odological maxim known as Ockhams Razor
entia non sunt multiplicanda praeter neces-
sitatem which may be glossed as asserting
Every set of real numbers which is that we should not postulate kinds of entity
bounded above by a real number has a beyond what is necessary (see Ockham).
least upper bound in the real numbers Although a popular ploy, this is problematic

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 68

r e a l i s m and antir ealism ab o ut abstrac t entities

for at least two reasons. First, it gives a clear numerically definite quantifications like
directive only when accompanied by some there are exactly two Fs (2xFx). These
answer to the obvious question: Necessary are logically equivalent to sentences invol-
for what? The equally obvious answer is: ving no number words at all, such as there
Necessary to account for all the (agreed) is something which is F and something else
facts, but it is doubtful that there is which is F and any F is identical with one or
sufficient agreement here to enable the other of these things (xy(xy z(Fz
nominalist to cut away abstracta as unnec- z=x z=y) ) ). Thus at some cost in length
essary. The realist is likely to suppose that the and readability, we may be able to reduce
relevant facts include facts of mathematics 2=3=5 to something nominalistically
which, taken at face value, do require the acceptable. But even if this kind of para-
existence of numbers, sets, etc. But second, phrase works for simple equations, it plainly
even if the facts in need of explanation can wont work for general arithmetical state-
be restricted, without begging the question, ments, such as ab (a + b = b + a),
to facts about the concrete, it is still unclear in which we quantify over numbers, with-
that the nominalist will be in position to out mentioning any in particular. Thus
wield the razor to advantage, since it may unless virtually the whole of arithmetic is
be argued that those facts admit of no satis- to lie beyond the nominalists reach, addi-
factory explanation without the aid of tional and more widely applicable methods
scientific (and especially physical) theories of paraphrase or re-interpretation will be
which make indispensable use of mathem- needed.
atics. This often called the QuinePutnam Eliminative structuralism offers a more
indispensability argument receives its promising strategy. On this account, arith-
clearest formulation in Putnam (1971). metic is not a theory about a particular
Since theories (especially mathematical infinite sequence of abstract objects the
theories) ostensibly involving reference to numbers 0,1,2,3, . . . but gives completely
abstracta appear to play an indispensable general information about those objects,
rle in our intellectual economy, nominalists if any, which exemplify a certain structure
can scarcely afford simply to reject them (viz. being a sequence having a first term,
outright; rather, they must explain how we and for each term, a unique next term,
may justifiably retain such theories, without and so no end of terms progressions, or
offending against nominalistic scruples. -sequences, in the usual jargon). Since, on
The standard nominalist response has this re-interpretation, no arithmetic sen-
been to seek ways of paraphrasing or tences assert the existence of any objects,
re-interpreting problematic statements and they are all nominalistically acceptable. A
theories in nominalistically acceptable terms well-known difficulty is that unless there
with the aim of showing that their appar- exists at least one -sequence, the elimina-
ent reference to and quantification over tive structuralists translations of all arith-
abstract entities is unnecessary or merely metic sentences, including those of false
apparent. This strategy has met with limited ones like 2+3=6, come out true. This leaves
success. The difficulties can be well illus- the nominalist facing a dilemma: to avoid this
trated by reference to arithmetic. Consider disaster, she must assert the existence of an
first simple equations, such as 2 + 3 = 5. -sequence but if she asserts that there
As a step towards eliminating its apparent ref- infinitely many abstract objects, she abandons
erence to numbers, we may paraphrase it nominalism, while if she asserts that there are
along the lines: If there are exactly two Fs infinitely many concrete objects, the viabil-
and exactly three Gs and no Fs are Gs, then ity of her translation-scheme depends upon
there are exactly five F-or-Gs (in symbols: an empirical hypothesis, and one which
(2xFx 3yGy x(Fx Gx) ) 5x(Fx may very well be false. Perhaps, as Hellman
Gx) ) . Although this still contains number (1989) argues, this dilemma can be avoided
words, they occur only in the context of by strengthening the structuralist translations

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 69

r ealism and antirealism about abstrac t entities

so that they make claims about what neces- 4. Vehicles of ontological commitment
sarily holds of any -sequence for then reference and quantification
the nominalist need only assert the possible
existence of an -sequence to avoid disaster, It was claimed above that a sufficient condi-
and perhaps the claim that there could be an tion for the existence of objects of a given kind,
-sequence is nominalistically unproblematic F, is the occurrence in true statements of
and otherwise acceptable. However, even if expressions functioning as singular terms
a nominalist version of arithmetic can be which, if they refer at all, refer to Fs. Such
salvaged in this way, it is doubtful whether terms are, we might say, vehicles of onto-
the strategy can be extended to more logical commitment. It might be objected
powerful mathematical theories such as set that the suggested condition cannot be
theory, since the needed possible existence sufficient as it stands, and that we should
claim would amount to the claim that there additionally require that the relevant singu-
could be a concrete model of transfinite set lar terms be ineliminable by reductive para-
theory, and this is surely false. phrase of the sort orthodox nominalists
Following a more radical course, Hartry have sought to supply. But this objection
Field (see Field, 1980, 1989) has argued is confused. Accepting as true statements
that nominalists can deny that mathem- in which certain expressions function as
atical theories are true, thereby avoiding singular terms commits us to the existence
commitment to their abstract ontology, but of corresponding objects, simply because
still accept them provided they are conserva- those statements cannot be true unless
tive in the sense that their conjunction with their ingredient expressions discharge their
non-mathematical (e.g., physical) theories semantic functions, and the semantic func-
entails no claims about non-mathematical tion of singular terms is to pick out objects.
entities which are not logical consequences Antirealists may agree, but object that
of those non-mathematical theories by this misses the real point, which is that if
themselves. Conservativeness in this sense, statements apparently involving singular
like logical consistency, does not require terms for abstract objects can indeed be
truth a theory can be conservative with- replaced by equivalent statements which do
out being true. The important uses of math- not, this shows that those terms are not
ematics in science, Field holds, are two: genuine singular terms at all, and that the
we use it to deduce the consequences of original statements, contrary to first appear-
scientific theories, and we use it, especially ances, involve no commitment to such
in physics, in actually formulating such objects. This antirealist counter assumes
theories. The assumption that standard that if statements apparently involving
mathematics is conservative, Field argues, ontological commitment to Fs are equivalent
is enough to justify its use in deducing, and to other statements apparently free of any
with the help of this assumption, we can, such commitment, it is the latter statements
he thinks, show that there are acceptable which should be reckoned as truly reflecting
nominalistic reformulations of such the- our ontological commitments, not the former.
ories. Fields view has attracted a barrage But why? An equivalence, as Alston
of objections, both technical and philo- (1958) points out in a perceptive discussion
sophical. Several critics have questioned of the issue, is just that what it shows, by
whether Fields reformulations of scientific itself, is only that if either of the two kinds
theories really are nominalistically acceptable. of statement involves a commitment to Fs,
Others have argued that he is committed to then both do. But to get to the conclusion that
the implausible view that while there exist statements of the first sort involve no genuine
no numbers or sets, their non-existence is reference (and hence commitment) to Fs,
a merely contingent matter. (See Maddy, we need a further premise one providing a
1980; Chihara, 1990; Hale and Wright, reason to regard the appearances presented
1992; Burgess and Rosen, 1997). by statements of that sort as misleading, in

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 70

r e a l i s m and antir ealism ab o ut abstrac t entities

contrast with those presented by statements Fs is an alternative vehicle of ontological
of the other sort. Suppose we could introduce commitment to Fs. Quine, famously, took it
terms for the directions of straight lines by to be the sole vehicle:
means of the Direction Equivalence:
The objects whose existence is implied
The direction of line a = the direction of line
in our discourse are finally just the
b iff lines a and b are parallel the idea
objects which must, for the truth of our
being to establish a use for such terms by
assertions, be . . . reckoned into the total-
fixing the truth-conditions of identity state-
ity of objects over which our variables of
ments involving them. (See Frege, 1884,
quantification range. To be is to be the
64.) The nominalist will regard the equi-
value of a variable. (Quine, 1952, 37)
valence as revealing that any apparent com-
mitment to the existence of abstract objects Quine sees quantification as the vehicle
carried by talk of directions is merely appar- of ontological commitment because he
ent. The realist will instead regard it as assumes that only ineliminable occurrences
disclosing an unobvious commitment to of singular terms would distinctively carry
the existence of directions implicit in talk ontological commitment, and believes that
of parallelism among lines. Of course, the there are no such terms, i.e., that singular
realist must agree that one could possess terms are everywhere eliminable. We may,
the concepts of straight line and parallelism he argues, always eliminate them by para-
without having that of direction indeed, one phrase using just general terms (predicates)
must be able to do so, if the latter concept is and quantification, either by the technique
to be explained by means of the Direction of Russells Theory of Definite Descrip-
Equivalence. His claim is that the com- tions (coupled with his doctrine that ordinary
mitment to directions is implicit in the sense proper names are abbreviated descrip-
that, once one has acquired the concept tions) or, if necessary, by an extension of
of direction in this way, one cannot con- it due to Quine himself whereby we may
sistently hold that there are straight lines replace any proper name by a corresponding
but no directions. (For further discussion, predicate understood as applying to that
see Wright, 1983, 5,10.) The realist object, if any, the name names thus
claims we should take apparent reference Socrates drinks, for example, may be para-
to abstracta at face value, in the absence of phrased as x(x socratizes & x drinks).
compelling reason to do otherwise. Resolu- Quine is also taking it for granted that pre-
tion of the issue in favor of an ontologically dicates or general terms carry no commitment
reductive interpretation of such equival- to corresponding entities. If this assumption
ences and so in favor of the antirealist to which we shall need to return were
requires making a case that there is com- granted, it would be at least plausible that
pelling reason to do otherwise. We shall quantification over Fs is the essential mark
return to this question. of commitment to their existence.
Our proposed sufficient condition for the However, while Quines eliminability
existence of Fs is clearly not a necessary con- thesis is, in one way, beyond dispute, its
dition. It may be that there are Fs whose significance is not. We may agree that,
existence we suspect not, and of which, starting from a base language containing
therefore, we do not speak. Perhaps, indeed, singular terms, we could employ Quines
we have no concept of them. Nor, evidently, recipe to construct a language in which all
is a readiness to make statements involving such terms were replaced by corresponding
singular terms for Fs needed for a com- predicates, but deny this purely syntactical
mitment to their existence. For without manoeuvre has any semantic or, more
employing any words which purport refer- widely, philosophical significance. It is quite
ence to particular Fs, we may simply assert unclear how one might learn the use or sat-
that there are Fs, or more generally, assert some isfaction conditions of Quines replacement
quantified statement whose truth requires predicates, in the absence of any means of
their existence. Roughly, quantification over making singular reference to the objects

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 71

r ealism and antirealism about abstrac t entities

which, if any, uniquely satisfy them. be no such knowledge. (See Benacerraf, 1973;
Relatedly, it does not seem one could Steiner, 1973; Kitcher, 1978). As against
explain the truth-conditions of quantified this, it may be claimed that even a broadly
sentences of Quines replacement language causal theory is open to objection on inde-
without treating variables as, in effect, pendent grounds; in particular, such a
functioning as temporary names of objects in theory would seem directly to rule a priori
the domain of quantification . (See Dummett, knowledge and while there is certainly a
1973, pp. 2236, 47680.) serious problem in explaining how such
knowledge is possible, it does not seem that
The Access Problem its impossibility should be so easily estab-
lished. However, as Field (1989, pp. 257,
Realists need to explain how we can know 2309) and others (Hart, 1977; Maddy,
about the abstract entities whose existence 1990, pp. 425) have pointed out, doubts
they assert how we can know that there about the capacity of realism to deliver a
are such things at all, and how we can credible epistemology do not have to be
know truths about them. The problem of grounded in the adoption of a specifically
providing such an explanation is part of causal analysis of knowledge. For even if a
what I shall call the access problem. It is causal constraint is not written into the
the fundamental problem for realism. If analysis, the problem of explaining how
realists could solve it, it is difficult to see we can acquire knowledge, or reliably
what, other than prejudice, would stand in form true beliefs about abstract objects,
the way of acceptance of their view. If, on the remains.
other hand, it could be shown that they Epistemological perplexity about, and
cannot solve it, that would be a decisive consequent suspicion of, abstract entities
objection, and would encourage, or even has other and more general sources, besides
enforce, an ontologically reductive reading causalist or, more generally, reliabilist
of the kind of equivalences between state- thinking in epistemology, which arguably
ments ostensibly about abstracta and others obstruct progress on the access problem.
apparently free of commitment to their One is that we tend to operate with a
existence discussed in the preceding section. wholly negative conception of abstract objects
In the absence of at least the outlines of a as outside space and time. This charac-
solution, or reason to believe one can be terization is obviously metaphorical, as well
found, it is hard to take realism seriously as negative there is, literally, nowhere
ontology without epistemology is just idle outside space and time. But this in itself
speculation. (See Hart, 1979; Bell, 1979). need not be particularly damaging, so long
Why do we or might we find the idea as we remind ourselves, when necessary
that we may have knowledge about abstract that is, when we feel tempted to think
objects so baffling? In explaining how we of abstract objects as in some queer sort
know much of what we know, we appeal to of limbo of the literal content of the
causal connections, such as those involved metaphor: roughly, that it makes no sense
in perception. This may encourage accept- to ask where an abstract object is, or when
ance of a broadly causal theory of know- it came into existence, or how long it will
ledge one which sees basic bits of last. It is, rather, the negative aspect of the
knowledge as involving a suitable causal characterization that impedes constructive
connection between knowers and the known thought. Of course, it is true that abstract
truths, and other knowledge as arising from objects arent located in space or time. And
this basis by a more or less complicated pro- it may be said that since that it enough to
cess of inference. Then, given that abstract ensure that there is an apparently intract-
objects stand in no spatial or temporal rela- able problem about how spatio-temporally
tions with us, and so in no causal relations, located knowers could know of their existence
it may seem not just that knowledge about or know anything about them, it is pointless
them eludes explanation, but that there can exercising ourselves over what more positive

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 72

r e a l i s m and antir ealism ab o ut abstrac t entities

characterization, if any, they can be given. put the objection is like this: even if one
But that just misses the present point: if we could give a realist account of the truth con-
focus exclusively on what abstract objects are ditions for mathematical statements (or any
not, with no thought about what they are or other class of statements supposed about
might be supposed to be, we can scarcely abstract objects), it would be impossible
expect anything but intellectual paralysis to explain how such statements, so under-
when we try to consider how we might get stood, could be known or reasonably believed;
to know about them. but in fact one cannot even give such a
The second factor is the idea that know- semantical account, since one cannot even
ledge of truths about objects of any kind so much as make reference to objects of
must involve contact with those objects. If the sort such an account takes them to be
contact is taken literally, so as to require about and if one cannot do that much,
some sort of physical connection or interac- one cannot so much as state realist truth-
tion perhaps of the sort that occurs in conditions. It helps to recast the objection in
normal sense perception, or even something this way, because doing so gives a clearer
more indirect the idea is obviously inimical view of the structure of the task that must
to realism, but equally not obviously one be addressed by a defensible form of realism.
that must be accepted. Of course, if it is The fundamental part of the access problem
given a sufficiently attenuated (and perhaps is not the knowledge problem (i.e., how, given
unavoidably metaphorical) construal, so that certain statements (e.g., mathematical
that possession of any sort of identifying ones) are about abstract objects, we could
knowledge of an object suffices for contact, know them to be true), but the reference
the idea reduces, near enough, to a truism problem (i.e., how they could be about such
one can hardly be credited with knowledge objects in the first place).
of truths about objects unless one knows That, then, is the problem the realist
which objects are in question and it need should tackle first. Although solving the re-
then cause the platonist no trouble, unless ference problem is merely a necessary, and
it is coupled with the further idea that not a sufficient, condition for a solution to
such contact is presupposed by and must the knowledge problem, one might expect a
be already in place before any knowledge good solution to the former to suggest how
of truths about objects can be had (cf. best to approach the latter. But how, if at all,
Russells famous principle that Every pro- may realists solve the reference problem?
position which we can understand must In my view (for a concise statement, see
be composed wholly of constituents with Hale and Wright, 2002, sect. 5), their best
which we are acquainted (see Russell, 1912, hope lies in rejecting the assumption that an
chs. 4, 5). ability to engage in identifying reference to,
Once we become locked into thinking or thought about, abstract objects is a
about the access problem within this strait- precondition for understanding statements
jacket, we can hardly avoid the further about them, as is suggested by the contact
thought that the access problem is not model and Russells Acquaintance Principle
just a problem about how we can know any- (see acquaintance). Positively, they should
thing about abstract objects, but goes wider argue that concepts of kinds of abstract
and deeper: how can we even so much as object may be introduced by fixing the
think about them at all. truth-conditions of complete sentences
Critics of realism may see this as just so involving terms for them, in accordance
much more grist to their mill: realism is in with Freges Context Principle (Only in
trouble on two counts, not just one, because the context of proposition does a word
it obstructs both a satisfactory epistemo- mean anything cf. Frege, 1884, 62).
logy and a workable theory of reference (cf. More specifically, they may then deploy
Benacerraf, 1973, p. 412; Field, 1989, p. 68). what have come to be known as abstraction
But this way of putting the difficulty obscures principles as a means of explaining both how
an important connection. The right way to terms for abstract objects are to be understood

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 73

r ealism and antirealism about abstrac t entities

and how basic truths about them may be Journal of Philosophy 89:3 (1992), 111
known a priori. Examples are the Direction 35.
Equivalence (see above, sect. 4) and Humes Hart, W.D.: The Epistemology of Abstract
Principle: Objects, Proceedings of the Aristotelian
Society, suppl. vol. 53 (1979), 153 65.
The number of Fs = the number of Gs iff
Hart, W.D.: Review of Mark Steiner, Math-
the Fs correspond oneone with the Gs
ematical Knowledge (Ithaca, NY: Cornell
Whether the access problem can be solved University Press, 1975), in Journal of
in this, or some other way, is a matter of Philosophy 74 (1997), 11829.
currently active debate. Hellman, G.: Mathematics Without Numbers
(Oxford: Clarendon Press, 1989).
Kitcher, P.: The Plight of the Platonist,
b i b l i og rap hy
Nos 12 (1978), 11936.
Alston, W.: Ontological Commitments, Lewis, D.: On the Plurality of Worlds (Oxford:
Philosophical Studies 9 (1958), 817. Blackwell, 1986).
Bell, D.: The Epistemology of Abstract Maddy, P.: Naturalism in Mathematics
Objects, Proceedings of the Aristotelian (Oxford: Clarendon Press, 1990).
Society, suppl. vol. 53 (1979), 13552. Maddy, P.: Realism in Mathematics (Oxford:
Benacerraf, P.: Mathematical Truth, Jour- Clarendon Press, 1980).
nal of Philosophy 70 (1973), 66180. Meinong, A.: The Theory of Objects, in
Bolzano, B.: Theory of Science, ed. and R. Chisholm, R., ed. Realism and the
trans. Rolf George (Oxford: Blackwell, Background to Phenomenology (London:
1972). From Wissenschaftslehre (Sulzbach, Allen & Unwin 1960; originally pub-
1837). lished 1904).
Burgess, J. P. and Rosen, G.: A Subject With Noonan, H.: Dummett on Abstract Ob-
No Object: Strategies for Nominalistic Inter- jects, Analysis 36:2 (1976), 4954.
pretation of Mathematics (Oxford: Claren- Priest, G.: Towards Non-Being: The Logic
don Press, 1997). and Metaphysics of Intentionality (Oxford:
Carnap, R.: Empiricism, Semantics, and Clarendon Press, 2005).
Ontology, Revue Internationale de Phil- Putnam, H.: Philosophy of Logic, New
osophie 4 (1950), 2040. York: Harper & Row, 1971; repr. in
Chihara, C.: Constructibility and Mathemat- Putnams Philosophical Papers Volume 1
ical Existence (Oxford: Clarendon Press, (Cambridge: Cambridge University Press
1990). 1975).
Dummett, M.: Frege: Philosophy of Language Quine, W.V.: Methods of Logic (London:
(London: Duckworth, 1973). Routledge and Kegan Paul, 1952). Quine,
Field, H.: Realism, Mathematics, and Modal- W.V.: Philosophy of Logic (Englewood
ity (Oxford: Blackwell, 1989). Cliffs, NJ: Prentice-Hall, 1970).
Field, H.: Science Without Numbers (Oxford: Quine, W.V.: Word & Object (Cambridge,
Blackwell, 1980). MA: MIT Press, 1960).
Frege, G.: Die Grundlagen der Arithmetik Routley, R.: Exploring Meinongs Jungle and
(Breslau, Poland: Wilhelm Koebner, 1884); Beyond (Canberra: Australian National
trans. into English by J.L. Austin as University, 1980).
The Foundations of Arithmetic (Oxford: Russell, B.: The Problems of Philosophy
Blackwell, 1959). (London: Oxford University Press, 1912).
Hale, B.: Abstract Objects (Oxford: Blackwell, Steiner, M.: Platonism and the Causal The-
1987). ory of Knowledge, Journal of Philosophy
Hale, B. and Wright, C.: Benacerrafs 70 (1973), 5766.
Dilemma Revisited, European Journal of Wright, C.: Freges Conception of Numbers as
Philosophy 10 (2002), 10129. Objects (Aberdeen: Aberdeen University
Hale, B. and Wright, C.: Nominalism and Press, 1983).
the Contingency of Abstract Objects, bob hale

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 74

s p a c e and time
Space and Time material things, the particle could be moving
from one cell to another of space itself.
This article discusses the following issues Newton argued for the existence of sub-
about space and time: whether they are stantival space with a famous thought
absolute or relative, whether they depend on experiment. Imagine a bucket suspended
minds, what their topological and metrical from a rope and filled with water. The rope
structures may be, Mctaggarts argu- is twisted and allowed to unwind, causing
ment against the reality of time, the ensuing the bucket to spin. At first the bucket moves
split between static and dynamic theories relative to the water, the water not yet
of time, problems with presentism, and the having begun to partake of the buckets
possibility of time travel. Our opening ques- motion, but eventually friction causes the
tions are posed in the following query water to rotate as well, and indeed to catch
from Kant: up with the bucket so that there is no
What, then, are space and time? Are they longer any relative motion between water
real existences? Are they only determina- and bucket. By the time this happens, some-
tions or relations of things, yet such as thing else happens as well: the surface of
would belong to things even if they were the water has become concave, the water
not intuited? Or are space and time such edging up the sides of the bucket. This
that they belong only to the form of intui- is explained in Newtonian mechanics as
tion, and therefore to the subjective con- a centrifugal-force effect, similar to what
stitution of our mind, apart from which happens when amusement park riders are
they could not be ascribed to anything pinned to the side of a rotating bottomless
whatsoever? (A23/B37) drum. Newtons argument now proceeds
as follows:
a b s ol u te o r r elative? 1. There is a time at which the water displays
centrifugal-force effects, but is not mov-
Newton regarded space as a real existence
ing relative to the bucket or any other
a vast aetherial container without walls,
material thing. (Why not relative to the
in which everything else that exists lives
ceiling, you ask? That is why the experi-
and moves and has its being. Leibniz
ment is a thought experiment: we are to
believed to the contrary that space is not a
imagine it performed in a universe with
genuine entity, but a mere faon de parler; he
no objects besides bucket, water, and
held that all talk of space is replaceable by
talk of material things and their relations
2. All centrifugal-force effects are induced by
to one another. For example, to say that
rotational motion
a thing has changed its place is merely
3. Therefore, there is a time at which the
to say that it has changed its distance or dir-
water is moving, but not relative to any
ection from some other thing chosen as a
material thing (from 1 and 2).
reference object. This is the issue that
4. Motion that is not relative to any mater-
divides partisans of absolute or substantival
ial thing is absolute motion, that is,
theories of space on the one hand from
motion with respect to space itself.
defenders of relative or relational theories
5. Therefore, the water is moving with
on the other.
respect to space itself (from 3 and 4)
To test his or her allegiance on this issue,
which must therefore exist.
the reader should answer the following
question: if the only material thing in exist- Newton thus argues that accelerated motion
ence were a single particle, would it make (the waters constant change of direction)
sense to say that it is moving? Leibniz would reveals itself in its effects and proves the
say no, since motion for him consists in existence of space, the existence of which then
change of relations (e.g., of distance) among grounds absolute uniform (non-accelerated)
two or more material things. Newton would motions, even though the latter do not
say yes, since even in the absence of other manifest themselves.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 75

space and t i me
Berkeley and Leibniz objected to the impure properties are being Fred (impure
conclusion of Newtons argument, but and intrinsic) and being married to Fred
without making clear which premise they (impure and relational). When Leibniz affirms
thought wrong. Mach objected to premise premise 2, he must mean that w and w
1, claiming that we simply do not know differ in no pure property, for Newtonians
how the water would behave in a universe could certainly maintain that w and w are
devoid of ceiling and stars (as though no distinguished by the fact that w is such
physicist ever extrapolated his laws to that part of the cosmos occupies cell 233
hypothetical situations, such as frictionless (an impure property), whereas in w, cell
planes!). A generally overlooked response 233 is empty. But that means when we
challenges premise 4: perhaps motion is get to premise 3, Leibniz must advance his
really absolute, that is, not a change in rela- Identity of Indiscernibles principle in the
tion to anything else at all, be it matter or following form: any two things must differ in
space. The possibility of this last response at least one pure property (and not merely in
shows that we should separate two issues that such properties as being identical with this
can be posed using the absolute vs. relative thing). Leibniz no doubt did wish to affirm the
formula: is space a substance or a system of principle in the required form, but if so, it
relations, and are motion, size, and various is open to counterexamples. Is it not con-
other spatial commodities absolute (intrinsic) ceivable that there be two spheres the same
or relational? in color, shape, composition, and every other
Leibniz argued that space is a pseudo- pure property you care to think of?
entity because its existence would generate The substantival vs. relational issue car-
distinctions without a difference or, more ries over to time. For Newton, time flows
precisely, exceptions to his principle of the equably without regard to anything exter-
identity of indiscernibles. Let w and w be nal; for Leibniz, time is nothing over and
two universes just alike in how all material above the sequence of events said to be in
things are related to one another, but differ- time. Newton (but not Leibniz) can make
ing in the alleged respect that in w the sense of the idea that the entire history of the
entire material cosmos has been moved six world (comprising the same events as now)
miles to the east or rotated through some might have begun earlier than it did.
angle. Leibnizs argument then proceeds as
real or i deal ?
1. If there were such a thing as substanti-
Another issue about space and time is
val space, w would be distinct from w.
whether they are ideal, that is, dependent for
2. But w and w are indiscernible they
their existence on minds. The most famous
share all their properties.
idealist about space and time in western
3. Things that are indiscernible are identi-
thought is Kant. Kant began his intellectual
cal. Putting it the other way around,
career as a Leibnizian, but was briefly con-
any two distinct things must differ in at
verted to Newtons view by considerations
least one property.
about incongruent counterparts objects
4. Hence, w = w after all (from 2 and 3).
that come in mirror image forms, like left and
5. Therefore, there is no such thing as sub-
right human hands. Kant thought the dif-
stantival space (from 1 and 4).
ference between incongruent counterparts
To evaluate this argument, we need to dis- could not be explicated using only relation-
tinguish two kinds of properties. A property ist resources, but had to consist in the differing
is pure if its being exemplified does not relations of the objects to space itself. By the
depend on the existence of any specific indi- time he wrote the Critique of Pure Reason,
vidual and impure otherwise. Examples of however, Kant had come around to the
pure properties are being red (which is pure third of the positions in the quotation
and intrinsic) and being next to some- above: space and time are merely forms of
thing red (pure and relational); examples of intuition, that is, ways in which human

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 76

s p a c e and time
beings order and arrange the things they potentially infinite always extendable
perceive; they are not features of things further into the past through our future
in themselves, or things as they exist outside discoveries but neither actually finite nor
the mind. actually infinite.
A characteristically Kantian reason for
believing that space is ideal is that no other
struc t ural quest i ons
hypothesis accounts for our knowledge of
geometry. Kant thought that geometry was The next group of questions about space
a body of synthetic and a priori truth a and time (or spacetime, in the Minkowskian
priori in that it is known in advance of melding of them) concerns their (or its)
experience, yet synthetic in that it is not metrical and topological structure. Are space
validated just by logic or the meanings of our and time infinitely divisible, or are there
concepts. How can that be? How can we smallest units? (Zenos paradoxes of motion
know even before we encounter them that are sometimes seen as set up so that the first
cubes on Mars will have 12 edges? Kants two apply if space and time are infinitely
answer is that (i) our form of intuition divisible and the second two if space and
makes us incapable of intuiting (perceiving time are quantized.) Does space obey the
or imagining) any cubes that do not have 12 laws of Euclidean geometry or those of one
edges and (ii) as prescribed by idealism, of the non-Euclidean geometries known to
no cubes or spatial objects exist anywhere be consistent since the nineteenth century?
except those that satisfy the conditions of our How many dimensions does space have?
intuiting them. Thus all cubes everywhere Could time have a beginning or an
have 12 edges and the other properties end? Must time be unilinear, or might it
imposed on them by our Euclidean form of branch into multiple paths or close back
intuition. upon itself in a loop?
Kant thought the ideality of space and time The dimensionality of space is represen-
was further confirmed by the antinomies tative of such questions. We all know about
pairs of opposed propositions in which one three dimensions of space a line possesses
or the other must be true if space and time one dimension, a plane two, and a solid
exist outside the mind, but both of which are three. What would it mean for space to
impossible. For example, does the world have a fourth dimension? (We are talking
have a beginning in time, or is it infinite now of a fourth spatial dimension, not time,
in its past duration? If things in time were even though time is sometimes considered
things in themselves, one of these alternatives as a fourth dimension.) Galileo offered one
would have to be true, yet both of them criterion: to say that space has n dimen-
boggle the mind. No beginning would mean sions is to say that n mutually perpendicu-
an infinity of events already elapsed, which lar lines (but no more) can meet in a single
Kant thought impossible because it would point. If our space were four-dimensional,
involve a completed infinity. (Think of a line could enter the corner of my desktop
Wittgensteins example of the man we at right angles to each of its three edges.
find saying, . . . 5, 4, 3, 2, 1; whew! Poincar offered another criterion: points
I just finished counting through all the neg- are zero-dimensional, and an entity is n-
ative integers.) A beginning would mean an dimensional iff n is the lowest number such
event for which there could not possibly be that any two points of the entity may be
a sufficient reason a blow to rationalist separated from each other by an entity of
aspirations, if not the outright impossibility n 1 dimensions. Thus, a line has one
Kant seemed to think it was. Kants solution dimension, because any two points of it can
was to hold that past events exist only in pre- be separated from each other by an inter-
sent or future memories or other evidence vening entity of zero dimensions (another
of them (for example, yet-to-be-perceived point); a plane has two dimensions, because
cosmic radiation). He thought this opened any two points within it may be separated
the possibility that the worlds history is by a circle enclosing one of them or a line

