blackwell-Companion-to-Metaphysics.pdf

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

A Companion to Metaphysics

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 today’s 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: 1. The Blackwell Companion to Philosophy, Second Edition Edited by Nicholas Bunnin and Eric Tsui-James

2. A Companion to Ethics Edited by Peter Singer

3. A Companion to Aesthetics, Second Edition Edited by Stephen Davies, Kathleen Higgins, Robert Hopkins, Robert Stecker, and David E. Cooper

4. A Companion to Epistemology Edited by Jonathan Dancy and Ernest Sosa

5. A Companion to Contemporary Political Philosophy (two-volume set), Second Edition Edited by Robert E. Goodin and Philip Pettit

6. A Companion to Philosophy of Mind Edited by Samuel Guttenplan

7. A Companion to Metaphysics, Second Edition Edited by Jaegwon Kim, Ernest Sosa and Gary S. Rosenkrantz

8. A Companion to Philosophy of Law and Legal Theory Edited by Dennis Patterson

9. A Companion to Philosophy of Religion Edited by Philip L. Quinn and Charles Taliaferro

10. A Companion to the Philosophy of Language Edited by Bob Hale and Crispin Wright

11. A Companion to World Philosophies Edited by Eliot Deutsch and Ron Bontekoe

12. A Companion to Continental Philosophy Edited by Simon Critchley and William Schroeder

13. A Companion to Feminist Philosophy Edited by Alison M. Jaggar and Iris Marion Young

14. A Companion to Cognitive Science Edited by William Bechtel and George Graham

15. A Companion to Bioethics Edited by Helga Kuhse and Peter Singer

16. A Companion to the Philosophers Edited by Robert L. Arrington

17. A Companion to Business Ethics Edited by Robert E. Frederick

18. A Companion to the Philosophy of Science Edited by W. H. Newton-Smith

19. A Companion to Environmental Philosophy Edited by Dale Jamieson

20. A Companion to Analytic Philosophy Edited by A. P. Martinich and David Sosa

21. A Companion to Genethics Edited by Justine Burley and John Harris 22. A Companion to Philosophical Logic Edited by Dale Jacquette

23. A Companion to Early Modern Philosophy Edited by Steven Nadler

24. A Companion to Philosophy in the Middle Ages Edited by Jorge J. E. Gracia and Timothy B. Noone

25. A Companion to African-American Philosophy Edited by Tommy L. Lott and John P. Pittman

26. A Companion to Applied Ethics Edited by R. G. Frey and Christopher Heath Wellman

27. A Companion to the Philosophy of Education Edited by Randall Curren

28. A Companion to African Philosophy Edited by Kwasi Wiredu

29. A Companion to Heidegger Edited by Hubert L. Dreyfus and Mark A. Wrathall

30. A Companion to Rationalism Edited by Alan Nelson

31. A Companion to Ancient Philosophy Edited by Mary Louise Gill and Pierre Pellegrin

32. A Companion to Pragmatism Edited by John R. Shook and Joseph Margolis

33. A Companion to Nietzsche Edited by Keith Ansell Pearson

34. A Companion to Socrates Edited by Sara Ahbel-Rappe and Rachana Kamtekar

35. A Companion to Phenomenology and Existentialism Edited by Hubert L. Dreyfus and Mark A. Wrathall

36. A Companion to Kant Edited by Graham Bird

37. A Companion to Plato Edited by Hugh H. Benson

38. A Companion to Descartes Edited by Janet Broughton and John Carriero

39. A Companion to the Philosophy of Biology Edited by Sahotra Sarkar and Anya Plutynski

40. A Companion to Hume Edited by Elizabeth S. Radcliffe

41. A Companion to the Philosophy of History and Historiography Edited by Aviezer Tucker

42. A Companion to Aristotle Edited by Georgios Anagnostopoulos

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

A Companion to Metaphysics Second Edition

Edited by

JAEGWON KIM, ERNEST SOSA, and

GARY S. ROSENKRANTZ

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. Rosenkrantz Edition history: Blackwell Publishers Ltd (1e, 1996) Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell’s publishing program has been merged with Wiley’s global Scientific, Technical, and Medical business to form Wiley-Blackwell. 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 www.wiley.com/wiley-blackwell. 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. Metaphysics–Dictionaries. I. Kim, Jaegwon. II. Sosa, Ernest. III. Rosenkrantz, Gary S. BD111.C626 2009 110′.3– dc22 2008032199 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

Contents

List of Contributors

vii

Introduction

xiii

Part I

Extended Essays

Part II Metaphysics From A to Z Index

1 95 637

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

Contributors

Felicia Ackerman Brown University

John Bigelow Monash University, Australia

Robert Ackermann University of Massachusetts, Amherst

Akeel Bilgrami Columbia University

Marilyn McCord Adams Christ Church, Oxford

John Biro University of Florida

Robert Merrihew Adams Yale University and Mansfield College, Oxford

Simon Blackburn University of Cambridge and the University of North Carolina at Chapel Hill

Jan A. Aertsen Thomas-Institut at the University of Cologne, Germany C. Anthony Anderson University of California, Santa Barbara Richard E. Aquila University of Tennessee D.M. Armstrong University of Sydney, Australia Keith Arnold University of Ottawa, Canada Bruce Aune University of Massachusetts, Amherst Thomas Baldwin University of York

Ned Block New York University Paul A. Boghossian New York University M.B. Bolton Rutgers University Michael E. Bratman Stanford University Harold I. Brown Northern Illinois University Douglas Browning University of Texas at Austin Panayot Butchvarov University of Iowa

George Bealer Yale University

Robert E. Butts†

Frederick Beiser Syracuse University

Alex Byrne Massachusetts Institute of Technology vii

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

c on t r i buto r s Steven M. Cahn City University of New York, Graduate School

Dagfinn Føllesdal Stanford University and University of Oslo, Norway

Keith Campbell University of Sydney, Australia

Matthew Davidson California State University, Bernardino

Albert Casullo University of Nebraska, Lincoln

Graeme Forbes University of Colarado at Boulder

Peter Caws George Washington University

Richard Fumerton University of Iowa

Arindam Chakrabarti University of Hawaii at Manoa John J. Compton Vanderbilt University, Tennessee John Corcoran State University of New York at Buffalo John Cottingham University of Reading Edwin Curley University of Michigan Helen Daly University of Arizona Harry Deutsch Illinois State University Cora Diamond University of Virginia Alan Donagan†

Richard M. Gale University of Pittsburgh Richard Gallimore University of North Carolina at Greensboro Patrick Gardiner† Brian Garrett Australian National University Don Garrett New York University Rolf George University of Waterloo, Canada Roger F. Gibson Washington University in Saint Louis Carl Ginet Cornell University Peter Godfrey-Smith Harvard University Thomas A. Goudge†

Rolf A. Eberle University of Rochester, New York

Jorge J.E. Gracia State University of New York at Buffalo

Catherine Z. Elgin Harvard University

John Greco Saint Louis University

Fred Feldman University of Massachusetts, Amherst

Reinhardt Grossmann Indiana University

viii

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

cont ri but ors Rick Grush University of California, San Diego

Paul Horwich New York University

Charles Guignon University of South Florida

Paul Hovda Reed College, Oregon

Bob Hale University of Sheffield

M.J. Inwood Trinity College, Oxford

Michael Hallett McGill University, Montreal

Frank Jackson Australian National University

Chad D. Hansen University of Hong Kong

Janine Jones University of North Carolina at Greensboro

Ross Harrison King’s College, Cambridge

Lynn S. Joy University of Notre Dame

W.D. Hart University of Illinois, Chicago

Robert Kane University of Texas at Austin

John Heil Washington University in Saint Louis

Christopher Kirwan Exeter College, Oxford

Mark Heller Syracuse University

Arnold Koslow Brooklyn College

Risto Hilpinen University of Miami

John Lachs Vanderbilt University, Tennessee

Eli Hirsch Brandeis University, Massachusetts

Karel Lambert University of California, Irvine

Herbert Hochberg University of Texas at Austin

Stephen Leeds University of Wisconsin, Milwaukee

Joshua Hoffman University of North Carolina at Greensboro

Keith Lehrer University of Arizona and University of Miami

Christopher Hookway University of Sheffield

Ramon M. Lemos†

James Hopkins King’s College, London

Jarrett Leplin University of North Carolina at Greensboro

Terence E. Horgan University of Arizona

Ernest LePore Rutgers University ix

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

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

Trenton Merricks University of Virginia

Barry Loewer Rutgers University

Phillip Mitsis New York University

Lawrence B. Lombard Wayne State University, Michigan Douglas C. Long University of North Carolina at Chapel Hill E.J. Lowe Durham University William Lyons Trinity College, Dublin Charles J. McCracken Michigan State University Neil McKinnon Monash University, Australia Brian P. McLaughlin Rutgers University Ernan McMullin University of Notre Dame Joseph Margolis Temple University, Pennsylvania Ned Markosian Western Washington University, Washington

J.M. Moravcsik Stanford University Alexander P.D. Mourelatos University of Texas at Austin Kevin Mulligan University of Geneva, Switzerland Daniel Nolan University of Nottingham Martha C. Nussbaum University of Chicago David S. Oderberg University of Reading Anthony O’Hear University of Buckingham Dominic J. O’Meara University of Fribourg, Switzerland Walter R. Ott, Jr. Virginia Polytechnic Institute and State University George S. Pappas Ohio State University

Gareth B. Matthews University of Massachusetts, Amherst

Terence Parsons University of California Los Angeles and University of California, Irvine

Anthonie Meijers Delft University of Technology, Netherlands

David Pears Christ Church, Oxford

Alfred R. Mele Florida State University

Terence Penelhum University of Calgary, Canada

Joseph Mendola University of Nebraska at Lincoln

Alvin Plantinga University of Notre Dame

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

cont ri but ors Ruth Anna Putnam Wellesley College, Massachusetts Diana Raffman University of Toronto Peter Railton University of Michigan Andrew J. Reck Tulane University, Louisiana Nicholas Rescher University of Pittsburgh Thomas Ricketts University of Pittsburgh Richard Robin Mount Holyoke College, Massachusetts Gideon Rosen Princeton University

Wesley C. Salmon† David H. Sanford Duke University Geoffrey Sayre-McCord University of North Carolina at Chapel Hill Richard Schacht University of Illinois at Urbana-Champaign Frederick F. Schmitt Indiana University Richard Schmitt Brown University George Schumm Ohio State University Jorge Secada University of Virginia

Jay F. Rosenberg†

Charlene Haddock Seigfried Purdue University

Gary S. Rosenkrantz University of North Carolina at Greensboro

David Shatz Yeshiva University, New York

David M. Rosenthal City University of New York, Graduate School

Fadlou Shehadi Rutgers University

David-Hillel Ruben Birbeck College, London and New York University in London Paul Rusnock University of Ottawa, Canada Nils-Eric Sahlin Lund University, Sweden

Donald W. Sherburne Vanderbilt University Sydney Shoemaker Cornell University Peter Simons University of Leeds Lawrence Sklar University of Michigan

R.M. Sainsbury University of Texas at Austin and King’s College, London

John Skorupski University of St Andrews

Nathan Salmon University of California, Santa Barbara

Robert C. Sleigh, Jr. University of Massachusetts, Amherst xi

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

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

Michael Tye University of Texas at Austin

Quentin Smith Western Michigan University

James Van Cleve University of Southern California

Elliott Sober University of Wisconsin, Madison

J. David Velleman New York University

Roy A. Sorensen Dartmouth College

Georgia Warnke University of California, Riverside

Robert Stalnaker Massachusetts Institute of Technology

Joan Weiner Indiana University

Howard Stein University of Chicago

Nicholas White University of Utah

Guy Stock University of Dundee Richard Swinburne University of Oxford Paul Teller University of California, Davis Neil Tennant Ohio State University Amie L. Thomasson University of Miami

S.G. Williams Worcester College, Oxford Kenneth P. Winkler Yale University Kwasi Wiredu University of South Florida Allen W. Wood Stanford University Larry Wright University of California, Riverside

James E. Tomberlin†

Edward N. Zalta Stanford University

Martin M. Tweedale University of Alberta, Canada

Dean Zimmerman Rutgers University

xii

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

Introduction 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 difficulties 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 philosophy; 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 pseudoquestions 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 Guttenplan’s A Companion to the Philosophy of Mind, for example, or in Peter Singer’s A Companion to Ethics, but which should be treated in this Companion, if only briefly and for the sake of a more complete and selfcontained 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 analytic 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 contributions 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

xiii

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

Part I

Extended Essays

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

Causation Making something happen, allowing or enabling something to happen, or preventing something from happening. Mental and extra-mental occurrences, of all spatial and temporal dimensions, great and small, have causes and are causes. Our awareness of the world and our action within the world depends at every stage on causal processes. Although not all explanations are causal, anything that can be explained in any way can be explained causally. Like other metaphysical concepts, the concept of causation applies very broadly. Yet this fundamental concept continues to elude metaphysical understanding. While there is some general philosophic agreement about causation, there is also considerable disagreement. Causal theories of knowledge, perception, memory, the mind, action, inference, meaning, reference, time, and identity through time, take a notion as fundamental that philosophers understand only incompletely. HUME is the dominant philosopher of cause and effect. A running commentary on Hume’s views and arguments, pro and con, could cover most contemporary philosophical concerns with causation (Hume, 1739, esp. Bk. I, Pt. III; Hume, 1748, esp. sects. IV, V, VII). According to Hume, it is not the experience of an individual causal transaction, but experience of other transactions, relevantly similar, that provides what causation involves in addition to priority and contiguity. Experiences of regularities or constant conjunctions condition our expectations. We project our conditioned feelings of inevitability on external objects as a kind of necessity that resides in the objects themselves (see Hume, 1748, sect. VII). Limitations of space preclude extensive quotation and discussion of these and other primary texts. A number of paragraphs in this entry begin with the statement of a view about causation. The next sentence then classifies the view as prevailing, majority, controversial, or minority. Some of these classifications may themselves be controversial. Their purpose is only to help organize the entry. Continuous causal paths connect causes with their effects. This is a prevailing view.

Causes and effects are often not contiguous. A switch on the wall is distant from the electric light overhead that it controls. Pulling a button on an alarm clock makes it ring six hours later. The New York performance of three musicians in 1937 contributes causally to what one hears on the Perth radio in 2007. Although intervals of space, time, or space–time separate the causes and effects in these examples, spatiotemporally continuous causal paths connect them. The path has no spatial or temporal gaps or breaks. (A rigorous definition of continuity requires the notion of a limit found in calculus textbooks.) The path is causal because for any two positions, a and c, on the path, there is an intermediate position b on the path such that either something at a causes something at b that causes something at c, or the causation runs in the other direction, cba. An explanation of what constitutes a causal path that does not use the notion of causation would serve as a reductive definition of causation. The explanation above, which uses the notion of causation explicitly, serves only to state a spatio-temporal necessary condition of causation. Causes and effects are events. This is a majority view (see Davidson, 1980). Idiomatic speech often mentions something other than a change, or non-change, or occurrence, as a cause or effect, as in “Richard makes me furious.” The question is whether an available paraphrase such as “Reading what Richard writes makes me become furious” brings events back into the picture as causes and effects. If both causes and effects are always of the same kind, then causal paths can continue indefinitely both from the past and into the future. On the other hand, the strategy of reducing all causal statements by paraphrase to statements about events does not convince philosophers who hold that sometimes facts, properties, or aspects of events are irreducible relata of causal relations (see Sanford, 1985). Some philosophers who concentrate on questions of agency and freedom entertain views of agent causation: in human action a person is an irreducible cause (see action theory). Although Lucy’s putting on her 3

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 causation, the ultimate cause of Lucy’s shoes being put on is Lucy herself. Causation is the transfer of something from cause to effect. This is a controversial view. In one version of this view, causation transfers some quantity subject to a conservation law of physics. Hans reichenbach propounded and Wesley salmon developed another version in terms the mark transmission of a “mark”, a modification that satisfies certain requirements. The transmission of a mark between processes is a transmission of structure. There are clear positive instances of this view. One controversy involves the generalization of these instances. Another questions whether the application of a notion such as “mark” requires some prior causal commitment. There is no element of genuine a priori reasoning in causal inference. This is a majority view. Most philosophers believe that Hume refuted the rationalists (see rationalism) before him (such as Spinoza, Descartes and, on this issue, Hobbes) and the idealists after him (such as McTaggart and Blanshard) who hold that causation is intrinsically intelligible. Given a determinate event, according to Hume, anything might happen next, so far as reason and logic are concerned. “The contrary of every matter of fact is still possible; because it can never imply a contradiction” (Hume, 1748, p. 25). Cause and effect are distinct existences, and “the mind never perceives any real connexion among distinct existences” (Hume, 1739, p. 636). Reason by itself cannot predict what will happen next after one billiard ball bumps into another. But from what should one attempt to make such predictions, from descriptions of the events in question? If so, which logical relations do or do not obtain will depend on the nature of the description. Any event has logically independent descriptions, and any two events have descriptions that are not logically independent (see Davidson, 1980, essay 1). The view that there is at least sometimes an intelligible connection between cause and effect does not rely on inventing clever descriptions. Rather, it concedes a lot to Hume without conceding everything. Just 4

from observing its sensible qualities, we cannot figure out a thing’s causal capacities. And when we do come to believe, from a much broader experience, what they are, our evidence does not entail our conclusion. It is still logically possible that anything will happen next. Our beliefs about the physical properties of belts and pulleys are fallible and based on more than an initial visual impression. Still, given the physical properties of the belt and pulley, the spatial relations between them, and the assumption that the belt moves in a certain direction, one can figure out which way the pulley rotates. Although one can draw on experience of similar set-ups that involve belts and pulleys when closing the final gap of causal inference, it is unnecessary to do so. Reason can bridge the gap unaided by additional experience (see Sanford, 1994). By the very nature of causation, effects are never earlier than their causes. This is a majority view. Mackie (1974, ch. 7) discusses the conceptual possibility of “backward causation” and provides further references. There are also serious philosophical discussions of the conceptual possibility of “time travel” in which in there are closed causal loops (see Lewis, 1986). By the very nature of causation, causes are always earlier than their effects. This is a controversial view. Other requirements of causal connection are symmetric in form; they do not distinguish effects from causes. Defining causal priority in terms of temporal priority thus has theoretical appeal. But there is also a theoretical drawback: the equally appealing account of temporal priority by reference to causation will be circular if the explanation of causal priority is to be temporal. Moreover, simultaneous causation appears not only to be possible, but actual. Physics assures us that much of this appearance is illusion. Since nothing transmits motion faster than the speed of light, the motion of one’s fingers, that grip the handle of a teaspoon, does not, strictly speaking, cause the simultaneous motion of the bowl of the spoon. Other cases of apparent simultaneous causation, however, do not involve bridging a spatial gap, as when

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

causation a moving belt turns a pulley with which it is in direct contact. We cannot directly perceive causal relations. This is a majority view that Hume influences greatly with his example of the impact of billiard balls. We can see motions and changes in motion in the balls. We can see that one ball touches the other immediately before the second begins to move. We cannot see that there is a causal relation between the two motions. Nor can we tell, just by observing the sensible qualities of a thing, what are its causal capacities and dispositions. Our sense of touch and our perceptions of the positions and movements of our limbs enable our direct perception of causal relations (see von Wright, 1971, pp. 66–74). This is a minority view. The causal relations between one’s arm movement and the movement of a cue stick one grasps is a more promising candidate for an object of direct perception than the impact of billiard balls that is merely seen. The conceptual fallacy (here so named) may be tempt one here. This is a mistaken inference of the form that since we cannot conceive of A without having the concept of B, therefore the existence of A requires the existence of B. It views ontological dependence as following from conceptual dependence. Granted the minority view that our conception of causation depends on our conceptions of ourselves as agents who make things happen in the physical world, and as patients affected by occurrences in the physical world, it does not follow that the existence of causation requires the occurrence of such interactions. Manipulations are causes. This is a prevailing view. Many languages have many verbs for specific manipulations such as cook, shake, turn, and hold that we understand as causal relations. The view is not strictly a truism since it is inconsistent with seriously held positions such as the following. (a) There really is no physical world; its appearance is an illusion; and from this it follows that there really are no genuine manipulations or physical causal relations. (b) Although there really are physical events, those we commonly but wrongly take as

cause–effect pairs are really coincident joint effects of a common cause, such as God. Current discussions of causation disregard such views and take it for granted that manipulations are causes. Causation depends on manipulation; a correct general account of causation is in terms of manipulation. This is a minority view. Just because one might reach this view by means of the conceptual fallacy discussed above, that does nothing to prove it false. When distinguished from a view about relations between concepts, however, the theory must deal, by appeal to analogy or imagination, with causal instances in which humans do not and sometimes cannot actually participate, such as those that involve clusters of galaxies. A correct general account of causation is in terms of intervention. This is a controversial view, which is currently the center of a robust research program (see Woodward, 2003). This program is careful to distinguish its technical term “intervention” from the ordinary term “manipulation”. Manipulations are performed by agents. While agents also intervene, some natural processes that involve no agents, directly or indirectly, are also called interventions. On the other hand, the notion of an intervention is explicitly causal. Its descriptions use the notion of a causal path. Not all of the descriptions in the literature are equivalent. Here is one description: INT is an intervention between two variables X and Y on the same causal path if and only if INT completely determines the value of X; every causal path between INT (or any cause of INT) and Y goes through X; and if there is a causal path between Z and Y that neither includes nor is included by the path between X and Y, INT does not affect Z. Adding fertilizer does not affect the amounts of water and light, which are relevant variables on causal paths that include the growth of tomatoes. According to this definition of intervention, does the addition of fertilizer then intervene on the causal path between nitrogen level and tomato growth? Weeds complicate the answer to 5

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 growth, a better tomato crop requires pulling some weeds and bringing the addition of fertilizer under the general description of intervention may also require it. Since intervention is a thoroughly causal notion, an interventionist account of a specific causal connection is not reductive in the sense of using only non-causal concepts. This need not render such accounts circular. The use of the notion of intervention to support the presence of a specific connection, such as between nitrogen and growth rate, need not assume its presence to begin with. This accords with the function of experiments. Experiment is a thoroughly causal notion, yet we use experiments to confirm and to disconfirm causal hypotheses. Theorems about interventions have a wide scope in understanding the roles of experiments in various sciences. This is a controversial view. From a precise definition of intervention and some strong assumptions about probabilistic relations between variables, theorists prove theorems about directed causal graphs. (There is no attempt here to summarize these results.) While the theorems themselves are neither trivial nor controversial, there is not a consensus about the manner and scope for their useful application to actual causal processes. Some generalizations that have no exceptions, and some statements of conditional probability, are causal laws. This is a prevailing view. Some universal laws are not causal because they are mathematical or logical laws. Some universal truths are not laws because they are mere “accidental” regularities. If all swimming birds eat fish, this does not imply that there is a law-like connection between birds” swimming and their eating fish. Finding evidence against an accidental regularity, whether quite surprising, or not at all surprising, does not upset our general theories about the world. Providing a general account of the difference between laws and accidental generalization is a major theoretical see law of nature undertaking. There are many competing theories about the character of physical laws, for example, the view that laws are relations between properties or universals. 6

All physical laws are causal laws. This is a majority view. Some philosophers deny that all laws of nature, for example Newton’s first law of motion, are causal laws. Consider a body traveling in a straight line, not changing direction or speeding up or slowing down. Where is the causation? Opinions divide on the adequacy of responses such as “Its motion from B to C is caused by its immediately prior motion from A to B.” Events related as cause and effect, when appropriately described, instantiate a physical law. This is a majority view. These appropriate descriptions typically use concepts different from the ones we ordinarily use in describing the causal transaction. Causation in the everyday world supervenes on causal relations that the fundamental laws of nature directly cover. If such supervenience is universal, there are no causal differences without differences of fundamental properties and spatio-temporal arrangements. A singular causal statement need not entail a law, but it does entail that there is a law that covers, probably as described differently, the events mentioned (see Davidson, 1980, essay 7). Causal attribution and the acceptance of corresponding conditional statements are closely related. This is a prevailing view. Hume connects causation with conditionals in this famous passage: Similar objects are always conjoined with similar. Of this we have experience. Suitable to this experience, therefore, we may define a cause to be an object, followed by another, and where all the objects similar to the first are followed by objects similar to the second. Or in other words, if the first object had not been, the second never had existed. (Hume, 1748, p. 76) What Hume puts “in other words” is scarcely a restatement of what goes before. It nevertheless expresses an important and influential claim, that a cause is necessary for its effect. Kate turned the key, and the engine started. But if the engine would have started at that very moment anyway, without Kate’s key turn, then Kate’s turning the key did not start the engine.

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

causation If-then statements about what would have happened if something else had occurred are called counterfactuals, contrary-to-fact or subjunctive conditionals. A conditional of the form “If a had not happened, then b would not have occurred” says that a is necessary for b: it is impossible for b to occur without a. If it is impossible for a to occur without b, then a is sufficient for b. For example, the downward movement of a lever of the first kind is sufficient for the upward movement of its other end. The necessity of a for b is often separate from the sufficiency for a for b; the thesis that a cause is both necessary and sufficient for its effect is quite strong. Events or conditions we single out as causes often are neither necessary nor sufficient for their effects. Adding Bob’s Super-Grow fertilizer speeded up the growth of the tomato plants, but it was not really necessary. Other brands would have had the same effect. Just by itself, moreover, it also was not sufficient; for other factors, independent of adding the fertilizer, such as light, water and the absence of large amounts of concentrated sulfuric acid, were also necessary for the quick growth of the plants. We can still use the notions of necessity and sufficiency to spell out the causal relevance of adding Bob’s Super-Grow to the plant’s rapid growth. It is presumably an inus condition of the growth; that is, it is an insufficient but non-redundant part of an unnecessary but sufficient condition of rapid growth (Mackie, 1974, p. 62). Inus conditions involve somewhat complicated counterfactual conditionals. The pair of simpler conditionals that express necessity and sufficiency, “If a had not happened, neither would b” and “If a had happened, then so would b” together express counterfactual dependence. Causation can be defined in terms of counterfactual dependence (Lewis, 1986, essays 17 and 21). This is a controversial view. Counterexamples provide one source of controversy. Counterexamples to a claim of the form A=B are in general examples of A that are not B or examples of B that are not A. Lucy threw a stone that broke a bottle. If Lucy had not thrown the stone, however, a stone would have broken the bottle

anyway. Dorothy was standing by, ready to throw a stone toward the bottle if Lucy did not. Standby causes, over-determination, prevention, and other examples serve as counterexamples to simple formulations of counterfactual conditional accounts. This leads to formulations that are less simple, which in turn stimulates the invention of examples of increasing complexity, and so on, back and forth. (See essays in Collins et al., 2004.) Opinions are divided about where this process is leading. Replacing the notion of counterfactual dependence with the notion of influence results in a counterfactual account that runs more smoothly. This is a minority view. One event influences another when each belongs to a range of similar events and there is a range of true counterfactuals of the form if event c (in the first range) had occurred, then event e (in the other range) would have occurred. A mass hanging on a spring influences its length, which varies systematically with the mass. (Within a certain range of values, the relation between mass and length is invariant. Invariance and intervention both figure in causal graph theory.) Adding acid to a base exemplifies causal influence. As more acid is added, more base is neutralized. There is, however, a causal relation in this process that seems not to fit the definition of influence. As more acid is added, it is not until all the base is neutralized that the next drop of acid causes a sudden, large increase in acidity (decrease in pH). It remains to be seen how the influence view accommodates this and similar “tipping point” examples in which a small event produces large effect by upsetting an equilibrium. Questions of causation, inductive support, laws of nature, and counterfactual conditionals are bound closely together. This is a prevailing view. The following distinctions are closely associated, and any one can explain the others: acceptable vs. unacceptable counterfactual conditionals; laws of nature vs. accidental generalizations; a particular observation’s inductively confirming vs. not confirming a hypothesis. Acceptable counterfactual conditionals, but not unacceptable ones, fall under laws (as Chisholm and Goodman have argued). On the other hand, 7

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, support acceptable counterfactuals. Laws, unlike accidental generalizations, are hypotheses that their instances confirm. These interconnections, although mutually explanatory, are arranged in a tight circle and thus evoke a sense of theoretical uneasiness. Philosophers who aspire to develop a theory of causation attempt to break out of the circle by explaining one distinction in the family without appeal to additional distinctions in the same family. Different theories attempt to break out in different places and also differ in their assignments of explanatory priority. For example, one theory holds that a relation between particulars is causal when it falls under a law, while another holds that a generalization is a law when particular causal relations fall under it. No views prevail about the best way to achieve equilibrium in these theoretical matters concerning causation. An adequate theory of causation should be in terms of Probability. This is a controversial view. When an event causes another, the occurrence of the cause often increases the probability of the occurrence of the other. However this is not always so. Attempts of formulate universal generalizations connecting probability with causation run up against examples such as the following (an earlier example with more details): Lucy aims a stone at a bottle. She throws it, and the stone breaks the bottle. Whenever they engage in the sport of throwing stones to break bottles, Dorothy throws a stone if Lucy doesn’t. Although Lucy often misses, Dorothy almost never misses. Lucy didn’t miss this time, however. Her throw broke the bottle. The probability that the bottle would break if she did not throw (and dead-eye Dorothy threw instead) is nevertheless higher than if she did throw. Qualifications of a probabilistic account can accommodate particular examples such as this one, but then, following a pattern of dialectic common in technical philosophy generally, and specifically with the associated counterfactual accounts of causation, new ingenuous counterexamples are not far behind. a is necessary for b if, and only if, b is sufficient for a. This is a prevailing view that 8

follows from the above standard explanations of necessary for and sufficient for. This view does not entail the stronger view that a is a necessary condition of b if, and only if, b is a sufficient condition of a. Causal examples, among others, show that “condition of” is not a symmetric relation. The presence of light, for example, is a causally necessary condition of the growth of tomatoes, which is not in turn a causally sufficient condition for the presence of light. No one attempts to produce light by growing tomatoes. A theory of the direction of conditionship can help account for the direction of causation (Sanford, 1975). A totality of conditions necessary for an occurrence is jointly sufficient for it. This is a controversial view, and not a logical truth, in the technical sense of sufficient spelt out above. There is an ordinary sense of sufficient, however, namely “enough, lacks nothing”. When everything necessary for b obtains, the aggregate is collectively sufficient for b’s occurrence, because jointly the members of the aggregate are enough – nothing necessary for b is missing (see Anscombe, 1981, p. 135). It is not a logical contradiction to maintain that an event did not occur even though nothing necessary for its occurrence was missing. This contention runs against the grain of the following controversial view: Something necessitates every event. This is a controversial view. Although what we call a “cause” often falls far short of being sufficient for its effect, it is common to assume that every effect has some, usually more complicated, sufficient cause. The main issue is not whether some occurrences are totally without causal antecedents, but whether, in the technical sense of sufficient, every event has a sufficient cause. If every event has a sufficient cause, and every cause is an event, then a classic version of Determinism is true. Every event is a link on a branching chain of causal necessitation that runs from the beginning to the end of the universe. The occurrence of any event is causally consistent with exactly one set of events causally connectible with it, whether these events are earlier or later.

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

causation Modern physics, for example in its treatment of atomic decay, discourages belief in determinism. Definitions that resemble Mackie’s definition of an inus condition provide for the possibility of causation without sufficiency: a is a suni condition of b, for example, if there is something x such that the disjunction a or x is a necessary condition of b, and x is not a necessary condition of b (Sanford, 1984, p. 58). Accounts of specific causal connections often refer to causal mechanism. This is a prevailing view. One of the early truly effective drugs was aspirin. As everyone knows, it relieves pain. What scientists did not know, but for years hoped to find, was the mechanism of aspirin’s effect. This goal is different from discovering a more general or more fundamental law. Many scientists try to understand mechanisms rather that find general laws that cover certain phenomena, and this is true not just in medicine, biology, and chemistry, but in many other special sciences. A general account of causation should refer to causal mechanisms rather than to causal laws. This is a minority view. Although operations of mechanisms, of whatever size, seem generally to involve three-dimensional motions, a general theory of causation as mechanism would want a more detailed account of what a mechanism is. Also, some causal connections are so direct that there seems to be no room for a mediating mechanism. Lucy threw a rock that hit a tree before it reached the wall. The tree interrupted the flight path of the rock. Where should one look for the mechanism of this causal interaction? In Plato’s dialogue “The Euthyphro” Socrates and Euthyphro reach a point where they agree that everything all the gods love is pious and that everything pious all the gods love. Socrates goes on to ask whether all the gods love pious things because they are pious, or whether things are pious because all the gods love them. We may call probing questions of this form Euthyphro Questions and proceed to ask them about treatments of causation that aspire to provide reductive accounts. Suppose that some theory is sufficiently refined that both conditionals of

these corresponding forms are true: when C causes E, a suitably situated relation R obtains; and when a suitably situated relations R obtains, C causes E. (This formulation is due to L. Paul.) The Euthyphro Question is whether (a) C causes E because R obtains or (b) R obtains because C causes E. A philosophical reductive definition, account, or analysis of causation should hope to give an answer of form (a). Some popular accounts appear to favor answers of form (b). Consider a counterfactual statements and a corresponding causal statement: If Kate had not turned the key, the engine would not have started. Kate’s turning the key caused the engine to start. It is more natural to say that the conditional is true because turning the key caused the engine to start rather than that turning the key caused the engine to start because the conditional is true. Some conditionals are true because of causal connections; causal connections do not obtain because conditionals are true (see Sanford, 2003, chs. 11–14). Similarly, causal connections explain the effectiveness of manipulation rather than the other way around. Causal connections also explain the effectiveness of interventions, although interventionist theory does not represent itself as reductive. Theories in terms of the transfer of something, or in terms of underlying mechanism, whatever their difficulties, promise to give appropriate answers to the Euthyphro Question. In Book II of the Physics, Aristotle discusses four kinds of aitia or causes. The present article deals only with efficient causes. In the “Second Analogy” of the Critique of Pure Reason (1781), Kant argues that all changes conform to the law of cause and effect. In “Of Induction”, Book III of A System of Logic (1843), J. S. Mill presents experimental methods for establishing causal relevance. In his 1912 lecture, “On the Notion of Cause”, Russell claims that the law of causation “is a relic of a bygone age”; but Russell’s own theoretical constructions in some later writings depend heavily on causal notions. 9

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 Anscombe, G.E.M.: Metaphysics and the Philosophy of Mind, Collected Philosophical Papers, vol. 2 (Minneapolis: University of Minnesota Press, 1981). Beauchamp, T. and Rosenberg, A.: Hume and the Problem of Causation (New York and Oxford: Oxford University Press, 1981). Collins, J., Paul, L., and Hall, N., ed.: Counterfactuals and Causation (Cambridge, MA: MIT Press, 2004) Davidson, D.: Essays on Actions and Events (Oxford: Oxford University Press, 980). Dowe, P.: “Causal Process,” in the Stanford Encyclopedia of Philosophy. Faye, J.: “Backward Causation,” in the Stanford Encyclopedia of Philosophy. Hausman, D.M.: Causal Asymmetries (Cambridge: Cambridge University Press, 1998). Hitchcock, C. “Probabilistic Causation,” in the Stanford Encyclopedia of Philosophy. Hume, D.: Enquiry Concerning Human Understanding (London, 1748): ed. L.A. SelbyBigge (Oxford: Oxford University Press, 1894); 3rd edn. rev. P.H. Nidditch (Oxford: Oxford University Press, 1975). Hume, D.: A Treatise of Human Nature, Book I (London, 1739); ed. L.A. Selby-Bigge (Oxford: Oxford University Press, 1888); 2nd edn. rev. P.H. Nidditch (Oxford: Oxford University Press, 1978). Lewis, D.K.: Philosophical Papers, vol. II (Oxford: Oxford University Press, 1986). Mackie, J.L.: The Cement of the Universe, 2nd edn. (Oxford: Oxford University Press, 1980; originally published 1974). Menzies, P.: “Counterfactual Theories of Causation,” in the Stanford Encyclopedia of Philosophy. Psillos, S.: Causation and Explanation (Chesham, Bucks.: Acumen; Montreal: McGill-Queens University Press, 2002). Salmon, W.C.: Scientific Explanation and the Causal Structure of the World, (Princeton, NJ: Princeton University Press, 1984). Sanford, D.H.: “Causal Relata,” in E. LePore and B. McLaughlin, ed., Actions and Events (Oxford and New York: Blackwell, 1985), 282–93. Sanford, D.H.: “Causation and Intelligibility,” Philosophy 69 (1994), 55–67. 10

Sanford, D.H.: “The Direction of Causation and the Direction of Conditionship,” Journal of Philosophy 73 (1975), 193–207. Sanford, D.H.: “The Direction of Causation and the Direction of Time,” Midwest Studies in Philosophy 9 (1984), 53–75. Sanford, D.H.: If P, then Q: Conditionals and the Foundations of Reasoning, 2nd edn. (London: Routledge, 2003; originally published 1989). Schaffer, J.: “The Metaphysics of Causation,” in the Stanford Encyclopedia of Philosophy. The Stanford Encyclopedia of Philosophy (http://plato.stanford.edu) is an online resource of substantial entries that typically have helpful bibliographies. Entries undergo periodic revision. Strawson, G.: The Secret Connexion: Causation, Realism and David Hume (Oxford: Oxford University Press, 1989). Woodward, J.F.: “Causation and Manipulability,” in the Stanford Encyclopedia of Philosophy. Woodward, J.F.: Making Things Happen: A Theory of Causal Explanation (New York: Oxford University Press, 2003). Wright, G.H. von: Explanation and Understanding (Ithaca, NY: Cornell University Press, 1971). david h. sanford

Fictional Entities The first question to be addressed about fictional entities is: are there any? The usual grounds given for accepting or rejecting the view that there are fictional entities come from linguistic considerations. We make many different sorts of claims about fictional characters in our literary discussions. How can we account for their apparent truth? Does doing so require that we allow that there are fictional characters we can refer to, or can we offer equally good analyses while denying that there are any fictional entities? While some have argued that we can offer a better analysis of fictional discourse if we accept that there are fictional characters, others have held that even if that’s 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 supposed that accepting such entities would involve us in contradictions and so must be avoided at all costs, while others have held that, even if contradiction can be averted, we should refrain from positing fictional entities if at all possible since they would be utterly mysterious, involve us in positing unexplained differences in “kinds of being”, or violate reasonable calls to parsimony. 1. Linguistic Considerations At least four sorts of fictional discourse may be distinguished: (1) Fictionalizing discourse (discourse within works of fiction), e.g., “[Holmes was] the most perfect reasoning and observing machine that the world has seen” in “A Scandal in Bohemia”. (2) Non-existence claims, e.g., “Sherlock Holmes does not exist”. (3) Internal discourse by readers about the content of works of fiction. This may be either intra-fictional (reporting the content of a single work of fiction, e.g., “Holmes solved his first mystery in his college years”,) or cross-fictional (comparing the contents of two works of fiction, e.g., “Anna Karenina is smarter than Emma Bovary”). (4) External discourse by readers and critics about the characters as fictional characters, e.g., “Holmes is a fictional character”, “Hamlet was created by Shakespeare”, “The Holmes character was modeled on an actual medical doctor Doyle knew”, “Holmes appears in dozens of stories”, “Holmes is very famous”. The puzzles for fictional discourse arise because many of the things we want to say about fictional characters seem in conflict with each other: How, for example, could Holmes solve a mystery if he doesn”t exist? How could Hamlet be born to Gertrude if he was created by Shakespeare? Any theory of fiction is obliged to say something about how we can understand these four kinds of claim in ways that resolve their apparent inconsistencies. And any theory of fictional

discourse will have import for whether or not we should accept that there are fictional entities we sometimes refer to, and if so, what sorts of thing they are and what is literally true of them. Given these very different types of fictional discourse, many different approaches have been developed, some of which accept and some of which deny that there are fictional entities. Many of the differences among them may be seen as products of differences in which of the four types of discourse each takes as its primary case and central motivator – though of course all are ultimately obliged to say how we should understand each type of discourse. Perhaps the most popular approach to fictional discourse has been to deny that there are any fictional entities, and to handle the linguistic evidence by adopting a pretense theory. It is plausible that authors in writing works of fiction (and so writing sentences of type (1)) are not making genuine assertions at all, but rather simply pretending to assert things about real people and places (Searle, 1979, p. 65). (Though see Martinich and Stroll, 2007, ch. 2, for challenges to this.) Inspired by this observation about discourse of type (1), full-blown pretense theories of fictional discourse (such as that developed by Kendall Walton) treat all four forms of fictional discourse as involving pretense and so as making no genuine reference to fictional entities. Discourse of type (3), on these views, involves readers “playing along” with the pretense “authorized” by the work of fiction, and so pretending that what is stated in works of fiction is true. Claims like “Holmes solved his first mystery in his college years” are “authorized” moves in the game of pretense licenced by the work, which is why we find them more acceptable than parallel claims like “Holmes drove a white Plymouth”. While that extension of the pretense view seems plausible enough, more difficulties arise for handling external discourse and non-existence claims. Walton takes external claims of type (4) to invoke new “ad hoc” “unofficial” games of pretense other than those authorized by the story, where, e.g., we pretend that “there are two kinds 11

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 characters” (1990, p. 423), or pretend that authors are like gods in being capable of creation, etc. Even apparently straightforward non-existence claims (type 2) are treated as involving pretense: first invoking a pretense that there is such a character to refer to (using the name “Sherlock Holmes”), and then in the same breath betraying that as mere pretense, with the addition of “doesn’t exist” (1990, p. 422). The full-blown pretense approach thus seems to implausibly take as pretenseful precisely the (type 2 and type 4) talk about fiction that is designed to step outside of the pretense and speak from the real-world perspective. It also offers contorted and ad hoc readings of what seem to be straightforward literal claims (cf. Thomasson, 2003). So while pretense theories do well at addressing internal and fictionalizing discourse, they are much less plausible adopted as across the board approaches – but if we can’t adopt them across the board, they can’t be used to avoid positing fictional entities. Various other approaches to fictional discourse have been proposed which don’t rely on taking pretense to be ubiquitous in fictional discourse, yet still avoid accepting that there are fictional entities. The best developed of these is Mark Sainsbury’s (2005) negative free logic approach, which takes as its central motivation the truth of claims of type (2): non-existence claims involving fictional names. On the negative free logic view, fictional names are nonreferring terms, and all simple sentences using non-referring terms are false. Thus “Holmes exists” is false (as “Holmes” doesn’t refer), and so its negation “Holmes doesn’t exist” is true (Sainsbury, 2005, p. 195), leaving us with a far simpler and more plausible account of the truth of nonexistence claims than pretense views provide. Internal discourse by readers can still be held to be true even though it involves non-referring names, since these claims are plausibly held to be implicitly prefixed with a fiction operator, where “According to the fiction, Holmes solved his first mystery in his college years” may be true even if the simple claim “Holmes solved his first mystery 12

in his college years” would be false. Crossfictional statements can be handled similarly by taking them to fall in the context of an “agglomerative” story operator that appeals to the total content of the relevant stories, taken together, e.g., “According to (Anna Karenina and Madame Bovary [taken agglomeratively] ), Anna Karenina was more intelligent than Emma Bovary” (Sainsbury, forthcoming). But like the pretense view, the negative free logic view has more difficulties accounting for the apparent truth of external claims of type (4), since their truth cannot be accounted for by taking them as implicitly reporting what is true according to the fiction. Various ad hoc ways of interpreting these claims have been tried, e.g., “Holmes is a fictional character”, may be read as reporting that, according to some fiction, Holmes exists (Sainsbury, forthcoming). But given the variety of external claims that must be rewritten in different ways, these remain the biggest thorn in the side of negative free logic theories. On the other side of the debate are those who argue that we can only or best handle fictional discourse by allowing that there are fictional entities and that at least sometimes our discourse refers to them. But even among those who accept that there are fictional entities there are widespread disagreements about what we should consider them to be and what is literally true of them. Some realist views about fiction are inspired by the apparent truth of internal claims of type (3), and so take fictional entities to be beings that (in some sense) have the properties the characters of the story are said to have, so that claims like “Holmes solved his first mystery in his college years” is true because there is a fictional entity, Holmes, who in some sense has this property. These views have taken many forms – with some taking the fictional entities to be possible people, others taking them to be Meinongian non-existent objects, and others still taking them to be pure abstract entities such as kinds. One natural approach inspired by the desire to accommodate the truth of type (3) 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 stories. Kripke expressed this idea when he wrote “Holmes does not exist, but in other states of affairs, he would have existed” (1963/1971, p. 65). But Kripke himself later (1972, p. 158) rejected this answer, and his rejection of it has generally been taken on board. His grounds for rejecting it come from considerations about reference: the name “Sherlock Holmes” is not a description (which could be fulfilled by various possible individuals); instead, if it refers at all, it picks out the individual to whom the speaker’s use of the name bears a historical connection, and it refers to that very individual across all possible worlds. So if there happened to be someone in the actual world who coincidentally was just as Holmes is said to be in the novels, that would not show that he was Holmes. Similarly, if there are individuals in other possible worlds who fulfill the descriptions in the books, that does not show that any of them is Holmes. Moreover, since there will be a great many different possible individuals who fulfill the descriptions, it seems there would be no non-arbitrary way of saying which of these is Holmes (Kripke, 1972, pp. 157–8). Given the problems with possibilist views, the most popular realist treatments of fictional entities have been not possibilist but Meinongian and abstractist views. Meinong himself was not interested in fiction per se, but rather sought to develop a general theory of the objects of speech and cognition (1904/1960). If there is knowledge, Meinong thought, there must be something known, if there is a judgment, there must be something judged, and so on. So, for example, if we know that the round square is round, there must be something (the round square) of which we know that it is round. Some of these objects of knowledge, however (like the round square) do not exist. Meinongian views thus take seriously the truth of internal (type (3)) sentences like “Holmes solved his first mystery in his college years”, and take fictional entities to be the non-existent objects truly described in such sentences – so on these views a fictional entity is the object that (in some sense) has all of the properties ascribed to

the character in the relevant work (or works) of fiction. The simple version of this approach encounters difficulties of the kind that led to Russell’s (1905/1990) criticisms of Meinong. For the stories ascribe to Holmes not only properties like being a person and solving mysteries, but also properties like existing, in conflict with the apparent truth that Holmes doesn”t exist. Indeed Meinongian theories take non-existence claims of type (2) to be straightforwardly true since, although there are the relevant fictional entities, they do not exist. So the Meinongian is in danger of contradiction by taking Holmes and the like both to exist (since Meinongian objects are supposed to have all of the properties ascribed to them) and not to exist (since they are nonexistent objects). The central achievement of neoMeinongians such as Terence Parsons (1980) and Edward Zalta (1983) has been to show how these contradictions may be avoided. Parsons avoids them by distinguishing two kinds of properties: nuclear properties (like being a man, being a detective, etc.) and extra-nuclear properties (like existing, being possible, etc.). He then holds that only the nuclear properties ascribed to the character in the story are actually possessed by the corresponding objects, so we do not have to conclude that Holmes exists. Nonetheless, we do need some way to mark the fact that there may be objects (arguably, like Macbeth’s dagger) that don’t exist according to the stories, as well as objects that (like Macbeth) are said to exist. To mark this, Parsons suggests that there are “watered down” nuclear properties corresponding to each extra-nuclear property, so that Holmes does not exist (extra-nuclear) but does have watered-down (nuclear) existence. Zalta (1983), following Ernst Mally, avoids contradiction by a different route: distinguishing two modes of predication: encoding and exemplifying. Fictional entities encode all of those properties they are said to have in the stories, but that does not mean that they exemplify them. So Holmes encodes existence but exemplifies non-existence, and contradiction is avoided. 13

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 fictional entities to be existing abstract objects of some sort rather than to be Meinongian non-existent objects. Nicholas Wolterstorff develops one such view, according to which fictional characters are “not persons of a certain kind, but personkinds” which do exist (1980, p. 144). On this view, authors do not refer to anyone when they write fictional stories; instead, they delineate a certain kind of person by describing certain sets of characteristics. The fictional character Holmes is not a person, but a certain kind of person, or “person-kind’, that has essentially within it those properties the work attributes to the character, e.g., being a man, being clever, being a detective. . . . As abstracta, of course kinds can’t literally have such properties as being clever or solving mysteries – but they can be defined by the properties essential within them. So on this view, type (3) claims such as “Holmes solved his first mystery in his college years” are true just in case the properties expressed by the predicate (solving one’s first mystery during one’s college years) are essential within the personkind Holmes (1980, p. 159). Many (but not all – see below) of the properties attributed to characters in external discourse, e.g., being famous, appearing in stories, may be properties these abstract person-kinds genuinely have rather than properties essential within the kind. But neither of these strategies helps Wolterstorff cope with (type 2) non-existence claims, for existence is ascribed to Holmes in the stories, and so is essential to that person-kind, and the abstract entity that is that personkind also exists. Wolterstorff suggests two alternative ways of understanding nonexistence claims: either as saying that the relevant person-kind has never been exemplified, or (acknowledging Kripke’s point) that the author was not referring to anyone when he used the name in writing the story (1980, p. 161). Despite their differences, possibilist, neoMeinongian, and abstractist views are alike in taking most seriously internal (type 3) claims about fictional characters, and as a result they face similar difficulties accounting 14

for the truth of at least some type (4) external claims. Whether fictional entities are taken to be unactualized possibilia, non-existent objects, or abstract kinds, it seems that in any of these cases the work of authors writing stories is completely irrelevant to whether or not there are these fictional entities: the relevant possibilia, non-existent objects, and abstract kinds were “around” just as much before as after acts of authoring, and so we can’t take seriously the idea that authors create fictional characters on any of these views. The best these views can do to account for the apparent truth of claims such as “Hamlet was created by Shakespeare” is to say that it is at least true that Shakespeare described or selected Hamlet from among all the available possibilia, non-existent objects, or abstract kinds and, by writing about that object, made it fictional. (Below I will return to discuss some metaphysical difficulties these views also face.) All of the views canvassed thus far – whether or not they accept that there are fictional entities – face difficulties accounting for the apparent truth of certain external (type 4) sentences. This has inspired several recent theorists to begin by taking this sort of discourse as the focal case – a view that requires accepting that there are fictional characters and that these are created by authors in the process of writing works of fiction. Since they take fictional characters to be products of the creative activities of authors, call these “artifactual” views of fiction. The phenomenologist Roman Ingarden suggested something like an artifactual view of fiction in his (1931) The Literary Work of Art, where he treats fictional characters (and the literary works in which they appear) as purely intentional objects – objects owing their existence and essence to consciousness. Saul Kripke (apparently independently) suggests that fictional entities are human creations in his unpublished 1973 John Locke lectures. He argues that fictional characters exist in the ordinary concrete world (not another possible world), but they do not exist “automatically” as pure 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 contingent and exist only given concrete activities of writing or telling stories. John Searle (1979, pp. 71–2) similarly claims that authors, in writing stories and pretending to refer to people, instead create fictional characters to which others can then refer. More recently, artifactual views of fiction have been defended by Schiffer (1996) and Salmon (1998), and developed at length by Thomasson (1999, 2003). (van Inwagen (1977, 1983, 2003) develops a similar view according to which fictional characters are theoretic entities of literary criticism, but he is noncommittal about whether or not they are created.) Artifactualist theories take external (type 4) claims about fictional characters – e.g., that Holmes is a fictional character created by Arthur Conan Doyle, who modeled Holmes on a medical doctor – to be literally true. On Thomasson’s view, fictional characters are abstract artifacts created by authors’ activities in writing or telling stories, and dependent for their ongoing existence on those stories (and copies or memories of them). The status of fictional characters as created, dependent, abstracta, she emphasizes, is like that of many social and cultural entities such as laws of state, symphonies, and works of literature themselves: none of them may be identified with any concrete entity, none has a definite spatial location, but all come into existence at a particular time given certain types of human activity. Most artifactualists, like Searle, take fictional characters to be created by authors pretending to refer to real people and places, and so take fictionalizing (type 1) discourse to involve mere pretended assertions. Artifactualists generally do not take (type 3) internal discourse to state literal truths about properties these fictional entities have; instead, they (like Sainsbury fictional entities) typically read these as shorthand for claims about what is true according to the fiction or (following Walton) about what is accepted in games of pretense authorized by the story. The greatest difficulty for artifactual views arises in handling (type 2) non-existence claims. Various strategies may be used

here: denials that Sherlock Holmes exists may be read as denials that there is any such person (Thomasson, 1999, p. 112), or any object answering the descriptions in the stories (van Inwagen, 2003, p. 146). Alternatively, these non-existence claims may be read as noting that past users of the name mistakenly supposed that the nameuse chain led back to a baptism rather than a work of fiction (van Inwagen, 2003, pp. 146–7; cf. Thomasson, 2003). If some such solution to the problem of nonexistence claims can be shown to be plausible and non ad hoc, artifactual theories may offer the best overall way to handle fictional discourse – a way which does require positing fictional entities. 2. Metaphysical Considerations None the less, many think that we have metaphysical grounds to resist positing fictional entities even if we can offer a somewhat better account of language by accepting that there are such entities and that we sometimes refer to them. These arguments have run in parallel to the developing theories of what fictional entities are. As we have seen, Russell originally claimed that Meinongian objects were “apt to infringe the law of contradiction” (1905/1990, 205); an objection that kept fictional entities largely undefended for over seventy years. While neo-Meinongians showed how to avoid contradiction, their views were none the less widely rejected for drawing a distinction between what objects exist and what objects there are (or over which we may quantify) – a distinction many philosophers claim to find incomprehensible (van Inwagen, 2003, pp. 138– 42). Abstractist and possibilist solutions, of course, are more acceptable to those already inclined to accept abstract objects, or possible worlds and the objects in them. But even if one accepts that there are platonistic abstracta or mere possibilia (see the extended essay on realism and antirealism about abstract entities), other problems arise in supposing that fictional characters are among them. As mentioned above, fictional characters are generally thought 15

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 actual world, but neither of these is true of either platonistically conceived abstracta (which are eternal and necessary) or of mere possibilia (which are not created by authors and are merely possible). Moreover, some stories are (intentionally or unintentionally) inconsistent, and so some of their characters can’t be treated as possible objects having all the properties ascribed in the story. Another metaphysical problem that arises for both possibilist and abstractist views comes from the fact that they (like the Meinongian views before them) take the descriptions in works of fiction to determine which object we are talking about: the fictional entity is the possible person or abstract entity that has, or has essential within it, all of the properties ascribed to the character in the story. But this leads to problems with the identity conditions for fictional characters (see Thomasson, 1999, ch. 5). For these views entail that no fictional character could have had any properties other than those they are ascribed. If the author made even a minor change in the work, so that the character is ascribed so much as one different property (however trivial), she would have written about a different possible person, or delineated a different person-kind. As a result, these views must hold that sequels, parodies, and even revised editions must always include entirely different characters from the original texts – in violation of our standard assumption that an author may change what she says about a given character, and that sequels may describe the further adventures of one and the same character. (Meinongian theories face similar difficulties with handling identity conditions.) Artifactualist views avoid metaphysical dificulties like these by taking fictional characters (like works of literature themselves) to be created by activities of authors and individuated primarily by their historical origin. The artifactualist typically treats historical continuity – not properties ascribed – as the primary factor for the identity of a fictional character. This leaves open the idea that an author might have described a 16

character somewhat differently than she did, and allows that a later author may ascribe new properties to a preexisting fictional character, provided she is familiar with that character and intends to refer back to it and ascribe it new properties (Thomasson, 1999, pp. 67–9). None the less, artifactualist views face other metaphysical objections. Although the artifactualist treats fictional characters as created entities, they are also clearly abstract in some sense: though not eternal and necessary like the Platonist’s abstracta, they still lack a spatio-temporal location (and are not material) (Thomasson, 1999; see also concrete/abstract). But the very idea that there may be created abstracta strikes some as hard to swallow. As Inwagen puts it “Can there really be abstract things that are made? Some might find it implausible to suppose that even God could literally create an abstract object” (2003, pp. 153– 4). Thomasson (1999) addresses these worries by noting that those who accept the existence of such ordinary social and cultural objects as laws, marriages, symphonies, and works of literature themselves are apparently already committed to the existence of created abstracta, so that no special problems arise in accepting created abstracta to account for fictional characters. Of course this “companions in guilt” argument leaves us with two choices: allow that there are abstract artifacts and accept the existence of fictional characters, literary works, laws, etc., or deny the existence of all of these and find some way of paraphrasing talk about the latter entities as well as about fictional characters. But those who would take the latter route should note that even accounting for fictional discourse itself is much more difficult if we cannot make reference to the stories in which they appear. A final and persistent metaphysical argument against fictional entities is that, since it would be much more parsimonious to deny the existence of fictional characters, we should do so if at all possible. The parsimony argument can be addressed in several ways. First, it is worth noting that even Occam’s razor only tells us that “it is vain 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 account of fictional discourse by accepting fictional entities, the antirealist about fictional entities is not really doing the same thing as the realist, with fewer entities. Second, as Thomasson (1999) notes, it is not obviously more parsimonious to do without fictional characters if we must posit abstract artifacts in some other arena, e.g. to make sense of our talk about novels, symphonies, laws of state, and the like. The most potentially powerful, though also the most controversial, response to parsimony-based arguments comes from a certain minimalist or “pleonastic” approach to their ontology proposed by Stephen Schiffer (1996). On Schiffer’s view, pretenseful uses of a fictional name in works of literature, e.g. “[Holmes was] the most perfect reasoning and observing machine that the world has seen”, automatically license us to introduce the singular term “the fictional character Sherlock Holmes” which may then be used in a hypostatizing way in literary discussions. Given those prior pretenseful uses, that singular term is guaranteed to refer to a fictional character. But if all that it takes for fictional names to be guaranteed to refer to characters is that these names be used pretensefully in works of literature, it is not at all clear that someone who accepts that there are pretenseful uses of these names in works of literature but denies that there are fictional characters is genuinely offering a more parsimonious view. Instead, as Thomasson argues (2003), such a person would be only twisting the ordinary rules of use for terms like “fictional character” by artificially inflating the conditions it takes for there to be such characters – not offering a genuinely more parsimonious ontology. 3. Broader Relevance The question of whether or not we should accept that there are fictional entities – and if so, what sort of thing they are – has been a recurrent topic throughout the history of analytic philosophy because of its broader relevance for a range of other philosophical issues. First, as we have seen in section 1, it has relevance for our theory of language.