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 77

space and t i me
running all the way across the plane be- A view that lies between the traditional and
tween them; and so on. It is a consequence the contemporary views is the convention-
of this criterion that in a four-dimensional alism of Poincar. Poincar thought that all
space, a two-dimensional entity would not the empirical data accommodated by non-
suffice to separate one point from another. Euclidean geometry plus standard physical
Thus a spherical shell enclosing point A but theory could equally well be accommodated
not point B would not suffice to separate A by Euclidean geometry together with
from B you could get from A to B without non-standard physical theory. For example,
penetrating the shell. measurements apparently indicating that
Such things defy visualization in a way that the ratio of circles to their diameters does
makes some people want to declare them not have the familiar value of could be
impossible. Those so inclined should read accommodated by a non-Euclidean geo-
E.A. Abbotts Victorian classic Flatland, metry in which this ratio is indeed other
in which the author describes a world of than , but they could also be accommod-
two-dimensional beings who are incapable ated by positing a heat gradient that causes
of rising out of their plane or visualizing our yardsticks to expand when laid along the
anything beyond it. A Flatlander may be diameter though not when laid along the cir-
imprisoned simply by enclosing him within cumference. We could thus always choose to
a circle or a polygon. Could a Flatlander but describe our world in Euclidean terms by
jump over the walls of his prison, he would complicating our physics. This position is at
be free, but he is incapable even of conceiv- odds with a hardy empiricism, in so far as it
ing such a motion as we are of any path denies that empirical results can settle the
from the interior to the exterior of a spherical structure of space, but it is also at odds with
shell that does not pass through the shell. an ambitious a priorism, in so far as it denies
The exhortation Upwards, not northwards! that decisions in favor of Euclid are deter-
falls on the Flatlanders ears as nonsense. minations of independent fact.
Abbotts intent, of course, is to soften us up
for the possibility that our own resistance
questions about t i me
to a fourth dimension may be as provincial
as that of the Flatlanders to a third. For issues specifically about time, the best
Questions about the structure of space point of departure is McTaggarts famous
and time give rise to meta-questions about argument of 1908 that time is unreal.
proper jurisdiction who is to answer Though few have accepted the conclusion of
them, and how? A traditional view is that this argument, nearly all students of time
space and time necessarily possess what- have taken over the distinctions McTaggart
ever structure they do, and that it ought to employed in formulating it.
be ascertainable a priori what this structure McTaggarts fundamental distinction is
is. Kant, for example, certainly believed that between the A-series and the B-series. An
space is necessarily three-dimensional and A-series is a series of events or moments
Euclidean. The prevalent contemporary possessing the characteristics of being past
view is that space and time have their (in varying degrees), present, or future; call
structures contingently, and that it is only these the A-characteristics. The B-series is
through the best science of the day that we a series of events or moments standing
can reach any reasonable opinion concern- in the relations of earlier-than, later-than,
ing what these structures are. This view and simultaneous with; call these the B-
was given impetus by Einsteins use of relations. The chief difference McTaggart
a non-Euclidean geometry in conjunction notes between the A-characteristics and the
with the General Theory of Relativity to B-relations is that the former are transient
explain gravitation; it is further exemplified while the latter are permanent: If M is ever
in the work of those physicists in search of earlier than N, it is always earlier. But an
a theory of everything who posit a space event, which is now present, was future,
of 11 dimensions. and will be past (LePoidevin and MacBeath,

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 78

s p a c e and time
1993, p. 24). In the ordinary way of think- copulas. If this is right, then in saying that
ing about time, McTaggart believes, an event an event has been future and will be past, we
becomes increasingly less future, is momen- are introducing a new A-series, this time
tarily present, and then slides ever farther into of moments. And this brings back our con-
the past. Yet all the while its B-relations to tradiction, because every moment, like
other events (e.g., its following the Battle of every event, is past, present, and future. If we
Waterloo and preceding the first landing on try to get rid of the contradiction by saying
the moon) are fixed. of moments what we said earlier about
McTaggarts overall argument against the events, our statement means that the
reality of time may be stated quite briefly: moment in question is future at a present
(I) time essentially involves an A-series; moment, and will be present and past at
(II) any A series involves a contradiction; different moments of future time. This, of
therefore, (III) therefore, time is unreal. course, is the same difficulty over again.
Behind each main premise is a subsidiary And so on infinitely (LePoidevin and
argument. The argument behind premise MacBeath, 1993, p. 33).
I is this: Why is McTaggart so convinced that
there is a contradiction in the A-series and
1. There can be no time without change.
a regress in any attempt to remove it? His
2. There can be no change without an A-
thought on these matters can be made
more understandable by presenting it
3. Therefore, there can be no time without
with the help of a metaphor. He begins
an A-series.
by supposing that the whole of history is
Both premises in this argument have been the laid out in a block comprising the B-series.
subject of interesting debate, but our focus He notes that in such a series, there is no
here will be on the argument behind main change and therefore no time, all events
premise II, which runs thus: simply sitting there alongside one another
on the B-axis. What can add time to such
1. The A-characteristics are mutually
a universe? We must bring in the A-
incompatible, yet
characteristics, letting the spotlight of
2. Every event in any A-series must have
presentness wash along the series in the
all of them, so
direction from earlier to later. But wait! If
3. Any A-series involves a contradiction.
the spotlight illuminates event e before it
McTaggart immediately anticipates an objec- illuminates event f, then the events of es
tion the reader will have to premise 2: being present and fs being present are both
it is not true that any event must have all there on the B-axis, permanently related
the A characteristics at once, but only that by the relation of earlier-than. Similarly, if
it must have them successively. An event the shadow of pastness falls on e before it
that is now present is not also past and falls on f, then es being past and fs being
future; rather, it was future and will be past permanently stand in the B-relation of
past. In reply, McTaggart claims that this earlier-than and are thus always there on
attempt to avoid the contradiction he the B-axis. What we are saying implies
alleges only raises it anew. What, he asks, that that e and f are both always past and
is meant by tensed verb forms such as always present surely a contradiction,
was and will be? His answer may be just as McTaggart alleges. If we seek to
given in the schema remove the contradiction by saying that
the spotlight of the present falls on es being
S {was, is now, will be} P iff for some
present before it falls on fs being present, we
moment m, S has P at m & m is {past,
are only embarking on a useless regress
present, future}
again just as McTaggart alleges.
where the italicized verbs are meant to be As noted above, few besides McTaggart
tenseless. He thus believes that tense can be have accepted his argument in toto, but
reduced to A-characteristics and tenseless many have accepted one half or the other.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 79

space and t i me
This gives rise to a great divide in the philo- called the A theory and the B theory,
sophy of time. One side accepts his first though the names can be misleading.
main premise while rejecting the second: There is an entire cluster of doctrines
the A-characteristics (or some surrogate that tend to go together under the banner of
for them) are indeed essential to time, but the A theory and another opposing cluster
there is nothing wrong with that. The other under the banner of the B theory. (Other
side accepts his second main premise while labels for the two sides are the dynamic
rejecting the first: there is indeed a defect versus the static theory and the theory of
in the A-series, but a B-series by itself is all passage or becoming versus the theory of
you need to have time. For obvious reasons, the four-dimensional manifold.) The rival
these two responses to McTaggart are often doctrines may be tabulated as follows:

The A Theory (Dynamic Time) The B Theory (Static Time)

A1. Tense is an irreducible and indispensable B1. Tense is reducible or eliminable; reality is
feature of thought and language, reflecting adequately describable without it.
a genuine feature of reality.
Corollary: some propositions change in truth Corollary: every true proposition is timelessly
value with the passage of time. true.

A2. The A-characteristics are successively B2. The A-characteristics are either delusive or
possessed by all events, and they are not reducible to the B-relations.
reducible to the B-relations.

A3. The present is ontologically privileged: B3. Past, present, and future are ontologically
things present have a reality not belonging on a par: things past and future are no less real
to things past or future. than things present.

A4. The future is open or indeterminate: some B4. The future is as fixed as the past; every
propositions about what is going to happen proposition must be true or false, and
in the future are not yet either true or false. propositions have their truth values eternally
(as noted in B1).

A5. Identity through time is endurance: B5. Identity through time is perdurance: a thing
numerically the same thing exists that lasts through time is a series of distinct
at many distinct times. temporal parts or stages, united by some
relation other than identity.

In row 1, we have the debate between those his boyhood is no more, and those like the
who take tense as primitive and those who Tralfamadorians in Vonneguts Slaughter-
seek to reduce it to something else (as house Five, who do not cry at funerals
Smart once did when he suggested that it because their departed loved one exists and
will rain just means rain occurs later breathes at an earlier moment. In row 4, we
than this utterance.) In row 2, we have have an issue that goes back to Aristotles
the debate between the A theory proper and discussion in De Interpretatione: must the
the B theory proper, which is sometimes proposition the captain will order a sea battle
too quickly equated with the debate in row tomorrow be true or false today, and if so, does
1. (Arguably, tenses are not equivalents of that mean the future is in some way fixed or
the A- characteristics, but superior substitutes fated? Finally, in row 5 we have the issue
for them.) In row 3, we have the issue (stated in David Lewiss terms) that divides
that divides presentists from eternalists those who believe in genuine continuants
those like Augustine, who laments that from those who accept an analysis of

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 80

s p a c e and time
identity through time like that of Williams, going by at the rate of one second per sec-
who once observed that each of us pro- ond? That is no rate at all. One second per
ceeds through time only as a fence proceeds hypersecond? That takes the first step in a pre-
across a farm that is, by having differ- posterous series of time orders. So time does
ent parts at different moments or regions not pass.
(Williams, 1951, p. 463). When the argument is formulated that
As noted, a philosopher who holds a view way, it presupposes a substantival theory of
in one of the columns will tend to hold the time as though there were drops of time
other views in that column as well. There passing through an hourglass. Perhaps,
is a certain amount of room for mixing then, the argument can be sidestepped by
and matching, however, and it should not combining belief in dynamic time with a
be assumed automatically that the proposi- rejection of substantival time. Such is the
tions in a given column must go together as combination espoused by Arthur Prior, the
a package deal. founder of tense logic. Prior represents
Indeed, no one should hold all of the tenses with operators, akin to modal oper-
propositions in column A, for a little ators: Peter will sneeze becomes It will be
reflection shows that A2 is inconsistent the case that Peter sneezes, symbolizable as
with A3. If presentism is true, there are no Fp, and Peter sneezed becomes It was
things or events that are not present, and thus the case that Peter sneezes, symbolizable as
no items possessed of pastness and futurity. Pp. The present tense is the default tense
So if A3 is true, A2 is false. and needs no operator. With this apparatus,
The best combination among A1A3 for it is possible to articulate many propositions
a friend of dynamic time is arguably A1 and about the structure of time. For example,
A3 without A2. Ironically, this would be an the density of time may be expressed as
A theory without the A-characteristics, (p)(Fp FFp). This formula would not be
so the common name is not well chosen. true if time were discrete, for if there were an
McTaggarts combination was just the immediately next moment and a proposi-
opposite, and this is arguably what led to tion p true at it but not thereafter, Fp would
the demise of time in his philosophy. His be true and FFp false. Prior denies that time
argument depends on reducing tense to the is a literal object, a sort of snake which
A-characteristics, and it also depends on either eats its tail or doesnt, either has ends
making the eternalist assumption that the or doesnt, either is made of separate seg-
earlier and later portions of the B-series are ments or isnt; rather, these issues can
equally real. A presentist could evade the be formulated using propositional variables
argument by denying that an event is there and tense operators in a way that makes no
before it becomes present; rather, the event reference to time or its parts (Prior, 1968,
simply becomes it comes into being and then p. 189).
as quickly passes out of being. Or better yet Returning now to the question of times
(since an ontology of things goes better with passage, Prior suggests that the metaphor can
presentism than an ontology of events), a be cashed out in tense logic as follows: there
thing becomes F and then is no longer F. are true instances of the schema Pp & ~p
The issue debated in rows 1 and 2 is it was the case that p, but is not now the case
sometimes put this way: does time pass, or that p. When the matter is put that way, it
is there simply a huge four-dimensional is no longer obvious how awkward ques-
manifold with time as one of its dimensions? tions about the rate of times passage are to
Some philosophers think the passage view be formulated.
may be refuted by asking a simple question:
how fast does time pass? If the first second of
probl ems f or present i sm
the year 2050 is getting closer to us, there
must be a rate at which it is doing this, Presentism is easily misunderstood. Pre-
yet any way of assigning the rate would sentists are not holocaust deniers; their
be nonsensical or absurd. Are the seconds insistence that nothing past exists is

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 81

space and t i me
compatible with their affirming truths for you is real for me, assuming that you are
about what happened using tense opera- real for me. This may be expressed equival-
tors. Nonetheless, presentism is not without ently as the assumption that the relation
its problems. Are there not past tense truths of being real-for is transitive:
about individuals who no longer exist, for
1. If x is real for y & y is real for z, then x is
example, that Lincoln was wise and wore a
real for z.
beard? But how can there be such truths if
Lincoln no longer exists to be a constituent Putnams other assumption is that in the
of propositions about him? On this question, context of Special Relativity, the presentists
Prior bites the bullet and says there are no core thesis that x is real iff x is present
singular truths about objects that no longer should be reformulated as x is real for y iff
exist, but only general truths it was once x is present for y and the latter in turn as
the case that there was a man who was x is simultaneous with y in the frame of y:
President during a civil war, etc., and who
2. Presentism implies: x is real for y iff x is
wore a beard. Other presentists find some
simultaneous with y in the frame of y.
presently existing entity for past-tense
truths to be about for example, the From 1 and 2, it follows that for presentists,
haecceity being Lincoln, a property that the simultaneity relation we have just men-
exists even if Lincoln does not, and which tioned is transitive:
was formerly co-instantiated with the prop-
3. Presentism implies: if x is simultaneous
erty of being wise.
with y in the frame of y & y is simul-
What some regard as the fatal blow for pre-
taneous with z in the frame of z, then x
sentism comes from the Special Theory of
is simultaneous with z in the frame of z.
Relativity. The theory is often presented as
resting on two postulates, the relativity of uni- According to Special Relativity, however,
form motion and the constancy of the speed
4. The relation in 3 (which Putnam calls
of light. Uniform motion is motion at a con-
simultaneity in the observers frame) is
stant speed in a constant direction. The first
not transitive.
postulate tells us that no experiment can
determine that an object is in a state of That is because if you pass right by me at
absolute uniform motion, from which it is a high relative speed, there will be events
often concluded that it makes no sense to simultaneous with you in your frame that are
ascribe uniform motion. (If two objects are not simultaneous with me in my frame,
moving uniformly relative to each other, it even though at the moment of passing, you
is as correct to say that one is moving and are simultaneous with me in my frame.
the other at rest as vice versa.) The second Putnam concludes that presentism is false,
postulate tells us that whether an observer and that I should acknowledge as real events
is moving towards or away from a beam of belonging to your present even though they
light, the lights speed with respect to the do not belong to mine.
observer will be the same. Einstein showed If presentists do not wish to accept this
that when these two postulates are com- conclusion, how should they respond to
bined, many surprising consequences fol- Putnams argument? There are three main
low, including the relativity of simultaneity: options. One is to reject the transitivity of
two events that are simultaneous in one the real-for relation, as advocated by Sklar;
observers frame of reference may be suc- in effect, this is to make reality itself as rela-
cessive in anothers frame, with no way of tive as simultaneity. A second is to reject
saying that either frame is uniquely correct. Putnams construal of x is present for y as
Putnam has offered an argument against x is simultaneous with y in the frame of y;
presentism based on Special Relativity and alternative relativistic reconstruals of the
two other assumptions. One assumption present-for relation have been canvassed
(which Putnam calls the principle of no by Hinchliff and Sider. The third is to ques-
privileged observers) is that what is real tion Special Relativity, as has been done by

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 82

s p a c e and time
Prior. This last response may strike some as metaphysics. An entrenched axiom is that
an audacious denial of physics to make no one can change the past. If we could
room for metaphysics, but it need not be travel to the past, why could we not change
that. It will probably not have escaped it, even in paradoxical ways such as by
the readers notice that insofar as Special killing ones grandfather or infant self?
Relativity says there is no such thing as abso- Science fiction writers sometimes take pains
lute uniform motion not just that it is to have their characters leave the past
undetectable by any experiment it ven- undisturbed; for example, they view dino-
tures beyond physics into philosophy. One saurs from magically suspended walkways
who questions the theory may be question- so as to leave no footprints. But of course
ing its verificationist auxiliary assumptions the mere presence of the time traveler as an
rather than anything that physics alone observer would constitute a change in the
can teach us. past if he had not been there the first (and
only) time around. Therefore, in consistent
time travel tales, the traveler always made
i s t i m e tr avel p o ssib le?
his visit the visit does not change the past,
This question turns in part on the issues in but was always part of it. (As Lewis has it, a
rows 3, 4, and 5. temporal stage of the traveler was perman-
The physics of the last century is sometimes ently present at the scene. Lewiss stage
thought to imply an answer of yes, for two view explains how it is possible for the
main reasons. First, the Special Theory of traveler to interact with his infant self: such
Relativity is sometimes thought to imply interaction occurs between stages of the
eternalism, as discussed above, and the same person that are contemporaneous in
eternalist view encourages us to take time external time but one later than the other
travel seriously. If the assassination of JFK is in personal time.) Because his actions are
there, several decades prior to us on the already woven into the past, a time traveler
time line, why couldnt we go there and cannot kill his grandfather or his infant
witness it? (Conversely, presentism is some- self; in history as it was, grandfather lived and
times thought to rule out time travel, on the the traveler failed to kill him, if he tried.
ground that if the past and the future are not This way of preserving the past from
there, there is literally nowhere to go.) change may arouse fears of fatalism. If in fact
Second, the General Theory of Relativity grandfather lived to sire my father, am I not
is now believed to imply the possibility of fated to fail in my attempts to kill him?
closed timelike curves, which might be And if in history as it happened, I emerged
exploited by time travelers. Einsteins field from a time machine in 1920 that I enter
equations enable one to calculate the space- (entered? will enter?) in 2020, am I not
time structures induced by various con- fated to enter the time machine in 2020,
figurations of matter, and in 1949, Gdel or at least at some time? To do otherwise
showed that there are possible configura- would be to do something at variance with
tions of matter that would generate closed past truth. In reply, some argue that time-
timelike curves temporal paths along which travel arguments for fatalism add nothing to
an event can precede other events which more general arguments for fatalism based
precede itself. An object part of whose lifeline on applying the law of bivalence to the
lay along such a curve could (in a sense) future, such as the following:
visit its own past. Interestingly, Gdels own
conclusion from his discovery was quite dif- 1. It was either true yesterday that I would
ferent: he thought real time could not violate push the nuclear button tomorrow or
the irreflexivity of precedence, so he took true yesterday that I would not.
the possibility of loops in time to show that 2. In the former case, I must push the but-
time is ideal in something like Kants sense. ton tomorrow
If permitted by physics, travel to the past 3. In the latter case, I must not push it.
may nonetheless be forbidden by logic or 4. Either way, only one course is open to me.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 83

subst anc e
A common reply to such Aristotelian wor- Prior, A.N.: Past, Present, and Future (Oxford:
ries is that all that follows from the supposi- Clarendon Press, 1967).
tion that it was true yesterday that I would Reichenbach, H.: The Philosophy of Space and
push the button tomorrow is that I will Time (New York: Dover, 1958).
push it, not that I must. It could be main- Sider, T.: Four-Dimensionalism: An Ontology
tained similarly that although in 2020 I of Persistence and Time (Oxford: Oxford
certainly will enter the time machine from University Press, 2001).
which I emerged in 1920, it is not true that Sklar, L.: Space, Time, and Spacetime (Berkeley:
I must. So a good case can be made that University of California Press, 1974).
time travel imposes fatalistic constraints on Contains discussions of Newton, Leibniz,
time travelers only if Aristotelian argu- and Poincar.
ments from bivalence impose fatalistic con- Sklar, L.: Up and Down, Left and Right,
straints on us all. So which is it, freedom for Past and Future, Nos 15 (1981), 111
time travelers or fate for us all? Space and time 29; repr. in LePoidevin and MacBeath
do not permit an answer to this question (1993).
here. Van Cleve, J.: If Meinong Is Wrong, Is
McTaggart Right? Philosophical Topics
See also the az entries on antinomies; 24 (1996), 23154.
change; continuant; fatalism; principle Van Cleve, J. and Frederick, R., ed.: The
of verifiability; smart, j.c.c.; space and Philosophy of Right and Left: Incongruent
time, temporal parts; zeno of elea. Counterparts and the Nature of Space
(Dordrecht: Kluwer, 1991).
Willliams, D.C.: The Myth of Passage,
b i b l i og rap hy
Journal of Philosophy 48 (1951), 45772.
Abbott, E.A.: Flatland (New York: Dover, Yourgrau, P.: Gdel Meets Einstein (Chicago:
1952; originally published in 1884). Open Court, 1999).
Broad, C.D.: Scientific Thought (New York:
james van cleve
Harcourt, Brace, 1920).
Hinchliff, M.: A Defense of Presentism in a
Relativistic Setting, Philosophy of Science
67, suppl. (2000), S57586. Substance
Kant, I.: Critique of Pure Reason, trans.
N. Kemp Smith (New York: St. Martins I Introduction
Press, 1965; originally published in 1781). In one metaphysically salient sense of the
LePoidevin, R. and MacBeath, M., ed.: The term substance, a substance is an indi-
Philosophy of Time (Oxford: Oxford Univer- vidual thing. From a commonsensical per-
sity Press, 1993). spective, it appears that the extension of
Lewis, D.: The Paradoxes of Time Travel, substance in this sense includes inan-
American Philosophical Quarterly 13 imate material objects, e.g., pieces of gold,
(1976), 14552; repr. in LePoidevin and mountains, and statues, as well as living
MacBeath (1993). things, e.g., people, frogs, and trees. (Note
McTaggart, J.: The Unreality of Time, in The that since a compound substance is a unified
Nature of Existence, vol. II (Cambridge: whole, its parts must stand in some
Cambridge University Press, 1927), sufficiently robust unifying relation to one
ch. 33; repr. in LePoidevin and MacBeath another, e.g., some appropriate causal or
(1993). functional relation; if there are simple (or
Markosian, N.: How Fast Does Time Pass? basic) substances, they do not have any
Philosophy and Phenomenological Research detachable parts, see part/whole.) A
53 (1993), 82944. belief in the existence of such individual
Putnam, H.: Time and Physical Geometry, substances is at core of our folk ontology.
Journal of Philosophy 64 (1967), 240 Moreover, various scientific theories seem
7. to be committed to their existence. The

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 84

s u b s t a nce
concept of an individual substance figures entail the existence of a quantity of material
prominently in Aristotles seminal work stuff of some kind.
in metaphysics and in much subsequent The existence of individual substances
important work in the field. It is this concept other than inanimate material objects and
that is the focus of this essay. living organisms is controversial. However,
Aristotles term primary ousia has often allowing for the possibility of such sub-
been translated as substance (or as primary stances, including non-physical substances,
substance) a practice which has caused con- it is extremely plausible that any conceivable
siderable confusion. This translation can substance is either spatially extended, spa-
be misleading, since although one ordinary tially located, or living (in a broad intuitive
meaning of substance is an individual sense of living). For example, spatially un-
thing, e.g., an inanimate material object or extended or spatially un-located substances
living organism, this is not what Aristotle which have thoughts, e.g., Cartesian souls,
means by primary ousia. A more accurate would qualify as living in virtue of their
and less misleading translation of primary having mental life, even if they lack bio-
ousia is primary being (or fundamental logical or physical life (see soul), whereas
entity, or basic entity). In the Categories apparently immaterial physical objects such
Aristotle argued that the primary beings as point-particles and mass-less extended
are individual things, e.g., living things, physical objects would have spatial location
and that essences are secondary beings. and/or spatial extension. (Hence, given the
However, in the later work, the Metaphysics, highly plausible assumption that, necessar-
he changed his view about primary beings, ily, life is either a physical process or a
and seems to have concluded that that the mental one, it is extremely plausible that
primary beings are forms, rather than indi- any conceivable substance either has spatial
vidual things. In the Metaphysics, Aristotle extension, spatial location, or thought.)
famously conceived of an individual thing According to Spinoza, there exists one and
as, in some sense, a combination of form only one individual substance, identical
and matter (see matter/form). Even if there with the universe, and this substance is
exists a technical usage of the term sub- neither a physical substance nor a Cartesian
stance in which it means primary being, soul; still, in Spinozas view, this substance
this is a different meaning than the more ordi- has both thought and spatial extension.
nary sense, that of individual thing.
But, according to another ordinary sense II The Analysis of Substantiality
of the term substance, a substance is a
quantity of material stuff of some kind, e.g., a In this section, we shall elucidate what we
quantity of gold, iron, oak, or lamb. But it is mean by an analysis of the concept of an
one thing to say that there exists a quantity individual substance, and then discuss the
of material stuff of some kind, and quite important further notion of the degree to
another to say that there exists an indi- which a philosophical analysis is ontologically
vidual substance, even if this individual neutral (see analysis).
substance is composed of a quantity of stuff We begin with what we mean by an
of the kind in question. For example, it is one analysis or analytical definition of a concept
thing to say that Mary has 50 pounds of or attribute, F-ness. Such an analysis provides
lamb, and quite another to say that Mary a set of conditions, SC, such that: (i) an
has a lamb that weighs 50 pounds. After all, a items (xs) satisfying SC is logically or
lamb necessarily possesses a certain form metaphysically necessary and sufficient for
and unity which a quantity of lamb need xs being F, and (ii) necessarily, if x is F,
not possess. Furthermore, it seems possible then xs being F can be explained by xs sat-
for there to be an individual substance which isfying SC. In this sense, it can be said that
has no proper parts, e.g., a non-physical an analytical definition of F-ness explicates
soul or a point-particle; yet, the existence F-ness. However, if being F is a part of being
of individual things of these sorts does not C, then xs being F cannot be explained by xs

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 85

subst anc e
satisfying SC on pain of vicious circularity. neutral with respect to universals, or to
In such a case, the proposed analytical Cartesian souls, or to the existence of God.
definition of F-ness is fatally flawed; e.g., the If an alternative analysis does not have
proposal to explicate what is just as what these entailments, and so is ontologically
conforms to just laws suffers from this sort of neutral with respect to universals, souls,
flaw. Circularity of this kind is vicious and God, then, to that extent, the second
because nothing can be explained by itself. analysis is more ontologically neutral than
Hence, necessarily, any purported or can- is the first analysis. Of course, it may be the
didate analytical definition that involves case that comparisons between competing
this sort of conceptual circularity fails to analyses are not completely straightforward.
satisfy condition (ii) for being an analytical It may happen, for example, that analysis
definition, above, and should be rejected. A1 is ontologically neutral with respect to
Applying this schema to substantiality, Fs and Gs, and not with respect to Ms and
let F be replaced by substance. It follows that Ns, while analysis A2 is ontologically neu-
in order to provide an analysis of being a sub- tral with respect to Ms and Ns, but not with
stance, an analytical definition must pro- respect to Fs and Gs. Many other permuta-
vide a set of conditions, SC, such that (i) an tions are possible. But at least sometimes,
items (xs) satisfying SC is logically or we will be able to say that one analysis is
metaphysically sufficient and necessary for more ontologically neutral than another. In
xs being a substance, and (ii) necessarily, if any case, one should be aware of the sorts
x is a substance, then xs being a substance of ontological commitments assumed by
can be explained by xs satisfying SC. any analysis.
A further important feature of philosoph- It is plausible to say, we believe, that the
ical analyses is to degree to which they are more ontologically neutral an analysis
ontologically neutral. The following Principle is, the better; more precisely, that all other
of Ontological Neutrality clarifies this notion: things being equal, analyses having a
higher degree of compatibility with the
(PON) An analysis, A, is ontologically
existence of entities of various Categories
neutral with respect to an ontological
are to be preferred, so long as the entities in
kind K (or to an entity E) =df. The
question are not known to be unintelligible,
adequacy of A does not entail either that
and plausible views about the nature, exist-
Ks exist or that Ks do not exist (or that E
ence conditions, and interrelationships of
exists or that E does not exist).
entities belonging to those categories are
By the adequacy of an ontological analysis, assumed. Why should this be so? Because
we mean that the analysis does not conflict which kinds of entities, and which entities,
with the data for that analysis. For exam- actually or possibly exist, is often a matter
ple, if one were trying to analyze what a of philosophical controversy. Witness the
concrete entity is, then ones analysis eternal debate over the existence of univer-
should imply that what intuitively are con- sals between realists and nominalists. Hence,
crete entities are concrete, and that what if one can analyze, say, the concept of sub-
intuitively are not concrete entities are not stance, without thereby being committed
concrete. (We shall ignore here the more either to the existence or non-existence of uni-
complicated situation that arises when no versals, then that is preferable, other things
analysis can be formulated that is in this being equal, to analyzing this concept in
sense adequate to the data, so that we have such a way as to be committed to the
to choose among proposed analyses none of existence or non-existence of universals.
which is entirely adequate.) It follows from This principle about ontological neutrality
PON that if in order to be adequate, a given seems to us just to be a special case of
analysis entails, for example, that univer- Ockhams Razor (see ockham). It also seems
sals do or do not exist, or that Cartesian to us likely that there are further principles
souls do or dont exist, or that God does for evaluating the ontological neutrality
or does not exist, then it is not ontologically of philosophical analyses, but we shall not

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 86

s u b s t a nce
attempt to provide a complete statement of A sympathetic reading of this attempt to
them in this article. analyze substance is that Aristotle is saying
In section IV, we shall defend a version that among entities, only individual sub-
of an independence analysis of the concept stances are able to persist through intrinsic
of substance which is ontologically neutral change. Hence, Aristotles analysis of sub-
with respect to a large variety of metaphys- stance in terms of change should be under-
ical entities: absolute and relational space stood as follows:
and time, space-time, universals, tropes, sets,
(D1) x is a substance =df. x is capable of
numbers, propositions, events, boundaries,
persisting through intrinsic change.
privations (see space and time; universals;
trope; class, set, collection; proposition, In the Categories, Aristotle lists other cat-
state of affairs; event theory; boundary) egories of being, for example, times, places,
and, among substances, living organisms, qualities, relations, and kinds. Note that
atoms, artefacts, and so forth. Other con- it does not seem plausible that such entities
temporary philosophers have offered compet- cannot persist through relational change,
ing versions of an independence analysis of as Aristotle appears to have noted. For
substance, for example, Lowe (2006) and example, at one moment a particular place
Chisholm (1996). Could there be more might be occupied by a body, while at
than one adequate analysis of substance? another time not. However, it does seem to
We see no a priori reason to rule out such be the case that entities of these sorts can-
a possibility. One measure of acceptability, not persist through intrinsic change, since
however, and one that ought not to be they cannot undergo intrinsic change (see
ignored, but is often ignored, is the degree to extrinsic/intrinsic).
which such competing analyses are onto- Nevertheless, there seem to be at least
logically neutral. two fairly plausible counterexamples to D1.
The first is an atomic body, that is, a phys-
III Historical Views of Substance ically indivisible body. Such substances do
not seem capable of undergoing intrinsic
The concept of an individual substance, change indeed, that was one of the reasons
thing, or object has held a very prominent for the first atomists, Democritus and
place in the history of metaphysics, perhaps Leucippus, to postulate such beings (see
because it holds such a prominent place in atomism; presocratics). Current atomic
our ordinary conceptual scheme. theory also regards its fundamental particles
In this section, we shall survey several in this way. Thus, if intrinsically unalterable
important approaches to analyzing the atoms are possible, then D1 fails to provide
notion of an individual substance. Among a logically necessary condition for some-
substance realists, there are independence, things being a substance.
inherence, change, and substratum theor- The second counterexample to D1 is pro-
ists. Also important to consider are those who vided by boundaries. For example, when a
would reduce substances to items belong- rubber ball bounces, its surface changes its
ing to some other ontological category, shape. Hence, if there are things like surfaces,
and those who argue for their elimination and surfaces can undergo intrinsic change,
altogether. then D1 fails to provide a logically sufficient
Aristotle, in the Categories, offers this condition for somethings being a substance.
account of substance in terms of change: Each of the preceding counterexamples
to D1 involves a kind of entity that Aristotle
It seems most distinctive of substance that did not include in his ontology. Hence,
what is numerically one and the same is able Aristotle could reply that there are no
to receive contraries. In no other case could such counterexamples. This points out how
one bring forward anything, numerically Aristotles D1 is not an ontologically neutral
one, which is able to receive contraries. analysis of the concept of substance: it is not
(Complete Works, Vol. I, p. 7) compatible with an ontology that allows for

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 87

subst anc e
the possible (or actual) existence of either asymmetry of the said-of relation, whereby
atomic, intrinsically unchangeable bodies, substance-kinds are said-of substances, but
or of boundaries such as surfaces. Especially not vice versa, establishes the intended
in the former instance, this seems to be a asymmetry of dependence that Aristotle has
serious problem for D1. in mind, whereby substance-kinds depend on
Aristotle provides a second account of substances, but not vice versa.
substance in the Categories: Similar difficulties attend the claim that,
because certain beings are in substances,
A substance that which is called a sub-
such beings asymmetrically depend upon
stance most strictly, primarily, and most of
those substances. Furthermore, it is not
all is that which is neither said of a subject
clear that on any reasonable understanding
nor in a subject, for example, the individual
of the in-relation employed in D2, sub-
man or the individual horse. (Complete
stances cannot be in anything. For exam-
Works, Vol. I, p. 4)
ple, it seems perfectly natural to assert that
This account of substance, then, seems to a particular body is in space and time.
analyze the notion of substance as follows: Aristotles attempt to analyze the concept
of substance in terms of the said-of-relation
(D2) x is a substance =df. x can be
and the in-relation seems to have arisen
neither said of nor in a subject.
from certain grammatical features of proper
The basic idea behind D2 is supposed to be names for individual substances. Such
that individual things or substances do not terms can function only as subjects in sen-
stand in certain relations of dependence to tences, and never as predicates. That this fact
other things, while things in other onto- about grammar can be used somehow to
logical categories do stand in certain depend- analyze the notion of substance while
ence relations to (at least) substances. For implying that substances are asymmetric-
example, Aristotle thinks that in the propo- ally independent of all other categories of
sition, Socrates is a man, the kind, Man, is said being is, however, an error. If substances do
of Socrates, implying that Man depends in enjoy this sort of independence, and it has
some sense on Socrates. He also thinks that been a persistent theme in metaphysics that
in the proposition, Socrates is hungry, the a correct analysis or understanding of sub-
quality, Hunger, is in Socrates, implying stance will have this implication, then we
that Hunger depends in some sense on must seek a different analysis of substan-
Socrates. tiality than D2.
One problem for the idea that D2 estab- In the later Metaphysics, Aristotle defends
lishes that substances possess a unique kind his hylomorphic account of substance,
of independence can be seen by looking at the according to which a substance is a com-
said-of relation. According to Aristotle, bination of form and matter. On one
what can be said-of substances are kinds interpretation, this is just a useful way of
(which Aristotle also calls secondary distinguishing, in the case of compound
beings), that is, the species and genera bodies, between the structure of the body and
under which a substance falls. And given his its constituent stuff. Such an analysis is
theory of universals, no substance-kind level-relative. If Aristotle meant to say that
exists unless it is instantiated by one or there could be pure (or prime) matter, stuff
more substances. Hence, given Aristotles without form, then this is of questionable
ontology, the existence of a substance-kind coherence. He is also ambivalent about
entails the existence of a substance, so that the possibility of the existence of pure form.
it might be said that substance-kinds In any case, Aristotles hylomorphism seems
depend on substances. On the other hand, no incompatible with the possible existence of
substance can exist unless it instantiates immaterial souls.
certain substance-kinds, so it also might be Descartes sought a different independ-
said that substances depend on substance- ence analysis of the concept of substance.
kinds. Thus, it is not at all clear that the For example, at one point he states,

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 88

s u b s t a nce
The answer is that the notion of substance is Spinozas definition presents many difficult
just this that it can exist all by itself, that problems of interpretation, but on the face of
is without the aid of any other substance. it, appears to analyze substance in terms of
(Philosophical Writings, Vol. II, p. 159) some sort of conceptual independence, with the
idea being that what is conceptually inde-
Hence, Descartes seems to be endorsing
pendent is also metaphysically independ-
the following analysis of the concept of
ent. Spinoza thought that his definition
implied that there was only one substance,
(D3) x is a substance =df. x can exist Nature, and that this substance exists
without the aid of any other substance. necessarily. There appear to be at least two
serious criticisms of Spinozas analysis of
This obviously wont do, since to try to
substance. First, it fails to account for the data
analyze the notion of a substance in terms
that any successful analysis must account
of being capable of existing without the aid
for. In this case, Spinozas analysis implies,
of any other substance, is viciously circular.
contrary to the data, that atoms, living
Moreover, D3 implausibly implies that if
organisms, and finite inanimate compound
God exists, then only God is a substance for
bodies are not substances only the universe
no created substance can exist without the
is. Thus, Spinoza has not succeeded in ana-
aid of God.
lyzing the ordinary concept of a substance;
At another point, Descartes avoids the
rather, he has substituted a radically revi-
circularity of D3 with the following statement:
sionary notion of his own. (This criticism
By substance, we can understand nothing applies as well to D3, above.) Second, it is
other than a thing which exists in such a way not clear that even the universe or nature
as to depend on no other thing for its existence. satisfies Spinozas definition, since in order to
(Philosophical Writings, Vol. II, p. 210) conceive of the universe, it seems, one must
conceive of one or more of the attributes
The implied analysis of the concept of sub-
of nature, e.g., extension.
stance is the following:
More recent independence analyses of
(D4) x is a substance =df. x exists and x the concept of substance attempt to con-
depends on no other entity for its existence. form largely to our intuitions about what
entities are substances while capturing a
D4 seems to avoid the circularity of D3, but
more complex sense in which substances
has problems of its own. The main one is
uniquely possess some sort of independence
that no entity is independent of every other
(e.g., Hoffman and Rosenkrantz, 1994;
entity. For example, for any entity, x, there
Lowe, 2006).
is a property, y, such that x has y essentially,
Some philosophers have tried to analyze
and thus depends on y in the sense of entail-
the concept of substance in terms of being
ing its existence. Another problem is that
a subject in which properties inhere. The
a compound body, which is a substance,
idea is that there are properties, and then
depends on its parts in the same sense.
there are things in which properties inhere,
Therefore, D4 does not appear to provide
namely, substances. For example, Descartes
a logically necessary condition for some-
seems to be embracing this theory when
things being a substance.
he says,
Spinoza is another proponent of an inde-
pendence theory of substance. His famous Substance. This term applies to every thing in
definition of substance reads as follows: which whatever we perceive immediately
resides, as in a subject, or to every thing by
By substance, I understand that which is
means of which whatever we perceive exists.
in itself and is conceived through itself; in
(Philosophical Writings, Vol. II, p. 114)
other words, that, the conception of which
does not need the conception of another thing The inherence theory, however, fails to
from which it must be formed. (Ethics and provide a sufficient condition for some-
Selected Letters, p. 31) things being a substance, for every entity

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 89

subst anc e
is a subject for its properties, and not only tropes that comprise the bundle. Another
substances. is to avoid difficulties that seem to derive
Realizing this, some philosophers have from the modal properties of the bundles
embraced the substratum or bare particu- and from their identity conditions. For ex-
lar theory of substance, according to ample, if a bundle is a (special kind of )
which a substance is a concrete individual collection of tropes, then since collections
that has no properties in itself, but instead have their parts essentially, how can a sub-
serves as that in which the properties stantial bundle undergo qualitative (or even
of ordinary objects inhere in some sense. A relational) change?
ball, on this theory, is not a substance, In addition to debates over the nature
but rather a whole constituted by a sub- or analysis of the concept of an individual sub-
stance/substratum and certain properties. stance, metaphysicians have differed over
(Alternatively, the ball is a substance, the kinds of individual substances
constituted by a substratum and certain that there are or could be. A familiar con-
properties the most effective criticism of sub- troversy of this sort is the one between
stratum theories applies to both versions.) materialists, dualists, and idealists. Another
Some have attributed this theory to aspect of this issue is, among material objects,
Descartes and/or Locke, and among more whether or not compound bodies exist,
recent philosophers, the substratum theory whether or not inanimate compound bodies
has been defended by Bergmann and, exist, and whether or not artefacts exist.
at one point, Russell. Van Inwagen, for example, has denied
An apparently devastating criticism of the reality of inanimate compound bodies
any sort of substratum theory is this: it is of any sort (while affirming the reality of
incoherent to postulate the existence of atomic bodies and organisms), and he has
something that lacks any properties. Nor challenged those who assert their existence
does the substratum theorist actually refrain to provide a satisfactory principle of unity
from attributing any properties to substrata, for such objects. Hoffman and Rosenkrantz
since he says that substrata are concrete, that attempt to do so both for inanimate com-
properties subsist or inhere in them, and so pound bodies and organisms, though not
forth. for artefacts (Hoffman and Rosenkrantz,
A final type of theory of substance is the 1997). Lowe (2006) and Thomasson (2007),
bundle theory. This, unlike the preceding on the other hand, defend the view that
theories, is a reductionist theory of substance, artefacts, understood as genuine substances,
that is, it implies that substances are aggre- belong in our ontology.
gates of entities belonging to another onto-
logical category. We shall concentrate here IV An Analysis of Substantiality
on the bundle theory that holds substances
to be aggregates of concrete attributes or All individual substances belong to the
tropes. Proponents of this sort of theory ontological category of Substance. In a
defend an ontology devoid of both universals broad sense, ontological categories are the
and irreducible substances a simplifying more general kinds of entities which (for all
move that they regard as a major strength we know) could exist. Examples of such
of the theory. Bundle theorists include Russell categories and sorts of entities which might
at a later stage of his career, Ayer, Hume, belong to them are the following: Place
Herbert Hochberg, and Castaeda. A (e.g., a volume of space), Time (e.g., an
recent and novel version of the bundle instant), Event (e.g., a process), Trope (in
theory that tries to distinguish between the sense of a concrete quality, e.g., the par-
those attributes essential to a substance and ticular wisdom of Socrates), Boundary (e.g.,
those accidental to it has been defended a surface), Privation (a concrete entity such
by Simons (1994). Bundle theories face as a hole, gap, or shadow), and Collection (in
several challenges. One is to explicate the rela- the sense of an arbitrary sum of any concrete
tion(s) that is (are) supposed to unify the entities, e.g., the Moon + the Empire State

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 90

s u b s t a nce
Building + Mount Everest). The foregoing people have an intuitive observational con-
examples of categories are species of Concrete cept of a material object, an observational con-
Entity. On the other hand, examples of cat- cept which does not presuppose the concept
egories which are species of non-concrete of a material substance. According to this
or Abstract Entity are Property (e.g., intuitive observational concept, a material
Wisdom), Relation (e.g., Betweenness), object or body is an entity which has certain
Proposition (e.g., that 2 + 2 = 4), Set (e.g., perceivable characteristics, including at
{ } ), and Number (e.g., 7). Intuitively, the least certain basic spatial characteristics,
foregoing species of concrete and abstract which can exist unperceived, and so forth.
entities are peers in the sense that all of them Similarly, people have available to them an
are at the same level of generality. We call this intuitive concept of a (Cartesian) soul as a
level of generality Level C, assuming a hier- non-spatial entity which has certain mental
archical tree-like taxonomy in which Entity characteristics. This intuitive concept does not
(the Level A category) is the summum genus, presuppose the concept of an immaterial or
Concrete Entity and Abstract Entity (the spiritual substance.
Level B categories) are the mutually By means of the aforementioned process
exclusive and exhaustive divisions of this of concept formation, one can see that souls
summum genus, and the various species of and bodies would belong to a common level
Concrete Entity and Abstract Entity are the C category because one can see that souls and
Level C categories (see concrete/abstract). bodies resemble one another in ontologic-
Since the category of Substance is a ally relevant respects. In particular, one
species of Concrete Entity, it is a Level C cat- can see that, like a body, a soul can endure,
egory. But how does one acquire the concept persist through qualitative change, exist inde-
of this Level C category? We address this pendently of other entities of the same kind,
question below. and so forth.
To begin, according to a plausible empiri- Since an immaterial physical object such
cist theory of concept formation, one can as a point-particle or a massless extended
acquire the concept of a genus by perceiving object resembles a body in these ontologically
instances of one or more species of that relevant respects, a physical object of this kind
genus and engaging in a process of abstrac- also would belong to the level C category in
tion. This plausible empiricist theory entails question.
that one may possess the concepts of certain However, people seem to be unable to
species before one possesses a concept of conceive of anything belonging to this level
the genus that subsumes them and this is C category other than a physical object
surely true. This process of concept forma- (including material objects and immaterial
tion involves ones observing certain relevant physical objects), a soul, and a Spinozistic sub-
similarities between the perceived instances stance. This is because we cannot conceive
of the genus while setting aside inessential of anything other than a physical object, a
dissimilarities between them. In particular, soul, and a Spinozistic substance that could
given that material objects or bodies are a endure, persist through qualitative change,
species of substance, one can acquire the exist independently of any other entities of
concept of a substance by abstracting from its kind, and so forth.
ones perceptions of bodies, for example, by In what follows, we seek to revive the
noticing that they are enduring entities, traditional idea that a substance is an inde-
that they persist through qualitative change, pendent or autonomous being. In particular,
that they exist independently of other entities we argue that the notion of a Level C cat-
of the same kind, and so forth, while setting egory can be utilized to analyze the concept
aside inessential observed differences between of substance in terms of a sort of ontological
them such as differences in shape and size. independence which uniquely characterizes
One reason why people can acquire the any possible substance.
concept of a substance via the abstractive Our proposed analysis of the concept of
process from perceptions of bodies is because substance entails that anything that could

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 91

subst anc e
belong to the category of Substance must existence of entities of another Level C cat-
meet certain independence conditions qua egory, for example, properties, in no case is
belonging to that category. In other words, this other category such that it could have
we shall argue that the concept of sub- a single instance throughout an interval of
stance can be analyzed in terms of independ- time. It follows that the category of Substance
ence conditions derived from an entitys satisfies clause (ii) of A.
belonging to a Level C category. Our ana- Clause (iii) of A entails that an entity, x,
lysis, A, stated below, consists of the con- is a substance only if x belongs to a Level C
junction of three independence conditions. category whose instantiation by an item is
independent of any other Level C category
(A) x is a substance =df. x belongs to a Level
(other than two special Level C categories
C category, C1, such that: (i) C1 could
referenced in clause (iii)) being instantiated
have a single instance throughout an
by a part of that item. In general, a part of a
interval of time, (ii) C1s instantiation
physical substance could only be a physical
does not entail the instantiation of
substance or a portion of physical stuff, and
another Level C category which satisfies (i),
a non-physical soul has no parts. Hence, it
and (iii) it is impossible that something
appears to be impossible for a substance to
belonging to C1 has a part which belongs
have a part that belongs to another Level C
to another Level C category (other than the
of the sort in question, for instance, a place,
categories of Concrete Proper Part and
a time, a boundary, an event, a trope, a pri-
Abstract Proper Part).
vation, a property, a relation, a proposition,
In condition (i), by an interval of time we and so on. Accordingly, the category of
mean a non-minimal time. And by C1s hav- Substance seems to satisfy (iii) of A.
ing a single instance throughout an interval A is compatible with either of two
of time, we mean that something instanti- assumptions. On the first, all individual
ates C1 throughout an interval of time, substances have contingent existence: each
and that there is no other instance of C1 in substance could fail to exist. On the second
that interval of time. assumption, there is a single necessarily
Although clause (i) of A entails that there existing substance, G, such as God, a sub-
could be a substance that is independent of stance which could not fail to exist. On
any other substance, it does not entail that either of these assumptions, it is possible for
every substance could be independent of there to be a substance, s, which exists
any other substance. For instance, clause (i) throughout some interval of time, t, without
of A is logically consistent with there being any other substance existing within t. On the
a compound substance that is dependent first assumption there could exist through-
upon its substantial parts. Hence, according out t nothing but a single contingent sub-
to clause (i) of A, an entity, x, (regardless stance. On the second assumption, if G
of whether x is simple or compound), is a exists in time, then there could exist
substance in virtue of xs belonging to a throughout t but a single necessary sub-
Level C category which could have a single stance; and if G exists outside of time, then
instance throughout an interval of time. there could exist throughout t but a single
Clause (i) of A characterizes a substance contingent substance.
in terms of an independence condition However, it might be objected that if
entailed by the instantiability of a certain there is an individual substance, then there
Level C category. must be other substances, namely, the (spa-
Clause (ii) of A entails that an entity, x, tial) parts of the individual substance in
is a substance only if xs instantiation of question. But it is only true that a compound
a Level C category is independent of the substance must be composed of other sub-
instantiation of another Level C category stances. It is possible for there to be a simple
which could have a single instance through- substance that has no other substance as a
out an interval of time. However, although (spatial) part, for instance, a non-spatial soul,
the existence of a substance may entail the a point-particle, an indivisible, spatially

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 92

s u b s t a nce
extended, substance, e.g., a Democritean satisfies the three clauses of A. On the other
atom. Note that an indivisible, spatially hand, it appear that the categories of Event,
extended substance has spatially extended Time, Place, Trope, Boundary, Collection,
parts. However, these parts cannot exist Property, Relation, Proposition, Set, and
independently of the whole of which they are Number could not have a single instance
parts. Yet, necessarily, a substance, s, is an throughout an interval of time. Let us
independent being in this sense: s can exist briefly explore the nature of these cate-
independently of any other contingently gories in order to give some indication of
existing substance, s*, unless s* is a proper how this observation can be supported.
part of s or s* helped generate s. Since a Consider first the categories of Property and
spatially extended proper part of an indi- Trope. Necessarily, either an abstract prop-
visible substance fails to satisfy this inde- erty, or a concrete trope, is an entity that
pendence requirement, such a proper part stands in lawful logical or causal relations to
does not qualify as an individual substance. others of its kind. For example, the existence
Rather, it is just a concrete proper part (of a of squareness (or of a particular squareness)
substance). Such an insubstantial proper entails the existence of straightness (or of
part would be an instance of the special a particular straightness). Similar argu-
Level C category of Concrete Proper Part. The ments apply to the categories of Relation,
wording in clause (iii) of A that excludes Proposition, Set, Number, and so on.
the category of Concrete Proper Part from With respect to the category of Place,
consideration accommodates this possibility necessarily, if space exists, then it has an
of an individual substance that has an intrinsic structure that it is compatible with
entity of another Level C category as a part. the occurrence of motion. This entails that,
One sort of part in addition to a spatial part necessarily, if space exists, then space con-
is a temporal part. Clearly, it is at least pos- tains at least two places.
sible for there to be an enduring substance In the case of the category of Time, neces-
that does not have another shorter-lasting sarily, if time exists, then it has an intrinsic
substance as a (temporal) part or sub-stage. structure that is compatible with creation,
(In contrast, necessarily, a temporally extended destruction, qualitative change, or relational
event has other shorter-lasting events as change. It follows that, necessarily, if time
temporal parts, stages.) Still, argu- exists, then there are at least two times.
ably, there could be a temporally extended With regard to the category of Boundary,
substance that does have other shorter- necessarily, every boundary is spatial or
lasting substances as temporal parts, e.g., temporal in character. The existence of
a four-dimensional physical object in a four- a boundary entails the existence of an
dimensional spacetime continuum. But, extended, continuous space or time which
possibly, there is an enduring, indivisible, contains infinitely many extended places
physical particle in three-dimensional space or times. Moreover, necessarily, whatever
(not four-dimensional spacetime) which is bounded has a dimension lacked by its
does not have another shorter-lasting, indi- boundary, e.g., a dimension of thickness,
visible, physical particle as a (temporal) area, length, or duration. Thus, necessarily,
part or sub-stage; or possibly, there is an if there is one (spatial or temporal) bound-
enduring non-spatial soul which does not ary, then there are infinitely many other
have another shorter-lasting soul as a (tem- spatial or temporal boundaries.
poral) part or sub-stage. Thus, it is possible Consider next the category of Event.
that throughout an interval of time, t, there Necessarily, an event that occurs over an
exists an indivisible substance and no other interval of time is a process. Necessarily, a
substance, for example, just one enduring process involves other sub-processes that
indivisible particle, or just one enduring are themselves events. Hence, necessarily,
non-spatial soul. if an event occurs over an interval of time,
On the basis of the preceding discussion, then there is another event that occurs
we conclude that the category of Substance within that temporal interval.

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 93

subst anc e
Finally, consider the category of concrete category entails the instantiation of another
entity, Collection. Necessarily, if a collec- Level C category that satisfies clause (i) of A,
tion, c1, exists throughout an interval of namely, the category of substance. After all,
time, t, then c1 has at least two parts, x and there cannot be a sense-datum unless there
y, both of which exist throughout t. In that is a perceiving substance. It follows that the
case, it appears that there must be a shorter category of Sense-Datum does not satisfy
time, t*, which is a sub-time of t and which clause (ii) of A. Moreover, there is no other
is a part of another collection, c2, for example, Level C category which satisfies A and
a shorter-lasting collection either composed which could be instantiated by a sense-
of t* and x, or composed of t* and y. Hence, datum. Thus, clause (ii) of A has the desir-
necessarily, if a collection exists throughout able implication that a sense-datum is an
an interval of time, then it appears that insubstantial entity (see sensa).
there is another collection which exists within Finally, consider the Level C category of
that interval of time. Privation. In this context, by a privation we
This suggests that the category of Collec- mean a concrete entity which is an absence
tion fails to satisfy clause (i) of A. However, or lack of one or more concrete entities, and
A also implies that collections are not sub- which is wholly extended between two or
stances in virtue of their failure to satisfy more bounding concrete entities, or else
clause (iii), a clause that requires that it is wholly extended between two or more
impossible for an entity of a Level C category bounding parts of a single concrete entity.
has as a part an entity of another Level C A privation in this sense is an insubstantial
category (with the exception of two special concrete entity. (So, a negative abstract
categories which are irrelevant here). After entity, e.g., the proposition that there are
all, something that belongs to a collection is no centaurs, does not qualify as a privation
a part of that collection, and it is evidently in the relevant sense.)
possible for something that belongs to a col- It seems that the category of Privation
lection to be an entity of a Level C category satisfies clause (i) of A. Consider, for example,
other than the category of Collection, e.g., an the possibility of there being nothing but
entity such as a substance, an event, or two temporally separated flashes and the
a place. period of darkness, d, between them. We
In sum, it appears that there could not may assume that in this possible situation
be just one entity of any of the foregoing d is the only privation throughout the
Level C categories (throughout an interval interval of time in question.
of time.) Moreover, in each case there is On the other hand, it can be argued that
no other Level C category which could be the category of Privation fails to satisfy clause
instantiated by an entity belonging to the (ii) of A for the following two reasons. First,
category in question, and which could have the category of Substance satisfies clause
a single instance throughout an interval of (i) of A and this Level C category is other than
time. Hence, (clause (i) of) A seems to have the category of Privation. Second, neces-
the desirable consequence that an entity sarily, if there is a privation, then there is a
that belongs to any of these categories is substance, e.g., a substance which flashes, a
insubstantial. Clauses (ii) and (iii) of A enable substance which is perforated, a substance
this proposed analysis to deal with insub- which is shadowed or which casts a shadow,
stantial entities of various other kinds. and so on; though, clearly, there could be
For example, suppose for the sake of a (basic) substance without there being a
argument that a purple after-image is an privation.
insubstantial entity of the irreducible category Still, some have claimed that there could
Sense-Datum. On this supposition, a sense- be a flash without there being a substance
datum is not an event, a property, a trope, that flashes, and thus it is controversial
a boundary, and so on. If so, then an after- whether the existence of a privation requires
image belongs to the Level C category of the existence of a substance. Fortunately, A
Sense-Datum. But the instantiation of this is neutral with respect to this controversy,

9780631199991_4_P1001.qxd 1/12/09 3:04 PM Page 94

s u b s t a nce
since, in any event, clause (iii) of A entails that Stoothoff, and Dugald Murdoch (Cam-
privations are not substances. To see this, bridge: Cambridge University Press, 1984).
note that privation, d, has as parts certain Hoffman, J. and Rosenkrantz, G. S.: How
(lightless) periods of time within d. These to Analyze Substance: A Reply to
parts belong to the category of Time, a Schneider, Ratio 20 (2007), 130 41.
Level C category other than the category of Hoffman, J. and Rosenkrantz, G. S.: Sub-
Privation. It follows that the category of stance Among Other Categories (Cambridge
Privation fails to satisfy clause (iii) of A. and New York: Cambridge University
In addition, there is no other Level C cat- Press, 1994).
egory which satisfies A and which could be Hoffman, J. and Rosenkrantz, G. S.: Sub-
instantiated by a privation. Hence, clause (iii) stance: Its Nature and Existence (London
of A has the desired consequence that a pri- and New York: Routledge, 1997).
vation is not a substantial entity. Locke, J.: An Essay Concerning Human
It appears that A provides a logically Understanding, ed. Roger Woolhouse
necessary and sufficient analysis of the con- (London: Penguin Books, 1997).
cept of substance in terms of a kind of onto- Loux, M.: Metaphysics: A Contemporary Intro-
logical independence. In the light of the duction, 3rd edn. (London and New York:
foregoing discussion, it also appears that Routledge, 2006).
this analysis is ontologically neutral to a Loux, M.: Substance and Attribute (Dord-
high degree, that is, compatible to a high recht: Reidel, 1978).
degree with the existence of entities belong- Lowe, J.: The Four-Category Ontology: A Meta-
ing to various intelligible categories, given physical Foundation for Natura Science
plausible views about the nature, existence (Oxford and New York: Oxford University
conditions, and interrelationships of entities Press, 2006).
belonging to those categories. Simons, P.: Farewell to Substance: A Dif-
ferentiated Leave-taking, Ratio N.S. XI
See also the az entry on substance. (1998), 23552.
Simons, P.: Particulars in Particular
Clothing: Three Trope Theories of Sub-
b i b l i og rap hy stance, Philosophy and Phenomenological
Research 54 (1994), 55376.
Aristotle: The Complete Works of Aristotle:
Spinoza: The Ethics and Selected Letters, trans.
The Revised Oxford Translation, ed. Jonathan
Samuel Shirley, ed. Seymour Feldman
Barnes (Princeton, NJ: Princeton Univer-
(Indianapolis, IN: Hackett Publishing
sity Press, 1984).
Company, 1982).
Bergmann, G.: Realism (Madison, WI: Uni-
Thomasson, A. L.: Ordinary Objects (Oxford
versity of Wisconsin, 1967).
and New York: Oxford University Press,
Campbell, K.: Abstract Particulars (Oxford:
Blackwell, 1990).
van Inwagen, P.: Material Beings (Ithaca,
Chisholm, R.: A Realistic Theory of Categories
NY and London: Cornell University Press,
(Cambridge and New York: Cambridge
University Press, 1996).
Descartes, R.: The Philosophical Writings of joshua hoffman
Descartes, trans. John Cottingham, Robert gary s. rosenkrantz

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 95

Part II

Metaphysics from A to Z

A Companion to Metaphysics, Second Edition Edited by Jaegwon Kim, Ernest Sosa, and Gary S. Rosenkrantz
2009 Blackwell Publishing Ltd. ISBN: 978-1-405-15298-3
9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 97

Abelard, Peter (10791142) French the items be defined in the same way. He
philosopher, logician and theologian. Born claimed that objects and the matter of which
near Nantes in France in 1079 Abelard they were composed were the same in the
studied logic in his youth under Roscelin, former sense but not in the latter.
notorious for his antirealist interpretation
of logic, and went on to become the most
wri t i ngs
sought-after teacher of logic in Europe.
Beyond logic Abelard involved himself in Dialectica (Dialectic), ed. L.M. de Rijk, 2nd
theological debates, and his interpretation of ed. (Assen: Van Gorcum, 1970).
the Holy Trinity, a topic which called forth Logica ingredientibus (Logic for Beginners),
his best work on the concept of sameness, was in Peter Abaelards Philosophische Schriften,
condemned twice by the church. Abelards fascicules 13, ed. B. Geyer, in Beitrge
life was a stormy one including the much zur Geschichte der Philosophie des Mittel-
celebrated romance with and marriage to alters, Vol. 21, fascicules 14 (Mnster:
Hlose, his subsequent castration by thugs Aschendorff, 191927).
hired by her uncle, and a bitter series of Logica nostrorum petitioni sociorum
disputes with William of Champeaux over (Logic in Response to the Request of Our
universals. Friends), in Peter Abaelards Philosoph-
It was the topics of universals and identity ische Schriften, fascicule 4, ed. B. Geyer,
that elicited Abelards main efforts in meta- in Beitrge zur Geschichte der Philosophie
physics. While arguing that no universal, des Mittelalters, Vol. 21, fascicules 14
i.e., nothing common to many, is any real (Mnster: Aschendorff, 191927). (A second
thing, that is has an existence independent edition by the same publisher appeared in
of the mental and linguistic activities that 1973.)
involve signification of things in the world, Theologia Christiana ed. E.M. Buytaert in
Abelard proposed that nevertheless there Corpus Christianorum Continuatio Media-
are status which serve as objective signi- evalis 1112 (Turnholt, Belgium: Brepols,
ficates of predicates that are true of many 1969).
distinct things. He gave the status much the martin m. tweedale
same treatment as he proposed for dicta,
which are the significates of sentences and
the primary bearers of truth and falsity. abstract see concrete/abstract
They are not things in the world, not even
psychological or linguistic things, but they accident see essence/accident
can exist and be known objectively.
Taking off from remarks by Aristotle in the acquaintance Acquaintance is a central
Topics Abelard distinguished different sorts notion in Russellian metaphysics, as well
of identity and distinctness. Most important as Russellian epistemology and philosophy
is the contrast between sameness in essence of language. Russell distinguishes know-
and sameness in property. The former means ledge by acquaintance from knowledge by
that the items in question have all their description, and characterizes the former
parts in common; the latter requires that as follows.