If we deny that there are fictional entities (and so deny that we ever refer to them), we must explain how we can have true statements involving non-referring terms. If we accept that there are fictional entities, we must explain how we can refer to nonexistent objects (if we take a Meinongian view), merely possible objects, or abstracta (whether Platonist or artifactual) – a task that is especially difficult for causal theories of reference, since none of these entities are obviously a part of the actual causal order. Issues regarding fictional entities also have broader relevance for work in metaphysics. If artifactualists like Thomasson are correct, then whether or not one accepts that there are fictional characters is closely connected to the issue of whether one accepts other mind-dependent social and cultural objects such as laws and nations, stories and symphonies. Moreover, our stance regarding fictional entities has central relevance for issues of ontological commitment and quantification: If the Meinongian is right, we can quantify over entities that don’t exist, and existence must be distinguished from quantification. If the minimalist is right, then the measure of ontological commitment is not whether or not we quantify over the relevant entities – for if we accept that there are authors who use fictional names pretensefully in writing works of fiction, we are already tacitly committed to fictional characters regardless of whether they explicitly quantify over them. See also the a–z entry on fictional truth, objects, and characters. bibl i ography Ingarden, R.: The Literary Work of Art, trans. George Grabowicz (Evanston, IL: Northwestern University Press, 1931). Kripke, S.: Naming and Necessity (Oxford: Blackwell, 1972). Kripke, S.: “Semantical Considerations on Modal Logic,” in Reference and Modality, ed. Leonard Linsky (Oxford: Oxford University Press, 1971; originally published 1963). Martinich, A.P. and Stroll, A.: Much Ado about Nonexistence: Fiction and Reference 17

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

f r e e w ill (Lanham, MD: Rowman and Littlefield, 2007). Meinong, A.: “On the Theory of Objects,” in Realism and the Background of Phenomenology ed. Roderick Chisholm (Atascadero, CA: Ridgeview, 1960; originally published 1904). Parsons, T.: Non-existent Objects (New Haven, CT: Yale University Press, 1980). Russell, B.: “On Denoting,” in The Philosophy of Language, 2nd edn. Ed. A.P. Martinich (New York: Oxford University Press, 1990; originally published 1905). Sainsbury, M.: Reference Without Referents (Oxford: Clarendon Press, 2005). Sainsbury, M.: “Serious Uses of Fictional Names” (forthcoming). Salmon, N.: “Nonexistence,” Noûs 32:3 (1998), 277–319. Schiffer, S.: “Language-Created LanguageIndependent Entities,” Philosophical Topics 24:1 (1996), 149–67. Searle, J.: Expression and Meaning: Studies in the Theory of Speech Acts (Cambridge: Cambridge University Press, 1979). Thomasson, A. L.: Fiction and Metaphysics (Cambridge: Cambridge University Press, 1999). Thomasson, A. L.: “Speaking of Fictional Characters,” Dialectica 57:2 (2003), 207– 26. van Inwagen, P.: “Creatures of Fiction,” American Philosophical Quarterly 14:4 (1977), 299–308. van Inwagen, P.: “Existence, Ontological Commitment, and Fictional Entities,” in The Oxford Handbook of Metaphysics, ed. Michael Loux and Dean Zimmerman (Oxford: Oxford University Press, 2003). van Inwagen, P.: “Fiction and Metaphysics,” Philosophy and Literature 7 (1983), 67– 77. Walton, K.: Mimesis as Make-Believe (Cambridge, MA: Harvard University Press, 1990). Wolterstorff, N.: Works and Worlds of Art (Oxford: Clarendon Press, 1980). Zalta, E.: Abstract Objects (Dordrecht: Reidel, 1983). amie l. thomasson

Free Will The metaphysical “problem of free will” has arisen in history whenever humans have reached a certain stage of self-consciousness about how profoundly the world may influence their behavior in ways unknown to them and beyond their control (Kane, 1996, pp. 95–6). Various authors describe this stage of self-consciousness as the recognition of a conflict between two perspectives we may have on ourselves and our place in the universe (e.g., Nagel, 1986). From a personal or practical standpoint, we believe we have free will when we view ourselves as agents capable of influencing the world in various ways through our choices or decisions. When faced with choices or decisions, open alternatives seem to lie before us – a “garden of forking paths” into the future, to use a popular image. We reason and deliberate among these alternatives and choose. We feel (1) it is “up to us” what we choose, and hence how we act; and this means we could have chosen to act otherwise. As Aristotle said, “when acting is ‘up to us’, so is not acting”. This “up-to-us-ness” also suggests that (2) the ultimate sources of our choices, and hence of our actions, lie in us and not outside us in factors beyond our control. Because of these features, free will is often associated with other valued notions such as moral responsibility, autonomy, genuine creativity, self-control, personal worth or dignity, and genuine desert for deeds or accomplishments (Kane, 1996, ch. 6). These two features of free will also lie behind various reactive attitudes that we naturally take toward the behavior of ourselves and others from a personal standpoint (P. Strawson, 1963). Gratitude, resentment, admiration, indignation, and other such reactive attitudes seem to depend upon the assumption that the acts for which we feel grateful, resentful, or admiring had their origins in the persons to whom these attitudes are directed. We feel that it was up to them whether they performed those 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 But something happens to this familiar picture of ourselves and other persons when we view ourselves from various impersonal, objective or theoretical perspectives (Nagel, 1986, p. 110). From such perspectives it may appear that our choices or decisions are not really “up to us”, but are determined or necessitated by factors unknown to us and beyond our control. The advent of doctrines of determinism in the history of philosophy is an indication that this worry has arisen. Doctrines of determinism have taken many forms. People have wondered whether their actions might be determined by Fate or by God, by the laws of physics or the laws of logic, by heredity or environment, by unconscious motives or hidden controllers, psychological or social conditioning, and so on. There is a core idea running through all these historical doctrines of determinism that shows why they are a threat to free will. All doctrines of determinism – whether logical, theological, physical, biological, psychological or social – imply that at any time, given the past and the laws of nature (see law of nature) and of logic, there is only one possible future. Whatever happens is therefore inevitable or necessary (it cannot but occur), given the past and the laws. Doctrines of determinism thus seem to threaten both features of free will mentioned earlier. If determinism is true, it seems that it would not be (1) “up to” agents what they chose from an array of alternative possibilities, since only one alternative future would be possible, given the past and laws. It also seems that, if determinism were true, the (2) sources or origins of choices and actions would not be in the agents themselves, but in something outside their control that determined their choices and actions (such as the decrees of fate, the foreordaining acts of God, heredity and upbringing or social conditioning). Those who believe, for these or other reasons, that free will and determinism are not compatible – and hence that free will could

not exist in a completely determined world – are incompatibilists about free will. The opposite view is taken by compatibilists, who hold that, despite appearances to the contrary, determinism poses no threat to free will, or at least to any free will “worth wanting” (Dennett, 1984). Compatibilists characteristically argue that all the freedoms we recognize and desire in ordinary life – e.g., freedoms from coercion or compulsion, from physical restraint, from addictions and political oppression – are really compatible with determinism. Even if the world should be deterministic, they argue, there would still be an important difference between persons who are free from constraints on their freedom of action (such as coercion, compulsion, addiction and oppression) and persons who are not free from such constraints; and we would prefer to be free from such constraints rather than not, even in a determined world. Compatibilism was espoused by some ancient philosophers, such as the Stoics, and also by Aristotle, according to some scholars. But it became especially influential in the modern era, defended in one form or another by philosophers such as Hobbes, Locke, Hume, and Mill, who saw compatibilism as a way of reconciling ordinary experience of being free with modern science. Compatibilism remains popular among philosophers and scientists today for similar reasons. By contrast, incompatibilists of the modern era, such as James, regard compatibilism as a “quagmire of evasion” or a “wretched subterfuge”, as Kant called the compatibilism of Hobbes and Hume. Compatibilists also characteristically warn against confusing determinism with fatalism, the view that whatever is going to happen, is going to happen, no matter what we do. Compatibilists, such as Mill, argue that what we decide and what we do will make a difference to how things turn out, even if determinism should be true. And since we do not know the future, we have to deliberate and try to decide upon the best course of action, whether determinism is 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 a n d i n c o mp atib ilism The “Compatibility Question” (“Is free will compatible or incompatible with determinism?”) has thus been central to modern debates about free will. And the popularity of compatibilism in the modern era has placed the burden of proof on incompatibilists to show why free will must be incompatible with determinism. Incompatibilists have tried to meet this challenge in various ways. The most widely discussed of their arguments for the incompatibility of free will and determinism in modern philosophy is called the “Consequence Argument”. It is stated informally by one of its defenders (van Inwagen, 1983, p. 16) as follows: If determinism is true, then our acts are the consequences of the laws of nature and events in the remote past. But it is not up to us what went on before we were born; and neither is it up to us what the laws of nature are. Therefore the consequences of these things (including our own acts) are not up to us. To say it is not “up to us” what “went on before we were born”, or “what the laws of nature are”, is to say that there is nothing we can now do to change the past or alter the laws of nature (it is beyond our control). If such things are beyond our control, and our present actions are necessary consequences of the past and laws of nature (as determinism entails), then altering the fact that our present actions occur would appear to be beyond our control as well. In short, if determinism is true, no one can do otherwise than he or she actually does; and if free will requires the power to do otherwise, or alternative possibilities, then no one would have free will. This argument has generated a large critical literature. Each premise and step has been questioned. (For useful summaries of the issues, see van Inwagen, 1983; Fischer, 1994; Ekstrom, 2000; Kapitan, in Kane, 2002). Compatibilists have usually challenged the argument in either of two ways. The first challenge comes from “classical compatibilists” (such as Hobbes, Hume, and 20

Mill) who defend hypothetical or conditional analyses of “can” and “can do otherwise”. According to such analyses, to say “we can do otherwise” means that “we would do otherwise, if we chose or wanted to do otherwise”. If such hypothetical analyses are correct, the conclusion of the Consequence Argument (“if determinism is true, no one can do otherwise” would fail. For, being able to do otherwise would merely entail that one would have done otherwise, if (contrary to fact) one had chosen or wanted to do otherwise, or if the past had been different in some way; and such a claim would be consistent with saying that one’s present action was determined by the actual past and laws. Much debate about the compatibility of free will and determinism has thus concerned such hypothetical analyses of “can” and “could have done otherwise” favored by classical compatibilists. Incompatibilists reject hypothetical analyses; and powerful objections have been made against them by J.L. Austin, R.M. Chisholm and K. Lehrer, among others. Yet hypothetical analyses continue to be defended by many compatibilists, e.g., Davidson and Lewis. (For an overview of the debates, see Berofsky, in Kane, 2002.) al t ernative possibi l i t i es and moral responsibi l i t y A more radical compatibilist challenge to the Consequence Argument consists in denying altogether the assumption that “free will requires the power to do otherwise, or alternative possibilities”. Call this assumption AP (for “alternative possibilities”). If AP is false – if free will does not in fact require the power to do otherwise – then the Consequence Argument, it is argued, would also fail to show that free will and determinism are not compatible. But on what grounds could one deny that free will requires the power to do otherwise? The answer lies in the connection between free will and moral responsibility. Freedom of will is not just any kind of freedom of action or freedom to do what you want. Freedom of will has a special relationship to responsibility or accountability for one’s actions. Indeed,

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

f ree will many philosophers actually define free will as that kind of freedom (whatever it may be) that is necessary to confer true moral responsibility (and hence genuine praiseworthiness and blameworthiness) on agents. The connection between free will and moral responsibility is important because a number of “new compatibilists”, including Frankfurt (1969), Dennett (1983), Fischer (1994), and Wallace (1994), have denied that moral responsibility requires the power to do otherwise, or alternative possibilities. They reject a principle that Frankfurt has called the Principle of Alternative Possibilities (PAP): Persons are morally responsible for what they have done, only if they could have done otherwise. If free will is the kind of freedom required for moral responsibility and PAP is false, then AP would be false as well: Free will (in the sense required for moral responsibility) would also not require the power to do otherwise or alternative possibilities. Two kinds of examples have been offered by new compatibilists to show the falsity of PAP. The most widely discussed of these two kinds of examples are called “Frankfurtstyle examples”, after Harry Frankfurt, who introduced the first such example in 1969. Frankfurt posited a controller named Black, whom we might suppose is a neurosurgeon with direct control over the brain of an agent Jones. Jones faces a choice between doing A (say, voting for a Democrat) and B (voting for a Republican). Black wants Jones to do A, but he does not want to intervene unless he has to. So if Black sees that Jones is going to choose A on his own, Black will not intervene. Only if Black sees that Jones is going to choose B will he intervene in Jones’s brain, making Jones choose A. Frankfurt asks us to consider the case where Jones chooses A on his own and Black does not intervene. In such a case, Frankfurt argues, Jones could well have been morally responsible for his choice, since Jones acted on his own and Black did not intervene. Yet Jones could not have done otherwise, since, if he had given any indication of choosing otherwise, Black would not have let him. So it seems that Jones can be morally responsible for his choice even

though he could not have done otherwise: and PAP is false. As with the Consequence Argument, an enormous literature has developed around these Frankfurt-style examples. (Overviews of the literature can be found in Fischer, 1994 and Widerker and McKenna, 2003.) Of many objections that have been made against the use of such examples to undermine PAP, the most discussed objection is one originally made by Kane (1985) and developed independently by Widerker (1995) and Ginet, among others. The objection insists that if some morally responsible (free will) choices are undetermined up to the moment they occur, as incompatibilists require, then a Frankfurt controller like Black could not know in advance which choice the agent Jones was going to make until the choice was actually made. In that case, if the controller did not intervene, the agent would have alternative possibilities; and if the controller did intervene, he would have to do so in advance to make the agent choose as he wished. But in that case, the controller would be responsible for the choice, not the agent. To meet this objection, a host of new, more sophisticated, Frankfurtstyle examples have been developed in the past decade by David Hunt, Eleonore Stump, Alfred Mele and David Robb, Derk Pereboom, and others. The jury is still out on the efficacy of these new Frankfurt-style examples. (For a discussion of this literature, see Widerker and McKenna, 2003.) hi erarc hical theories and other new compat i bilist vi ews New compatibilists, such as Frankfurt, have also put forward novel (compatibilist) accounts of free will, according to which free will does not require the power to do otherwise. In a seminal essay, Frankfurt (1971) argues that persons, unlike other animals, “have the capacity for reflective self-evaluation that is manifested in the formation of second-order desires” (p. 7) – desires to have or not to have various first-order desires. Free will and responsibility require that we assess our first-order desires or motives and form “second-order volitions” about which 21

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 action. Our “wills” – the first-order desires that move us to action – are free, according to Frankfurt, when they are in conformity with our second-order volitions, so that we have the will (first-order desires) we want (second-order desires) to have and in that sense we “identify” with our will. Such a theory of free will is called “hierarchical” for obvious reasons. Classical compatibilism is deficient, according to hierarchical theorists such as Frankfurt, because it gives us only a theory of freedom of action (being able to do what we will) without a theory of freedom of will in terms of the conformity of first-order motives to higher-order motives (being able, so to speak, to will what we will). Hierarchical theories remain compatibilist, however, since they define free will in terms of a conformity (or “mesh”) between desires at different levels without requiring that desires at any level be undetermined. Other new compatibilist accounts of free will, such as those of Watson (1975) and Wolf (1990), are also “mesh theories”, but they reject Frankfurt’s hierarchical view. For Watson, the relevant mesh required for free agency is not between higher and lower-order desires, but between an agent’s “valuational system” (beliefs about what is good or ought to be done), which has its source in the agent’s reason, and the “motivational system” (desires and other motives), which has its source in appetite. Watson thus revives the ancient Platonic opposition between reason and desire – arguing that freedom consists in a certain conformity of desire to reason. Wolf’s “reason view” takes this approach in another direction that also has ancient roots. She argues that freedom consists in being able to do the right thing for the right reasons, which requires in turn the ability to appreciate “the True and the Good”. Wolf’s theory thus has a stronger normative component than other compatibilist theories. Other new compatibilist approaches to freedom with a normative component include those of Michael Slote, Paul Benson, and Philip Pettit and Michael Smith. Still other new compatibilist theories, e.g., those 22

of P. Strawson (1962) and Wallace (1994), emphasize the role of “reactive attitudes”, such as gratitude, resentment and indignation, in our understanding of freedom and responsibility. (For critical surveys of many of these “new compatibilist” theories, see the essays by Haji and Russell, in Kane, 2002). Another significant new compatibilist approach to free will is semi-compatibilism, whose chief defender is Fischer (1994; see also Fischer and Ravizza, 1998). Fischer is convinced by Frankfurt-style examples and other considerations that moral responsibility does not require alternative possibilities. But he also argues that freedom does require forking paths into the future, and hence alternative possibilities; and he is convinced by the Consequence Argument that determinism rules out alternative possibilities. The result of these competing considerations is “semi-compatibilism”: moral responsibility is compatible with determinism, but freedom (in the sense that requires alternative possibilities) is not compatible with determinism. hard determini sm and hard inc ompatibi l i sm Incompatibilists have also put forth new accounts of free will in modern philosophy and new defenses of its incompatibility with determinism. Incompatibilism, however, may take two opposing forms: Incompatibilists who affirm the existence of free will and hence deny the truth of determinism are called libertarians in modern free will debates. By contrast, incompatibilists who affirm determinism, and thus deny the existence of free will, have traditionally been called hard determinists. Hard determinism will be considered here first and then libertarianism. Classical hard determinism (as held by d’Holbach, for example) consists of three theses: (i) free will (in the strong sense required for ultimate responsibility and desert) is not compatible with determinism; (ii) there is no free will in this strong sense because (iii) all events are determined by natural causes (i.e., determinism is true). 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 Honderich (1988), Pereboom (2001), and G. Strawson (1986) tend to accept theses (i) and (ii), but remain non-committal about (iii) – whether universal determinism is true. These modern skeptics about free will are aware that, with the advent of quantum physics in the twentieth century, it is far less clear that the physical world is the deterministic system imagined by classical Newtonian physics. In the eighteenth century, with Newtonian physics in mind, LaPlace famously imagined that a superintelligence, knowing all the forces of nature and the exact positions and momenta of particles at any one time, could predict with certainty every future event in the minutest detail. Today it is customary to distinguish predictability or this sort from determinism. For it is known that even in some classical physical systems (such as those that exhibit chaotic behavior), future behavior may not be predictable, even though such systems continue to be deterministic. Modern quantum physics complicates this classical picture even further (at least on standard interpretations of it), by insisting that no superintelligence could know the exact positions and momenta of all particles at any moment because the particles do not have both exact positions and momenta at the same time (the Heisenberg uncertainty principle); and hence their future behavior is not predictable or determined. Yet issues of determinism and indeterminism in physics remain unsettled because there is continuing controversy about the interpretation of quantum physics and about its metaphysical implications. As a consequence, modern skeptics about free will usually remain non-committal about the truth of universal determinism (thesis iii), preferring to leave that debate to the physicists. But these modern skeptics about free will continue to hold the first two theses of classical hard determinism, namely that (i) free will – in the “true responsibilityentailing” sense – is incompatible with determinism and that (ii) there is, and can be, no incompatibilist (or libertarian) free will of this true responsibility-entailing kind.

One of these modern skeptics about free will, Pereboom calls this successor view to hard determinism, hard incompatibilism, which is a useful designation for those who hold theses (i) and (ii), but remain noncommittal about thesis (iii). Why do hard incompatibilists continue to believe that incompatibilist or libertarian free will does not exist, if they are unsure of the truth of universal determinism? There are several reasons. First, while hard incompatibilists remain non-committal about indeterminism in physics generally, they tend to believe that human behavior is regular and determined for the most part. If indeterminism does exist in the microphysical world, in the behavior of elementary particles, its macroscopic effects on human behavior, they argue, would be negligible and of no significance for free will. Second, hard incompatibilists are convinced by developments in sciences other than physics – in biology (greater knowledge of genetic influences), neuroscience, psychology, psychiatry, social and behavior sciences – that more of our behavior than previously believed is determined by causes unknown to us. For example, controversial neuroscientific experiments of Libet (2002) and others have led psychologists, such as Wegner (2002), to argue that our familiar experiences of conscious willing may be an “illusion”. New research in the neurosciences in general has had an increasing impact on free will debates. (For discussions of this impact, see Walter, 2001; Dennett, 2003; Wegner, 2002; and the essays in Libet et al., 1999. For discussions of the implications of quantum physics and other developments in the physical and behavioral sciences for free will, see the essays of Hodgson and Bishop, in Kane, 2002; Earman, 1986; and the essays in Atmanspacher and Bishop, 2004.) There is a third reason why hard incompatibilists are skeptical of an indeterminist or libertarian free will. They insist that if quantum indeterminism at the micro-level did sometimes have macroscopic effects on the human brain or behavior, such indeterminism would be of “no help” to believers in libertarian free will, since such indeterminism 23

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

f r e e w ill would not enhance, but would only diminish, freedom and responsibility. Suppose a choice was the result of a quantum jump or other undetermined event in a person’s brain, they argue. Such undetermined effects in the brain or body would be unpredictable and occur by chance, like the sudden occurrence of a thought or the spasmodic jerking of an arm – quite the opposite of what we take free and responsible actions to be. Undetermined events in the brain and body would therefore undermine our freedom rather than enhance it. Hard incompatibilists have also been concerned with the impact their denial of free will would have for morality and the meaning of life. Some of them, such as Honderich and Pereboom, argue that we can still live meaningful lives without the illusion of libertarian free will, though some important “life-hopes” and attitudes would have to change. For example, we could no longer believe that criminal punishment was ultimately deserved. Yet, we could still incarcerate criminals to deter them and others from committing future crimes or to reform them. But other philosophers, such as Smilansky (2000), who also believe libertarian free will is impossible, argue that the effects on society and moral life would be dire if most people became convinced that we do not have an incompatibilist or libertarian free will. Smilansky provocatively suggests that while we do not have such an incompatibilist free will, we must continue to foster the illusion which most ordinary persons share that we do have such a free will for the sake of morality and social order. l i b e r t ar ian views o f fr ee will Libertarianism is the name usually given to those who hold that (i) free will and determinism are incompatible, (ii′) free will (in this incompatibilist sense) exists and so (iii′) determinism is false. Libertarianism about free will in this sense should not be confused with the political doctrine of libertarianism, the view that governments should be limited to protecting the liberties of individuals so long as the individuals do not interfere with the liberties of others. Libertarianism about 24

free will and political libertarianism share a name – from the Latin liber, meaning free – and they share an interest in freedom. But libertarians about free will are not necessarily committed to political libertarianism and may (and many do) hold differing political views. To defend libertarianism about free will, one has to do more than merely argue for the incompatibility of free will and determinism. One must also show that we can actually have a free will that is incompatible with determinism. Many philosophers, including both hard determinists and compatibilists, have argued that an incompatibilist free will of the kind that libertarians affirm is not even possible or intelligible and that it has no place in the modern scientific picture of the world. Critics of libertarianism note that libertarians have often invoked obscure and mysterious forms of agency or causation to defend their view. In order to explain how free actions can escape the clutches of physical causes and laws of nature, libertarians have sometimes posited a disembodied mind or soul in the manner of Descartes, which is outside of the physical realm and not governed by physical laws, yet capable of influencing physical events. Other libertarians, such as Kant, have appealed to a noumenal self, outside space and time, not subject to scientific causes and explanations. Still other libertarians, such as Reid, appeal to a special kind of agent- or immanent causation that is irreducible to ordinary forms of causation (see the extended essay) in terms of events common to the sciences. Appeals such as these, and other appeals by libertarians to uncaused causes or unmoved movers, have invited charges of obscurity or mystery against libertarian views of free will by their opponents. Even some of the greatest defenders of libertarianism, such as Kant, have argued that we need to believe in libertarian freedom to make sense of morality and true responsibility. But we cannot completely understand such a freedom in theoretical and scientific terms. The problem that usually provokes skepticism about libertarian free will has to do 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 seem to be compatible with indeterminism either. Events that are undetermined, such as quantum jumps in atoms, happen merely by chance. So if free actions must be undetermined, as libertarians claim, it seems that they too would happen by chance. But how can chance events be free and responsible actions? To defend their view, libertarians must not only show that free will is incompatible with determinism, they must also show how free will can be compatible with indeterminism. Libertarian accounts of free will have taken a number of different forms in the attempt to address this problem. It is now customary to distinguish three main types of libertarian theories: (1) non-causalist (or simple indeterminist) views; (2) causal indeterminist or event-causal views; and (3) agent-causal views. Non-causalist or simple indeterminist libertarian views rely on a distinction between two ways of explaining events, explanations in terms of reasons and purposes (desires, beliefs and intentions) and explanations in terms of causes. Non-causalists, such as Ginet (1990) and McCann (1998), argue that free actions can be explained in terms of the agent’s reasons for action (desires, beliefs, etc.), without being caused or determined, because explanations in terms of reasons are not causal explanations. Noncausalist views raise important questions of action theory about the nature of action, about the distinction between actions and other events (see event theory), about whether reasons for action can be causes of action, and so on. Critics of non-causalist or simple indeterminist views note that, for non-causalists, free actions are literally uncaused events, and the critics raise questions about how events that are uncaused can be under the control of agents. Agent-causalist libertarians follow Reid in postulating a special kind of causation by an agent or substance that does not consist in causation by events or states of affairs, as is common for forms of causation studied by the sciences. Agent-causalists, such as Chisholm and O’Connor (2000), insist against simple indeterminists that, while free actions may

be uncaused by events, they are not uncaused by anything. Free actions are caused by the agents themselves in a sui generis way that is not reducible to causation by states or events of any kinds involving the agent, physical or mental. Other agent-causalists, such as Clarke (2003), allow that reasons for action (such as desires and beliefs) can causally influence choices and actions. But he also postulates a special non-event causation by agents to explain what tips the balance between reasons for one choice or action rather than another. Critics of agentcausal theories, such as Watson, argue that appeals to a special kind of non-event causation by substances are no less mysterious than Kantian appeals to noumenal selves or Cartesian appeals to disembodied minds to explain undetermined free choices. Agentcausalists have attempted to rebut such charges in various ways. (For an overview of the debates see the essays of O’Connor et al., in Kane, 2002; Clarke, 2003, ch. 10). Causal indeterminist or event-causal views (the third kind of libertarian theory) are of more recent origin. Such views were first suggested, though not developed in detail, in the 1970s by David Wiggins and Robert Nozick as an alternative to non-causalist and agent-causal views. The first fully developed causal indeterminist view was that of Kane (1985, 1996). Causal indeterminists attempt to explain undetermined choices without appealing to claims that reasons cannot be causes of actions and without appealing to “extra factors” such as noumenal selves, disembodied minds, or non-event agent causes to explain free actions. Causal indeterminists allow that free actions may be caused by reasons, intentions and other states and processes of the agent, but not deterministically caused. “Undetermined”, they point out, need not mean “uncaused”: Reasons can cause actions nondeterministically or probabilistically, so that, while libertarian freedom must be indeterminist, it need not be “contra-causal”. Causal indeterminist or event-causal libertarian views come in two varieties. “Deliberative” views (first suggested by Dennett and later developed by Mele, 2006 and Ekstrom, 2000) place the indeterminism 25

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, in the undetermined “coming to mind” of thoughts, memories and other considerations that influence subsequent choice. By contrast, so called “centered” causal indeterminist views (such as that of Kane) insist that indeterminism can in some cases persist right up to the moment of choice itself. An important criticism of causal indeterminist views of the centered variety is the so-called “luck objection”, an objection that has been used against other libertarian views as well, agent-causal and non-causalist. (See Mele, 2006 and Haji, 2003 for extended discussions of this objection.) Mele puts the luck objection this way: Suppose John fails to resist the temptation to tell a lie. If his choice to lie is a free act in the libertarian sense then it must have been undetermined up to the moment it was made. This means John could have done otherwise (could have chosen not to lie), given exactly the same past up to that moment (since indeterminism implies “same past, different possible outcomes”). Thus we can imagine a counterpart of John, John*, in an alternative possible world with exactly the same past who did resist temptation and chose not to lie. Mele argues that, since there is nothing about the powers, capacities, states of mind, moral character and so on that is different in the pasts of John and John* right up to the moment they chose that could explain the difference in their choices, then the difference in their choices would have been merely a matter of luck. John* got lucky in attempting to resist temptation, while John did not; and it would not be fair to reward one and punish the other for what was merely a matter of luck. A considerable literature has been generated by this “luck objection”. Causal indeterminists and other libertarians have tried to answer it in various ways, but many believe it cannot be answered. u l t i m a te r esp o nsib ility One final topic concerning incompatibilist and libertarian views of free will deserves mention. Most arguments for the incompatibility of free will and determinism, like the 26

Consequence Argument, have appealed to the requirement of alternative possibilities or AP, or branching paths into the future. But a number of modern incompatibilists about free will, have argued that another requirement of free will, a requirement of ultimate responsibility or UR, is as important as AP, perhaps even more important, to debates about the incompatibility of free will and determinism. The basic idea of UR is this: To be ultimately responsible for an action, an agent must be responsible for anything that is a sufficient cause or motive for the action’s occurring. If, for example, a choice issues from and can be sufficiently explained by an agent’s character and motives (together with background conditions), then to be ultimately responsible for the choice, the agent must be at least in part responsible by virtue of choices or actions voluntarily performed in the past for having the character and motives he or she now has. Compare Aristotle’s claim that if a man is responsible for the wicked acts that flow from his character, he must be responsible for forming the wicked character from which these acts flowed. The importance of this UR condition was first noted in recent free will debates independently by Kane (1985) and G. Strawson (1986). Both agreed that UR could not be satisfied in a deterministic world, so it provided a further argument for the incompatibility of free will and determinism that did not appeal to AP. But Kane and Strawson disagreed about whether UR was an intelligible or satisfiable condition. Kane, a libertarian, attempted to show that UR could be satisfied. While Strawson, a hard incompatibilist, argued that UR was an unsatisfiable condition since it would either require an impossible infinite regress of past voluntary actions by which we formed our later characters or it would require some initial character-forming acts that were not determined by prior character. Such initial acts would either be determined by something external to the agent or would occur merely by chance. This regress argument, which Strawson called the “Basic Argument”, poses a significant challenge 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 been an important part of current free will debates. The requirement of ultimate responsibility or UR has played another role in free will debates. Some incompatibilists, now called “source incompatibilists” (including some hard incompatibilists, such as Pereboom, and some libertarians, such as Eleonore Stump and Linda Zagzebski) argue that UR is the primary condition required for an incompatibilist free will and that alternative possibilities (AP) are not required for free will at all. “Source incompatibilists” of this sort are now often distinguished from “leeway incompatibilists”, who hold the more traditional view that AP is the primary reason why free will and determinism are incompatible. Disputes between these two views about the comparative importance of UR and AP for free will have thus also become a significant part of modern debates about free will.

b i b l i og rap hy Atmanspacher, H. and Bishop, R., ed.: Between Chance and Choice: Inter-disciplinary Perspectives on Determinism (Thorverton, Devon: Imprint Academic, 2002). Clarke, R.: Libertarian Accounts of Free Will (Oxford: Oxford University Press, 2003). Dennett, D.: Elbow Room (Cambridge, MA: MIT Press, 1984). Dennett, D.: Freedom Evolves (New York: Vintage Books, 2003). Earman, J.: A Primer on Determinism (Dordrecht: Reidel, 1986). Ekstrom, L. W.: Free Will: A Philosophical Study (Boulder, CO: Westview, 2000). Fischer, J. M.: The Metaphysics of Free Will: A Study of Control (Oxford: Blackwell, 1994). Fischer, J. M. and Ravizza, M.: Responsibility and Control: A Theory of Moral Responsibility (Cambridge: Cambridge University Press, 1998). Frankfurt, H.: “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68 (1971): 5–20.

Ginet, C.: On Action (Cambridge: Cambridge University Press, 1990). Haji, I.: Deontic Morality and Control (Cambridge: Cambridge University Press, 2002). Honderich, T.: A Theory of Determinism, 2 vols. (Oxford: Clarendon Press, 1988). Kane, R., ed. Free Will and Values (Albany, NY: State University of New York Press, 1985). Kane, R., ed.: The Oxford Handbook of Free Will (Oxford and New York: Oxford University Press, 2002). Kane, R., ed.: The Significance of Free Will (Oxford: Oxford University Press, 1996). Libet, B., Freeman, A., and Sutherland, K., ed.: The Volitional Brain: Towards a Neuroscience of Free Will (Thorverten, Devon: Imprint Academic, 1999). McCann, H.: The Works of Agency: On Human Action, Will and Freedom. (Ithaca, NY: Cornell University Press, 1998). Mele, A.: Free Will and Luck (Oxford: Oxford University Press, 2006). Nagel, T.: The View from Nowhere (New York: Oxford University Press, 1986). O’Connor, T.: Persons and Causes: The Metaphysics of Free Will (New York: Oxford University Press, 2000). Pereboom, D.: Living Without Free Will (Cambridge: Cambridge University Press, 2001). Smilansky, S.: Free Will and Illusion (Oxford: Clarendon Press, 2000). Strawson, G.: Freedom and Belief (Oxford: Oxford University Press, 1986). Strawson, P. F.: “Freedom and Resentment,” Proceedings of the British Academy 48 (1962), 1–25. van Inwagen, P.: An Essay on Free Will (Oxford: Clarendon Press, 1983). Wallace, R. J.: Responsibility and the Moral Sentiments (Cambridge, MA: Harvard University Press, 1994). Walter, H.: Neurophilosophy of Free Will (Cambridge MA: MIT Press, 2001). Watson, G. “Free Agency,” Journal of Philosophy 72 (1975), 205–20. Widerker, D.: “Libertarianism and Frankfurt’s Attack on the Principle of Alternative Possibilities,” The Philosophical Review 104 (1975), 247– 61. 27

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 Responsibility and Alternative Possibilities (Aldershot, Hants: Ashgate Publishers, 2003). Wegner, D.: The Illusion of Conscious Will (Cambridge MA: MIT Press, 2002). Wolf, S.: Freedom Within Reason (Oxford: Oxford University Press, 1990). robert kane

Individuation For reasons which will become clear, it is appropriate to begin a general account of individuation with some discussion of sortal terms and concepts. The expression “sortal” is a coinage of John Locke’s (Locke, 1975, III, III, p. 15). He held a sortal name to signify the complex general idea of a certain sort of things (Locke, 1975, III, VI, p. 1). Prime examples of sortal terms, sometimes also called substantival general terms, are “cat”, “apple”, “mountain”, and “star”. Sortal terms may be contrasted with adjectival terms, such as “red”, “round”, and “heavy”. It is commonly said that the key distinction between sortal and adjectival terms is that while both possess criteria of application, only the former possess criteria of IDENTITY (Dummett, 1981, pp. 547–8). A criterion of application for a general term tells us what it applies to. In other words, it determines the extension of the term: the SET of entities all and only the members of which are correctly described by the term, such as the set of cats in the case of the sortal term “cat” and the set of red things in the case of the adjectival term “red”. A criterion of identity for a sortal term tells us what determines whether or not one thing that the term applies to is the same as, or numerically identical with, another thing that the term applies to: whether or not, for instance, the cat that is now sitting on the mat is the same cat as the cat that was formerly sleeping on the sofa. Where “K” is a sortal term, the general form of a criterion of identity will be this: If x and y are Ks, then x is identical with y if and only if x is RK-related to y (Lowe, 1989b). Here “RK” denotes a certain equivalence relation on Ks – a relation which 28

must, of course, be distinct from identity itself in order for the criterion in question to be informative and non-circular. (An equivalence relation is one that is reflexive, symmetrical, and transitive.) An adjectival term lacks a criterion of identity because there is no single condition that things to which it applies must satisfy in order to be identical (other than, trivially, identity 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 depends at least in part on what sort or kind of red things they are – and then the relevant criterion of identity will be that supplied by the relevant sortal term, be it, say, “cat”, “apple”, or “star”. Sortal concepts are what sortal terms express or convey – although, of course, we shouldn’t assume that for every sortal concept there exists a sortal term (much less a sortal term in every natural language) which expresses or conveys it. Another name for sortal concepts is individuative concepts, for reasons that will become plain when I come, in a moment, to introduce the notion – or, rather, the notions – of individuation. But what, quite generally, are concepts supposed to be? Of course, this in itself is a highly contentious question. Here I shall simply state one widely held view of the matter, which is that a concept is a way of thinking of some thing or things (Lowe, 2006, pp. 85–6). Since thought is a mental process, this means, in effect, that concepts are mental properties of a certain kind. For properties or qualities, quite generally, are ways of being – ways things are (Lowe, 2006, pp. 14, 90–1). For example, roundness is a way of being shaped and redness is a way of being colored. By the same token, concepts, being ways of thinking of things, are ways of being and hence properties – and, more specifically, mental properties, since thought is a mental process. So much for the ontology of concepts. But we speak of thinkers grasping or failing to grasp concepts. We may take this simply to be a matter of their being able, or not being able, to think of things in certain ways. Someone who grasps the 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 – in a certain way. What way is that? Well, of course, such a person is able to think of certain living organisms as being cats. And what does this involve? Well, among other things, it involves being able to think of these organisms as possessing certain characteristic properties, such as furriness and warmbloodedness, and – most importantly for present purposes – as satisfying a certain criterion of identity. We needn’t suppose, however, that a person who grasps the concept of a cat must be able to articulate such a criterion in an explicit form, in line with the general form of a criterion of identity stated earlier. Indeed, it is notoriously difficult – even for philosophers – to formulate clear and uncontroversial criteria of identity for many kinds of things, even when we seem to have a good implicit grasp of such criteria that is manifested in our ability to make confident identity-judgments concerning things of those kinds. So far, we have discussed sortal terms and sortal concepts. In addition, however, there are sorts or kinds, which sortal terms and concepts purportedly designate. I say “purportedly” for this reason if for no other: even granted that a sortal term or concept may designate a really existing sort of things, we can hardly insist that it must do so. The point is exactly parallel to one that may be made concerning adjectival terms and concepts or, more generally, predicates and predicative concepts: that they may, but need not, designate anything. For example, it is natural to suppose that “red” denotes a certain color property or quality, redness. But, for familiar reasons, it may disputed whether there really are any color properties at all. It would, of course, be quite extravagant to suppose that cats don’t exist, but the history of human thought is replete with examples of sortal terms that failed to designate anything, such as “mermaid”, “dragon”, “unicorn”, and “centaur”. What, however, should we say concerning the sortal terms that do designate or denote something: what, exactly, do they denote? Various answers are possible, one being that they denote, in plural fashion, all of the various particular things to which they are

applicable: so that the sortal term “cat”, for instance, denotes the cats – all of them – that exist (or, perhaps, all that do, did, or will exist). Another view and more popular view is that a sortal term that has denotation denotes a sort or kind of things conceived as a type of UNIVERSAL , which has as its particular instances all of the particular things to which it is applicable. According to this view, the sortal term “cat” denotes a substantial universal or kind, whose particular instances are all the individual cats that do, did, or will exist (Lowe, 1989a, pp. 157– 63). Now let us turn to another key notion that needs to be clarified for present purposes: that of an object. This is a philosophical term of art, which admits of various different interpretations, some narrower than others. In its very broadest use, “object” is interchangeable with the very general all-purpose term “entity”. In this sense, anything whatever that does or could exist is an “object”, including numbers, properties, propositions, events, surfaces, waves, holes, and places, as well as common-or-garden material objects, such as apples, tables, and rocks. However, I propose to use the term “object” more narrowly to mean an entity that does at least possess determinate identity conditions and the kind of unity that makes it something that is, at least in principle, countable (Lowe, 1998, pp. 58–61, Lowe, 2006, pp. 75– 6). Some of the items listed earlier do not indisputably qualify as objects by this criterion: for example, waves do not. In what follows we shall restrict our attention for the most part to material objects. However, it is important to recognize that the notion of a material object is still a very broad notion indeed. Crucially, material objects do not collectively constitute a sort or kind in the sense discussed earlier. In other words, “material object” is not a sortal term and does not express or convey a sortal concept. The reason is simple enough: it is simply not the case that all material objects are governed by the same criterion of identity. Thus, for example, both cats and mountains are material objects, but they do not share the same criterion of identity. All it takes for something to qualify as a material object 29

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 sense, and (b) be composed of matter. Both cats and mountains qualify by this standard, as do many other material objects governed by yet other criteria of identity, such as tables and stars. The next important thing to notice is this. Although two different sortal terms, each designating a different sort or kind of things, may convey different criteria of identity for the individual objects to which they apply, this is not necessarily the case and, indeed, is very often not the case. Very often, two such sortal terms convey exactly the same criterion of identity. This is the case, for instance, with the sortal terms “cat” and “dog” – and, indeed, with all sortal terms denoting kinds of living organism (Lowe, 1998, p. 45). Particulars of all these kinds share the same criterion of identity, which is that of living organisms in general. So it is likewise with all kinds of material artefact, for instance, such as tables and computers: they all share the same criterion of identity, which differs from that governing living organisms. But why, it may be asked, must we suppose that all living organisms, say – and certainly all animals – share the same criterion of identity? For the following reason. “Animal” – unlike, for instance, “material object” – does at least appear to be a sortal term in good standing, conveying a criterion of identity for the objects to which it applies. After all, we can always intelligibly ask whether an individual animal encountered on one occasion is or is not identical with another individual animal encountered on another occasion – and in order to determine the answer to such a question, we do not necessarily need to know what sort or sorts of animal these individuals are. Indeed, we may well be uncertain, at least at an early stage of our inquiries, whether we are confronted with just one sort of animal or two, because the individual animals encountered on the two occasions may exhibit very considerable morphological differences, as in the case of a tadpole and a mature frog. However, cats and dogs, say, clearly are both sorts of animal. (Indeed, they are clearly different sorts of animal.) But in that case the sortal terms “cat” and “animal” must 30

convey the same criterion of identity, as must the sortal terms “dog” and “animal”, on pain of incoherence. For it is not even metaphysically possible that objects of kinds governed by different criteria of identity should be identical (Lowe, 1989a, ch. 4). Hence if the sortal terms “cat” and “animal”, say, conveyed different criteria of identity, no individual cat could be identified with any individual animal, which is plainly absurd. But if “cat” and “dog” must, for the foregoing reason, both convey the same criterion of identity as “animal” does, then they must clearly convey the same criterion of identity as each other – and the same applies in the case of all other sortal terms denoting animal kinds. This, then, is why I maintain that all animal kinds share the same criterion of identity. The foregoing discussion, if it is along the right lines, reveals that general names – as Locke would have called them – fall into at least three distinct classes. First, there are non-adjectival general terms like “material object” which are certainly not sortal terms, because they do not convey any criterion of identity whatever. Second, there are regular sortal terms, such as “cat”, “dog”, “mountain”, “star”, and “table”, which not only convey a criterion of identity but also purportedly denote certain distinct sorts or kinds. Intermediate in generality between these two classes of general names are non-adjectival general terms like “living organism” and “material artefact”, which do convey a criterion of identity but are too general to qualify as regular sortal terms. What these terms designate are not, properly speaking, specific sorts or kinds but, rather, certain ontological CATEGORIES – or, more precisely, certain categories of object (compare Dummett, 1981, p. 583). What cats and dogs and all other such sorts or kinds have in common is that they are all kinds of living organism. The individual members of all these kinds all belong to the same ontological category, the hallmark of this fact precisely being that they are all governed by the same criterion of identity. In effect, then, we can identify those general terms that denote ontological categories – categorial 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 criteria of identity for the objects to which they apply. And categorial terms fall, in respect of their degree of generality, in between regular sortal terms and transcategorial terms, such as “material object”. I must emphasize that criteria of identity, on this view, are not empirically discoverable principles, but are, rather, a priori ontological principles which delimit what is and is not metaphysically possible for the objects governed by them (Lowe, 1998, ch. 8). With this stage-setting in place, we can now at last introduce the term “individuation” itself. This term has two senses (Lowe, 2003). In one sense – which we may call the cognitive sense – individuation is a cognitive achievement, consisting in the singling out of an object in thought (compare Wiggins, 2001, pp. 6–7). In this sense, it is we, or thinkers quite generally, who individuate objects, whenever we single them out in thought. But in a quite different sense – which we may call the ontological sense – individuation has nothing to do with cognition or thinkers, but is simply a certain kind of metaphysical determination relation between entities. In this sense, an object is individuated by one or more other entities, its individuator or individuators. An object’s individuators, in this ontological sense, are the entities which determine which object it is. A simple example drawn from the domain of abstract objects will serve for illustrative purposes. A set, then, is individuated, in the ontological sense, by the entities that are its members, at least in all cases in which it has members (not, thus, in the case of the empty set). If a set has members, its members, and these entities alone, determine which set it is. Turning to the case of material objects, we can see that material objects of some kinds are individuated by their material parts (at least at some level of decomposition): for example, a heap of stones is individuated by the stones that make it up, because which heap it is is determined by which stones make it up. (Note, however, that a heap of stones is not individuated by the subatomic particles that make it up at any given time, which is why it is important to specify the relevant level of

decomposition.) Material objects of some other kinds, however, are apparently not individuated by their material parts (at any level of decomposition). Living organisms seem to be a case in point, for they can undergo a change of any of the material parts that they possess at any time during their careers. It is not even clear that – as some philosophers suggest (for instance, Kripke 1980) – living organisms are individuated by the material parts that they possess at their moment of origin, since it seems that these too could always have been different (Lowe, 1998, pp. 165– 6). It should be clear from these examples that metaphysical principles of individuation are closely related to criteria of identity. But they should not be confused with them. A metaphysical principle of individuation tells us what determines the identity of an object, in the sense that it tells us what determines which object it is. A criterion of identity, by contrast, tells us what determines whether an object belonging to a given ontological category is or is not identical with another such object. In the latter case, we are concerned with identity conceived as a relation, whereas in the former case we are concerned with “identity” in the sense of individual ESSENCE (to use a traditional term). Identity in this sense, or individual essence, is – as John Locke aptly put it – “the very being of any thing, whereby it is, what it is”, this being, according to Locke, the “proper original signification” of the word “essence” (Locke, 1975, III, III, p. 15). Plausibly, every object is individuated in the ontological sense. In every case, something – some entity or entities – individuates the object in question, in the sense of determining which object it is. To suppose that there are unindividuated objects seems incoherent. For an unindividuated object would be an object concerning which there was no fact of the matter as to which object it was, and it is very hard to see how this could be the case. Now, clearly, objects of different ontological categories are individuated, in the ontological sense, in very different ways. For example, mountains and islands are individuated, at least partly, by their geographical locations. But living organisms and 31

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, the claim that every object is individuated might raise in the minds of some critics the worry that an infinite regress is thereby threatened. The thought would be that if every object is individuated and, moreover, is individuated by one or more objects, then there is no end to individuation and so, perhaps, no object at all really gets individuated. However, there are two ways, at least, to counter this worry. One is to point out that it was implied earlier only that every object is individuated by some entity or entities, but not that the entities in question must always themselves be objects. Indeed, I said that mountains and islands are partly individuated by their geographical locations, but geographical locations are doubtfully objects at all and are certainly not material objects. Another point to bear in mind is that nothing said so far implies that objects may never be self-individuating. In fact, it is plausible to claim that what we may call material SUBSTANCES are indeed selfindividuating, including living organisms. According to this view, for example, what determines which animal a given animal is is nothing other than that very animal. The idea that some objects are selfindividuating is certainly far from being absurd (Lowe, 2003). Indeed, in some cases it seems extremely compelling: for instance, in the case of the empty set. For, given that every set is individuated and that sets which have members are individuated purely by those members, we seem to have little option but to say that the empty set individuates itself, for it has no members to individuate it in the only way that other sets are individuated. In opposition to this view, it might be suggested that the empty set is in fact individuated by a certain property that it alone possesses and possesses necessarily – the property of being the only set that has no members – and that since this property is an entity that is distinct from the empty set itself, that set is not self-individuating. However, this assumes that the predicate “is the only set that has no members”, which undoubtedly applies uniquely to the empty set, does indeed denote a certain property which that set possesses. But, as has 32

already been noted, we cannot uncritically assume that every predicate denotes a property, if by a property we mean some really existing entity, be it a universal or a so-called trope. There is no obvious reason to suppose that the predicate now in question denotes a property in this sense. That being so, it is hard to see what we can say about the empty set other than that it is selfindividuating. It alone is the only entity that determines which set it is, since nothing else does. However, it may seem that, because the empty set is an abstract object, we can draw few lessons from its case when considering the individuation of material objects. But that conclusion would be too hasty. For what makes it plausible to say that the empty set is self-individuating is the fact that it is an object that does not appear to depend for its identity on anything other than itself (on this notion of identity-dependence, see Lowe, 1998, pp. 147–9). But this also seems to be a characteristic of what we are calling material substances, including living organisms. We may take it to be an essential feature of such substances that, even though they are composed of matter, they are capable of changing their material parts and, indeed, could have been made up, at any given time, of material parts numerically distinct from those that actually make them up at that time. This, if true, is why they do not depend for their identity upon such parts, in the way that something like a heap or pile of stones does. But, given that they do not depend for their identity upon their material parts, it is not clear what else they could depend on for their identity, other than simply themselves. Perhaps, in the end, saying that material substances are self-individuating is not so very different from saying, as some metaphysicians do, that they are individuated by their so-called HAECCEITIES or thisnesses (Rosenkrantz, 1993). According to this view, what determines which animal a given animal, a, is is that animal’s haecceity – its property of being that animal or, in other words, its property of being identical with a. But, it may be asked, are there really 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 identical with a really an entity that is distinct from a itself? Some metaphysicians may find it hard to believe so. But if haecceities are not genuine entities in their own right, it is difficult to see what it can mean to say that animal a is individuated by its “property” of being identical with a, other than simply to say that a individuates itself – that a itself is the only entity that determines which animal a is. And this is, I think, a perfectly coherent thing to say. Nor should it be supposed that once we say this about some objects, we shall be obliged to say it about all objects. For we have already seen that there are plenty of objects, such as piles of stones and sets that have members, which are plainly not self-individuating. Furthermore, it seems very reasonable to say that at least some entities must be self-individuating, on pain of the sort of infinite regress that was mooted earlier. (Thus, if there are haecceities, must not they be self-individuating?) So why not say this about material substances, together with, perhaps, other objects such as the empty set? Anyway, let us adopt it as a working assumption in what follows that material substances, including animals and other living organisms, are indeed selfindividuating in the ontological sense. So far, however, I have said very little about individuation in the cognitive sense, but this notion too raises important metaphysical issues, concerning the nature of thought. What I did say is that individuation in this sense is a cognitive achievement, consisting in the singling out of an object in thought by a thinker, that is, by a person. A sortalist, in this connection, is a theorist who maintains that a thinker can successfully single out an object in thought on a given occasion only as an object of some specific sort, that is, as falling under or satisfying some specific sortal concept – a concept that the thinker in question must therefore grasp and apply in individuating that object on that occasion, in the cognitive sense of “individuate”. An anti-sortalist, correspondingly, is a theorist who denies the foregoing claim. On the face of it, the sortalist thesis as I have just formulated it is clearly too strong. For, it may be urged, a thinker can surely successfully

single out an object in thought before having any conception of what sort of object it is. For example, in thinking about a particular animal – let’s call it Tom – a thinker surely need not be able to single out Tom as being, say, a cat, as opposed to a dog, or a pig. Maybe so. But can a thinker successfully single out in thought a particular animal, such as Tom, without even grasping that Tom is an animal, or at least a living organism? Is it possible, for example, for a thinker successfully to single out in thought a particular animal, Tom, while grasping only that Tom is a material object? It is hard to see how this can be possible. For, it seems, one cannot successfully single out an object in thought without grasping which object it is that one has thus singled out. However, this is the point at which the cognitive and the metaphysical notions of individuation come together in a crucial way. Which object a given object is is something that is determined by that object’s individuator or individuators and, as we have seen, objects of different types have different types of individuator. Material objects as such have no single type of individuator, because material objects as such do not constitute an ontological category but, rather, fall into many diverse ontological categories, such as living organisms, material artefacts, and geological formations. Turning aside, for a moment, from the case of material objects, consider the following question: can we intelligibly suppose that a thinker could successfully single out in thought an abstract object, such as a set – for example, the set of prime numbers smaller than 10, {2, 3, 5, 7}, or the set of planets closer to the sun than Jupiter, {Mercury, Venus, Earth, Mars} – without grasping that the object in question is indeed a set and thereby grasping its criterion of identity and principle of individuation? For, assuming as we now are that the set in question is non-empty and therefore individuated by its members, how could a thinker know which object this set is without grasping what it is that determines which object it is, namely, its members – something that, it seems, requires the thinker to grasp that what this object is is a set. But if a thinker 33

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 thinking about, how can he really be said to have “singled out that object in thought”? To single out an object in thought is, at the very least, to think something about that very object. But how can a thinker’s thoughts be said to fasten upon a certain object in particular, as opposed to some other object, if that thinker cannot even be said to know which object it is that he is thinking about? It doesn’t appear that matters are fundamentally different in the case of thoughts about material, as opposed to abstract, objects. Accordingly, it is hard to see how a thinker could successfully single out a material object in thought while conceiving of it as nothing more specific than a material object – that is, as an object composed of matter. For conceiving of an object in this way would leave entirely open the question of what determined which object it is – and yet, without his having a grasp of what a correct answer to that question would be it is hard to see how a thinker could be said to know which object he was, supposedly, thinking about. Thus, while we should be happy to allow that a thinker can successfully single out a material object in thought without conceiving of it as belonging to some quite specific sort or kind, such as the kind cat, or the kind table, or the kind mountain, we should insist that he must grasp, at least implicitly, to what ontological category the object in question belongs – such as living organism, or material artefact, or geological formation. This is not at all to imply, of course, that the thinker need be able to have a linguistic command of such categorial terms as these, only that he must have at least an implicit grasp of the relevant criteria of identity and principles of individuation. For without such a grasp the thinker cannot really be said to know what it is that he is, supposedly, thinking about. And without knowing that, he cannot really be said to have singled out an object in thought. However, it is unlikely that this claim will go entirely unchallenged. One kind of challenge that is likely to be raised against it focuses on the perceptual capacities of thinkers. Against sortalists, it is sometimes complained that their position improbably 34

requires us to suppose that thinkers cannot perceive objects which do not fall under sortal concepts grasped by them. It is then pointed out that very frequently we find ourselves perceiving some object while simply having no idea at all what sort of object it is that we are perceiving. This may happen when, for example, an archaic artefact of unknown purpose is dug up and we ask ourselves, “What on earth is this – a drinking vessel, perhaps, or an oil lamp, or something designed to be used in a religious rite?” (compare Campbell, 2002, pp. 70–1). We undoubtedly see and feel the object, however, and can talk about it intelligibly. So is this not a case in which we have managed to single out the object in thought but without having a sortal conception of it, quite contrary to the sortalist thesis? The first thing that must be said about this type of example is that, of course, we have already conceded that the sortalist thesis is too strong. The most that we should say is that we cannot single out an object in thought without having, at least implicitly, a categorial conception of it, and thereby having at least an implicit grasp of the criterion of identity that the object satisfies. We could call this the categorialist thesis, as opposed to the stronger sortalist thesis. The latter is stronger, because it implies the former, but the reverse is not the case. Now, in the foregoing archaeological example, no challenge to the categorialist thesis was even threatened, since we were supposing the discoverers of the mysterious object in question to be convinced, at least, that what they had found was a material artefact of some kind – and material artefacts constitute an ontological category. In this context, it is vitally important to distinguish between thought and perception. The categorialist thesis is the claim that a thinker cannot successfully single out an object in thought without conceiving of that object as falling under a certain ontological category and thereby grasping a corresponding criterion of identity that he conceives it to satisfy. But perceiving is not thinking and there is no reason at all why the categorialist should not accept that a person can perceive an object without having any 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 compelling reasons to accept precisely this. For it is evident that many non-human animals perceive objects in their immediate environment, even though it would be utterly extravagant to suppose that those animals are capable of categorizing those objects ontologically or grasping the relevant criteria of identity for those objects. A dog, for instance, can surely see its feeding bowl, without recognizing that what it sees is a material artefact. But, conceding this, let us then ask: can the dog successfully single out that object in thought? Can the dog think about its feeding bowl – that very object, as distinct from any other? There seems to be no compelling reason to suppose that it can. We may conclude now with a final question: what cognitive significance, if any, is there in the fact – assuming that it is a fact – that material substances are, in the ontological sense of individuation, selfindividuating? There seems to be considerable cognitive significance in this fact. For what it apparently implies is that it is sufficient for a thinker to be able to single out such a substance in thought that that thinker should have perceived that substance at some time, knowing on that occasion that what he was perceiving was an instance of a certain category of material substance, and have retained a memory of this experience. For instance, having seen a certain animal, a, knowing that what I was then seeing was an animal, and remembering this perceptual encounter with a, I can subsequently have singular thoughts about a – that is to say, I can continue to single out a in thought. In other words, I continue to know which animal a is, even if I never again have perceptual contact with a. To put this another way, I continue to grasp a’s individual essence. For, we are supposing, what individuates a, in the ontological sense, is just a itself – it is just a itself that determines which object a is. Hence, my perceptual encounter with a, provided that it is informed by a grasp of the category of object to which a belongs – and thus a grasp of a’s general essence, which it shares with all other members of the same category –

makes me acquainted with a’s individuator. But if one grasps an object’s principle of individuation and is also acquainted with the entities which, according to that principle, are its individuators, then one knows which object it is. For example, if I grasp the principle of individuation for sets and am acquainted with the prime numbers smaller than 10, then I know which set, and hence which object, the set of prime numbers smaller than 10 is. What is special about material substances – together, maybe, with some other objects, such as the empty set – is that a thinker does not need to be acquainted with anything else in order to be acquainted with such a substance’s individuator, so that a grasp of such a substance’s general essence together with perceptual acquaintance with that substance provides a thinker with a grasp of that substance’s individual essence and thereby an ability to single out that substance in thought, that is, to individuate it in the cognitive sense. But whether this is the only way in which a thinker can acquire a grasp of the individual essence of a material substance is another and difficult question. See also the a–z entry on individuation. bibl i ography Campbell, J.: Reference and Consciousness (Oxford: Clarendon Press, 2002). Dummett, M.A.E.: Frege: Philosophy of Language, 2nd edn. (London: Duckworth, 1981). Kripke, S.A.: Naming and Necessity (Oxford: Blackwell, 1980). Locke, J.: An Essay Concerning Human Understanding, ed. P. H. Nidditch (Oxford: Clarendon Press, 1975 [1690]). Lowe, E.J.: The Four-Category Ontology: A Metaphysical Foundation for Natural Science (Oxford: Clarendon Press, 2006). Lowe, E.J.: “Individuation,” in M.J. Loux and D.W. Zimmerman, ed., The Oxford Handbook of Metaphysics (Oxford: Oxford University Press, 2003), 75–95. Lowe, E.J.: Kinds of Being: A Study of Individuation, Identity and the Logic of Sortal Terms (Oxford: Blackwell, 1989a). 35

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: Substance, Identity, and Time (Oxford: Clarendon Press, 1998). Lowe, E.J.: “What Is a Criterion of Identity?,” The Philosophical Quarterly 39 (1989b), 1–21. Rosenkrantz, G.S.: Haecceity: An Ontological Essay (Dordrecht: Kluwer, 1993). Wiggins, D.: Sameness and Substance Renewed (Cambridge: Cambridge University Press, 2001). e.j. lowe

The Mind/Body Problem Reinhardt Grossmann calls the mind “the great garbage bin of ontology” (1983, p. 256). What seems real but lacks physical respectability we consign to the mind. A long tradition places “secondary qualities” (colors, tastes, sounds, odors) in the mind. These are thought, not to be “out there”, but to be “subjective” transitory occurrences in the minds of observers (see quality, primary/ secondary). Hume regarded causation (see the extended essay) as a psychological “projection”, and it seems natural to distinguish the world as experienced from the world as it is. The idea that minds incorporate non-worldly, non-physical elements, however, evidently places minds outside the physical realm. What science casts asunder, philosophers must piece together. Hence the mind–body problem. Although he did not invent the mind– body problem, Descartes (1596–1650) is responsible for its modern formulation (see Matson, 1966). Immediately after proving his existence by noting that the thought expressed by “I exist” must be true if I can so much as consider whether it is true (Meditation 2), Descartes asks, “What am I?” He answers, “a thing that thinks”, a thinking substance. Descartes regards planets and trees, not as substances, but as modes, ways extended matter is organized. On the one hand we have extended substance and its modes: material bodies (see matter). On the other hand, we have thinking substances, minds, and their modes: thoughts, images, feeling (see soul). Just as minds and bodies 36

are irreconcilable (bodies are extended in space, minds are non-spatial), so modes of thought and modes of extension are incommensurable. Now we are faced with a problem: how could mental goings-on have material effects; how could material occurrences affect the mind? This is Descartes’s mind–body problem. In fact there are two problems here. The first arises from the difficulty of understanding how spatial and non-spatial entities could engage causally. The difficulty is especially pressing for Descartes who regards mental and physical substances as operating on very different laws or principles. A second difficulty arises from our conception of the physical world as a self-contained closed system. Physical events have, we suppose, purely physical causes. If non-physical minds affect the physical world, it looks as though they would have to initiate or intervene in physical processes. Were that so, the physical world would not be a closed system governed by physical law – a daunting prospect that threatens the garbage-bin status of the mental. The self-contained nature of the physical world could be expressed in terms of a conservation principle. Descartes, writing before Newton, imagined that what was conserved was motion. Minds could not initiate or inhibit motion in the physical world. Minds could, however, have physical effects without violating physical closure by altering the direction taken by material particles. This solution unraveled with Newton’s introduction of force, which moved physics from Cartesian kinematics to a dynamical system. Nowadays we think that what is conserved is mass–energy. In either case Descartes’s account of mind–body interaction is no longer viable. Malebranche (1638–1715), a Cartesian, sees the problem and rejects interaction. According to Malebranche (and there are suggestions of such a view in Descartes), not only is there no mental–physical causation, there is no purely physical causation. Whatever happens is the result of God’s 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. The world resembles a succession of images on a movie screen. In a movie, events on the screen succeed each other. Their cause, a projector, is not a member of the sequence, however, but something entirely outside it. For Malebranche, God does not cause, but “occasions” events in the world. God does this, not by intervening in worldly processes, but making it the case at every instant that a world exists containing those processes (see occasion, occasionalism). We should not be shocked by the thought that mental events are causally impotent: physical events are in the same boat! Leibniz (1646–1716) depicts a world comprising an infinity of independent substances each reflecting the world from a unique point of view. On this conception the physical world amounts to a “virtual world” made up of these points of view. Events unfold in each substance independently but in perfect harmony with events in every other substance. Causal interaction is a harmless illusion. Both Malebranche and Leibniz skirt the mind–body problem by rejecting mental– physical interaction altogether. If there is no mind–body interaction, there is no mind– body problem. Such maneuvers, however, exact a heavy price. Can we reasonably abandon the idea that physical events are causally connected? Could we ever be satisfied with an account of the world according to which mental occurrences – perceptual experiences, for instance – are not brought about by physical occurrences, and thoughts and decisions never give rise to actions and utterances? Must we settle for the idea that mind–body interaction is illusory? For Descartes, mental and physical substances are, God aside, mutually exclusive and exhaustive. Each kind of substance has a distinctive attribute: mental substances think, but are not extended; physical substances are extended, but do not think. Mental and physical properties are modes of these attributes, determinate ways be being extended or thinking. Being spherical and being red are ways of being extended. An experience of a spherical red object, in contrast is a mode of thought, a way of being conscious.