A Companion to Metaphysics, Second Edition Edited by Jaegwon Kim, Ernest Sosa, and Gary S. Rosenkrantz 97
2009 Blackwell Publishing Ltd. ISBN: 978-1-405-15298-3
9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 98

a c q u a i ntance
(1) We shall say we have acquaintance the conjunction of his view that objects of
with anything of which we are directly acquaintance include universals and his
aware, without the intermediary of any view that objects of acquaintance, are such
process of inference or any knowledge of that their existence cannot be doubted (see
truths (Russell, 1959, p. 46, italics in Russell, 1959, chs. 9 and 10). James Van
original). Cleve has mentioned that any philosopher
(2) it is possible, without absurdity, to doubt holding an indubitability thesis will need
whether there is a table at all, whereas to formulate it so as to avoid the conclusion
it is not possible to doubt the sense- that we have indubitable knowledge of
data (Russell, 1959, p. 47). The table anything that is in fact philosophically con-
is not an object of acquaintance, but troversial. But the way to do this in the
the sense-data are, and this condition is present case seems to be, for example, to
supposed to provide a general contrast replace such claims as Russells that it is not
between objects of acquaintance and possible to doubt the sense-data (Russell,
other things. 1959, p. 47) with claims to the effect that
(3) All our knowledge, both knowledge it is not possible to doubt that one seems
of things and knowledge of truths, rests to see something blue, or that one is in pain,
upon acquaintance as its foundation etc. This no longer involves reference to
(Russell, 1959, p. 48). any object of acquaintance whose existence
(4) Russell also specifies objects of acquain- cannot be doubted.
tance by extension. For Russell, only an object of acquaintance
We have acquaintance in sensation can be the referent of a logically proper name,
with the data of the outer senses, i.e., a name that refers directly, without
and in introspection with the data of describing, and whose sole semantic function
what may be called the inner sense is to stand for it referent. By his principle
thoughts, feeling, desires, etc.; we of acquaintance, Every proposition which
have acquaintance in memory with we can understand must be composed wholly
things which have been data either of of constituents with which we are acquainted
the outer senses or of the inner sense. (Russell, 1959, p. 58, italics in original).
Further, it is probable, though not Donnellan offers a useful formalization of
certain, that we have acquaintance this notion of a constituent, when he says that
with Self, as that which is aware of if, and only if, Socrates is a constituent of
things or has desires towards things. the proposition expressed by the sentence
In addition to our acquaintance Socrates is snub-nosed, this proposition
with particular existing things, we might be represented as an ordered pair
also have acquaintance with . . . consisting of Socrates the actual man,
universals (Russell, 1959, pp. 512, of course, not his name and the predicate
italics in original). (or property, perhaps), being snub-nosed
(Donnellan, 1974, p. 225).
These four specifications cannot be Russell grants that his principle of
assumed to coincide. What, if anything, is acquaintance entails that much of a per-
the foundation of our knowledge and what, sons language is private (in the sense that
if anything, is known directly are, of it is logically impossible for anyone else to
course, themselves matters of philosophical apprehend the propositions expressed by
controversy. And the second specification the speaker) as well as ephemeral (in the
has its own special problem, since universals sense that it is logically impossible for any-
and sense data, far from being indubitable, one to apprehend at time t2 the proposition
are just the sorts of entities whose existence he expressed at t1. (For Russell on ephemer-
many philosophers doubt. Russell recog- ality, see Russell, 1956, pp. 2014.) But
nizes that many people doubt or deny the Russell overstates the extent of privacy
existence of universals, but he does not seem his principle of acquaintance requires. He
to recognize the problem this fact raises for says

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 99

action theory
When one person uses a word, he does not Donnellan, K.S.: Speaking of Nothing, The
mean by it the same thing as another Philosophical Review 83 (1974), 331; repr.
person means by it . . . It would be abso- in Naming, Necessity and Natural Kinds,
lutely fatal if people meant the same things ed. S.P. Schwartz (Ithaca, NY: Cornell
by their words . . . the meaning you attach University Press, 1977), 216 44.
to your words must depend on the nature Russell, B.: The Philosophy of Logical
of the objects you are acquainted with, and Atomism, in his Logic and Knowledge, ed.
since different people are acquainted with R.C. Marsh (London: George Allen and
different objects, they would not be able to Unwin, 1956), 177281.
talk to each other unless they attached Russell, B.: The Problems of Philosophy (New
quite different meanings to their words. York: Oxford University Press, 1959).
We should have to talk only about logic. felicia ackerman
(Russell, 1956, p. 195)
By Russells own lights, this claim is
action theory Action theory deals with
overstated, since he does not limit objects
that concept of action that applies only to
of acquaintance to sense data, oneself, and
beings who have wills. The questions it
entities of logic, such as sets. He also
addresses include: (1) what is the mark of
includes universals. Thus, on the principle
action? (2) How should actions be individu-
of acquaintance, we would not have to
ated? (3) What makes an action inten-
talk only about logic in order to attach
tional? (4) Is freedom of action compatible
the same meanings to our words. We could
with determinism? (5) What makes true
also talk about blueness, roundness, etc.,
the sort of explanation peculiar to action,
and we could discuss such propositions as
namely, that the agent did the action for
the proposition that blue is more like purple
certain reasons?
than either is like orange. But this quali-
fication is unlikely to assuage the doubt of
opponents of the principle of acquaintance, t he mark of ac t i on
especially since the argument Russell offers What distinguishes an action from other
for the principle is drastically inadequate. sorts of events of which a person may be
He says the subject, such as sensations, perceptions,
it is scarcely conceivable that we can feelings, unbidden thoughts, tremblings,
make a judgment or enter a supposition reflex actions? Two main sorts of answer
without knowing what it is that we are have been offered. According to one, what
judging or supposing about. We must marks an event (say, a movement of ones
attach some meaning to the words we body) as an action is something extrinsic to
use, if we are to speak significantly and not the event, namely its having been caused in
utter mere noise; and the meaning we the right sort of way by the subjects desires
attach to our words must be something (or intentions) and beliefs (see Goldman,
with which we are acquainted. (Russell, 1970, ch. 3, and Davidson, 1980, essay
1958, p. 58, italics in original) 1). The right sort of causal connection is
important, because, for example, the fact
Of course this is not really an argument. that a desire to have another drink results
It begs the question (see Ackerman, 1987). in the subjects falling down does not make
that event an action. This sort of view seems,
b i b l i og rap hy however, not to cover spontaneous actions
whose occurrence is not explained by any
Ackerman, F.: An Argument for a Modified antecedent motives of the agent.
Russellian Principle of Acquaintance, The other kind of account finds the mark
in Philosophical Perspectives, Vol. 1, Meta- of an action in the intrinsic nature of the
physics, ed. J. Tomberlin (Atascadero, CA: event, rather than in something external
Ridgeview, 1987), 50112. to it. The idea is that an event is an action

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 100

a c t i o n theo r y
because it is, or begins with, a special sort of action is normally thought of as a more
event. Some hold that the special event is concrete entity than an exemplifying of a
an occurrence of a quite special sort of cau- property (so that (1) and (2) describe the
sation, where an event is caused, not by same concrete action in terms of different
another event, but by the agent herself; an intrinsic properties), one action can be a
agent is the only sort of enduring thing proper part of a distinct action in which
that can be the subject of this special kind something that is not an action, namely, a
of causation (see Taylor, 1966; Chisholm, consequence of the first action, is an addi-
1976). Others hold that the special event is tional part (so that (3) picks out a larger
mental; for some, what makes it special is its action of which (1) (2) is merely the initial
functional role (Davis, 1979, chs. 12), and part, and (4) picks out a still larger action
for others, it is its phenomenal character of which (4) is merely the initial part) (see
(Ginet, 1990, ch. 2). In actions that go on Thomson, 1977; Ginet, 1990, ch. 3).
to become voluntary bodily exertions this
event is a willing (or volition) to act. Some
t he int ent i onal i t y of ac t i on
(for example, Hornsby, 1980) think that
the content of this volition may be anything Smith swung the racket intentionally and
that the agent was trying to do in the in so doing inadvertently hit his opponent
action. Reflection on our experience of vol- with it. Smiths hitting his opponent with
untary bodily exertion suggests, however, the racket was not intentional, but it could
that there is in it something to be called have been. Whether an action is intentional
volition that is quite distinct from intention or not often makes a big difference for the
and the content of which is limited to the sort of evaluation it deserves. What deter-
immediately present exertion of the body mines whether an action is intentional or
(see Ginet, 1990, ch. 2). not (under a given description)? This can
be divided into two questions, depending on
whether or not the action description in
t h e i n d ividuatio n o f actio n
question is basic. An action description, of the
Suppose that just now I moved my right form Ss A-ing, is basic just in case there
index finger and thereby pressed a key and is no other, non-equivalent action description,
thereby put a character on the computer Ss B-ing, such that it is true that S A-ed
screen. Each of the following is a description by B-ing. With respect to a basic description,
of an action I performed: (1) I moved a it is plausible to hold that whatever makes
finger; (2) I moved my right index finger; an event it fits an action also makes it in-
(3) I pressed a key; and (4) I put a char- tentional under that description. (This is
acter on the screen. How many different especially plausible if the basic descriptions
actions do these four descriptions pick out? attribute mental acts of volition.)
One view holds that they pick out four The question with respect to non-basic
different actions; because an action is an descriptions is more difficult. One might
exemplifying of an action property by an think that it would have been sufficient
agent at a time, and our four descriptions for Smiths hitting his opponent with the
express four different action properties (see racket intentionally that he intended of his
Goldman, 1970). Another view holds that voluntary bodily movement that by it he
they all describe the same action in terms would cause the racket to hit his opponent.
of different properties; my action was just But suppose he was too far from the oppon-
the minimal thing by (or in) doing which ent for the swing to hit as he intended;
I did the things attributed to me by all the however, his grip loosened as he swung and
descriptions (see Davidson, 1980, essay 3; the racket flew out of his hand and hit
Hornsby, 1980). (On some views this basic the opponent. We cannot then say that his
action is the bodily movement, on others it hitting his opponent with the racket was
is a volition.) Between these extreme views, intentional. Perhaps it is sufficient for his
one may take the position that, although an actions being intentional if he caused the

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 101

action theory
racket to hit his opponent in the way he From (2) it follows that (3) one can have it
intended. But this appears not to be neces- open to one at a given time to perform a cer-
sary. Suppose Smith stumbled slightly as he tain action a, only if, for any truth entirely
swung, causing him to hit the opponent about the past, p, one has it open to one
slightly below the spot he intended to hit; in to make it the case that: p and one does a.
this case, though he did not hit him in just From (1), (3), and determinism it follows
the way he intended, it seems that he still hit that one never has it open to one to do any-
him intentionally. In light of such difficulties thing other than what one actually does.
(there are others), it is clear that it will not Suppose that at 2 oclock it seemed to me to
be a simple matter to devise a satisfactory be open to me to raise my hand then, but I
necessary and sufficient condition for an did not do so. If determinism is true then there
actions being intentional under a description. is a true proposition, p, which is entirely
(For one complex proposal, see Ginet, 1990, about the past relative to 2 oclock and such
ch. 4.) that it follows from the laws of nature that:
if p then I did not raise my hand at 2
oclock. From (3) it follows that it was open
f r e e a c tio n and d eter minism
to me to make it the case that I did raise my
I have freedom of action at a given time just hand at 2 oclock only if it was open to me
in case more than one alternative action to make it the case that: p and I raised my
is then open to me (see the extended essay hand at 2 oclock. But, given (1), it could
on free will). We continually have the not have been open to me to make that
impression of having more than one altern- proposition true, for it contradicts the laws
ative action open to us (indeed, a great of nature. Therefore, if determinism is true
many alternative actions normally seem (and so also are (1) and (3)), then (contrary
open to us: consider all the different ways to my impression) it was not open to me
that, as it seems to me, I could next move to do a at t.
my right hand). determinism is the thesis This argument is obviously valid and so
that, given the state of the world at any philosophers who resist the conclusion that
particular time, the laws of nature (see law determinism is incompatible with freedom of
of nature) determine everything that hap- action (and many do) must reject either (1)
pens thereafter down to the last detail. Some or (2). Arguments against (2) are possible,
philosophers have argued that our impres- but the more popular, and perhaps more
sion of freedom is always an illusion if promising, line is to attack (1) (see Fischer,
determinism is true or if, though false, 1988; Lewis, 1981). (1) could be put this
it fails to be false in the right places (see way: if it follows from the laws of nature
van Inwagen, 1983, ch. 3; and Ginet, 1990, that if p then q, then it is never in anyones
ch. 5). (This last disjunct is important be- power to make it the case that: p and not-q.
cause, although contemporary physics may This principle seems appealing because it
give us good reason to think that determin- seems that we ordinarily feel compelled to
ism is false, it does not give us good reason make inferences in accordance with it. For
to think it is false in the right places: as example, if I know that Xs brain state at t
yet we do not even know precisely what is such as to nomically necessitate Xs being
the right places are.) unconscious for at least one minute after t,
The essential premises of the argument then that seems good enough to infer that
that determinism is incompatible with free- it is not open to X at t to voluntarily raise
dom of action are two: (1) No one ever has his or her arm during the minute after t. To
it open to him or her to make true a propo- account for the apparent cogency of such
sition that contradicts the laws of nature. an inference, while rejecting (1), one might
(2) No one ever has it open to him or her to suggest that what really underlies its valid-
determine how the past was, i.e., to make true ity is not (1) but a more complex principle,
one rather than another of contrary pro- something like the following: if p nomically
positions that are entirely about the past. necessitates that X does not act in a certain

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 102

a c t i o n theo r y
way at t, and the necessitation does not run in believing any reasons explanations.
through Xs internal processes in the way that Perhaps we need not know what the relevant
it does in normal seemingly free action, then causal laws are in order to be justified in
it is not open to X to act in that way at t. giving a reasons explanation in a particular
This more complex principle, says the critic case. We must, however, be justified in
of (1), will account for all the acceptable believing that there are laws that govern
inferences that seem to invoke (1). But will the case (whether or not their contents are
it? Imagine a possible world where deter- known to us); and it might well be doubted
minism is true and Martians control all of whether there is any case for which we are
Xs actions over a long period through con- justified in believing even this.
trolling Xs normal psychological processes The nomic view nevertheless has a strong
of motivation and deliberation. If you are appeal for many philosophers. This may be
inclined to think this would mean that X has because they find it hard to see what else,
no more freedom of action than a puppet, if not a nomic connection, could make a
then, it seems, you are inclined to operate genuine explanatory connection between
with (1) and not just the more complex motives and action. This is a fair question,
principle (for the latter would not justify which one must answer if one wants to
that inference). make a good case against the nomic view.
One must specify a condition that is clearly
sufficient for the explanatory connection,
t h e n a tur e o f actio n exp lained
does not imply a nomic connection, and
b y r e a so ns
is easy to know is present (especially for the
Typically when one acts one has motives or subject). Here is a sketch of how one might
reasons for acting in the way one does and try to do that (see Ginet, 1990, ch. 6 for a
one acts in that way for those reasons. For fuller exposition).
example, my reason for opening the window Suppose that concurrently with my action
was that I wanted to let out the smoke. I of opening the window I remembered my
opened the window in order to let out the antecedent desire to rid the room of smoke
smoke, that is, because I intended thereby to and I intended of that action I was engaging
let out the smoke. in that I would thereby satisfy that desire.
The main metaphysical issue concerning These conditions seem clearly sufficient to
explanations of this sort is whether they make the explanatory connection between
are essentially nomic, that is, whether the the desire and the action, to make it true that
truth of one of them entails that the case I opened the window because I wanted to rid
be subsumable under causal laws which the room of smoke; and just as clearly they
dictate that whenever motives of the same seem to be compatible with there being no
sort as those the explanation cites occur in true causal laws which dictate that always
sufficiently similar circumstances they (the a desire of that same sort in sufficiently sim-
motives and the relevant circumstances) ilar circumstances must produce the same
causally necessitate an action of the same sort of action. That is, they give a non-nomic
sort (see Ayer, 1946; and Davidson, 1980, sufficient condition for a reasons explana-
essays 1, 11, for expressions of this view). The tion of an action. (Of course the obtaining of
nomic view of reasons explanations would a non-nomic sufficient condition does not
tend to be confirmed if we knew (or had rule out the possibility of a nomic suffici-
good evidence for) the relevant laws in most ent condition, perhaps even for the same
cases of true reasons explanations. But we explanation of the same action.)
do not. Indeed, it may be that, as yet, there That reasons explanations need not be
is no true reasons explanation of any action nomic is important for the view that freedom
for which anyone knows causal laws that of action is incompatible with determinism.
govern the explanation. Of course, this igno- Otherwise, that view would be committed
rance does not show that the nomic view to the counterintuitive proposition that
is wrong or that, on it, we are not justified no free action (one for which there were

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 103

adverbial theory
alternatives open to the agent) could have inferred that he is related, through his
a reasons explanation. experience, to a red, square sense-datum.
The sense-datum theory leads to a number
of perplexing questions. For example, can
b i b l i og rap hy
sense-data exist unsensed? Can two persons
Chisholm, R.M.: The Agent as Cause, in experience numerically identical sense-data?
Action Theory, ed. M. Brand and D. Walton Do sense-data have surfaces which are not
(Dordrecht: Reidel, 1976), 199211. sensed? What are sense-data made of? Are
Davidson, D.: Essays on Actions and Events they located? Historically, the desire to avoid
(Oxford: Clarendon Press, 1980). questions like these was one reason for the
Davis, L.: Theory of Action (Englewood Cliffs, development of the adverbial theory.
NJ: Prentice-Hall, 1979). This position that having a perceptual
Fischer, J.M.: Freedom and Miracles, Nos experience is a matter of sensing in a certain
22 (1988), 23552. manner rather than sensing a peculiar
Ginet, C.: On Action (New York: Cambridge immaterial object is arrived at by reflect-
University Press, 1990). ing on the fact that, on standard views,
Goldman, A.I.: A Theory of Human Action appearance, after-images, and so on, cannot
(Englewood Cliffs, NJ: Prentice-Hall, 1970). exist when not sensed by some person. The
Hornsby, J.: Actions (London: Routledge and explanation the adverbial theorist offers for
Kegan Paul, 1980). this fact is that statements which purport
Lewis, D.: Are we free to break the laws?, to be about appearances, after-images, and
Theoria 47 (1981), 11221. so on, are in reality statements about the way
Taylor, R.: Action and Purpose (Englewood or mode in which some person is sensing.
Cliffs, NJ: Prentice-Hall, 1966). Hence, a statement of the general form,
Thomson, J.J.: Acts and Other Events (Ithaca, Person, P, has an F sense-impression, or
NY: Cornell University Press, 1977). P has an F sensation, is reconstructed
van Inwagen, P.: An Essay on Free Will adverbially as, P senses F-ly, or as it is some-
(Oxford: Clarendon Press, 1983). times put, P senses in an F manner. This
carl ginet transformation has a number of gram-
matical parallels. Patrick has a noticeable
stutter, for example, is equivalent to Patrick
actuality see potentiality/actuality stutters noticeably, and Patrick stutters
in a noticeable manner. Similarly Jane
adverbial theory The adverbial theory does a charming waltz, may be transcribed
is, at root, the view that to have a perceptual as, Jane waltzes charmingly. It should
experience is to sense in a certain manner. be obvious that the adverbial view can
Traditionally, the most popular analysis of account for the facts of hallucination and
perceptual experience has been the opposing illusion. If, for example, I am correctly des-
sense-datum theory (see sensa). According cribed as having a visual sensation of some-
to this theory, having a perceptual experi- thing blue then blue in this description
ence amounts to standing in a relation is taken upon analysis to function as an
of direct perceptual awareness to a special adverb which expresses a mode of my sens-
immaterial entity. In particular cases this ing. Hence, my having the sensation does
entity is called an after-image or a mirage not require that there be a blue physical
or an appearance, and, in the general case, object (or anything else for that matter)
a sense-impression or a sense-datum. The in my general vicinity it suffices that I
sense-datum is required, so it is normally sense bluely.
argued, in order to explain the facts of Although the adverbial theory began as,
hallucination and illusion: since a person and is still most strongly associated with, the
can have a visual sensation of a red square, analysis of perceptual experience, it has also
say, even when there is no real red, square been applied elsewhere. For example, it is
object in his general vicinity, it is typically often held by adverbialists that our ordinary

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 104

a l f a r ab i
talk of bodily sensations is misleading, and Chisholm, R.M.: Perceiving (Ithaca, NY:
that in reality there are no such items as pains Cornell University Press, 1957).
and itches to which persons are related Sellars, W.: Science and Metaphysics (London:
when they have a pain or feel an itch. Routledge and Kegan Paul, 1968), 928.
Rather statements about bodily sensations Tye, M.: The Adverbial Approach to Visual
have an underlying adverbial structure. Experience, The Philosophical Review 93
Jones has an intense pain, for example, is (1984), 195225.
analyzed as Jones is pained intensely; Tye, M.: The Metaphysics of Mind (Cambridge:
hence it is about the way in which Jones is Cambridge University Press, 1989).
pained. The motivation for this approach
runs parallel to the one for perceptual michael tye
experience: countenancing pains and other
sensory objects in our ontology generates
a host of philosophical puzzles. For example, Alfarabi [al-Farabc] (c.870950) Islamic
are pains really located about the body as logician, metaphysician, political philosopher,
our ordinary pain talk suggests? If so, then also wrote commentaries on Aristotles
presumably they are material objects. Why, logical treatises and expositions of Platos
then, are they never revealed by surgical and Aristotles philosophies.
examination of the appropriate limbs? Can Alfarabi was the first to raise the question
pains exist in parts of the body without of how the philosopher writing in Arabic
their being felt? Can two persons ever which has no copula, can do logic and supply
feel one and the same pain? All these precise vocabulary for the Greek concept
puzzles dissolve once the adverbial view of being. He proposes to use derivatives of
is adopted. wjd (to find) for all the functions of to be,
Some philosophers have argued that the in a stipulative fashion, including the most
adverbial theory can even be extended to general sense of being (Shehadi, 1982,
the analysis of belief and desire discourse. pp. 45 51).
Thus, having the belief that snow is white, Existents are divided by Alfarabi into the
say, is not a matter of bearing the having possible and the necessary. In the case of
relation to a particular belief, but rather possible beings, existence is not a property
a matter of believing in a certain way. and cannot be part of their essence (see
Whether this extension is defensible, and essence/accident; essence and essential-
indeed whether the adverbial theory is viable ism). Asked whether Man exists has a
anywhere, depends ultimately on how the predicate, Alfarabi replied that for the logi-
theory is further spelled out. Recent work (see cian, exists is a predicate in the proposition.
Tye, 1989) has supplied a clear semantics and But it is not a predicate to the investigator
metaphysics for the theory with the result into the nature of things. However, in the
that the adverbial approach is no longer case of the First, existence is Its essence, for
open to the charge that it is just a rather It is the being necessary through itself.
trivial grammatical transformation without In Islamic philosophy Neoplatonic (see
any real constraints. Indeed, once fully elu- Neoplatonism) emanationism gets its first full
cidated, the adverbial theory is seen to be a statement by Alfarabi. Islamic Neo-platonists
very powerful and well-founded approach were influenced by an Arabic translation of
which has the resources to answer all the a pseudo Theology of Aristotle which was in
more obvious objections. fact a summary of sections of Plotinus
Enneads, as well as by a translation of the
Liber de causis.
b i b l i og rap hy
The First is one, uncomposed, and bey-
Ducasse, C.J.: Moores Refutation of Ideal- ond human knowledge. From its activity
ism, in Philosophy of G.E. Moore, ed. of thinking itself emerges the First Intel-
P.A. Schilpp (Chicago: Northwestern lect which thinks itself as well as its source.
University Press, 1942), 22551. The emanations proceed until the Tenth

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 105

anal ysi s
Intellect, each intellect with its correspond- Recently, Alston has defended a realist
ing cosmic sphere. conception of truth according to which (1)
Of special interest is Alfarabis trans- a statement is true if and only if what the
formation of Aristotles active intellect into statement says to be the case actually is
a separate entity between humankind and the case, and (2) truth is an important or
the First, one of the separate substances significant feature of reality. It often matters,
above the terrestrial sphere. While it still and we do care, whether our beliefs are
makes the knowable known, its cosmo- true. Alston has also defended a form of
logical status prepares the way for the metaphysical realism against a number of
eschatological and mystical roles that it objections including the idea that there is
plays in Islamic philosophy. a unique description of the world, a com-
mitment to the causal theory of reference, and
physicalism. See also experience; realism;
b i b l i og rap hy theories of truth.
Fakhry, M.: A History of Islamic Philosophy
(New York: Columbia University Press, wri t i ngs
Hammond, R.: The Philosophy of Alfarabi and Epistemic Justification (Ithaca, NY: Cornell
its Influence on Medieval Thought (New University Press, 1989).
York: Hobson Book Press, 1947). Perceiving God (Ithaca, NY: Cornell Univer-
Rescher, N.: Al-Farabi: An Annotated sity Press, 1991).
Bibliography (Pittsburgh: University of A Realist Conception of Truth (Ithaca, NY:
Pittsburgh Press, 1962). Cornell University Press, 1996).
Rescher, N.: Al- al-Farabc on the Question: The Reliability of Sense Perception (Ithaca,
Is Existence a Predicate?, in his Studies in NY: Cornell University Press, 1993).
the History of Arabic Logic (Pittsburgh, richard gallimore
PA: University of Pittsburgh Press, 1963)
Shehadi, F.: Metaphysics in Islamic Philosophy analysis Consider the following proposition.
(Delmar, NY: Caravan Books, 1982).
(1) To be an instance of knowledge is to
fadlou shehadi be an instance of justified true belief not
essentially grounded in any falsehood.
(1) exemplifies a central sort of philosoph-
Alston, William P. (1921 ) is an Ameri-
ical analysis. Analyses of this sort can be
can philosopher who has made significant
characterized as follows:
contributions to epistemology, philosophy
of religion, and the realismantirealism (a) The analysans and analysandum are
debate among other areas. Alstons work necessarily coexistensive, i.e., every in-
in epistemology has focused primarily on stance of one is an instance of the other.
Foundationalism, the nature of epistemic (b) The analysans and analysandum are
justification, the internalismexternalism knowable a priori to be coextensive.
controversy, sense perception, and religious (c) The analysandum is simpler than the
epistemology. In philosophy of religion, analysans (a condition whose necessity
Alston has argued that putative perceptual is recognized in classical writings on
experience of GOD is epistemically on a analysis, such as Langford, 1942).
par with putative perceptual experience of (d) The analysans does not have the
ordinary material objects. Alston uses this analysandum as a constituent.
argument along with a detailed account (e) A proposition that gives a correct ana-
of mystical experience, to defend the lysis can be justified by the philosophical
importance of experiential grounds for the example-and-counter-example method,
justification of religious belief. i.e., by generalizing from intuitions about

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 106

a n a l y s is
the correct answers to questions about and reformatory analysis, which seeks to
a varied and wide-ranging series of reduce sloppiness and imprecision by replac-
simple described hypothetical test cases, ing a concept considered in some way
such as If such-and-such were the case, defective with one considered in the relev-
would you call this a case of knowledge? ant way improved. Reformatory analysis
Thus, such an analysis is a philosophical makes no claim of conceptual identity
discovery, rather than something that between analysans and analysandum and
must be obvious to ordinary users of the hence gives rise to no paradox of analysis.
terms in question. Aside from the possibility of paradox,
philosophers have raised various objections
Condition (d) rules out circularity. But since to analysis as a philosophical method. It
many valuable quasi-analyses are partly is a commonplace to object that analysis
circular (e.g., knowledge is justified true belief is not all of philosophy. But, of course, the
supported by known reasons not essentially claim that analysis is a viable method does
involving any falsehood), it seems best to not amount to saying that it is the only
distinguish between full analysis, for which one. Wittgenstein (see Wittgenstein, 1968,
(d) is a necessary condition, and partial especially sects. 3967) has raised objections
analysis, for which it is not. to the atomist metaphysics and epistemology
This core notion of analysis fits the intui- underlying Russellian new-level analysis
tive idea the term analysis suggests, which (see logical atomism; russell). But most of
is that something is analyzed by breaking these objections do not apply to other types
it down into its parts (see Moore, 1903, of analysis. It can also be objected that it is
sects. 8 and 10). But Moore also holds that virtually impossible to produce an example
analysis is a relation solely between con- of an analysis that is both philosophically
cepts, rather than one involving entities of interesting and generally accepted as true.
other sorts, such as linguistic expressions, But virtually all propositions philosophers
and that in a true analysis, analysans and put forth suffer from this problem. (See
analysandum will be the same concept (see Reschler, 1978; Ackerman, 1992a.) The
Moore, 1942). These views give rise to what hypothetical example-and-counterexample
is nowadays generally called the paradox method the sort of analysis (1) exemplifies is
of analysis: how can analyses such as (1) fundamental in philosophical inquiry, even
be informative? Philosophers have proposed if philosophers cannot reach agreement on
various solutions, such as relaxing the re- analyses and often even individually can-
quirement that analysans and analysandum not give full analyses and have to settle for
are the same concept (Langford, 1942), and less, such as one-way conditionals, partially
denying that (1) is genuinely informative circular accounts, and accounts (like that of
to someone who fully grasps the concepts being a game) that are justified in the same
involved (Sosa, 1983). general way as analyses but that are too
Regardless of how this paradox is to be open-ended even to purport to yield neces-
handled, there are types of analysis other sary and sufficient conditions.
than that exemplified by (1). One such type
of analysis involves an analysans and
bibl i ography
analysandum that are clearly epistemically
equivalent and that hence do not raise the Ackerman, F.: Analysis and Its Paradoxes,
paradox discussed here, although they do in The Scientific Enterprise: The Israel
raise a different paradox (see Ackerman, Colloquium Studies in History, Philosophy,
1990). Other types of analyses include new- and Sociology of Science, vol. 4, ed. E.
level analysis, which aims at providing Ullman-Margalit (Norwell, MA: Kluwer,
metaphysical insight through metaphysical 1992b).
reduction (for example, the analysis of sen- Ackerman, F.: Analysis, Language, and Con-
tences about physical objects into sentences cepts: The Second Paradox of Analysis, in
about sense data (see Urmson, 1956, ch. 3), Philosophical Perspectives, vol. 4, Philosophy

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 107

ansc ombe, g.e.m.

of Mind and Action, ed. J. Tomberlin causation to be that of derivativeness, exem-
(Atascadero, CA: Ridgeview, 1990), 535 plified by making a noise, pushing, wetting.
43. Her view that these causal notions do not
Ackerman, F.: Philosophical Knowledge, involve universality or necessity might be
in A Companion to Epistemology, ed. questioned, so she examines different sorts
J. Dancy and E. Sosa (Oxford: Blackwell, of examples, like Feynmans case of a bomb
1992a), 3425. which may be caused to explode by some
Langford, C.H.: The Notion of Analysis in radioactive emission. The absence of neces-
Moores Philosophy, in The Philosophy sitation is irrelevant to the causing of the
of G.E. Moore, ed. P.A. Schilpp (Evanston, subsequent explosion. Anscombe also ex-
IL: Northwestern University Press, 1942), amines the relevance of non-necessitating
319 43. causes to freedom of the will (see the
Moore, G.E.: Principia Ethica (New York extended essay on free will).
London: Cambridge University Press, She has discussed the subject of causa-
1903). tion in several other essays. An important
Moore, G.E.: A Reply to My Critics, in The theme is the different kinds of causal relation
Philosophy of G.E. Moore, ed. P.A. Schilpp (see, for example, 1974a). In Times, Begin-
(Evanston, IL: Northwestern University nings and Causes (1974b), she examines
Press, 1942), 6607. Humes claim that it is logically possible for
Rescher, N.: Philosophical Disagreement, something to begin to exist without a cause.
Review of Metaphysics 22 (1978), 217 She develops an argument of Hobbess to
51. show how judgments about beginnings of
Sosa, E.: Classical Analysis, Journal of existence depend on the application of causal
Philosophy 53 (1983), 695710. knowledge.
Urmson, J.O.: Philosophical Analysis (Oxford: Among the other topics in metaphysics
Oxford University Press, 1956). which she has discussed is that of the self. In
Wittgenstein, L.: Philosophical Investigations, The First Person (1975), she argues that
3rd edn., ed. and trans. G.E.M. Anscombe Descartess view of the self would be correct
(New York: Macmillan, 1968). if I were genuinely a referring expression,
felicia ackerman but that it is not a referring expression.
Metaphysical problems concerning time and
substance are the focus of some of her other
Anscombe, G.E.M. (19192001) G.E.M. essays.
Anscombe is a philosopher of great range,
many of whose important contributions to
philosophy lie in metaphysics and in fields
which substantially overlap metaphysics, Aristotle: The Search for Substance, in
especially philosophy of logic and philosophy G.E.M. Anscombe and P.T. Geach, Three
of mind. Philosophers (Oxford: Blackwell, 1963),
In Causality and Determination (1971) 563.
she questioned a central assumption made Causality and Determination (inaugural
in virtually all philosophical writing about lecture at Cambridge University, Cam-
causation (see the extended essay), namely, bridge, 1971); repr. in (1981), vol. II,
If an effect occurs in one case and a similar 13347.
effect does not occur in an apparently similar Collected Papers (Oxford: Blackwell, 1981;
case, there must be a further relevant differ- Minneapolis: University of Minnesota
ence. The most disparate views of causation, Press, 1981), 3 vols.; papers bearing on
from Aristotles and Spinozas to Hobbess, metaphysics are in Vol. I, From Par-
Humes, and Russells, all accept that cau- menides to Wittgenstein and Vol. II,
sation involves universality or necessity or Metaphysics and the Philosophy of Mind.
both; but Anscombe argues that such views The First Person, in Mind and Language:
cannot stand up. She shows the core idea in Wolfson College Lectures 1974, ed.

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 108

a n s e l m o f canter b ur y, st
S. Guttenplan (Oxford: Clarendon Press, Sweetness and Pleasant Texture, in its own
1975), 4565; repr. in (1981), vol. II, ineffable manner. Nevertheless, God is
2136. supremely simple, omne et unum, totum et
Memory, Experience and Causation, in solum bonum, a being a more delectable
Contemporary British Philosophy, ed. H.D. than which cannot be conceived.
Lewis (London: George Allen and Unwin, God is both the efficient cause of everything
1974[a] ) 1529; repr. in (1981), vol. II, else and the paradigm of all created
12030. natures, the latter ranking as better in so far
Times, Beginnings and Causes, Proceedings as they are less imperfect ways of resem-
of the British Academy 60 (1974[b] ) 253 bling God. Such natures have a teleological
70; repr. in (1981), vol. II, 14862. structure, which is at once internal to them
(a created f is a true (defective) f to the
extent that it exemplifies (falls short of) that
b i b l i og rap hy for which fs were made) and established
Diamond, C. and Teichman, J., ed.: Intention by God. From teleology, Anselm infers
and Intentionality: Essays in Honour of a general obligation on all created natures
G.E.M. Anscombe (Brighton: Harvester (non-rational as well as personal): since
Press, 1979). they owe their being and well-being to God
as their cause, so they owe their being and
cora diamond well-being to God in the sense of having
an obligation to praise Him by fulfilling
their teloi.
Anselm of Canterbury, St. (10331109) Anselms distinctive action theory reasons
Scholastic philosopher and Archbishop that if the telos of rational natures is unend-
of Canterbury, born at Aosta, Italy. Like ing beatific intimacy with God, their powers
Augustine before him, Anselm is a Christian of reason and will have been given to
Platonist in metaphysics (see Platonism). promote that end. Thus, the wills freedom
In the Monologion, he deploys a cosmo- must be telos-promoting, and since sin
logical argument for the existence of the is deviation from the telos should not be
source of all goods, which is Good per se and defined as a power for opposites (the power
thus supremely good, identical with what to sin and the power not to sin), but rather
exists per se and is the Supreme Being. In as the power to preserve justice for its own
Proslogion c.ii, Anselm advances his famous sake (see the extended essay on free will).
ontological proof: namely, that a being a Choices are imputable only if spontaneous
greater than which cannot be conceived (from the agent itself). Since creatures
exists in the understanding, since even a have their natures from God and not from
fool understands the phrase when he hears themselves, they cannot act spontaneously
it; but if it existed in the intellect alone, a by the necessity of their natures. To enable
greater could be conceived which existed in creatures to be just of themselves, God
reality. A parallel reductio in c.iii concludes endowed them with two motivational drives
that a being a greater than which cannot toward the good the affectio commodi, or
be conceived exists necessarily. And in his tendency to will things for the sake of their
Reply to Gaunilo, he offers a modal argu- benefit to the agent itself: and the affectio
ment for Gods necessary existence, based on justitiae, or tendency to will things because
the premise that whatever does not exist is of their own intrinsic value. It is up to the
such that if it did exist, its non-existence creature whether or not to align them (by let-
would be possible. God is essentially what- ting the latter temper the former). Anselms
ever it is other things being equal better motivational theory contrasts sharply with
to be than not to be, and hence living, wise, Aquinass Aristotelian account, but was
powerful, true, just, blessed, immaterial, taken up and developed by Duns Scotus.
immutable and eternal per se; even the par-
adigm of sensory goods Beauty, Harmony, See also god.

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 109

ant i nomi es
writings This embarrassment of riches constitutes
the data for Kants meta-argument in favor
Adams, R.M.: The Logical Structure of
of the critical point of view: instead of aim-
Anselms Arguments, The Philosophical
ing at knowledge of a mind-independent
Review 80 (1971), 2854.
reality, we should abandon the classical
Anselm of Canterbury, Vols. 13 (Toronto
metaphysical enterprise and restrict the
and New York: Edwin Mellor Press,
objects of knowledge to appearances. We can
then see that the antinomies are a product
Henry, D.P.: The Logic of Saint Anselm
of transcendental illusion which arises from
(Oxford: Clarendon Press, 1967).
the temptation to apply the principles that
Hopkins, J.: Anselm of Canterbury, Vol. 4:
constitute the framework for knowledge
Hermeneutical and Textual Problems in the
of phenomenal reality to noumenal reality
Complete Treatises of St. Anselm (Toronto
(see noumenal/phenomenal).
and New York: Edwin Mellen Press, 1976).
Contemporary philosophers do not share
Kane, G.S.: Anselms Doctrine of Freedom of
Kants awe at the cogency of the clash-
the Will (New York and Toronto: Edwin
ing arguments. Indeed, cosmologists and
Mellen Press, 1981).
infinitistic mathematicians dismiss the pros
S. Anselmi. Opera Omnia, ed. F.S. Schmitt
and cons about the extent of space and time
(Edinburgh: Thomas Nelson and Sons,
as amateurish fallacies. However, Kants
194661), vols. IVI.
unconvincing choice of examples does not
marilyn mccord adams undermine the philosophical interest of the
concept of an antinomy. After all, apparent
needs to be relativized to epistemic agents. An
antinomies An antinomy is a pair of antinomy for an eighteenth-century figure
apparently impeccable arguments for oppo- need not be an antinomy for a twentieth-
site conclusions. Obviously, the arguments century thinker.
cannot both be sound because a proposition In any case, there certainly are argu-
and its contradictory must have opposite mentative deadlocks. Recently, a Japanese
truth values. Thus the two appearances group of topologists announced a result that
of cogency are not all things considered contradicted the result of an American group
judgments because conflicting appearances of topologists. Since both proofs involved
cancel out. The challenge posed by an anti- complex calculations, they exchanged proofs
nomy is at the level of adjudication and to check for mistakes. Despite their high
diagnosis. We know that at least one arm of motivation and logical acumen, neither team
the antinomy is fallacious. But which? And has been able to find an error in the others
exactly where does it go wrong? reasoning. The JapaneseAmerican dead-
Antinomy is most closely associated lock is not an antinomy if it is caused by
with Immanuel Kants attack on meta- a slight but subtle slip. The appearance of
physics. In the Critique of Pure Reason, he lays cogency must be due to a deep error not
out parallel arguments literally side by side a mistake due to bad luck or ignorance sur-
to emphasize their utter deadlock. As long mountable by merely mechanical methods.
as we assume that things-in-themselves Metaphysicians have a particular interest
are objects of knowledge, we can mount a in antinomies that turn on false existential
knock-down argument for the thesis that presuppositions. The Barber paradox fea-
the world has beginning in time and a tures a village in which a barber shaves all
knock-down argument for the antithesis and only those people who do not shave
that the world has no beginning. Metaphy- themselves. Does the barber shave himself?
sicians can prove that we are free by expos- First argument: if the barber shaves himself,
ing the absurdity of an actual infinity of then he is a self-shaver. But he only shaves
past events and metaphysicians can dis- those who do not shave themselves. There-
prove our freedom by demonstrating the fore, the barber does not shave himself.
incoherency of a break in the causal order. Second argument: if the barber does not

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 110

a n t i r e alism
shave himself, then he is among the non- Sorensen, R.: Pseudo-Problems (London:
self-shavers. But he shaves all those who do Routledge, 1993).
not shave themselves. Therefore, the barber
roy a. sorensen
does shave himself! The lesson to be learned
from this modest antinomy is that the bar-
ber cannot exist. antirealism By antirealism we mean
More ambitious resolutions of antinomies here semantic antirealism, of the kind
aim at a more dramatic impact on our advanced by Dummett in numerous writ-
ontology or cosmology. The paradox of the ings. The main thesis of semantic antireal-
stone (can God make a stone so large that ism is that we do not have to regard every
He Himself could not lift it?) is used to declarative statement of our language as
disprove Gods existence. The Buddhists determinately true or false independently of
use antinomies to disprove the existence of our means of coming to know what its truth
the self. The Eleatics (see presocratics) and value is. That is, the semantic antirealist
nineteenth-century idealists (see idealism) refuses to accept the principle of bivalence.
deployed antinomies against the assump-
tion that material things exist and that they
are spatially related. ant i real i sm is not a form of
Other antinomies turn on false dicho- i deal i sm or nomi nal i sm
tomies. For example, the old arguments for and Semantic antirealism is to be distinguished
against infinite space tended to assume that from ontological antirealism. Ontological
finite and unbounded were mutually antirealism casts doubt on the existence
exclusive terms. Albert Einsteins applica- of objects. It comes in varying degrees. The
tion of Riemannian geometry makes sense ontological antirealist may doubt the exist-
of a spherical universe that is finite but ence of any objects in the external world
unbounded. So besides subtracting entities (idealism); or, more modestly, doubt the
and relationships from metaphysical sys- existence of the unobservable entities posited
tems, antinomies enrich these systems by by science (van Fraassens constructive
stimulating the discovery of new entities empiricism (1980)); or, more traditionally,
and possibilities. doubt the existence of abstract objects, such
An antinomy cannot prove anything on as numbers (see number), or of universals
its own. Indeed, its internal conflict makes it (nominalism). Semantic antirealism is com-
a paradigm of dialectic impotence. However, patible with both Platonistic (see Platonism)
the meta-arguments that grapple with anti- and nominalistic views about numbers. In the
nomies are powerful tools of metaphysical case of mathematics, G. Kreisels dictum is
inquiry. often stressed: what one is concerned with
is not so much the existence of mathem-
See also aporia; sorites arguments; trans- atical objects, as the objectivity of math-
cendental arguments; Zeno. ematical statements.

b i b l i og rap hy ant i real i sm is compat i ble

wi t h nat ural i sm
al-Asm, S.J.: The Origins of Kants Arguments
in The Antinomies (Oxford: Oxford Univer- Indeed, one might even maintain that it is
sity Press, 1972). a consequence of naturalism. By naturalism
Kant, I.: Critique of Pure Reason (Riga, 1781); we mean the metaphysical view that all
trans. N. Kemp Smith (London: Macmillan, things, events, states and processes are
1933). material or physical. Naturalism asserts
Quine, W.V.: The Ways of Paradox (Cambridge, supervenience, but does not claim reduc-
MA: Harvard University Press, 1976). tionism (see reduction, reductionism).
Sainsbury, R.M.: Paradoxes (Cambridge: Cam- It asserts that all mental, moral, semantic
bridge University Press, 1988). and social facts supervene on material or

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 111

physical facts. The physical facts, that is, (3) Truth is bivalent.
fix the mental, moral, semantic and social
facts. But naturalism does not claim that (4) Grasp of meaning cannot be mani-
psychology, moral theory, semantics or the fested fully in observable behavior.
social sciences can be reduced to physics. On
the contrary, each of these special sciences Both Quine and the antirealist agree on the
is autonomous. Each presents important first premise. Quine holds that meaning
aspects of reality in its own terminology. (via translation) is indeterminate, but that
Indeed, antirealism itself is a theory whose truth is bivalent. The antirealist, by con-
content would be lost were it not formu- trast, holds that meaning is determinate,
lated in its own special terms, terms which but that truth is not bivalent.
defy reduction to physics.
ant i real i sm enjoins a
a n t i r e a lism str esses molec ular, as opposed t o
o b s e r v a ble b ehavio r an hol i stic, theory of meani ng
a s t h e s o ur ce o f meaning The antirealist believes in determinate sen-
The antirealist is centrally concerned with tential contents. He or she adopts a com-
grasp of meanings, or contents (see content); positional approach. One familiar ground
and with the conditions under which for this comes from theoretical linguistics,
speakers and thinkers can acquire such which rightly stresses our recursive, gener-
grasp and display it. It lays great stress on ative or creative capacity to understand new
what have become known as the acquisition sentences as we encounter them. Another
and manifestation arguments. These argu- ground is that the opposing holistic view
ments are used to cast doubt on the claim, (see holism) simply cannot account for lan-
concerning sentences in any given area of guage learning. We do, it would appear,
discourse, that their meanings consist in master language fragments progressively as
verification-transcendent truth conditions. learners, and are able to isolate or excise
For, if they did, so these arguments con- them for theoretical study later on. Meanings
clude, speakers of the language would never of words remain relatively stable under
be able fully to acquire or display grasp increase of vocabulary and during develop-
of meaning. The observable conditions sur- ments in our ability to produce and under-
rounding their discourse, and their own stand more complicated utterances. These
observable behavior, prevent such overly considerations point to a compositional
enriched contents from being grasped and approach.
assigned to sentences. The acquisition and
manifestation arguments, as developed by ant i real i sm is conc erned
Dummett, show most clearly the influence wi t h normativi t y
of the later Wittgenstein on Dummetts
thinking. As we have just seen, the antirealist main-
tains determinacy of meaning. Precision
about contents brings with it commitment to
a n t i r e a lism co ntr asted normative connections among them: their
w i t h qu ineanism justification conditions and their entailments.
One way to understand antirealism is to One of the main aims of antirealism is to
consider how Quine and the antirealist react give an accurate picture of such contents as
to an argument on which they both agree. the speaker or thinker can genuinely grasp
The argument has three premises and a or entertain in thought, and convey in
conclusion that they both reject: language. This means that antirealism has
to have some answer to skeptical problems
(1) Meaning is given by truth conditions. about the objectivity of rule following. For it
(2) Meaning is determinate. is only by conforming to, or keeping faith

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 112

a n t i r e alism
with, rules for the use of expressions that informative claim. Here the situation is
the speaker can claim to have mastered very different from mathematics. For in
their meanings. mathematics, once a statement is proved
it remains proved. In empirical discourse,
however, statements are defeasible. That is,
a n t i r e alism favo r s r efo r mism
they can be justified on a certain amount
r a t h e r than q uietism
of evidence; but may have to be retracted
In particular, the antirealist critique of or even denied on the basis of new evidence
genuinely graspable meanings can be accreting upon the old. (A modern way of
brought to bear on the meanings of the putting this is to say that they are governed
logical expressions of our language: the con- by a non-monotonic logic.) There is also
nectives and the quantifiers. The observable the familiar problem from the philosophy
conditions of their use (especially in math- of science, that no general claim about
ematics) concern the discovery, construc- natural kinds (see natural kind) can ever
tion, presentation and appraisal of proofs. conclusively be proved. At best, such claims
Central features of the use of logical expres- can be conclusively refuted; but no amount
sions in particular, their introduction of humanly accessible evidence can entail
rules serve to fix their meanings. Other them. The combination of defeasibility with
features need to be justified as flowing from this familiar asymmetry between proof
the central features. We can justify the and refutation makes particularly prob-
elimination rules, because these are in a lematic the provision of a satisfactory
certain sense in balance or harmony with antirealist account of meaning for empirical
the introduction rules. But on this model discourse.
of meanings and how one comes to grasp
them, there does not appear to be any
antirealism t ends t o be
justification for the strictly classical rules
pi ecemeal, rat her than gl obal
of reasoning, especially as they concern
negation. There does not appear to be any Most writers on antirealism try to explore
justification for the Law of Excluded Middle its strengths and weaknesses on particular
(either a or not-a) or for the Law of Double areas of discourse: mathematics, statements
Negation Elimination (from not-not-a infer a) about other minds, statements about the
or any of their equivalents. Thus the anti- past, counterfactual statements (see coun-
realist response has been to favor logical terfactuals), and so on. In each area one
reform: crucially, to drop the strictly clas- looks critically at the observational basis on
sical negation rules and opt for intuitionistic which one can acquire grasp of meaning. One
logic. Thus intuitionism is the main form of examines the criterial structure governing
mathematical antirealism (see intuitionism how speakers venture, and are taken at,
in logic and mathematics). When the their words. One tries (if necessary) to deflate
antirealist generalizes from the mathem- any overly realistic classical conception
atical case, with its conditions of construc- of how, in response to each such area of
tive proof, he or she looks for appropriate discourse, a mind-independent region of
conditions of warranted assertability. reality might inaccessibly yet determinately
be. The realist sometimes complains that
the antirealist is guilty of epistemic hubris
t h e c h allenge o f an
in taking the human mind to be the meas-
a n t i r e a list acco unt o f
ure of reality. The antirealist responds by
e m p i r i cal d isco ur se
charging the realist with semantic hubris
In moving to empirical discourse, and in claiming to grasp such propositional
especially statements about other minds, contents as could be determinately truth-
one has to attend closely to the criteria in valued independently of our means of com-
accordance with which one ventures any ing to know what those truth values are.

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 113

a n t i r e a lism is no t a cr ude (4) stress on the compositionality of mean-
f or m o f ver ificatio nism ing, thereby allowing meaningful though
undecidable sentences;
There was an old principle of the logical
(5) advocacy of some kind of logical reform,
positivists (see logical positivism) which,
making ones logic more intuitionistic or
over the years, fell into deserved disrepute.
This was the verificationist principle that
(6) a generally naturalistic metaphysical
every meaningful declarative sentence was,
outlook, and a quietist demurral from
in principle, decidable. That is, in grasping
extreme skeptical misgivings or theses in
its meaning a speaker would have recourse
to a method which, if applied correctly,
would within a finite time yield the correct
verdict as to the truth or falsity of the sen-
mai n alleged weaknesses i n
tence. Despite its emphasis on assertibility
t he antirealist position
conditions, antirealism lays claim to no
such principle.
(1) Alleged failure to do justice to the intu-
ition that the world is robustly independ-
a n t i r e a lism str esses ent of human cognitive faculties;
c om p o s i tio nally (2) alleged failure to appreciate the strength
of independent arguments to the effect
Antirealism stresses, instead of the posi- that translation is indeterminate, that
tivists naive decidability principle, various there can be no firm analytic/synthetic
canonical ways of establishing statements distinction, that meaning (such as it is)
with prominent occurrences of expressions is graspable at best only holistically;
whose antirealistically licit meaning is at (3) alleged failure to appreciate that, in so
issue. (An example of this would be dominant far as meaning is determined (by the
occurrences of logical operators, in the con- antirealists own lights) by the use we
text of their introduction rules.) Various make of our expressions, we should
such expressions could then be combined accordingly accept classical rules of
into a sentence which is meaningful but inference (such as Double Negation
which the antirealistic need not claim is Elimination) as justified by the very use
decidable. The sentence will be meaningful we make of them;
by virtue of the way those expressions are (4) alleged instability in the antirealists
combined within it, and by virtue of their own argumentative strategy: why stop
central meanings as conferred by those special at intuitionism, for example? Why not go
contexts. It is at this point that modern all the way to strict finitism? Why treat
antirealism is crucially influenced by the of decidability in principle rather than
contribution of Frege to logical semantics. feasible decidability?
(5) alleged failure to understand the
s u m m a r y o f main featur es o f semantic contribution of the negation
t h e a n t i r ealist p o sitio n operator in embedded contexts;
(6) alleged failure to appreciate that there
(1) refusal to accept the principle of are, even within the constraints set
bivalence; by the antirealist, resources enough to
(2) behaviorist emphasis on the epistemo- secure the realists grasp of verification-
logy of linguistic understanding: acqui- transcendent propositional contents;
sition and manifestation arguments; (7) alleged failure to appreciate that the
(3) confidence in the determinacy of sen- semantic issue of logical reform is
tence meaning, leading to a molecular independent of the metaphysical and
as opposed to an holistic theory of epistemological issues at the heart of
meaning; antirealism.