Many of Descartes’s contemporaries and most of his successors rejected this picture. A mental substance might be a substance with mental properties; a physical substance, one with physical properties. This leaves open another possibility: some substances might have both mental and physical properties, a dualism of properties, not substances. Perhaps mental properties are just distinctive properties of certain complex physical systems. Property dualism can be developed in various ways. According to Epiphenomenalists – T. H. Huxley (1825–1895), for instance – mental occurrences are by-products of brain processes. They resemble squeaks made by a complex machine that play no role in the machine’s operation. When you bark your shin, you feel a pain. This feeling is a result of a chain of events in your nervous system leading from your shin to a region of your brain. In the simplest case, the neurological event that “gives rise to” your painful sensation also produces bodily motions that might otherwise be thought to be caused by the sensation. Conscious states and bodily motions are correlated, not because consciousness is causally efficacious, but because conscious states and bodily motions have common causes. On the one hand, epiphenomenalism enables us to sidestep worries about mental goings-on intervening in the physical world thereby violating closure. On the other hand, we are left with two significant worries. First, as in the case of Malebranche and Leibniz, we will need to abandon the idea that mentality makes a difference in what we do. You might worry about this, not merely because it seems on the face of it implausible, but because it is hard to see how consciousness could possibly bestow any sort of evolutionary advantage on creatures possessing it. True, consciousness could be an evitable by-product of evolutionarily adaptive physical processes, but it is hard to believe that consciousness itself is evolutionarily irrelevant (Nichols and Grantham, 2000). A second worry concerns the production of conscious experiences. These are caused by physical processes in the brain, but how is this supposed to work? What exactly is 37

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 event? Epiphenomenalists tell us that consciousness “arises from” the brain, but what is this “arising from” relation? Mental events presumably involve mental properties, but where are these properties? They seem not to be among those we discover when we probe the brain. Are they invisible? Are they somehow “outside” space or space–time? The Cartesian problem concerned how extended and non-extended things could interact. The problem arises anew for epiphenomenalism in relation to the production of mental properties or events. The situation appears bleak. We have a robust conviction that, although mental and physical properties are utterly different, interaction between minds and bodies is commonplace. The difficulty is to square this with closure, our conviction that the physical world as a whole is causally closed, mass–energy is conserved. (For a dissenting view, see Lowe, 1996.) One elegant solution is to deny the existence of minds and mental properties altogether. If there are no minds, no mental properties, there is no mind–body problem. Hobbes (1588–1679) argued that we are nothing more than elaborate machines. In a way, Hobbes is just extending Descartes’s official view. Descartes held that most human behavior and all behavior of non-human creatures could be explained mechanically. Only in the case of behavior resulting from rational mental processes (most notably linguistic behavior), do we need to posit mental causes. If ratiocination, however, were just a matter of calculation (think of a computing machine to get a feel for what Hobbes has in mind) we would have no need to imagine that our bodies are controlled by minds with distinctive mental properties. A conception of this kind, materialism, can be developed in two ways (see physicalism/ materialism). First, you might think, as Hobbes does, that mental states and properties are “reducible to”, that is identifiable with, physical states and processes. On this view, minds turn out to be brains, mental states and properties turn out to be physical 38

states and properties. Second, you might simply deny that there are minds or mental states or properties (Churchland, 1981; Stich, 1996). To see the difference, consider the discovery of DNA and its consequences for genetics. We now think we can map genes onto complex molecular structures, thereby “reducing” genes to DNA (see reduction, reductionism). Compare reduction of this kind to the demise of phlogiston. Seventeenthcentury chemists explained combustion by supposing that flammable materials contained phlogiston, a fluid driven out when the materials were heated. Advances in chemistry rendered phlogiston superfluous. Phlogiston was not reduced to more fundamental goings-on, but stricken from the scientific inventory. Eliminativists believe a similar fate lies in store for the mind. According to eliminativists, talk of mental states and properties belongs to an outmoded “folk theory” of human and animal behavior. At one time we explained natural occurrences by supposing objects were animated by spirits. Such explanations were gradually supplanted by explanations adverting exclusively to physical processes. Nevertheless, we persist in regarding human bodies (and the bodies of most animals) as animated by spirits. We comprehend the behavior of intelligent creatures by supposing they are conscious of their surroundings and do what they believe will subserve their interests. Advances in neuroscience, however, promise to undermine “folk psychology” and its posits just as chemical discoveries undermined phlogiston. You might worry that this way of framing the issues stacks the deck. Consider ordinary beliefs about ordinary objects: tables, trees, volcanoes. Physics and chemistry assure us that these things are at bottom just clouds of particles. We can explain the behavior of these particles without positing the ordinary entities, and there is no prospect of smoothly reducing the ordinary things to respectable physical– chemical kinds. Should we eliminate tables, trees, volcanoes? Mightn’t it be better to see talk of tables, trees, and volcanoes as reflecting an inventory of genuine objects that happen to 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 deep story about the world, a world that includes the fundamental things and includes as well tables, trees, and volcanoes. These are not add-ons any more that the forest is something in addition to the trees. Whether or not you are moved by such considerations, even tough-minded philosophers have found eliminativism hard to swallow. We can explain away – “eliminate”, consign to the garbage bin – ghosts by supposing that they are illusions, but it is hard to see how this could work with states of consciousness. Illusions seem ineluctably mental. An illusory feeling of pain is still a feeling. Conceiving of mental phenomena as “only in the mind” is scarcely a recipe for their elimination. The problem of reconciling illusions with the physical world is just the mind–body problem all over again. Materialism dissolves the mind-body problem by subtracting the mental as a distinct category (see physicalism/materialism). Others, idealists, move in the opposite direction: all that exists are minds and their contents. The physical world is, as George Berkeley (1685–1753) would put it, a “mere appearance”. (For a more recent variant, see Foster, 1982.) One advantage of idealism is that it is not hard to see how physical objects could turn out to be illusory. A disadvantage is that idealism addresses the world in a way deeply at odds with tenor, if not the substance, of modern science. The sense is that idealism “works”, but only by tossing out the baby with the bath water. The urge for scientific respectability underlies the advent of psychological behaviorism during the first half of the twentieth century. Behaviorists were intent upon distancing themselves from reliance on introspective techniques to study states of consciousness prominent in the nineteenth century. By their lights this meant providing toughminded “operational” characterizations of important concepts and shunning anything that might prove objectively unverifiable (Skinner, 1963). The result was psychology minus the mental trappings. Behavior was to be explained by contingencies of “reinforcement” and “operant conditioning”. We

are conditioned by our involvement with the world to do as we do. The mechanisms are simple but, in combination, yield complex responses. Meanwhile, philosophers, inspired by Wittgenstein’s (1953, §38) insistence that “philosophical problems arise when language goes on holiday”, were crafting a philosophical version of behaviorism. Gilbert Ryle campaigned against the “Cartesian myth”, the conception of minds as “ghosts in the machine”. The mistake, thought Ryle, was to regard mental events as private, inwardly observable goings-on that, while not quite physical, had physical causes (incoming stimuli) and effects (bodily motions). Ryle thought this picture stemmed from a “category mistake” (see categories): representing “the facts of mental life as if they belonged to one logical type or category . . . when they actually belong to another” (1949, p. 16). A child, watching a parade, is told that a regiment is marching past. Puzzled, the child remarks, “I see soldiers, but where is the regiment?” The child thinks a regiment is something alongside or “over and above” the soldiers, a peculiar sort of object. So it is with us and the mind. Scrutinizing the body, we fail to observe the mind and conclude that minds must be organs like the brain but invisible to outside observers. Rather, Ryle thinks, talk of minds and states of mind is a way of indicating what intelligent agents do or would do under various circumstances. Thoughts and feelings are not inner states. Your thinking of Vienna is just a matter of your being disposed to remark on Vienna or respond with “Vienna” when queried. Neither Wittgenstein nor Ryle denied that there were inner states, only that states of mind were identifiable with such states. Their aim was to challenge the picture of mental goings on as being causally related to physical goings on. Your forming an intention to stroll does not cause your subsequent strolling. Puzzling over mind–body interaction in such cases manifests a category confusion. Your intention “illuminates” or “makes sense of” your subsequent action. Actions, which presumably have purely physical causes, are understood “in light of” 39

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 mistake is to see these states as ghostly internal causes of behavior. Despite attempts to move us away from the Cartesian model of minds as inner control centers, philosophical behaviorism came under fire from philosophers who found behaviorist analyses of mental states implausible. Such analyses seek to reduce talk of mental states to talk of behavior or behavioral dispositions (see disposition). If you believe the ice is thin, you will avoid skating on it, or at least be disposed to avoid skating on it – but only assuming that you want not to fall through. Your wanting not to fall through could be analyzed behaviorally, but only by mentioning still further states of mind. What we do or would do depends, it would seem, on interrelations among beliefs and desires, and this resists reductive analysis. Whatever states of mind are, they do seem to affect behavior causally and to be causally responsive to perceptual inputs from the environment. In the 1950s, U.T. Place (1956) and J.J.C. Smart (1959), colleagues at the University of Adelaide, put forward a mind–brain identity thesis. Mental states, although not analyzable in physical terms, might nevertheless be identified with states of the brain: sensations are brain-processes. This is not something that could be worked out solely by attending introspectively to one’s own states of mind, any more than one could work out that lightning is an electrical discharge or that water is H2O, merely by reflecting on familiar properties of lightning and water. Identities of this kind are discoverable only after careful scientific study. When we investigate the brain, we discover that it has the kind of administrative standing in the processing of incoming stimulation and the production of behavior we associate with the mind. The simplest explanation for this coincidence of roles is that the brain is the mind, mental states are states of the brain. Plenty of scientists and non-philosophers have thought this for a long time, why not philosophers? Philosophers see the task of reconciling mental and physical properties as fraught with difficulty. The “feel” of a state 40

of mind, its “what-it’s-like-ness”, its “subjectivity” (Nagel, 1974), seem utterly unlike any physical properties we might hope to discover in the brain. Smart noted that this was so with lightning and electrical discharges, water and H2O. In both cases properties encountered in experience differed from those we discover via scientific investigation, yet this does not prevent us from identifying lightning and water with electrical discharges and H2O, respectively. In the case of water and lightning, however, we compare properties of the appearance of water or lightning with properties of the stuff that gives rise to the appearance. In the case of minds and brains, the roles are reversed. What we are trying to explain are the appearances. It would be futile to suggest that we are aware only of the appearances of states of mind. Philosophical behaviorism succumbed to pressure from the identity theory and transformed itself into functionalism. The stumbling block for psychological behaviorism came with the advent of the computing machine and the Chomskeyean revolution in linguistics. Chomsky (1966) argued that behaviorist categories were hopelessly inadequate to account for human linguistic capacities. At the same time, computing machines were coming to be seen as affording explanatorily tractable models of intelligent behavior. Alan Turing (1950), echoing Hobbes, argued that intelligence could be understood as computation. It would be possible in principle to build a mind by programming a machine that would “process symbols” so as to mimic an intelligent human being. Turing proposed a test for intelligence, the “imitation game”. Start with two people, A and B, a man and a woman, communicating via teletype with a third person, the interrogator. The interrogator queries A and B in an effort to determine which is the woman. The woman must answer truthfully, but the man can prevaricate. A wins the game when he convinces the interrogator that he is B. Now, imagine a cleverly programmed digital computer replacing A. If the machine succeeds in fooling the interrogator 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 intelligent. Despite important advances in technology, events have not born out Turing’s optimistic prediction that machines would pass his test by the turn of the century. Still, work in artificial intelligence (AI) has progressed on several, less adventurous fronts. Although attacks on AI (most famously by Hubert Dreyfus, 1972 and John Searle, 1980) have been inconclusive, philosophical enthusiasm for the thesis that the nature of the mind can be captured by a computer program has waned. One question is whether consciousness might supply some needed spark, and this brings us back to the fundamental mind–body problem. The advent of the digital computer encouraged philosophers to separate what could be called “hardware” questions from questions about “software”. Perhaps we should view the mind, not as a physical machine, but more abstractly, as a program running on a physical machine, the brain. What is important is not the mind’s physical “implementation”, but networks of internal relationships that mediate inputs and outputs. So long as this pattern is preserved, whatever the nature of the underlying “hardware”, we have a mind. This is one way of thinking about functionalism (Fodor, 1968). Functionalists note that we are comfortable ascribing states of mind to very different kinds of physical system. A human being, an octopus, and a Martian could all be said to feel pain, although physical states that might be thought to “realize” pain in each could be very different. This thought led to the thesis that states of mind are “multiply realizable”. A property – the pain property, for instance – that has different physical realizers cannot be identified with any of those realizers. This sounds like old-fashioned dualism. But realized properties are realized physically. In this regard they are shaped by, and dependent on, physical goings-on. Functionalists focus on structure. What matters to a mind is not the medium in which it is embodied (flesh and blood, silicon and metal, ectoplasm), but its organization. Thus construed, functionalism is sometimes

traced to Aristotle, who, at times, seemed to be thinking along these lines (De Anima Book II, 1–3). One difficulty for any such view is that it seems possible to imagine systems that preserve the same patterns of internal relations as minds, but are not minds. Ned Block (1978) imagines the population of China organized in the way an intelligent system might be organized. Although the Chinese nation is a functional duplicate of a conscious agent, it is hard to think that the nation, as opposed to the individuals who make it up, constitutes a conscious mind. The functionalist picture is one of “higherlevel” mental properties realized by, but distinct from “lower-level” physical realizers. The result is “non-reductive physicalism”: minds and their properties are grounded in the physical world, but not reducible to their physical grounds. A similar picture has been inspired by Donald Davidson’s “anomalous monism”. Davidson (1970) describes the mental as “supervening” on the physical. Davidson borrows the notion of supervenience from R.M. Hare, who had borrowed it from G.E. Moore. Both Hare and Moore were concerned with issues in ethics. Both, though for different reasons, held that, although moral assertions could not be translated into non-moral, “natural” assertions, moral differences required nonmoral differences. If St. Frances is good, an agent indistinguishable from St. Frances in relevant non-moral respects – a “molecular duplicate” of St. Frances – must be good as well. Davidson applied this idea to the relation between mental and physical descriptions: agents alike physically (agents answering to all the same physical descriptions) must be alike mentally (must answer to the very same mental descriptions). Reduction fails – in both ethics and psychology – because agents could be alike morally or mentally, yet differ physically. Supervenience fits nicely with multiple realization, so nicely that some philosophers began to think of supervenience as providing an account of the realizing relation. Considerable effort was expended on refining the supervenience concept. The result was a proliferation of kinds and grades of supervenience and much discussion as to which 41

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 physical properties (Kim, 1990). Supervenience, however, is a purely formal, modal notion. If you know that the As supervene on the Bs (moral truths supervene on natural truths, mental truths supervene on physical truths), you know that the Bs in some fashion necessitate the As. But what is responsible for this necessitation? What is it about the Bs that necessitates the As? There are a number of possibilities: (1) the As are the Bs; (2) the As are made up of the Bs; (3) the Bs include the As as parts; (4) the As are caused by the Bs; (5) the As and the Bs have a common cause. None of these fit what proponents of supervenience or multiple realizability appear to have in mind, however. Sydney Shoemaker (1980) has suggested that “causal powers” “bestowed” by mental properties are a subset of powers “bestowed” by a variety of physical realizing properties. When one of these physical properties is on the scene, the mental property is thereby on the scene, option (3) above. Derk Pereboom (2002), invoking the idea that a statue, although “constituted by” a particular lump of bronze, is not identical with the lump, argues that instances of mental properties are wholly constituted by, but not identifiable with their physical realizers, option (2). These accounts of the realization relation locate mental properties within the physical causal nexus. It is hard to see, however, how any such account could preserve the thought that mental properties are really distinct from their realizers while mingling their causal powers with powers of the realizers. Powers comprising a subset of a thing’s physical powers would seem to be physical powers; and powers of a statue are hard to distinguish from powers of the bronze that “constitutes” the statue. Non-reductive physicalism has proved popular because it promises to preserve the distinctiveness and autonomy of the mental, while anchoring it firmly in the physical world. However, non-reductive physicalism has come under fire from Jaegwon Kim (2005) and others for failing adequately to accommodate mental causation, the centerpiece of the mind–body problem. If 42

mental properties are distinct, higher-level properties, how are they supposed to figure in causal relations involving lower-level physical goings-on? So long as we embrace closure, it appears that physical events – bodily motions, for instance – must have wholly physical causes. The prospect of mental properties making a causal difference in the physical world is evidently inconsistent with mental properties’ being irreducible to physical properties and the physical world’s being causally closed. We must choose, it seems, between epiphenomenalism – mental properties, although real, are physically impotent – and systematic overdetermination – some events have mental causes as well as physically sufficient causes. Kim argues that over-determination is a false option. We thus face a choice between epiphenomenalism, on the one hand, and, on the other hand, the abandonment of the non-reductivist hypothesis. Mental properties are either reducible to physical properties or epiphenomenal. Perhaps, Kim suggests, most mental properties are reducible. Those that are not, qualitative properties of conscious experiences, for instance, the qualia, must be epiphenomenal: real, but causally impotent. This is close to the line advanced by David Chalmers (1996) in a ringing defense of the irreducible nature of qualia. Chalmers divides mental attributes into those characterizable in “information processing” terms and those that are essentially conscious. The former “logically supervene” on fundamental physical features of organisms: a system with the right sort of functional organization will be intelligent and, in general, psychologically explicable. consciousness, on the other hand, although determined by the physical facts, is not reducible. To facilitate the distinction he has in mind, Chalmers imagines zombies, creatures resembling us but altogether lacking in conscious experiences (Kirk, 1974). Such creatures are impossible “in our world”, that is, given actual laws of nature. The conceivability of zombies, however, suggests that laws governing the production of conscious qualities are fundamental in 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. Think of such laws as analogous to Euclidian axioms. Laws governing consciousness resemble the parallel postulate in being independent of the rest. Their presence or absence has no effect on physical goings-on. Outwardly, a zombie world is indistinguishable from ours. Both Kim and Chalmers render conscious qualities – qualia – epiphenomenal, perfectly real, but physically irrelevant. The result is what Kim calls “modest physicalism” – physicalism plus a “mental residue” – a conception reminiscent of Descartes’s idea that much human behavior is explicable on mechanical principles alone. The difference is that, whereas Descartes embraced interactionism – mental properties are causally potent – Kim and Chalmers regard consciousness as qualitatively remarkable but causally inert. Other philosophers with physicalist leanings are not so ready to throw in the towel. What exactly are mental qualities, the socalled qualia? Describe a dramatic sensory scene: a sunset viewed from a tropical beach. Your description will invoke a panoply of vivid qualities: colors, odors, sounds. Were we to look inside your head, however, we would observe none of this. Colin McGinn asks “how Technicolor phenomenology could arise from grey soggy matter” (1989, p. 349). As C.D. Broad reminds us, properties of brains seem utterly different from properties of our conscious experiences. Let us suppose, for the sake of argument, that whenever it is true to say that I have a sensation of a red patch it is also true to say that a molecular movement of a certain specific kind is going on in a certain part of my brain. There is one sense in which it is plainly nonsensical to attempt to reduce the one to the other. There is something which has the characteristic of being an awareness of a red patch. There is something which has the characteristic of being a molecular movement. It would surely be obvious even to the most “advanced thinker” who ever worked in a physiological laboratory that, whether these “somethings” are the same or different, there are two different characteristics (Broad, 1925, p. 622).

Suppose, however, we distinguish properties of things experienced from properties of experiences. The sunset is red, the breeze balmy, the sand warm, and the waves murmur softly. Colors sounds, odors, and the like are not properties of our experiences of such things, but properties of things we experience, or at any rate properties we represent such things as possessing. The point was made by J.J.C. Smart (1959) in his original discussion of mind–brain identity, and, more recently, others have sought to demystify qualia by arguing that what have been regarded as irreducible qualities of conscious experiences are, in reality, only qualities we represent things as having (Harman, 1990 and Lycan, 1996). Were that so, there would be no insurmountable gulf between mental properties, including properties of conscious experiences, and unexceptional physical properties. Much of the mystery of consciousness might be due to confusion over what experiential properties could be (see experience). Here we have representation playing the garbage-bin role: embarrassing or inconvenient features of the world are consigned to representations of the world. Still, it is difficult to shake the idea that representings are themselves permeated with irreducibly mental qualities. Your being in pain might involve your representing a bodily state as painful, but this representing is, or certainly seems to be, qualitatively loaded. What we might hope to learn from all this? The mind–body problem takes hold only when we respect the integrity of both the physical and the mental. More often than not this has meant accommodating the mental to the physical, thereby privileging the physical. The ideal solution would involve finding a niche for the mental within the physical realm, but that seems hopeless, no more promising than reduction or elimination. Perhaps we are deluding ourselves. Perhaps we have erred in letting Descartes set the agenda and assuming at the outset that the mental and the physical are mutually exclusive. Suppose, instead, it turned out that the mental/physical distinction were not metaphysically deep. In that case, we would have no mystery as 43

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) properties could have physical (in the sense of non-mental) causes or effects. Consider Davidson’s “anomalous monism”. Davidson is commonly read as holding that mental properties depend on, but are not reducible to physical properties. A mental event is an event with a mental property; a physical event is an event with a physical property. This leaves open the possibility of “token identity” without “type identity”: one and the same event could be both mental and physical by virtue of possessing a mental property and a (distinct) physical property. The problem of mental causation arises because we think that events have the effects they have solely in virtue of their physical properties. Mental properties “piggyback” on physical properties, but appear causally inefficacious. Although this picture is widely attributed to Davidson, it is pretty clearly not what Davidson has in mind. Davidson speaks of descriptions and predicates, not properties. An event is mental, he holds, if it answers to (“satisfies”) a mental description; it is physical if it satisfies a physical predicate. One and the same event, including the event’s causally efficacious constituent properties, could answer to both a mental and a physical description. For Davidson, the mental– physical distinction is classificatory, not metaphysical. Everything in the world could be given a physical description and so counts as physical. Some portions of the world could also be described using mental terms. truthmakers for applications of mental predicates will be fully describable using a physical vocabulary. This is so despite the fact that, owing to very different application conditions, there is no prospect of analyzing mental predicates in physical terms. A view of this kind treats “mental” and “physical” as classificatory designations, not fundamental metaphysical categories. In this regard it resembles Spinoza’s “neutral monism”. Spinoza (1632–1677) held that there is but a single substance possessing multiple “attributes”, including the mental and the physical. Finite physical or mental entities are modes of these attributes, ways 44

of being mental or physical. Spinoza’s attributes differ from Descartes’s, however, in being attributes of a single substance and in being, at a deeper level, unified. In singling out attributes, we are “abstracting” in Locke’s sense, engaging in “partial consideration” of a substance. Abstraction is a mental act, but what is abstracted is in no way mind-dependent. These are deep metaphysical waters, but the mind–body problem cries out for a deep solution. Perhaps it is time to abandon the Cartesian presumption that the mental and the physical differ in a fundamental way, along with all the many attempts at reconciliation beholden to the Cartesian presumption. As noted, such attempts have tended to privilege the physical. The mental is seen as reducible to or dependent on the physical in some way. For Davidson and Spinoza, the physical is in no regard privileged. We have one world, variously propertied, describable in various ways, with various degrees of specificity. To imagine that dramatic differences in our modes of classification must reflect fundamental metaphysical discontinuities is to mistake features of our representations of the world for features of the world. Or so Spinoza and Davidson think. Whether a move to monism represents progress or merely one more philosophical byway leading nowhere remains to be seen. Meanwhile, philosophers will continue to till familiar soil in familiar ways in hopes of bringing forth some new solution to the mind–body problem. bi bliography Block, N.J.: “Troubles with Functionalism,” in C.W. Savage, ed., Perception and Cognition: Issues in the Foundations of Psychology, Minnesota Studies in the Philosophy of Science 9 (Minneapolis: University of Minnesota Press, 1978), 261–325. Broad, C.D.: The Mind and Its Place in Nature (Paterson, NJ: Littlefield Adams, 1960; originally published 1925). Chalmers, D.: The Conscious Mind: In Search of a Fundamental Theory (New York: 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 in the History of Rationalist Thought (New York: Harper and Row, 1966). Churchland, P.S.: “Eliminative Materialism and the Propositional Attitudes,” Journal of Philosophy 78 (1981), 67–90. Dreyfus, H.L.: What Computers Can’t Do: A Critique of Artificial Reason (New York: Harper and Row, 1972). Feyerabend, P.K. and Maxwell, G., ed.: Mind, Matter and Method: Essays in Philosophy and Science in Honor of Herbert Feigl (Minneapolis: University of Minnesota Press, 1966). Fodor, J.A.: Psychological Explanation: An Introduction to the Philosophy of Psychology (New York: Random House, 1968). Foster, J.: The Case for Idealism (London: Routledge and Kegan Paul, 1982). Grossmann, R.: The Categorial Structure of the World (Bloomington: Indiana University Press, 1983). Harman, G.: “The Intrinsic Quality of Experience,” Philosophical Perspectives 4 (1990), 31–52. Kim, J.: Physicalism, or Something Near Enough (Princeton, NJ: Princeton University Press, 2005). Kim, J.: “Supervenience as a Philosophical Concept,” Metaphilosophy 12 (1990), 1– 27. Kirk, R.: “Zombies versus Materialists,” Proceedings of the Aristotelian Society, suppl. vol. 48 (1974), 135–52. Lowe, E.J.: Subjects of Experience (Cambridge: Cambridge University Press, 1996). Lycan, W.G.: Consciousness and Experience (Cambridge, MA: MIT Press, 1996.) McGinn, C.: “Can We Solve the Mind–Body Problem?,” Mind 98 (1989), 349–66. Matson, W.I.: “Why Isn’t the Mind–Body Problem Ancient?,” in Feyerabend and Maxwell (1966), 92–102. Nagel, T.: “What Is it Like To Be a Bat?,” Philosophical Review 83 (1974), 435–50. Nichols, S. and Grantham, T.: “Adaptive Complexity and Phenomenal Consciousness,” Philosophy of Science 67 (2000), 648–70. Pereboom, D.: “Robust Nonreductive Materialism,” Journal of Philosophy 99 (2002), 499–531.

Place, U.T.: “Is Consciousness a Brain Process?,” The British Journal of Psychology 47 (1956), 44–50. Ryle, G.: The Concept of Mind (London: Hutchinson, 1949). Searle, J.R.: “Minds, Brains, and Programs,” Behavioral and Brain Sciences 3 (1980), 417–24. Shoemaker, S.: “Causality and Properties,” in Peter van Inwagen, ed., Time and Cause (Dordrecht: Reidel, 1980), 109– 35. Skinner, B.F.: “Behaviorism at Fifty,” Science 140 (1963), 951–8. Smart, J.J.C.: “Sensations and Brain Processes,” Philosophical Review 68 (1959), 141–56. Stich, S.P.: Deconstructing the Mind (New York: Oxford University Press, 1996). Turing, A.M.: “Computing Machinery and Intelligence,” Mind 59 (1950), 434– 60. Wittgenstein, L.: Philosophical Investigations, trans. G.E.M. Anscombe (Oxford: Basil Blackwell, 1968; originally published 1953). john heil

Modality and Possible Worlds Propositions are evaluated not only as true or false, but as necessarily or contingently true or false. That seven plus five equals 12 is necessary; that George W. Bush was the President of the United States in 2008 is contingently true, and that Saul Kripke has seven sons is merely possible. What sort of fact makes it true that these propositions have the modal status that they have? The problem is sometimes put in epistemological terms: empiricists, for example, ask how experience could give us reason to believe that a proposition is not just true, but necessary. But the real problem behind this question is not epistemological, and not dependent on any thesis about the sources of our knowledge. Even if an oracle gave us unlimited access to matters of fact about the world, we would still face the question, what could make it the case that some fact was not just true, but had to be true? 45

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 this problem, modal propositions are made true by relations of ideas or linguistic conventions: not by the way the world is, but by the way we conceive or describe it. But (on this view) what is necessary is not that we conceive or describe the world as we do. If it is necessary that all uncles are male, it is not because it is necessary that we should have adopted certain conventions to use the worlds “uncle” and “male” in certain ways. What is said to be a matter of convention – that a certain sentence be used to say something that is true no matter what the facts are – is different from what is said to be necessary, which is the proposition itself that this sentence is conventionally used to express. So how can linguistic conventions, or facts about the way we conceive of things, explain necessity and contingency? In any case, it is hard to see how some statements widely thought to be necessary could be true by convention. How could the way we talk or think make it true and necessary that something (a number, or God, for example) should exist, or that a particular thing (Hillary Clinton, say) should be a member of a particular kind (human being)? The way we have put the problem is already contentious, since it assumes that the things that are said to be true or false, and necessary or contingent, are propositions (see proposition, state of affairs). An adequate theory of modality must give some account of what propositions are, or of whatever the bearers of truth and necessity are taken to be. One way to begin that is motivated by the empiricist’s idea that necessity has its source in relations of ideas or in the meanings of words is with a predicate, not of propositions, but of the sentences of some language (see empiricism). Paradigms of necessary truths, according to this approach, are statements that are logical truths, or truths in virtue of meaning. The significance of this alternative starting point can be illustrated by looking at W.V. Quine’s criticisms of modal logic which began with the assumption that our most basic modal concepts are applied to linguistic expressions, rather than to what they express. 46

Quine distinguished three grades of modal involvement (Quine, 1953). (He was skeptical even of the first, but saw them as increasingly problematic.) The first grade was a necessity predicate of sentences: being logically true, or perhaps being analytic. The second grade was a move from a predicate of sentences to an operator on sentences – from “uncles are male is necessary” to “necessarily, uncles are male”. Quine argued that the move involved a use-mention confusion, since operators are to be interpreted in terms of functions whose arguments are the values of expressions, and not the expressions themselves. To stipulate that a sentence of the form “necessarily p” shall be true whenever the sentence that is in the place of “p” satisfies the necessity predicate constrains the interpretation of the operator, but does not determine it. Quine argued that the move from the first to the second grade of modal involvement, while based on a use-mention mistake, was in itself relatively harmless, until one made the further move to the third grade, which was to allow the operator to operate on open as well as closed sentences – that is, to allow quantification into modal contexts. The first move disguised the fact that modal contexts were really quotational, and so that quantification into modal context was, implicitly, quantification into a quotation. One could repair the damage and avoid incoherence, he argued, only by making metaphysical commitments that he and the empiricist developers of modal logic that he was criticizing would agree are unacceptable. It is true that modern modal logic began (with C.I. Lewis) as a project of analyzing logical necessity, and deducibility, so Quine’s analysis is appropriate as an ad hominem argument against his intended targets (C.I. Lewis and Rudolf Carnap). But modal concepts in general have much wider application. We may be concerned with what must or might happen, in various senses, and with what would or might have happened under various conditions that did not, in fact, obtain, with the dependence and independence of facts on other facts, and these concerns arise in our attempts to understand 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 modal concepts more generally, it seems appropriate to begin with something like facts, states of affairs, or propositions as the things to which modal predicates are applied. If we begin with a predicate of propositions, rather than sentences, then Quine’s three grades of modal involvement look quite different. Suppose we assume, about propositions, only that if we ascribe a well-defined predicate to a determinate entity that is within the range of the predicate, we will have expressed a proposition. Then the move from the first to the second grade of modal involvement looks unproblematic: the proposition expressed by a sentence of the form “necessarily p” will be as well defined as the predicate of propositions with which we began. And the move to the third grade – to an operator on open sentences into which one may quantify – looks unproblematic as well, for the following reason: if an operator on propositions is well-defined, then so is a corresponding operator on propositional functions (functions from individuals to propositions). Suppose the necessity operator, “” is interpreted with a function that takes (for example) the proposition that Socrates is human to the proposition that it is necessary that Socrates is human. Suppose that the open sentence “x is human” expresses a function from individuals to propositions. Then “it is necessary that x is human” will express the propositional function whose value, for any individual a, is the proposition that the proposition which is the value of “x is human” for argument a is necessary. But even if the move through the grades of involvement is unproblematic, given the assumption that what we start with is a predicate of propositions, a clear account of modality still need an account of propositions (about which Quine was famously skeptical). There are many conceptions of proposition, and lots of controversies about how this notion is best understood, but fortunately we can go some way toward an account of modality while making only minimal assumptions about exactly what propositions are. Whatever propositions are, all who are willing to talk at all about such things will

agree that they have truth conditions, and that their truth conditions are essential to them. Any theory of propositions will say that the class of propositions determines a structure that can be characterized by some familiar interdefinable relations: entailment, incompatibility, consistency, etc. If we start with a notion of consistency or compatibility, as a property of sets of propositions, we can define the other relations that are required in terms of it. We assume that consistency will satisfy the following property: if a set of propositions is consistent, then so is any subset of it. It will be assumed, in a minimal theory of propositions, that every proposition has a contradictory, where the notion of a contradictory is definable in terms of the consistency relation as follows: proposition x is a contradictory of proposition y, if and only if, first, the set {x,y} is inconsistent, and second, any consistent set of propositions is either consistent with x, or consistent with y. A set of propositions Γ entails a proposition x if and only if the set Γ∪{y} is inconsistent, where y is a contradictory of x. Two propositions will be equivalent if and only if they are mutually entailing. A minimal theory might identify equivalent propositions. Even if finer distinctions between propositions are required for some purposes, we can go some way toward a theory of modality while ignoring such distinctions. It is clear from the requirements of a minimal theory of propositions that the most basic modal properties are not something added onto a minimal theory of propositions, but are constitutive of it. Intuitively, the consistency of a set is the possibility that the members of the set all be true together, and a necessary truth is a proposition that is entailed by every set. This is possibility in the widest sense; more restrictive notions of possibility and other modal properties and relations might be defined with additions to the basic structure. Necessity, according to a familiar slogan going back at least to Leibniz, is truth in all possible worlds, and the notion of a possible world has played a prominent role in contemporary treatments of modality, both in formal semantic models, and in the informal 47

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. (See Kripke, 1963 for an exposition of the model theory, and Kripke, 1980 for an influential treatments of philosophical problems in metaphysics and the philosophy of language that uses the possible worlds framework.) The notion is a controversial one, and there are substantive disagreements about how it should be understood, and about whether any notion of possible world should play a central role in an account of modality. But at least a minimal concept of possible world can be defined within the basic minimal theory of proposition. Within that theory, we can define maximal consistent classes of propositions: classes that are consistent, and that for every proposition contain either that proposition or its contradictory. One might identify a possible (state of the) world with these maximal sets of propositions. Or in an alternative formulation, one might take a set of possible worlds as the primitive basis of one’s theory, and define the propositions as sets of them. Whichever primitive notion one begins with, there will be, in a minimal theory, a one-one correspondence between sets of possible worlds and coarse-grained propositions. (See Adams, 1974 for an analysis of possible worlds in terms of propositions, and Stalnaker, 2003, ch. 1, for a discussion of the relation between propositions and possible worlds.) The point of spelling out this minimal theory of propositions, possible worlds and basic modal properties and relations is to set up a framework in which the substantive metaphysical questions about modality can be sharpened and clarified. We will consider questions about the nature of possible worlds and their role in a metaphysical account of modality, but it is useful first to see the minimal framework as an attempt to provides only a paraphrase of problematic modal claims in a language in which ambiguities and equivocations are more easily avoided, and in which the structure of modal claims and questions are more perspicuously displayed. The thesis that necessity is truth in all possible worlds is like Quine’s thesis that to be is to be the value of a bound variable. The Quinean thesis is not a substantive 48

claim about ontology, but an attempt to get clearer about what such claims come to. I think the thesis that necessity is truth in all possible worlds should be understood in a similar spirit. The paraphrase of modal claims and questions into the language of possible worlds solves some of the more superficial puzzles about referential opacity and merely possible individuals by diagnosing scope ambiguities and by separating questions about names and words from questions about the individuals, kinds and properties that the names and words are used to designate. And it brings to the surface and gives new form to the underlying metaphysical questions about the nature of modality. In the context of this simple framework, I will consider a number of interrelated metaphysical problems about modality. First, if possible worlds are to be taken as basic entities in our ontology, what kind of thing are they? What is it that makes it true that there are the possible worlds that there are? Different philosophers who take possible worlds to be fundamental to an explanation of modality give radically different answers to these questions. David Lewis argued that we should take other possible worlds literally as concrete particular universes, spatio-temporally disconnected from our own (Lewis, 1986). Most other philosophers who take possible worlds seriously explain them as possible states of the world, ways the world might be. (See Kripke, 1980; Plantinga, 2003; Stalnaker, 2003 for actualist accounts. See Divers, 2002 for a survey of a range of accounts of possible worlds.) These contrasting answers take on different explanatory burdens and give different response to various more specific problems about modality. Second, can we give an account of modality that is reductive in some sense, and if so, what is being reduced to what? David Lewis argued that his realist analysis of modality in terms of possible worlds was a reduction of modal to non-modal notions, but others have disputed this. Alternatively, one might try to reduce the notion of a possible world to something more basic. Is a reduction of modal 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 succeeds, in its own terms, in giving one? Third, might there have been things that do not in fact exist? If so, what can be said about what is merely possible? What ontological commitments are required to make sense of the possibility of things that do not actually exist? The Lewisian modal realist has no problem here (at least no new problem), but the actualist needs either to explain what we are really talking about when we seem to be talking about things that might, but don’t exist, or else to reject the thesis that there might have been things other than those there are. Fourth, whether or not there might have been things that do not actually exist, it seems obvious that the things there are might have been different in various ways from the ways they in fact are. Does this imply that the same things exist in many possible worlds? Is there a problem about the identification of individuals across possible worlds, and if so what is it? There are different theoretical accounts of the relations between the individuals that exist in different possible worlds, and of the relations between particular things and the properties and relations that they exemplify. 1. Modal realism vs. actualism. The basic contrast between possibilist, or modal realist accounts of modality on the one hand and actualist accounts on the other is central to many of the more specific issues in the metaphysics of modality. According to the modal realist, there are literally many universes, individuated by the spatial and temporal relations between things in them. Two things count as worldmates – denizens of the same possible world – if and only if they are spatio-temporal relations between them. But for the actualist, everything that is real is actually real. Possible worlds are possible ways that a world might have been. The difference between the two kinds of theory comes out in the contrasting answers that they give to the following general challenge to the coherence of the idea of a merely possible world: A merely possible world is a world that is not actual, which is to say a world that does not exist. But the possible worlds analysis of modality is committed to the existence of

merely possible worlds, which seems to mean that it is committed to the existence of things that do not exist. Any response to this challenge that seeks to defend the coherence of the account must distinguish a sense in which merely possible worlds exist from a sense in which they do not, and there are two very different strategies for making this distinction. The modal realist answers the question by distinguishing two different ranges for the quantifier – one unrestricted and one restricted. When we talk about absolutely everything that exists, we include a plurality of possible worlds (as well as merely possible donkeys, people and things). But we most often use the quantifiers so that they range over a restricted domain: “everyone” might, for example, mean all the people invited to the party. Even when we are making very general claims, we are often (according to the modal realist response to this challenge) restricting our quantifiers to things in our vicinity, broadly construed. Our vicinity, on this construal, includes that part of reality that is spatio-temporally connected with us. In this broad but restricted sense, there is only one possible universe that exists: the one we are in. But the other universes, like the actual people who were not invited to the party, are equally real. For the actualist, the distinction is of a different kind. According to this theory, the only things that exists, in the most absolute and unrestricted sense, are actual things. The relevant distinction is not in the range of the quantifier, but in the kind of thing that one is talking about. Possible worlds, properly construed as things that there are many of, are more accurately labeled “possible states of the world”, and states are the kind of thing that may be instantiated or exemplified. (For a given state, there may or may not be something that is in that state). We might distinguish a notion of “possible world” meaning a thing that exemplifies a given possible state of the world from a notion of “possible world” as the state itself – something that is perhaps not exemplified. Using “possible world” in the first sense, there is only one of them (within the domain of absolutely everything), while using it in 49

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) many, only one of which is exemplified. The modal realist doctrine can be separated into two theses, one metaphysical and one semantic. The metaphysical thesis is that there is a rich plurality of spatio-temporally disconnected universes, rich enough to obey certain principles of recombination. (Roughly, for any two things in different universes, there will be a universe that contains intrinsic duplicates of both.) The semantic thesis is that statements about what is necessary and possible are properly interpreted by quantifiers that range over these universes (A sentence of the form “Possibly P” is true if and only if P is true in one of these universes.) Judged separately, both theses seem highly implausible. What reason do we have to believe in this extravagant ontology? And even if we did, what does it have to do with what is necessary and possible in the actual world? But while Lewis granted the prima facie implausibility of his doctrine (he took what he called “the incredulous stare” to be the most serious challenge to his metaphysical view), he argued that the two parts of the doctrine must be judged together, and that together they provide an indispensable foundation for a rich family of modal concepts. Despite its initial implausibility, the fruitfulness of the doctrine that provides this foundation is sufficient reason to believe that it is true. Crucial to this defense of modal realism is the thesis that the rich family of modal concepts cannot rest on a more modest foundation. To use Lewis’s rhetoric, the claim is that we cannot have the “paradise” that this family of concepts brings “on the cheap.” In this context, Lewis criticizes several versions of the actualist alternative to modal realism, arguing that none of them is up to the job. All of the actualist accounts that Lewis considers assume that “possible worlds” must be representations of a world: either linguistic representations, something like scale models, or perhaps just simple and primitive representations. I think Lewis is right that a notion of possible world as representation cannot provide an adequate foundation for our modal concepts, but there are other alternatives that Lewis does not consider. 50

Possible states of the world, or ways a world might be, are not representations of a world, but properties that a world might have. While properties allow for the distinction between existing and being exemplified that the actualist needs to distinguish the sense in which merely possible worlds exist from the sense in which they do not, properties differ crucially from representations in the following way: representations, whether pictorial, linguistic, mental, or of some other form, face a problem of intentionality; it makes sense to ask, of a representation, what is it that explains why the representation has the representational content that it has? There is no analogous question about properties. One cannot intelligibly ask, of a property, what makes it that particular property, rather than some other one? I think Lewis’s critique trades on the fact that the actualists do not have an answer to a question like this about the things they are calling “possible worlds”. In the context of Lewis’s overall metaphysical picture, the thesis that possible states of the world are a kind of property does not provide an alternative to modal realism, since on Lewis’s account, properties are classes, individuated by their extensions, and so a total way a world might be will be a unit set, with the world that is that way as its member (see class, collection, set). On this account of properties, there will be many possible total states of the world only if there are many things that are in those states. It is an irony of Lewis’s modal realism that the metaphysically extravagant doctrine is grounded in Quinean ontological austerity – a rejection of any notion of property or attribute that cannot be identified with its extension. The actualist is committed to a more robust notion of property, and so needs an explanation of what properties are. 2. Reduction. Possible worlds, construed as concrete universes, are the fundamental primitive elements of the modal realist theory, and are clearly prior, in the order of explanation, to propositions, which are identified with sets of possible worlds (see concrete/abstract). Necessity and possibility 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 relations between propositions. Is Lewis right to claim that this theory provides a reductive account of modality – an explanation of the modal in terms of the non-modal? This is a delicate question, since it is debatable what basic concepts count as modal, but I think Lewis’s claim is a reasonable one. The reason is that the metaphysical component of the theory (the hypothesis of a plurality of parallel universes) is intelligible independently of the semantic analysis of modal concepts in terms of it. The parallel universes are individuated by spatio-temporal relations, and if it is fair to claim that the notion of a spatio-temporal relation is a non-modal notion, than he theory seems to offer a metaphysical characterization of the structure of reality in terms of concepts that modal skeptics should be willing to accept (even if they reject the substantive metaphysical claims made with those concepts). So I would concede Lewis’s claim that he offers a reduction, but maintain that it is debatable whether this is a cost or a benefit of the overall account. It is not just the metaphysical commitments of the theory that elicit the incredulous stare; the semantic analysis of modal notions in terms of it also seems implausible, since it defines modal concepts in terms of things that, intuitively, seem to have nothing to do with modality, even if one were to accept the metaphysics. The intuitive resistance to the semantic component of the doctrine may derive from the judgment that modal notions are fundamental, and not properly reduced to something more basic. Compare the way one might react to a project of giving a reductive analysis of truth to something more basic (warranted assertability, perhaps, or what will be believed at the end of inquiry). Even if such a story could be spelled out in noncircular terms, one might judge that the analysis mistakenly categorizes a substantive claim as a definition. Even if it were correct that, at the end of inquiry all and only truths would be believed, this would not give us an account of what truth is (see theories of truth). Similarly, I think it is reasonable to think that even if a principle of plenitude were true, so that everything that

might happen does happen, somewhere and sometime, perhaps in a parallel universe, it would still be wrong to say that this is what possibility consists in. I also agree with Lewis that no actualist attempt to explain possible worlds in nonmodal terms (for example, as linguistic representations) can succeed. But most versions of modal actualism are not attempts to explain the modal in terms of the non-modal, since the basic notions of this kind of theory – whether they are propositions or total ways a world might be – are characterized in terms that presuppose modal notions. In fact, I think the notion of a property, which is used to say what kind of thing a possible state of the world is, is itself a modal notion: one grasps what property one is talking about to the extent that one has a sense for what it would be for that property to be exemplified, which is to understand a certain possibility. 3. Merely possible things. It seems at least prima facie reasonable to believe that there might have existed things that do not in fact exist. For example, Saul Kripke might have had seven sons, and if he had, then seven people who do not in fact exist would have existed (assuming that Saul Kripke actually has no sons). In the possible worlds framework, the general thesis is modeled by the claim that the domains of some possible worlds contain individuals that are not in the domain of the actual world. The modal realist has no problem with this thesis, since the actual world is just one place among others. Non-actual things are just things that are located in one of the other places. But for the actualist, the domain of the actual world includes everything that exists at all, in any sense, so it seems that actualism is at least prima facie committed to the thesis that everything that might exist does exist. This is the most serious challenge to the actualist conception. I will describe four strategies that different actualists use to respond to it: The first actualist response begins by noting that to understand talk of possible individuals, we need a distinction that parallels the distinction between two senses of the term “possible world”: just as actualists 51

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 a world that is that way, so they must distinguish individuating properties that an individual might have from the individuals that has those properties. While actualists are committed to the thesis that there are no things that might exist, but do not, they can allow that there are (and necessarily are) properties that are necessary and sufficient to determine a unique individual, but that are in fact uninstantiated. More precisely, the view is that there are properties X that meet the following condition: it is necessary that if there exists something that instantiates X, then that thing is necessarily identical to anything that instantiates X. The domains of the different possible worlds are to be understood, according to this response, not literally as domains of individuals, but as domains of properties of this kind – individual essences, or haecceities (see haeccity). Alvin Plantinga, who develops and defends this response, calls the domains “essential domains” (Plantinga, 2003). The basic structure of the orthodox Kripke semantics for quantified modal logic, with variable domains, is unchanged by this move; the difference is in the interpretation of the formal models. This response to the problem is simple, formally conservative, and successful, on its own terms, in reconciling actualism with intuitions about what might have been true. But it requires what some regard as a metaphysical extravagance: a belief in a special kind of property that carries with it the particularity of an individual, but that is also conceptually separable from the individual. We may have no problem understanding the property of being identical to Socrates, but one might reasonably think that this is an object-dependent property – a property that would exist only if Socrates did. But the haecceitist response to the problem holds that while we use the person Socrates to fix the reference of the property of being identical to Socrates (or a property that is necessarily equivalent to it), the property itself would exist even if he did not. And furthermore, there actually exist, on this account, properties of this kind that would be instantiated by Saul Kripke’s seven sons in the possible worlds in which 52

he had seven sons – properties whose reference could be fixed, in such a world, with a predicate of the form “being identical to this individual”, where “this individual” refers to one of the seven sons. Plantinga grants that we may not have the resources, even in principle, to refer to particular uninstantiated haecceities, but we can talk about them in general terms, and that, he argues, is good enough. The second response (defended in Williamson, 2002 and in Linsky and Zalta, 1994) is to reject the intuition that gives rise to the problem – that there might have existed things that do not in fact exist (as well as the intuition that there are some things that exist only contingently). This response avoids the problem, and as its defenders emphasize, it also allows for a modal logic that is much simpler than what is required when the domains vary from possible world to possible world. But of course it takes on the burden of explaining the divergence between the theory and conflicting intuitions about modal truths that seem compelling. How can it be made plausible that apparently temporary and contingent beings such as ourselves exist eternally and necessarily? How can we accept that there actually are things that might have been Saul Kripke’s seven sons? The defenders of this strategy respond to the challenge by acknowledging that people and ordinary physical objects are only temporarily and contingently concrete things, with a spatio-temporal location. In possible worlds and at times when one is inclined to say that the people and things do not exist, we should instead say that they exist, but lack the features that we are inclined to think are essential to being a person or a physical object. They are in no place, at those worlds and times, and are neither concrete things, nor abstract objects, but particular things that have the potentiality to be concrete things. This may seem a gratuitously extravagant metaphysics, but Williamson argues that it is entailed by principles that it is difficult to reject. The most controversial of the premises of Williamson’s argument is the thesis that singular propositions (and identity 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 the things they are about. As we have seen, Plantinga rejects this thesis (which he labels “existentialism”), but I suggested that this is a serious cost of his account. But as Williamson shows, if we accept it, and also accept that a proposition is true only if it exists, then we must conclude that the singular proposition that Socrates does not exist could not be true, and this seems to imply that Socrates must exist. One might try to avoid the uncomfortable choice between Plantinga’s haecceities and Williamson’s objects of pure potentiality by rejecting a presupposition that both positions apparently share. The need either for primitive individual essences or for object dependent propositions would be avoided if we were able to reduce individuals to their properties. So a third actualist response to our problem is to adopt some kind of bundle theory of individuals. If this kind of account of individuals were defensible, then we could characterize possibilia in terms of ordinary universal properties and relations, rather than in terms of primitive haecceities, and no propositions would be dependent on particular individuals. But this kind of metaphysical doctrine has a problem accounting for the potentialities and counterfactual properties of particular individuals; we will say more about this problem below. There is a fourth response that accepts the irreducibility of individuals to their properties and relations, and the objectdependence of singular propositions. It takes at face value the intuition that there might have been things other than those there are, and it avoids a commitment to individual essences. I think this is the best response to the problem, though it has its own counterintuitive consequences. The problems for this strategy come from an immediate consequence of the combination of the object-dependence of singular propositions with the contingent existence of individuals: that some propositions themselves are things that exist only contingently. If possible worlds are identified with maximal propositions, or maximal sets of propositions, then possible worlds themselves will be contingent objects. Propositions that are maximal in the sense

that they entail every (actual) proposition or its contradictory may fail to be maximal in another sense: they may entail existential propositions without entailing any singular propositions that witness the existential claim. That is, this response claims that there may be cases where an existential proposition (Such as the proposition that Saul Kripke had a seventh son) is possibly true even though there is no singular proposition (no proposition of the form “x is Saul Kripke’s seventh son”) that is possibly true. I think this is right, but making sense of it requires a more radical reinterpretation of the standard semantic models than do the theories of Plantinga or of Williamson, Zalta and Linsky. And this response must accept the consequence that there are propositions (Such as the proposition that Socrates never existed) which are true with respect to some possible worlds in which the proposition itself does not exist. (Different versions of the fourth response have been defended in Fine (2005), and Adams (1981). 4. Modal properties. Even if we ignore merely possible individuals, there are problems with attributions of modal properties to actual individuals. De re modal claims, claims about what could or could not have been true of some particular things seem especially problematic since it is not clear how they could be true by convention, or by virtue of the relation of ideas. The possible worlds picture seems to offer a straightforward paraphrase of such claims: to say that David Lewis might have been a plumber, but could not have been a fried egg, is to say that there is a possible world in which David Lewis was a plumber, but no possible world in which he was a fried egg. But is the plumber in the other possible world really the same person as our own David Lewis? What is it about him explains his metaphysical incapacity to be a fried egg? Modal realists and actualists answer these questions in different ways. If possible worlds are ways the world might have been then there is no implausibility in accepting the straightforward assumption that Lewis himself inhabits other worlds, since this is only to say that among the ways the world might have been but was not 53

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. Kripke (1980) attempts to demystify counterfactual suppositions about particular individuals, arguing that nothing prevents us from simply stipulating, in specifying the counterfactual situation we are talking about, that it is a situation in which David Lewis is a plumber. But Kripke acknowledged that we might also specify a counterfactual situation in a way that does not explicitly identify a particular individual – in terms of the qualitative characteristics, origin, or constitutive parts of the individual, and that in such a case, we might then ask whether the individual we have specified is or might be some particular actual individual. It remains puzzling exactly what determines the answers to such questions. If possible worlds are understood as other places, as the modal realist understands them, then it is no longer plausible to think that the inhabitants of the actual world will also be found in other possible worlds. The Lewisian modal realist explains modal properties of individuals – their capacities and dispositions, essential and accidental characteristics – in terms of the existence, in other possible worlds, of counterparts of the individual – individuals in other possible worlds who are similar, in relevant respects, to the given individual. According to counterpart theory, David Lewis himself existed only in the actual world, but he might have been a plumber in virtue of the fact that there is a possible worlds in which a person who is like him in certain specific respects was a plumber. Actualists may also use counterpart theory, but for them there is no conflict between the counterpart analysis and the thesis that the plumber in the other possible world really is our own David Lewis. An actualist counterpart theorist may say, as Alvin Plantinga does, that the domains of other possible worlds should be thought of, not as sets of individuals, but as some kind of property that would have been instantiated by an individual, were the state of the world to have been realized. For the haecceitist, the relevant properties are individual essences, but an anti-haecceitist actualist might take the relevant individuating properties to be 54

bundles of qualities, and reduce individual essences to such bundles and counterpart relations between them. An actualist might also adopt a counterpart framework, with a primitive counterpart relation, for methodological reasons: the aim would be simply to provide a framework that is neutral on controversial theses about essential and accidental properties, a framework in which puzzles about identity across times and worlds can be formulated in a perspicuous way. (Actualist counterpart theory is discussed in some papers in Stalnaker, 2003.) Whether one is an actualist or a modal realist, and however one explains the apparent possibility of things that do not in fact exist, and the relation between particular individuals and the properties and relations that they exemplify, and might exemplify, there will remain a general puzzle about the nature and source of modal truth. If necessity is true in all possible worlds, what explains why there are just the possible worlds that there are? Both actualists and modal realists resist the idea that we can explain modal facts as conventional or semantic facts: conventions may determine that our words express certain propositions, but the propositions themselves are necessary or contingent, independently of the words that are used to express them. But actualists and modal realists also agree that to express substantive propositions is to distinguish between the possibilities – to locate the actual world in the space of all possible worlds – and this seems to imply that it is not possible to give a substantive characterization of what is common to all possible worlds. We won’t have a clear grasp of the concept of metaphysical possibility until we see a way to resolve this tension. bibl i ography Adams, R.: “Actualism and Thisness,” Synthese 49 (1981), 3– 41. Adams, R.: “Theories of Actuality,” Noûs 8 (1974), 211–31. Divers, J.: Possible Worlds (London and New York: Routledge, 2002). Fine, K.: Modality and Tense: Philosophical Papers (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, MA: Harvard University Press, 1980). Kripke, S.: “Semantical Considerations on Modal Logic,” Acta Philosophial Fennica 15 (1963), 83–94. Lewis, D.: On the Plurality of Worlds (Oxford: Blackwell, 1986). Linsky, B. and Zalta, E.: “In Defense of the Simplest Quantified Modal Logic,” Philosophical Perspectives 8: Logic and Language, ed. J. Tomberlin (Atascadero, CA: Ridgeview, 1994), 431–58. Plantinga, A.: Essays in the Metaphysics of Modality (Oxford: Oxford University Press, 2003). Stalnaker, R.: Ways a World Might Be: Metaphysical and Anti-metaphysical Essays (Oxford: Oxford University Press, 2003). Williamson, T.: “Necessary Existents,” in A. O’Hear, Logic, Thought, and Language (Cambridge: Cambridge University Press, 2002), 233–51. robert stalnaker