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 114

a n t i r e alism ab o ut ab str act enti t i es

w r i t i n gs in the mo der n r ealism See also realism; the extended essay on
v s . a n tir ealism deb ate realism and antirealism about abstract
Michael Dummett put forward his classic
challenge to the principle of bivalence in his
bi bliography
essay Truth. His defense of intuitionistic
logic as the correct logic on an antirealist Dummett, M.A.E. (with the assistance of
construal of mathematics was given in his R. Minio): The Elements of Intuitionism
essay The philosophical basis of intuition- (Oxford: Clarendon Press, 1977).
istic logic. This treatment was amplified in Dummett, M.A.E.: Truth and Other Enigmas
the chapter on philosophical reflections in his (London: Duckworth, 1978).
book The Elements of Intuitionism (1977). He Kripke, S.: Wittgenstein on Rules and Private
explored the implications of antirealism for Language (Oxford: Blackwell, 1982).
statements about the past in his essay The McDowell, J.: Anti-Realism and the Epi-
Reality of the Past. Dummetts essays are stemology of Understanding, in Meaning
collected in his book Truth and Other and Understanding, ed. H. Parret and
Enigmas (1978). J. Bouveresse (Berlin and New York: De
Dag Prawitz has provided an excellent Gruyter, 1981), 225.
exposition and amplification of Dummetts McDowell, J.: Truth Conditions, Verifica-
line of argument in his paper Meaning and tionism and Bivalence, in Truth and
Proofs: On the Conflict Between Classical Meaning, ed. G. Evans and J. McDowell
and Intuitionistic Logic (1977). Crispin (Oxford: Clarendon Press, 1976), 42 6.
Wright has written widely on antirealism Peacocke, C.: Thoughts: An Essay on Content
in mathematics, on statements about the (Oxford: Blackwell, 1986).
past and on statements about other minds. Prawitz, D.: Meaning and Proofs: On the
He has also treated the problems of criteria, Conflict Between Classical and Intui-
defeasibility and the objectivity of rule fol- tionistic Logic, Theoria 43 (1977), 240.
lowing. See his book Wittgenstein on the Smart, J.J.C.: Realism v. Idealism, Philo-
Foundations of Mathematics (1980), and his sophy 61 (1986), 295312.
collection of essays Realism, Meaning and Strawson, P.F.: Scruton and Wright on
Truth (1986). Neil Tennant, in his book Anti-realism, etc., Proceedings of the
Anti-Realism and Logic (1987), has extended Aristotelian Society 77 (1976), 1522.
the antirealist critique and the logical reform Tennant, N.: Anti-Realism and Logic (Oxford:
it arguably entails in favor of the system of Clarendon Press, 1987).
intuitionistic relevant logic. He also explores Wright, C.J.G.: Realism, Meaning and Truth,
antirealism as a consequence of naturalized 2nd ed. (Oxford: Blackwell, 1986).
epistemology. John Mcdowell has pursued Wright, C.J.G.: Truth and Objectivity (Cam-
subtle variations on realistic and antirealistic bridge, MA: Harvard University Press,
themes in his essays Anti-Realism and 1992).
the Epistemology of Understanding (1981), Wright, C.J.G.: Wittgenstein on the Founda-
and Truth Conditions, Verificationism and tions of Mathematics (London: Duckworth,
Bivalence (1970). 1980).
Opposition by the realists has been led neil tennant
most notably by Peter Strawson (1976),
Christopher Peacocke (1986) and J.J.C.
Smart (1986). Saul Kripke gave great antirealism about abstract entities
impetus to the debate about the objectivity see the extended essay on realism and
of rule following with the publication of antirealism about abstract entities
his provocative monograph Wittgenstein on
Rules and Private Language (1982). Kripke aporia An apory is a small set of individu-
adopts an antirealistic construal of content- ally plausible but jointly inconsistent pro-
attribution statements in his sceptical solution. positions. Aporia gained initial popularity

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 115

appearance / reality
from Chisholms demonstration of how to be F and x is F that he had to show
they help to motivate and structure philo- to be untenable. When Descartes, and after
sophical issues. For instance, he regiments the him, most thinkers of the modern era,
problem of ethical knowledge with a set struggled with the skeptics challenge, the
containing the following three members: threat posed by that challenge was the pos-
sibility that that same gap was too great, that
(1) We have knowledge of certain ethical
no reliable evidence about reality was ever
furnished by what appeared in experience.
(2) Experience and reason do not yield
In part inspired by that challenge, one
such knowledge.
empiricist strain (see empiricism), strangely
(3) There is no source of knowledge other
echoed in a late flowering of rationalism,
than experience and reason.
concludes that what appears to the well-
To avoid inconsistency, thinkers need to functioning mind (in perception or in
reject at least one member of the set. Thus reasoning) is, and must be, the real, and it
the skeptic denies (1), the naturalist rejects must be just as it appears. Found in both
(2), while the intuitionist argues against Berkeleys and Hegels form of idealism,
(3). The aporetic cluster provides each posi- this maneuver closes the gap the Sophists
tion with a ready-made argument. For the had ruled out, but does so from the opposite
negation of any member of the set is the side. Where the Sophist insists that whatever
conclusion of an argument containing appears must be real the idealist argues that
the remaining members as premises. Since only what is real can appear. For both, the
members of the original set are jointly real must be just as it appears to be; either
inconsistent, the argument will be valid. way, the commonsense distinction is ren-
And since the members are individually dered philosophically moot (and needs to
plausible, the audience will also find each be explained, or explained away, in com-
premise of the argument persuasive. plicated and, to some, implausible, ways).
In both its everyday and its philosophical
b i b l i og rap hy versions, the appearance/reality distinction
must be seen as a completely general one.
Chisholm, R.M.: Theory of Knowledge, 2nd While its most obvious illustrations involve
edn. (Englewood Cliffs, NJ: Prentice Hall, sense perception, it extends naturally to all
1977). dimensions of thought and experience. It
roy a. sorensen may seem to someone that two and two add
up to five. Arguably, it may just seem to one
that one desires or fears something. Hence
appearance/reality Nothing is more com- it is a mistake to draw the distinction by
monplace than the remark that things are not identifying one side with one metaphysical
always what they seem. We all know that category and the other with another, for
a thing can appear to be some way and yet example, the real with the material and
be really quite otherwise. Unlike some other appearances with the mental.
distinctions philosophers are enamoed of, What, then, is at the heart of the common-
the distinction between appearance and sense distinction, and what, if anything, is
reality is firmly rooted in everyday experience philosophically interesting about it?
and discourse. It is not surprising, then, I mentioned the perennial skeptical
that it has, since the dawn of philosophy, worry about whether appearances can tell
served to structure debates about what us whether there are things other than
there is to know and how, if at all, it can appearances and, if so, what they are like.
be known. Skepticism is an epistemological position.
When Socrates objected to the relativism But the very idea that there is a way things
of the Sophists, with its ugly moral con- are, whether or not one can know what
sequences, it was their refusal to allow that that way is, expresses a metaphysical belief,
there could be a gap between x appears usually labeled realism. Thus skepticism

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 116

a qu i n a s, st. tho mas

itself involves a metaphysical component. there are dangers in this reification of them
What account can we give of the appear- (see hypostasis, reification). First, it can
ance/reality distinction that does justice to lead to intractable metaphysical problems
both these components? Here is where the that are in fact avoidable. Second, it misleads
notion of evidence can provide the needed us as to the true nature of the distinction
general framework. between appearance and reality. Only when
An appearance is always an appearance understood as involving the relation be-
to someone, just as a piece of evidence is tween epistemological and ontological con-
always evidence for someone. The former cepts can it both retain the intuitive content
notion, in fact, represents a special case of of the commonsense distinction and yield
the latter. But the concept of evidence also a general philosophical problem that is not
involves the thought of something for which the artefact of some special metaphysical
the evidence is evidence. Thought of in this doctrine.
way, so does the idea of an appearance, as
the appearance of something. Even Kant,
bi bliography
who insists on the empirical reality of what
he calls appearances, arguably sometimes Austin, J.L.: Sense and Sensibilia (Oxford:
treats them as representing, albeit in a Oxford University Press, 1962).
special and highly problematical sense, a Ayer, A.J.: The Foundations of Empirical
transcendent reality (see noumenal/ Knowledge (London: Macmillan, 1940).
phenomenal). It is a conceptual truth that Burnyeat, M.: The Skeptical Tradition (Berkeley:
even the best evidence must fall short of University of California Press, 1983).
certainty (else it would not be evidence Grayling, A.C.: The Refutation of Scepticism
for something other than itself). In the same (La Salle, IL: Open Court, 1985).
way, the very concept of an appearance john biro
requires it to be distinct from that of which
it is an appearance. This is why the idealist
attempt to identify reality with appearances, Aquinas, St. Thomas (1224/574) The
no matter how the latter are idealized, is a philosophy of St. Thomas Aquinas was
mistake. It involves a non-evidential, hence strongly influenced by Aristotle and by
a non-epistemic, conception of appearances; the Islamic philosophers Avicenna and
in doing so, it loses contact with the point Averroes, whose works became available
of the commonsense distinction out of which in Latin translations at the beginning of the
the philosophical one grows. thirteenth century. But Aquinass meta-
What makes the appearance/reality dis- physical thought contains a number of
tinction both important and slippery is that elements that are not to be found in his
it straddles the division between epistemo- leading sources.
logy and metaphysics. Other well-worn
philosophical distinctions are either inter-
t he subj ect mat t er
nal to one or another of the traditional divi-
of metaphysi c s
sions of the subject (particular/universal,
necessary/contingent, a priori/a posteriori, Aristotles divergent statements on the nature
or concrete/abstract) or indifferent to of first philosophy led to an intensive dis-
them (extrinsic/intrinsic, specific/general, cussion of the subject matter of metaphysics
objective/subjective). Thinking of the among medieval thinkers. In Metaphysics
appearance/reality distinction in the evid- iv c.1 (1003a2132), Aristotle speaks of a
ential way as suggested here can save us science which studies being as being and
from mistaking it for a metaphysical one, one opposes it to other sciences which investigate
between two different kinds of entity. There beings from a particular point of view, for
may be good reasons for thinking that there instance, in so far as they are mobile. The
are appearances, as opposed to just the science of being as being, by contrast, is uni-
various ways the things there are appear. But versal. But in Book vi c.1 (1026a2332),

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 117

aqui nas, st . t homas

Aristotle distinguishes three theoretical sci- the transcendentals, becomes understand-
ences physics, mathematics and the divine able, for transcendentia are the universal
science and calls theology the first science, properties of being as such (see Aertsen,
because it is concerned with immobile and 1988). The term transcendental suggests
immaterial beings. The medieval discussion a kind of surpassing or going beyond. What
is focused on the question how Aristotles is transcended is the special modes of being
theological conception of first philosophy is which Aristotle called the categories.
related to the conception of metaphysics as While for the latter the categories are the
the universal science of being (see Zimmerm- most general genera of being, Aquinas
ann, 1965). considers them as special modes of being, as
In the prologue to his Commentary on the contractions of that which is: not every being
Metaphysics, Aquinas argues that meta- is a substance, or a quantity, or a quality, or
physics is concerned with both being as being a relation, etc. By contrast, the transcen-
and the immaterial substances, although not dentals express general modes of being.
in the same way. He develops his synthesis They transcend the categories, not because
with the help of the logician Aristotle, for they refer to a reality beyond the categories
Aquinass argument is based on the theory but because they are not limited to one
of science in the Posterior Analytics. The determinate category. Unlike the categories,
unity of a science consists in the unity of its the transcendentals do not exclude each
subject (subjectum). What is sought in every other, but are interchangeable or convertible
science are the proper causes of its subject. (convertibilis) with being and each other.
Now the immaterial substances are the In De veritate 1.1, Aquinas presents his
universal causes of being. Therefore, being most complete account of the transcend-
in general (ens commune) and the properties entals, of which the most important are
belonging to it are the subject of meta- being, one, true and good. Being is the first
physics. God is studied in this science only transcendental. The other transcendentals,
in so far as he is the cause of the subject although convertible with being, add con-
of metaphysics, that is, in so far as he is ceptually something to being, in the sense
the cause of being as such. God is not the that they express a mode of it which is
subject, but rather the end of metaphysical not yet made explicit by the term being
investigation. By this feature metaphysics itself. The general mode of being expressed
is distinguished from Christian theology by one pertains to every being in itself (in
(the theology of sacred scripture), for the se); one adds to being a negation, for
subject matter of this science is God himself it signifies that being is undivided. True
(cf. Summa theologiae 1.1.7). and good are relational transcendentals:
From this account it appears that Aquinas they express the conformity (convenientia) of
does not adopt the theological conception every being to something else. The condition
of metaphysics that was prevalent among for this relation is something whose nature
the Greek commentators on Aristotle. Accord- it is to accord with every being. Such is,
ing to them first philosophy is the science of Aquinas argues, the human soul, which
the most eminent being, the divine being. according to Aristotle (De anima iii c.8,
Aquinass view is ontological: metaphysics is 431b21) is in a sense all things. In the soul
the scientia communis, for its subject matter there is both a cognitive power and an
is being in general. appetitive power. The conformity to the
appetite or will is expressed by the term
good, the conformity to the intellect by
t h e d o c tr ine o f the
the term true. Truth as transcendental
t r a n s c e nd entals
signifies the intelligibility of things.
Against the background of Aquinass onto- Aquinass innovation in the doctrine of the
logical conception of metaphysics, the sig- transcendentals is the correlation he intro-
nificance of a doctrine that was developed duces between the human soul and being. He
in the thirteenth century, the doctrine of understands the transcendentals true and

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 118

a qu i n a s, st. tho mas

good in relation to the faculties of a spiritual that is purely potential and is brought into
substance. This understanding means an actuality through a form. Aquinas regards
acknowledgment of the special place human it as one of Aristotles great merits that with
being has among other beings in the world. his doctrine of the potentiality of matter
A human being is marked by a transcen- he made it possible to acknowledge a sub-
dental openness; its object is being in general. stantial change, or generation.
This openness is the condition of the pos- Aquinas emphasizes, however, that the
sibility of metaphysics. final step had not yet been taken, for the
The doctrine of the transcendentals plays generation, too, presupposes something, in
a central role in Aquinass metaphysics. It keeping with a common supposition of Greek
integrates the theory of knowledge (truth) thought: nothing comes from nothing
into an ontology and it provides the (ex nihilo nihil fit). The philosophers of the
foundation for the first principle of morality: first and second phases considered the origin
good is to be done and pursued, and evil of being under some particular aspect,
avoided (Summa theologiae III.94.2). The namely, either as this being or as such
doctrine is also fundamental for philosophical being. As a result, the causes to which they
theology. Within the framework of a reflec- attributed the becoming of things were
tion on the divine names Being, Unity, particular. Their causality is restricted to
Truth and Good Aquinas discusses the one category of being: accident (as in the first
relation between the transcendentals and place), or substance (as in the second).
God. Because the transcendentals are self- The third phase began when some
evidently knowable, and because they do not thinkers raised themselves to the considera-
express a limited, categorical mode of being, tion of being as being. In this metaphysical
they are seen as providing the basis for the analysis they assigned a cause to things
possibility of rational knowledge of God. not only in so far as they are such (by
accidental forms) and these (by substan-
tial forms), but also as considered according
t h e h i sto r y o f the q uestio n
to all that belongs to their being. The origin
of being
considered by the metaphysician is tran-
In Summa theologiae 1.44.2, Aquinas sketches scendental, it concerns being as such, not
the history of philosophical reflection about merely being as analyzed into natural cat-
the origin of being. This text can be regarded egories. The procession of all being from the
as the medieval origin of the question Why universal cause is not a generation, because
is there something and not rather nothing? it no longer presupposes anything in that
Three main phases can be distinguished in which is caused. It is creation ex nihilo.
the progression of philosophy as Aquinas A striking feature of Aquinass view of
sees it. the progress of philosophy is that the idea of
The first step was taken by the creation appears as the result of the internal
presocratics. They held that matter is development of human thought, independ-
the substance of things and that all forms ent of revelation. In the context of the idea
are accidents. They posited one or more of creation Aquinas elaborates two central
substrata (water, fire, etc.) which they ideas of his metaphysics: the composition of
regarded as the ungenerated and inde- essence and existence in created things, and
structible principle of all things. To the the doctrine of participation.
extent to which they acknowledged chance
in the substratum, it consisted only in alter-
t he composi t i on of essenc e and
ation, a change of its accidental forms (see
exi stence participation
The second stage in the progress of philo- The distinction between essence and existence
sophy was reached when philosophers made (esse) was introduced by Islamic thinkers
a distinction between matter and sub- in order to explain the contingent character
stantial form. They posited a prime matter of caused beings (see essence/accident).

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 119

aqui nas, st . t homas

Existence does not belong to the essence of being. One of the most significant innova-
what is caused, for it has received its being tions in Thomistic scholarship since the
from something else. The relation between Second World War has been the discov-
essence and existence was interpreted by ery of the Platonist Thomas (see Plato,
Avicenna according to the model of sub- Platonism). Pioneering studies were the
stance and accident: esse is an accident works of Fabro (1961) and Geiger (1942),
superadded to essence. which showed the central role of the
Aquinas teaches the real composition of Platonic notion of participation in Aquinass
essence and existence in all creatures already metaphysics, a notion that was sharply
in one of his earliest works, De ente et essen- criticized by Aristotle. Aquinas interprets
tia. In chapter 4 he discusses the essence the idea of creation philosophically in
of the separated substances or spiritual terms of participation. The relation of cre-
creatures. This issue engaged Aquinas a atures to the first cause is the relation of
great deal he even devoted a particular participation in being.
treatise to it. De substantiis separatis for it
concerns the ontological structure of finite
substances. This structure cannot consist wri t i ngs
in the composition of form and matter, since The critical edition of Aquinass works, the
spiritual substances are separated from Leonine edition, is still unfinished: Opera
matter. Yet although such substances are omnia. Iussu impensaque Leonis XIII, P.M.
pure forms, they do not have complete edita (Rome: Vatican Polyglot Press,
simplicity. All creatures are composed of 1882). For a complete listing of the vari-
essence and existence, because they have ous editions, see J.A. Weisheipl, Friar
their esse not of themselves, but from God. Thomas dAquino (Washington, DC: The
According to Aquinas, however, existence Catholic University of America Press,
is not an accident superadded to essence. 1983), 355 404.
Existence and essence are related to each
other as act to potency. He extends the
notions of act and potency, which were bibl i ography
correlative with the notions of form and Aertsen, J.A.: Nature and Creature. Thomas
matter in Aristotle, to being as such. In a Aquinass Way of Thought (Leiden: Brill,
famous text in his De potentia (7.2 ad. 9), 1988).
Aquinas states: That which I call esse is the Fabro, C.: Participation et causalit selon S.
actuality (actualitas) of all acts, and for this Thomas dAquin (Louvain and Paris:
reason it is the perfection of all perfections. Nauwelaerts, 1961).
For Aquinas to be is not a bare fact, but the Geiger, L.-B.: La Participation dans la philo-
ultimate act through which a thing achieves sophie de S. Thomas dAquin (Paris: Vrin,
its perfection. Every excellence of any thing 1942); 2nd edn. (Paris: Vrin, 1953).
belongs to it according to its esse. For man Gilson, E.: Being and Some Philosophers
would have no excellence as a result of his (Toronto: Pontifical Institute of Mediaeval
wisdom unless through it he were wise Studies, 1949); 2nd edn. (Toronto: Pon-
(Summa contra Gentiles 1.28). It was Gilson tifical Institute of Mediaeval Studies,
(1949) in particular, who has emphasized 1952).
the existential character of Aquinass meta- Wippel, J.F.: Metaphysical Themes in Thomas
physics against the dominant essentialist Aquinas (Washington, DC: The Catholic
tradition in modern philosophy. University of America Press, 1984).
Closely connected with the distinction of Zimmermann, A.: Ontologie oder Metaphysik?
essence and existence in created things is Die Diskussion ber den Gegenstand der
Aquinass doctrine of participation. No finite Metaphysik im 13. und 14. Jahrhundert
being is its esse, but has it. Only in God are (Leiden: Brill, 1965).
essence and existence identical: he is essen-
tially Being. All other things participate in jan a. aertsen

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 120

a r c h e t yp e
archetype From the Greek JK, Aristotle (384322 bc) Greek philosopher
a pattern or model. The word is applied to born in Stagira. Aristotles writings can be
the reality whether in the mind of God said to have set the agenda for the western
in nature itself, or in a third, abstract tradition in metaphysics. Indeed, meta-
realm to which a conception is referred. physics is a term derived from a first century
Archetypes sometimes play a causal role in bc edition of Aristotles work, in which a
originating those conceptions; their refer- collection of his writings was put together
ence or truth is then assured (or at the very under the title Ta Meta ta Phusika, which
least argued for) by their causal ancestry. means simply What comes after the writings
The Greek word was applied by Platonists on nature (ta phusika). Since the writings
(though not by Plato himself, who spoke thus put together concerned topics that
instead of O) to the forms (see seemed in certain ways related substance
Platonism). Later Platonists placed these and being, change and explanation, unity
forms or archetypes in the mind of God. and plurality, potentiality and actuality,
Philosophers of the seventeenth and eigh- non-contradiction, the nature of the eternal
teenth centuries conceived of them more and unchanging these topics were sub-
broadly. Descartes described the external sequently taken to be the subject matter of
cause of an idea as like an archetype. metaphysics, which increasingly became
Locke applied the word to the things the a separate department of philosophy. But
mind intends [its ideas] to stand for, and Aristotle himself did not group these topics
to which it refers them. Berkeley applied together. He does have a conception of
the word, with some reluctance, to ideas in the study of being qua being the study
the mind of God, which were, he argued, of what is true of all things that are, as such
no less serviceable as archetypes than the that links some of the contents of the
corporeal substances of the materialists. Metaphysics. But there is dispute about what
that study is, and how much of the work it
b i b l i og rap hy includes. Nor are Aristotles inquiries into the
topics we now call metaphysical confined
Berkeley, G.: Philosophical Correspondence to the work called Metaphysics. There is
Between Berkeley and Samuel Johnson an especially close link between that work
172930 (New York, 1929); repr. in and his inquiries into natural change and
The Works of George Berkeley, vol. 2, ed. explanation.
A.A. Luce and T.E. Jessop (London:
Thomas Nelson, 194857), 27194.
Berkeley, G.: Third Dialogue, in Three subst anc e, change and i dent i t y
Dialogues between Hylas and Philonous Aristotle once remarked that the central
(London, 1713); repr. in The Works of concern of previous philosophers, when they
George Berkeley, Bishop of Cloyne, vol. 2, asked questions about what being is, was
ed. A.A. Luce and T.E. Jessop (London: really, at bottom, a question about what
Thomas Nelson, 194857), 22763. substance is. (The term we translate sub-
Descartes, R.: Third Meditation, in Medita- stance is ousia, a verbal noun formed from
tions on First Philosophy (Paris, 1641); the participle of the verb to be.) For it
repr. in The Philosophical Writings of Des- is this, he continues, that some claim to
cartes, vol. 2, ed. and trans. J. Cottingham, be one in number, some more than one, and
R. Stoothoff, and D. Murdoch (Cambridge: some limited, others unlimited. He himself
Cambridge University Press, 1984), 29. devotes much effort to the task of finding
Locke, J.: An Essay Concerning Human Under- an adequate account of substance, and
standing (London, 1689); ed. P.H. Nidditch on defending the priority of substance to
(Oxford: Clarendon Press, 1975), Book II, other items such as qualities and materials.
ch. xxxi, sects. 13. It is not, however, intuitively obvious what
Plato: Republic V, VI (many versions). Aristotle means by the question, What is
kenneth p. winkler substance?, all the more since the term

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 121

ari stotle
ousia is primarily an Aristotelian term, with Socrates persistence as the same entity,
no clear history. We must search in his and properties (such as, perhaps, the ability
arguments and examples for an understand- to metabolize food, or the ability to think
ing of his motivation and goal: to what real and choose) whose presence is constitutive
puzzles does such a search respond? of the individuals identity, whose depar-
As Aristotle characterizes earlier inquiries ture would mean the end of an individual.
into substance, they focus on two questions, The identity question has a special urgency
rather closely related: (1) a question about where living creatures are concerned, since
the explanation of change; and (2) a question it is connected with complicated ethical and
about identity. We observe many changes political issues, for example, the determina-
in the world around us, such as the cycle of tion of death and the moral status of the
the seasons, the birth, growth and death foetus. Thus Aristotle holds that a creature
of living creatures. Early Greek mythology dies whenever it loses one of the properties
explained these changes by invoking the in the second group (the essential pro-
capricious will of anthropomorphic beings; perties); and he holds that the foetus at an
early philosophers, instead, looked for law- early stage of life is not a human being, and
like explanations. In the process, they had to does not exhibit identity with the human
ask themselves, first, what sorts of entities are being that may in due course come to be,
relatively stable and persisting, the things since it does not yet have all the essential
to which changes happen and in terms of properties of the human being.
whose underlying stability change could be In one way, these two questions seem to
coherently explained. (Plato had cogently point in opposite directions, identifying dif-
argued that coherent talk about change ferent things as the substance of a thing.
presupposes at least some stability: for a For the question about persistence through
change has to be the change of something, change might lead us to hold that material
and that thing cannot at the same time be stuffs are the basic substances of the things
ceasing to be the thing it is, or we will not they compose, seeing that these stuffs (for
be able to say anything about it.) The search example, the materials that make up the
for substance is, in part, a search for these body of Socrates) pre-exist the birth of
most basic persisting entities (see the extended Socrates and post-date his death. On the
essay on persistence), which Aristotle calls other hand, for this very reason they do not
substrata or subjects (two different trans- give the answer to questions about necessary
lations of his Greek term hupokeimenon, and sufficient conditions for Socrates iden-
literally that which underlies). tity. We are inclined, there, to look in the
The second question is what Aristotle direction of the structure characteristic of
calls the What is it? question. It may be Socrates species, his human make-up and
illustrated by countless common examples. functioning. For it seems that it is the dis-
Suppose I am considering some particular ruption of those modes of organization that
thing in my experience, say, Socrates. I have spells the end of his existence.
a sense that, in order to pursue my curiosity On the other hand, looked at in another
about this thing further, I must have some way the two questions seem to be closely
answer to the question, What is this? I intertwined. An adequate theory of change
want to know what it is about this thing that must single out, as its substrates, things
makes it the thing it is, what enables me to that are not only relatively enduring, but also
single it out as a distinct particular and definite and distinct. Unless we can indi-
mark it off from its surroundings, to reiden- viduate an item from its surroundings and
tify it later as the same thing I encountered say something about what it is, it will be
earlier. But to know this I need, it seems, difficult to make it the cornerstone of an
to separate the attributes of the thing into explanatory enterprise. And a good answer
two groups: properties (such as a sun-tan, to the What is it? question, asked about
or knowledge of history) that may come a particular such as Socrates, must tell us,
to be present, or depart, without affecting among other things, what changes Socrates

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 122

a r i s t o tle
can endure (as a substrate) and still remain of ten categories or (literally) predica-
one and the same. tions is an attempt to enumerate different
As Aristotle sees it, his predecessors went ways we might characterize a particular
wrong because they pursued one prong of the in our experience: we might speak about
substance inquiry to the neglect or distortion its substantial nature, its quantity, its
of the other. Early natural scientists, seeing qualit(ies), its relation(s), its place, time,
that material stuffs were the most persist- position, state, activity, passivity. At the
ing things around, surviving the deaths of same time, Aristotle also introduces a four-
humans and animals, held that these were fold distinction of things that are, separating
the real substance of things and the best (1) universals in the substance category,
answer to the What is it? question, when called Secondary Substance e.g., human
asked about particular substances. What being, horse; (2) particulars in non-
Socrates really is, is the materials that substance categories, such as this item of
compose him. This leads to paradoxical knowledge, this instance of pink color; (3)
conclusions: no substance ever perishes, universals in non-substance categories,
and substances continue to exist although such as knowledge, color; (4) particulars in
their parts are widely dispersed in space and substance categories, called Primary Sub-
time. Above all, this view fails to capture stance, e.g., this human being, this horse. The
a distinction that is fundamental in our motivation for these distinctions emerges
discourse and practices, namely the dis- when Aristotle explains the fundamental
tinction between property change (alloiosis) classifying role of natural-kind universals.
and real coming-into-being and going- His point is that we do not pick things
out-of-being (genesis and phthora), between out and trace them through time as bare
Socrates getting a sun-tan and the death unclassified matter; fundamental to our
of Socrates. practices of identifying and explaining is
Platonists (see Platonism), on the other the ability to say to what kind the thing
hand, focus on the identity question, and belongs. (His later writings give natural kinds
on the universals that are, as they see it, the a special place here, since artefacts have
best answer to that question. Each aspect of comparably unclear criteria of identity.)
Socrates is explained by his participation When we point at Socrates and say, What
in some universal form, such as the form is it? we are asking about a particular,
of Justice, which is imagined as existing and it is that particular thing that exists;
apart from particulars and as explaining classifying universals have no existence
the possession of that property in all the apart from particulars. But the universal
particulars that have it. Aristotle finds is of fundamental importance in coming
fault with this emphasis on the universal, to grips with the particulars identity
because it fails to come to grips with the and not just any universal, but the one,
material changing character of the indi- human being, that gives the kind to
vidual substance. Nor, in its Platonic form at which he belongs from birth to death. To
least, does this approach even succeed in answer, Socrates is a sitting thing, or
separating universals such as the Human, Socrates is a white thing, is a less reveal-
which must be true of Socrates as long as he ing answer, parasitic on our ability already
exists, from universals such as the White, to pick out Socrates as a human being. In
which he might lose (getting a suntan) short: the category of substance, which
while still remaining the same individual. includes the natural-kind universals and
In his early work, the Categories, Aristotle the particulars that fall under them, has
focuses on two tasks: demarcating the role priority over the other categories in both
of particulars and universals in answering explaining and identifying. Within this
What is it? questions about things, and category, particulars in a sense take priority,
defending the central role of natural-kind as the most basic substrates of change;
concepts in answering both change and but they get their identity from the uni-
identity questions. The famous enumeration versal under which they fall.