Persistence Introduction Things change. This much looks like a metaphysical and observational datum. By the proposition that things change we typically mean that things survive change – not all changes, but most. In other words, we live in a world in which there is both change and sameness. My car was red; I have given it a coat of green paint; now it is green. One of the car’s qualities has changed, and to that extent the car itself has changed. But we would all accept that it is still the same car. The standard way of putting this philosophically – though not a way we often describe it – is to say that the car has persisted through a change, in this case of color. Yet it is not just change that compels metaphysicians to wonder about persistence. It may be that, as Aristotle held, time is the measure of change, and so without change there could be no time and hence no persistence – since persistence occurs in time (extra-temporal existence, such as God’s, would not on this view be a kind of

persistence) – yet things persist even when undergoing no macroscopic change such as that of color. What is it for a material object to persist pure and simple? Is this a misconceived question because persistence is too basic a phenomenon to yield to analysis? Or can the metaphysician say something informative about what it involves? Is there something inherently strange, or even paradoxical, about the concept of persistence such that we ought to deny that anything really persists? Or can we retain the idea of persistence and instead deny that anything changes since it is change, rather than persistence itself, that raises insoluble problems? Spatio-temporal continuity The standard approach to analyzing persistence is in terms of spatio-temporal continuity (Coburn, 1971; Swinburne, 1968/1981). The idea is that an object F persists through a temporal interval if and only if it traces a spatio-temporally continuous path through that interval. Tracing a spatio-temporally continuous path is then defined in terms of overlap between every pair of adjacent spatio-temporal regions enclosing F during the interval. The approach is intuitively plausible inasmuch as we do tend to think of persistence in terms of some kind of continuity, or perhaps continuous history, involving the persisting object. We tend to associate diachronic distinctness (distinctness over time), not just synchronic distinctness (distinctness at a time), with breaks in continuity, for example between my car and my house: there is no single continuous path traced by both of them; their spatiotemporal histories are discontinuous. It turns out, however, that it is far more difficult to spell out an adequate continuity criterion of identity over time than it seems, since it has to rule out obvious counterexamples. For instance, take a single-celled organism such as an amoeba, which reproduces by binary fission. It produces daughter organisms neither of which are, it seems, identical to the original; yet one can trace a continuous path between the pre-fission amoeba and each of its descendants. Or consider a marine flatworm, cut into two 55

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. Perhaps the continuity in these cases is too weak, since we could specify a strong form of continuity such that the difference in overlap between any adjacent regions was indefinitely small, which does not seem to obtain in an instantaneous division of the sort just mentioned. But strong continuity also has counterexamples both to necessity and sufficiency for diachronic identity. As to necessity, consider an instantaneous loss or gain of parts: a tree has a branch lopped off, yet it still persists, though the continuity is only weak. As to sufficiency, consider the infinite series beginning with a tree, all of the other members of which are decreasingly smaller parts of the tree – the lump of plant matter minus a millimetre of wood, the lump minus two millimetres, and so on, where the series fully constituted is a real continuum of spatio-temporal parts measured along some dimension, such as length or width. We can specify a strongly continuous path, but we do not want to say that the tree, although strongly continuous with its parts, is identical to any of them. The obvious move, at least to counter whole–part tracing confusions, is to place some sort of sortal restriction on what must be in the path: since parts of trees are not trees, a series such as that just given would not constitute a single persisting object. But there are other examples that make difficulty even here (Shoemaker, 1979; Forbes, 1985, pp. 152–9; the examples go back in some respects to Kripke’s unpublished lectures on identity over time from 1978). Consider a homogeneous rotating sphere. Take one of its segments, i.e., one of its physical parts that moves with the sphere. This is clearly a single, persisting object, namely a sphere segment. Now imagine a light that constantly illuminates one single region with the same surface area as the segment. The segment passes through the illuminated region one time for every complete rotation. But during the period of one rotation, an infinite series of distinct sphere segments also pass through that single, illuminated region. They occupy a strongly continuous path, and they all fall under the same sortal sphere segment, yet they are not 56

a single, persisting sphere segment. Such cases are easily multiplied and provide a formidable challenge to continuity theories of identity. Perhaps the causal relations and counterfactual dependence between the segments is relevantly different from those between the single segment at one time in its history and at another, but spelling this out is no easy matter. Ought we to take a different approach to analyzing persistence? We might take a cue from Butler’s famous criticism of the memory criterion of personal identity over time (Butler, 1975, p. 100), that “consciousness of personal identity presupposes, and therefore cannot constitute, personal identity”. We might argue (Oderberg, 1993; Merricks, 1998) that spatio-temporal continuity gives us evidence of persistence but does not constitute it. For continuity always presupposes identity, inasmuch as the objects related by continuity (the car at t1, the car at t2) are themselves persisting objects – so how can continuity be used to analyze persistence if persistence is always part of what is described in describing a case of continuity? If continuity can only have evidential force – being a symptom of persistence but not a criterion, as it is sometimes put – then the evidence will be defeasible, and in some cases easily so. Where it is absent, moreover, we may still have good grounds for believing identity to obtain: imagine the radical disassembly and reassembly of an object, or its vanishing and reappearing. (Is the latter a metaphysical impossibility? It is certainly conceivable.) Absent any other viable analyses of persistence, we might take it to be a brute fact, an unanalyzable phenomenon. We usually know it when we see it, though we do make mistakes of reidentification. Temporal parts A defender of continuity, however, will not be content with the circularity objection to the proposed analysis. We do not have to think of the relata of the continuity relation as themselves persisting objects: they are, rather, temporal parts of persisting objects, 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 occupied by temporal parts, and each temporal part is itself analyzable in terms of parts of shorter duration, also on a continuous path. The circularity is only apparent, since the termination of the analysis lies in parts that do not themselves persist – we might think of these as points in space and time, occupied by certain qualities. Temporal part theory, often known as four-dimensionalism or (misusing an old word) the theory of perdurance, has many defenders. (See, for a small sample: Quine, 1950; Lewis, 1986; Forbes, 1987; Heller, 1990; Armstrong, 1997.) The general idea is that just as persisting objects have spatial parts (the wheels on my car, the branches of the tree) so they also have temporal parts (that part or stage of my car from t1 to t2, the part or stage of the tree from t3 to t4). The temporal parts are usually designated by a kind of hyphenated singular term: “my car-from-Monday-to-Wednesday”, “the tree-from-Thursday-to-Saturday”, and so on for any persistent and for any times however specified. The idea has intuitive appeal, since we know that objects have spatial parts, and space and time are in many ways similar. Moreover, it seems that contemporary space– time physics, with the theory of relativity at its core, is at least congenial to temporal parts if not committed to them. There is, it might be claimed, no space and no time – only space–time. Space–time is an ontological unity, with objects spread out across both the three spatial and one temporal dimensions, all of which are features of a single “block”, with objects being (on the favorite metaphor) something like “worms” stretched out across the block, divisible into “segments”. These segments, speaking accurately, are supposed to be spatio-temporal parts: there are no purely spatial parts, and no purely temporal parts, but these spatiotemporal parts are what are called temporal parts on the four-dimensionalist way of looking at the universe. How can the four-dimensionalist get persistence from temporal parts? He might simply say that a series of temporal parts constitutes a persisting thing if and only if

it tracks our best intuitive, pre-theoretical judgments about what persisting things there are; in other words, temporal part theory should leave our reidentification practices undisturbed. Ordinary persistence aside, moreover, if we think that a certain object might, say, vanish and reappear after an interval, we could count the series of its temporal parts before disappearance and after reappearance as constituting a single persistent. This view of persistence could be supplemented, or to some extent modified, by a mixture of ontology and evolutionary theory along these lines: every materially occupied portion of space–time, no matter how heterogeneous, constitutes an object. We humans, for survival purposes, “gerrymander” certain portions of space–time and call these particular, privileged “worms” the persisting objects. (See, for example, Quine, 1981.) The debate about the existence of temporal parts and their putative explanatory role as regards persistence continues unabated. The intuitive appeal of four-dimensionalism has captured the imagination of many metaphysicians, but it has to face some serious objections. (For some of the critics, see Geach, 1972; Chisholm, 1976: Appendix A; Thomson, 1983; Oderberg, 1993; Lowe, 1999, pp. 114–18.) For instance, the thought that just because an object has spatial parts so it must have temporal parts is specious. For in order to generate a sufficiently convincing analogy between space and time to motivate the thought, it turns out that one has to presuppose the existence of temporal parts in the first place (as can be seen in Taylor, 1955; discussed in Oderberg, 1993, pp. 97–103; see also Meiland, 1966). Second, is it true that space–time physics commits us to temporal parts? To say that Minkowskian space–time geometry has shown, as Minkowski himself thought, that space and time are “mere shadows” of an underlying unified space–time (Minkowski, 1952, p. 75) could be seen as a metaphysical step too far since the spatial and temporal dimensions are given differing mathematical treatments in relativity theory. One should, in addition, be careful about drawing metaphysical conclusions from 57

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”, “space–time worm”, and the like, since where these terms appear in their work persistence is nearly always presupposed as a more fundamental concept rather than explained or analyzed in those terms. It is in fact very difficult to motivate a metaphysic of temporal parts from space–time physics (Rea, 1998). It might further be argued, independently of considerations from the physics of space– time, that the very concept of a temporal part of a persisting object is of dubious coherence. To be sure, the temporal part skeptic does not deny that some things have temporal parts: events paradigmatically have them (the first half of the battle, the last five minutes of the opera), as do processes (the first hour of a compound’s dissolution in water) and histories (the medieval history of Portugal; the first half of my life). What the skeptic denies is that persisting objects have such parts, and while events and processes involve objects that persist, they themselves do not persist. So what sense can be made of the very idea that a persisting object could have temporal parts? The hyphenated singular terms mentioned above are a philosophers’ invention; not all such inventions are bad, of course, but we should not infer from their existence that what they purport to refer to exists as well. For how could we – how even could God – distinguish between a putative temporal part of an object and a temporal part of that object’s history (or career, as it is sometimes called) with exactly the same temporal boundaries? The critic needs to be more precise, though. The history of my car from Monday to Tuesday involves more than just the car itself: there are all of its relations to other objects that need to be included in that history. The putative temporal part of the car itself from Monday to Tuesday, however, is supposed to involve only what is within the car’s spatial boundaries. The critic, however, can reply as follows. Call that part (not temporal, not exactly spatial – let’s think of it as quasi-spatial) of my car’s Mondaythrough-Tuesday history that involves only the car itself and its intrinsic features its intrinsic history. Hence we factor out, for 58

instance, that part of its history involving its being parked by the kerbside or its being owned by me; we include its being green, curved on top, and having five windows. Now this intrinsic history is a genuine part of the car’s total history. If you wanted to, and you knew my car well enough, you could write a rather boring narrative of its history from dawn on Monday until dusk on Tuesday. Now, says the skeptic, what is to distinguish this history of the car from Monday to Tuesday from the supposed carfrom-Monday-to-Tuesday? Could God, let alone we, tell them apart? But, comes the reply, the temporal part of the car is a physical object, whereas the temporal part of its history is not. It is the physical temporal part that makes the history true. If the car’s history involves its being green on Monday and receiving a coat of red paint on Tuesday, the Monday-Tuesday temporal part will have a green sub-part existing on Monday and a red sub-part existing on Tuesday. Yet the skeptic will insist that there just is no ontological room for such objects. What makes the car’s history what it is from Monday to Tuesday is just the car itself and what is true or false of it: it is green on Monday, red on Tuesday. This is what makes it the case that it has a history with the following temporal parts – the Monday history, in which it is green, and the Tuesday history, in which it is red (and, of course, the temporal part overlapping these in which it is changed from red to green). What room is there for temporal parts of the car itself ? Among various other objections, a couple more are worth raising. Remember that to avoid circularity in the analysis of persistence, temporal parts will have to terminate in instantaneous entities of which all the rest (those with duration) are composed. Yet what sense can be made of an instantaneous temporal part? Calling it a space–time point (with or without qualities “associated” with or “true” of it) does not clarify matters. If instantaneous stages are really that – durationless – then how can they constitute an object with duration any more than dimensionless points can constitute a region with dimension? Wasn’t Zeno right all along? One 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 are no more than limits of a process of potential division, but it is not as though such things have any actuality. The notion might be a good one, but can the fourdimensionalist appeal to it given that he has non-circularly to analyze persistence in terms of stages? No one who accepts the distinction between actual and potential infinity would want to analyze a line in terms of dimensionless points. But the friend of temporal parts needs just such an analysis if he is to avoid being left with an unanalyzed remainder of persisting, i.e., non-instantaneous, temporal parts. Another point concerns whether fourdimensionalism denies the phenomenon it seeks to explain. If persistents are just sums of stages, does anything really persist in the first place? Rather than genuine persistence, doesn’t the friend of temporal parts offer us no more than a series of creations and annihilations, with new matter literally springing into existence ex nihilo all the time (Thomson, 1983)? The temporal parts theorist might bite the bullet here, taking his account to be eliminative; though he would be committed to implausible claims about creation and annihilation. Or he might say that this interpretation is true only if he is a presentist about time, according to which only the present moment and what happens in it are real. More congenial to his position, though, is eternalism, according to which all moments of time and what happens in them are equally real. Matter, on the latter view, does not keep vanishing and springing into existence; rather, the sum of stages making up a persistent is, as it were, given “all at once” – not simultaneously, but with equal reality. There is no temporal becoming: the space-time worm just exists with its spatio-temporal dimensions. Wherever one of the segments is in space-time, so is the persisting object present, just as my car is present wherever one of its spatial parts is. The skeptic still worries that an eternalist view also denies persistence: there is no persistence where there is just creation and annihilation, but equally no persistence where the object is viewed simply as a block in space–time. Moreover, on both presentist and eternalist

interpretations, even if persistence or something approximating it is maintained, can the equally basic phenomenon of change can be accounted for? We will return to this shortly. First, let us briefly consider a couple of other accounts that can be given of persistence. Stage theory A view quite similar to standard fourdimensionalism/temporal parts theory/ perdurance is sometimes called “stage theory” or “exdurance” (Haslanger, 2003). According to this theory (Sider, 2001; Hawley, 2001), there are indeed temporal parts of things other than events, processes, and histories, and there are space-time worms consisting of series of such parts. The basic four-dimensional framework is accepted. What the stage theorist denies, though, is that any of these worms are identical to what we identify as ordinary persisting things. When we talk about persistents we are not talking about worms but about the temporal parts themselves. What we think of as persistents are no more than stages. Of the various motivations for this approach, an important one is to avoid what is seen as a problem about spatiotemporal coincidence. So, assuming (perhaps rashly) that personal fission is possible via a split-brain operation and transplant, suppose that I undergo this procedure and two people each get half of my brain. When they awake, each is psychologically continuous with me. Call the new persons Bill and Ben. Bill is to be tortured, Ben is to live in pleasure. Should I be worried about what will happen to me after the operation, or not concerned, or both, or neither? On the standard four-dimensionalist model, the most likely interpretation of events is that two worms exist before, during, and after the operation – the one including the temporal parts of Bill post-fission and me pre-fission, these being connected by psychological continuity; and the one including the continuous stages of me and Ben. But this means that pre-fission, there are two space– time worms overlapping – the me-Bill worm, and the me-Ben worm. But if persistents are worms, and persistents include persons, then 59

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 overlapping – coinciding in space and time – before the fission occurs. How, though, can two persons be in the very same place at the same time? And what thoughts am I likely to have – a single ambiguous thought about future pain and future pleasure for me, or two thoughts? Wouldn’t the pronoun “I” be ambiguous pre-fission? Could two persons share a single thought, ambiguous or not? Surely I wouldn’t notice any ambiguity when thinking about my fate. This interpretation of fission has too many problems, according to the stage theorist. What we need to say is that there is a single person pre-fission, and that person is a stage, and that stage will be both continuous with Bill-stages and continuous with Ben-stages. In this sense it is true to say, for the stage theorist, that I will be Bill and I will be Ben, but there is only one of me prior to fission. To say that I will be Bill (and will be Ben) is akin to what a counterpart theorist such as Lewis says about modal statements. The counterpart theorist interprets a statement such as “I could have been smarter” as meaning that in some possible world there is a counterpart of me who is smarter than I am (in the actual world). Similarly, for the stage theorist to say that I will be Bill is to mean, properly interpreted, that the stage that I am is continuous with some future Bill-stage. Since the relation is duplicable, it can hold simultaneously of the stage that I am and both future Bill-stages and future Ben-stages. To critics, the stage view fares little better than standard four-dimensionalism. Prefission (at t), I can truly say that after fission (at t1) I will be Bill and that after fission I will be Ben (but not, on a necessary revision of standard reasoning about tense, that I will at t1 be Bill and Ben: see Sider 2001, pp. 201–2). Yet if I will be Bill at t1, is it not the case that there must at t1 be a stage that is me, and that that stage is a Bill-stage? Yet the same is true for myself and Ben: if I will be him, then at t1 there must be a Ben-stage that is me. Yet I cannot be both of them since they are distinct! And by the hypothesis of equal continuity there is no good reason to say that I am either. All the stage 60

theorist allows is that at t1 there is Bill, who was me, and Ben, who was me, and these relations are explained in terms of the duplicable relation of continuity. It is hard to see how genuine identity – and hence persistence in anything like a recognizable form – gets into the picture. If, before fission, I really am a certain stage, then how, without violating the necessity of identity, can it be the case that there is any stage after fission that is me if the pre-fission stage that I am now no longer exists then? So isn’t it the case, on the stage view (assuming necessity of identity) that I simply cease to exist at fission? In which case I will not be Bill and I will not be Ben, in any recognizable sense of those propositions. The stage theorist will accept that what he posits is not genuine identity; as one such theorist puts it, “claims of identity between things at different times make sense, even though they are false” (Hawley, 2001, p. 156). Yet the stage theorist wants to preserve the commonsense belief that persons and other persistents do exist. So on the theory, persons (for example) do exist but no identity statements about them are literally true! Well, it takes time to have a thought – even the simplest and most fleeting of thoughts – but it cannot then be literally true that it is I who has any thoughts. For persons are stages, and the only real stages are instantaneous ones. One can call a three-hour stage of me a stage, but it only has the title honorifically, or we might say derivatively. It literally has no thoughts, rather it is made up of instantaneous stages with mental properties suitably related in some unspecified (and arguably unspecifiable) way. But what possible mental properties could an instantaneous stage have, if even the briefest of thoughts takes time? Presumably, moreover, there is only one of me. But if I am a stage, which stage? It seems that stage theory retains all of the vices of standard fourdimensionalism but loses any virtues, for at least the standard worm theorist hold that there is precisely one of me, and that this single person is a four-dimensional sum of stages. On the stage view, it looks as though eliminativism about persons and other persistents 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 to help itself to the commonsense belief that persons, cars, trees, and the other familiar objects of our universe do indeed exist. Endurance Theorists of persistence usually speak as well of “endurantism”, taken as the view of virtually all those who deny that objects have (or are) temporal parts, so rejecting fourdimensionalism of any stripe. There are many such metaphysicians, but whether there is a theory around which they all rally is dubious. The believer in endurance rejects four-dimensionalism for the reasons already given and more. Stage theory, as we have seen, denies literal persistence. Standard worm theory takes there to be persistents – sums of continuous stages – but, according to the critic, it denies the reality of change (see, for example, Lombard, 1994, replying to Heller, 1992; see also Oderberg, 2004). All there is, on fourdimensionalism, is replacement of one temporal part by another, or addition of one temporal part to another – but neither replacement nor addition are genuine change. When my red car is painted green, a red car-stage (or series of such stages) is replaced or added to by a green car-stage (or series of such stages). As Peirce put it, “Phillip is drunk and Phillip is sober would be absurd, did not time make the Phillip of this morning another Phillip than the Phillip of last night” (Peirce, 1931, 1.494). The endurance “theorist” wants to retain both the commonsense belief that there is literal persistence and the commonsense belief that there is genuine change throughout that persistence. In other words, one and the same object literally has a property and loses it. No four-dimensionalist theory can hold on to both of these beliefs. True, for the worm theorist my car does exist at every time at which any of its temporal parts do, but the properties it gains and loses are gained and lost only in virtue of there being distinct stages that have and do not have those properties respectively. My house also exists at every place at which it its spatial parts do, and many of its intrinsic

properties are had only because one or more of its parts has those properties: it is warm because parts of it are warm, it is brick because it has parts made of brick, and so on. Now whether the worm theorist can account for all the properties of a thing in terms of properties of its temporal parts is highly questionable (see Zimmerman, 1998), but even restricting ourselves to those that she can so account for, why shouldn’t we say that my house changes across space as well, since it has distinct spatial parts with different properties? My house does in one sense vary across space, but the endurantist holds that not all variation is change, and that something crucial is lost when change is defined (as it so often is in metaphysics texts) as the mere having of a property by an object at one time and its lacking it at another. For if that is all there is to change, then objects change across space as well, since why on this view should having a property at a time be significantly different from having it at a place? But change is a fundamentally dynamic phenomenon, involving a real transition of a thing itself from one state to another. Mere addition, replacement, and/ or distinctness of parts do not capture this phenomenon. The endurantist is often asserted to hold as a theoretical commitment that a persistent is “wholly” present at different times (Lewis, 1986). If there is any theory here, it is the denial of four-dimensionalism. But it is not the assertion of a recondite metaphysical state, merely the belief that one and the same object literally exists through time, itself having properties at some times that it loses at others. There are, though, theoretical consequences of this commonsense view. For example, endurance rules out any approach to fission cases that posits coinciding pre-fission objects, or violation of the standard logic of identity (the idea that I will be both Bill and Ben is a non-starter), or any relation weaker than strict identity as capturing “what matters” as between me and my post-fission descendants. Since identity cannot hold between myself and both Bill and Ben, the only option for the endurantist is to deny that I continue to 61

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. Since the fission case has psychological continuity built into it, this means denying that I really will be psychologically continuous with both Bill and Ben and hence denying that genuine fission is possible, on the assumption that psychological continuity is present wherever personal identity is and vice versa. (The endurantist could hold the weaker position that psychological continuity is defeasible evidence of personal identity, and simply claim that in the case of myself, Bill, and Ben, the evidence is defeated.) Change: metaphysics and semantics What, then, of the phenomenon of change? In contemporary discussion, following Lewis (1986), theorists set up what is called the “problem of temporary intrinsics”. The idea is that the following propositions are incompatible: (1) that objects (such as my car or me) persist through change; (2) that, across the same dimension of change, the intrinsic properties involved in an object’s change are incompatible (being red and green, or red and non-red, round and square, round and non-round, etc.); (3) no object can possess incompatible properties. The problem is how, in analyzing change, all of these very plausible claims can be held true. Kant, for one, gave voice to a worry about how there could be a “combination of contradictorily opposed predicates in one and the same object” (Kemp Smith, 1933, A32/ B48, p. 76). A concern about the way in which this problem has been tackled is that two distinct issues have tended to be conflated: the semantic one of how to represent sentences describing change in such a way that they do not state a contradiction, and the metaphysical one of how change should be understood in such a way that no contradiction is assumed or implied. Semantics and metaphysics are not the same thing, and so the identity theorist needs to be careful to separate these issues. Taking the metaphysical one first, the obvious target is the third proposition. Does the Law of Non-contradiction state that nothing can have incompatible properties? Not at all. 62

The locus classicus for the law, followed by virtually all philosophers ever since, is Aristotle, who affirms: “[T]he same attribute cannot at the same time belong and not belong to the same subject and in the same respect” (Metaphysics, Book Gamma, sect. 3, 1005b19; Ross, 1928; emphasis added). When my green car is painted red it certainly does not, at any one time, possess incompatible properties – the change itself ensures that the law is not violated, nor could it be. Hence it looks as though the “problem of temporary intrinsics” is spurious: why would anyone want to affirm the third proposition unless they had not thought carefully about the Law of Non-contradiction in the first place, or they wanted something to puzzle about for the sake of it? Similarly, though less obviously, change does not involve any violation of Leibniz’s Law. This law (more precisely that half of the law called the Indiscernibility of Identicals) states that if x and y are identical then they share all their properties. Some writers (e.g., Heller, 1992) argue that if an object such as my car is red at t and green at t1, then my car at t is discernible from my car at t1, the first being red and the second non-red. But this cannot be, so the properties must be possessed by numerically distinct temporal parts (united into a single four-dimensional worm). Yet Leibniz’s Law is entailed by the Law of Non-contradiction: no object can both possess a property and lack it at the same time and in the same respect. So if x and y are identical, i.e., the same object, that object x (y) cannot be F and not-F at the same time and in the same respect. So it must be that if x is F, then it (y) is F at the same time and in the same respect, and vice versa. So if change violates Leibniz’s Law, it violates the Law of Non-Contradiction, which it cannot do. For if change meant that x and y really did not share all their properties, though they were one and the same object, a contradiction would result. To say that Leibniz’s Law still allows discernibility at different times, so is not entailed by The Law of Non-contradiction, is to miss the point. For if y has a property at t1 incompatible with a property that it (x) has at t, there will be a 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 won’t have it at t. In other words, Leibniz’s Law does not mean that objects cannot change: if such change is just what we mean by “discernibility at different times”, that is harmless. But if we intend something more by the expression, e.g., that the property y has at t1 is a property that x (=y) lacks at t1, even given that x exists at t, we end up in contradiction. So: my car at t is red at t, and my car at t is green at t1; my car at t1 is green at t1, and my car at t1 is red at t. My car does not at any time possess incompatible properties, though it does so at different times. But having incompatible properties at different times does not mean that x has any property that y lacks, and conversely. This is where the semantic problem rears its head, though for all its interest we can only consider it briefly. The problem is how, semantically, to represent my car’s change in such a way that no contradiction is stated or implied. What do these expressions such as “at t”, “at t1”, and so on, mean? Does it matter where they are placed in a sentence stating property possession? There are at least three alternative proposals for dealing with this. The four-dimensionalist (including the stage theorist) applies the metaphysics to the semantics: the temporal qualifiers are affixed to the subject terms so as to block a contradiction. My car-at-t is red and my car-at-t1 is green. The subject terms denote temporal parts of my car, so no one thing literally possesses incompatible properties at any time. For reasons already given, the temporal part skeptic will reject this approach. Another is called adverbialism (Johnston, 1987; Hanslanger, 1989): the temporal qualifiers are attached to the copula. Hence my car is-in-the-t-way red and my car is-in-the-t1-way green. The same object possesses incompatible properties but in different ways, and these different ways of property possession remove the contradiction. It is difficult to get a grip on whether adverbialism has any metaphysical implications, and if so what they are. One semantic criticism is that the adverbialist has to give an account of temporal adverb dropping. Since we can usually drop adverbs and preserve truth (Fred runs

fast, therefore Fred runs), why can’t we drop temporal adverbs? But if we do so we end up with a contradiction again – my car is red and non-red – so must the adverbialist say the temporal adverbs cannot be dropped simply to avoid contradiction? If so, the justification for the solution looks circular: introduce temporal adverbs to block contradiction, but then exclude the standard semantic rule for dropping adverbs because otherwise there would be a contradiction. (See further Oderberg, 2004, and also Merricks, 1994 for criticism of adverbialism.) A third approach is called “sententialism” (Oderberg, 2004; see also Myro, 1986, who uses sentential temporal operators but for a different purpose). Taking the separation of semantics from metaphysics seriously, the sententiaist holds that the temporal operators in sentences describing change are affixed to atemporal predications. At t, my car is red and at t1, my car is non-red. The temporal operators create something like an opaque context: not strictly, since the context is still extensional (it doesn’t matter what co-referring subject term I use for the sentence to be true), but the operators cannot be dropped. Why not? The main reason is that atemporal predications for changeable objects are incomplete – they do not state facts about the objects, only incomplete information that needs supplementation to make sense. So no inference can sensibly be made for such objects from a temporal to an atemporal predication. This is reinforced by the semantic fact that when we make predications that are not explicitly temporal – “My car is red” – there is always taken to be an implicit reference to the present, since otherwise the statement would be radically incomplete and truth unevaluable. Hence semantics is on the side of the sententialist, whereas the adverbialist has the standard rule in favor of adverb dropping to contend with. See also the a–z entries on being and becoming; change; continuant; continuity; identity; persons and personal identity; space and time; temporal parts, stages. 63

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 Armstrong, D.M.: A World of States of Affairs (Cambridge: Cambridge University Press, 1997). Butler, J.: “Of Personal Identity,” Appendix 1 of The Analogy of Religion (1736); repr. in Personal Identity, ed. J. Perry (Berkeley: University of California Press, 1975). Chisholm, R.: Person and Object (La Salle, IL: Open Court, 1976). Coburn, R.: “Identity and Spatiotemporal Continuity,” in M. Munitz, ed., Identity and Individuation (New York: New York University Press, 1971), 51–101. Forbes, G.: “Is There a Problem about Persistence?,” Proceedings of the Aristotelian Society, suppl. vol. 61 (1987), 137–55; repr. in Haslanger and Kurtz (2006). Forbes, G.: The Metaphysics of Modality (Oxford: Clarendon Press, 1985). Geach, P.: “Some Problems about Time,” in his Logic Matters (Oxford: Blackwell, 1972). Haslanger, S.: “Endurance and Temporary Intrinsics,” Analysis 49 (1989), 119–25. Haslanger, S.: “Persistence through Time,” in M.J. Loux and D.W. Zimmerman, ed., The Oxford Handbook of Metaphysics (Oxford: Oxford University Press, 2003), 315–54. Haslanger, S. and Kurtz, R.M., ed.: Persistence: Contemporary Readings (Cambridge, MA: Bradford Books/MIT Press, 2006). Hawley, K.: How Things Persist (Oxford: Clarendon Press, 2001). Heller, M.: The Ontology of Physical Objects (Cambridge: Cambridge University Press, 1990). Heller, M.: “Things Change,” Philosophy and Phenomenological Research 52 (1992), 695– 704. Johnston, M.: “Is There a Problem about Persistence?,” Proceedings of the Aristotelian Society, suppl. vol. 61 (1987), 107–35. Kemp Smith, N. (trans.): The Critique of Pure Reason, by Immanuel Kant (London: Macmillan, 1933; originally published 1781 (A), 1787 (B)). Lewis, D.: On the Plurality of Worlds (Oxford: Blackwell, 1986). Lombard, L.: “The Doctrine of Temporal Parts and the ‘No-Change’ Objection,” 64

Philosophy and Phenomenological Research 54 (1994), 365–73. Lowe, E.J.: The Possibility of Metaphysics: Substance, Identity, and Time (Oxford: Clarendon Press, 1999). Meiland, J.W.: “Temporal Parts and Spatiotemporal Analogies,” American Philosophical Quarterly 3 (1966), 64–70. Merricks, T.: “Endurance and Indiscernibility,” The Journal of Philosophy 91 (1994), 165–84. Merricks, T.: “There Are No Criteria of Identity Over Time,” Noûs 32 (1998), 106–24. Minkowski, H.: “Space and Time,” in H.A. Lorentz, A. Einstein, H. Minkowski, and H. Weyl, The Principle of Relativity (New York: Dover, 1952); address originally given in 1908. Myro, G.: “Identity and Time,” in R. Grandy and R. Warner, ed., Philosophical Grounds of Rationality (Oxford: Clarendon Press, 1986), 383–409. Oderberg, D.S.: The Metaphysics of Identity Over Time (New York: St. Martin’s Press, 1993). Oderberg, D.S.: “Temporal Parts and the Possibility of Change,” Philosophy and Phenomenological Research 69 (2004), 686–708. Peirce, C.S.: Collected Papers, vol. I (Cambridge, MA: Harvard University Press, 1931). Quine, W.V.: “Identity, Ostension, and Hypostasis,” The Journal of Philosophy 47 (1950), 621–33; repr. in his From a Logical Point of View (New York: Harper and Row, 1961), and in Haslanger and Kurtz (2006). Quine, W.V.: “Worlds Away,” in his Theories and Things (Cambridge, MA: Harvard University Press, 1981). Rea, M.C.: “Temporal Parts Unmotivated,” The Philosophical Review 107 (1998), 225–60. Ross, W.D. (trans. and ed.): Aristotle’s Metaphysics, vol. VIII of The Works of Aristotle Translated into English, 2nd edn. (Oxford: Clarendon Press, 1928). Shoemaker, S.: “Identity, Properties, and Causality,” in P. French, T. Uehling, and 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 Minnesota Press, 1979), 321–42; also in his Identity, Cause, and Mind (New York: Oxford University Press, 2003). Sider, T.: Four-Dimensionalism: An Ontology of Persistence and Time (New York: Oxford University Press, 2001). Swinburne, R.: Space and Time (New York: St Martin’s Press, 1968; 2nd edn., 1981). Taylor, R.: “Spatial and Temporal Analogies and the Concept of Identity,” The Journal of Philosophy 52 (1955), 599–612. Thomson, J.J.: “Parthood and Identity Across Time,” The Journal of Philosophy 80 (1983), 201–20; repr. in Haslanger and Kurtz (2006). Zimmerman, D.W.: “Temporal Parts and Humean Supervenience: The Incompatibility of Two Humean Doctrines,” Australasian Journal of Philosophy 76 (1998), 265–88. david s. oderberg

Realism and Antirealism about Abstract Entities 1. What Is Realism? Realism about abstract entities, in its most general form, asserts – and antirealism denies – that there are such things. This simple formulation calls for some explanatory comment. (1) Neither realism, nor its denial, need be an all or nothing affair. Realists have asserted, and their opponents have denied, the existence abstract entities of several different kinds – universals, mathematical entities such as numbers and sets, propositions, and various others (see number; class, collection, set; proposition, state of affairs). Realists may be selective about the kinds of abstract entities whose existence they assert, and antirealists may likewise be selective – denying the existence of one kind of abstract entities, while remaining agnostic about, or even accepting, the existence of another. Quine, for example, was a realist – for most of his career – about sets and numbers, while steadfastly refusing

to accept the existence of propositions and properties or attributes. (See Quine, 1960, chs. 6, 7 and 1970, ch. 1). (2) Realism and antirealism, though mutually exclusive, need not exhaust the possibilities. One may just be agnostic about whether or not there are abstracta (of some given kind). More interestingly, unwillingness to assert or deny the existence of abstracta might stem from a conviction that the issue is either hopelessly unclear or confused. Carnap’s view that philosophers’ questions about the existence of numbers, propositions, etc., are not genuinely factual or “theoretical” questions at all, but misleadingly formulated “practical” questions – calling for a decision whether or not to adopt a certain “linguistic framework” rather than an answer assessable as true or false – can be seen as exemplifying the latter position. (See Carnap, 1950.) (3) Abstract – as opposed to concrete – entities are commonly taken to be those, if any, which occupy neither space nor time. Thus they contrast both with physical entities, which occupy both space and time (e.g., tables, tennis matches, vapor trails, and more exotic entities like sub-atomic particles and force-fields), and with those entities, if any, which occupy time but not space (e.g., mental events, processes and states, on some dualist views), or space but not time (possible examples: the Greenwich Meridian, the North Pole, and spatial points and regions generally). This explanation is not unproblematic. While numbers and sets, and many other standard examples of the abstract, have neither spatial nor temporal location. It makes no sense to ask where the number 17 is, or when it came into existence, or how long it will last. But it is not clear that this holds for all abstract entities – one might, for example, argue that literary and musical works (as distinct from copies or performances) are abstract entities, but that they have not always existed – rather, they came into existence when first composed, and so are not wholly atemporal. We shall, however, assume its approximate correctness here. (See concrete/abstract. For skepticism about the distinction, see Lewis, 1986, pp. 81– 6; for 65

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 Dummett, 1973, ch. 14; Noonan, 1976; Hale, 1987, ch. 3). (4) Realism is standardly taken to involve the further claim that the existence of abstracta is objective, this being understood in terms of mind-independence (see objectivity). This is both natural and plausible, but not unproblematic. Indeed, the same examples illustrate the difficulty – novels and symphonies are (complex) abstract objects, but ones which would not have existed without a good deal of mental activity on the part of their makers. Perhaps the simplest way around this difficulty is to distinguish a strong form of realism which asserts that there are mind-independent abstract entities, and weaker forms which assert the existence of abstracta, but not their mindindependence. The strong realist thesis is itself open to more and less demanding interpretations, which differ over how mind-independence is understood. A minimum condition for the existence of certain entities to be mind-independent is that they would exist even if there were no minds, and so exist independently of our actual knowledge or beliefs about them. But we may distinguish an extreme form of realism, according to which the existence of abstracta is entirely independent, even in principle, of the possibility of our knowing of it, from more moderate forms which maintain that there are abstract entities which would exist even if there were no thinkers, but which accept an epistemological constraint to the effect that their existence must be detectable, at least in principle. (5) Alexius Meinong (1904) (see nonexistent objects) denies existence (German: Existenz) to abstract entities, but maintains that they have a different kind of being, sometimes called “subsistence” (German: Bestand). A closely related, but subtly different view – sometimes called “noneism” – is defended in Routley (1980) and Priest (2005). Meinong’s doctrine is standardly classed as a kind of realism, but in our terms, Meinongian Realism counts, somewhat paradoxically, as a form of antirealism. It is tempting to suppose that Meinong is using the word “exists” in a restricted way, 66

so as to apply only to what occupies space and time, or perhaps only to what is capable of causal interaction. If this were so, the disagreement between his Realism and realism as characterized here would be largely if not entirely verbal, at least as far as the ontological status of abstract entities is concerned. One might be similarly tempted to think that the disagreement between realists and antirealists in our sense is likewise a merely verbal one, in which antirealists are simply evincing a prejudice in favor of restricting application of the word “exists” to what is concrete. But while some antirealist polemics encourage such a view, the temptation should, in this case, be resisted. There is a genuine issue, and it concerns knowledge. If his position is to be taken seriously, the realist must claim that we can have at least some knowledge about some abstract entities. But then, given that such entities lack spatio-temporal location, and so must be incapable of standing in any causal or other natural relations to us, however remote, he faces a challenge to explain how knowledge about them is possible. We shall return to this issue. 2. Some Realist Views One may, as noted, be a realist about one kind of abstract entities but not about others. We illustrate with three examples. Universals One of the earliest and most famous realist doctrines is Plato’s Theory of Forms, which asserts the existence such things as the Beautiful and the Just in themselves, over and above particular beautiful objects and just acts which, in Plato’s view, more or less imperfectly exemplify them. Although Plato’s usual term for the Forms (ειδοσ) is often translated as “Idea”, it is clear that he takes them to be abstract entities existing independently both of our mental activity and of their instantiation in sensible particulars (see plato). In support of this view, it may be argued that there is something which different just acts, for example, have in common, in virtue of which they are all rightly said to be just, and that what they have in common does not depend for its existence upon any of those 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 place, but what they have in common has itself no spatio-temporal location. The detailed interpretation of Plato’s theory and his arguments for it remain matters of scholarly controversy, but there is no doubt that his promulgation of the theory initiated a dispute over the nature and existence of universals – often conceived, in opposition to particulars, as entities such as general properties which may be wholly present at different times and places, or instantiated by many distinct particular objects – which has been actively pursued in much subsequent philosophy (see universals and particulars). Propositions Much as realists about universals argue for them by appealing to the existence of something common to different particular objects or events which all satisfy some general description or predicate (e.g., “blue”, “square”, “just”, etc.), so some philosophers have argued that when different speakers or thinkers say or think, say, that 172+1 is even, or that Julius Caesar was assassinated, there is something common to their distinct linguistic performances or psychological acts or states. What they share is a common content – what is said or thought, as distinct from the saying or thinking of it. In other words, they all assert, or assent in thought to, the same proposition. Propositions in themselves, in contrast with the linguistic performances or psychological acts or states in which they expressed or encoded, have no spatial or temporal location, and hence are abstract objects (A classic statement of realism about propositions is Bolzano, 1972). Numbers, Sets and other Mathematical Entities The sentences of pure mathematics almost invariably involve expressions (simple or complex singular terms) whose ostensible role is to make reference to numbers of some kind, or sets, or other mathematical entities, along with quantifiers binding variables understood as ranging over such entities. Simple examples are: 2+7=9 Every set of real numbers which is bounded above by a real number has a least upper bound in the real numbers

For every set X there exists a set Y whose members are exactly the subsets of X If these and similar sentences, taken at facevalue, are true, then there must be numbers, sets, etc., to which they refer or over which they quantify. But such sentences are widely accepted as true, and are accepted as they stand, without benefit of some reinterpretation which dispels the appearance of reference to or quantification over numbers, sets, etc. Here we have the premises of an argument which makes at least a prima facie case for the existence of numbers, sets, and other mathematical entities – abstract entities, surely, if any are – and hence a case for realism (or Platonism, as it is often called – see Platonism) about mathematics. 3. Antirealism Realism’s traditional opponents have been nominalists (see nominalism). Thus in the medieval dispute over universals, the nominalists insisted that there exist only particular entities, and that the application of the same general term (or name – hence the label “nominalism”) to many distinct particulars does not require the existence of a common non-linguistic entity which is somehow present in each of them, but is sufficiently explained by reference to similarities between them. Likewise, in the modern dispute over the existence of abstract entities in mathematics, nominalists argue that the acceptance of mathematical theories involves no unavoidable commitment to the existence of numbers, functions, sets, or any other ostensibly abstract entities. Before we consider some of the main strategies by which nominalists have sought to avoid such commitment, we shall briefly review their reasons for thinking it is necessary or desirable to avoid it. Nominalists have often recommended their rejection of abstracta on grounds of ontological economy, invoking the methodological maxim known as Ockham’s Razor – entia non sunt multiplicanda praeter necessitatem – which may be glossed as asserting that we should not postulate kinds of entity beyond what is necessary (see Ockham). Although a popular ploy, this is problematic 67

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 directive only when accompanied by some answer to the obvious question: “Necessary for what?” The equally obvious answer is: “Necessary to account for all the (agreed) facts”, but it is doubtful that there is sufficient agreement here to enable the nominalist to cut away abstracta as unnecessary. The realist is likely to suppose that the relevant facts include facts of mathematics which, taken at face value, do require the existence of numbers, sets, etc. But second, even if the facts in need of explanation can be restricted, without begging the question, to facts about the concrete, it is still unclear that the nominalist will be in position to wield the razor to advantage, since it may be argued that those facts admit of no satisfactory explanation without the aid of scientific (and especially physical) theories which make indispensable use of mathematics. This – often called the Quine–Putnam indispensability argument – receives its clearest formulation in Putnam (1971). Since theories (especially mathematical theories) ostensibly involving reference to abstracta appear to play an indispensable rôle in our intellectual economy, nominalists can scarcely afford simply to reject them outright; rather, they must explain how we may justifiably retain such theories, without offending against nominalistic scruples. The standard nominalist response has been to seek ways of paraphrasing or re-interpreting problematic statements and theories in nominalistically acceptable terms – with the aim of showing that their apparent reference to and quantification over abstract entities is unnecessary or merely apparent. This strategy has met with limited success. The difficulties can be well illustrated by reference to arithmetic. Consider first simple equations, such as “2 + 3 = 5”. As a step towards eliminating its apparent reference to numbers, we may paraphrase it along the lines: “If there are exactly two Fs and exactly three Gs and no Fs are Gs, then there are exactly five F-or-Gs” (in symbols: (∃2xFx ∧ ∃3yGy ∧ ¬∃x(Fx ∧ Gx) ) → ∃5x(Fx ∨ Gx) ) . Although this still contains number words, they occur only in the context of

numerically definite quantifications like “there are exactly two Fs” (∃2xFx). These are logically equivalent to sentences involving no number words at all, such as “there is something which is F and something else which is F and any F is identical with one or other of these things” (∃x∃y(x≠y ∧ ∀z(Fz↔ z=x ∨ z=y) ) ). Thus at some cost in length and readability, we may be able to reduce “2=3=5” to something nominalistically acceptable. But even if this kind of paraphrase works for simple equations, it plainly won’t work for general arithmetical statements, such as ∀a∀b (a + b = b + a), in which we quantify over numbers, without mentioning any in particular. Thus unless virtually the whole of arithmetic is to lie beyond the nominalist’s reach, additional and more widely applicable methods of paraphrase or re-interpretation will be needed. Eliminative structuralism offers a more promising strategy. On this account, arithmetic is not a theory about a particular infinite sequence of abstract objects – the numbers 0,1,2,3, . . . – but gives completely general information about those objects, if any, which exemplify a certain structure (viz. being a sequence having a first term, and for each term, a unique next term, and so no end of terms – progressions, or ω-sequences, in the usual jargon). Since, on this re-interpretation, no arithmetic sentences assert the existence of any objects, they are all nominalistically acceptable. A well-known difficulty is that unless there exists at least one ω-sequence, the eliminative structuralist’s translations of all arithmetic sentences, including those of false ones like 2+3=6, come out true. This leaves the nominalist facing a dilemma: to avoid this disaster, she must assert the existence of an ω-sequence – but if she asserts that there infinitely many abstract objects, she abandons nominalism, while if she asserts that there are infinitely many concrete objects, the viability of her translation-scheme depends upon an empirical hypothesis, and one which may very well be false. Perhaps, as Hellman (1989) argues, this dilemma can be avoided 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 necessarily holds of any ω-sequence – for then the nominalist need only assert the possible existence of an ω-sequence to avoid disaster, and perhaps the claim that there could be an ω-sequence is nominalistically unproblematic and otherwise acceptable. However, even if a nominalist version of arithmetic can be salvaged in this way, it is doubtful whether the strategy can be extended to more powerful mathematical theories such as set theory, since the needed possible existence claim would amount to the claim that there could be a concrete model of transfinite set theory, and this is surely false. Following a more radical course, Hartry Field (see Field, 1980, 1989) has argued that nominalists can deny that mathematical theories are true, thereby avoiding commitment to their abstract ontology, but still accept them provided they are conservative in the sense that their conjunction with non-mathematical (e.g., physical) theories entails no claims about non-mathematical entities which are not logical consequences of those non-mathematical theories by themselves. Conservativeness in this sense, like logical consistency, does not require truth – a theory can be conservative without being true. The important uses of mathematics in science, Field holds, are two: we use it to deduce the consequences of scientific theories, and we use it, especially in physics, in actually formulating such theories. The assumption that standard mathematics is conservative, Field argues, is enough to justify its use in deducing, and with the help of this assumption, we can, he thinks, show that there are acceptable nominalistic reformulations of such theories. Field’s view has attracted a barrage of objections, both technical and philosophical. Several critics have questioned whether Field’s reformulations of scientific theories really are nominalistically acceptable. Others have argued that he is committed to the implausible view that while there exist no numbers or sets, their non-existence is a merely contingent matter. (See Maddy, 1980; Chihara, 1990; Hale and Wright, 1992; Burgess and Rosen, 1997).

4. Vehicles of ontological commitment – reference and quantification It was claimed above that a sufficient condition for the existence of objects of a given kind, F, is the occurrence in true statements of expressions functioning as singular terms which, if they refer at all, refer to Fs. Such terms are, we might say, vehicles of ontological commitment. It might be objected that the suggested condition cannot be sufficient as it stands, and that we should additionally require that the relevant singular terms be ineliminable by reductive paraphrase of the sort orthodox nominalists have sought to supply. But this objection is confused. Accepting as true statements in which certain expressions function as singular terms commits us to the existence of corresponding objects, simply because those statements cannot be true unless their ingredient expressions discharge their semantic functions, and the semantic function of singular terms is to pick out objects. Antirealists may agree, but object that this misses the real point, which is that if statements apparently involving singular terms for abstract objects can indeed be replaced by equivalent statements which do not, this shows that those terms are not genuine singular terms at all, and that the original statements, contrary to first appearances, involve no commitment to such objects. This antirealist counter assumes that if statements apparently involving ontological commitment to Fs are equivalent to other statements apparently free of any such commitment, it is the latter statements which should be reckoned as truly reflecting our ontological commitments, not the former. But why? An equivalence, as Alston (1958) points out in a perceptive discussion of the issue, is just that – what it shows, by itself, is only that if either of the two kinds of statement involves a commitment to Fs, then both do. But to get to the conclusion that statements of the first sort involve no genuine reference (and hence commitment) to Fs, we need a further premise – one providing a reason to regard the appearances presented 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 of the other sort. Suppose we could introduce terms for the directions of straight lines by means of the Direction Equivalence: The direction of line a = the direction of line b iff lines a and b are parallel – the idea being to establish a use for such terms by fixing the truth-conditions of identity statements involving them. (See Frege, 1884, §64.) The nominalist will regard the equivalence as revealing that any apparent commitment to the existence of abstract objects carried by talk of directions is merely apparent. The realist will instead regard it as disclosing an unobvious commitment to the existence of directions implicit in talk of parallelism among lines. Of course, the realist must agree that one could possess the concepts of straight line and parallelism without having that of direction – indeed, one must be able to do so, if the latter concept is to be explained by means of the Direction Equivalence. His claim is that the commitment to directions is implicit in the sense that, once one has acquired the concept of direction in this way, one cannot consistently hold that there are straight lines but no directions. (For further discussion, see Wright, 1983, §§5,10.) The realist claims we should take apparent reference to abstracta at face value, in the absence of compelling reason to do otherwise. Resolution of the issue in favor of an ontologically reductive interpretation of such equivalences – and so in favor of the antirealist – requires making a case that there is compelling reason to do otherwise. We shall return to this question. Our proposed sufficient condition for the existence of Fs is clearly not a necessary condition. It may be that there are Fs whose existence we suspect not, and of which, therefore, we do not speak. Perhaps, indeed, we have no concept of them. Nor, evidently, is a readiness to make statements involving singular terms for Fs needed for a commitment to their existence. For without employing any words which purport reference to particular Fs, we may simply assert that there are Fs, or more generally, assert some quantified statement whose truth requires their existence. Roughly, quantification over 70

Fs is an alternative vehicle of ontological commitment to Fs. Quine, famously, took it to be the sole vehicle: The objects whose existence is implied in our discourse are finally just the objects which must, for the truth of our assertions, be . . . reckoned into the totality of objects over which our variables of quantification range. To be is to be the value of a variable. (Quine, 1952, §37) Quine sees quantification as the vehicle of ontological commitment because he assumes that only ineliminable occurrences of singular terms would distinctively carry ontological commitment, and believes that there are no such terms, i.e., that singular terms are everywhere eliminable. We may, he argues, always eliminate them by paraphrase using just general terms (predicates) and quantification, either by the technique of Russell’s Theory of Definite Descriptions (coupled with his doctrine that ordinary proper names are “abbreviated” descriptions) or, if necessary, by an extension of it due to Quine himself whereby we may replace any proper name by a corresponding predicate understood as applying to that object, if any, the name names – thus “Socrates drinks”, for example, may be paraphrased as “∃x(x socratizes & x drinks)”. Quine is also taking it for granted that predicates or general terms carry no commitment to corresponding entities. If this assumption – to which we shall need to return – were granted, it would be at least plausible that quantification over Fs is the essential mark of commitment to their existence. However, while Quine’s eliminability thesis is, in one way, beyond dispute, its significance is not. We may agree that, starting from a base language containing singular terms, we could employ Quine’s recipe to construct a language in which all such terms were replaced by corresponding predicates, but deny this purely syntactical manoeuvre has any semantic or, more widely, philosophical significance. It is quite unclear how one might learn the use or satisfaction conditions of Quine’s replacement predicates, in the absence of any means of 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. Relatedly, it does not seem one could explain the truth-conditions of quantified sentences of Quine’s replacement language without treating variables as, in effect, functioning as temporary names of objects in the domain of quantification . (See Dummett, 1973, pp. 223–6, 476–80.) The Access Problem Realists need to explain how we can know about the abstract entities whose existence they assert – how we can know that there are such things at all, and how we can know truths about them. The problem of providing such an explanation is part of what I shall call the access problem. It is the fundamental problem for realism. If realists could solve it, it is difficult to see what, other than prejudice, would stand in the way of acceptance of their view. If, on the other hand, it could be shown that they cannot solve it, that would be a decisive objection, and would encourage, or even enforce, an ontologically reductive reading of the kind of equivalences between statements ostensibly about abstracta and others apparently free of commitment to their existence discussed in the preceding section. In the absence of at least the outlines of a solution, or reason to believe one can be found, it is hard to take realism seriously – ontology without epistemology is just idle speculation. (See Hart, 1979; Bell, 1979). Why do we – or might we – find the idea that we may have knowledge about abstract objects so baffling? In explaining how we know much of what we know, we appeal to causal connections, such as those involved in perception. This may encourage acceptance of a broadly causal theory of knowledge – one which sees basic bits of knowledge as involving a suitable causal connection between knowers and the known truths, and other knowledge as arising from this basis by a more or less complicated process of inference. Then, given that abstract objects stand in no spatial or temporal relations with us, and so in no causal relations, it may seem not just that knowledge about them eludes explanation, but that there can

be no such knowledge. (See Benacerraf, 1973; Steiner, 1973; Kitcher, 1978). As against this, it may be claimed that even a broadly causal theory is open to objection on independent grounds; in particular, such a theory would seem directly to rule a priori knowledge – and while there is certainly a serious problem in explaining how such knowledge is possible, it does not seem that its impossibility should be so easily established. However, as Field (1989, pp. 25–7, 230–9) and others (Hart, 1977; Maddy, 1990, pp. 42–5) have pointed out, doubts about the capacity of realism to deliver a credible epistemology do not have to be grounded in the adoption of a specifically causal analysis of knowledge. For even if a causal constraint is not written into the analysis, the problem of explaining how we can acquire knowledge, or reliably form true beliefs about abstract objects, remains. Epistemological perplexity about, and consequent suspicion of, abstract entities has other and more general sources, besides causalist or, more generally, reliabilist thinking in epistemology, which arguably obstruct progress on the access problem. One is that we tend to operate with a wholly negative conception of abstract objects as “outside” space and time. This characterization is obviously metaphorical, as well as negative – there is, literally, nowhere outside space and time. But this in itself need not be particularly damaging, so long as we remind ourselves, when necessary – that is, when we feel tempted to think of abstract objects as “in” some queer sort of limbo – of the literal content of the metaphor: roughly, that it makes no sense to ask where an abstract object is, or when it came into existence, or how long it will last. It is, rather, the negative aspect of the characterization that impedes constructive thought. Of course, it is true that abstract objects aren’t located in space or time. And it may be said that since that it enough to ensure that there is an apparently intractable problem about how spatio-temporally located knowers could know of their existence or know anything about them, it is pointless exercising ourselves over what more positive 71

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. But that just misses the present point: if we focus exclusively on what abstract objects are not, with no thought about what they are or might be supposed to be, we can scarcely expect anything but intellectual paralysis when we try to consider how we might get to know about them. The second factor is the idea that knowledge of truths about objects of any kind must involve “contact” with those objects. If “contact” is taken literally, so as to require some sort of physical connection or interaction – perhaps of the sort that occurs in normal sense perception, or even something more indirect – the idea is obviously inimical to realism, but equally not obviously one that must be accepted. Of course, if it is given a sufficiently attenuated (and perhaps unavoidably metaphorical) construal, so that possession of any sort of identifying knowledge of an object suffices for contact, the idea reduces, near enough, to a truism – one can hardly be credited with knowledge of truths about objects unless one knows which objects are in question – and it need then cause the platonist no trouble, unless it is coupled with the further idea that such “contact” is presupposed by and must be already in place before any knowledge of truths about objects can be had (cf. Russell’s famous principle that “Every proposition which we can understand must be composed wholly of constituents with which we are acquainted” (see Russell, 1912, chs. 4, 5). Once we become locked into thinking about the access problem within this straitjacket, we can hardly avoid the further thought – that the access problem is not just a problem about how we can know anything about abstract objects, but goes wider and deeper: how can we even so much as think about them at all. Critics of realism may see this as just so much more grist to their mill: realism is in trouble on two counts, not just one, because it obstructs both a satisfactory epistemology and a workable theory of reference (cf. Benacerraf, 1973, p. 412; Field, 1989, p. 68). But this way of putting the difficulty obscures an important connection. The right way to 72

put the objection is like this: even if one could give a realist account of the truth conditions for mathematical statements (or any other class of statements supposed about abstract objects), it would be impossible to explain how such statements, so understood, could be known or reasonably believed; but in fact one cannot even give such a semantical account, since one cannot even so much as make reference to “objects” of the sort such an account takes them to be about – and if one cannot do that much, one cannot so much as state realist truthconditions. It helps to recast the objection in this way, because doing so gives a clearer view of the structure of the task that must be addressed by a defensible form of realism. The fundamental part of the access problem is not the knowledge problem (i.e., how, given that certain statements (e.g., mathematical ones) are about abstract objects, we could know them to be true), but the reference problem (i.e., how they could be about such objects in the first place). That, then, is the problem the realist should tackle first. Although solving the reference problem is merely a necessary, and not a sufficient, condition for a solution to the knowledge problem, one might expect a good solution to the former to suggest how best to approach the latter. But how, if at all, may realists solve the reference problem? In my view (for a concise statement, see Hale and Wright, 2002, sect. 5), their best hope lies in rejecting the assumption that an ability to engage in identifying reference to, or thought about, abstract objects is a precondition for understanding statements about them, as is suggested by the “contact” model and Russell’s Acquaintance Principle (see acquaintance). Positively, they should argue that concepts of kinds of abstract object may be introduced by fixing the truth-conditions of complete sentences involving terms for them, in accordance with Frege’s Context Principle (“Only in the context of proposition does a word mean anything” – cf. Frege, 1884, §62). More specifically, they may then deploy what have come to be known as “abstraction principles” as a means of explaining both how 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 known a priori. Examples are the Direction Equivalence (see above, sect. 4) and Hume’s Principle: The number of Fs = the number of Gs iff the Fs correspond one–one with the Gs Whether the access problem can be solved in this, or some other way, is a matter of currently active debate. b i b l i og rap hy Alston, W.: “Ontological Commitments,” Philosophical Studies 9 (1958), 8–17. Bell, D.: “The Epistemology of Abstract Objects,” Proceedings of the Aristotelian Society, suppl. vol. 53 (1979), 135–52. Benacerraf, P.: “Mathematical Truth,” Journal of Philosophy 70 (1973), 661–80. Bolzano, B.: Theory of Science, ed. and trans. Rolf George (Oxford: Blackwell, 1972). From Wissenschaftslehre (Sulzbach, 1837). Burgess, J. P. and Rosen, G.: A Subject With No Object: Strategies for Nominalistic Interpretation of Mathematics (Oxford: Clarendon Press, 1997). Carnap, R.: “Empiricism, Semantics, and Ontology,” Revue Internationale de Philosophie 4 (1950), 20–40. Chihara, C.: Constructibility and Mathematical Existence (Oxford: Clarendon Press, 1990). Dummett, M.: Frege: Philosophy of Language (London: Duckworth, 1973). Field, H.: Realism, Mathematics, and Modality (Oxford: Blackwell, 1989). Field, H.: Science Without Numbers (Oxford: Blackwell, 1980). Frege, G.: Die Grundlagen der Arithmetik (Breslau, Poland: Wilhelm Koebner, 1884); trans. into English by J.L. Austin as The Foundations of Arithmetic (Oxford: Blackwell, 1959). Hale, B.: Abstract Objects (Oxford: Blackwell, 1987). Hale, B. and Wright, C.: “Benacerraf’s Dilemma Revisited,” European Journal of Philosophy 10 (2002), 101–29. Hale, B. and Wright, C.: “Nominalism and the Contingency of Abstract Objects,”

Journal of Philosophy 89:3 (1992), 111– 35. Hart, W.D.: “The Epistemology of Abstract Objects,” Proceedings of the Aristotelian Society, suppl. vol. 53 (1979), 153– 65. Hart, W.D.: Review of Mark Steiner, Mathematical Knowledge (Ithaca, NY: Cornell University Press, 1975), in Journal of Philosophy 74 (1997), 118–29. Hellman, G.: Mathematics Without Numbers (Oxford: Clarendon Press, 1989). Kitcher, P.: “The Plight of the Platonist,” Noûs 12 (1978), 119–36. Lewis, D.: On the Plurality of Worlds (Oxford: Blackwell, 1986). Maddy, P.: Naturalism in Mathematics (Oxford: Clarendon Press, 1990). Maddy, P.: Realism in Mathematics (Oxford: Clarendon Press, 1980). Meinong, A.: “The Theory of Objects,” in R. Chisholm, R., ed. Realism and the Background to Phenomenology (London: Allen & Unwin 1960; originally published 1904). Noonan, H.: “Dummett on Abstract Objects,” Analysis 36:2 (1976), 49–54. Priest, G.: Towards Non-Being: The Logic and Metaphysics of Intentionality (Oxford: Clarendon Press, 2005). Putnam, H.: “Philosophy of Logic,” New York: Harper & Row, 1971; repr. in Putnam’s Philosophical Papers Volume 1 (Cambridge: Cambridge University Press 1975). Quine, W.V.: Methods of Logic (London: Routledge and Kegan Paul, 1952). Quine, W.V.: Philosophy of Logic (Englewood Cliffs, NJ: Prentice-Hall, 1970). Quine, W.V.: Word & Object (Cambridge, MA: MIT Press, 1960). Routley, R.: Exploring Meinong’s Jungle and Beyond (Canberra: Australian National University, 1980). Russell, B.: The Problems of Philosophy (London: Oxford University Press, 1912). Steiner, M.: “Platonism and the Causal Theory of Knowledge,” Journal of Philosophy 70 (1973), 57–66. Wright, C.: Frege’s Conception of Numbers as Objects (Aberdeen: Aberdeen University Press, 1983). bob hale 73

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

s p a c e and time

Space and Time This article discusses the following issues about space and time: whether they are absolute or relative, whether they depend on minds, what their topological and metrical structures may be, Mctaggart’s argument against the reality of time, the ensuing split between static and dynamic theories of time, problems with presentism, and the possibility of time travel. Our opening questions are posed in the following query from Kant: What, then, are space and time? Are they real existences? Are they only determinations or relations of things, yet such as would belong to things even if they were not intuited? Or are space and time such that they belong only to the form of intuition, and therefore to the subjective constitution of our mind, apart from which they could not be ascribed to anything whatsoever? (A23/B37) a b s ol u te o r r elative? Newton regarded space as a real existence – a vast aetherial container without walls, in which everything else that exists lives and moves and has its being. Leibniz believed to the contrary that space is not a genuine entity, but a mere façon de parler; he held that all talk of space is replaceable by talk of material things and their relations to one another. For example, to say that a thing has “changed its place” is merely to say that it has changed its distance or direction from some other thing chosen as a reference object. This is the issue that divides partisans of absolute or substantival theories of space on the one hand from defenders of relative or relational theories on the other. To test his or her allegiance on this issue, the reader should answer the following question: if the only material thing in existence were a single particle, would it make sense to say that it is moving? Leibniz would say no, since motion for him consists in change of relations (e.g., of distance) among two or more material things. Newton would say yes, since even in the absence of other 74

material things, the particle could be moving from one cell to another of space itself. Newton argued for the existence of substantival space with a famous thought experiment. Imagine a bucket suspended from a rope and filled with water. The rope is twisted and allowed to unwind, causing the bucket to spin. At first the bucket moves relative to the water, the water not yet having begun to partake of the bucket’s motion, but eventually friction causes the water to rotate as well, and indeed to “catch up” with the bucket so that there is no longer any relative motion between water and bucket. By the time this happens, something else happens as well: the surface of the water has become concave, the water edging up the sides of the bucket. This is explained in Newtonian mechanics as a centrifugal-force effect, similar to what happens when amusement park riders are pinned to the side of a rotating bottomless drum. Newton’s argument now proceeds as follows: 1. There is a time at which the water displays centrifugal-force effects, but is not moving relative to the bucket – or any other material thing. (Why not relative to the ceiling, you ask? That is why the experiment is a thought experiment: we are to imagine it performed in a universe with no objects besides bucket, water, and rope.) 2. All centrifugal-force effects are induced by rotational motion 3. Therefore, there is a time at which the water is moving, but not relative to any material thing (from 1 and 2). 4. Motion that is not relative to any material thing is absolute motion, that is, motion with respect to space itself. 5. Therefore, the water is moving with respect to space itself (from 3 and 4) – which must therefore exist. Newton thus argues that accelerated motion (the water’s constant change of direction) reveals itself in its effects and proves the existence of space, the existence of which then grounds absolute uniform (non-accelerated) motions, even though the latter do not 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 conclusion of Newton’s argument, but without making clear which premise they thought wrong. Mach objected to premise 1, claiming that we simply do not know how the water would behave in a universe devoid of ceiling and stars (as though no physicist ever extrapolated his laws to hypothetical situations, such as frictionless planes!). A generally overlooked response challenges premise 4: perhaps motion is really absolute, that is, not a change in relation to anything else at all, be it matter or space. The possibility of this last response shows that we should separate two issues that can be posed using the “absolute vs. relative” formula: is space a substance or a system of relations, and are motion, size, and various other spatial commodities absolute (intrinsic) or relational? Leibniz argued that space is a pseudoentity because its existence would generate distinctions without a difference or, more precisely, exceptions to his principle of the identity of indiscernibles. Let w and w′ be two universes just alike in how all material things are related to one another, but differing in the alleged respect that in w′ the entire material cosmos has been moved six miles to the east or rotated through some angle. Leibniz’s argument then proceeds as follows: 1. If there were such a thing as substantival space, w would be distinct from w′. 2. But w and w′ are indiscernible – they share all their properties. 3. Things that are indiscernible are identical. Putting it the other way around, any two distinct things must differ in at least one property. 4. Hence, w = w′ after all (from 2 and 3). 5. Therefore, there is no such thing as substantival space (from 1 and 4). To evaluate this argument, we need to distinguish two kinds of properties. A property is pure if its being exemplified does not depend on the existence of any specific individual and impure otherwise. Examples of pure properties are being red (which is pure and intrinsic) and being next to something red (pure and relational); examples of

impure properties are being Fred (impure and intrinsic) and being married to Fred (impure and relational). When Leibniz affirms premise 2, he must mean that w and w′ differ in no pure property, for Newtonians could certainly maintain that w and w′ are distinguished by the fact that w is such that part of the cosmos occupies cell 233 (an impure property), whereas in w′, cell 233 is empty. But that means when we get to premise 3, Leibniz must advance his Identity of Indiscernibles principle in the following form: any two things must differ in at least one pure property (and not merely in such properties as being identical with this thing). Leibniz no doubt did wish to affirm the principle in the required form, but if so, it is open to counterexamples. Is it not conceivable that there be two spheres the same in color, shape, composition, and every other pure property you care to think of? The substantival vs. relational issue carries over to time. For Newton, time “flows equably without regard to anything external;” for Leibniz, time is nothing over and above the sequence of events said to be in time. Newton (but not Leibniz) can make sense of the idea that the entire history of the world (comprising the same events as now) might have begun earlier than it did. real or i deal ? Another issue about space and time is whether they are ideal, that is, dependent for their existence on minds. The most famous idealist about space and time in western thought is Kant. Kant began his intellectual career as a Leibnizian, but was briefly converted to Newton’s view by considerations about “incongruent counterparts” – objects that come in mirror image forms, like left and right human hands. Kant thought the difference between incongruent counterparts could not be explicated using only relationist resources, but had to consist in the differing relations of the objects to space itself. By the time he wrote the Critique of Pure Reason, however, Kant had come around to the third of the positions in the quotation above: space and time are merely “forms of intuition,” that is, ways in which human 75

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 perceive; they are not features of things in themselves, or things as they exist outside the mind. A characteristically Kantian reason for believing that space is ideal is that no other hypothesis accounts for our knowledge of geometry. Kant thought that geometry was a body of synthetic and a priori truth – a priori in that it is known in advance of experience, yet synthetic in that it is not validated just by logic or the meanings of our concepts. How can that be? How can we know even before we encounter them that cubes on Mars will have 12 edges? Kant’s answer is that (i) our form of intuition makes us incapable of intuiting (perceiving or imagining) any cubes that do not have 12 edges and (ii) as prescribed by idealism, no cubes or spatial objects exist anywhere except those that satisfy the conditions of our intuiting them. Thus all cubes everywhere have 12 edges and the other properties imposed on them by our Euclidean form of intuition. Kant thought the ideality of space and time was further confirmed by the antinomies – pairs of opposed propositions in which one or the other must be true if space and time exist outside the mind, but both of which are impossible. For example, does the world have a beginning in time, or is it infinite in its past duration? If things in time were things in themselves, one of these alternatives would have to be true, yet both of them boggle the mind. No beginning would mean an infinity of events already elapsed, which Kant thought impossible because it would involve a “completed infinity.” (Think of Wittgenstein’s example of the man we find saying, “. . . −5, −4, −3, −2, −1; whew! I just finished counting through all the negative integers.”) A beginning would mean an event for which there could not possibly be a sufficient reason – a blow to rationalist aspirations, if not the outright impossibility Kant seemed to think it was. Kant’s solution was to hold that past events exist only in present or future memories or other evidence of them (for example, yet-to-be-perceived cosmic radiation). He thought this opened the possibility that the world’s history is 76

potentially infinite – always extendable further into the past through our future discoveries – but neither actually finite nor actually infinite. struc t ural quest i ons The next group of questions about space and time (or spacetime, in the Minkowskian melding of them) concerns their (or its) metrical and topological structure. Are space and time infinitely divisible, or are there smallest units? (Zeno’s paradoxes of motion are sometimes seen as set up so that the first two apply if space and time are infinitely divisible and the second two if space and time are quantized.) Does space obey the laws of Euclidean geometry or those of one of the non-Euclidean geometries known to be consistent since the nineteenth century? How many dimensions does space have? Could time have a beginning or an end? Must time be unilinear, or might it branch into multiple paths or close back upon itself in a loop? The dimensionality of space is representative of such questions. We all know about three dimensions of space – a line possesses one dimension, a plane two, and a solid three. What would it mean for space to have a fourth dimension? (We are talking now of a fourth spatial dimension, not time, even though time is sometimes considered as a fourth dimension.) Galileo offered one criterion: to say that space has n dimensions is to say that n mutually perpendicular lines (but no more) can meet in a single point. If our space were four-dimensional, a line could enter the corner of my desktop at right angles to each of its three edges. Poincaré offered another criterion: points are zero-dimensional, and an entity is ndimensional iff n is the lowest number such that any two points of the entity may be separated from each other by an entity of n – 1 dimensions. Thus, a line has one dimension, because any two points of it can be separated from each other by an intervening entity of zero dimensions (another point); a plane has two dimensions, because any two points within it may be separated 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 between them; and so on. It is a consequence of this criterion that in a four-dimensional space, a two-dimensional entity would not suffice to separate one point from another. Thus a spherical shell enclosing point A but not point B would not suffice to separate A from B – you could get from A to B without penetrating the shell. Such things defy visualization in a way that makes some people want to declare them impossible. Those so inclined should read E.A. Abbott’s Victorian classic Flatland, in which the author describes a world of two-dimensional beings who are incapable of rising out of their plane or visualizing anything beyond it. A Flatlander may be imprisoned simply by enclosing him within a circle or a polygon. Could a Flatlander but jump over the walls of his prison, he would be free, but he is incapable even of conceiving such a motion – as we are of any path from the interior to the exterior of a spherical shell that does not pass through the shell. The exhortation “Upwards, not northwards!” falls on the Flatlander’s ears as nonsense. Abbott’s intent, of course, is to soften us up for the possibility that our own resistance to a fourth dimension may be as provincial as that of the Flatlanders to a third. Questions about the structure of space and time give rise to meta-questions about proper jurisdiction – who is to answer them, and how? A traditional view is that space and time necessarily possess whatever structure they do, and that it ought to be ascertainable a priori what this structure is. Kant, for example, certainly believed that space is necessarily three-dimensional and Euclidean. The prevalent contemporary view is that space and time have their structures contingently, and that it is only through the best science of the day that we can reach any reasonable opinion concerning what these structures are. This view was given impetus by Einstein’s use of a non-Euclidean geometry in conjunction with the General Theory of Relativity to explain gravitation; it is further exemplified in the work of those physicists in search of a “theory of everything” who posit a space of 11 dimensions.

A view that lies between the traditional and the contemporary views is the conventionalism of Poincaré. Poincaré thought that all the empirical data accommodated by nonEuclidean geometry plus standard physical theory could equally well be accommodated by Euclidean geometry together with non-standard physical theory. For example, measurements apparently indicating that the ratio of circles to their diameters does not have the familiar value of π could be accommodated by a non-Euclidean geometry in which this ratio is indeed other than π, but they could also be accommodated by positing a heat gradient that causes our yardsticks to expand when laid along the diameter though not when laid along the circumference. We could thus always choose to describe our world in Euclidean terms by complicating our physics. This position is at odds with a hardy empiricism, in so far as it denies that empirical results can settle the structure of space, but it is also at odds with an ambitious a priorism, in so far as it denies that decisions in favor of Euclid are determinations of independent fact. questions about t i me For issues specifically about time, the best point of departure is McTaggart’s famous argument of 1908 that time is unreal. Though few have accepted the conclusion of this argument, nearly all students of time have taken over the distinctions McTaggart employed in formulating it. McTaggart’s fundamental distinction is between the A-series and the B-series. An A-series is a series of events or moments possessing the characteristics of being past (in varying degrees), present, or future; call these the A-characteristics. The B-series is a series of events or moments standing in the relations of earlier-than, later-than, and simultaneous with; call these the Brelations. The chief difference McTaggart notes between the A-characteristics and the B-relations is that the former are transient while the latter are permanent: “If M is ever earlier than N, it is always earlier. But an event, which is now present, was future, and will be past” (LePoidevin and MacBeath, 77

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 thinking about time, McTaggart believes, an event becomes increasingly less future, is momentarily present, and then slides ever farther into the past. Yet all the while its B-relations to other events (e.g., its following the Battle of Waterloo and preceding the first landing on the moon) are fixed. McTaggart’s overall argument against the reality of time may be stated quite briefly: (I) time essentially involves an A-series; (II) any A series involves a contradiction; therefore, (III) therefore, time is unreal. Behind each main premise is a subsidiary argument. The argument behind premise I is this: 1. There can be no time without change. 2. There can be no change without an Aseries. 3. Therefore, there can be no time without an A-series. Both premises in this argument have been the subject of interesting debate, but our focus here will be on the argument behind main premise II, which runs thus: 1. The A-characteristics are mutually incompatible, yet 2. Every event in any A-series must have all of them, so 3. Any A-series involves a contradiction. McTaggart immediately anticipates an objection the reader will have to premise 2: it is not true that any event must have all the A characteristics at once, but only that it must have them successively. An event that is now present is not also past and future; rather, it was future and will be past. In reply, McTaggart claims that this attempt to avoid the contradiction he alleges only raises it anew. What, he asks, is meant by tensed verb forms such as “was” and “will be”? His answer may be given in the schema S {was, is now, will be} P iff for some moment m, S has P at m & m is {past, present, future} where the italicized verbs are meant to be tenseless. He thus believes that tense can be reduced to A-characteristics and tenseless 78

copulas. If this is right, then in saying that an event has been future and will be past, we are introducing a new A-series, this time of moments. And this brings back our contradiction, because every moment, like every event, is past, present, and future. If we try to get rid of the contradiction by saying of moments what we said earlier about events, our statement “means that the moment in question is future at a present moment, and will be present and past at different moments of future time. This, of course, is the same difficulty over again. And so on infinitely” (LePoidevin and MacBeath, 1993, p. 33). Why is McTaggart so convinced that there is a contradiction in the A-series and a regress in any attempt to remove it? His thought on these matters can be made more understandable by presenting it with the help of a metaphor. He begins by supposing that the whole of history is laid out in a block comprising the B-series. He notes that in such a series, there is no change and therefore no time, all events simply sitting there alongside one another on the B-axis. What can add time to such a universe? We must bring in the Acharacteristics, letting the spotlight of presentness wash along the series in the direction from earlier to later. But wait! If the spotlight illuminates event e before it illuminates event f, then the events of e’s being present and f’s being present are both there on the B-axis, permanently related by the relation of earlier-than. Similarly, if the shadow of pastness falls on e before it falls on f, then e’s being past and f’s being past permanently stand in the B-relation of earlier-than and are thus always there on the B-axis. What we are saying implies that that e and f are both always past and always present – surely a contradiction, just as McTaggart alleges. If we seek to remove the contradiction by saying that the spotlight of the present falls on e’s being present before it falls on f’s being present, we are only embarking on a useless regress – again just as McTaggart alleges. As noted above, few besides McTaggart have accepted his argument in toto, but 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 philosophy of time. One side accepts his first main premise while rejecting the second: the A-characteristics (or some surrogate for them) are indeed essential to time, but there is nothing wrong with that. The other side accepts his second main premise while rejecting the first: there is indeed a defect in the A-series, but a B-series by itself is all you need to have time. For obvious reasons, these two responses to McTaggart are often

called “the A theory” and “the B theory,” though the names can be misleading. There is an entire cluster of doctrines that tend to go together under the banner of the A theory and another opposing cluster under the banner of the B theory. (Other labels for the two sides are the dynamic versus the static theory and the theory of passage or becoming versus the theory of the four-dimensional manifold.) The rival doctrines may be tabulated as follows:

The A Theory (Dynamic Time)

The B Theory (Static Time)

A1. Tense is an irreducible and indispensable feature of thought and language, reflecting a genuine feature of reality. Corollary: some propositions change in truth value with the passage of time.

B1. Tense is reducible or eliminable; reality is adequately describable without it.

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

B2. The A-characteristics are either delusive or reducible to the B-relations.

A3. The present is ontologically privileged: things present have a reality not belonging to things past or future.

B3. Past, present, and future are ontologically on a par: things past and future are no less real than things present.

A4. The future is open or indeterminate: some propositions about what is going to happen in the future are not yet either true or false.

B4. The future is as fixed as the past; every proposition must be true or false, and propositions have their truth values eternally (as noted in B1).

A5. Identity through time is endurance: numerically the same thing exists at many distinct times.

B5. Identity through time is perdurance: a thing that lasts through time is a series of distinct temporal parts or stages, united by some relation other than identity.

In row 1, we have the debate between those who take tense as primitive and those who seek to reduce it to something else (as Smart once did when he suggested that “it will rain” just means “rain occurs later than this utterance”.) In row 2, we have the debate between the A theory proper and the B theory proper, which is sometimes too quickly equated with the debate in row 1. (Arguably, tenses are not equivalents of the A- characteristics, but superior substitutes for them.) In row 3, we have the issue that divides presentists from eternalists – those like Augustine, who laments that

Corollary: every true proposition is timelessly true.

his boyhood is no more, and those like the Tralfamadorians in Vonnegut’s Slaughterhouse Five, who do not cry at funerals because their departed loved one exists and breathes at an earlier moment. In row 4, we have an issue that goes back to Aristotle’s discussion in De Interpretatione: must the proposition the captain will order a sea battle tomorrow be true or false today, and if so, does that mean the future is in some way fixed or fated? Finally, in row 5 we have the issue (stated in David Lewis’s terms) that divides those who believe in genuine continuants from those who accept an analysis of 79

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, who once observed that “each of us proceeds through time only as a fence proceeds across a farm” – that is, by having different parts at different moments or regions (Williams, 1951, p. 463). As noted, a philosopher who holds a view in one of the columns will tend to hold the other views in that column as well. There is a certain amount of room for mixing and matching, however, and it should not be assumed automatically that the propositions in a given column must go together as a package deal. Indeed, no one should hold all of the propositions in column A, for a little reflection shows that A2 is inconsistent with A3. If presentism is true, there are no things or events that are not present, and thus no items possessed of pastness and futurity. So if A3 is true, A2 is false. The best combination among A1–A3 for a friend of dynamic time is arguably A1 and A3 without A2. Ironically, this would be an “A theory” without the A-characteristics, so the common name is not well chosen. McTaggart’s combination was just the opposite, and this is arguably what led to the demise of time in his philosophy. His argument depends on reducing tense to the A-characteristics, and it also depends on making the eternalist assumption that the earlier and later portions of the B-series are equally real. A presentist could evade the argument by denying that an event is there before it becomes present; rather, the event simply becomes – it comes into being and then as quickly passes out of being. Or better yet (since an ontology of things goes better with presentism than an ontology of events), a thing becomes F and then is no longer F. The issue debated in rows 1 and 2 is sometimes put this way: does time pass, or is there simply a huge four-dimensional manifold with time as one of its dimensions? Some philosophers think the passage view may be refuted by asking a simple question: how fast does time pass? If the first second of the year 2050 is getting closer to us, there must be a rate at which it is doing this, yet any way of assigning the rate would be nonsensical or absurd. Are the seconds 80

going by at the rate of one second per second? That is no rate at all. One second per hypersecond? That takes the first step in a preposterous series of time orders. So time does not pass. When the argument is formulated that way, it presupposes a substantival theory of time – as though there were drops of time passing through an hourglass. Perhaps, then, the argument can be sidestepped by combining belief in dynamic time with a rejection of substantival time. Such is the combination espoused by Arthur Prior, the founder of tense logic. Prior represents tenses with operators, akin to modal operators: “Peter will sneeze” becomes “It will be the case that Peter sneezes”, symbolizable as Fp, and “Peter sneezed” becomes “It was the case that Peter sneezes”, symbolizable as Pp. The present tense is the default tense and needs no operator. With this apparatus, it is possible to articulate many propositions about the structure of time. For example, the density of time may be expressed as (p)(Fp → FFp). This formula would not be true if time were discrete, for if there were an immediately next moment and a proposition p true at it but not thereafter, Fp would be true and FFp false. Prior denies that time is a literal object, “a sort of snake which either eats its tail or doesn’t, either has ends or doesn’t, either is made of separate segments or isn’t;” rather, these issues can be formulated using propositional variables and tense operators in a way that makes no reference to time or its parts (Prior, 1968, p. 189). Returning now to the question of time’s passage, Prior suggests that the metaphor can be cashed out in tense logic as follows: there are true instances of the schema Pp & ~p – it was the case that p, but is not now the case that p. When the matter is put that way, it is no longer obvious how awkward questions about the rate of time’s passage are to be formulated. probl ems f or present i sm Presentism is easily misunderstood. Presentists are not holocaust deniers; their 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 about what happened using tense operators. Nonetheless, presentism is not without its problems. Are there not past tense truths about individuals who no longer exist, for example, that Lincoln was wise and wore a beard? But how can there be such truths if Lincoln no longer exists to be a constituent of propositions about him? On this question, Prior bites the bullet and says there are no singular truths about objects that no longer exist, but only general truths – it was once the case that there was a man who was President during a civil war, etc., and who wore a beard. Other presentists find some presently existing entity for past-tense truths to be about – for example, the haecceity being Lincoln, a property that exists even if Lincoln does not, and which was formerly co-instantiated with the property of being wise. What some regard as the fatal blow for presentism comes from the Special Theory of Relativity. The theory is often presented as resting on two postulates, the relativity of uniform motion and the constancy of the speed of light. Uniform motion is motion at a constant speed in a constant direction. The first postulate tells us that no experiment can determine that an object is in a state of absolute uniform motion, from which it is often concluded that it makes no sense to ascribe uniform motion. (If two objects are moving uniformly relative to each other, it is as correct to say that one is moving and the other at rest as vice versa.) The second postulate tells us that whether an observer is moving towards or away from a beam of light, the light’s speed with respect to the observer will be the same. Einstein showed that when these two postulates are combined, many surprising consequences follow, including the relativity of simultaneity: two events that are simultaneous in one observer’s frame of reference may be successive in another’s frame, with no way of saying that either frame is uniquely correct. Putnam has offered an argument against presentism based on Special Relativity and two other assumptions. One assumption (which Putnam calls the principle of “no privileged observers”) is that what is real

for you is real for me, assuming that you are real for me. This may be expressed equivalently as the assumption that the relation of being real-for is transitive: 1. If x is real for y & y is real for z, then x is real for z. Putnam’s other assumption is that in the context of Special Relativity, the presentist’s core thesis that x is real iff x is present should be reformulated as “x is real for y iff x is present for y” and the latter in turn as “x is simultaneous with y in the frame of y”: 2. Presentism implies: x is real for y iff x is simultaneous with y in the frame of y. From 1 and 2, it follows that for presentists, the simultaneity relation we have just mentioned is transitive: 3. Presentism implies: if x is simultaneous with y in the frame of y & y is simultaneous with z in the frame of z, then x is simultaneous with z in the frame of z. According to Special Relativity, however, 4. The relation in 3 (which Putnam calls “simultaneity in the observer’s frame”) is not transitive. That is because if you pass right by me at a high relative speed, there will be events simultaneous with you in your frame that are not simultaneous with me in my frame, even though at the moment of passing, you are simultaneous with me in my frame. Putnam concludes that presentism is false, and that I should acknowledge as real events belonging to your present even though they do not belong to mine. If presentists do not wish to accept this conclusion, how should they respond to Putnam’s argument? There are three main options. One is to reject the transitivity of the real-for relation, as advocated by Sklar; in effect, this is to make reality itself as relative as simultaneity. A second is to reject Putnam’s construal of “x is present for y” as “x is simultaneous with y in the frame of y”; alternative relativistic reconstruals of the present-for relation have been canvassed by Hinchliff and Sider. The third is to question Special Relativity, as has been done by 81

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 an audacious denial of physics to make room for metaphysics, but it need not be that. It will probably not have escaped the reader’s notice that insofar as Special Relativity says there is no such thing as absolute uniform motion – not just that it is undetectable by any experiment – it ventures beyond physics into philosophy. One who questions the theory may be questioning its verificationist auxiliary assumptions rather than anything that physics alone can teach us. i s t i m e tr avel p o ssib le? This question turns in part on the issues in rows 3, 4, and 5. The physics of the last century is sometimes thought to imply an answer of yes, for two main reasons. First, the Special Theory of Relativity is sometimes thought to imply eternalism, as discussed above, and the eternalist view encourages us to take time travel seriously. If the assassination of JFK is there, several decades prior to us on the time line, why couldn’t we go there and witness it? (Conversely, presentism is sometimes thought to rule out time travel, on the ground that if the past and the future are not there, there is literally nowhere to go.) Second, the General Theory of Relativity is now believed to imply the possibility of closed timelike curves, which might be exploited by time travelers. Einstein’s field equations enable one to calculate the spacetime structures induced by various configurations of matter, and in 1949, Gödel showed that there are possible configurations of matter that would generate closed timelike curves – temporal paths along which an event can precede other events which precede itself. An object part of whose lifeline lay along such a curve could (in a sense) visit its own past. Interestingly, Gödel’s own conclusion from his discovery was quite different: he thought real time could not violate the irreflexivity of precedence, so he took the possibility of loops in time to show that time is ideal in something like Kant’s sense. If permitted by physics, travel to the past may nonetheless be forbidden by logic or 82

metaphysics. An entrenched axiom is that no one can change the past. If we could travel to the past, why could we not change it, even in paradoxical ways such as by killing one’s grandfather or infant self? Science fiction writers sometimes take pains to have their characters leave the past undisturbed; for example, they view dinosaurs from magically suspended walkways so as to leave no footprints. But of course the mere presence of the time traveler as an observer would constitute a change in the past if he had not been there the “first” (and only) time around. Therefore, in consistent time travel tales, the traveler “always” made his visit – the visit does not change the past, but was always part of it. (As Lewis has it, a temporal stage of the traveler was permanently present at the scene. Lewis’s stage view explains how it is possible for the traveler to interact with his infant self: such interaction occurs between stages of the same person that are contemporaneous in “external” time but one later than the other in “personal” time.) Because his actions are already woven into the past, a time traveler cannot kill his grandfather or his infant self; in history as it was, grandfather lived and the traveler failed to kill him, if he tried. This way of preserving the past from change may arouse fears of fatalism. If in fact grandfather lived to sire my father, am I not fated to fail in my attempts to kill him? And if in history as it happened, I emerged from a time machine in 1920 that I enter (entered? will enter?) in 2020, am I not fated to enter the time machine in 2020, or at least at some time? To do otherwise would be to do something at variance with past truth. In reply, some argue that timetravel arguments for fatalism add nothing to more general arguments for fatalism based on applying the law of bivalence to the future, such as the following: 1. It was either true yesterday that I would push the nuclear button tomorrow or true yesterday that I would not. 2. In the former case, I must push the button tomorrow 3. In the latter case, I must not push it. 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 worries is that all that follows from the supposition that it was true yesterday that I would push the button tomorrow is that I will push it, not that I must. It could be maintained similarly that although in 2020 I certainly will enter the time machine from which I emerged in 1920, it is not true that I must. So a good case can be made that time travel imposes fatalistic constraints on time travelers only if Aristotelian arguments from bivalence impose fatalistic constraints on us all. So which is it, freedom for time travelers or fate for us all? Space and time do not permit an answer to this question here. See also the a–z entries on antinomies; change; continuant; fatalism; principle of verifiability; smart, j.c.c.; space and time, temporal parts; zeno of elea. b i b l i og rap hy Abbott, E.A.: Flatland (New York: Dover, 1952; originally published in 1884). Broad, C.D.: Scientific Thought (New York: Harcourt, Brace, 1920). Hinchliff, M.: “A Defense of Presentism in a Relativistic Setting,” Philosophy of Science 67, suppl. (2000), S575–86. Kant, I.: Critique of Pure Reason, trans. N. Kemp Smith (New York: St. Martin’s Press, 1965; originally published in 1781). LePoidevin, R. and MacBeath, M., ed.: The Philosophy of Time (Oxford: Oxford University Press, 1993). Lewis, D.: “The Paradoxes of Time Travel,” American Philosophical Quarterly 13 (1976), 145–52; repr. in LePoidevin and MacBeath (1993). McTaggart, J.: “The Unreality of Time,” in The Nature of Existence, vol. II (Cambridge: Cambridge University Press, 1927), ch. 33; repr. in LePoidevin and MacBeath (1993). Markosian, N.: “How Fast Does Time Pass?” Philosophy and Phenomenological Research 53 (1993), 829–44. Putnam, H.: “Time and Physical Geometry,” Journal of Philosophy 64 (1967), 240– 7.

Prior, A.N.: Past, Present, and Future (Oxford: Clarendon Press, 1967). Reichenbach, H.: The Philosophy of Space and Time (New York: Dover, 1958). Sider, T.: Four-Dimensionalism: An Ontology of Persistence and Time (Oxford: Oxford University Press, 2001). Sklar, L.: Space, Time, and Spacetime (Berkeley: University of California Press, 1974). Contains discussions of Newton, Leibniz, and Poincaré. Sklar, L.: “Up and Down, Left and Right, Past and Future,” Noûs 15 (1981), 111– 29; repr. in LePoidevin and MacBeath (1993). Van Cleve, J.: “If Meinong Is Wrong, Is McTaggart Right?” Philosophical Topics 24 (1996), 231–54. Van Cleve, J. and Frederick, R., ed.: The Philosophy of Right and Left: Incongruent Counterparts and the Nature of Space (Dordrecht: Kluwer, 1991). Willliams, D.C.: “The Myth of Passage,” Journal of Philosophy 48 (1951), 457–72. Yourgrau, P.: Gödel Meets Einstein (Chicago: Open Court, 1999). james van cleve

Substance I – Introduction In one metaphysically salient sense of the term “substance”, a substance is an individual thing. From a commonsensical perspective, it appears that the extension of “substance” in this sense includes inanimate material objects, e.g., pieces of gold, mountains, and statues, as well as living things, e.g., people, frogs, and trees. (Note that since a compound substance is a unified whole, its parts must stand in some sufficiently robust unifying relation to one another, e.g., some appropriate causal or functional relation; if there are simple (or basic) substances, they do not have any detachable parts, see part/whole.) A belief in the existence of such individual substances is at core of our “folk ontology”. Moreover, various scientific theories seem to be committed to their existence. The 83

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 prominently in Aristotle’s seminal work in metaphysics and in much subsequent important work in the field. It is this concept that is the focus of this essay. Aristotle’s term “primary ousia” has often been translated as substance (or as primary substance) a practice which has caused considerable confusion. This translation can be misleading, since although one ordinary meaning of “substance” is an individual thing, e.g., an inanimate material object or living organism, this is not what Aristotle means by “primary ousia”. A more accurate and less misleading translation of “primary ousia” is primary being (or fundamental entity, or basic entity). In the Categories Aristotle argued that the primary beings are individual things, e.g., living things, and that essences are secondary beings. However, in the later work, the Metaphysics, he changed his view about primary beings, and seems to have concluded that that the primary beings are forms, rather than individual things. In the Metaphysics, Aristotle famously conceived of an individual thing as, in some sense, a combination of form and matter (see matter/form). Even if there exists a technical usage of the term “substance” in which it means primary being, this is a different meaning than the more ordinary sense, that of individual thing. But, according to another ordinary sense of the term “substance”, a substance is a quantity of material stuff of some kind, e.g., a quantity of gold, iron, oak, or lamb. But it is one thing to say that there exists a quantity of material stuff of some kind, and quite another to say that there exists an individual substance, even if this individual substance is composed of a quantity of stuff of the kind in question. For example, it is one thing to say that Mary has 50 pounds of lamb, and quite another to say that Mary has a lamb that weighs 50 pounds. After all, a lamb necessarily possesses a certain form and unity which a quantity of lamb need not possess. Furthermore, it seems possible for there to be an individual substance which has no proper parts, e.g., a non-physical soul or a point-particle; yet, the existence of individual things of these sorts does not 84

entail the existence of a quantity of material stuff of some kind. The existence of individual substances other than inanimate material objects and living organisms is controversial. However, allowing for the possibility of such substances, including non-physical substances, it is extremely plausible that any conceivable substance is either spatially extended, spatially located, or living (in a broad intuitive sense of “living”). For example, spatially unextended or spatially un-located substances which have thoughts, e.g., Cartesian souls, would qualify as living in virtue of their having mental life, even if they lack biological or physical life (see soul), whereas apparently immaterial physical objects such as point-particles and mass-less extended physical objects would have spatial location and/or spatial extension. (Hence, given the highly plausible assumption that, necessarily, life is either a physical process or a mental one, it is extremely plausible that any conceivable substance either has spatial extension, spatial location, or thought.) According to Spinoza, there exists one and only one individual substance, identical with the universe, and this substance is neither a physical substance nor a Cartesian soul; still, in Spinoza’s view, this substance has both thought and spatial extension. II – The Analysis of Substantiality In this section, we shall elucidate what we mean by an analysis of the concept of an individual substance, and then discuss the important further notion of the degree to which a philosophical analysis is ontologically neutral (see analysis). We begin with what we mean by an analysis or analytical definition of a concept or attribute, F-ness. Such an analysis provides a set of conditions, SC, such that: (i) an item’s (x’s) satisfying SC is logically or metaphysically necessary and sufficient for x’s being F, and (ii) necessarily, if x is F, then x’s being F can be explained by x’s satisfying SC. In this sense, it can be said that an analytical definition of F-ness explicates F-ness. However, if being F is a part of being C, then x’s being F cannot be explained by x’s

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

subst anc e satisfying SC on pain of vicious circularity. In such a case, the proposed analytical definition of F-ness is fatally flawed; e.g., the proposal to explicate what is just as what conforms to just laws suffers from this sort of flaw. Circularity of this kind is vicious because nothing can be explained by itself. Hence, necessarily, any purported or candidate analytical definition that involves this sort of conceptual circularity fails to satisfy condition (ii) for being an analytical definition, above, and should be rejected. Applying this schema to substantiality, let F be replaced by substance. It follows that in order to provide an analysis of being a substance, an analytical definition must provide a set of conditions, SC, such that (i) an item’s (x’s) satisfying SC is logically or metaphysically sufficient and necessary for x’s being a substance, and (ii) necessarily, if x is a substance, then x’s being a substance can be explained by x’s satisfying SC. A further important feature of philosophical analyses is to degree to which they are ontologically neutral. The following Principle of Ontological Neutrality clarifies this notion: (PON) An analysis, A, is ontologically neutral with respect to an ontological kind K (or to an entity E) =df. The adequacy of A does not entail either that Ks exist or that Ks do not exist (or that E exists or that E does not exist). By the adequacy of an ontological analysis, we mean that the analysis does not conflict with the data for that analysis. For example, if one were trying to analyze what a concrete entity is, then one’s analysis should imply that what intuitively are concrete entities are concrete, and that what intuitively are not concrete entities are not concrete. (We shall ignore here the more complicated situation that arises when no analysis can be formulated that is in this sense adequate to the data, so that we have to choose among proposed analyses none of which is entirely adequate.) It follows from PON that if in order to be adequate, a given analysis entails, for example, that universals do or do not exist, or that Cartesian souls do or don’t exist, or that God does or does not exist, then it is not ontologically

neutral with respect to universals, or to Cartesian souls, or to the existence of God. If an alternative analysis does not have these entailments, and so is ontologically neutral with respect to universals, souls, and God, then, to that extent, the second analysis is more ontologically neutral than is the first analysis. Of course, it may be the case that comparisons between competing analyses are not completely straightforward. It may happen, for example, that analysis A1 is ontologically neutral with respect to Fs and Gs, and not with respect to Ms and Ns, while analysis A2 is ontologically neutral with respect to Ms and Ns, but not with respect to Fs and Gs. Many other permutations are possible. But at least sometimes, we will be able to say that one analysis is more ontologically neutral than another. In any case, one should be aware of the sorts of ontological commitments assumed by any analysis. It is plausible to say, we believe, that the more ontologically neutral an analysis is, the better; more precisely, that all other things being equal, analyses having a higher degree of compatibility with the existence of entities of various Categories are to be preferred, so long as the entities in question are not known to be unintelligible, and plausible views about the nature, existence conditions, and interrelationships of entities belonging to those categories are assumed. Why should this be so? Because which kinds of entities, and which entities, actually or possibly exist, is often a matter of philosophical controversy. Witness the eternal debate over the existence of universals between realists and nominalists. Hence, if one can analyze, say, the concept of substance, without thereby being committed either to the existence or non-existence of universals, then that is preferable, other things being equal, to analyzing this concept in such a way as to be committed to the existence or non-existence of universals. This principle about ontological neutrality seems to us just to be a special case of Ockham’s Razor (see ockham). It also seems to us likely that there are further principles for evaluating the ontological neutrality of philosophical analyses, but we shall not 85

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 them in this article. In section IV, we shall defend a version of an independence analysis of the concept of substance which is ontologically neutral with respect to a large variety of metaphysical entities: absolute and relational space and time, space-time, universals, tropes, sets, numbers, propositions, events, boundaries, privations (see space and time; universals; trope; class, set, collection; proposition, state of affairs; event theory; boundary) and, among substances, living organisms, atoms, artefacts, and so forth. Other contemporary philosophers have offered competing versions of an independence analysis of substance, for example, Lowe (2006) and Chisholm (1996). Could there be more than one adequate analysis of substance? We see no a priori reason to rule out such a possibility. One measure of acceptability, however, and one that ought not to be ignored, but is often ignored, is the degree to which such competing analyses are ontologically neutral. III – Historical Views of Substance The concept of an individual substance, thing, or object has held a very prominent place in the history of metaphysics, perhaps because it holds such a prominent place in our ordinary conceptual scheme. In this section, we shall survey several important approaches to analyzing the notion of an individual substance. Among substance realists, there are independence, inherence, change, and substratum theorists. Also important to consider are those who would reduce substances to items belonging to some other ontological category, and those who argue for their elimination altogether. Aristotle, in the Categories, offers this account of substance in terms of change: It seems most distinctive of substance that what is numerically one and the same is able to receive contraries. In no other case could one bring forward anything, numerically one, which is able to receive contraries. (Complete Works, Vol. I, p. 7) 86

A sympathetic reading of this attempt to analyze substance is that Aristotle is saying that among entities, only individual substances are able to persist through intrinsic change. Hence, Aristotle’s analysis of substance in terms of change should be understood as follows: (D1) x is a substance =df. x is capable of persisting through intrinsic change. In the Categories, Aristotle lists other categories of being, for example, times, places, qualities, relations, and kinds. Note that it does not seem plausible that such entities cannot persist through relational change, as Aristotle appears to have noted. For example, at one moment a particular place might be occupied by a body, while at another time not. However, it does seem to be the case that entities of these sorts cannot persist through intrinsic change, since they cannot undergo intrinsic change (see extrinsic/intrinsic). Nevertheless, there seem to be at least two fairly plausible counterexamples to D1. The first is an atomic body, that is, a physically indivisible body. Such substances do not seem capable of undergoing intrinsic change – indeed, that was one of the reasons for the first atomists, Democritus and Leucippus, to postulate such beings (see atomism; presocratics). Current atomic theory also regards its fundamental particles in this way. Thus, if intrinsically unalterable atoms are possible, then D1 fails to provide a logically necessary condition for something’s being a substance. The second counterexample to D1 is provided by boundaries. For example, when a rubber ball bounces, its surface changes its shape. Hence, if there are things like surfaces, and surfaces can undergo intrinsic change, then D1 fails to provide a logically sufficient condition for something’s being a substance. Each of the preceding counterexamples to D1 involves a kind of entity that Aristotle did not include in his ontology. Hence, Aristotle could reply that there are no such counterexamples. This points out how Aristotle’s D1 is not an ontologically neutral analysis of the concept of substance: it is not 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 atomic, intrinsically unchangeable bodies, or of boundaries such as surfaces. Especially in the former instance, this seems to be a serious problem for D1. Aristotle provides a second account of substance in the Categories: A substance – that which is called a substance most strictly, primarily, and most of all – is that which is neither said of a subject nor in a subject, for example, the individual man or the individual horse. (Complete Works, Vol. I, p. 4) This account of substance, then, seems to analyze the notion of substance as follows: (D2) x is a substance =df. x can be neither said of nor in a subject. The basic idea behind D2 is supposed to be that individual things or substances do not stand in certain relations of dependence to other things, while things in other ontological categories do stand in certain dependence relations to (at least) substances. For example, Aristotle thinks that in the proposition, Socrates is a man, the kind, Man, is said of Socrates, implying that Man depends in some sense on Socrates. He also thinks that in the proposition, Socrates is hungry, the quality, Hunger, is in Socrates, implying that Hunger depends in some sense on Socrates. One problem for the idea that D2 establishes that substances possess a unique kind of independence can be seen by looking at the said-of relation. According to Aristotle, what can be said-of substances are kinds (which Aristotle also calls “secondary beings”), that is, the species and genera under which a substance falls. And given his theory of universals, no substance-kind exists unless it is instantiated by one or more substances. Hence, given Aristotle’s ontology, the existence of a substance-kind entails the existence of a substance, so that it might be said that substance-kinds depend on substances. On the other hand, no substance can exist unless it instantiates certain substance-kinds, so it also might be said that substances depend on substancekinds. Thus, it is not at all clear that the

asymmetry of the said-of relation, whereby substance-kinds are said-of substances, but not vice versa, establishes the intended asymmetry of dependence that Aristotle has in mind, whereby substance-kinds depend on substances, but not vice versa. Similar difficulties attend the claim that, because certain beings are “in” substances, such beings asymmetrically depend upon those substances. Furthermore, it is not clear that on any reasonable understanding of the in-relation employed in D2, substances cannot be “in” anything. For example, it seems perfectly natural to assert that a particular body is “in” space and time. Aristotle’s attempt to analyze the concept of substance in terms of the said-of-relation and the in-relation seems to have arisen from certain grammatical features of proper names for individual substances. Such terms can function only as subjects in sentences, and never as predicates. That this fact about grammar can be used somehow to analyze the notion of substance while implying that substances are asymmetrically independent of all other categories of being is, however, an error. If substances do enjoy this sort of independence, and it has been a persistent theme in metaphysics that a correct analysis or understanding of substance will have this implication, then we must seek a different analysis of substantiality than D2. In the later Metaphysics, Aristotle defends his hylomorphic account of substance, according to which a substance is a combination of form and matter. On one interpretation, this is just a useful way of distinguishing, in the case of compound bodies, between the structure of the body and its constituent stuff. Such an analysis is level-relative. If Aristotle meant to say that there could be pure (or prime) matter, stuff without form, then this is of questionable coherence. He is also ambivalent about the possibility of the existence of pure form. In any case, Aristotle’s hylomorphism seems incompatible with the possible existence of immaterial souls. Descartes sought a different independence analysis of the concept of substance. For example, at one point he states, 87

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 just this – that it can exist all by itself, that is without the aid of any other substance. (Philosophical Writings, Vol. II, p. 159) Hence, Descartes seems to be endorsing the following analysis of the concept of substance: (D3) x is a substance =df. x can exist without the aid of any other substance. This obviously won’t do, since to try to analyze the notion of a substance in terms of being capable of existing without the aid of any other substance, is viciously circular. Moreover, D3 implausibly implies that if God exists, then only God is a substance – for no created substance can exist without the aid of God. At another point, Descartes avoids the circularity of D3 with the following statement: By substance, we can understand nothing other than a thing which exists in such a way as to depend on no other thing for its existence. (Philosophical Writings, Vol. II, p. 210) The implied analysis of the concept of substance is the following: (D4) x is a substance =df. x exists and x depends on no other entity for its existence. D4 seems to avoid the circularity of D3, but has problems of its own. The main one is that no entity is independent of every other entity. For example, for any entity, x, there is a property, y, such that x has y essentially, and thus depends on y in the sense of entailing its existence. Another problem is that a compound body, which is a substance, depends on its parts in the same sense. Therefore, D4 does not appear to provide a logically necessary condition for something’s being a substance. Spinoza is another proponent of an independence theory of substance. His famous definition of substance reads as follows: By substance, I understand that which is in itself and is conceived through itself; in other words, that, the conception of which does not need the conception of another thing from which it must be formed. (Ethics and Selected Letters, p. 31) 88

Spinoza’s definition presents many difficult problems of interpretation, but on the face of it, appears to analyze substance in terms of some sort of conceptual independence, with the idea being that what is conceptually independent is also metaphysically independent. Spinoza thought that his definition implied that there was only one substance, Nature, and that this substance exists necessarily. There appear to be at least two serious criticisms of Spinoza’s analysis of substance. First, it fails to account for the data that any successful analysis must account for. In this case, Spinoza’s analysis implies, contrary to the data, that atoms, living organisms, and finite inanimate compound bodies are not substances – only the universe is. Thus, Spinoza has not succeeded in analyzing the ordinary concept of a substance; rather, he has substituted a radically revisionary notion of his own. (This criticism applies as well to D3, above.) Second, it is not clear that even the universe or nature satisfies Spinoza’s definition, since in order to conceive of the universe, it seems, one must conceive of one or more of the attributes of nature, e.g., extension. More recent independence analyses of the concept of substance attempt to conform largely to our intuitions about what entities are substances while capturing a more complex sense in which substances uniquely possess some sort of independence (e.g., Hoffman and Rosenkrantz, 1994; Lowe, 2006). Some philosophers have tried to analyze the concept of substance in terms of being a subject in which properties inhere. The idea is that there are properties, and then there are things in which properties inhere, namely, substances. For example, Descartes seems to be embracing this theory when he says, Substance. This term applies to every thing in which whatever we perceive immediately resides, as in a subject, or to every thing by means of which whatever we perceive exists. (Philosophical Writings, Vol. II, p. 114) The inherence theory, however, fails to provide a sufficient condition for something’s 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 substances. Realizing this, some philosophers have embraced the substratum or bare particular theory of substance, according to which a substance is a concrete individual that has no properties in itself, but instead serves as that in which the properties of ordinary objects inhere in some sense. A ball, on this theory, is not a substance, but rather a whole constituted by a substance/substratum and certain properties. (Alternatively, the ball is a substance, constituted by a substratum and certain properties – the most effective criticism of substratum theories applies to both versions.) Some have attributed this theory to Descartes and/or Locke, and among more recent philosophers, the substratum theory has been defended by Bergmann and, at one point, Russell. An apparently devastating criticism of any sort of substratum theory is this: it is incoherent to postulate the existence of something that lacks any properties. Nor does the substratum theorist actually refrain from attributing any properties to substrata, since he says that substrata are concrete, that properties subsist or inhere in them, and so forth. A final type of theory of substance is the bundle theory. This, unlike the preceding theories, is a reductionist theory of substance, that is, it implies that substances are aggregates of entities belonging to another ontological category. We shall concentrate here on the bundle theory that holds substances to be aggregates of concrete attributes or tropes. Proponents of this sort of theory defend an ontology devoid of both universals and irreducible substances – a simplifying move that they regard as a major strength of the theory. Bundle theorists include Russell at a later stage of his career, Ayer, Hume, Herbert Hochberg, and Castañeda. A recent and novel version of the bundle theory that tries to distinguish between those attributes essential to a substance and those accidental to it has been defended by Simons (1994). Bundle theories face several challenges. One is to explicate the relation(s) that is (are) supposed to unify the

tropes that comprise the bundle. Another is to avoid difficulties that seem to derive from the modal properties of the bundles and from their identity conditions. For example, if a bundle is a (special kind of ) collection of tropes, then since collections have their parts essentially, how can a substantial bundle undergo qualitative (or even relational) change? In addition to debates over the nature or analysis of the concept of an individual substance, metaphysicians have differed over the kinds of individual substances that there are or could be. A familiar controversy of this sort is the one between materialists, dualists, and idealists. Another aspect of this issue is, among material objects, whether or not compound bodies exist, whether or not inanimate compound bodies exist, and whether or not artefacts exist. Van Inwagen, for example, has denied the reality of inanimate compound bodies of any sort (while affirming the reality of atomic bodies and organisms), and he has challenged those who assert their existence to provide a satisfactory principle of unity for such objects. Hoffman and Rosenkrantz attempt to do so both for inanimate compound bodies and organisms, though not for artefacts (Hoffman and Rosenkrantz, 1997). Lowe (2006) and Thomasson (2007), on the other hand, defend the view that artefacts, understood as genuine substances, belong in our ontology. IV – An Analysis of Substantiality All individual substances belong to the ontological category of Substance. In a broad sense, ontological categories are the more general kinds of entities which (for all we know) could exist. Examples of such categories and sorts of entities which might belong to them are the following: Place (e.g., a volume of space), Time (e.g., an instant), Event (e.g., a process), Trope (in the sense of a concrete “quality”, e.g., the particular wisdom of Socrates), Boundary (e.g., a surface), Privation (a concrete entity such as a hole, gap, or shadow), and Collection (in the sense of an arbitrary sum of any concrete entities, e.g., the Moon + the Empire State 89

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

s u b s t a nce Building + Mount Everest). The foregoing examples of categories are species of Concrete Entity. On the other hand, examples of categories which are species of non-concrete or Abstract Entity are Property (e.g., Wisdom), Relation (e.g., Betweenness), Proposition (e.g., that 2 + 2 = 4), Set (e.g., { } ), and Number (e.g., 7). Intuitively, the foregoing species of concrete and abstract entities are peers in the sense that all of them are at the same level of generality. We call this level of generality Level C, assuming a hierarchical tree-like taxonomy in which Entity (the Level A category) is the summum genus, Concrete Entity and Abstract Entity (the Level B categories) are the mutually exclusive and exhaustive divisions of this summum genus, and the various species of Concrete Entity and Abstract Entity are the Level C categories (see concrete/abstract). Since the category of Substance is a species of Concrete Entity, it is a Level C category. But how does one acquire the concept of this Level C category? We address this question below. To begin, according to a plausible empiricist theory of concept formation, one can acquire the concept of a genus by perceiving instances of one or more species of that genus and engaging in a process of abstraction. This plausible empiricist theory entails that one may possess the concepts of certain species before one possesses a concept of the genus that subsumes them – and this is surely true. This process of concept formation involves one’s observing certain relevant similarities between the perceived instances of the genus while setting aside inessential dissimilarities between them. In particular, given that material objects or bodies are a species of substance, one can acquire the concept of a substance by abstracting from one’s perceptions of bodies, for example, by noticing that they are enduring entities, that they persist through qualitative change, that they exist independently of other entities of the same kind, and so forth, while setting aside inessential observed differences between them such as differences in shape and size. One reason why people can acquire the concept of a substance via the abstractive process from perceptions of bodies is because 90

people have an intuitive observational concept of a material object, an observational concept which does not presuppose the concept of a material substance. According to this intuitive observational concept, a material object or body is an entity which has certain perceivable characteristics, including at least certain basic spatial characteristics, which can exist unperceived, and so forth. Similarly, people have available to them an intuitive concept of a (Cartesian) soul as a non-spatial entity which has certain mental characteristics. This intuitive concept does not presuppose the concept of an immaterial or spiritual substance. By means of the aforementioned process of concept formation, one can see that souls and bodies would belong to a common level C category because one can see that souls and bodies resemble one another in ontologically relevant respects. In particular, one can see that, like a body, a soul can endure, persist through qualitative change, exist independently of other entities of the same kind, and so forth. Since an immaterial physical object such as a point-particle or a massless extended object resembles a body in these ontologically relevant respects, a physical object of this kind also would belong to the level C category in question. However, people seem to be unable to conceive of anything belonging to this level C category other than a physical object (including material objects and immaterial physical objects), a soul, and a Spinozistic substance. This is because we cannot conceive of anything other than a physical object, a soul, and a Spinozistic substance that could endure, persist through qualitative change, exist independently of any other entities of its kind, and so forth. In what follows, we seek to revive the traditional idea that a substance is an independent or autonomous being. In particular, we argue that the notion of a Level C category can be utilized to analyze the concept of substance in terms of a sort of ontological independence which uniquely characterizes any possible substance. Our proposed analysis of the concept of 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 meet certain independence conditions qua belonging to that category. In other words, we shall argue that the concept of substance can be analyzed in terms of independence conditions derived from an entity’s belonging to a Level C category. Our analysis, A, stated below, consists of the conjunction of three independence conditions. (A) x is a substance =df. x belongs to a Level C category, C1, such that: (i) C1 could have a single instance throughout an interval of time, (ii) C1’s instantiation does not entail the instantiation of another Level C category which satisfies (i), and (iii) it is impossible that something belonging to C1 has a part which belongs to another Level C category (other than the categories of Concrete Proper Part and Abstract Proper Part). In condition (i), by an interval of time we mean a non-minimal time. And by C1’s having a single instance throughout an interval of time, we mean that something instantiates C1 throughout an interval of time, and that there is no other instance of C1 in that interval of time. Although clause (i) of A entails that there could be a substance that is independent of any other substance, it does not entail that every substance could be independent of any other substance. For instance, clause (i) of A is logically consistent with there being a compound substance that is dependent upon its substantial parts. Hence, according to clause (i) of A, an entity, x, (regardless of whether x is simple or compound), is a substance in virtue of x’s belonging to a Level C category which could have a single instance throughout an interval of time. Clause (i) of A characterizes a substance in terms of an independence condition entailed by the instantiability of a certain Level C category. Clause (ii) of A entails that an entity, x, is a substance only if x’s instantiation of a Level C category is independent of the instantiation of another Level C category which could have a single instance throughout an interval of time. However, although the existence of a substance may entail the

existence of entities of another Level C category, for example, properties, in no case is this other category such that it could have a single instance throughout an interval of time. It follows that the category of Substance satisfies clause (ii) of A. Clause (iii) of A entails that an entity, x, is a substance only if x belongs to a Level C category whose instantiation by an item is independent of any other Level C category (other than two special Level C categories referenced in clause (iii)) being instantiated by a part of that item. In general, a part of a physical substance could only be a physical substance or a portion of physical stuff, and a non-physical soul has no parts. Hence, it appears to be impossible for a substance to have a part that belongs to another Level C of the sort in question, for instance, a place, a time, a boundary, an event, a trope, a privation, a property, a relation, a proposition, and so on. Accordingly, the category of Substance seems to satisfy (iii) of A. A is compatible with either of two assumptions. On the first, all individual substances have contingent existence: each substance could fail to exist. On the second assumption, there is a single necessarily existing substance, G, such as God, a substance which could not fail to exist. On either of these assumptions, it is possible for there to be a substance, s, which exists throughout some interval of time, t, without any other substance existing within t. On the first assumption there could exist throughout t nothing but a single contingent substance. On the second assumption, if G exists in time, then there could exist throughout t but a single necessary substance; and if G exists outside of time, then there could exist throughout t but a single contingent substance. However, it might be objected that if there is an individual substance, then there must be other substances, namely, the (spatial) parts of the individual substance in question. But it is only true that a compound substance must be composed of other substances. It is possible for there to be a simple substance that has no other substance as a (spatial) part, for instance, a non-spatial soul, a point-particle, an indivisible, spatially 91