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 123

ari stotle
f or m a n d matter The famous twelfth book of the Metaphysics
then gives an account of god as an immortal
So far, Aristotle has said nothing about immaterial substance whose entire form is
the coming-to-be and passing-away of sub- thinking, and whose entire being is actual-
stances. Nor has he spoken about the ity rather than potentiality. God imparts
matter that composes them. To these tasks movement to the universe by being an
he turns in Physics i 79 and in Metaphysics object of passionate love to the heavenly
vii. He acknowledges that living substances bodies, who are themselves imagined to
are essentially enmattered structures: they be living thinking beings.
cannot continue as the things they are
without suitable matter to make them up
bei ng qua bei ng
and perform their life-activities. On the
and the basic
other hand, he insists that matter all by
princ i ples of thought
itself cannot give us the identity of a particu-
lar: for it is a mere lump or heap with- In Book iv of the Metaphysics, Aristotle
out the form or structure that it constitutes. defends the idea of a general study of the
Nor, indeed, despite matters purported attributes of things that are as such, or of
claim to be the substrate par excellence, does being qua being an idea that he seemed
matter even turn out to be as continuous to attack in some earlier writings as
as form, with respect to the individual insufficiently attentive to the multiplicity of
species member: for the matter that composes types of being. Here, by contrast, he argues
Socrates is changing continually, as he eats that the many ways in which we speak of
and excretes, while he himself remains being have more than a verbal unity: for
one and the same. all are understood through an inquiry into
Looking more closely into the question of substance, which is in some sense the basic
what does provide Socrates with his identity type of being in our explanation and under-
over time, Aristotles answer is that this is standing of the world. Aristotles project
his essence, and that this essence is a here has been understood in two very dif-
particular instance of characteristic species ferent ways. Some interpreters understand
organization or form (see hylomorphism), him to be calling for a general study of
not different in kind from that of other substances, focusing in particular on living
species members, but a countably different creatures, and for an illumination of pro-
instance, tracing a distinct career through perties, of activity and passivity, and so
time and space. (There are many different forth, that would be based upon that
interpretations of Aristotles final position understanding. Others have understood
on the contribution of the universal and him to be referring to god as the primary and
the particular in identity, but this one has central substance, a study of which is the
broad support.) focal point for all study of substance. The fact
In later books of the Metaphysics Aristotle that the relevant texts of the Metaphysics
investigates the role of form in making a derive from different periods in Aristotles
thing a unity, providing still further argu- life and are not edited into their present
ments against thinking of material stuffs as order by him makes resolution of this
what a thing is. Introducing the important question very difficult. One can at least say,
ideas of capability or potentiality (dunamis) however, that in the central books in which
and activity or actuality (energeia), he argues Aristotle does in fact investigate the nature
for the explanatory priority of a things actual of substance (Books viiix), there is no dis-
nature to its potentialities. Aristotle here cussion of god, and no sign that we need
begins to think about matter as a set of to understand the nature of god before
potentialities for functioning, which can answering questions about forms relation to
be explicated only when we have grasped matter. The same is true of the De anima,
the actual functional structure of the entity where bodiless substance is an anomaly,
that matter composes. briefly mentioned, in the works systematic

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 124

a r i s t o tle
study of the necessary interrelatedness of save as true the greatest number and the
form and matter (see hylomorphism). most basic of those appearances. This
Aristotle then goes on to argue that in procedure can be seen at work in many of
any inquiry whatever, a basic role is played his inquiries, both in natural science and
by two logical principles: the principle in ethics.
of non-contradiction and the principle of On the other hand, in the Posterior
the excluded middle. Formulating Non- Analytics Aristotle presents an account of
Contradiction as the principle that contra- the structure of scientific understanding,
dictory predicates cannot apply to a single and the goal of inquiry, that seem, at first,
subject at the same time in the same distinctly different. He argues that an inquirer
respect, Aristotle argues that this is the can claim episteme, or scientific understand-
most secure starting point of all, concern- ing, only when he has been able to arrange
ing which it is impossible to be in error. the results of inquiry into a deductive
Confronting an opponent who claims to explanatory system, internally consistent
doubt the principle (apparently a relativist and hierarchically ordered, depending on
who holds that if x seems F to observer O, x first principles that are true, necessary, basic
simply is F, and if x to observer P not to and explanatory of the other truths of the
be F, x simply is not F), Aristotle argues science in question. By itself this need not
that this opponent himself refutes himself, conflict with Aristotles emphasis elsewhere
if he utters any coherent sentence, or even on sorting out the record of experience:
any definite word. For any meaningful for he is simply adding the point that this
utterance must, in putting something definite sorting-out must be one that yields a
forward, at the same time implicitly rule out systematic grasp and the ability to give
something at the very least, the contra- explanations. But in Posterior Analytics ii 19,
dictory of what is put forward. He adds that Aristotle makes some remarks about the
if the opponent is silent and refuses to say nature of his first principles that seem to go
anything definite, he loses this way too: in a different direction: for he holds that,
for he is pretty much like a vegetable, after experience provides us with the
and it is ridiculous to look for words material of a science, its first principles
to address to someone who doesnt use must be grasped by a faculty which he calls
words. Moreover, even definite action with- nous. In traditional mediaeval interpretations
out words reveals a commitment to Non- of Aristotle, this has been understood to be
Contradiction: for when one acts one must a faculty of intellectual intuition that seizes
have some definite belief about what one on first principles a priori, and thus sets the
is aiming to do, and such beliefs, proposi- science on an extra-experiential foundation.
tional in form, presuppose a commitment Recent interpretations of the passage,
to Non-Contradiction. however, have pointed out that this is not
a plausible way of understanding what
is meant by nous in Aristotle (or, indeed,
m e t h odo lo gy: ap p ear ances
in the ordinary vocabulary of cognition
a n d u n der stand ing
from which he derives the term). Nous is
In passages such as the one from Metaphysics insight based upon experience; and what
iv just discussed, Aristotle appears to derive Aristotle is saying is that true under-
support for what he calls the most basic standing is not achieved until, in addition
principle of all simply by showing its depth to the grasp and use of principles, we gain
and ubiquity in our discourse and practices. understanding of the fundamental explana-
And elsewhere he states that in all inquiries tory role. This is exactly what the person
the aim should in fact be, first to set down who follows Aristotles arguments about
the appearances by which he seems to Non-Contradiction does derive: so there is
mean the record of human experience on no need to see the Posterior Analytics as
the issue and then, working through the in tension with that passage or others
puzzles this record presents, to go on to in which the method of philosophy is

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 125

ari stotle
understood to involve a systematization from design such as was developed later by
of experience. the Stoics. Instead, Aristotles interest is in the
plastic and self-maintaining, self-nourishing
character of living systems: in a variety of
n a t u r e and exp lanatio n
circumstances, they will behave in the way
Aristotles account of explanation, in the best suited to realize and then maintain
second book of his Physics, is closely linked their forms and structures. And under-
to his arguments about substance. He standing this will enable us to grasp their
identifies four different types of explanation doings in a unified way predicting, for
that are standardly given when we ask the example, that a plants roots will grow in
question Why? about some entity or event the direction of the water supply, wherever
in our experience. (These are often called that happens to be. Teleological explana-
the four causes, but it would be better tions do not invoke mysterious notions;
to think of them as the four becauses.) they grow from a biologists observation
First, we often enumerate the material that organic systems function in integrated
constituents of a thing; but this, Aristotle and form-preserving ways.
argues, explains nothing about a thing unless Aristotles passionate interest in biology
we have already said what sort of thing it animates much of his metaphysical writing.
is. The second sort of explanation, which He spent about twenty years of his career
cites the things form or structure, is in that doing first-hand biological research, much
sense prior to the first. The third sort, which of it very fine. And his biological writings pro-
Aristotle calls the origin of change, and vide rich insight into metaphysical issues
which is often called efficient cause, cor- such as the relation of form and matter and
responds rather closely to our notion of the nature of functional explanation. To
causal explanation: asked why something students who evidently preferred theology to
happened, or why a thing is as it is, we the study of worms and shellfish, he makes
cite some other event or agency that acted a reply that might perhaps serve as an
in such a way as to produce it. excellent introduction to Aristotles tem-
Finally, Aristotle introduces the explana- perament as metaphysician and philosopher
tion that for the sake of which, often of nature:
called teleological explanation (see tele- We must not enter upon the study of
ology). Here we say that the reason x the lesser animals with childish disgust. For
happened was for the sake of y, where y is in every natural thing there is something
in the future. It is not difficult to understand wonderful. There is a story which tells how
the relevance of this sort of explanation in some foreigners once wanted to meet
the context of intentional human action Heraclitus. When they entered, they saw
(He did this in order to get that). What is him warming himself in front of the stove.
harder to understand is the role Aristotle They hesitated; but he told them, Come in;
gives it in explaining the growth and devel- dont be afraid; there are gods here too.
opment of living creatures of all sorts,
including many (such as plants) that are
wri t i ngs
not, in his view, capable of intentional
action. He recommends that we should Categories, On Interpretation, Physics, De anima
give accounts of the development of a seed, (On the Soul), Parts of Animals, Generation
for example, or of various life processes in of Animals, Metaphysics.
a mature plant, by saying that they happen Translations: the best general collection is The
for the sake of the form or structure of the Collected Works of Aristotle, ed. J. Barnes,
plant. Aristotle is at pains to insist that he is 2 vols. (Princeton, NJ: Princeton Univer-
not invoking any causal factors external to sity Press, 1984). See also the comment-
the nature of the organism in each case. aries and translations in the Clarendon
It seems wrong to see any implications of Aristotle Series, esp. those of Categories
a grand teleology of nature or an argument and On Interpretation by J.L. Ackrill, of

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 126

a r m s t r o ng, d avid malet

Metaphysics, ivvi, by C. Kirwan (Oxford: Owens, J.: The Doctrine of Being in the
Clarendon Press, 1963, 1971), of Parts of Aristotelian Metaphysics (Toronto: Univer-
Animals I and Generation of Animals I by sity of Toronto Press, 1978).
David Balme (1972). Ross, W.D.: Aristotle, 5th edn. (London:
A useful collection of good translations Methuen, 1949).
can be found in A New Aristotle Reader, Witt, C.: Substance and Essence in Aristotle: An
ed. J.L. Ackrill (Oxford: Clarendon Press, Interpretation of Metaphysics VIIIX (Ithaca,
1987). NY: Cornell University Press, 1989).
Editions and commentaries: W.D. Ross, martha c. nussbaum
Aristotles Metaphysics (Oxford: Clarendon
Press, 1924); M. Frede and G. Patzig,
Aristoteles: Metaphysik Z, 2 vols. (Munich: Armstrong, David Malet (1926 )
C.H. Beck, 1988); G. Fine, On Ideas (Oxford: Australian philosopher, born in Melbourne
Clarendon Press, 1993) (on the fragments and educated at the University of Sydney
of Aristotles lost Peri Ideon, a critique of and Exeter College, Oxford. After Oxford, he
Platos theory of forms). spent a brief period teaching at Birkbeck
College in the University of London, then
seven years at the University of Melbourne.
b i b l i og rap hy
He held John Andersons chair as Challis
Ackrill, J.L.: Aristotle the Philosopher (Oxford: Professor of Philosophy in Sydney from 1964
Oxford University Press, 1981). until his retirement at the end of 1991.
Barnes, J.: Aristotle (Oxford: Oxford University Armstrongs work in philosophy ranges
Press, 1982). over many of the main issues in epistemo-
Burnyeat, M.: Aristotle on Understanding logy and metaphysics, where he has helped
Knowledge, in Aristotle on Science: The to shape philosophys agenda and terms of
Posterior Analytics, ed. E. Berti (Padua: debate. Several themes run through it all: it
Antenori, 1981), 97139. is always concerned to elaborate and defend
Furth, M., Substance, Form, and Psyche: a philosophy which is ontically economical,
An Aristotelian Metaphysics (Cambridge: synoptic, and compatibly continuous with
Cambridge University Press, 1988). established results in the natural sciences.
Gill, M.L.: Aristotle on Substance: The Paradox Accordingly, he has argued for a naturalism
of Unity (Princeton, NJ: Princeton Univer- which holds all reality to be spatio-temporal,
sity Press, 1989). for a materialism (see physicalism, mater-
Hartman, E.: Substance, Body, and Soul ialism) which aims to account for all
(Princeton, NJ: Princeton University Press, mental phenomena without appeal beyond
1978). the categories of physical being, and for
Irwin, T.H.: Aristotles First Principles an empiricism which both vindicates and
(Oxford: Clarendon Press, 1988). draws strength from the methods and suc-
Kosman, A.: Substance, Being, and Ener- cesses of the natural sciences.
geia, Oxford Studies in Ancient Philosophy In Perception and the Physical World (1961),
2 (1984), 12149. he confronted then-fashionable phenomen-
Lesher, J.: The Role of Nous in Aristotles alist tendencies (see phenomenalism) with
Posterior Analytics, Phronesis 18 (1973), a direct realism which had no place for
44 68. sense data or other mentalistic items (see
Nussbaum, M.: Aristotle, in Ancient Writers: sensa). He urged the objections to sense
Greece and Rome, vol. 1, ed. T.J. Luce data from their indeterminacy, their hidden
(New York: Charles Scribners Sons, 1982), features, and the identification problems they
377 416. face. He began also to develop a realist
Owen, G.E.L.: Logic, Science, and Dialectic: account of secondary qualities (see quality,
Collected Papers in Ancient Philosophy primary/secondary).
(London and Ithaca, NY: Cornell Univer- A Materialist Theory of the Mind (1968) was
sity Press, 1986). the first full-dress presentation of central-state

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 127

arnauld, antoine
materialism, which identifies states of mind actual cosmic constituents. Here again,
with states of the central nervous system Armstrongs doctrine about universals, as
(see the mind/body problem). The theory is abstractions from states of affairs on an
as naturalistic as the behaviorism it aspired equal footing with particulars, stands him
to supplant, yet much more plausible and sci- in good stead.
entifically fruitful as a philosophy of mind.
Armstrong presents an analysis of mental See also logical atomism; the extended essay
phenomena in terms of what they are apt on modalities and possible worlds.
to cause, or be caused by, then proceeds to
claim that the most likely items to fit those
places in the causal networks of human
perception, feeling, memory and action are A Combinatorial Theory of Possibility (Cam-
structures, states and processes in the cen- bridge: Cambridge University Press, 1989).
tral nervous system. The view is refined in A Materialist Theory of the Mind (London:
further essays. With hindsight, Armstrongs Routledge and Kegan Paul, 1968).
philosophy of mind counts as a typetype The Nature of Mind and Other Essays (Brisbane:
identity theory, a precursor of contempor- Queensland University Press, 1980).
ary functionalism. Perception and the Physical World (London:
During the 1970s, Armstrong turned his Routledge and Kegan Paul, 1961).
attention to the problem of universals. In Universals and Scientific Realism, 2 vols. Vol. 1
Universals and Scientific Realism (1978) he Nominalism and Realism; Vol. 2, A Theory
built a case for an immanent realism in of Universals (Cambridge: Cambridge
which universals, and particulars (see University Press, 1978).
universals and particulars) are equally What Is a Law of Nature? (Cambridge:
abstractions from states of affairs. The work Cambridge University Press, 1983).
has three principal themes: first, all the
widely accepted varieties of nominalism are
bibl i ography
deeply implausible. Second, an empiricist
naturalism need not, and should not, bear Bacon, J.B., Campbell, K., and Reinhardt,
the nominalist burden. Third, to establish L., ed.: Ontology, Causality, and Mind; Essays
the actual existence of any universal calls for in Honour of D.M. Armstrong (Cambridge:
a substantive enquiry for which the funda- Cambridge University Press, 1993).
mental sciences alone are equipped. keith campbell
This scientific realism about universals
was promptly put to work in developing a
philosophy of the laws which apparently Arnauld, Antoine (161294) A French
govern the cosmos. What Is a Law of Nature? Roman Catholic theologian and philo-
(1983) argues that the regularity theories sopher. Arnauld was born in Paris into a
of law, deriving from Hume, are all fatally family associated with Jansenism. Angelique
flawed (see law of nature). It goes on to urge Arnauld, his sister, was abbess of port-royal,
that laws relating particular states of affairs which became, under her direction, a cen-
rest on a relation of necessitation holding ter of Jansenism. One aspect of Jansenism
between the universals involved. is adherence to whatever view of the rela-
Armstrongs next major project was A tion of divine grace to human freedom is
Combinatorial Theory of Possibility (1989). expressed in Augustinus, a work written
Here he attempts to build, from a foundation by Cornelius Jansen and published pos-
in the thought of Wittgensteins Tractatus, thumously in 1640. Numerous Roman
an account of modality in which a spatio- Catholics, including various popes, believed
temporal naturalism is upheld. Non-actual that the Jansenist account of grace is incom-
possibilities do not exist, nor are they given patible with the Roman Catholic dogma that
ersatz treatment. The attempt makes use of divine grace can always be resisted by a
the idea of fictive reorderings of strictly free agent. Much of Arnaulds theological

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 128

a r n a u ld , anto ine
writings is devoted to a defense of the Descartess thesis that nothing occurs
Jansenist account of divine grace and the in the soul of which it is not conscious.
claim that it is consistent with Roman Except for this last thesis, which Arnauld
Catholic dogma. Another important segment regarded as inessential to Descartess pro-
of Arnaulds theological writings concerns gram, his criticisms were aimed at Descartess
the role of the sacraments in the process of arguments, not the conclusions of those
absolution, where Arnauld emphasized the arguments.
attitude that the penitent must bring to the Arnauld criticized some of Descartess
process if the sacrament is to absolve. doctrines because of their theological im-
In connection with a school associated plications. The majority of Arnaulds criti-
with Port-Royal Arnauld wrote or co-wrote cisms of Malebranche center on what he
three important textbooks that influenced viewed as Malebranches speculative and
seventeenth-century thought: Grammaire innovative contributions to theology. But
gnrale et raisonne (1660), La Logique, in the process, Arnauld formulated a theory
ou lart de penser (1662) and Nouveaux of perception, which he presented as a mere
lments de gomtrie (1667). recasting of Descartess theory, but which,
In his Jansenist phase Arnauld offered and in fact, involves many ideas original to
argued in favor of an historical approach Arnauld. Arnaulds theory of perception
to theology on the ground that the essential is contained in two works aimed at
theological truths could be extracted from the Malebranche: Des vraies et des fausses ides
work of the Fathers of the Church and, in (1683) and Dfense de M. Arnauld, contre la
particular, at least with respect to matters rponse au livre des vraies et des fausses ides
of divine grace and freedom, from the work (1684). In these works, Arnauld articulated
of Augustine. He, therefore, strongly opposed and defended a subtle form of a direct real-
what he took to be the innovative, specula- ist position, based on an act theory of ideas,
tive philosophical theology of leibniz and in which ideas are identified with represen-
malebranche. Criticism of Malebranche gen- tative acts of the mind rather than objects
erated the majority of Arnaulds positive of the mind that serve as intermediaries
contributions to philosophy. between an act of the mind and the external
While Arnauld was a conservative in reality thereby represented.
theology, he believed that scholastic philo-
sophy had been exposed as inadequate by
the seventeenth-century scientific revolution
and Cartesian mechanics (see descartes). La Logique, ou lart de penser (Paris, 1662);
In philosophy, Arnauld regarded himself ed. and trans. J. Dickoff and P. James
as a Cartesian, specifically associating him- (Indianapolis, IN: Bobbs-Merrill, 1964).
self with Descartess theses concerning the Oeuvres de Messire Antoine Arnauld, docteur de
nature and origin of ideas, the idea of God, la maison et socit de Sorbonne, 43 vols.
the distinction between the soul and the (Paris, 17751839); repr. Brussels:
body, and the nature of matter. This may Culture et Civilisation, 1967).
seem odd, given Arnaulds famous criti- On True and False Ideas, New Objections
cisms of Descartess Meditations on First to Descartes Meditations and Descartes
Philosophy, including a brilliant critique of Replies (Cologne, 1683); trans. E.J. Kremer
Descartess arguments intended to prove (Lewiston, NY, Queenston, ON, and
that the soul and body are distinct sub- Lampeter, Wales: Edwin Mellen Press,
stances (see the extended essay on the 1990).
mind/body problem), a critique of one of
Descartess arguments for the existence of
bi bliography
God, a query concerning the possibility
of avoiding circularity, given Descartess Nadler, S.M.: Arnauld and the Cartesian Philo-
way of establishing the principle of clear sophy of Ideas (Princeton, NJ: Princeton
and distinct perception, and a criticism of University Press, 1989).

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 129

associat i onism
Ndiaye, A.R.: La Philosophie dAntoine Arnauld of association, associationism as a psy-
(Paris: J. Vrin, 1991). chological program achieved its greatest
influence in eighteenth- and nineteenth-
robert c. sleigh, jr. century Britain. Locke was the first to use
the term association of ideas, but he used
artefact Any object produced to design it only to describe a cause of error, in which
by skilled action. Artefacts are continuants, accidental or logically irrelevant relations
that is, objects persisting in time: an event among ideas usurp the role of logical rela-
such as a pianists performance is itself an tions. Berkeley put association to more posi-
action and not the persisting product of tive and extensive use in An Essay towards
one. Artefacts are not exclusively human: a New Theory of Vision (1709), arguing
consider a beavers dam, or the cosmos that visual perception of distance is the
viewed by creationists. But the most elabor- result of an association between certain
ate artefacts we know, requiring conscious kinds of visual ideas and certain kinds of
planning, training and sophisticated forms non-resembling tactile ideas, an association
of representation, are human: levels of resulting from their repeated conjunction
culture are even measured by the kinds of in past experience. David Humes cognitive
artefacts people produce, from stone axes psychology gives a fundamental role to
to moon rockets. Artefacts contrast with three principles of association: contiguity,
natural objects: Aristotle considered arte- resemblance and causation, the latter based
facts, defined by function rather than an on constant conjunction. Hume uses these
autonomous principle of unity and persist- relations to explain both the formation of
ence, not to be substances. Mechanistic complex ideas from simpler ideas, and the
world views tend to blur this distinction. succession of ideas in thought. For Berkeley
The identity conditions (see individuation) and Hume, in particular, the association of
of artefacts are, however, vaguer and more ideas provided a way of explaining mental
convention-bound than those of natural phenomena without presupposing intellectual
objects: the puzzle of the Ship of Theseus insight into the metaphysical structure of
notably concerns an artefact. the world. David Hartley (170557), a
physician and Humes contemporary, also
sought to explain a variety of mental
b i b l i og rap hy phenomena associationistically, proposing
to explain the influence of associative rela-
Aristotle: Physics, ed. W.D. Ross (Oxford: tions through their relation to vibrations
Clarendon Press, 1950), Bk. 2, 192. in the brain. Later associationists included
Hobbes, T.: De corpore (London, 1655); in his Thomas Brown (17781820), James Mill
Opera philosophica, ed. W. Molesworth (17731836), John Stuart Mill, and
(London, 1839), vol. 1, Part II, ch. 11. Alexander Bain (18181903).
Wiggins, D.: Sameness and Substance (Oxford:
Blackwell, 1980), esp. ch. 3.
bi bliography
peter simons
Bain, A.: Mental Science (New York, 1868);
(New York: Arno Press, 1973).
associationism Associationism is the Bain, A.: The Senses and the Intellect, 4th
attempt to explain mental phenomena edn. (New York: Appleton, 1894).
through relations among mental contents Berkeley, G.: The Works of George Berkeley,
and representations particularly relations Bishop of Cloyne, ed. A.A. Luce and T.E.
such as contiguity or simultaneity, resem- Jessop, 9 vols. (London and New York:
blance and constant conjunction that Nelson, 194857).
cause them to become associated with one Brown, T.: Inquiry into the Relation of Cause
another. Although Aristole, Hobbes, and and Effect (Edinburgh, 1806); repr. as
Spinoza, among others, described phenomena The Doctrine of Mr. Hume: Concerning the

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 130

Relation of Cause and Effect (New York: to be from Non-Being or abruptly cease to
Garland, 1983). be. And where one sort of Being appears to
Hartley, D.: Observations on Man, His Frame, become another sort, the difference must
His Duty, and His Expectations (London, itself count as Being, so that there is
1749); (New York: Garland, 1971). no real change. Being is thus ultimately
Hume, D.: An Enquiry Concerning Human immutable and one. For a physics, that is,
Understanding (London, 1777); in Enquiries an account of the regularities of perceived
Concerning Human Understanding and change, to be possible, this paradoxical con-
Concerning the Principles of Morals, ed. clusion had to be overcome.
L.A. Selby-Bigge (Oxford: Clarendon The atomism of Leucippus and Demo-
Press, 1893); 3rd edn. rev. P.H. Nidditch critus retained something of Parmenides
(Oxford: Clarendon Press, 1975). sharp dichotomy being Being and Non-Being,
Hume, D.: A Treatise of Human Nature while modifying it in two fundamental
(London, 183940); ed. L.A. Selby-Bigge respects. Instead of one Being, there is an
(Oxford: Clarendon Press, 1888); 2nd edn. infinite multitude of indistinguishable beings,
rev. P.H. Nidditch (Oxford: Clarendon each (like the Parmenidean original) one
Press, 1978). and immutable. And instead of Non-Being,
Locke, J.: An Essay Concerning Human Under- there is the Void in which atoms can move.
standing (London, 1690); ed. P.H. Nidditch The Void is almost Non-Being; indeed,
(Oxford: Clarendon Press, 1975). Democritus calls it Nothing. But it is just
Mill, J.: Analysis of the Phenomena of the sufficient to make change possible, though
Human Mind (London: Longmans, 1829). only one kind of change, local motion. Thus
Mill, J.S.: An Examination of Sir William all change must (despite appearances) reduce
Hamiltons Philosophy (London and Boston: to local motion of entities that themselves
Spencer, 1865). must be imperceptibly small since no local
motion is actually perceived when, for
don garrett
example, a leaf changes color. Likewise, the
manifold qualitative differences between
atomism Atomism takes the world to be perceptible things must reduce to differ-
made up of indivisible and imperceptibly ences of atomic configuration, size and
small material units. (Atomos in Greek means shape. And since the analysis is a perfectly
indivisible.) The diverse qualities of per- general one, it must extend to all things,
ceptible bodies are to be explained by the to soul, for example, whose atomic con-
simple quantitative properties of the atoms stituents presumably are small and round so
composing them. Perceptible changes are to that they can direct the vital functions of the
be understood as rearrangements of the living body. Atomism in this pure form
underlying atoms. In its origins, atomism thus entails a strongly reductionist form of
was primarily a metaphysical doctrine; it materialism. Its appeal is to the coherence of
was not, indeed, until the early nineteenth its very general account of change, though
century that the atomic hypothesis was there are hints of a more specific sort of
linked tightly enough to the explanation of warrant also; evaporation and condensa-
specific empirical data to count as physical tion are said to be explained by different
theory in the familiar modern sense. degrees of packing of atoms, for example.
Though atomism itself was not immedi-
ately influential, the atomic metaphor can be
e a r l y a to mism
found everywhere in the philosophic think-
The first atomist doctrines are best understood ing of Parmenides successors. One finds
as a response to the challenge of Parmenides hints of it in Empedocles four elements, in
analysis of change. Parmenides argued that, Anaxagoras seeds, and in Platos elemental
despite the evidence of our senses, our rea- geometrical shapes. Aristotle proposed an
son compels us to conclude that change is alternative analysis of change in terms of
illusory. Being obviously cannot just come matter, form and privation that countered

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 131

at omism
Parmenides doctrine without yielding to eventually, mass. Other properties (the
the reductionism and lack of teleology that secondary ones) would then have to be
he found so objectionable in the atomist explicable in terms of the primary ones (see
proposal. Yet Aristotle also objected to quality, primary/secondary).
Anaxagoras assumption that physical things This requires explanation in terms of
can be divided without limits. There are, he something like atoms. Since, however, the
said, least natural parts. The limits of divis- atoms do not have to be strictly indivisible,
ibility depend on the kind of thing being the term corpuscle was preferred. But how
divided. were these invisible corpuscles to be known?
This suggestion was the occasion for a How, in practice, could their sizes, shapes,
vast and ingenious elaboration among later and motions explain such a property as yel-
Aristotelian commentators of the doctrine lowness? Locke was pessimistic about the
of the minima naturalia, that is, of the con- prospects of linking the two sorts of proper-
ceptual limits of physical divisibility. Averroes ties in a demonstrative science, though he
and his later followers seem to have been the suggested that plausible analogies might
first to present these least parts as separately yield at least a weak kind of probability.
existent, indeed as potentially capable by Meanwhile, chemists were trying to
their intermixing of explaining the qualit- understand chemical combination in cor-
ative changes we today call chemical. Such puscular and quantitative terms. Robert
Renaissance Aristotelians as Julius Caesar Boyle (162791) proposed that the cor-
Scaliger (14841558) and Agostino Nifo puscles constituting the chemical elements
(14731538) propounded a doctrine of could combine to form complex corpuscles
minima which was close to atomism in that yielded chemical compounds. He con-
significant ways, since the minima were ceded that the former might themselves
regarded as real constituents whose manner be primary concretions, composites made
of union explains the properties of sensible up of Democritean atoms. But in practice,
bodies. What separated these philosophers these primary concretions could be regarded
from Democritean atomism was their com- as basic from the point of view of the
mitment to matter-form composition, and chemist because they remained unaltered
especially to the role of substantial form in through chemical change. The problem was
making the product of the union of minima how to decide which concretions were
into a qualitatively new kind of thing. primary, how to distinguish element from
compound. Boyle could not discover a
consistent way to carry this all-important
t r a n s i t io n
distinction through.
With the seventeenth century the trans- By the end of the century the separation
formation of a philosophic doctrine into a between philosophers and scientists (as the
physical theory began. Most of the natural latter would come to be called) was widen-
philosophers of the century subscribed to ing. Scientists were convinced of the under-
the corpuscular philosophy. Though it lying corpuscular character of the world,
had roots in classical atomism (here the role but they had no real evidence (as evidence
of Gassendi in modifying and popularizing in natural science was coming to be under-
the ancient doctrine was important) and in stood) in support of their hypothesis. There
minima theory (the main spokesman here was as yet no satisfactory atomic theory.
being Daniel Sennert (15721657)), the
more important motivation came from the
atomi c theory
new science of mechanics. If mechanics
were to be as all-explanatory as its expo- Atomic theory took shape only very gradu-
nents expected it to be, the primary prop- ally, and in two different parts of natural
erties of things had to be those which science, in chemistry first and later in the
made things subject to mechanical law: physics of gases. The Newtonian project
size, shape, mobility, solidity and, perhaps of organizing chemical research around

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 132

a u g u s t ine o f hip p o , st.

short-range laws of force operating between acceptable as a calculational device but
corpuscles proved fruitless (see Newton). no more. The debate was once more philo-
Careful weighing of the products of chemical sophical, though numerous scientific issues
combination ultimately, in the hands of were also involved. Only after Einstein made
Antoine Lavoisier (174394), yielded the use of the molecular hypothesis in 1905 6
first victory. Aided by the assumption that to derive in a strikingly detailed way the
weight is conserved through chemical main parameters of Brownian motion did
change, Lavoisier provided for the first time the critics concede. Not that scientific real-
a reliable way of distinguishing element ism would from henceforward be immune
from compound, enabling him to identify to challenge!
many of the commonest elements. Joseph From Democritus to Einstein is a long road,
Louis Proust (17541826) proposed that and the atom of modern quantum theory
each compound is made up of elements bears little resemblance to the immutable
combined in a constant way. But it was qualityless particle of the first atomists. But
John Dalton (17661844) in A New System the claim that the world around us consists
of Chemical Philosophy (1808) who drew of a swarm of imperceptible entities whose
from the ancient notion of atom the crucial properties can causally explain the proper-
clue. He proposed that the simplest under- ties of that larger world evokes echoes all
lying structure that would explain the along that road. The transition from meta-
empirically established laws of definite physical doctrine to physical theory has no
proportions (a compound contains fixed clearer illustrative example.
proportions by weight of its constituents)
and of equivalent proportions (the ratio of the
bibl i ography
weights of a and b that react with a given
amount of c is independent of c), was an Furley, D.J.: Two Studies in the Greek Atomists
atomic one. Each atom of an element is like (Princeton, NJ: Princeton University Press,
any other atom of that element; each element 1967).
is constituted by a different kind of atom. Kirk, G.S., Raven, J.E., and Schofield, M.:
Compounds are formed by a simple and The Presocratic Philosophers, 2nd edn.
uniform juxtaposition of elemental atoms (Cambridge: Cambridge University Press,
in compound particles (molecules). The key 1983), 40233.
to chemical analysis is thus the determina- Lasswitz, K.: Geschichte der Atomistik vom
tion of relative atomic weights. Mittelalter bis. Newton, 2 vols. (Hamburg:
This turned out to be a more difficult Voss, 1890).
matter than Dalton had anticipated, and the Nash, L.: The Atomic-Molecular Theory (Cam-
contributions of many other researchers bridge, MA: Harvard University Press,
(notable among them Joseph Louis Gay- 1950).
Lussac (17881850), Amedeo Avogadro Nye, M.J.: Molecular Reality (London:
(17761856) and Stanislao Cannizzaro Macdonald, 1972).
(18261910)) were needed before the atomic van Melsen, A.G.: From Atomos to Atom
model of chemical change was established (Pittsburgh, PA: Duquesne University
to the satisfaction of chemists generally. Press, 1952).
The kinetic theory of gases followed in ernan mcmullin
physics; many of the physical properties of
gases were shown to be derivable from the
hypothesis that gases are made up of vast Augustine of Hippo, St. (354 430)
numbers of molecules in rapid motion. Theologian, born in North Africa.
Despite this convergence of chemistry and Augustine drew his metaphysics from the
physics, empiricists like Mach argued that the Platonic philosophers, who said that the
notable successes of the atomic hypothesis did true God is at once the author of things,
not warrant belief in the actual existence the illuminator of truth, and the giver of
of atoms and molecules. Atomic theory was happiness (City of God 8.5). He knew Latin

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 133

august i ne of hippo, st .
versions of Plotinus and of his disciple and bodies and become animals is altogether
editor Porphyry (ad c.232c.303). These mysterious (City of God 21.10.1).
pagan Platonists Neoplatonists (see His celebrated investigation of time in
neoplatonism) to us were the chief instru- Confessions 11 and City of God 1112 meets
ment of his rescue from Manichean dualism the pagan challenge against creationism.
and from Ciceronian skepticism at the time Why then?, with a response developed from
when, as a 31-year-old teacher in Milan, he Philo Judaeus (c.20 bcad c.50) that God
resumed the Christianity of his childhood made time too; follows Plotinus and antici-
and planned the little African philosophical pates Boethius in a perplexing account of
community whose life was to be cut short eternity; and wrestles with Aristotles puzzle
by his ordination (ad 391) four years later. how times can exist, since they are all past,
His philosophical education was Latin, and future or durationless (Augustines specula-
narrow, enriched during his career as a tive solution, arising from his insight that we
Christian controversialist only by the Bible. measure times by memorizing their length,
According to Augustine there are three is that they are affections of the mind).
natures, i.e., kinds of substance: corpor- His various writings on free will (see
eal, which are mutable in time and place; the extended essay) provided materials
spiritual, mutable in time only; and God, for both parties in the Reformation debates,
immutable (De Genesi ad litteram 8.20.39). for example between Erasmus and Luther,
Souls are not corporeal since they see and which set the scene for modern treatments
judge similitudes which are not corpor- of the subject. He failed to find a consistent
eal; therefore God is not corporeal either (City response to the contrary pressures on him,
of God 8.5). Among non-corporeal beings arguing (e.g., in De correptione et gratia against
are angels and demons, but at most one the Pelagians) that Gods prevenient grace
God since only what is supreme is divine cannot be resisted, but refusing to repudiate
(De vera religione 25.46). Everything is from his earlier argument (e.g., in De libero arbi-
God, since all good things are from God and trio against the Manichees) that some evils
everything is good (De natura boni 3); mir- are, and others punish, sins freely committed.
acles differ from natural events only in not pro-
ceeding by an ordinary route (De Trinitate
3.6.11). The perfectly ordinary course of
nature is the regular (and planned) unfold- Augustines works are in Patrologiae cursus
ing of causal or seminal reasons (De Genesi completus, series latina, ed. J.P. Migne,
ad litteram 9.17.32), which date from the vols. 3247 (Paris, 184455) (PL); many
creation when God completed his work are also in Corpus scriptorum ecclesiastico-
(ibid. 6.11.1819). These reasons do not all rum latinorum (Vienna: Tempsky, 1866 )
necessitate (ibid. 6.15.26). At some places (CSEL), and in Corpus christianorum,
Augustines conception of God seems to series latina (Turnhout, Belgium: Brepols,
combine the two roles, cause of truth and 1953 ) (CCL). Various of his works are
cause of knowledge, assigned by Plato to the translated into English in: A Select Library
form of the good: the latter role makes God of the Nicenc and Post-Nicene Fathers of the
the only teacher (De magistro), illuminator Christian Church, ed. P. Schaff (New York:
of truths as the sun illuminates visible The Christian Literature Co., first series
things (De libero arbitrio 2.13.36); the former 18868; repr. Grand Rapids, MI: Wm B.
makes him truth itself (ibid.). Eerdmans, 197180) (NPNF); Library of
Following Varro (11627 bc), Augustine Christian Classics ed. J. Baillie, J.T. McNeill,
proposed that the question what a man is and H.P. van Dusen (Philadelphia: West-
is the question whether he is both [a body and minster Press, 1953 ) (LCC): Fathers
a soul], or only a body, or only a soul (De of the Church, ed. R.J. Deferrari et al.
moribus ecclesiae catholicae 4.6). He chose (Washington, DC: Catholic University
the first answer, but felt forced to conclude of America Press, 1947 ) (FC); Ancient
that the way in which spirits adhere to Christian Writers, ed. J. Quasten and

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 134

a v e r r oes
J.C. Plumpe (Westminster, MD: Newman (Cambridge: Cambridge University Press,
Press, 1946 ) (ACW); Basic Writings 1967), chs. 217.
of Saint Augustine (New York: Random Sorabji, R.R.K.: Time, Creation and the Con-
House, 1948) (BW). A useful compendium tinuum (London: Duckworth, 1983).
of excerpts in translation is: The Essential
Augustine, ed. V.J. Bourke, 2nd edn. christopher kirwan
(Indianapolis, IN: Hackett, 1974). The
list below is of works cited; numbers
denote volumes. Averroes, [Ibn Rushd] (112698) Spanish-
City of God (De civitate Dei contra paganos, Islamic philosopher who lived in Cordoba
ad 41326): PL 41, CSEL 40, CCL 47 and Seville, a thoroughgoing Aristotelian,
8, and elsewhere; trans. NPNF 2, FC 8, wrote commentaries on most of Aristotles
14, 24; text and translation also in Loeb works, but is better known in Islam as the
Classical Library (London: Heinemann defender of philosophy against the attacks by
and Cambridge, MA: Harvard University al-Ghazali (10581111), in The Incoherence
Press, 196672). of the Philosophers and as a reconciler of
Confessions (Confessiones, ad 397401): PL philosophy and religion.
32, CSEL 33, and elsewhere; trans. NPNF The Aristotelian commentaries were based
1, LCC 7, FC 21, BW, and elsewhere; on excellent translations that gave reliable
text and (old) translation also in Loeb access to Aristotle without Neoplatonic eyes
Classical Library (London: Heinemann; (see alfarabi; neoplatonism), and thus
Cambridge, MA: Harvard University Press, played an important role in the Latin and
1912). Jewish Aristotelian tradition.
De correptione et gratia (ad 426): PL 44; In his Incoherence of the Incoherence
trans. NPNF 5, FC 2. Averroes takes up Ghazalis attacks on
De genesi ad litteram (ad 40114); PL 34, Alfarabi and Avicenna. To safeguard Gods
CSEL 28.1; trans. ACW 412. omnipotence Ghazali had rejected their
De libero arbitrio (ad 388, 3915): PL 32, claim of a necessary connection between
CSEL 74, CCL 29; trans. LCC 6, ACW 22, cause and effect. According to Ghazali, such
FC 59, and elsewhere. necessity is not given in observation. All
De magistro (ad 389): PL 32, CSEL 77, CCL we see is a temporal sequence between, say,
29; trans. LCC 6, ACW 9, FC 59, BW 1, fire and cotton burning. God, the only agent,
and elsewhere. causes the occurrence of fire, the burning of
De moribus ecclesiae catholicae (ad 3879): PL cotton and the coincidence which it becomes
32; trans. NPNF 4, FC 56, BW 1. our habit to expect.
De natura boni (ad 399): PL 42, CSEL 25.2; Against this Averroes argued that to
trans. NPNF 4, LCC 6, BW 1. deny cause is to deny knowledge. It is also
De trinitate (ad 399419): PL 42, CCL 50, to deny human agency and the distinction
50A; trans. NPNF 3, FC 45. between the voluntary and the involun-
De vera religione (ad 391): PL 34, CSEL 77, tary. Further, it violates the view that things
CCL 32; trans. LCC 6, and elsewhere. have a real nature. Finally, if there is no
regularity nor design in creation, we cannot
infer a wise Agent.
b i b l i og rap hy
Resting on Aristotelian grounds. Averroes
Kirwan, C.A.: Augustine (London and New criticized Avicenna for confusing the logical
York: Routledge, 1989). and metaphysical features of being by mak-
Kirwan, C.A.: Augustine on Souls and ing the definitional separation of essence
Bodies, in Logica, mente e persona, ed. and existence characteristic also of existing
A. Alberti (Florence: Olschki, 1990), things, thus espousing an un-Aristotelian
207 41. essentialism (see essence and essentialism).
Markus, R.A.: Later Greek and Early Medi- A similar confusion is said to occur with
eval Philosophy, ed. A.H. Armstrong respect of the numerical and the metaphysical

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 135

avi c enna
one. (See Shehadi, 1982, pp. 93111 for a On the relation between essence and
fairer view of Avicenna.) existence in Avicenna one must distinguish
On the doctrine of creation Averroes three contexts in which these could be related
argues that creation ex nihilo of both world (see essence/accident; essence and essen-
and time does not have Quranic support. On tialism). First, the logical, where in any
the contrary, some verses (11:6, 41:10) definition, say, of a horse, existence must
suggest that matter and time pre-existed be excluded from the essence of a horse.
with God, making Aristotles God consistent Excluded also is any property that is not
with Scripture. part of what a horse is, even universal. For
although a horse qua essence is universal,
i.e., applies to many, being universal is not
part of what makes a horse a horse.
Tahafut al-Tahafut, ed. M. Bouyges (Bey- Second, the metaphysical context: essence
routh: Imprimerie Catholique, 1930); and existence are inseparable in individual
trans. S. van den Bergh The Incoherence of things. While existence and one are dis-
the Incoherence, 2 vols. (London: Luzac, tinct from the meaning of horse, they are
1954). metaphysically part of what makes a horse
this horse, and are not accidental to it qua
b i b l i og rap hy substance.
Third is the theological context. Follow-
Fakhry, M.: A History of Islamic Philosophy ing Alfarabi, Avicenna divides beings into the
(New York: Columbia University Press, possible in itself, though necessary through
1970). another, and the necessary in itself. The
Fakhry, M.: Islamic Occasionalism and Its existence of the former is contingent and
Critique by Averroes and Aquinas (London: its non-existence possible, while the non-
Allen and Unwin, 1958). existence of the Necessary Being is impossible.
Kogan, B.S.: Averroes and the Metaphysics of God gives existence to all contingent beings.
Causation (Albany, NY: SUNY Press, And while existence is a necessary feature
1985). of a thing qua substance, it is accidental
Mehren, I.: Etudes sur la philosophie to it qua contingent.
dAverroes concernant ses rapports avec Avicenna reproduces the emanationist
celle dAvicenna et de Ghazzali, Muson scheme of Alfarabi. The soul being an ema-
VII (18889). nation of the Active Intellect turns to this
Shehadi, F.: Metaphysics in Islamic Philo- intermediary between humans and God for
sophy (Delmar, NY: Caravan Books, 1982). knowledge and mystical illumination.
fadlou shehadi
bibl i ography
Avicenna [Ibn Sina] (9801037) Islamic Fakhry, M.: A History of Islamic Philosophy
philosopher. Avicenna was the most sys- (New York: Columbia University Press,
tematic and sophisticated, as well as the 1970).
most influential of Islamic philosophers, Goichon, A.M.: La Distinction de lessence
although much of his thought is already in et de lexistence daprs Ibn Sina (Paris:
Alfarabi. Descle, 1937).
Being is a primary intuition of the soul. Goodman, L.: Avicenna (London: Routledge,
It can be known without the mediation of 1992).
any other concept, and it cannot be defined Gutas, D.: Avicenna and the Aristotelian Tradi-
without circularity. Even thing, its coequal tion: Introduction to Reading Avicennas Phi-
in extension, presupposes being and cannot losophical Works (Leiden: E.J. Brill, 1988).
be used in explaining it without circularity. Shehadi, F.: Metaphysics in Islamic Philosophy
Being is the most general concept; its oppo- (Delmar, NY: Caravan Books, 1982).
site is the absolute nothing. fadlou shehadi

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 136

avowals The verb to avow has been meaningful, and synthetic statements were
adopted by many philosophers of mind as understood to be ultimately verifiable in sense
the translation of the German verb ussern. experience. One intention of the verification
The usual alternative translations are to criterion was to rule out as meaningless the
express or to utter. wordy, but empirically uncheckable claims
In Wittgensteins later work avowals of metaphysicians in the Hegelian tradition.
are the keystone of a new philosophy of But while the criterion did allow those who
mind, founded on the rejection of the held it to dismiss much of Hegels Science of
Cartesian idea that a person discloses the Logic (181216), say, without the trouble of
contents of his mind by identifying inner reading it, it had the not so welcome effect
objects and describing them (see descartes). of rendering meaningless such unverifiable
According to Wittgenstein, an avowal of an statements as Every event has a cause or
intention is not based on a self-examination even For every action, there is an equal
which parallels the investigation of the and opposite reaction. Even the proposal
world around us: it is only marginally liable Ayer made to treat these statements as
to error, and in certain cases is an artificial heuristic aids to living and to scientific
expression of the intention replacing a enquiry implicitly admitted their meaning-
natural one (e.g., a raised fist). Each of fulness. For reasons outlined in later edi-
these three points makes its contribution to tions of Language, Truth and Logic the
the new philosophy of mind, which some of verification criterion was dropped by Ayer,
Wittgensteins followers have accepted in and metaphysics, at least in a certain sense,
its entirety and which, perhaps, nobody re-admitted to the canon of meaningful
can totally reject. But the third point may be discourse.
the most important one, because it shows Ayer remained skeptical to the end of
how language can develop directly out of his life concerning the pretensions of some
behavior which antedates it. This makes it metaphysicians to inform us of any sup-
possible to explain how we can learn, and rasensible reality, or to delineate the most
communicate with, mentalistic language, general characteristics of being as such.
which were things that remained mysterious Nevertheless, in another sense, in much of
when intentions, feelings, and so on, were his philosophy subsequent to Language,
treated as private objects. So it prepares the Truth and Logic he was engaged in meta-
way for a naturalistic, rather than an intel- physical enquiry. Although the motivation
lectualist answer to skepticism about other of his philosophy was largely epistemo-
minds. logical, concerning the status of our claims
to knowledge, many of its conclusions were
metaphysical, concerning what there actu-
b i b l i o g r ap hy
ally is. Indeed, throughout the whole of his
Malcolm, N.: Nothing Is Hidden (Oxford: philosophical career, Ayer was concerned
Blackwell, 1986), esp. ch. 8. about the nature of physical objects
Wittgenstein, L.: Philosophical Investigations, in particular. There is, in fact, an interesting
trans. G.E.M. Anscombe, 3rd edn. (London: transition in Ayers work from the phenom-
Macmillan, 1969). enalistic stance (see phenomenalism) of
david pears Language, Truth and Logic to the sophisti-
cated realism of The Central Questions of
Philosophy (1976).
Ayer, Alfred Jules (191089) British Ayer always rejected what he called
philosopher. Ayer was famous for the naive realism. That is to say, he denied that
attack on metaphysics in his Language, objects are just as they appear. He was fur-
Truth and Logic (1936). According to the ther convinced that there was an inference
verification criterion of meaning (see logical involved in any transition from appearance
positivism; principle of verifiability), only to object, on the grounds that there is always
analytic or synthetic statements were more involved in assertions about objects

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 137

ayer, al f red j ules

than is available to us in our perceptions. from percepts to objects. He also insists that
What, then, is the relation between the under the dominion of the theory our ima-
objects and the perceptions? gination has led us to, the existence of phys-
Ayer came to reject phenomenalism on the ical objects becomes a matter of objective
grounds that the percepts that are presented fact, and he denies the possibility of any
even to the totality of observers are too straightforward phenomenalist reduction.
scanty to answer to our conception of the At the same time, the suspicion remains
physical world. He also rejected the causal that there is a sense in Ayers story in
theory of perception, largely because that which sense qualia (see sensa), rather than
theory would render the causes of our objects, are the basic stuff of the world. On
perceptions unobservable occupants of an this point Ayer himself would probably
unobservable space. Instead he proposed have said as he did on related issues that
what he called a construction, in which the the matter is ultimately undecidable. It is
subject of experience is initially presented just a matter of decision, based on experien-
with a mass of sensory data; he then begins tial coherence of any story we tell. If this was
to perceive patterns within this data, which indeed his attitude, it would certainly be in
tend to cluster in predictable ways. At a a direct line of descent from his earlier repu-
certain stage in the process, the clusters diation of metaphysics as meaningless.
or visuo-tactual continuants as Ayer calls
them are cut loose from their moorings and wri t i ngs
regarded as having an existence quite inde-
pendent of their being perceived. The Central Questions of Philosophy (Har-
Our common-sense view of the world is mondsworth: Pelican, 1976).
thus seen as a theory relative to our percep- Language, Truth and Logic (1936); 2nd edn.
tions; but it is a theory, which once accepted, (London: Victor Gollancz, 1946).
ontologically downgrades the perceptions The Problem of Knowledge (Harmondsworth:
on which it was originally based. Pelican, 1956).
It cannot be said that everything about this
construction is clear. Ayer denies that he bibl i ography
is telling a psychological story about how
children actually learn about the physical Foster, J.: Ayer (London: Routledge and Kegan
world, but he insists that an exercise of the Paul, 1985).
imagination is required in the passage anthony ohear

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 138

Baker, Lynne Rudder (1944 ) defends if they lose their essential properties (includ-
a position called Practical Realism, in which ing relational ones).
she intends to do justice to the common
sense conception of reality. In her view,
wri t i ngs
reality as disclosed by everyday life and
natural language is not second-class, to be Explaining Attitudes. A Practical Approach to
replaced by science in the long run. The the Mind (Cambridge: Cambridge Univer-
world of intentional agents, social institutions, sity Press, 1995).
medium-sized natural objects, and Artefacts The Metaphysics of Everyday Life: An Essay in
is ontologically irreducible. Practical Realism (Cambridge: Cambridge
In the mind/body debate Baker is a lead- University Press, 2007).
ing critic of reductionist (see the extended Persons and Bodies. A Constitution View
essay on the mind/body problem; reduc- (Cambridge: Cambridge University Press,
tionism) and eliminativist approaches to 2000).
mental states. She rejects the idea that Saving Belief. A Critique of Physicalism (Prin-
mental states are identical with, constituted ceton, NJ: Princeton University Press,
by, supervene on (see supervenience), or 1987).
grounded in brain states. They are con-
ceived as global states of whole persons. bibl ography
The criterion for having a belief is the
truth of relevant Counterfactuals. Meijers, A.W.M. (ed.): Explaining Beliefs:
Whether a person S has a particular belief Lynne Rudder Baker and Her Critics
is determined by what S does, says, and (Stanford, CA: CSLI Publications, 2001).
thinks, and what S would do, say, and think anthonie meijers
in various circumstances, where what S
would do may itself be specified intention-
ally (Baker, 1995, p. 154). Having a belief bare particular Bare particulars are the
is a relational property, where Relations individuators of concrete objects. The basis
are seen as both real and causally efficacious. for this contention can be articulated only
For Baker a person (see persons and if the problem of individuation is placed in
personal identity) is most fundamentally the broader context of the issues raised by the
a being with a first-person perspective. She relationship between a concrete object and
conceives the relation between a person its properties. For it is an antecedent com-
and her body in terms of the general meta- mitment to property realism (see universals)
physical relation of constitution. Persons and anti-essentialism (see essence/accident;
are constituted by, but not identical to, or essence and essentialism) that provide key
separate from their bodies. Constitution is premises in the argument for bare particulars.
a relation between individual things or An account of the relationship must
aggregates, not between mereological parts explain two features: (1) some objects have
(see part/whole) or between properties. properties in common with other objects;
It is also a contingent relation: persons, yet (2) no object is identical with any other
artefacts or natural objects go out of existence object. According to realism, the properties

138 A Companion to Metaphysics, Second Edition Edited by Jaegwon Kim, Ernest Sosa, and Gary S. Rosenkrantz
2009 Blackwell Publishing Ltd. ISBN: 978-1-405-15298-3
9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 139

bare particul ar
of objects are universals. Hence, if two is impossible for the same bare particular to
objects have a property in common, say be a constituent of two different momentary
redness, the redness of one object is identical concrete objects. Since difference in bare
with the redness of the other. The realist has particulars is sufficient to insure difference
three basic options available for explaining in any two momentary objects, even two
(2): (a) non-identical objects differ in the with all properties in common, the theory
universals they instantiate; (b) non-identical is not committed to the necessary truth of
objects differ in some feature other than the the identity of indiscernibles. There are two
universals they instantiate; and (c) the non- important features of the relationship be-
identity of objects is primitive. tween a bare particular and the universals
Option (a) is associated with the view, it instantiates. Universals exist only if
often called the bundle theory, that concrete instantiated by some bare particular and
objects are complex entities whose sole con- bare particulars exist only if they instantiate
stituents are universals. Proponents of bare some universal. Neither is capable of inde-
particulars, such as Bergmann, reject this pendent existence. Furthermore, the instanti-
view on the grounds that (i) it is committed ation of a universal by a bare particular
to the necessary truth of the principle of the requires a nexus. A nexus, unlike a relation,
identity of indiscernibles; but (ii) the can unite two distinct entities into a complex
principle is not a necessary truth. Option without some further relation. Hence, a
(c) is rejected on the grounds that the non- momentary object is a complex entity, often
identity of objects is insufficiently funda- called a fact, whose constituents must include
mental to be taken as primitive. a bare particular, a nexus, and a universal.
Theories exercising option (b) fall into There are two familiar objections to bare
two broad categories depending on their particulars. The first alleges that theories
explanation of identity through time. Con- invoking them are incompatible with an
crete objects typically change their proper- empiricist epistemology (see empiricism).
ties over time while remaining the same This objection rests on the claim, articu-
object. One explanation is that there is a lated in Allaire (1963), that, according to
constituent of every object which endures empiricism, the basic entities of an ontolo-
through time and remains unchanged despite gical theory must be entities with which we
the changes in the object itself. Furthermore, are directly acquainted. This claim, how-
this enduring constituent, often called a ever, ties empiricism to phenomenalism in
substance, has some of its properties essen- a manner few contemporary empiricists
tially. Since proponents of bare particulars are would accept. The second alleges that the the-
anti-essentialists, they reject this explana- ory is incoherent since its central thesis,
tion. Instead, they maintain that a concrete Bare particulars instantiate properties,
object is a temporal series of momentary is equivalent to Entities which have no
objects which stand in some complex con- properties have properties which is self-
tingent lawlike relations (see temporal parts, contradictory. But, as Loux (1978) points
stages). The endurance of an object through out, a bare particular is not an entity which
time is explained in terms of the obtaining has no properties but one none of whose
of these contingent relations among the properties is essential. Loux, however, main-
momentary objects. Change is explained by tains that the latter thesis is itself problem-
differences in the properties of successive atic since bare particulars have essentially the
momentary objects. property of having no properties essentially. This
Bare particulars are the individuators of contention rests on the assumption that
momentary concrete objects. Such particu- the predicate has no properties essentially
lars differ from substances in two significant designates a property, and this is denied by
ways: (1) they are momentary entities rather Bergmann (1967) and Armstrong (1978).
than continuants; and (2) they have no Critics also allege that bare particulars
essential properties. The particularity of are unnecessary and have little explanatory
bare particulars consists in the fact that it value. Proponents of the theory maintain

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 140

b a s i c a ctio n
that an adequate account of the non-identity The term was introduced in Danto (1963),
of two concrete objects must ground it where the following analysis is offered: B is
in some difference in the constituents of a basic action of a if and only if (i) B is an action
the objects. Yet, they also maintain that the and (ii) whenever a performs B, there is
non-identity of two bare particulars is a no other action A performed by a such that
primitive fact. It is not evident, as Hochberg B is caused by A. This analysis fails for a
(1965) contends, that explaining the non- variety of reasons (see Goldman, 1970;
identity of objects in terms of constituents Hornsby, 1980). The fundamental problem
whose non-identity is primitive is more (or a symptom thereof) is that the difference
illuminating than maintaining that the between basic and non-basic actions does
non-identity of the objects themselves is not hinge on causal transactions of the
primitive. Proponents of the bundle theory, kind specified in (ii). Typically, when an
such as Russell (1948) and Casullo (1988). agent does one thing by doing another, the
argue that it is a contingent truth that latter is more basic than the former. If by
concrete particulars are complexes of uni- moving her right index finger upward Jane
versals and that this view does not com- flips a switch, and by flipping the switch
man them to the necessary truth of the illuminates the room, Janes moving her
identity of indiscernibles. They also maintain finger upward is more basic than her flipping
that the purported counterexamples to the switch, and both are more basic than
the necessary truth of the principle are her illuminating the room. However, Janes
questionable. moving her finger does not cause her flipping
the switch. (It does cause the switchs mov-
ing upward, but the latter event must be
b i b l i og rap hy
distinguished from Janes flipping the switch.)
Allaire, E.B.: Bare Particulars, Philo- Nor does her flipping the switch cause her
sophical Studies 14 (1963), 18. illuminating the room. Indeed, Janes flipping
Armstrong, D.M.: Universals and Scientific the switch and her illuminating the room
Realism, 2 vols., Vol. 1, Nominalism and might not be caused by any action of hers.
Realism (Cambridge: Cambridge University Still, they are not basic actions. Just how
Press, 1978). Janes actions are related is controversial.
Bergmann, G.: Realism: A Critique of Brentano Some philosophers say that they are the same
and Meinong (Madison, WI: University of action under different descriptions; others
Wisconsin Press, 1967). that they are distinct actions related by
Casullo, A.: A Fourth Version of the Bundle causal generation, as opposed to causa-
Theory, Philosophical Studies 54 (1988), tion; yet others that the more basic actions
12539. are components of the less basic (see action
Hochberg, H.: Universals, Particulars, and theory).
Predication, Review of Metaphysics 19 In the same vein, an action caused by
(1965), 87102. another action of the agents might never-
Loux, M.: Substance and Attribute (Dordrecht: theless be a basic action. Suppose that
Reidel Publishing Company, 1978). Janes turning on her computer caused a
Russell, B.: Human Knowledge: Its Scope and power surge that cut off the electricity in
Limits (New York: Simon and Schuster, her study, with the effect that, moments
1948). later, she illuminated her utility room so
albert casullo that she could see the fuse box. Presumably,
Jane illuminated the room by performing
some basic action or other; and this basic
basic action Basic actions, broadly char- action has Janes turning on her computer
acterized, differ from non-basic actions in as a causal antecedent.
not being performed by way of the agents Influential analyses of basic act-types
performing another action. and act-tokens that avoid these difficulties

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 141

bei ng and bec oming

are offered in Goldman (1970, pp. 67, 72). teleological counterparts prevents an epis-
Goldmans proposals are framed in terms temic regress. If, to a under any description
of his own theory of act-individuation. at all, we always had to a under a causally
Neutral approximations are: more basic description, we would never act;
similarly, if intentionally a-ing-under-d, for
An act-type, b, is a basic act-type for an any d, required the possession of meansend
agent, S, at a time if, and only if, (a) given knowledge identifying a under a teleologically
normal conditions, if S wanted on bal- more basic description, we would be lost
ance to do a b, S would do so; and (b) the in thought.
truth of (a) does not depend upon Ss
cause-and-effect knowledge nor upon
any knowledge of Ss of the form x-ing bi bliography
may be done by y-ing. Danto, A.C.: Basic Actions, American Philos-
An action, b, done by S is a basic act-token ophical Quarterly 2 (1965), 141 8.
if, and only if, (a) b instantiates an act-type Danto, A.C.: What We Can Do, Journal of
that is basic for S; (b) Ss b-ing is caused, Philosophy 60 (1963), 435 45.
in the characteristic way, by an action-plan Goldman, A.I.: A Theory of Human Action
of S; (c) S does not b by doing anything, (Englewood Cliffs, NJ: Prentice-Hall, 1970).
a, that satisfies clauses (a) and (b). Hornsby, J.: Actions (London: Routledge and
Kegan Paul, 1980).
Some philosophers have denied that there alfred r. mele
are basic actions. Suppose that Janes illu-
minating the room which is not a basic
action is the same action (under another being and becoming The idea of being
description) as her moving her finger. Given functions primarily in three contrast-
this identity, one might argue, Janes mov- contexts: (1) being/non-being, with the con-
ing her finger is not a basic action either. trast of the non-existent or unreal; (2) being/
However, assuming the theory of action seeming, with the contrast of that which is
presupposed by this argument, the notion of merely suppositional, imaginary or visionary;
basic action may be relativized to action- and (3) being/becoming with a view to the
descriptions. Janes a-ing, under the descrip- origination of that which is not or not
tion illuminating the room, is not a basic heretofore. Becoming in this third context
action; but her a-ing might be basic under (Greek: einai/genesis) is a matter of a shift from
another description perhaps moving her non-being to being, which can be either
finger. absolute via a transition from non-being
An important distinction between causally to being (an origination) or its reverse (an
and teleologically basic action, framed in annihilation) or relative via a change from one
terms of action-descriptions, is developed state or condition of being to another.
in Hornsby (1980). Hornsby contends, con- With respect to being/becoming, the
cerning any pair of descriptions, d and d, of ancient Greeks puzzled over the aporetically
an action a, that d is causally more basic, if inconsistent triad (see aporia): (1) absolute
the effect that is introduced by [a under becoming involves a transition from non-
description d] causes the effect that is intro- being to being; (2) absolute becoming occurs:
duced by [a under description d] (1980, there are some things that exist (now or
p. 71). By comparison, d is teleologically sometime) that did not do so at an earlier
more basic than d if, and only if, in virtue time; (3) becoming presupposes being: only
of as occurrence, a statement to the effect something that is already in being can
that S intentionally a-ed-under-d by a-ing- undergo any sort of alteration or transition.
under-d is true (ibid., p. 78). The aptness of These propositions are incompatible as they
causally basic descriptions blocks a vicious stand. For, by (2), there is something, say
causal regress, while the appropriateness of x, that instantiates absolute becoming

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 142

b e i n g and b eco ming

something that exists but yet did not do so overthrowing the static mathematics of the
at an earlier juncture and only came into Greeks as reflected in Zenos paradoxes.
being at some particular time. But then, Leibniz envisioned the prospect of bringing
by (3), x must have had some pre-existent the domain of becoming into the range of
state prior to that time, contrary to (1)s a rigorous science. His monads (see monad,
stipulation of the nature of absolute becom- monadology) are centers of activity, preserv-
ing. The dialectic of this perplexity is encap- ing their programmatic identity of lawful
sulated in the paradoxes of Zeno. development through an ever dynamic
Different theorists resolved the problem course of perpetual becoming.
differently. Heraclitus (c.535 c.475 bc) Harking back to the preoccupation of early
rejected (3), maintaining that becoming is all- Greek philosophers with being/becoming,
predominant and exhaustive: only becoming the German philosopher Heidegger rep-
occurs and nothing is (has being) but every- roached the post-Platonic philosophical tra-
thing is perpetually becoming. The Eleatics, dition with a neglect (forgetfulness) of being
by contrast, rejected (2), denying all becom- (Dasein). According to Heidegger, philoso-
ing and insisting that everything just phers have been so concerned with explana-
unchangingly is, with change relegated to tion with pursuing a theoretical account
the condition of an illusion of sorts. The of how things have become what they are
atomists (see atomism) Leucippus (fifth cen- that they neglect the immediate experience