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 atom. Note that an indivisible, spatially extended substance has spatially extended parts. However, these parts cannot exist independently of the whole of which they are parts. Yet, necessarily, a substance, s, is an independent being in this sense: s can exist independently of any other contingently existing substance, s*, unless s* is a proper part of s or s* helped generate s. Since a spatially extended proper part of an indivisible substance fails to satisfy this independence requirement, such a proper part does not qualify as an individual substance. Rather, it is just a concrete proper part (of a substance). Such an insubstantial proper part would be an instance of the special Level C category of Concrete Proper Part. The wording in clause (iii) of A that excludes the category of Concrete Proper Part from consideration accommodates this possibility of an individual substance that has an entity of another Level C category as a part. One sort of part in addition to a spatial part is a temporal part. Clearly, it is at least possible for there to be an enduring substance that does not have another shorter-lasting substance as a (temporal) part or sub-stage. (In contrast, necessarily, a temporally extended event has other shorter-lasting events as temporal parts, stages.) Still, arguably, there could be a temporally extended substance that does have other shorterlasting substances as temporal parts, e.g., a four-dimensional physical object in a fourdimensional space–time continuum. But, possibly, there is an enduring, indivisible, physical particle in three-dimensional space (not four-dimensional space–time) which does not have another shorter-lasting, indivisible, physical particle as a (temporal) part or sub-stage; or possibly, there is an enduring non-spatial soul which does not have another shorter-lasting soul as a (temporal) part or sub-stage. Thus, it is possible that throughout an interval of time, t, there exists an indivisible substance and no other substance, for example, just one enduring indivisible particle, or just one enduring non-spatial soul. On the basis of the preceding discussion, we conclude that the category of Substance 92

satisfies the three clauses of A. On the other hand, it appear that the categories of Event, Time, Place, Trope, Boundary, Collection, Property, Relation, Proposition, Set, and Number could not have a single instance throughout an interval of time. Let us briefly explore the nature of these categories in order to give some indication of how this observation can be supported. Consider first the categories of Property and Trope. Necessarily, either an abstract property, or a concrete trope, is an entity that stands in lawful logical or causal relations to others of its kind. For example, the existence of squareness (or of a particular squareness) entails the existence of straightness (or of a particular straightness). Similar arguments apply to the categories of Relation, Proposition, Set, Number, and so on. With respect to the category of Place, necessarily, if space exists, then it has an intrinsic structure that it is compatible with the occurrence of motion. This entails that, necessarily, if space exists, then space contains at least two places. In the case of the category of Time, necessarily, if time exists, then it has an intrinsic structure that is compatible with creation, destruction, qualitative change, or relational change. It follows that, necessarily, if time exists, then there are at least two times. With regard to the category of Boundary, necessarily, every boundary is spatial or temporal in character. The existence of a boundary entails the existence of an extended, continuous space or time which contains infinitely many extended places or times. Moreover, necessarily, whatever is bounded has a dimension lacked by its boundary, e.g., a dimension of thickness, area, length, or duration. Thus, necessarily, if there is one (spatial or temporal) boundary, then there are infinitely many other spatial or temporal boundaries. Consider next the category of Event. Necessarily, an event that occurs over an interval of time is a process. Necessarily, a process involves other sub-processes that are themselves events. Hence, necessarily, if an event occurs over an interval of time, then there is another event that occurs 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 entity, Collection. Necessarily, if a collection, c1, exists throughout an interval of time, t, then c1 has at least two parts, x and y, both of which exist throughout t. In that case, it appears that there must be a shorter time, t*, which is a sub-time of t and which is a part of another collection, c2, for example, a shorter-lasting collection either composed of t* and x, or composed of t* and y. Hence, necessarily, if a collection exists throughout an interval of time, then it appears that there is another collection which exists within that interval of time. This suggests that the category of Collection fails to satisfy clause (i) of A. However, A also implies that collections are not substances in virtue of their failure to satisfy clause (iii), a clause that requires that it is impossible for an entity of a Level C category has as a part an entity of another Level C category (with the exception of two special categories which are irrelevant here). After all, something that belongs to a collection is a part of that collection, and it is evidently possible for something that belongs to a collection to be an entity of a Level C category other than the category of Collection, e.g., an entity such as a substance, an event, or a place. In sum, it appears that there could not be just one entity of any of the foregoing Level C categories (throughout an interval of time.) Moreover, in each case there is no other Level C category which could be instantiated by an entity belonging to the category in question, and which could have a single instance throughout an interval of time. Hence, (clause (i) of) A seems to have the desirable consequence that an entity that belongs to any of these categories is insubstantial. Clauses (ii) and (iii) of A enable this proposed analysis to deal with insubstantial entities of various other kinds. For example, suppose for the sake of argument that a purple after-image is an insubstantial entity of the irreducible category Sense-Datum. On this supposition, a sensedatum is not an event, a property, a trope, a boundary, and so on. If so, then an afterimage belongs to the Level C category of Sense-Datum. But the instantiation of this

category entails the instantiation of another Level C category that satisfies clause (i) of A, namely, the category of substance. After all, there cannot be a sense-datum unless there is a perceiving substance. It follows that the category of Sense-Datum does not satisfy clause (ii) of A. Moreover, there is no other Level C category which satisfies A and which could be instantiated by a sensedatum. Thus, clause (ii) of A has the desirable implication that a sense-datum is an insubstantial entity (see sensa). Finally, consider the Level C category of Privation. In this context, by a privation we mean a concrete entity which is an absence or lack of one or more concrete entities, and which is wholly extended between two or more bounding concrete entities, or else wholly extended between two or more bounding parts of a single concrete entity. A privation in this sense is an insubstantial concrete entity. (So, a negative abstract entity, e.g., the proposition that there are no centaurs, does not qualify as a privation in the relevant sense.) It seems that the category of Privation satisfies clause (i) of A. Consider, for example, the possibility of there being nothing but two temporally separated flashes and the period of darkness, d, between them. We may assume that in this possible situation d is the only privation throughout the interval of time in question. On the other hand, it can be argued that the category of Privation fails to satisfy clause (ii) of A for the following two reasons. First, the category of Substance satisfies clause (i) of A and this Level C category is other than the category of Privation. Second, necessarily, if there is a privation, then there is a substance, e.g., a substance which flashes, a substance which is perforated, a substance which is shadowed or which casts a shadow, and so on; though, clearly, there could be a (basic) substance without there being a privation. Still, some have claimed that there could be a flash without there being a substance that flashes, and thus it is controversial whether the existence of a privation requires the existence of a substance. Fortunately, A is neutral with respect to this controversy, 93

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 privations are not substances. To see this, note that privation, d, has as parts certain (lightless) periods of time within d. These parts belong to the category of Time, a Level C category other than the category of Privation. It follows that the category of Privation fails to satisfy clause (iii) of A. In addition, there is no other Level C category which satisfies A and which could be instantiated by a privation. Hence, clause (iii) of A has the desired consequence that a privation is not a substantial entity. It appears that A provides a logically necessary and sufficient analysis of the concept of substance in terms of a kind of ontological independence. In the light of the foregoing discussion, it also appears that this analysis is ontologically neutral to a high degree, that is, compatible to a high degree with the existence of entities belonging to various intelligible categories, given plausible views about the nature, existence conditions, and interrelationships of entities belonging to those categories. See also the a–z entry on substance. b i b l i og rap hy Aristotle: The Complete Works of Aristotle: The Revised Oxford Translation, ed. Jonathan Barnes (Princeton, NJ: Princeton University Press, 1984). Bergmann, G.: Realism (Madison, WI: University of Wisconsin, 1967). Campbell, K.: Abstract Particulars (Oxford: Blackwell, 1990). Chisholm, R.: A Realistic Theory of Categories (Cambridge and New York: Cambridge University Press, 1996). Descartes, R.: The Philosophical Writings of Descartes, trans. John Cottingham, Robert

Stoothoff, and Dugald Murdoch (Cambridge: Cambridge University Press, 1984). Hoffman, J. and Rosenkrantz, G. S.: “How to Analyze Substance: A Reply to Schneider,” Ratio 20 (2007), 130– 41. Hoffman, J. and Rosenkrantz, G. S.: Substance Among Other Categories (Cambridge and New York: Cambridge University Press, 1994). Hoffman, J. and Rosenkrantz, G. S.: Substance: Its Nature and Existence (London and New York: Routledge, 1997). Locke, J.: An Essay Concerning Human Understanding, ed. Roger Woolhouse (London: Penguin Books, 1997). Loux, M.: Metaphysics: A Contemporary Introduction, 3rd edn. (London and New York: Routledge, 2006). Loux, M.: Substance and Attribute (Dordrecht: Reidel, 1978). Lowe, J.: The Four-Category Ontology: A Metaphysical Foundation for Natura Science (Oxford and New York: Oxford University Press, 2006). Simons, P.: “Farewell to Substance: A Differentiated Leave-taking,” Ratio N.S. XI (1998), 235–52. Simons, P.: “Particulars in Particular Clothing: Three Trope Theories of Substance,” Philosophy and Phenomenological Research 54 (1994), 553–76. Spinoza: The Ethics and Selected Letters, trans. Samuel Shirley, ed. Seymour Feldman (Indianapolis, IN: Hackett Publishing Company, 1982). Thomasson, A. L.: Ordinary Objects (Oxford and New York: Oxford University Press, 2007). van Inwagen, P.: Material Beings (Ithaca, NY and London: Cornell University Press, 1990). joshua hoffman gary s. rosenkrantz

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

Part II

Metaphysics from A to Z

9780631199991_4_P2001.qxd 1/12/09 3:04 PM Page 97

Abelard, Peter (1079–1142) French philosopher, logician and theologian. Born near Nantes in France in 1079 Abelard studied logic in his youth under Roscelin, notorious for his antirealist interpretation of logic, and went on to become the most sought-after teacher of logic in Europe. Beyond logic Abelard involved himself in theological debates, and his interpretation of the Holy Trinity, a topic which called forth his best work on the concept of sameness, was condemned twice by the church. Abelard’s life was a stormy one including the much celebrated romance with and marriage to Héloïse, his subsequent castration by thugs hired by her uncle, and a bitter series of disputes with William of Champeaux over universals. It was the topics of universals and identity that elicited Abelard’s main efforts in metaphysics. While arguing that no universal, i.e., nothing common to many, is any “real” thing, that is has an existence independent of the mental and linguistic activities that involve signification of things in the world, Abelard proposed that nevertheless there are status which serve as objective significates of predicates that are true of many distinct things. He gave the status much the same treatment as he proposed for dicta, which are the significates of sentences and the primary bearers of truth and falsity. They are not things in the world, not even psychological or linguistic things, but they can exist and be known objectively. Taking off from remarks by Aristotle in the Topics Abelard distinguished different sorts of identity and distinctness. Most important is the contrast between sameness in “essence” and sameness in property. The former means that the items in question have all their parts in common; the latter requires that

the items be defined in the same way. He claimed that objects and the matter of which they were composed were the same in the former sense but not in the latter. wri t i ngs Dialectica (“Dialectic”), ed. L.M. de Rijk, 2nd ed. (Assen: Van Gorcum, 1970). Logica ingredientibus (“Logic for Beginners”), in Peter Abaelards Philosophische Schriften, fascicules 1–3, ed. B. Geyer, in Beiträge zur Geschichte der Philosophie des Mittelalters, Vol. 21, fascicules 1–4 (Münster: Aschendorff, 1919–27). Logica “nostrorum petitioni sociorum” (“Logic in Response to the Request of Our Friends”), in Peter Abaelards Philosophische Schriften, fascicule 4, ed. B. Geyer, in Beiträge zur Geschichte der Philosophie des Mittelalters, Vol. 21, fascicules 1–4 (Münster: Aschendorff, 1919–27). (A second edition by the same publisher appeared in 1973.) Theologia Christiana ed. E.M. Buytaert in Corpus Christianorum Continuatio Mediaevalis 11–12 (Turnholt, Belgium: Brepols, 1969). martin m. tweedale abstract see concrete/abstract accident see essence/accident acquaintance Acquaintance is a central notion in Russellian metaphysics, as well as Russellian epistemology and philosophy of language. Russell distinguishes knowledge by acquaintance from knowledge by description, and characterizes the former as follows.

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 with anything of which we are directly aware, without the intermediary of any process of inference or any knowledge of truths” (Russell, 1959, p. 46, italics in original). (2) “it is possible, without absurdity, to doubt whether there is a table at all, whereas it is not possible to doubt the sensedata” (Russell, 1959, p. 47). The table is not an object of acquaintance, but the sense-data are, and this condition is supposed to provide a general contrast between objects of acquaintance and other things. (3) “All our knowledge, both knowledge of things and knowledge of truths, rests upon acquaintance as its foundation” (Russell, 1959, p. 48). (4) Russell also specifies objects of acquaintance by extension. “We have acquaintance in sensation with the data of the outer senses, and in introspection with the data of what may be called the inner sense – thoughts, feeling, desires, etc.; we have acquaintance in memory with things which have been data either of the outer senses or of the inner sense. Further, it is probable, though not certain, that we have acquaintance with Self, as that which is aware of things or has desires towards things. In addition to our acquaintance with particular existing things, we also have acquaintance with . . . universals’ (Russell, 1959, pp. 51–2, italics in original). These four specifications cannot be assumed to coincide. What, if anything, is the foundation of our knowledge and what, if anything, is known “directly” are, of course, themselves matters of philosophical controversy. And the second specification has its own special problem, since universals and sense data, far from being indubitable, are just the sorts of entities whose existence many philosophers doubt. Russell recognizes that many people doubt or deny the existence of universals, but he does not seem to recognize the problem this fact raises for 98

the conjunction of his view that objects of acquaintance include universals and his view that objects of acquaintance, are such that their existence cannot be doubted (see Russell, 1959, chs. 9 and 10). James Van Cleve has mentioned that any philosopher holding an indubitability thesis will need to formulate it so as to avoid the conclusion that we have indubitable knowledge of anything that is in fact philosophically controversial. But the way to do this in the present case seems to be, for example, to replace such claims as Russell’s that “it is not possible to doubt the sense-data” (Russell, 1959, p. 47) with claims to the effect that it is not possible to doubt that one seems to see something blue, or that one is in pain, etc. This no longer involves reference to any object of acquaintance whose existence cannot be doubted. For Russell, only an object of acquaintance can be the referent of a logically proper name, i.e., a name that refers directly, without describing, and whose sole semantic function is to stand for it referent. By his principle of acquaintance, “Every proposition which we can understand must be composed wholly of constituents with which we are acquainted” (Russell, 1959, p. 58, italics in original). Donnellan offers a useful formalization of this notion of a constituent, when he says that if, and only if, Socrates is a constituent of the proposition expressed by the sentence “Socrates is snub-nosed”, this proposition “might be represented as an ordered pair consisting of Socrates – the actual man, of course, not his name – and the predicate (or property, perhaps), being snub-nosed” (Donnellan, 1974, p. 225). Russell grants that his principle of acquaintance entails that much of a person’s language is private (in the sense that it is logically impossible for anyone else to apprehend the propositions expressed by the speaker) as well as ephemeral (in the sense that it is logically impossible for anyone to apprehend at time t2 the proposition he expressed at t1. (For Russell on ephemerality, see Russell, 1956, pp. 201–4.) But Russell overstates the extent of privacy his principle of acquaintance requires. He says

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

action theory When one person uses a word, he does not mean by it the same thing as another person means by it . . . It would be absolutely fatal if people meant the same things by their words . . . the meaning you attach to your words must depend on the nature of the objects you are acquainted with, and since different people are acquainted with different objects, they would not be able to talk to each other unless they attached quite different meanings to their words. We should have to talk only about logic. (Russell, 1956, p. 195) By Russell’s own lights, this claim is overstated, since he does not limit objects of acquaintance to sense data, oneself, and entities of logic, such as sets. He also includes universals. Thus, on the principle of acquaintance, we would not “have to talk only about logic” in order to attach the same meanings to our words. We could also talk about blueness, roundness, etc., and we could discuss such propositions as the proposition that blue is more like purple than either is like orange. But this qualification is unlikely to assuage the doubt of opponents of the principle of acquaintance, especially since the argument Russell offers for the principle is drastically inadequate. He says it is scarcely conceivable that we can make a judgment or enter a supposition without knowing what it is that we are judging or supposing about. We must attach some meaning to the words we use, if we are to speak significantly and not utter mere noise; and the meaning we attach to our words must be something with which we are acquainted. (Russell, 1958, p. 58, italics in original) Of course this is not really an argument. It begs the question (see Ackerman, 1987). b i b l i og rap hy Ackerman, F.: “An Argument for a Modified Russellian Principle of Acquaintance,” in Philosophical Perspectives, Vol. 1, Metaphysics, ed. J. Tomberlin (Atascadero, CA: Ridgeview, 1987), 501–12.

Donnellan, K.S.: “Speaking of Nothing,” The Philosophical Review 83 (1974), 3–31; repr. in Naming, Necessity and Natural Kinds, ed. S.P. Schwartz (Ithaca, NY: Cornell University Press, 1977), 216– 44. Russell, B.: “The Philosophy of Logical Atomism,” in his Logic and Knowledge, ed. R.C. Marsh (London: George Allen and Unwin, 1956), 177–281. Russell, B.: The Problems of Philosophy (New York: Oxford University Press, 1959). felicia ackerman action theory Action theory deals with that concept of action that applies only to beings who have wills. The questions it addresses include: (1) what is the mark of action? (2) How should actions be individuated? (3) What makes an action intentional? (4) Is freedom of action compatible with determinism? (5) What makes true the sort of explanation peculiar to action, namely, that the agent did the action for certain reasons? t he mark of ac t i on What distinguishes an action from other sorts of events of which a person may be the subject, such as sensations, perceptions, feelings, unbidden thoughts, tremblings, reflex actions? Two main sorts of answer have been offered. According to one, what marks an event (say, a movement of one’s body) as an action is something extrinsic to the event, namely its having been caused in the right sort of way by the subject’s desires (or intentions) and beliefs (see Goldman, 1970, ch. 3, and Davidson, 1980, essay 1). The right sort of causal connection is important, because, for example, the fact that a desire to have another drink results in the subject’s falling down does not make that event an action. This sort of view seems, however, not to cover spontaneous actions whose occurrence is not explained by any antecedent motives of the agent. The other kind of account finds the mark of an action in the intrinsic nature of the event, rather than in something external to it. The idea is that an event is an action 99

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 event. Some hold that the special event is an occurrence of a quite special sort of causation, where an event is caused, not by another event, but by the agent herself; an agent is the only sort of enduring thing that can be the subject of this special kind of causation (see Taylor, 1966; Chisholm, 1976). Others hold that the special event is mental; for some, what makes it special is its functional role (Davis, 1979, chs. 1–2), and for others, it is its phenomenal character (Ginet, 1990, ch. 2). In actions that go on to become voluntary bodily exertions this event is a willing (or volition) to act. Some (for example, Hornsby, 1980) think that the content of this volition may be anything that the agent was trying to do in the action. Reflection on our experience of voluntary bodily exertion suggests, however, that there is in it something to be called volition that is quite distinct from intention and the content of which is limited to the immediately present exertion of the body (see Ginet, 1990, ch. 2). t h e i n d ividuatio n o f actio n Suppose that just now I moved my right index finger and thereby pressed a key and thereby put a character on the computer screen. Each of the following is a description of an action I performed: (1) “I moved a finger”; (2) “I moved my right index finger”; (3) “I pressed a key”; and (4) “I put a character on the screen.” How many different actions do these four descriptions pick out? One view holds that they pick out four different actions; because an action is an exemplifying of an action property by an agent at a time, and our four descriptions express four different action properties (see Goldman, 1970). Another view holds that they all describe the same action in terms of different properties; my action was just the minimal thing by (or in) doing which I did the things attributed to me by all the descriptions (see Davidson, 1980, essay 3; Hornsby, 1980). (On some views this basic action is the bodily movement, on others it is a volition.) Between these extreme views, one may take the position that, although an 100

action is normally thought of as a more concrete entity than an exemplifying of a property (so that (1) and (2) describe the same concrete action in terms of different intrinsic properties), one action can be a proper part of a distinct action in which something that is not an action, namely, a consequence of the first action, is an additional part (so that (3) picks out a larger action of which (1)– (2) is merely the initial part, and (4) picks out a still larger action of which (4) is merely the initial part) (see Thomson, 1977; Ginet, 1990, ch. 3). t he int ent i onal i t y of ac t i on Smith swung the racket intentionally and in so doing inadvertently hit his opponent with it. Smith’s hitting his opponent with the racket was not intentional, but it could have been. Whether an action is intentional or not often makes a big difference for the sort of evaluation it deserves. What determines whether an action is intentional or not (under a given description)? This can be divided into two questions, depending on whether or not the action description in question is basic. An action description, of the form “S’s A-ing”, is basic just in case there is no other, non-equivalent action description, “S’s B-ing”, such that it is true that S A-ed by B-ing. With respect to a basic description, it is plausible to hold that whatever makes an event it fits an action also makes it intentional under that description. (This is especially plausible if the basic descriptions attribute mental acts of volition.) The question with respect to non-basic descriptions is more difficult. One might think that it would have been sufficient for Smith’s hitting his opponent with the racket intentionally that he intended of his voluntary bodily movement that by it he would cause the racket to hit his opponent. But suppose he was too far from the opponent for the swing to hit as he intended; however, his grip loosened as he swung and the racket flew out of his hand and hit the opponent. We cannot then say that his hitting his opponent with the racket was intentional. Perhaps it is sufficient for his action’s 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 intended. But this appears not to be necessary. Suppose Smith stumbled slightly as he swung, causing him to hit the opponent slightly below the spot he intended to hit; in this case, though he did not hit him in just the way he intended, it seems that he still hit him intentionally. In light of such difficulties (there are others), it is clear that it will not be a simple matter to devise a satisfactory necessary and sufficient condition for an action’s being intentional under a description. (For one complex proposal, see Ginet, 1990, ch. 4.) f r e e a c tio n and d eter minism I have freedom of action at a given time just in case more than one alternative action is then open to me (see the extended essay on free will). We continually have the impression of having more than one alternative action open to us (indeed, a great many alternative actions normally seem open to us: consider all the different ways that, as it seems to me, I could next move my right hand). determinism is the thesis that, given the state of the world at any particular time, the laws of nature (see law of nature) determine everything that happens thereafter down to the last detail. Some philosophers have argued that our impression of freedom is always an illusion if determinism is true or if, though false, it fails to be false in the right places (see van Inwagen, 1983, ch. 3; and Ginet, 1990, ch. 5). (This last disjunct is important because, although contemporary physics may give us good reason to think that determinism is false, it does not give us good reason to think it is false in the right places: as yet we do not even know precisely what the right places are.) The essential premises of the argument that determinism is incompatible with freedom of action are two: (1) No one ever has it open to him or her to make true a proposition that contradicts the laws of nature. (2) No one ever has it open to him or her to determine how the past was, i.e., to make true one rather than another of contrary propositions that are entirely about the past.

From (2) it follows that (3) one can have it open to one at a given time to perform a certain action a, only if, for any truth entirely about the past, p, one has it open to one to make it the case that: p and one does a. From (1), (3), and determinism it follows that one never has it open to one to do anything other than what one actually does. Suppose that at 2 o’clock it seemed to me to be open to me to raise my hand then, but I did not do so. If determinism is true then there is a true proposition, p, which is entirely about the past relative to 2 o’clock and such that it follows from the laws of nature that: if p then I did not raise my hand at 2 o’clock. From (3) it follows that it was open to me to make it the case that I did raise my hand at 2 o’clock only if it was open to me to make it the case that: p and I raised my hand at 2 o’clock. But, given (1), it could not have been open to me to make that proposition true, for it contradicts the laws of nature. Therefore, if determinism is true (and so also are (1) and (3)), then (contrary to my impression) it was not open to me to do a at t. This argument is obviously valid and so philosophers who resist the conclusion that determinism is incompatible with freedom of action (and many do) must reject either (1) or (2). Arguments against (2) are possible, but the more popular, and perhaps more promising, line is to attack (1) (see Fischer, 1988; Lewis, 1981). (1) could be put this way: if it follows from the laws of nature that if p then q, then it is never in anyone’s power to make it the case that: p and not-q. This principle seems appealing because it seems that we ordinarily feel compelled to make inferences in accordance with it. For example, if I know that X’s brain state at t is such as to nomically necessitate X’s being unconscious for at least one minute after t, then that seems good enough to infer that it is not open to X at t to voluntarily raise his or her arm during the minute after t. To account for the apparent cogency of such an inference, while rejecting (1), one might suggest that what really underlies its validity is not (1) but a more complex principle, something like the following: if p nomically necessitates that X does not act in a certain 101

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 through X’s internal processes in the way that it does in normal seemingly free action, then it is not open to X to act in that way at t. This more complex principle, says the critic of (1), will account for all the acceptable inferences that seem to invoke (1). But will it? Imagine a possible world where determinism is true and Martians control all of X’s actions over a long period through controlling X’s normal psychological processes of motivation and deliberation. If you are inclined to think this would mean that X has no more freedom of action than a puppet, then, it seems, you are inclined to operate with (1) and not just the more complex principle (for the latter would not justify that inference). t h e n a tur e o f actio n exp lained b y r e a so ns Typically when one acts one has motives or reasons for acting in the way one does and one acts in that way for those reasons. For example, my reason for opening the window was that I wanted to let out the smoke. I opened the window in order to let out the smoke, that is, because I intended thereby to let out the smoke. The main metaphysical issue concerning explanations of this sort is whether they are essentially nomic, that is, whether the truth of one of them entails that the case be subsumable under causal laws which dictate that whenever motives of the same sort as those the explanation cites occur in sufficiently similar circumstances they (the motives and the relevant circumstances) causally necessitate an action of the same sort (see Ayer, 1946; and Davidson, 1980, essays 1, 11, for expressions of this view). The nomic view of reasons explanations would tend to be confirmed if we knew (or had good evidence for) the relevant laws in most cases of true reasons explanations. But we do not. Indeed, it may be that, as yet, there is no true reasons explanation of any action for which anyone knows causal laws that govern the explanation. Of course, this ignorance does not show that the nomic view is wrong or that, on it, we are not justified 102

in believing any reasons explanations. Perhaps we need not know what the relevant causal laws are in order to be justified in giving a reasons explanation in a particular case. We must, however, be justified in believing that there are laws that govern the case (whether or not their contents are known to us); and it might well be doubted whether there is any case for which we are justified in believing even this. The nomic view nevertheless has a strong appeal for many philosophers. This may be because they find it hard to see what else, if not a nomic connection, could make a genuine explanatory connection between motives and action. This is a fair question, which one must answer if one wants to make a good case against the nomic view. One must specify a condition that is clearly sufficient for the explanatory connection, does not imply a nomic connection, and is easy to know is present (especially for the subject). Here is a sketch of how one might try to do that (see Ginet, 1990, ch. 6 for a fuller exposition). Suppose that concurrently with my action of opening the window I remembered my antecedent desire to rid the room of smoke and I intended of that action I was engaging in that I would thereby satisfy that desire. These conditions seem clearly sufficient to make the explanatory connection between the desire and the action, to make it true that I opened the window because I wanted to rid the room of smoke; and just as clearly they seem to be compatible with there being no true causal laws which dictate that always a desire of that same sort in sufficiently similar circumstances must produce the same sort of action. That is, they give a non-nomic sufficient condition for a reasons explanation of an action. (Of course the obtaining of a non-nomic sufficient condition does not rule out the possibility of a nomic sufficient condition, perhaps even for the same explanation of the same action.) That reasons explanations need not be nomic is important for the view that freedom of action is incompatible with determinism. Otherwise, that view would be committed to the counterintuitive proposition that 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 a reasons explanation. b i b l i og rap hy Chisholm, R.M.: “The Agent as Cause,” in Action Theory, ed. M. Brand and D. Walton (Dordrecht: Reidel, 1976), 199–211. Davidson, D.: Essays on Actions and Events (Oxford: Clarendon Press, 1980). Davis, L.: Theory of Action (Englewood Cliffs, NJ: Prentice-Hall, 1979). Fischer, J.M.: “Freedom and Miracles,” Noûs 22 (1988), 235–52. Ginet, C.: On Action (New York: Cambridge University Press, 1990). Goldman, A.I.: A Theory of Human Action (Englewood Cliffs, NJ: Prentice-Hall, 1970). Hornsby, J.: Actions (London: Routledge and Kegan Paul, 1980). Lewis, D.: “Are we free to break the laws?,” Theoria 47 (1981), 112–21. Taylor, R.: Action and Purpose (Englewood Cliffs, NJ: Prentice-Hall, 1966). Thomson, J.J.: Acts and Other Events (Ithaca, NY: Cornell University Press, 1977). van Inwagen, P.: An Essay on Free Will (Oxford: Clarendon Press, 1983). carl ginet actuality see potentiality/actuality adverbial theory The adverbial theory is, at root, the view that to have a perceptual experience is to sense in a certain manner. Traditionally, the most popular analysis of perceptual experience has been the opposing sense-datum theory (see sensa). According to this theory, having a perceptual experience amounts to standing in a relation of direct perceptual awareness to a special immaterial entity. In particular cases this entity is called an after-image or a mirage or an appearance, and, in the general case, a sense-impression or a sense-datum. The sense-datum is required, so it is normally argued, in order to explain the facts of hallucination and illusion: since a person can have a visual sensation of a red square, say, even when there is no real red, square object in his general vicinity, it is typically

inferred that he is related, through his experience, to a red, square sense-datum. The sense-datum theory leads to a number of perplexing questions. For example, can sense-data exist unsensed? Can two persons experience numerically identical sense-data? Do sense-data have surfaces which are not sensed? What are sense-data made of? Are they located? Historically, the desire to avoid questions like these was one reason for the development of the adverbial theory. This position – that having a perceptual experience is a matter of sensing in a certain manner rather than sensing a peculiar immaterial object – is arrived at by reflecting on the fact that, on standard views, appearance, after-images, and so on, cannot exist when not sensed by some person. The explanation the adverbial theorist offers for this fact is that statements which purport to be about appearances, after-images, and so on, are in reality statements about the way or mode in which some person is sensing. Hence, a statement of the general form, “Person, P, has an F sense-impression”, or “P has an F sensation”, is reconstructed adverbially as, “P senses F-ly”, or as it is sometimes put, “P senses in an F manner.” This transformation has a number of grammatical parallels. “Patrick has a noticeable stutter”, for example, is equivalent to “Patrick stutters noticeably”, and “Patrick stutters in a noticeable manner.” Similarly “Jane does a charming waltz”, may be transcribed as, “Jane waltzes charmingly.” It should be obvious that the adverbial view can account for the facts of hallucination and illusion. If, for example, I am correctly described as having a visual sensation of something blue then “blue” in this description is taken upon analysis to function as an adverb which expresses a mode of my sensing. Hence, my having the sensation does not require that there be a blue physical object (or anything else for that matter) in my general vicinity – it suffices that I sense bluely. Although the adverbial theory began as, and is still most strongly associated with, the analysis of perceptual experience, it has also been applied elsewhere. For example, it is often held by adverbialists that our ordinary 103

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 that in reality there are no such items as pains and itches to which persons are related when they have a pain or feel an itch. Rather statements about bodily sensations have an underlying adverbial structure. “Jones has an intense pain”, for example, is analyzed as “Jones is pained intensely”; hence it is about the way in which Jones is pained. The motivation for this approach runs parallel to the one for perceptual experience: countenancing pains and other sensory objects in our ontology generates a host of philosophical puzzles. For example, are pains really located about the body as our ordinary pain talk suggests? If so, then presumably they are material objects. Why, then, are they never revealed by surgical examination of the appropriate limbs? Can pains exist in parts of the body without their being felt? Can two persons ever feel one and the same pain? All these puzzles dissolve once the adverbial view is adopted. Some philosophers have argued that the adverbial theory can even be extended to the analysis of belief and desire discourse. Thus, having the belief that snow is white, say, is not a matter of bearing the “having” relation to a particular belief, but rather a matter of believing in a certain way. Whether this extension is defensible, and indeed whether the adverbial theory is viable anywhere, depends ultimately on how the theory is further spelled out. Recent work (see Tye, 1989) has supplied a clear semantics and metaphysics for the theory with the result that the adverbial approach is no longer open to the charge that it is just a rather trivial grammatical transformation without any real constraints. Indeed, once fully elucidated, the adverbial theory is seen to be a very powerful and well-founded approach which has the resources to answer all the more obvious objections. b i b l i og rap hy Ducasse, C.J.: “Moore’s Refutation of Idealism,” in Philosophy of G.E. Moore, ed. P.A. Schilpp (Chicago: Northwestern University Press, 1942), 225–51. 104

Chisholm, R.M.: Perceiving (Ithaca, NY: Cornell University Press, 1957). Sellars, W.: Science and Metaphysics (London: Routledge and Kegan Paul, 1968), 9–28. Tye, M.: “The Adverbial Approach to Visual Experience,” The Philosophical Review 93 (1984), 195–225. Tye, M.: The Metaphysics of Mind (Cambridge: Cambridge University Press, 1989). michael tye

Alfarabi [al-Farabc] (c.870–950) Islamic logician, metaphysician, political philosopher, also wrote commentaries on Aristotle’s logical treatises and expositions of Plato’s and Aristotle’s philosophies. Alfarabi was the first to raise the question of how the philosopher writing in Arabic which has no copula, can do logic and supply precise vocabulary for the Greek concept of being. He proposes to use derivatives of wjd (to find) for all the functions of “to be”, in a stipulative fashion, including the most general sense of “being” (Shehadi, 1982, pp. 45– 51). Existents are divided by Alfarabi into the possible and the necessary. In the case of possible beings, existence is not a property and cannot be part of their essence (see essence/accident; essence and essentialism). Asked whether “Man exists” has a predicate, Alfarabi replied that for the logician, “exists” is a predicate in the proposition. But it is not a predicate to the investigator into the nature of things. However, in the case of the First, existence is Its essence, for It is the being necessary through itself. In Islamic philosophy Neoplatonic (see Neoplatonism) emanationism gets its first full statement by Alfarabi. Islamic Neo-platonists were influenced by an Arabic translation of a pseudo Theology of Aristotle which was in fact a summary of sections of Plotinus’ Enneads, as well as by a translation of the Liber de causis. The First is one, uncomposed, and beyond human knowledge. From its activity of thinking itself emerges the First Intellect which thinks itself as well as its source. 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 corresponding cosmic sphere. Of special interest is Alfarabi’s transformation of Aristotle’s active intellect into a separate entity between humankind and the First, one of the separate substances above the terrestrial sphere. While it still makes the knowable known, its cosmological status prepares the way for the eschatological and mystical roles that it plays in Islamic philosophy. b i b l i og rap hy Fakhry, M.: A History of Islamic Philosophy (New York: Columbia University Press, 1970). Hammond, R.: The Philosophy of Alfarabi and its Influence on Medieval Thought (New York: Hobson Book Press, 1947). Rescher, N.: Al-Farabi: An Annotated Bibliography (Pittsburgh: University of Pittsburgh Press, 1962). Rescher, N.: “Al- al-Farabc on the Question: Is Existence a Predicate?,” in his Studies in the History of Arabic Logic (Pittsburgh, PA: University of Pittsburgh Press, 1963) 39–42. Shehadi, F.: Metaphysics in Islamic Philosophy (Delmar, NY: Caravan Books, 1982). fadlou shehadi

Alston, William P. (1921– ) is an American philosopher who has made significant contributions to epistemology, philosophy of religion, and the realism–antirealism debate among other areas. Alston’s work in epistemology has focused primarily on Foundationalism, the nature of epistemic justification, the internalism–externalism controversy, sense perception, and religious epistemology. In philosophy of religion, Alston has argued that putative perceptual experience of GOD is epistemically on a par with putative perceptual experience of ordinary material objects. Alston uses this argument along with a detailed account of mystical experience, to defend the importance of experiential grounds for the justification of religious belief.

Recently, Alston has defended a realist conception of truth according to which (1) a statement is true if and only if what the statement says to be the case actually is the case, and (2) truth is an important or significant feature of reality. It often matters, and we do care, whether our beliefs are true. Alston has also defended a form of metaphysical realism against a number of objections including the idea that there is a unique description of the world, a commitment to the causal theory of reference, and physicalism. See also experience; realism; theories of truth. wri t i ngs Epistemic Justification (Ithaca, NY: Cornell University Press, 1989). Perceiving God (Ithaca, NY: Cornell University Press, 1991). A Realist Conception of Truth (Ithaca, NY: Cornell University Press, 1996). The Reliability of Sense Perception (Ithaca, NY: Cornell University Press, 1993). richard gallimore analysis Consider the following proposition. (1) To be an instance of knowledge is to be an instance of justified true belief not essentially grounded in any falsehood. (1) exemplifies a central sort of philosophical analysis. Analyses of this sort can be characterized as follows: (a) The analysans and analysandum are necessarily coexistensive, i.e., every instance of one is an instance of the other. (b) The analysans and analysandum are knowable a priori to be coextensive. (c) The analysandum is simpler than the analysans (a condition whose necessity is recognized in classical writings on analysis, such as Langford, 1942). (d) The analysans does not have the analysandum as a constituent. (e) A proposition that gives a correct analysis can be justified by the philosophical example-and-counter-example method, i.e., by generalizing from intuitions about 105

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 a varied and wide-ranging series of simple described hypothetical test cases, such as “If such-and-such were the case, would you call this a case of knowledge?” Thus, such an analysis is a philosophical discovery, rather than something that must be obvious to ordinary users of the terms in question. Condition (d) rules out circularity. But since many valuable quasi-analyses are partly circular (e.g., knowledge is justified true belief supported by known reasons not essentially involving any falsehood), it seems best to distinguish between full analysis, for which (d) is a necessary condition, and partial analysis, for which it is not. This core notion of analysis fits the intuitive idea the term “analysis” suggests, which is that something is analyzed by breaking it down into its parts (see Moore, 1903, sects. 8 and 10). But Moore also holds that analysis is a relation solely between concepts, rather than one involving entities of other sorts, such as linguistic expressions, and that in a true analysis, analysans and analysandum will be the same concept (see Moore, 1942). These views give rise to what is nowadays generally called “the” paradox of analysis: how can analyses such as (1) be informative? Philosophers have proposed various solutions, such as relaxing the requirement that analysans and analysandum are the same concept (Langford, 1942), and denying that (1) is genuinely informative to someone who fully grasps the concepts involved (Sosa, 1983). Regardless of how this paradox is to be handled, there are types of analysis other than that exemplified by (1). One such type of analysis involves an analysans and analysandum that are clearly epistemically equivalent and that hence do not raise the paradox discussed here, although they do raise a different paradox (see Ackerman, 1990). Other types of analyses include newlevel analysis, which aims at providing metaphysical insight through metaphysical reduction (for example, the analysis of sentences about physical objects into sentences about sense data (see Urmson, 1956, ch. 3), 106

and reformatory analysis, which seeks to reduce sloppiness and imprecision by replacing a concept considered in some way defective with one considered in the relevant way improved. Reformatory analysis makes no claim of conceptual identity between analysans and analysandum and hence gives rise to no paradox of analysis. Aside from the possibility of paradox, philosophers have raised various objections to analysis as a philosophical method. It is a commonplace to object that analysis is not all of philosophy. But, of course, the claim that analysis is a viable method does not amount to saying that it is the only one. Wittgenstein (see Wittgenstein, 1968, especially sects. 39–67) has raised objections to the atomist metaphysics and epistemology underlying Russellian new-level analysis (see logical atomism; russell). But most of these objections do not apply to other types of analysis. It can also be objected that it is virtually impossible to produce an example of an analysis that is both philosophically interesting and generally accepted as true. But virtually all propositions philosophers put forth suffer from this problem. (See Reschler, 1978; Ackerman, 1992a.) The hypothetical example-and-counterexample method the sort of analysis (1) exemplifies is fundamental in philosophical inquiry, even if philosophers cannot reach agreement on analyses and often even individually cannot give full analyses and have to settle for less, such as one-way conditionals, partially circular accounts, and accounts (like that of being a game) that are justified in the same general way as analyses but that are too open-ended even to purport to yield necessary and sufficient conditions. bibl i ography Ackerman, F.: “Analysis and Its Paradoxes,” in The Scientific Enterprise: The Israel Colloquium Studies in History, Philosophy, and Sociology of Science, vol. 4, ed. E. Ullman-Margalit (Norwell, MA: Kluwer, 1992b). Ackerman, F.: “Analysis, Language, and Concepts: The Second Paradox of Analysis,” in 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 (Atascadero, CA: Ridgeview, 1990), 535– 43. Ackerman, F.: “Philosophical Knowledge,” in A Companion to Epistemology, ed. J. Dancy and E. Sosa (Oxford: Blackwell, 1992a), 342–5. Langford, C.H.: “The Notion of Analysis in Moore’s Philosophy,” in The Philosophy of G.E. Moore, ed. P.A. Schilpp (Evanston, IL: Northwestern University Press, 1942), 319– 43. Moore, G.E.: Principia Ethica (New York– London: Cambridge University Press, 1903). Moore, G.E.: “A Reply to My Critics,” in The Philosophy of G.E. Moore, ed. P.A. Schilpp (Evanston, IL: Northwestern University Press, 1942), 660–7. Rescher, N.: “Philosophical Disagreement,” Review of Metaphysics 22 (1978), 217– 51. Sosa, E.: “Classical Analysis,” Journal of Philosophy 53 (1983), 695–710. Urmson, J.O.: Philosophical Analysis (Oxford: Oxford University Press, 1956). Wittgenstein, L.: Philosophical Investigations, 3rd edn., ed. and trans. G.E.M. Anscombe (New York: Macmillan, 1968). felicia ackerman Anscombe, G.E.M. (1919–2001) G.E.M. Anscombe is a philosopher of great range, many of whose important contributions to philosophy lie in metaphysics and in fields which substantially overlap metaphysics, especially philosophy of logic and philosophy of mind. In “Causality and Determination” (1971) she questioned a central assumption made in virtually all philosophical writing about causation (see the extended essay), namely, “If an effect occurs in one case and a similar effect does not occur in an apparently similar case, there must be a further relevant difference.” The most disparate views of causation, from Aristotle’s and Spinoza’s to Hobbes’s, Hume’s, and Russell’s, all accept that causation involves universality or necessity or both; but Anscombe argues that such views cannot stand up. She shows the core idea in

causation to be that of derivativeness, exemplified by making a noise, pushing, wetting. Her view that these causal notions do not involve universality or necessity might be questioned, so she examines different sorts of examples, like Feynman’s case of a bomb which may be caused to explode by some radioactive emission. The absence of necessitation is irrelevant to the causing of the subsequent explosion. Anscombe also examines the relevance of non-necessitating causes to freedom of the will (see the extended essay on free will). She has discussed the subject of causation in several other essays. An important theme is the different kinds of causal relation (see, for example, 1974a). In “Times, Beginnings and Causes” (1974b), she examines Hume’s claim that it is logically possible for something to begin to exist without a cause. She develops an argument of Hobbes’s to show how judgments about beginnings of existence depend on the application of causal knowledge. Among the other topics in metaphysics which she has discussed is that of the self. In “The First Person” (1975), she argues that Descartes’s view of the self would be correct if “I” were genuinely a referring expression, but that it is not a referring expression. Metaphysical problems concerning time and substance are the focus of some of her other essays. writings “Aristotle: The Search for Substance,” in G.E.M. Anscombe and P.T. Geach, Three Philosophers (Oxford: Blackwell, 1963), 5–63. “Causality and Determination” (inaugural lecture at Cambridge University, Cambridge, 1971); repr. in (1981), vol. II, 133–47. Collected Papers (Oxford: Blackwell, 1981; Minneapolis: University of Minnesota Press, 1981), 3 vols.; papers bearing on metaphysics are in Vol. I, From Parmenides to Wittgenstein and Vol. II, Metaphysics and the Philosophy of Mind. “The First Person,” in Mind and Language: Wolfson College Lectures 1974, ed. 107

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, 1975), 45–65; repr. in (1981), vol. II, 21–36. “Memory, ‘Experience’ and Causation,” in Contemporary British Philosophy, ed. H.D. Lewis (London: George Allen and Unwin, 1974[a] ) 15–29; repr. in (1981), vol. II, 120–30. “Times, Beginnings and Causes,” Proceedings of the British Academy 60 (1974[b] ) 253– 70; repr. in (1981), vol. II, 148–62. b i b l i og rap hy Diamond, C. and Teichman, J., ed.: Intention and Intentionality: Essays in Honour of G.E.M. Anscombe (Brighton: Harvester Press, 1979). cora diamond

Anselm of Canterbury, St. (1033–1109) Scholastic philosopher and Archbishop of Canterbury, born at Aosta, Italy. Like Augustine before him, Anselm is a Christian Platonist in metaphysics (see Platonism). In the Monologion, he deploys a cosmological argument for the existence of the source of all goods, which is Good per se and thus supremely good, identical with what exists per se and is the Supreme Being. In Proslogion c.ii, Anselm advances his famous ontological proof: namely, that a being a greater than which cannot be conceived exists in the understanding, since even a fool understands the phrase when he hears it; but if it existed in the intellect alone, a greater could be conceived which existed in reality. A parallel reductio in c.iii concludes that a being a greater than which cannot be conceived exists necessarily. And in his Reply to Gaunilo, he offers a modal argument for God’s necessary existence, based on the premise that whatever does not exist is such that if it did exist, its non-existence would be possible. God is essentially whatever it is – other things being equal – better to be than not to be, and hence living, wise, powerful, true, just, blessed, immaterial, immutable and eternal per se; even the paradigm of sensory goods – Beauty, Harmony, 108

Sweetness and Pleasant Texture, in its own ineffable manner. Nevertheless, God is supremely simple, omne et unum, totum et solum bonum, a being a more delectable than which cannot be conceived. God is both the efficient cause of everything else and the paradigm of all created natures, the latter ranking as better in so far as they are less imperfect ways of resembling God. Such natures have a teleological 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 for which f’s were made) and established by God. From teleology, Anselm infers a general obligation on all created natures (non-rational as well as personal): since they owe their being and well-being to God as their cause, so they owe their being and well-being to God in the sense of having an obligation to praise Him by fulfilling their teloi. Anselm’s distinctive action theory reasons that if the telos of rational natures is unending beatific intimacy with God, their powers of reason and will have been given to promote that end. Thus, the will’s freedom must be telos-promoting, and – since sin is deviation from the telos – should not be defined as a power for opposites (the power to sin and the power not to sin), but rather as the power to preserve justice for its own sake (see the extended essay on free will). Choices are imputable only if spontaneous (from the agent itself). Since creatures have their natures from God and not from themselves, they cannot act spontaneously by the necessity of their natures. To enable creatures to be just of themselves, God endowed them with two motivational drives toward the good – the affectio commodi, or tendency to will things for the sake of their benefit to the agent itself: and the affectio justitiae, or tendency to will things because of their own intrinsic value. It is up to the creature whether or not to align them (by letting the latter temper the former). Anselm’s motivational theory contrasts sharply with Aquinas’s Aristotelian account, but was taken up and developed by Duns Scotus. See also god.

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

ant i nomi es writings Adams, R.M.: “The Logical Structure of Anselm’s Arguments,” The Philosophical Review 80 (1971), 28–54. Anselm of Canterbury, Vols. 1–3 (Toronto and New York: Edwin Mellor Press, 1974–6). Henry, D.P.: The Logic of Saint Anselm (Oxford: Clarendon Press, 1967). Hopkins, J.: Anselm of Canterbury, Vol. 4: Hermeneutical and Textual Problems in the Complete Treatises of St. Anselm (Toronto and New York: Edwin Mellen Press, 1976). Kane, G.S.: Anselm’s Doctrine of Freedom of the Will (New York and Toronto: Edwin Mellen Press, 1981). S. Anselmi. Opera Omnia, ed. F.S. Schmitt (Edinburgh: Thomas Nelson and Sons, 1946–61), vols. I–VI. marilyn mccord adams

antinomies An antinomy is a pair of apparently impeccable arguments for opposite conclusions. Obviously, the arguments cannot both be sound because a proposition and its contradictory must have opposite truth values. Thus the two appearances of cogency are not “all things considered” judgments because conflicting appearances cancel out. The challenge posed by an antinomy is at the level of adjudication and diagnosis. We know that at least one arm of the antinomy is fallacious. But which? And exactly where does it go wrong? “Antinomy” is most closely associated with Immanuel Kant’s attack on metaphysics. In the Critique of Pure Reason, he lays out parallel arguments literally side by side to emphasize their utter deadlock. As long as we assume that things-in-themselves are objects of knowledge, we can mount a knock-down argument for the thesis that the world has beginning in time and a knock-down argument for the antithesis that the world has no beginning. Metaphysicians can prove that we are free by exposing the absurdity of an actual infinity of past events and metaphysicians can disprove our freedom by demonstrating the incoherency of a break in the causal order.

This embarrassment of riches constitutes the data for Kant’s meta-argument in favor of the critical point of view: instead of aiming at knowledge of a mind-independent reality, we should abandon the classical metaphysical enterprise and restrict the objects of knowledge to appearances. We can then see that the antinomies are a product of transcendental illusion which arises from the temptation to apply the principles that constitute the framework for knowledge of phenomenal reality to noumenal reality (see noumenal/phenomenal). Contemporary philosophers do not share Kant’s awe at the cogency of the clashing arguments. Indeed, cosmologists and infinitistic mathematicians dismiss the pros and cons about the extent of space and time as amateurish fallacies. However, Kant’s unconvincing choice of examples does not undermine the philosophical interest of the concept of an antinomy. After all, “apparent” needs to be relativized to epistemic agents. An antinomy for an eighteenth-century figure need not be an antinomy for a twentiethcentury thinker. In any case, there certainly are argumentative deadlocks. Recently, a Japanese group of topologists announced a result that contradicted the result of an American group of topologists. Since both proofs involved complex calculations, they exchanged proofs to check for mistakes. Despite their high motivation and logical acumen, neither team has been able to find an error in the other’s reasoning. The Japanese–American deadlock is not an antinomy if it is caused by a slight but subtle slip. The appearance of cogency must be due to a “deep error” – not a mistake due to bad luck or ignorance surmountable by merely mechanical methods. Metaphysicians have a particular interest in antinomies that turn on false existential presuppositions. The Barber paradox features a village in which a barber shaves all and only those people who do not shave themselves. Does the barber shave himself? First argument: if the barber shaves himself, then he is a self-shaver. But he only shaves those who do not shave themselves. Therefore, the barber does not shave himself. Second argument: if the barber does not 109

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 nonself-shavers. But he shaves all those who do not shave themselves. Therefore, the barber does shave himself! The lesson to be learned from this modest antinomy is that the barber cannot exist. More ambitious resolutions of antinomies aim at a more dramatic impact on our ontology or cosmology. The paradox of the stone (can God make a stone so large that He Himself could not lift it?) is used to disprove God’s existence. The Buddhists use antinomies to disprove the existence of the self. The Eleatics (see presocratics) and nineteenth-century idealists (see idealism) deployed antinomies against the assumption that material things exist and that they are spatially related. Other antinomies turn on false dichotomies. For example, the old arguments for and against infinite space tended to assume that “finite” and “unbounded” were mutually exclusive terms. Albert Einstein’s application of Riemannian geometry makes sense of a “spherical” universe that is finite but unbounded. So besides subtracting entities and relationships from metaphysical systems, antinomies enrich these systems by stimulating the discovery of new entities and possibilities. An antinomy cannot prove anything on its own. Indeed, its internal conflict makes it a paradigm of dialectic impotence. However, the meta-arguments that grapple with antinomies are powerful tools of metaphysical inquiry. See also aporia; sorites arguments; transcendental arguments; Zeno. b i b l i og rap hy al-Asm, S.J.: The Origins of Kant’s Arguments in The Antinomies (Oxford: Oxford University Press, 1972). Kant, I.: Critique of Pure Reason (Riga, 1781); trans. N. Kemp Smith (London: Macmillan, 1933). Quine, W.V.: The Ways of Paradox (Cambridge, MA: Harvard University Press, 1976). Sainsbury, R.M.: Paradoxes (Cambridge: Cambridge University Press, 1988). 110

Sorensen, R.: Pseudo-Problems (London: Routledge, 1993). roy a. sorensen antirealism By “antirealism” we mean here semantic antirealism, of the kind advanced by Dummett in numerous writings. The main thesis of semantic antirealism is that we do not have to regard every declarative statement of our language as determinately true or false independently of our means of coming to know what its truth value is. That is, the semantic antirealist refuses to accept the principle of bivalence. ant i real i sm is not a form of i deal i sm or nomi nal i sm Semantic antirealism is to be distinguished from ontological antirealism. Ontological antirealism casts doubt on the existence of objects. It comes in varying degrees. The ontological antirealist may doubt the existence of any objects in the external world (idealism); or, more modestly, doubt the existence of the unobservable entities posited by science (van Fraassen’s constructive empiricism (1980)); or, more traditionally, doubt the existence of abstract objects, such as numbers (see number), or of universals (nominalism). Semantic antirealism is compatible with both Platonistic (see Platonism) and nominalistic views about numbers. In the case of mathematics, G. Kreisel’s dictum is often stressed: what one is concerned with is not so much the existence of mathematical objects, as the objectivity of mathematical statements. ant i real i sm is compat i ble wi t h nat ural i sm Indeed, one might even maintain that it is a consequence of naturalism. By naturalism we mean the metaphysical view that all things, events, states and processes are material or physical. Naturalism asserts supervenience, but does not claim reductionism (see reduction, reductionism). It asserts that all mental, moral, semantic and social facts supervene on material or

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

antirealism physical facts. The physical facts, that is, fix the mental, moral, semantic and social facts. But naturalism does not claim that psychology, moral theory, semantics or the social sciences can be reduced to physics. On the contrary, each of these special sciences is autonomous. Each presents important aspects of reality in its own terminology. Indeed, antirealism itself is a theory whose content would be lost were it not formulated in its own special terms, terms which defy reduction to physics. a n t i r e a lism str esses o b s e r v a ble b ehavio r a s t h e s o ur ce o f meaning The antirealist is centrally concerned with grasp of meanings, or contents (see content); and with the conditions under which speakers and thinkers can acquire such grasp and display it. It lays great stress on what have become known as the acquisition and manifestation arguments. These arguments are used to cast doubt on the claim, concerning sentences in any given area of discourse, that their meanings consist in verification-transcendent truth conditions. For, if they did, so these arguments conclude, speakers of the language would never be able fully to acquire or display grasp of meaning. The observable conditions surrounding their discourse, and their own observable behavior, prevent such overly enriched contents from being grasped and assigned to sentences. The acquisition and manifestation arguments, as developed by Dummett, show most clearly the influence of the later Wittgenstein on Dummett’s thinking. a n t i r e a lism co ntr asted w i t h qu ineanism One way to understand antirealism is to consider how Quine and the antirealist react to an argument on which they both agree. The argument has three premises and a conclusion that they both reject: (1) Meaning is given by truth conditions. (2) Meaning is determinate.

(3) Truth is bivalent. (4) Grasp of meaning cannot be manifested fully in observable behavior. Both Quine and the antirealist agree on the first premise. Quine holds that meaning (via translation) is indeterminate, but that truth is bivalent. The antirealist, by contrast, holds that meaning is determinate, but that truth is not bivalent. ant i real i sm enjoins a molec ular, as opposed t o an hol i stic, theory of meani ng The antirealist believes in determinate sentential contents. He or she adopts a compositional approach. One familiar ground for this comes from theoretical linguistics, which rightly stresses our recursive, generative or creative capacity to understand new sentences as we encounter them. Another ground is that the opposing holistic view (see holism) simply cannot account for language learning. We do, it would appear, master language fragments progressively as learners, and are able to isolate or excise them for theoretical study later on. Meanings of words remain relatively stable under increase of vocabulary and during developments in our ability to produce and understand more complicated utterances. These considerations point to a compositional approach. ant i real i sm is conc erned wi t h normativi t y As we have just seen, the antirealist maintains determinacy of meaning. Precision about contents brings with it commitment to normative connections among them: their justification conditions and their entailments. One of the main aims of antirealism is to give an accurate picture of such contents as the speaker or thinker can genuinely grasp or entertain in thought, and convey in language. This means that antirealism has to have some answer to skeptical problems about the objectivity of rule following. For it is only by conforming to, or keeping faith 111

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 the speaker can claim to have mastered their meanings. a n t i r e alism favo r s r efo r mism r a t h e r than q uietism In particular, the antirealist critique of genuinely graspable meanings can be brought to bear on the meanings of the logical expressions of our language: the connectives and the quantifiers. The observable conditions of their use (especially in mathematics) concern the discovery, construction, presentation and appraisal of proofs. Central features of the use of logical expressions – in particular, their introduction rules – serve to fix their meanings. Other features need to be justified as flowing from the central features. We can justify the elimination rules, because these are in a certain sense in balance or harmony with the introduction rules. But on this model of meanings and how one comes to grasp them, there does not appear to be any justification for the strictly classical rules of reasoning, especially as they concern negation. There does not appear to be any justification for the Law of Excluded Middle (either a or not-a) or for the Law of Double Negation Elimination (from not-not-a infer a) or any of their equivalents. Thus the antirealist response has been to favor logical reform: crucially, to drop the strictly classical negation rules and opt for intuitionistic logic. Thus intuitionism is the main form of mathematical antirealism (see intuitionism in logic and mathematics). When the antirealist generalizes from the mathematical case, with its conditions of constructive proof, he or she looks for appropriate conditions of warranted assertability. t h e c h allenge o f an a n t i r e a list acco unt o f e m p i r i cal d isco ur se In moving to empirical discourse, and especially statements about other minds, one has to attend closely to the criteria in accordance with which one ventures any

112

informative claim. Here the situation is very different from mathematics. For in mathematics, once a statement is proved it remains proved. In empirical discourse, however, statements are defeasible. That is, they can be justified on a certain amount of evidence; but may have to be retracted or even denied on the basis of new evidence accreting upon the old. (A modern way of putting this is to say that they are governed by a non-monotonic logic.) There is also the familiar problem from the philosophy of science, that no general claim about natural kinds (see natural kind) can ever conclusively be proved. At best, such claims can be conclusively refuted; but no amount of humanly accessible evidence can entail them. The combination of defeasibility with this familiar asymmetry between proof and refutation makes particularly problematic the provision of a satisfactory antirealist account of meaning for empirical discourse. antirealism t ends t o be pi ecemeal, rat her than gl obal Most writers on antirealism try to explore its strengths and weaknesses on particular areas of discourse: mathematics, statements about other minds, statements about the past, counterfactual statements (see counterfactuals), and so on. In each area one looks critically at the observational basis on which one can acquire grasp of meaning. One examines the criterial structure governing how speakers venture, and are taken at, their words. One tries (if necessary) to deflate any overly realistic classical conception of how, in response to each such area of discourse, a mind-independent region of reality might inaccessibly yet determinately be. The realist sometimes complains that the antirealist is guilty of epistemic hubris in taking the human mind to be the measure of reality. The antirealist responds by charging the realist with semantic hubris in claiming to grasp such propositional contents as could be determinately truthvalued independently of our means of coming to know what those truth values are.

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

antirealism a n t i r e a lism is no t a cr ude f or m o f ver ificatio nism There was an old principle of the logical positivists (see logical positivism) which, over the years, fell into deserved disrepute. This was the verificationist principle that every meaningful declarative sentence was, in principle, decidable. That is, in grasping 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 sentence. Despite its emphasis on assertibility conditions, antirealism lays claim to no such principle. a n t i r e a lism str esses c om p o s i tio nally Antirealism stresses, instead of the positivists’ naive decidability principle, various canonical ways of establishing statements with prominent occurrences of expressions whose antirealistically licit meaning is at issue. (An example of this would be dominant occurrences of logical operators, in the context of their introduction rules.) Various such expressions could then be combined into a sentence which is meaningful but which the antirealistic need not claim is decidable. The sentence will be meaningful by virtue of the way those expressions are combined within it, and by virtue of their central meanings as conferred by those special contexts. It is at this point that modern antirealism is crucially influenced by the contribution of Frege to logical semantics. s u m m a r y o f main featur es o f t h e a n t i r ealist p o sitio n (1) refusal to accept the principle of bivalence; (2) behaviorist emphasis on the epistemology of linguistic understanding: acquisition and manifestation arguments; (3) confidence in the determinacy of sentence meaning, leading to a molecular as opposed to an holistic theory of meaning;

(4) stress on the compositionality of meaning, thereby allowing meaningful though undecidable sentences; (5) advocacy of some kind of logical reform, making one’s logic more intuitionistic or constructive; (6) a generally naturalistic metaphysical outlook, and a quietist demurral from extreme skeptical misgivings or theses in epistemology.

mai n alleged weaknesses i n t he antirealist position (1) Alleged failure to do justice to the intuition that the world is robustly independent of human cognitive faculties; (2) alleged failure to appreciate the strength of independent arguments to the effect that translation is indeterminate, that there can be no firm analytic/synthetic distinction, that meaning (such as it is) is graspable at best only holistically; (3) alleged failure to appreciate that, in so far as meaning is determined (by the antirealist’s own lights) by the use we make of our expressions, we should accordingly accept classical rules of inference (such as Double Negation Elimination) as justified by the very use we make of them; (4) alleged instability in the antirealist’s own argumentative strategy: why stop at intuitionism, for example? Why not go all the way to strict finitism? Why treat of decidability in principle rather than feasible decidability? (5) alleged failure to understand the semantic contribution of the negation operator in embedded contexts; (6) alleged failure to appreciate that there are, even within the constraints set by the antirealist, resources enough to secure the realist’s grasp of verificationtranscendent propositional contents; (7) alleged failure to appreciate that the semantic issue of logical reform is independent of the metaphysical and epistemological issues at the heart of antirealism.

113

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 v s . a n tir ealism deb ate Michael Dummett put forward his classic challenge to the principle of bivalence in his essay “Truth”. His defense of intuitionistic logic as the correct logic on an antirealist construal of mathematics was given in his essay “The philosophical basis of intuitionistic logic”. This treatment was amplified in the chapter on philosophical reflections in his book The Elements of Intuitionism (1977). He explored the implications of antirealism for statements about the past in his essay “The Reality of the Past”. Dummett’s essays are collected in his book Truth and Other Enigmas (1978). Dag Prawitz has provided an excellent exposition and amplification of Dummett’s line of argument in his paper “Meaning and Proofs: On the Conflict Between Classical and Intuitionistic Logic” (1977). Crispin Wright has written widely on antirealism in mathematics, on statements about the past and on statements about other minds. He has also treated the problems of criteria, defeasibility and the objectivity of rule following. See his book Wittgenstein on the Foundations of Mathematics (1980), and his collection of essays Realism, Meaning and Truth (1986). Neil Tennant, in his book Anti-Realism and Logic (1987), has extended the antirealist critique and the logical reform it arguably entails in favor of the system of intuitionistic relevant logic. He also explores antirealism as a consequence of naturalized epistemology. John Mcdowell has pursued subtle variations on realistic and antirealistic themes in his essays “Anti-Realism and the Epistemology of Understanding” (1981), and “Truth Conditions, Verificationism and Bivalence” (1970). Opposition by the realists has been led most notably by Peter Strawson (1976), Christopher Peacocke (1986) and J.J.C. Smart (1986). Saul Kripke gave great impetus to the debate about the objectivity of rule following with the publication of his provocative monograph Wittgenstein on Rules and Private Language (1982). Kripke adopts an antirealistic construal of contentattribution statements in his “sceptical solution”. 114

See also realism; the extended essay on realism and antirealism about abstract entities. bi bliography Dummett, M.A.E. (with the assistance of R. Minio): The Elements of Intuitionism (Oxford: Clarendon Press, 1977). Dummett, M.A.E.: Truth and Other Enigmas (London: Duckworth, 1978). Kripke, S.: Wittgenstein on Rules and Private Language (Oxford: Blackwell, 1982). McDowell, J.: “Anti-Realism and the Epistemology of Understanding,” in Meaning and Understanding, ed. H. Parret and J. Bouveresse (Berlin and New York: De Gruyter, 1981), 225. McDowell, J.: “Truth Conditions, Verificationism and Bivalence,” in Truth and Meaning, ed. G. Evans and J. McDowell (Oxford: Clarendon Press, 1976), 42– 6. Peacocke, C.: Thoughts: An Essay on Content (Oxford: Blackwell, 1986). Prawitz, D.: “Meaning and Proofs: On the Conflict Between Classical and Intuitionistic Logic,” Theoria 43 (1977), 2–40. Smart, J.J.C.: “Realism v. Idealism,” Philosophy 61 (1986), 295–312. Strawson, P.F.: “Scruton and Wright on Anti-realism, etc.,” Proceedings of the Aristotelian Society 77 (1976), 15–22. Tennant, N.: Anti-Realism and Logic (Oxford: Clarendon Press, 1987). Wright, C.J.G.: Realism, Meaning and Truth, 2nd ed. (Oxford: Blackwell, 1986). Wright, C.J.G.: Truth and Objectivity (Cambridge, MA: Harvard University Press, 1992). Wright, C.J.G.: Wittgenstein on the Foundations of Mathematics (London: Duckworth, 1980). neil tennant antirealism about abstract entities see the extended essay on realism and antirealism about abstract entities aporia An apory is a small set of individually plausible but jointly inconsistent propositions. Aporia gained initial popularity

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

appearance / reality from Chisholm’s demonstration of how they help to motivate and structure philosophical issues. For instance, he regiments the problem of ethical knowledge with a set containing the following three members: (1) We have knowledge of certain ethical facts. (2) Experience and reason do not yield such knowledge. (3) There is no source of knowledge other than experience and reason. To avoid inconsistency, thinkers need to reject at least one member of the set. Thus the skeptic denies (1), the naturalist rejects (2), while the intuitionist argues against (3). The aporetic cluster provides each position with a ready-made argument. For the negation of any member of the set is the conclusion of an argument containing the remaining members as premises. Since members of the original set are jointly inconsistent, the argument will be valid. And since the members are individually plausible, the audience will also find each premise of the argument persuasive. b i b l i og rap hy Chisholm, R.M.: Theory of Knowledge, 2nd edn. (Englewood Cliffs, NJ: Prentice Hall, 1977). roy a. sorensen appearance/reality Nothing is more commonplace than the remark that things are not always what they seem. We all know that a thing can appear to be some way and yet be really quite otherwise. Unlike some other distinctions philosophers are enamoed of, the distinction between appearance and reality is firmly rooted in everyday experience and discourse. It is not surprising, then, that it has, since the dawn of philosophy, served to structure debates about what there is to know and how, if at all, it can be known. When Socrates objected to the relativism of the Sophists, with its ugly moral consequences, it was their refusal to allow that there could be a gap between “x appears

to be F” and “x is F” that he had to show to be untenable. When Descartes, and after him, most thinkers of the modern era, struggled with the skeptic’s challenge, the threat posed by that challenge was the possibility that that same gap was too great, that no reliable evidence about reality was ever furnished by what appeared in experience. In part inspired by that challenge, one empiricist strain (see empiricism), strangely echoed in a late flowering of rationalism, concludes that what appears to the wellfunctioning mind (in perception or in reasoning) is, and must be, the real, and it must be just as it appears. Found in both Berkeley’s and Hegel’s form of idealism, this maneuver closes the gap the Sophists had ruled out, but does so from the opposite side. Where the Sophist insists that whatever appears must be real the idealist argues that only what is real can appear. For both, the real must be just as it appears to be; either way, the commonsense distinction is rendered philosophically moot (and needs to be explained, or explained away, in complicated and, to some, implausible, ways). In both its everyday and its philosophical versions, the appearance/reality distinction must be seen as a completely general one. While its most obvious illustrations involve sense perception, it extends naturally to all dimensions of thought and experience. It 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 it is a mistake to draw the distinction by identifying one side with one metaphysical category and the other with another, for example, the real with the material and appearances with the mental. What, then, is at the heart of the commonsense distinction, and what, if anything, is philosophically interesting about it? I mentioned the perennial skeptical worry about whether appearances can tell us whether there are things other than appearances and, if so, what they are like. Skepticism is an epistemological position. But the very idea that there is a way things are, whether or not one can know what that way is, expresses a metaphysical belief, usually labeled realism. Thus skepticism 115

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. What account can we give of the appearance/reality distinction that does justice to both these components? Here is where the notion of evidence can provide the needed general framework. An appearance is always an appearance to someone, just as a piece of evidence is always evidence for someone. The former notion, in fact, represents a special case of the latter. But the concept of evidence also involves the thought of something for which the evidence is evidence. Thought of in this way, so does the idea of an appearance, as the appearance of something. Even Kant, who insists on the “empirical” reality of what he calls “appearances”, arguably sometimes treats them as representing, albeit in a special and highly problematical sense, a “transcendent” reality (see noumenal/ phenomenal). It is a conceptual truth that even the best evidence must fall short of certainty (else it would not be evidence for something other than itself). In the same way, the very concept of an appearance requires it to be distinct from that of which it is an appearance. This is why the idealist attempt to identify reality with appearances, no matter how the latter are idealized, is a mistake. It involves a non-evidential, hence a non-epistemic, conception of appearances; in doing so, it loses contact with the point of the commonsense distinction out of which the philosophical one grows. What makes the appearance/reality distinction both important and slippery is that it straddles the division between epistemology and metaphysics. Other well-worn philosophical distinctions are either internal to one or another of the traditional divisions of the subject (particular/universal, necessary/contingent, a priori/a posteriori, or concrete/abstract) or indifferent to them (extrinsic/intrinsic, specific/general, objective/subjective). Thinking of the appearance/reality distinction in the evidential way as suggested here can save us from mistaking it for a metaphysical one, one between two different kinds of entity. There may be good reasons for thinking that there are appearances, as opposed to just the various ways the things there are appear. But 116

there are dangers in this reification of them (see hypostasis, reification). First, it can lead to intractable metaphysical problems that are in fact avoidable. Second, it misleads us as to the true nature of the distinction between appearance and reality. Only when understood as involving the relation between epistemological and ontological concepts can it both retain the intuitive content of the commonsense distinction and yield a general philosophical problem that is not the artefact of some special metaphysical doctrine. bi bliography Austin, J.L.: Sense and Sensibilia (Oxford: Oxford University Press, 1962). Ayer, A.J.: The Foundations of Empirical Knowledge (London: Macmillan, 1940). Burnyeat, M.: The Skeptical Tradition (Berkeley: University of California Press, 1983). Grayling, A.C.: The Refutation of Scepticism (La Salle, IL: Open Court, 1985). john biro Aquinas, St. Thomas (1224/5–74) The philosophy of St. Thomas Aquinas was strongly influenced by Aristotle and by the Islamic philosophers Avicenna and Averroes, whose works became available in Latin translations at the beginning of the thirteenth century. But Aquinas’s metaphysical thought contains a number of elements that are not to be found in his leading sources. t he subj ect mat t er of metaphysi c s Aristotle’s divergent statements on the nature of first philosophy led to an intensive discussion of the subject matter of metaphysics among medieval thinkers. In Metaphysics iv c.1 (1003a21–32), Aristotle speaks of a science which studies being as being and opposes it to other sciences which investigate beings from a particular point of view, for instance, in so far as they are mobile. The science of being as being, by contrast, is universal. But in Book vi c.1 (1026a23–32),

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

aqui nas, st . t homas Aristotle distinguishes three theoretical sciences – physics, mathematics and the “divine science” – and calls theology the first science, because it is concerned with immobile and immaterial beings. The medieval discussion is focused on the question how Aristotle’s theological conception of first philosophy is related to the conception of metaphysics as the universal science of being (see Zimmermann, 1965). In the prologue to his Commentary on the Metaphysics, Aquinas argues that metaphysics is concerned with both being as being and the immaterial substances, although not in the same way. He develops his synthesis with the help of the logician Aristotle, for Aquinas’s argument is based on the theory of science in the Posterior Analytics. The unity of a science consists in the unity of its subject (subjectum). What is sought in every science are the proper causes of its subject. Now the immaterial substances are the universal causes of being. Therefore, being in general (ens commune) and the properties belonging to it are the subject of metaphysics. God is studied in this science only in so far as he is the cause of the subject of metaphysics, that is, in so far as he is the cause of being as such. God is not the subject, but rather the end of metaphysical investigation. By this feature metaphysics is distinguished from Christian theology (“the theology of sacred scripture”), for the subject matter of this science is God himself (cf. Summa theologiae 1.1.7). From this account it appears that Aquinas does not adopt the theological conception of metaphysics that was prevalent among the Greek commentators on Aristotle. According to them first philosophy is the science of the most eminent being, the divine being. Aquinas’s view is ontological: metaphysics is the scientia communis, for its subject matter is being in general. t h e d o c tr ine o f the t r a n s c e nd entals Against the background of Aquinas’s ontological conception of metaphysics, the significance of a doctrine that was developed in the thirteenth century, the doctrine of

the transcendentals, becomes understandable, for transcendentia are the universal properties of being as such (see Aertsen, 1988). The term “transcendental” suggests a kind of surpassing or going beyond. What is transcended is the special modes of being which Aristotle called “the categories”. While for the latter the categories are the most general genera of being, Aquinas considers them as special modes of being, as contractions of that which is: not every being is a substance, or a quantity, or a quality, or a relation, etc. By contrast, the transcendentals express general modes of being. They transcend the categories, not because they refer to a reality beyond the categories but because they are not limited to one determinate category. Unlike the categories, the transcendentals do not exclude each other, but are interchangeable or convertible (convertibilis) with being and each other. In De veritate 1.1, Aquinas presents his most complete account of the transcendentals, of which the most important are being, one, true and good. Being is the first transcendental. The other transcendentals, although convertible with being, add conceptually something to being, in the sense that they express a mode of it which is not yet made explicit by the term “being” itself. The general mode of being expressed by “one” pertains to every being in itself (in se); “one” adds to being a negation, for it signifies that being is undivided. “True” and “good” are relational transcendentals: they express the conformity (convenientia) of every being to something else. The condition for this relation is something whose nature it is to accord with every being. Such is, Aquinas argues, the human soul, which according to Aristotle (De anima iii c.8, 431b21) “is in a sense all things”. In the soul there is both a cognitive power and an appetitive power. The conformity to the appetite or will is expressed by the term “good”, the conformity to the intellect by the term “true”. Truth as transcendental signifies the intelligibility of things. Aquinas’s innovation in the doctrine of the transcendentals is the correlation he introduces between the human soul and being. He understands the transcendentals true and 117

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 substance. This understanding means an acknowledgment of the special place human being has among other beings in the world. A human being is marked by a transcendental openness; its object is being in general. This openness is the condition of the possibility of metaphysics. The doctrine of the transcendentals plays a central role in Aquinas’s metaphysics. It integrates the theory of knowledge (“truth”) into an ontology and it provides the foundation for the first principle of morality: “good is to be done and pursued, and evil avoided” (Summa theologiae I–II.94.2). The doctrine is also fundamental for philosophical theology. Within the framework of a reflection on the divine names “Being”, “Unity”, “Truth” and “Good” Aquinas discusses the relation between the transcendentals and God. Because the transcendentals are selfevidently knowable, and because they do not express a limited, categorical mode of being, they are seen as providing the basis for the possibility of rational knowledge of God. t h e h i sto r y o f the q uestio n of being In Summa theologiae 1.44.2, Aquinas sketches the history of philosophical reflection about the origin of being. This text can be regarded as the medieval origin of the question “Why is there something and not rather nothing?” Three main phases can be distinguished in the progression of philosophy as Aquinas sees it. The first step was taken by the presocratics. They held that matter is the “substance” of things and that all forms are accidents. They posited one or more substrata (water, fire, etc.) which they regarded as the ungenerated and indestructible principle of all things. To the extent to which they acknowledged chance in the substratum, it consisted only in “alteration”, a change of its accidental forms (see matter/form). The second stage in the progress of philosophy was reached when philosophers made a distinction between “matter” and “substantial form”. They posited a prime matter 118

that is purely potential and is brought into actuality through a form. Aquinas regards it as one of Aristotle’s great merits that with his doctrine of the potentiality of matter he made it possible to acknowledge a substantial change, or “generation”. Aquinas emphasizes, however, that the final step had not yet been taken, for the generation, too, presupposes something, in keeping with a common supposition of Greek thought: “nothing comes from nothing” (ex nihilo nihil fit). The philosophers of the first and second phases considered the origin of being under some particular aspect, namely, either as “this” being or as “such” being. As a result, the causes to which they attributed the becoming of things were particular. Their causality is restricted to one category of being: accident (as in the first place), or substance (as in the second). The third phase began when “some thinkers raised themselves to the consideration of being as being”. In this metaphysical analysis they assigned a cause to things not only in so far as they are “such” (by accidental forms) and “these” (by substantial forms), but also as considered according to all that belongs to their being. The origin considered by the metaphysician is transcendental, it concerns being as such, not merely being as analyzed into natural categories. The procession of all being from the universal cause is not a generation, because it no longer presupposes anything in that which is caused. It is creation ex nihilo. A striking feature of Aquinas’s view of the progress of philosophy is that the idea of creation appears as the result of the internal development of human thought, independent of revelation. In the context of the idea of creation Aquinas elaborates two central ideas of his metaphysics: the composition of essence and existence in created things, and the doctrine of participation. t he composi t i on of essenc e and exi stence – participation The distinction between essence and existence (esse) was introduced by Islamic thinkers in order to explain the contingent character 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 what is caused, for it has received its being from something else. The relation between essence and existence was interpreted by Avicenna according to the model of substance and accident: esse is an accident superadded to essence. Aquinas teaches the real composition of essence and existence in all creatures already in one of his earliest works, De ente et essentia. In chapter 4 he discusses the essence of the “separated substances” or spiritual creatures. This issue engaged Aquinas a great deal – he even devoted a particular treatise to it. De substantiis separatis – for it concerns the ontological structure of finite substances. This structure cannot consist in the composition of form and matter, since spiritual substances are “separated” from matter. Yet although such substances are pure forms, they do not have complete simplicity. All creatures are composed of essence and existence, because they have their esse not of themselves, but from God. According to Aquinas, however, existence is not an accident superadded to essence. Existence and essence are related to each other as act to potency. He extends the notions of act and potency, which were correlative with the notions of form and matter in Aristotle, to being as such. In a famous text in his De potentia (7.2 ad. 9), Aquinas states: “That which I call esse is the actuality (actualitas) of all acts, and for this reason it is the perfection of all perfections.” For Aquinas to be is not a bare fact, but the ultimate act through which a thing achieves its perfection. “Every excellence of any thing belongs to it according to its esse. For man would have no excellence as a result of his wisdom unless through it he were wise” (Summa contra Gentiles 1.28). It was Gilson (1949) in particular, who has emphasized the existential character of Aquinas’s metaphysics against the dominant “essentialist” tradition in modern philosophy. Closely connected with the distinction of essence and existence in created things is Aquinas’s doctrine of participation. No finite being is its esse, but has it. Only in God are essence and existence identical: he is essentially Being. All other things participate in

being. One of the most significant innovations in Thomistic scholarship since the Second World War has been the discovery of the “Platonist” Thomas (see Plato, Platonism). Pioneering studies were the works of Fabro (1961) and Geiger (1942), which showed the central role of the Platonic notion of participation in Aquinas’s metaphysics, a notion that was sharply criticized by Aristotle. Aquinas interprets the idea of creation philosophically in terms of participation. The relation of creatures to the first cause is the relation of participation in being. wri t i ngs The critical edition of Aquinas’s works, the Leonine edition, is still unfinished: Opera omnia. Iussu impensaque Leonis XIII, P.M. edita (Rome: Vatican Polyglot Press, 1882–). For a complete listing of the various editions, see J.A. Weisheipl, Friar Thomas d’Aquino (Washington, DC: The Catholic University of America Press, 1983), 355– 404. bibl i ography Aertsen, J.A.: Nature and Creature. Thomas Aquinas’s Way of Thought (Leiden: Brill, 1988). Fabro, C.: Participation et causalité selon S. Thomas d’Aquin (Louvain and Paris: Nauwelaerts, 1961). Geiger, L.-B.: La Participation dans la philosophie de S. Thomas d’Aquin (Paris: Vrin, 1942); 2nd edn. (Paris: Vrin, 1953). Gilson, E.: Being and Some Philosophers (Toronto: Pontifical Institute of Mediaeval Studies, 1949); 2nd edn. (Toronto: Pontifical Institute of Mediaeval Studies, 1952). Wippel, J.F.: Metaphysical Themes in Thomas Aquinas (Washington, DC: The Catholic University of America Press, 1984). Zimmermann, A.: Ontologie oder Metaphysik? Die Diskussion über den Gegenstand der Metaphysik im 13. und 14. Jahrhundert (Leiden: Brill, 1965). jan a. aertsen 119

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

a r c h e t yp e archetype From the Greek JρχKτυπον, a pattern or model. The word is applied to the reality – whether in the mind of God in nature itself, or in a third, abstract realm – to which a conception is referred. Archetypes sometimes play a causal role in originating those conceptions; their reference or truth is then assured (or at the very least argued for) by their causal ancestry. The Greek word was applied by Platonists (though not by Plato himself, who spoke instead of παραδε Oγματα) to the forms (see Platonism). Later Platonists placed these forms or archetypes in the mind of God. Philosophers of the seventeenth and eighteenth centuries conceived of them more broadly. Descartes described the external cause of an idea as “like an archetype”. Locke applied the word to the things the mind “intends [its ideas] to stand for, and to which it refers them”. Berkeley applied the word, with some reluctance, to ideas in the mind of God, which were, he argued, no less serviceable as archetypes than the corporeal substances of the materialists. b i b l i og rap hy Berkeley, G.: Philosophical Correspondence Between Berkeley and Samuel Johnson 1729–30 (New York, 1929); repr. in The Works of George Berkeley, vol. 2, ed. A.A. Luce and T.E. Jessop (London: Thomas Nelson, 1948–57), 271–94. Berkeley, G.: “Third Dialogue,” in Three Dialogues between Hylas and Philonous (London, 1713); repr. in The Works of George Berkeley, Bishop of Cloyne, vol. 2, ed. A.A. Luce and T.E. Jessop (London: Thomas Nelson, 1948–57), 227–63. Descartes, R.: “Third Meditation,” in Meditations on First Philosophy (Paris, 1641); repr. in The Philosophical Writings of Descartes, vol. 2, ed. and trans. J. Cottingham, R. Stoothoff, and D. Murdoch (Cambridge: Cambridge University Press, 1984), 29. Locke, J.: An Essay Concerning Human Understanding (London, 1689); ed. P.H. Nidditch (Oxford: Clarendon Press, 1975), Book II, ch. xxxi, sects. 1–3. Plato: Republic V, VI (many versions). kenneth p. winkler 120

Aristotle (384–322 bc) Greek philosopher born in Stagira. Aristotle’s writings can be said to have set the agenda for the western tradition in metaphysics. Indeed, “metaphysics” is a term derived from a first century bc edition of Aristotle’s work, in which a collection of his writings was put together under the title Ta Meta ta Phusika, which means simply “What comes after the writings on nature” (ta phusika). Since the writings thus put together concerned topics that seemed in certain ways related – substance and being, change and explanation, unity and plurality, potentiality and actuality, non-contradiction, the nature of the eternal and unchanging – these topics were subsequently taken to be the subject matter of “metaphysics”, which increasingly became a separate department of philosophy. But Aristotle himself did not group these topics together. He does have a conception of “the study of being qua being” – the study of what is true of all things that are, as such – that links some of the contents of the Metaphysics. But there is dispute about what that study is, and how much of the work it includes. Nor are Aristotle’s inquiries into the topics we now call metaphysical confined to the work called Metaphysics. There is an especially close link between that work and his inquiries into natural change and explanation. subst anc e, change and i dent i t y Aristotle once remarked that the central concern of previous philosophers, when they asked questions about what “being” is, was really, at bottom, a question about what substance is. (The term we translate “substance” is ousia, a verbal noun formed from the participle of the verb “to be”.) “For it is this,” he continues, “that some claim to be one in number, some more than one, and some limited, others unlimited.” He himself devotes much effort to the task of finding an adequate account of “substance”, and on defending the priority of substance to other items such as qualities and materials. It is not, however, intuitively obvious what Aristotle means by the question, “What is 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 no clear history. We must search in his arguments and examples for an understanding of his motivation and goal: to what real puzzles does such a search respond? As Aristotle characterizes earlier inquiries into substance, they focus on two questions, rather closely related: (1) a question about the explanation of change; and (2) a question about identity. We observe many changes in the world around us, such as the cycle of the seasons, the birth, growth and death of living creatures. Early Greek mythology explained these changes by invoking the capricious will of anthropomorphic beings; early philosophers, instead, looked for lawlike explanations. In the process, they had to ask themselves, first, what sorts of entities are relatively stable and persisting, the things to which changes happen and in terms of whose underlying stability change could be coherently explained. (Plato had cogently argued that coherent talk about change presupposes at least some stability: for a change has to be the change of something, and that thing cannot at the same time be ceasing to be the thing it is, or we will not be able to say anything about it.) The search for substance is, in part, a search for these most basic persisting entities (see the extended essay on persistence), which Aristotle calls “substrata” or “subjects” (two different translations of his Greek term hupokeimenon, literally “that which underlies”). The second question is what Aristotle calls the “What is it?” question. It may be illustrated by countless common examples. Suppose I am considering some particular thing in my experience, say, Socrates. I have a sense that, in order to pursue my curiosity about this thing further, I must have some answer to the question, “What is this?” I want to know what it is about this thing that makes it the thing it is, what enables me to single it out as a distinct particular and mark it off from its surroundings, to reidentify it later as the same thing I encountered earlier. But to know this I need, it seems, to separate the attributes of the thing into two groups: properties (such as a sun-tan, or knowledge of history) that may come to be present, or depart, without affecting

Socrates’ persistence as the same entity, and properties (such as, perhaps, the ability to metabolize food, or the ability to think and choose) whose presence is constitutive of the individual’s identity, whose departure would mean the end of an individual. The identity question has a special urgency where living creatures are concerned, since it is connected with complicated ethical and political issues, for example, the determination of death and the moral status of the foetus. Thus Aristotle holds that a creature dies whenever it loses one of the properties in the second group (the “essential” properties); and he holds that the foetus at an early stage of life is not a human being, and does not exhibit identity with the human being that may in due course come to be, since it does not yet have all the essential properties of the human being. In one way, these two questions seem to point in opposite directions, identifying different things as “the substance” of a thing. For the question about persistence through change might lead us to hold that material stuffs are the basic substances of the things they compose, seeing that these stuffs (for example, the materials that make up the body of Socrates) pre-exist the birth of Socrates and post-date his death. On the other hand, for this very reason they do not give the answer to questions about necessary and sufficient conditions for Socrates’ identity. We are inclined, there, to look in the direction of the structure characteristic of Socrates’ species, his human make-up and functioning. For it seems that it is the disruption of those modes of organization that spells the end of his existence. On the other hand, looked at in another way the two questions seem to be closely intertwined. An adequate theory of change must single out, as its substrates, things that are not only relatively enduring, but also definite and distinct. Unless we can individuate an item from its surroundings and say something about what it is, it will be difficult to make it the cornerstone of an explanatory enterprise. And a good answer to the “What is it?” question, asked about a particular such as Socrates, must tell us, among other things, what changes Socrates 121

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 one and the same. As Aristotle sees it, his predecessors went wrong because they pursued one prong of the substance inquiry to the neglect or distortion of the other. Early natural scientists, seeing that material stuffs were the most persisting things around, surviving the deaths of humans and animals, held that these were the real substance of things and the best answer to the “What is it?” question, when asked about particular substances. What Socrates really is, is the materials that compose him. This leads to paradoxical conclusions: no substance ever perishes, and substances continue to exist although their parts are widely dispersed in space and time. Above all, this view fails to capture a distinction that is fundamental in our discourse and practices, namely the distinction between property change (alloiosis) and real coming-into-being and goingout-of-being (genesis and phthora), between Socrates getting a sun-tan and the death of Socrates. Platonists (see Platonism), on the other hand, focus on the identity question, and on the universals that are, as they see it, the best answer to that question. Each aspect of Socrates is explained by his “participation” in some universal “form”, such as the form of Justice, which is imagined as existing apart from particulars and as explaining the possession of that property in all the particulars that have it. Aristotle finds fault with this emphasis on the universal, because it fails to come to grips with the material changing character of the individual substance. Nor, in its Platonic form at least, does this approach even succeed in separating universals such as the Human, which must be true of Socrates as long as he exists, from universals such as the White, which he might lose (getting a suntan) while still remaining the same individual. In his early work, the Categories, Aristotle focuses on two tasks: demarcating the role of particulars and universals in answering “What is it?” questions about things, and defending the central role of natural-kind concepts in answering both change and identity questions. The famous enumeration 122

of ten “categories” or (literally) “predications” is an attempt to enumerate different ways we might characterize a particular in our experience: we might speak about its substantial nature, its quantity, its qualit(ies), its relation(s), its place, time, position, state, activity, passivity. At the same time, Aristotle also introduces a fourfold distinction of “things that are”, separating (1) universals in the substance category, called Secondary Substance – e.g., human being, horse; (2) particulars in nonsubstance categories, such as this item of knowledge, this instance of pink color; (3) universals in non-substance categories, such as knowledge, color; (4) particulars in substance categories, called Primary Substance, e.g., this human being, this horse. The motivation for these distinctions emerges when Aristotle explains the fundamental classifying role of natural-kind universals. His point is that we do not pick things out and trace them through time as bare unclassified matter; fundamental to our practices of identifying and explaining is the ability to say to what kind the thing belongs. (His later writings give natural kinds a special place here, since artefacts have comparably unclear criteria of identity.) When we point at Socrates and say, “What is it?” we are asking about a particular, and it is that particular thing that exists; classifying universals have no existence apart from particulars. But the universal is of fundamental importance in coming to grips with the particular’s identity – and not just any universal, but the one, “human being”, that gives the kind to which he belongs from birth to death. To answer, “Socrates is a sitting thing”, or “Socrates is a white thing”, is a less revealing answer, parasitic on our ability already to pick out Socrates as a human being. In short: the category of substance, which includes the natural-kind universals and the particulars that fall under them, has priority over the other categories in both explaining and identifying. Within this category, particulars in a sense take priority, as the most basic substrates of change; but they get their identity from the universal 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 So far, Aristotle has said nothing about the coming-to-be and passing-away of substances. Nor has he spoken about the matter that composes them. To these tasks he turns in Physics i 7–9 and in Metaphysics vii. He acknowledges that living substances are essentially enmattered structures: they cannot continue as the things they are without suitable matter to make them up and perform their life-activities. On the other hand, he insists that matter all by itself cannot give us the identity of a particular: for it is a mere “lump” or “heap” without the form or structure that it constitutes. Nor, indeed, despite matter’s purported claim to be the substrate par excellence, does matter even turn out to be as continuous as form, with respect to the individual species member: for the matter that composes Socrates is changing continually, as he eats and excretes, while he himself remains one and the same. Looking more closely into the question of what does provide Socrates with his identity over time, Aristotle’s answer is that this is his “essence”, and that this essence is a particular instance of characteristic species organization or “form” (see hylomorphism), not different in kind from that of other species members, but a countably different instance, tracing a distinct career through time and space. (There are many different interpretations of Aristotle’s final position on the contribution of the universal and the particular in identity, but this one has broad support.) In later books of the Metaphysics Aristotle investigates the role of form in making a thing a unity, providing still further arguments against thinking of material stuffs as what a thing is. Introducing the important ideas of capability or potentiality (dunamis) and activity or actuality (energeia), he argues for the explanatory priority of a thing’s actual nature to its potentialities. Aristotle here begins to think about matter as a set of potentialities for functioning, which can be explicated only when we have grasped the actual functional structure of the entity that matter composes.

The famous twelfth book of the Metaphysics then gives an account of god as an immortal immaterial substance whose entire form is thinking, and whose entire being is actuality rather than potentiality. God imparts movement to the universe by being an object of passionate love to the heavenly bodies, who are themselves imagined to be living thinking beings. bei ng qua bei ng and the basic princ i ples of thought In Book iv of the Metaphysics, Aristotle defends the idea of a general study of the attributes of things that are as such, or of “being qua being” – an idea that he seemed to attack in some earlier writings as insufficiently attentive to the multiplicity of types of being. Here, by contrast, he argues that the many ways in which we speak of “being” have more than a verbal unity: for all are understood through an inquiry into substance, which is in some sense the basic type of being in our explanation and understanding of the world. Aristotle’s project here has been understood in two very different ways. Some interpreters understand him to be calling for a general study of substances, focusing in particular on living creatures, and for an illumination of properties, of activity and passivity, and so forth, that would be based upon that understanding. Others have understood him to be referring to god as the primary and central substance, a study of which is the focal point for all study of substance. The fact that the relevant texts of the Metaphysics derive from different periods in Aristotle’s life and are not edited into their present order by him makes resolution of this question very difficult. One can at least say, however, that in the central books in which Aristotle does in fact investigate the nature of substance (Books vii–ix), there is no discussion of god, and no sign that we need to understand the nature of god before answering questions about form’s relation to matter. The same is true of the De anima, where bodiless substance is an anomaly, briefly mentioned, in the work’s systematic 123

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 form and matter (see hylomorphism). Aristotle then goes on to argue that in any inquiry whatever, a basic role is played by two logical principles: the principle of non-contradiction and the principle of the excluded middle. Formulating NonContradiction as the principle that contradictory predicates cannot apply to a single subject at the same time in the same respect, Aristotle argues that this is “the most secure starting point of all”, concerning which “it is impossible to be in error”. Confronting an opponent who claims to doubt the principle (apparently a relativist who holds that if x seems F to observer O, x simply is F, and if x to observer P not to be F, x simply is not F), Aristotle argues that this opponent himself refutes himself, if he utters any coherent sentence, or even any definite word. For any meaningful utterance must, in putting something definite forward, at the same time implicitly rule out something – at the very least, the contradictory of what is put forward. He adds that if the opponent is silent and refuses to say anything definite, he loses this way too: for he is “pretty much like a vegetable”, and it is “ridiculous to look for words to address to someone who doesn’t use words”. Moreover, even definite action without words reveals a commitment to NonContradiction: for when one acts one must have some definite belief about what one is aiming to do, and such beliefs, propositional in form, presuppose a commitment to Non-Contradiction. m e t h odo lo gy: ap p ear ances a n d u n der stand ing In passages such as the one from Metaphysics iv just discussed, Aristotle appears to derive support for what he calls “the most basic principle of all” simply by showing its depth and ubiquity in our discourse and practices. And elsewhere he states that in all inquiries the aim should in fact be, first to “set down the appearances” – by which he seems to mean the record of human experience on the issue – and then, working through the puzzles this record presents, to go on to 124

“save” as true “the greatest number and the most basic” of those “appearances”. This procedure can be seen at work in many of his inquiries, both in natural science and in ethics. On the other hand, in the Posterior Analytics Aristotle presents an account of the structure of scientific understanding, and the goal of inquiry, that seem, at first, distinctly different. He argues that an inquirer can claim episteme, or scientific understanding, only when he has been able to arrange the results of inquiry into a deductive explanatory system, internally consistent and hierarchically ordered, depending on first principles that are true, necessary, basic and explanatory of the other truths of the science in question. By itself this need not conflict with Aristotle’s emphasis elsewhere on sorting out the record of experience: for he is simply adding the point that this sorting-out must be one that yields a systematic grasp and the ability to give explanations. But in Posterior Analytics ii 19, Aristotle makes some remarks about the nature of his first principles that seem to go in a different direction: for he holds that, after experience provides us with the material of a science, its first principles must be grasped by a faculty which he calls nous. In traditional mediaeval interpretations of Aristotle, this has been understood to be a faculty of intellectual intuition that seizes on first principles a priori, and thus sets the science on an extra-experiential foundation. Recent interpretations of the passage, however, have pointed out that this is not a plausible way of understanding what is meant by nous in Aristotle (or, indeed, in the ordinary vocabulary of cognition from which he derives the term). Nous is insight based upon experience; and what Aristotle is saying is that true understanding is not achieved until, in addition to the grasp and use of principles, we gain understanding of the fundamental explanatory role. This is exactly what the person who follows Aristotle’s arguments about Non-Contradiction does derive: so there is no need to see the Posterior Analytics as in tension with that passage or others 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 of experience. n a t u r e and exp lanatio n Aristotle’s account of explanation, in the second book of his Physics, is closely linked to his arguments about substance. He identifies four different types of explanation that are standardly given when we ask the question “Why?” about some entity or event in our experience. (These are often called the “four causes”, but it would be better to think of them as the “four becauses”.) First, we often enumerate the material constituents of a thing; but this, Aristotle argues, explains nothing about a thing unless we have already said what sort of thing it is. The second sort of explanation, which cites the thing’s form or structure, is in that sense prior to the first. The third sort, which Aristotle calls “the origin of change”, and which is often called “efficient cause”, corresponds rather closely to our notion of causal explanation: asked why something happened, or why a thing is as it is, we cite some other event or agency that acted in such a way as to produce it. Finally, Aristotle introduces the explanation “that for the sake of which”, often called “teleological explanation” (see teleology). Here we say that the reason x happened was for the sake of y, where y is in the future. It is not difficult to understand the relevance of this sort of explanation in the context of intentional human action (“He did this in order to get that”). What is harder to understand is the role Aristotle gives it in explaining the growth and development of living creatures of all sorts, including many (such as plants) that are not, in his view, capable of intentional action. He recommends that we should give accounts of the development of a seed, for example, or of various life processes in a mature plant, by saying that they happen “for the sake of” the form or structure of the plant. Aristotle is at pains to insist that he is not invoking any causal factors external to the nature of the organism in each case. It seems wrong to see any implications of a grand teleology of nature or an argument

from design such as was developed later by the Stoics. Instead, Aristotle’s interest is in the plastic and self-maintaining, self-nourishing character of living systems: in a variety of circumstances, they will behave in the way best suited to realize and then maintain their forms and structures. And understanding this will enable us to grasp their doings in a unified way – predicting, for example, that a plant’s roots will grow in the direction of the water supply, wherever that happens to be. Teleological explanations do not invoke mysterious notions; they grow from a biologist’s observation that organic systems function in integrated and form-preserving ways. Aristotle’s passionate interest in biology animates much of his metaphysical writing. He spent about twenty years of his career doing first-hand biological research, much of it very fine. And his biological writings provide rich insight into metaphysical issues such as the relation of form and matter and the nature of functional explanation. To students who evidently preferred theology to the study of worms and shellfish, he makes a reply that might perhaps serve as an excellent introduction to Aristotle’s temperament as metaphysician and philosopher of nature: We must not enter upon the study of the lesser animals with childish disgust. For in every natural thing there is something wonderful. There is a story which tells how some foreigners once wanted to meet Heraclitus. When they entered, they saw him warming himself in front of the stove. They hesitated; but he told them, “Come in; don’t be afraid; there are gods here too.” wri t i ngs Categories, On Interpretation, Physics, De anima (On the Soul), Parts of Animals, Generation of Animals, Metaphysics. Translations: the best general collection is The Collected Works of Aristotle, ed. J. Barnes, 2 vols. (Princeton, NJ: Princeton University Press, 1984). See also the commentaries and translations in the Clarendon Aristotle Series, esp. those of Categories and On Interpretation by J.L. Ackrill, of 125

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

a r m s t r o ng, d avid malet Metaphysics, iv–vi, by C. Kirwan (Oxford: Clarendon Press, 1963, 1971), of Parts of Animals I and Generation of Animals I by David Balme (1972). A useful collection of good translations can be found in A New Aristotle Reader, ed. J.L. Ackrill (Oxford: Clarendon Press, 1987). Editions and commentaries: W.D. Ross, Aristotle’s Metaphysics (Oxford: Clarendon Press, 1924); M. Frede and G. Patzig, Aristoteles: Metaphysik Z, 2 vols. (Munich: C.H. Beck, 1988); G. Fine, On Ideas (Oxford: Clarendon Press, 1993) (on the fragments of Aristotle’s lost Peri Ideon, a critique of Plato’s theory of forms). b i b l i og rap hy Ackrill, J.L.: Aristotle the Philosopher (Oxford: Oxford University Press, 1981). Barnes, J.: Aristotle (Oxford: Oxford University Press, 1982). Burnyeat, M.: “Aristotle on Understanding Knowledge,” in Aristotle on Science: The Posterior Analytics, ed. E. Berti (Padua: Antenori, 1981), 97–139. Furth, M., Substance, Form, and Psyche: An Aristotelian Metaphysics (Cambridge: Cambridge University Press, 1988). Gill, M.L.: Aristotle on Substance: The Paradox of Unity (Princeton, NJ: Princeton University Press, 1989). Hartman, E.: Substance, Body, and Soul (Princeton, NJ: Princeton University Press, 1978). Irwin, T.H.: Aristotle’s First Principles (Oxford: Clarendon Press, 1988). Kosman, A.: “Substance, Being, and Energeia,” Oxford Studies in Ancient Philosophy 2 (1984), 121–49. Lesher, J.: “The Role of Nous in Aristotle’s Posterior Analytics,” Phronesis 18 (1973), 44– 68. Nussbaum, M.: “Aristotle,” in Ancient Writers: Greece and Rome, vol. 1, ed. T.J. Luce (New York: Charles Scribner’s Sons, 1982), 377– 416. Owen, G.E.L.: Logic, Science, and Dialectic: Collected Papers in Ancient Philosophy (London and Ithaca, NY: Cornell University Press, 1986). 126

Owens, J.: The Doctrine of Being in the Aristotelian Metaphysics (Toronto: University of Toronto Press, 1978). Ross, W.D.: Aristotle, 5th edn. (London: Methuen, 1949). Witt, C.: Substance and Essence in Aristotle: An Interpretation of Metaphysics VII–IX (Ithaca, NY: Cornell University Press, 1989). martha c. nussbaum Armstrong, David Malet (1926– ) Australian philosopher, born in Melbourne and educated at the University of Sydney and Exeter College, Oxford. After Oxford, he spent a brief period teaching at Birkbeck College in the University of London, then seven years at the University of Melbourne. He held John Anderson’s chair as Challis Professor of Philosophy in Sydney from 1964 until his retirement at the end of 1991. Armstrong’s work in philosophy ranges over many of the main issues in epistemology and metaphysics, where he has helped to shape philosophy’s agenda and terms of debate. Several themes run through it all: it is always concerned to elaborate and defend a philosophy which is ontically economical, synoptic, and compatibly continuous with established results in the natural sciences. Accordingly, he has argued for a naturalism which holds all reality to be spatio-temporal, for a materialism (see physicalism, materialism) which aims to account for all mental phenomena without appeal beyond the categories of physical being, and for an empiricism which both vindicates and draws strength from the methods and successes of the natural sciences. In Perception and the Physical World (1961), he confronted then-fashionable phenomenalist tendencies (see phenomenalism) with a direct realism which had no place for sense data or other mentalistic items (see sensa). He urged the objections to sense data from their indeterminacy, their hidden features, and the identification problems they face. He began also to develop a realist account of secondary qualities (see quality, primary/secondary). A Materialist Theory of the Mind (1968) was 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 with states of the central nervous system (see the mind/body problem). The theory is as naturalistic as the behaviorism it aspired to supplant, yet much more plausible and scientifically fruitful as a philosophy of mind. Armstrong presents an analysis of mental phenomena in terms of what they are apt 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 structures, states and processes in the central nervous system. The view is refined in further essays. With hindsight, Armstrong’s philosophy of mind counts as a type–type identity theory, a precursor of contemporary functionalism. During the 1970s, Armstrong turned his attention to the problem of universals. In Universals and Scientific Realism (1978) he built a case for an immanent realism in which universals, and particulars (see universals and particulars) are equally abstractions from states of affairs. The work has three principal themes: first, all the widely accepted varieties of nominalism are deeply implausible. Second, an empiricist naturalism need not, and should not, bear the nominalist burden. Third, to establish the actual existence of any universal calls for a substantive enquiry for which the fundamental sciences alone are equipped. This scientific realism about universals was promptly put to work in developing a philosophy of the laws which apparently govern the cosmos. What Is a Law of Nature? (1983) argues that the regularity theories of law, deriving from Hume, are all fatally flawed (see law of nature). It goes on to urge that laws relating particular states of affairs rest on a relation of necessitation holding between the universals involved. Armstrong’s next major project was A Combinatorial Theory of Possibility (1989). Here he attempts to build, from a foundation in the thought of Wittgenstein’s Tractatus, an account of modality in which a spatiotemporal naturalism is upheld. Non-actual possibilities do not exist, nor are they given ersatz treatment. The attempt makes use of the idea of fictive reorderings of strictly

actual cosmic constituents. Here again, Armstrong’s doctrine about universals, as abstractions from states of affairs on an equal footing with particulars, stands him in good stead. See also logical atomism; the extended essay on modalities and possible worlds. writings A Combinatorial Theory of Possibility (Cambridge: Cambridge University Press, 1989). A Materialist Theory of the Mind (London: Routledge and Kegan Paul, 1968). The Nature of Mind and Other Essays (Brisbane: Queensland University Press, 1980). Perception and the Physical World (London: Routledge and Kegan Paul, 1961). Universals and Scientific Realism, 2 vols. Vol. 1 Nominalism and Realism; Vol. 2, A Theory of Universals (Cambridge: Cambridge University Press, 1978). What Is a Law of Nature? (Cambridge: Cambridge University Press, 1983). bibl i ography Bacon, J.B., Campbell, K., and Reinhardt, L., ed.: Ontology, Causality, and Mind; Essays in Honour of D.M. Armstrong (Cambridge: Cambridge University Press, 1993). keith campbell Arnauld, Antoine (1612–94) A French Roman Catholic theologian and philosopher. Arnauld was born in Paris into a family associated with Jansenism. Angelique Arnauld, his sister, was abbess of port-royal, which became, under her direction, a center of Jansenism. One aspect of Jansenism is adherence to whatever view of the relation of divine grace to human freedom is expressed in Augustinus, a work written by Cornelius Jansen and published posthumously in 1640. Numerous Roman Catholics, including various popes, believed that the Jansenist account of grace is incompatible with the Roman Catholic dogma that divine grace can always be resisted by a free agent. Much of Arnauld’s theological 127

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 Jansenist account of divine grace and the claim that it is consistent with Roman Catholic dogma. Another important segment of Arnauld’s theological writings concerns the role of the sacraments in the process of absolution, where Arnauld emphasized the attitude that the penitent must bring to the process if the sacrament is to absolve. In connection with a school associated with Port-Royal Arnauld wrote or co-wrote three important textbooks that influenced seventeenth-century thought: Grammaire générale et raisonnée (1660), La Logique, ou l’art de penser (1662) and Nouveaux éléments de géométrie (1667). In his Jansenist phase Arnauld offered and argued in favor of an historical approach to theology on the ground that the essential theological truths could be extracted from the work of the Fathers of the Church and, in particular, at least with respect to matters of divine grace and freedom, from the work of Augustine. He, therefore, strongly opposed what he took to be the innovative, speculative philosophical theology of leibniz and malebranche. Criticism of Malebranche generated the majority of Arnauld’s positive contributions to philosophy. While Arnauld was a conservative in theology, he believed that scholastic philosophy had been exposed as inadequate by the seventeenth-century scientific revolution and Cartesian mechanics (see descartes). In philosophy, Arnauld regarded himself as a Cartesian, specifically associating himself with Descartes’s theses concerning the nature and origin of ideas, the idea of God, the distinction between the soul and the body, and the nature of matter. This may seem odd, given Arnauld’s famous criticisms of Descartes’s Meditations on First Philosophy, including a brilliant critique of Descartes’s arguments intended to prove that the soul and body are distinct substances (see the extended essay on the mind/body problem), a critique of one of Descartes’s arguments for the existence of God, a query concerning the possibility of avoiding circularity, given Descartes’s way of establishing the principle of clear and distinct perception, and a criticism of 128

Descartes’s thesis that nothing occurs in the soul of which it is not conscious. Except for this last thesis, which Arnauld regarded as inessential to Descartes’s program, his criticisms were aimed at Descartes’s arguments, not the conclusions of those arguments. Arnauld criticized some of Descartes’s doctrines because of their theological implications. The majority of Arnauld’s criticisms of Malebranche center on what he viewed as Malebranche’s speculative and innovative contributions to theology. But in the process, Arnauld formulated a theory of perception, which he presented as a mere recasting of Descartes’s theory, but which, in fact, involves many ideas original to Arnauld. Arnauld’s theory of perception is contained in two works aimed at Malebranche: Des vraies et des fausses idées (1683) and Défense de M. Arnauld, contre la réponse au livre des vraies et des fausses idées (1684). In these works, Arnauld articulated and defended a subtle form of a direct realist position, based on an act theory of ideas, in which ideas are identified with representative acts of the mind rather than objects of the mind that serve as intermediaries between an act of the mind and the external reality thereby represented. writings La Logique, ou l’art de penser (Paris, 1662); ed. and trans. J. Dickoff and P. James (Indianapolis, IN: Bobbs-Merrill, 1964). Oeuvres de Messire Antoine Arnauld, docteur de la maison et société de Sorbonne, 43 vols. (Paris, 1775–1839); repr. Brussels: Culture et Civilisation, 1967). On True and False Ideas, New Objections to Descartes’ Meditations and Descartes’ Replies (Cologne, 1683); trans. E.J. Kremer (Lewiston, NY, Queenston, ON, and Lampeter, Wales: Edwin Mellen Press, 1990). bi bliography Nadler, S.M.: Arnauld and the Cartesian Philosophy of Ideas (Princeton, NJ: Princeton University Press, 1989).

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

associat i onism Ndiaye, A.R.: La Philosophie d’Antoine Arnauld (Paris: J. Vrin, 1991). robert c. sleigh, jr.

artefact Any object produced to design by skilled action. Artefacts are continuants, that is, objects persisting in time: an event such as a pianist’s performance is itself an action and not the persisting product of one. Artefacts are not exclusively human: consider a beaver’s dam, or the cosmos viewed by creationists. But the most elaborate artefacts we know, requiring conscious planning, training and sophisticated forms of representation, are human: levels of culture are even measured by the kinds of artefacts people produce, from stone axes to moon rockets. Artefacts contrast with natural objects: Aristotle considered artefacts, defined by function rather than an autonomous principle of unity and persistence, not to be substances. Mechanistic world views tend to blur this distinction. The identity conditions (see individuation) of artefacts are, however, vaguer and more convention-bound than those of natural objects: the puzzle of the Ship of Theseus notably concerns an artefact. b i b l i og rap hy Aristotle: Physics, ed. W.D. Ross (Oxford: Clarendon Press, 1950), Bk. 2, 192. Hobbes, T.: De corpore (London, 1655); in his Opera philosophica, ed. W. Molesworth (London, 1839), vol. 1, Part II, ch. 11. Wiggins, D.: Sameness and Substance (Oxford: Blackwell, 1980), esp. ch. 3.

of association, associationism as a psychological program achieved its greatest influence in eighteenth- and nineteenthcentury Britain. Locke was the first to use the term “association of ideas”, but he used it only to describe a cause of error, in which accidental or logically irrelevant relations among ideas usurp the role of logical relations. Berkeley put association to more positive and extensive use in An Essay towards a New Theory of Vision (1709), arguing that visual perception of distance is the result of an association between certain kinds of visual ideas and certain kinds of non-resembling tactile ideas, an association resulting from their repeated conjunction in past experience. David Hume’s cognitive psychology gives a fundamental role to three “principles of association”: contiguity, resemblance and causation, the latter based on “constant conjunction”. Hume uses these relations to explain both the formation of complex ideas from simpler ideas, and the succession of ideas in thought. For Berkeley and Hume, in particular, the association of ideas provided a way of explaining mental phenomena without presupposing intellectual insight into the metaphysical structure of the world. David Hartley (1705–57), a physician and Hume’s contemporary, also sought to explain a variety of mental phenomena associationistically, proposing to explain the influence of associative relations through their relation to “vibrations” in the brain. Later associationists included Thomas Brown (1778–1820), James Mill (1773–1836), John Stuart Mill, and Alexander Bain (1818–1903). bi bliography

peter simons

associationism Associationism is the attempt to explain mental phenomena through relations among mental contents and representations – particularly relations such as contiguity or simultaneity, resemblance and constant conjunction – that cause them to become associated with one another. Although Aristole, Hobbes, and Spinoza, among others, described phenomena

Bain, A.: Mental Science (New York, 1868); (New York: Arno Press, 1973). Bain, A.: The Senses and the Intellect, 4th edn. (New York: Appleton, 1894). Berkeley, G.: The Works of George Berkeley, Bishop of Cloyne, ed. A.A. Luce and T.E. Jessop, 9 vols. (London and New York: Nelson, 1948–57). Brown, T.: Inquiry into the Relation of Cause and Effect (Edinburgh, 1806); repr. as The Doctrine of Mr. Hume: Concerning the 129

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

atomism Relation of Cause and Effect (New York: Garland, 1983). Hartley, D.: Observations on Man, His Frame, His Duty, and His Expectations (London, 1749); (New York: Garland, 1971). Hume, D.: An Enquiry Concerning Human Understanding (London, 1777); in Enquiries Concerning Human Understanding and Concerning the Principles of Morals, ed. L.A. Selby-Bigge (Oxford: Clarendon Press, 1893); 3rd edn. rev. P.H. Nidditch (Oxford: Clarendon Press, 1975). Hume, D.: A Treatise of Human Nature (London, 1839–40); ed. L.A. Selby-Bigge (Oxford: Clarendon Press, 1888); 2nd edn. rev. P.H. Nidditch (Oxford: Clarendon Press, 1978). Locke, J.: An Essay Concerning Human Understanding (London, 1690); ed. P.H. Nidditch (Oxford: Clarendon Press, 1975). Mill, J.: Analysis of the Phenomena of the Human Mind (London: Longmans, 1829). Mill, J.S.: An Examination of Sir William Hamilton’s Philosophy (London and Boston: Spencer, 1865). don garrett atomism Atomism takes the world to be made up of indivisible and imperceptibly small material units. (Atomos in Greek means indivisible.) The diverse qualities of perceptible bodies are to be explained by the simple quantitative properties of the atoms composing them. Perceptible changes are to be understood as rearrangements of the underlying atoms. In its origins, atomism was primarily a metaphysical doctrine; it was not, indeed, until the early nineteenth century that the atomic hypothesis was linked tightly enough to the explanation of specific empirical data to count as physical theory in the familiar modern sense. e a r l y a to mism The first atomist doctrines are best understood as a response to the challenge of Parmenides’ analysis of change. Parmenides argued that, despite the evidence of our senses, our reason compels us to conclude that change is illusory. Being obviously cannot just come 130

to be from Non-Being or abruptly cease to be. And where one sort of Being appears to become another sort, the difference must itself count as Being, so that there is no real change. Being is thus ultimately immutable and one. For a “physics”, that is, an account of the regularities of perceived change, to be possible, this paradoxical conclusion had to be overcome. The atomism of Leucippus and Democritus retained something of Parmenides’ sharp dichotomy being Being and Non-Being, while modifying it in two fundamental respects. Instead of one Being, there is an infinite multitude of indistinguishable beings, each (like the Parmenidean original) one and immutable. And instead of Non-Being, there is the Void in which atoms can move. The Void is almost Non-Being; indeed, Democritus calls it Nothing. But it is just sufficient to make change possible, though only one kind of change, local motion. Thus all change must (despite appearances) reduce to local motion of entities that themselves must be imperceptibly small since no local motion is actually perceived when, for example, a leaf changes color. Likewise, the manifold qualitative differences between perceptible things must reduce to differences of atomic configuration, size and shape. And since the analysis is a perfectly general one, it must extend to all things, to soul, for example, whose atomic constituents presumably are small and round so that they can direct the vital functions of the living body. Atomism in this “pure” form thus entails a strongly reductionist form of materialism. Its appeal is to the coherence of its very general account of change, though there are hints of a more specific sort of warrant also; evaporation and condensation are said to be explained by different degrees of “packing” of atoms, for example. Though atomism itself was not immediately influential, the atomic metaphor can be found everywhere in the philosophic thinking of Parmenides’ successors. One finds hints of it in Empedocles’ four elements, in Anaxagoras’ seeds, and in Plato’s elemental geometrical shapes. Aristotle proposed an alternative analysis of change in terms of 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 the reductionism and lack of teleology that he found so objectionable in the atomist proposal. Yet Aristotle also objected to Anaxagoras’ assumption that physical things can be divided without limits. There are, he said, least natural parts. The limits of divisibility depend on the kind of thing being divided. This suggestion was the occasion for a vast and ingenious elaboration among later Aristotelian commentators of the doctrine of the minima naturalia, that is, of the conceptual limits of physical divisibility. Averroes and his later followers seem to have been the first to present these “least parts” as separately existent, indeed as potentially capable by their intermixing of explaining the qualitative changes we today call chemical. Such Renaissance Aristotelians as Julius Caesar Scaliger (1484–1558) and Agostino Nifo (1473–1538) propounded a doctrine of minima which was close to atomism in significant ways, since the minima were regarded as real constituents whose manner of union explains the properties of sensible bodies. What separated these philosophers from Democritean atomism was their commitment to matter-form composition, and especially to the role of substantial form in making the product of the union of minima into a qualitatively new kind of thing. t r a n s i t io n With the seventeenth century the transformation of a philosophic doctrine into a physical theory began. Most of the natural philosophers of the century subscribed to the “corpuscular philosophy”. Though it had roots in classical atomism (here the role of Gassendi in modifying and popularizing the ancient doctrine was important) and in minima theory (the main spokesman here being Daniel Sennert (1572–1657)), the more important motivation came from the “new science” of mechanics. If mechanics were to be as all-explanatory as its exponents expected it to be, the primary properties of things had to be those which made things subject to mechanical law: size, shape, mobility, solidity and, perhaps

eventually, mass. Other properties (the “secondary” ones) would then have to be explicable in terms of the primary ones (see quality, primary/secondary). This requires explanation in terms of something like atoms. Since, however, the atoms do not have to be strictly indivisible, the term “corpuscle” was preferred. But how were these invisible corpuscles to be known? How, in practice, could their sizes, shapes, and motions explain such a property as yellowness? Locke was pessimistic about the prospects of linking the two sorts of properties in a demonstrative science, though he suggested that plausible analogies might yield at least a weak kind of probability. Meanwhile, chemists were trying to understand chemical combination in corpuscular and quantitative terms. Robert Boyle (1627–91) proposed that the corpuscles constituting the chemical elements could combine to form complex corpuscles that yielded chemical compounds. He conceded that the former might themselves be “primary concretions”, composites made up of Democritean atoms. But in practice, these primary concretions could be regarded as basic from the point of view of the chemist because they remained unaltered through chemical change. The problem was how to decide which concretions were primary, how to distinguish element from compound. Boyle could not discover a consistent way to carry this all-important distinction through. By the end of the century the separation between philosophers and scientists (as the latter would come to be called) was widening. Scientists were convinced of the underlying corpuscular character of the world, but they had no real evidence (as evidence in natural science was coming to be understood) in support of their hypothesis. There was as yet no satisfactory atomic theory. atomi c theory Atomic theory took shape only very gradually, and in two different parts of natural science, in chemistry first and later in the physics of gases. The Newtonian project of organizing chemical research around 131

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 corpuscles proved fruitless (see Newton). Careful weighing of the products of chemical combination ultimately, in the hands of Antoine Lavoisier (1743–94), yielded the first victory. Aided by the assumption that weight is conserved through chemical change, Lavoisier provided for the first time a reliable way of distinguishing element from compound, enabling him to identify many of the commonest elements. Joseph Louis Proust (1754–1826) proposed that each compound is made up of elements combined in a constant way. But it was John Dalton (1766–1844) in A New System of Chemical Philosophy (1808) who drew from the ancient notion of atom the crucial clue. He proposed that the simplest underlying structure that would explain the empirically established laws of definite proportions (a compound contains fixed proportions by weight of its constituents) and of equivalent proportions (the ratio of the weights of a and b that react with a given amount of c is independent of c), was an atomic one. Each atom of an element is like any other atom of that element; each element is constituted by a different kind of atom. Compounds are formed by a simple and uniform juxtaposition of elemental atoms in compound particles (molecules). The key to chemical analysis is thus the determination of relative atomic weights. This turned out to be a more difficult matter than Dalton had anticipated, and the contributions of many other researchers (notable among them Joseph Louis GayLussac (1788–1850), Amedeo Avogadro (1776–1856) and Stanislao Cannizzaro (1826–1910)) were needed before the atomic model of chemical change was established to the satisfaction of chemists generally. The kinetic theory of gases followed in physics; many of the physical properties of gases were shown to be derivable from the hypothesis that gases are made up of vast numbers of molecules in rapid motion. Despite this convergence of chemistry and physics, empiricists like Mach argued that the notable successes of the atomic hypothesis did not warrant belief in the actual existence of atoms and molecules. Atomic theory was 132

acceptable as a calculational device but no more. The debate was once more philosophical, though numerous scientific issues were also involved. Only after Einstein made use of the molecular hypothesis in 1905– 6 to derive in a strikingly detailed way the main parameters of Brownian motion did the critics concede. Not that scientific realism would from henceforward be immune to challenge! From Democritus to Einstein is a long road, and the atom of modern quantum theory bears little resemblance to the immutable qualityless particle of the first atomists. But the claim that the world around us consists of a swarm of imperceptible entities whose properties can causally explain the properties of that larger world evokes echoes all along that road. The transition from metaphysical doctrine to physical theory has no clearer illustrative example. bibl i ography Furley, D.J.: Two Studies in the Greek Atomists (Princeton, NJ: Princeton University Press, 1967). Kirk, G.S., Raven, J.E., and Schofield, M.: The Presocratic Philosophers, 2nd edn. (Cambridge: Cambridge University Press, 1983), 402–33. Lasswitz, K.: Geschichte der Atomistik vom Mittelalter bis. Newton, 2 vols. (Hamburg: Voss, 1890). Nash, L.: The Atomic-Molecular Theory (Cambridge, MA: Harvard University Press, 1950). Nye, M.J.: Molecular Reality (London: Macdonald, 1972). van Melsen, A.G.: From Atomos to Atom (Pittsburgh, PA: Duquesne University Press, 1952). ernan mcmullin Augustine of Hippo, St. (354– 430) Theologian, born in North Africa. Augustine drew his metaphysics from “the Platonic philosophers, who said that the true God is at once the author of things, the illuminator of truth, and the giver of 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 editor Porphyry (ad c.232–c.303). These pagan Platonists – “Neoplatonists” (see neoplatonism) to us – were the chief instrument of his rescue from Manichean dualism and from Ciceronian skepticism at the time when, as a 31-year-old teacher in Milan, he resumed the Christianity of his childhood and planned the little African philosophical community whose life was to be cut short by his ordination (ad 391) four years later. His philosophical education was Latin, and narrow, enriched during his career as a Christian controversialist only by the Bible. According to Augustine there are three “natures”, i.e., kinds of substance: corporeal, which are mutable in time and place; spiritual, mutable in time only; and God, immutable (De Genesi ad litteram 8.20.39). Souls are not corporeal since they see and judge “similitudes” which are not corporeal; therefore God is not corporeal either (City of God 8.5). Among non-corporeal beings are angels and demons, but at most one God since only what is supreme is divine (De vera religione 25.46). Everything is from God, since all good things are from God and everything is good (De natura boni 3); miracles differ from natural events only in not proceeding “by an ordinary route” (De Trinitate 3.6.11). The “perfectly ordinary course of nature” is the regular (and planned) unfolding of causal or seminal reasons (De Genesi ad litteram 9.17.32), which date from the creation when God “completed” his work (ibid. 6.11.18–19). These reasons do not all necessitate (ibid. 6.15.26). At some places Augustine’s conception of God seems to combine the two roles, cause of truth and cause of knowledge, assigned by Plato to the form of the good: the latter role makes God the only teacher (De magistro), illuminator of truths as the sun illuminates visible things (De libero arbitrio 2.13.36); the former makes him “truth itself” (ibid.). Following Varro (116–27 bc), Augustine proposed that “the question what a man is is the question whether he is both [a body and a soul], or only a body, or only a soul” (De moribus ecclesiae catholicae 4.6). He chose the first answer, but felt forced to conclude that “the way in which spirits adhere to

bodies and become animals is altogether mysterious” (City of God 21.10.1). His celebrated investigation of time in Confessions 11 and City of God 11–12 meets the pagan challenge against creationism. “Why then?”, with a response developed from Philo Judaeus (c.20 bc–ad c.50) that God made time too; follows Plotinus and anticipates Boethius in a perplexing account of eternity; and wrestles with Aristotle’s puzzle how times can exist, since they are all past, future or durationless (Augustine’s speculative solution, arising from his insight that we measure times by memorizing their length, is that they are affections of the mind). His various writings on free will (see the extended essay) provided materials for both parties in the Reformation debates, for example between Erasmus and Luther, which set the scene for modern treatments of the subject. He failed to find a consistent response to the contrary pressures on him, arguing (e.g., in De correptione et gratia against the Pelagians) that God’s prevenient grace cannot be resisted, but refusing to repudiate his earlier argument (e.g., in De libero arbitrio against the Manichees) that some evils are, and others punish, sins freely committed. writings Augustine’s works are in Patrologiae cursus completus, series latina, ed. J.P. Migne, vols. 32–47 (Paris, 1844–55) (PL); many are also in Corpus scriptorum ecclesiasticorum latinorum (Vienna: Tempsky, 1866– ) (CSEL), and in Corpus christianorum, series latina (Turnhout, Belgium: Brepols, 1953– ) (CCL). Various of his works are translated into English in: A Select Library of the Nicenc and Post-Nicene Fathers of the Christian Church, ed. P. Schaff (New York: The Christian Literature Co., first series 1886–8; repr. Grand Rapids, MI: Wm B. Eerdmans, 1971–80) (NPNF); Library of Christian Classics ed. J. Baillie, J.T. McNeill, and H.P. van Dusen (Philadelphia: Westminster Press, 1953– ) (LCC): Fathers of the Church, ed. R.J. Deferrari et al. (Washington, DC: Catholic University of America Press, 1947– ) (FC); Ancient Christian Writers, ed. J. Quasten and 133

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

a v e r r oes J.C. Plumpe (Westminster, MD: Newman Press, 1946– ) (ACW); Basic Writings of Saint Augustine (New York: Random House, 1948) (BW). A useful compendium of excerpts in translation is: The Essential Augustine, ed. V.J. Bourke, 2nd edn. (Indianapolis, IN: Hackett, 1974). The list below is of works cited; numbers denote volumes. City of God (De civitate Dei contra paganos, ad 413–26): PL 41, CSEL 40, CCL 47– 8, and elsewhere; trans. NPNF 2, FC 8, 14, 24; text and translation also in Loeb Classical Library (London: Heinemann and Cambridge, MA: Harvard University Press, 1966–72). Confessions (Confessiones, ad 397–401): PL 32, CSEL 33, and elsewhere; trans. NPNF 1, LCC 7, FC 21, BW, and elsewhere; text and (old) translation also in Loeb Classical Library (London: Heinemann; Cambridge, MA: Harvard University Press, 1912). De correptione et gratia (ad 426): PL 44; trans. NPNF 5, FC 2. De genesi ad litteram (ad 401–14); PL 34, CSEL 28.1; trans. ACW 41–2. De libero arbitrio (ad 388, 391–5): PL 32, CSEL 74, CCL 29; trans. LCC 6, ACW 22, FC 59, and elsewhere. De magistro (ad 389): PL 32, CSEL 77, CCL 29; trans. LCC 6, ACW 9, FC 59, BW 1, and elsewhere. De moribus ecclesiae catholicae (ad 387–9): PL 32; trans. NPNF 4, FC 56, BW 1. De natura boni (ad 399): PL 42, CSEL 25.2; trans. NPNF 4, LCC 6, BW 1. De trinitate (ad 399–419): PL 42, CCL 50, 50A; trans. NPNF 3, FC 45. De vera religione (ad 391): PL 34, CSEL 77, CCL 32; trans. LCC 6, and elsewhere. b i b l i og rap hy Kirwan, C.A.: Augustine (London and New York: Routledge, 1989). Kirwan, C.A.: “Augustine on Souls and Bodies,” in Logica, mente e persona, ed. A. Alberti (Florence: Olschki, 1990), 207– 41. Markus, R.A.: Later Greek and Early Medieval Philosophy, ed. A.H. Armstrong 134

(Cambridge: Cambridge University Press, 1967), chs. 21–7. Sorabji, R.R.K.: Time, Creation and the Continuum (London: Duckworth, 1983). christopher kirwan

Averroes, [Ibn Rushd] (1126–98) SpanishIslamic philosopher who lived in Cordoba and Seville, a thoroughgoing Aristotelian, wrote commentaries on most of Aristotle’s works, but is better known in Islam as the defender of philosophy against the attacks by al-Ghazali (1058–1111), in The Incoherence of the Philosophers and as a reconciler of philosophy and religion. The Aristotelian commentaries were based on excellent translations that gave reliable access to Aristotle without Neoplatonic eyes (see alfarabi; neoplatonism), and thus played an important role in the Latin and Jewish Aristotelian tradition. In his Incoherence of the Incoherence Averroes takes up Ghazali’s attacks on Alfarabi and Avicenna. To safeguard God’s omnipotence Ghazali had rejected their claim of a necessary connection between cause and effect. According to Ghazali, such necessity is not given in observation. All we see is a temporal sequence between, say, fire and cotton burning. God, the only agent, causes the occurrence of fire, the burning of cotton and the coincidence which it becomes our habit to expect. Against this Averroes argued that to deny cause is to deny knowledge. It is also to deny human agency and the distinction between the voluntary and the involuntary. Further, it violates the view that things have a real nature. Finally, if there is no regularity nor design in creation, we cannot infer a wise Agent. Resting on Aristotelian grounds. Averroes criticized Avicenna for confusing the logical and metaphysical features of being by making the definitional separation of essence and existence characteristic also of existing things, thus espousing an un-Aristotelian essentialism (see essence and essentialism). A similar confusion is said to occur with 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. 93–111 for a fairer view of Avicenna.) On the doctrine of creation Averroes argues that creation ex nihilo of both world and time does not have Qur’anic support. On the contrary, some verses (11:6, 41:10) suggest that matter and time pre-existed with God, making Aristotle’s God consistent with Scripture. writings Tahafut al-Tahafut, ed. M. Bouyges (Beyrouth: Imprimerie Catholique, 1930); trans. S. van den Bergh The Incoherence of the Incoherence, 2 vols. (London: Luzac, 1954). b i b l i og rap hy Fakhry, M.: A History of Islamic Philosophy (New York: Columbia University Press, 1970). Fakhry, M.: Islamic Occasionalism and Its Critique by Averroes and Aquinas (London: Allen and Unwin, 1958). Kogan, B.S.: Averroes and the Metaphysics of Causation (Albany, NY: SUNY Press, 1985). Mehren, I.: “Etudes sur la philosophie d’Averroes concernant ses rapports avec celle d’Avicenna et de Ghazzali,” Muséon VII (1888–9). Shehadi, F.: Metaphysics in Islamic Philosophy (Delmar, NY: Caravan Books, 1982).

On the relation between essence and existence in Avicenna one must distinguish three contexts in which these could be related (see essence/accident; essence and essentialism). First, the logical, where in any definition, say, of a horse, existence must be excluded from the essence of a horse. Excluded also is any property that is not 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. Second, the metaphysical context: essence and existence are inseparable in individual things. While “existence” and “one” are distinct from the meaning of “horse”, they are metaphysically part of what makes a horse this horse, and are not accidental to it qua substance. Third is the theological context. Following Alfarabi, Avicenna divides beings into the possible in itself, though necessary through another, and the necessary in itself. The existence of the former is contingent and its non-existence possible, while the nonexistence of the Necessary Being is impossible. God gives existence to all contingent beings. And while existence is a necessary feature of a thing qua substance, it is accidental to it qua contingent. Avicenna reproduces the emanationist scheme of Alfarabi. The soul being an emanation of the Active Intellect turns to this intermediary between humans and God for knowledge and mystical illumination.

fadlou shehadi bibl i ography Avicenna [Ibn Sina] (980–1037) Islamic philosopher. Avicenna was the most systematic and sophisticated, as well as the most influential of Islamic philosophers, although much of his thought is already in Alfarabi. Being is a primary intuition of the soul. It can be known without the mediation of any other concept, and it cannot be defined without circularity. Even “thing”, its coequal in extension, presupposes being and cannot be used in explaining it without circularity. Being is the most general concept; its opposite is the absolute nothing.

Fakhry, M.: A History of Islamic Philosophy (New York: Columbia University Press, 1970). Goichon, A.M.: La Distinction de l’essence et de l’existence d’après Ibn Sina (Paris: Desclée, 1937). Goodman, L.: Avicenna (London: Routledge, 1992). Gutas, D.: Avicenna and the Aristotelian Tradition: Introduction to Reading Avicenna’s Philosophical Works (Leiden: E.J. Brill, 1988). Shehadi, F.: Metaphysics in Islamic Philosophy (Delmar, NY: Caravan Books, 1982). fadlou shehadi 135

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

avowals avowals The verb “to avow” has been adopted by many philosophers of mind as the translation of the German verb äussern. The usual alternative translations are “to express” or “to utter”. In Wittgenstein’s later work avowals are the keystone of a new philosophy of mind, founded on the rejection of the Cartesian idea that a person discloses the contents of his mind by identifying inner objects and describing them (see descartes). According to Wittgenstein, an avowal of an intention is not based on a self-examination which parallels the investigation of the world around us: it is only marginally liable to error, and in certain cases is an artificial expression of the intention replacing a natural one (e.g., a raised fist). Each of these three points makes its contribution to the new philosophy of mind, which some of Wittgenstein’s followers have accepted in its entirety and which, perhaps, nobody can totally reject. But the third point may be the most important one, because it shows how language can develop directly out of behavior which antedates it. This makes it possible to explain how we can learn, and communicate with, mentalistic language, which were things that remained mysterious when intentions, feelings, and so on, were treated as private objects. So it prepares the way for a naturalistic, rather than an intellectualist answer to skepticism about other minds. b i b l i o g r ap hy Malcolm, N.: Nothing Is Hidden (Oxford: Blackwell, 1986), esp. ch. 8. Wittgenstein, L.: Philosophical Investigations, trans. G.E.M. Anscombe, 3rd edn. (London: Macmillan, 1969). david pears Ayer, Alfred Jules (1910–89) British philosopher. Ayer was famous for the attack on metaphysics in his Language, Truth and Logic (1936). According to the verification criterion of meaning (see logical positivism; principle of verifiability), only analytic or synthetic statements were 136

meaningful, and synthetic statements were understood to be ultimately verifiable in sense experience. One intention of the verification criterion was to rule out as meaningless the wordy, but empirically uncheckable claims of metaphysicians in the Hegelian tradition. But while the criterion did allow those who held it to dismiss much of Hegel’s Science of Logic (1812–16), say, without the trouble of reading it, it had the not so welcome effect of rendering meaningless such unverifiable statements as “Every event has a cause” or even “For every action, there is an equal and opposite reaction.” Even the proposal Ayer made to treat these statements as heuristic aids to living and to scientific enquiry implicitly admitted their meaningfulness. For reasons outlined in later editions of Language, Truth and Logic the verification criterion was dropped by Ayer, and metaphysics, at least in a certain sense, re-admitted to the canon of meaningful discourse. Ayer remained skeptical to the end of his life concerning the pretensions of some metaphysicians to inform us of any suprasensible reality, or to delineate the most general characteristics of being as such. Nevertheless, in another sense, in much of his philosophy subsequent to Language, Truth and Logic he was engaged in metaphysical enquiry. Although the motivation of his philosophy was largely epistemological, concerning the status of our claims to knowledge, many of its conclusions were metaphysical, concerning what there actually is. Indeed, throughout the whole of his philosophical career, Ayer was concerned about the nature of physical objects in particular. There is, in fact, an interesting transition in Ayer’s work from the phenomenalistic stance (see phenomenalism) of Language, Truth and Logic to the sophisticated realism of The Central Questions of Philosophy (1976). Ayer always rejected what he called naive realism. That is to say, he denied that objects are just as they appear. He was further convinced that there was an inference involved in any transition from appearance to object, on the grounds that there is always 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. What, then, is the relation between the objects and the perceptions? Ayer came to reject phenomenalism on the grounds that the percepts that are presented even to the totality of observers are too scanty to answer to our conception of the physical world. He also rejected the causal theory of perception, largely because that theory would render the causes of our perceptions unobservable occupants of an unobservable space. Instead he proposed what he called a construction, in which the subject of experience is initially presented with a mass of sensory data; he then begins to perceive patterns within this data, which tend to cluster in predictable ways. At a certain stage in the process, the clusters or “visuo-tactual continuants” as Ayer calls them are “cut loose from their moorings” and regarded as having an existence quite independent of their being perceived. Our common-sense view of the world is thus seen as a theory relative to our perceptions; but it is a theory, which once accepted, ontologically downgrades the perceptions on which it was originally based. It cannot be said that everything about this construction is clear. Ayer denies that he is telling a psychological story about how children actually learn about the physical world, but he insists that “an exercise of the imagination” is required in the passage

from percepts to objects. He also insists that under the dominion of the theory our imagination has led us to, the existence of physical objects becomes a matter of objective fact, and he denies the possibility of any straightforward phenomenalist reduction. At the same time, the suspicion remains that there is a sense in Ayer’s story in which sense qualia (see sensa), rather than objects, are the basic stuff of the world. On this point Ayer himself would probably have said – as he did on related issues – that the matter is ultimately undecidable. It is just a matter of decision, based on experiential coherence of any story we tell. If this was indeed his attitude, it would certainly be in a direct line of descent from his earlier repudiation of metaphysics as meaningless. wri t i ngs The Central Questions of Philosophy (Harmondsworth: Pelican, 1976). Language, Truth and Logic (1936); 2nd edn. (London: Victor Gollancz, 1946). The Problem of Knowledge (Harmondsworth: Pelican, 1956). bibl i ography Foster, J.: Ayer (London: Routledge and Kegan Paul, 1985). anthony o’hear

137

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

Baker, Lynne Rudder (1944– ) defends a position called Practical Realism, in which she intends to do justice to the common sense conception of reality. In her view, reality as disclosed by everyday life and natural language is not second-class, to be replaced by science in the long run. The world of intentional agents, social institutions, medium-sized natural objects, and Artefacts is ontologically irreducible. In the mind/body debate Baker is a leading critic of reductionist (see the extended essay on the mind/body problem; reductionism) and eliminativist approaches to mental states. She rejects the idea that mental states are identical with, constituted by, supervene on (see supervenience), or grounded in brain states. They are conceived as global states of whole persons. The criterion for having a belief is the truth of relevant Counterfactuals. “Whether a person S has a particular belief is determined by what S does, says, and thinks, and what S would do, say, and think in various circumstances, where “what S would do” may itself be specified intentionally” (Baker, 1995, p. 154). Having a belief is a relational property, where Relations are seen as both real and causally efficacious. For Baker a person (see persons and personal identity) is most fundamentally a being with a first-person perspective. She conceives the relation between a person and her body in terms of the general metaphysical relation of constitution. Persons are constituted by, but not identical to, or separate from their bodies. Constitution is a relation between individual things or aggregates, not between mereological parts (see part/whole) or between properties. It is also a contingent relation: persons, artefacts or natural objects go out of existence 138

if they lose their essential properties (including relational ones). wri t i ngs Explaining Attitudes. A Practical Approach to the Mind (Cambridge: Cambridge University Press, 1995). The Metaphysics of Everyday Life: An Essay in Practical Realism (Cambridge: Cambridge University Press, 2007). Persons and Bodies. A Constitution View (Cambridge: Cambridge University Press, 2000). Saving Belief. A Critique of Physicalism (Princeton, NJ: Princeton University Press, 1987). bibl ography Meijers, A.W.M. (ed.): Explaining Beliefs: Lynne Rudder Baker and Her Critics (Stanford, CA: CSLI Publications, 2001). anthonie meijers bare particular Bare particulars are the individuators of concrete objects. The basis for this contention can be articulated only if the problem of individuation is placed in the broader context of the issues raised by the relationship between a concrete object and its properties. For it is an antecedent commitment to property realism (see universals) and anti-essentialism (see essence/accident; essence and essentialism) that provide key premises in the argument for bare particulars. An account of the relationship must explain two features: (1) some objects have properties in common with other objects; yet (2) no object is identical with any other object. According to realism, the properties

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 139

bare particul ar of objects are universals. Hence, if two objects have a property in common, say redness, the redness of one object is identical with the redness of the other. The realist has three basic options available for explaining (2): (a) non-identical objects differ in the universals they instantiate; (b) non-identical objects differ in some feature other than the universals they instantiate; and (c) the nonidentity of objects is primitive. Option (a) is associated with the view, often called the bundle theory, that concrete objects are complex entities whose sole constituents are universals. Proponents of bare particulars, such as Bergmann, reject this view on the grounds that (i) it is committed to the necessary truth of the principle of the identity of indiscernibles; but (ii) the principle is not a necessary truth. Option (c) is rejected on the grounds that the nonidentity of objects is insufficiently fundamental to be taken as primitive. Theories exercising option (b) fall into two broad categories depending on their explanation of identity through time. Concrete objects typically change their properties over time while remaining the same object. One explanation is that there is a constituent of every object which endures through time and remains unchanged despite the changes in the object itself. Furthermore, this enduring constituent, often called a substance, has some of its properties essentially. Since proponents of bare particulars are anti-essentialists, they reject this explanation. Instead, they maintain that a concrete object is a temporal series of momentary objects which stand in some complex contingent lawlike relations (see temporal parts, stages). The endurance of an object through time is explained in terms of the obtaining of these contingent relations among the momentary objects. Change is explained by differences in the properties of successive momentary objects. Bare particulars are the individuators of momentary concrete objects. Such particulars differ from substances in two significant ways: (1) they are momentary entities rather than continuants; and (2) they have no essential properties. The particularity of bare particulars consists in the fact that it

is impossible for the same bare particular to be a constituent of two different momentary concrete objects. Since difference in bare particulars is sufficient to insure difference in any two momentary objects, even two with all properties in common, the theory is not committed to the necessary truth of the identity of indiscernibles. There are two important features of the relationship between a bare particular and the universals it instantiates. Universals exist only if instantiated by some bare particular and bare particulars exist only if they instantiate some universal. Neither is capable of independent existence. Furthermore, the instantiation of a universal by a bare particular requires a nexus. A nexus, unlike a relation, can unite two distinct entities into a complex without some further relation. Hence, a momentary object is a complex entity, often called a fact, whose constituents must include a bare particular, a nexus, and a universal. There are two familiar objections to bare particulars. The first alleges that theories invoking them are incompatible with an empiricist epistemology (see empiricism). This objection rests on the claim, articulated in Allaire (1963), that, according to empiricism, the basic entities of an ontological theory must be entities with which we are directly acquainted. This claim, however, ties empiricism to phenomenalism in a manner few contemporary empiricists would accept. The second alleges that the theory is incoherent since its central thesis, “Bare particulars instantiate properties”, is equivalent to “Entities which have no properties have properties” which is selfcontradictory. But, as Loux (1978) points out, a bare particular is not an entity which has no properties but one none of whose properties is essential. Loux, however, maintains that the latter thesis is itself problematic since bare particulars have essentially the property of having no properties essentially. This contention rests on the assumption that the predicate “has no properties essentially” designates a property, and this is denied by Bergmann (1967) and Armstrong (1978). Critics also allege that bare particulars are unnecessary and have little explanatory value. Proponents of the theory maintain 139

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 of two concrete objects must “ground” it in some difference in the constituents of the objects. Yet, they also maintain that the non-identity of two bare particulars is a primitive fact. It is not evident, as Hochberg (1965) contends, that explaining the nonidentity of objects in terms of constituents whose non-identity is primitive is more illuminating than maintaining that the non-identity of the objects themselves is primitive. Proponents of the bundle theory, such as Russell (1948) and Casullo (1988). argue that it is a contingent truth that concrete particulars are complexes of universals and that this view does not comman them to the necessary truth of the identity of indiscernibles. They also maintain that the purported counterexamples to the necessary truth of the principle are questionable. b i b l i og rap hy Allaire, E.B.: “Bare Particulars,” Philosophical Studies 14 (1963), 1–8. Armstrong, D.M.: Universals and Scientific Realism, 2 vols., Vol. 1, Nominalism and Realism (Cambridge: Cambridge University Press, 1978). Bergmann, G.: Realism: A Critique of Brentano and Meinong (Madison, WI: University of Wisconsin Press, 1967). Casullo, A.: “A Fourth Version of the Bundle Theory,” Philosophical Studies 54 (1988), 125–39. Hochberg, H.: “Universals, Particulars, and Predication,” Review of Metaphysics 19 (1965), 87–102. Loux, M.: Substance and Attribute (Dordrecht: Reidel Publishing Company, 1978). Russell, B.: Human Knowledge: Its Scope and Limits (New York: Simon and Schuster, 1948). albert casullo basic action Basic actions, broadly characterized, differ from non-basic actions in not being performed by way of the agent’s performing another action.

140

The term was introduced in Danto (1963), where the following analysis is offered: “B is a basic action of a if and only if (i) B is an action and (ii) whenever a performs B, there is no other action A performed by a such that B is caused by A.” This analysis fails for a variety of reasons (see Goldman, 1970; Hornsby, 1980). The fundamental problem (or a symptom thereof) is that the difference between basic and non-basic actions does not hinge on causal transactions of the kind specified in (ii). Typically, when an agent does one thing by doing another, the latter is more basic than the former. If by moving her right index finger upward Jane flips a switch, and by flipping the switch illuminates the room, Jane’s moving her finger upward is more basic than her flipping the switch, and both are more basic than her illuminating the room. However, Jane’s moving her finger does not cause her flipping the switch. (It does cause the switch’s moving upward, but the latter event must be distinguished from Jane’s flipping the switch.) Nor does her flipping the switch cause her illuminating the room. Indeed, Jane’s flipping the switch and her illuminating the room might not be caused by any action of hers. Still, they are not basic actions. Just how Jane’s actions are related is controversial. Some philosophers say that they are the same action under different descriptions; others that they are distinct actions related by “causal generation”, as opposed to causation; yet others that the more basic actions are components of the less basic (see action theory). In the same vein, an action caused by another action of the agent’s might nevertheless be a basic action. Suppose that Jane’s turning on her computer caused a power surge that cut off the electricity in her study, with the effect that, moments later, she illuminated her utility room so that she could see the fuse box. Presumably, Jane illuminated the room by performing some basic action or other; and this basic action has Jane’s turning on her computer as a causal antecedent. Influential analyses of basic “act-types” 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). Goldman’s proposals are framed in terms of his own theory of act-individuation. Neutral approximations are: An act-type, b, is a basic act-type for an agent, S, at a time if, and only if, (a) given normal conditions, if S wanted on balance to do a b, S would do so; and (b) the truth of (a) does not depend upon S’s “cause-and-effect knowledge” nor upon any knowledge of S’s of the form “x-ing may be done by y-ing”. An action, b, done by S is a basic act-token if, and only if, (a) b instantiates an act-type that is basic for S; (b) S’s b-ing “is caused, in the characteristic way, by an action-plan of S”; (c) S does not b by doing anything, a, that satisfies clauses (a) and (b). Some philosophers have denied that there are basic actions. Suppose that Jane’s illuminating the room – which is not a basic action – is the same action (under another description) as her moving her finger. Given this identity, one might argue, Jane’s moving her finger is not a basic action either. However, assuming the theory of action presupposed by this argument, the notion of basic action may be relativized to actiondescriptions. Jane’s a-ing, under the description “illuminating the room”, is not a basic action; but her a-ing might be basic under another description – perhaps “moving her finger”. An important distinction between causally and teleologically basic action, framed in terms of action-descriptions, is developed in Hornsby (1980). Hornsby contends, concerning any pair of descriptions, d and d′, of an action a, that d is causally more basic, “if the effect that is introduced by [a under description d] causes the effect that is introduced by [a under description d′]” (1980, p. 71). By comparison, d is teleologically more basic than d′ if, and only if, “in virtue of” a’s occurrence, a statement to the effect that S intentionally a-ed-under-d′ by a-ingunder-d is true (ibid., p. 78). The aptness of causally basic descriptions blocks a vicious causal regress, while the appropriateness of

teleological counterparts prevents an epistemic regress. If, to a under any description at all, we always had to a under a causally more basic description, we would never act; similarly, if intentionally a-ing-under-d, for any d, required the possession of meansend knowledge identifying a under a teleologically more basic description, we would be lost in thought. bi bliography Danto, A.C.: “Basic Actions,” American Philosophical Quarterly 2 (1965), 141– 8. Danto, A.C.: “What We Can Do,” Journal of Philosophy 60 (1963), 435– 45. Goldman, A.I.: A Theory of Human Action (Englewood Cliffs, NJ: Prentice-Hall, 1970). Hornsby, J.: Actions (London: Routledge and Kegan Paul, 1980). alfred r. mele being and becoming The idea of being functions primarily in three contrastcontexts: (1) being/non-being, with the contrast of the non-existent or unreal; (2) being/ seeming, with the contrast of that which is merely suppositional, imaginary or visionary; and (3) being/becoming with a view to the origination of that which is not or not heretofore. Becoming in this third context (Greek: einai/genesis) is a matter of a shift from non-being to being, which can be either absolute via a transition from non-being to being (an origination) or its reverse (an annihilation) or relative via a change from one state or condition of being to another. With respect to being/becoming, the ancient Greeks puzzled over the aporetically inconsistent triad (see aporia): (1) absolute becoming involves a transition from nonbeing to being; (2) absolute becoming occurs: there are some things that exist (now or sometime) that did not do so at an earlier time; (3) becoming presupposes being: only something that is already in being can undergo any sort of alteration or transition. These propositions are incompatible as they stand. For, by (2), there is something, say x, that instantiates absolute becoming

141

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 at an earlier juncture and only came into being at some particular time. But then, by (3), x must have had some pre-existent state prior to that time, contrary to (1)’s stipulation of the nature of absolute becoming. The dialectic of this perplexity is encapsulated in the paradoxes of Zeno. Different theorists resolved the problem differently. Heraclitus (c.535– c.475 bc) rejected (3), maintaining that becoming is allpredominant and exhaustive: only becoming occurs and nothing is (has being) but everything is perpetually becoming. The Eleatics, by contrast, rejected (2), denying all becoming and insisting that everything just unchangingly is, with change relegated to the condition of an illusion of sorts. The atomists (see atomism) Leucippus (fifth century bc) and Democritus (c.460– c.370 bc) made yet another resolution by rejecting (1): for them absolute becoming is simply the rearrangement of pre-existing (and totally unchanging) units, the atoms. Plato struck a compromise position: with the Eleatics he saw the real as unchanging (namely, the realm of ideas), while with Heraclitus he accepted the world of sensory experience as ever changing but also as pervaded by illusion. Because genuine knowledge is confined to what really is, only mere opinion about the changeable world of sense is possible, so that any authentic knowledge of the material world is impossible. (See Presocratics.) The Eleatic idea that being excludes all becoming has an ever-renewed appeal. In modern times its main exponent has been McTaggart (“The universe is eternally the same and eternally perfect. The movement is only in our minds” (McTaggart, 1896, p. 71), and by Bradley, whose Appearance and Reality (1897) maintains the selfcontradictory and consequently unreal character of change and time. The problem of how an unchanging reality can accommodate the mental changes that occur in the domain of appearance is left obscure by these neo-Eleatics. The philosopher of becoming par excellence is Leibniz, who saw the calculus as providing a mathematics of change, thereby 142

overthrowing the static mathematics of the Greeks as reflected in Zeno’s paradoxes. Leibniz envisioned the prospect of bringing the domain of becoming into the range of a rigorous science. His monads (see monad, monadology) are centers of activity, preserving their programmatic identity of lawful development through an ever dynamic course of perpetual becoming. Harking back to the preoccupation of early Greek philosophers with being/becoming, the German philosopher Heidegger reproached the post-Platonic philosophical tradition with a neglect (forgetfulness) of being (Dasein). According to Heidegger, philosophers have been so concerned with explanation – with pursuing a theoretical account of how things have become what they are – that they neglect the immediate experience of our human presence in the world. This charge comes down to complaining that, after Plato, philosophy turned away from feeling to thinking, from the path of art (of experience and aisthesis) to the path of science (of understanding and episteme). Whether through disapproval or failed understanding, such a doctrine that theorizing is a betrayal of authenticity abandons the Leibnizian vision of an integration of sensibility and intellect. A cognate position is that of the pessimistic philosophical tradition of Spain which is captivated by the paradoxicalseeming idea that all being is caught up in a process of becoming that leads inexorably to non-being – to death or annihilation. Unable in actual fact to escape this allpervasive destruction, humanity seeks escape in thought, be it by way of art (poems, after all, can be more durable than mausoleums), or by way of science or religion. The preordained ultimate failure of these efforts at evasion makes them at once futile and noble, a paradox which lies at the core of Unamuno’s classic, The Tragic Sense of Life (1913). See also change. bibl i ography Bradley, F.H.: Appearance and Reality, 2nd edn. (New York: Macmillan, 1897).

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 143

bennett, jonathan McTaggart, J.M.E.: Studies in Hegelian Dialectic (1896); 2nd edn. (Cambridge: Cambridge University Press, 1922). Unamuno, M. de: The Tragic Sense of Life (1913); (London: Macmillan, 1921). nicholas rescher Bennett, Jonathan (1930– ) Born in New Zealand, he has been Lecturer at the University of Cambridge, and Professor at the University of British Columbia, and Syracuse University. Now retired, he lives on Bowen Island off the coast of British Columbia. Consistently adopting an austere empiricism through a career spanning more than half a century, Bennett has written on a wide spectrum of metaphysical topics, including causation (see the extended essay), conditionals and modality, consciousness and self, facts and events, identity, physical objects, space and time, and substance (see event theory; the extended essay on modality and possible worlds). Bennett’s writings are models of analytic philosophy, constantly pursuing clarity and precision in his claims and the arguments with which they are supported. Bennett’s approach tends to be dialectical: he is at his best when he in engaging an interlocutor. He has an acute sensitivity for the strengths and nuances of his opponents’ positions, as well as a commendably honest openness to the problems with his own arguments. It is not uncommon for his proposals to preserve the strengths of the views from which he is departing. In an early paper, while adopting a conventionalist account of necessity, he writes that the “traditional picture of proof which has been associated with the platonic-proposition account of meaning is an entirely correct picture;” the mistake is to believe that if a behaviorist theory is right then this picture “must be in some way softened or blurred”. Among his more important contemporary interlocutors and influences are Quine and Paul Grice, and more recently Daniel Dennett and David Lewis. Perhaps Bennett’s most distinctive philosophical reflections, however, have arisen in conversation with the canonical early modern philosophers from Descartes to

Kant. His approach to these historical figures, whom he engages as direct conversational partners, studying their texts “in the spirit of a colleague, an antagonist, a student, a teacher,” has drawn charges of anachronism. That these charges are not well-grounded and that Bennett is sensitive to the task of historical recovery is shown in his excellent translation (with Peter Remnant) of Leibniz’s Nouveaux essais sur l’entendement humain. His commitment to early modern philosophy as a tool of teaching and learning is shown by his retirement activities as presented at www. earlymoderntexts.com. In his masterpiece Kant’s Analytic, Bennett sought to provide a post-Quinean empiricist reading of the earlier part of the first Critique. He articulates a phenomenalism which gets rid of noumena, and suggests that synthetic a priori truths are best understood as “unobvious analytic truths about the conditions under which . . . certain concepts can have a significant use” (see noumenal/ phenomenal). Bennett has no sympathy for atomic sense data, or for any “apprehended representations,” and proposes that we substitute “sensory states” in their stead. A famous chapter of that book inquires after the conditions under which a perceptual world bears the application of objective concepts. Bennett takes off from Strawon’s demonstration that merely out of auditory sensations one can develop spatial concepts. To allow the use of objective concepts, an auditory universe must possess, at a minimum, a certain identifiable ordering so that sounds can be located by their relation to other sounds. Bennett uses the time sequence to provide such order. Once locations can be established in the auditory universe, distinctions between misperceptions, hallucinations, and veridical hearings can start taking hold. Bennett then introduces the possibility of changes in the position of the sounds, as well as qualitative changes in the sounds themselves, and processes of generation and corruption. These complexities increase the hold of the objectivity concepts on the auditory universe. Throughout, Bennett points to the structural parallels between 143

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 144

b e n n e t t, jo nathan applying concepts to the auditory universe and to our actual phenomenal world. Bennett’s Strawsonian thought experiment establishes that “spatiality is sufficient for objectivity.” It does not conclusively prove that if there is objectivity then there must be spatiality. This is one of those synthetic a priori truths (in Bennett’s sense) that can only be supported by arguments which “consist in certain ways of assembling facts about meaning” and whose point is not to put forward claims which if true must be necessarily so: “The impossibility that there should be an objective but nonspatial world does not matter. What matters is a fact, if it is a fact, about the way in which objectivity and spatiality are connected in our conceptual scheme.” Later in the book Bennett establishes that the use of objective concepts is sufficient for self-consciousness. But he also reconstructs in Kant a plausible connection from selfconsciousness to objectivity, and therefore suggests that at least for humans it appears that self-consciousness is sufficient for spatiality. Bennett uses the thought-experiment to provide evidence for a corollary of Quine’s view that truth involves conceptual efficiency: namely, that legitimate concepts are simplifiers or abbreviators. Indeed, the simplifying power of objectivity concepts when compared to descriptions in purely sensorial terms increases with their hold on the auditory universe; the place which objectivity concepts gain in the thoughtexperiment could be established by considering instead their abbreviating advantages. In more recent work, Bennett has articulated the outlines of a Spinozistic “field metaphysics,” according to which space is the unique ontologically fundamental entity and physical bodies are second-level constructs out of spatiotemporally continuous “strings” of regions of the one substance of which certain qualities are predicated (see spinoza). This metaphysics is “neutral with respect to time: it can be displayed in terms of regions at instants or . . . throughout periods.” Bennett maintains that this ontology can illuminate issues such as whether two distinct bodies can occupy 144

exactly the same location at the same time: consider intersecting spatiotemporal strings, such that when they merge their degree-admitting qualities correspondingly intensify in the appropriate region. (This is not a surprising result given that in Bennett’s field metaphysics bodies are event-like entities supervening on properties of regions of space.) Another use to which Bennett puts the field metaphysics is to help us “see that space might contain items other than portions of matter and constructs out of them,” entities like forces and waves which are not in or of bodies. This, he comments, “seems to be precisely what has drawn contemporary physics in the direction of field theory.” wri t i ngs b ook s

Events and Their Names (Indianapolis, IN: Hackett, 1988). Kant’s Analytic (Cambridge: Cambridge University Press, 1966). Kant’s Dialectic (Cambridge: Cambridge University Press, 1974). Learning from Six Philosophers, 2 vols. (Oxford: Oxford University Press, 2001). Linguistic Behaviour (Cambridge: Cambridge University Press, 1976). Locke, Berkeley, Hume: Central Themes (Cambridge: Cambridge University Press, 1971). A Philosophical Guide to Conditionals (Oxford: Oxford University Press, 2003). A Study of Spinoza’s Ethics (Indianapolis, IN: Hackett, 1984). a rt i c l e s

“Comments on Dennett from a Cautious Ally”, Behavioral and Brain Science 16 (1993), 381–5. “Identity and Cardinality: Geach and Frege” (with William Alston), Philosophical Review 93 (1984), pp. 553–67. “On Being Forced to a Conclusion”, Proceedings of the Aristotelian Society, suppl. vol. 35 (1961), 15–34. jorge secada

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 145

bergmann, gustav Bentham, Jeremy (1748–1832) British philosopher who studied law. As a thinker whose “fundamental axiom” was the principle of utility, Bentham was a practical thinker. Hence traditional metaphysical questions, such as the existence of material objects, did not particularly interest him. He thought that we should suppose that they existed because “no bad consequences” could possibly arise from the supposition. The center of Bentham’s interest was not such abstract questions but the practical topic of law. However, his deep investigations into the nature of law led him, in spite of himself, into original metaphysical analysis. For he had to account for such legal entities as property, rights, duties and laws. Bentham said that metaphysics was “to know and to be able to make others know what it is we mean”. Doing metaphysics was making things comprehensible; metaphysical theories were theories of meaning. So, faced with explaining rights or duties, Bentham’s task was to explain how such terms as “right” or “duty” can have meaning. The traditional approach to an analysis of obligation or duty, as in Locke, would be to say that it was a complex idea, composed of simple ideas. Part of this Bentham adopts, in that he thinks analysis should terminate with simple ideas which are, or refer to, objects of direct perception, and which can be immediately understood. Prominent amongst such simple ideas for Bentham are pleasure and pain. These ideas he takes to be immediately comprehensible and universally understood. However Bentham saw that the method of directly analyzing a term like “obligation” into simple ideas does not work. So instead he invented the method he called paraphrasis. In this the term to be analyzed is placed in a sentence (for example “John is under an obligation to . . .”). This whole sentence is then taken to be equivalent to another sentence, which does not contain the term being analyzed but, rather, terms referring to more directly perceptible entities. So, for obligation, the analysis is by sentences mentioning sanctions (for example, “John is threatened with pain if he . . .”). We have now reached pain, a directly perceptible entity.

Entities which need such an analysis, Bentham called fictitious; entities which can be perceived, or inferred from perception, Bentham called real. Examples of fictitious entities are right, obligation, privilege, legal possession, property; and (away from the law) motion, quality, necessity, certainty. Examples of real entities are not just tables and chairs but also pains. Fictitious entities, whose meaning can be unfolded by analysis in terms of real entities, and which, Bentham says, are essential for the purposes of language and communication, are importantly different from what he calls fabulous entities. Fabulous entities, such as the Devil, or the golden mountain, are quite simply non-existent. They can be imagined, but there is nothing in reality corresponding to them. By contrast, if someone is said to be under a duty, although there is literally no duty there, this may still express something true, a truth which can only be properly unfolded when it is analyzed in terms of the sanctions threatened to the person who is said to be under the duty. writings Bentham’s Theory of Fictions, ed. C.K. Ogden (London: Kegan Paul, 1932). Chrestomathia (1817); ed. M.J. Smith and W.H. Burston (Oxford: Oxford University Press, 1983), appendix IV. Essay on Language, in Works, ed. J. Bowring (Edinburgh: Tait, 1843) vol. VIII. Essay on Logic, in Works, ed. J. Bowring (Edinburgh: Tait, 1843) vol. VIII. Fragment on Government (1776); ed. J.H. Burns and H.L.A. Hart (Cambridge: Cambridge University Press, 1988), ch. V. Fragment on Ontology, in Works ed. J. Bowring (Edinburgh: Tait, 1843) vol. VIII. Of Laws in General, ed. H.L.A. Hart (London: Athlone, 1970), Appendices B, C. ross harrison Bergmann, Gustav (1906–87) AustrianAmerican philosopher. Bergmann was one of the younger members of the Vienna Circle (see logical positivism). In 1938 he had to leave Austria and, with the help of Otto 145

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 146

b e r g m a nn, gustav Neurath (1882–1945), emigrated to the United States. In the United States. Herbert Feigl, another member of the Vienna Circle, secured a position for Bergmann at the University of Iowa initially as an assistant to Kurt Lewin in his endeavor to use topological methods in theoretical psychology. Bergmann taught at the University of Iowa until his retirement. At the beginning, like most members of the Vienna Circle, Bergmann was interested in the philosophy of science and, particularly, in the foundations of psychology. He eventually published a short book on philosophy of science (Bergmann, 1957). But around 1945 his interest turned more and more to metaphysics and to what he called “the heart of metaphysics”, namely, ontology. He used to tell an anecdote about how the title of his article “A positivistic metaphysics of consciousness” (1945) upset Carnap when he was visiting Iowa City. Much of Bergmann’s work in metaphysics is contained in two collections of articles (Bergmann, 1959, 1964). Unlike most members of the Vienna Circle and their students, Bergmann defended realism (see Platonism) against nominalism and mind/ body dualism against materialism (see physicalism, materialism; the extended essay on the mind/body problem). His philosophy, as a consequence, revolved around two main topics: The proper ontological analysis of ordinary perceptual objects and the intentional relationship between a mind and its objects (see intentionality). Bergmann held that the dialectic of numerical difference and qualitative sameness among perceptual objects requires an ontology of bare particulars (see bare particular) and universals. An ordinary object, a white billiard ball, is a complex entity, containing a bare particular and the universal whiteness. According to this analysis, an ordinary object turns out to be a fact. Bergmann’s ontology, therefore, embraces facts as well as particulars and universals. Particulars are called “bare” because they do not have natures. They resemble the haecceitates (see haecceity) of Duns Scotus. Universals can be subdivided into properties and relations. Among relations, a special place is occupied by the 146

nexus of exemplification, which holds between particulars and their universals, and the nexus of intentionality, which holds between minds and what they are about. A mind, according to Bergmann, consists essentially of mental acts. He follows in this regard in the footsteps of another school of Austrian philosophers, namely, Brentano and his students. Every mental act has two characteristic properties: a property that determines the kind of act it is, say, a remembering, or a desiring, and a property that determines which particular object an act intends. It is this latter property, the “content” of the act, which stands in a unique and unanalysable relation, the intentional nexus, to the object of the act. But a mind may intend an object that does not exist, and this creates one of the most intractable problems of the philosophy of mind, a problem that Bergmann discussed for many years: How can the intentional nexus hold between a mental act (“content”) and something that does not exist, that is not there at all? Bergmann discusses this problem and other problems of the philosophy of mind in his main work (Bergmann, 1967). His solution to the “problem of non-existent objects” is that states of affairs, the “objects” of mental acts, may exist in either of two modes: the mode of actuality or the mode of potentiality. A non-existent state of affairs, therefore, does exist, but exists in the mode of potentiality (see potentiality/actuality). Bergmann’s ontology culminates in the investigation of the ontological ground of logic. He finds this foundation in the world’s form. In particular, he attributes ontological status, subsistence, to the so-called “quantifiers” (generality and existence) and “connectives” (conjunction, disjunction, etc.) (see Bergmann, 1962). During the last years of his philosophical work, Bergmann developed a completely new ontology, an ontology explained in a posthumously published book (see Bergmann, 1991). wri t i ngs “Generality and Existence,” Theoria 28 (1962), 1–26.

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 147

berkel ey, george Logic and Reality (Madison, WI: University of Wisconsin Press, 1964). Meaning and Existence (Madison, WI: University of Wisconsin Press, 1959). New Foundations of Ontology, ed. William Heald (Madison, WI: University of Wisconsin Press, 1991). Philosophy of Science (Madison, WI: University of Wisconsin Press, 1957). “A Positivistic Metaphysics of Consciousness,” Mind 45 (1945), 193–226. Realism: A Critique of Brentano and Meinong (Madison, WI: University of Wisconsin Press, 1967). reinhardt grossmann

Bergson, Henri (1859–1941) French philosopher. Bergson formulated a new and impressive conception of metaphysics early in the twentieth century. It attracted wide attention not only for its content but also because of its opposition to the prevailing classical view that metaphysics is the inquiry into the universe as a whole. This inquiry was taken to be a purely intellectual one, aimed at embodying its results in a coherent system of ideas or basic truths about reality. Bergson rejected this view of metaphysics, because it mistakenly assumed that the human intellect is a truth-finding capacity, whereas it is in fact a capacity which has evolved to promote man’s practical action in the world. Because of its role the intellect treats what it deals with as individual entities in space, and seeks to understand them in mathematical terms. Hence the entities are regarded as static and immobile. But this is not the way the world is presented to us in immediate experience. Here we are aware of a continuous flowing of things and events in time. This time, however, is not the mathematical time of the physical sciences, but is what Bergson calls “duration” or real time. Scientific time is a fiction, albeit a useful one. Metaphysical time can only be obtained by having recourse to “intuition” or introspection of our immediate experience, not by the employment of the intellect in using abstract concepts. Hence the scientific picture of the universe as a mechanistic

and deterministic system in imaginary. It leads to a misconception of both freedom and creativity. Such phenomena are intimately related to the claim that evolution results from a vital impulse (élan vital) or current of consciousness that has penetrated matter and has given rise to a multiplicity of interwoven potentialities which constitute the evolutionary process. The above metaphysical doctrines are persuasively presented by Bergson in his writings, often with the aid of striking metaphors and analogies. But these may not satisfy many contemporary readers who are perplexed by the scarcity of logical arguments and supporting reasons for the doctrines being advanced. wri t i ngs Creative Evolution (London and New York: Macmillan, 1911). The Creative Mind (New York: The Wisdom Library, 1946). Introduction to Metaphysics (New York: Liberal Arts Press, 1949). Mind Energy (London and New York, 1914, 1920). Time and Free Will (London and New York: Macmillan and Co., 1910). thomas a. goudge

Berkeley, George (1685–1753) Irish philosopher. Berkeley argued forcefully against the existence of matter or material substance. His arguments, Hume later wrote, “admit of no answer and produce no conviction”, but Berkeley himself was convinced that the denial of matter (or immaterialism) was closer to common sense, more remote from skepticism, and friendlier to recent developments in science than the materialism (see physicalism) he found in Descartes, Malebranche, Locke and Newton. Berkeley did not deny the existence of bodies; instead, he construed statements about bodies as claims about perceptions or ideas. “The table I write on, I say, exists,” he wrote, “that is, I see and feel it; and if I were out of my study I should say it existed, meaning thereby 147

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 148

b e r k e l ey, geo r ge that if I was in my study I might perceive it, or that some other spirit actually does perceive it” (Works, vol. 2, p. 42). In 1707–8, as a student and fellow at Trinity College, Dublin, Berkeley completed two notebooks now known as the Philosophical Commentaries. They announce and argue for what Berkeley then called “the Principle”: to be is to be perceived, or to perceive (or will, or act). The houses, mountains and rivers whose esse is percipi (whose being is being perceived) have no existence apart from the minds or spirits who perceive or act upon them. t h e or y o f visio n Berkeley’s first important book, An Essay Towards a New Theory of Vision (1709), is an attempt to explain how we see the distance, size and orientation of objects, on the assumption (carried over from earlier writers) that they are not seen directly. Berkeley assumes that distance, size, and orientation are directly perceived by touch. We see distance, for example, only because experience invests visual appearances with tangible meaning. We come to associate tangible ideas (including ideas of our own body and its movements) with ideas attending vision – the sensation arising from the “turn” of the eyes, the “confusedness” of the visual appearance, and the strain of holding these appearances in focus. Correlations between the two kinds of ideas are contingent or “arbitrary”; Berkeley argues against the view, attributed (perhaps unfairly) to Descartes, that we compute an object’s distance by a kind of “innate geometry”, inferring it from the size of the angle formed by the two “optic axes” as they meet at the eye. Berkeley assumes in the Essay that the objects of touch do not depend for their existence on the mind. “Not that to suppose that vulgar error”, he later wrote, “was necessary for establishing the notion therein laid down, but because it was beside my purpose to examine and refute it in a discourse concerning vision” (Works, vol. 2, p. 59). Berkeley also speaks of the objects of touch as ideas, but he warns that the Essay does not assume that ideas are mind148

dependent. “When I speak of tangible ideas,” he explains, “I take the word idea for any immediate object of sense or understaning, in which large signification it is commonly used by the moderns” (Works, vol. 1, p. 188). Because the ideas attending vision are arbitrary signs of the objects of touch, Berkeley views them as a language – a visual language in which God speaks to us of tangible objects to come. Thus the Essay affords the first glimpse of Berkeley’s substitute for the image of nature as machine, blindly obeying laws laid down, long ago, by a now-indifferent God. For Berkeley, nature is a text or speech, renewed at every moment, and bespeaking a continuing providence. Like a text or speech, its signs have no power over what they signify. They are useful to us not because of what they bring about, but because of the divine intentions they communicate. abstrac t ideas Berkeley’s main work, A Treatise Concerning the Principles of Human Knowledge (1710), begins with an attempt to untangle what its author calls “the fine and subtle net of abstract ideas”. “It is agreed on all hands,” Berkeley reports, “that the qualities or modes of things do never really exist each of them apart by itself, and separated from all others, but are mixed, as it were, and blended together, several in the same object” (Works, vol. 2, p. 27). Yet according to the doctrine of abstraction (as contained, according to Berkeley, in Locke’s Essay Concerning Human Understanding (1689), the mind can, for example, form an idea of an object’s color apart from its other modes or qualities. When the mind later observes that different colors are alike, it can form “an idea of colour in abstract which is neither red, nor blue, nor white, nor any other determinate colour”. And “by the same precision or mental separation”, Berkeley reports, it is alleged to form abstract ideas of composite things, such as the idea of a human being, or of a triangle in general (Works, vol. 2, p. 28). Berkeley’s main argument against abstract ideas rests on the premise that we cannot conceive of the impossible. Because the object of an abstract idea cannot exist in isolation,

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 149

berkel ey, george it cannot be conceived in isolation. Locke himself had insisted that every existing thing, whether substance or mode, is particular. He inferred from this that every idea is particular. Berkeley urges the further conclusion that every idea is of a particular. Yet he does not deny the possibility of abstract or general thinking. When we think of human nature in general, he suggests, we consider or attend to a single aspect of a fully determinate idea. And when we prove theorems in geometry, we take a single triangle as the impartial representative of them all. Berkeley traces the doctrine of abstraction to the assumption that every significant word stands for an idea – an assumption he denies. He observes that a word need not excite an idea on every occasion of its use, and he argues that some words, used to express emotion or incite action, do not stand for ideas at all. In Alciphron (1732), Berkeley argues that words such as force and grace owe their meaning not to ideas we can conceive in isolation, but to their place in a system of signs with a bearing on practice or experience.

The medium of Berkeley’s philosophy is argument: he uses it not only to persuade, but to expound and clarify. Of the many arguments he offers against the existence of matter, the following five are central:

perceived” is “perfectly unintelligible” (Works, vol. 2, p. 42). (3) The notion of matter is either selfcontradictory or empty. It is selfcontradictory if sensible qualities are said to exist in it (so that they require nothing else for their existence), because sensible qualities are ideas, and ideas cannot exist without the mind. If we try to escape the contradiction by saying, vaguely, that matter is a substratum or “support” of qualities unknown, we deplete the notion of content. (4) It is impossible even to conceive of bodies “unthought of or without the mind”. “The mind, taking no notice of itself, is deluded to think” it can do so, but the attempt is self-defeating, because the bodies the mind brings forward as examples “are apprehended by or exist in it” (Works, vol. 2, pp. 50–1). (5) Even if matter exists we cannot know that it does. We cannot know it by sense, because we immediately perceive only our own ideas. Nor can we know it by reason – that is, by demonstrative or probable argument. We cannot know it by demonstrative argument because there is no necessary connection between matter and ideas. And we cannot know it by probable argument (that is, by explanatory inference) because we cannot comprehend the action of matter on the mind.

(1) It is a dictate of common sense that we immediately perceive such things as houses, mountains, and rivers. But philosophy teaches that we immediately perceive only our own ideas. (In the Principles, Berkeley tends to assume, without argument, that ideas are minddependent.) It follows that houses, mountains and rivers are ideas, and that they do not exist “without” (i.e., independently of) the mind. (2) If we inquire into the meaning of the word exist when it is applied to sensible things, we discover that it means only that they are perceived or perceivable. Hence the existence of unthinking things “without . . . relation to their being

Berkeley develops these arguments in both the Principles and the Three Dialogues between Hylas and Philonous (1713). In the Principles, the conflict between immaterialism and common sense is, at least at times, openly acknowledged. Berkeley writes at one point, for example, that belief in things without the mind is “strangely [i.e., greatly] prevailing amongst men”. In the Dialogues he is more concerned to emphasize the harmony between the two. The Dialogues also fills a gap in arguments (1) and (3), by arguing for the assumption that the immediate objects of perception are mind-dependent ideas. Philonous (Berkeley’s spokesman) seeks to establish this by an appeal to perceptual relativity: because the immediate

i m m a t e r ialism

149

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 150

b e r k e l ey, geo r ge objects of perception vary with changes in us, he argues, they must exist only in the mind. But neither Philonous nor Berkeley infers from this that qualities themselves exist there. They reach this further conclusion by arguing that mind-dependent ideas cannot represent mind-independent qualities. This is because one thing can represent another only if they are alike, and “an idea can be like nothing but an idea” (Works, vol. 2, p. 44). Berkeley repudiates a version of the distinction between primary and secondary qualities, according to which the primary qualities (such as extension and figure) are independent of the mind although the secondary qualities (color and taste, for example) are not (see quality, primary/ secondary). He argues that we cannot conceive of an object bereft of all secondary qualities; it follows that the primary qualities exist where the secondary do – “in the mind and nowhere else”. A purely geometrical conception of body is, he maintains, an illegitimate abstraction; our conception of body is forever marked or stained by its origin in sense. In De motu (1721) and Siris (1744), however, Berkeley shows that he is willing, for scientific purposes, to consider body as including only primary qualities. This distinction between primary and secondary qualities is pragmatic rather than metaphysical: the primary qualities, more useful in prediction and control, become objects of selective attention. Berkeley regards bodies or real things as sensations or “ideas of sense”. Although they are more regular, vivid, and constant than ideas of imagination, they are, he cautions, “nevertheless ideas” (Works, vol. 2, p. 54). Their reality – their greater strength, order and coherence – is no argument that they exist without the mind. Berkeley often suggests that bodies are clusters or collections of simpler ideas, but he provides little guidance as to how these clusters should be understood. Perhaps they are literal collections; if so, they seem to include ideas of several senses, existing in different minds at different times. But other passages suggest that Berkeley is a phenomenalist, holding that statements about 150

bodies are equivalent in meaning (or at the very least in truth conditions) to statements about what we perceive, or would perceive, under certain circumstances (see phenomenalism). substance and spi ri t Phenomenalism is sometimes described as “Berkeley without God”. But Berkeley’s phenomenalism is theocentric: statements about what we would perceive are true, he thinks, only because of the standing volitions of the deity. Berkeley recognizes that divine agency cannot be “blind”. Hence God’s sustaining activity has two aspects: he wills that we have certain ideas under certain circumstances, and he perceives the ideas he wills. Ideas of sense, Berkeley observes, are independent of our will. This fact, coupled with the wisdom and power they exhibit, constitutes Berkeley’s main argument for God’s existence. In Alciphron, Berkeley’s presentation of the argument emphasizes the languagelike character of our experience. The inference to God’s existence is akin to the inference from speech or writing to the existence of other finite minds. Berkeley argues that the only substance is spirit. He does not abandon the traditional view that perceived qualities or modes need a substratum. But the substratum in which they exist is a mind or spirit. Yet even though color and shape, for example, exist in the mind, they cannot be predicated of the mind. They are in the mind “not by way of mode or attribute, but by way of idea” (Works, vol. 2, p. 61). How does Berkeley know there are substances? He thinks our own substance-hood is known immediately and reflexively. But he insists that we have no idea of substance, because spirits are active beings, and ideas, being passive and inert, cannot resemble them. In the second edition of the Principles (1734), Berkeley explains that we have notions of mind or spirit. This is not a theory of representation, but a way of saying that we understand words such as mind or soul. The basis seems to be our understanding of the word I – our reflexive

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 151

bl anshard, brand awareness of our own selves. We are also said to have notions of relations, because according to Berkeley they involve an act of the comparing mind. science Berkeley argues that the only true causes are spirits; corporeal “causes” are marks or signs. We say, of course, that fire heats, and water cools, but “in such things we ought to think with the learned, and speak with the vulgar” (Works, vol. 2, p. 62). The only true cause at work in nature is God, “the author of nature”, and a scientific law is a rule of the language in which he speaks. t h e m e t ap hysics o f sir is Passages in Siris have convinced some readers that late in life, Berkeley turned to Platonism – to a belief in the existence of objects of pure intellect more real than sensible things, of which they are the patterns or archetypes. But Siris is often a tentative book, one whose “hoary maxims” are proposed not “as principles, but barely as hints” (Works, vol. 2, p. 157), and its objects of pure intellect are not, in any case, archetypes of sensible objects, but spirits, or aspects of spirit. writings The Works of George Berkeley, Bishop of Cloyne, ed. A.A. Luce and T.E. Jessop (London: Thomas Nelson, 1948–57). b i b l i og rap hy Atherton, M.: Berkeley’s Revolution in Vision (Ithaca, NY: Cornell University Press, 1990). Dancy. J.: Berkeley: An Introduction (Oxford: Blackwell, 1987). Grayling, A.C.: Berkeley: The Central Arguments (La Salle, IL: Open Court, 1985). Luce, A.A.: Berkeley and Malebranche (London: Oxford University Press, 1934). Pitcher, G.: Berkeley (London: Routledge and Kegan Paul, 1977).

Tipton, I.C.: Berkeley: The Philosophy of Immaterialism (London: Methuen, 1974). Urmson, J.O.: Berkeley (Oxford: Oxford University Press, 1982). Warnock, G.J.: Berkeley, 3rd edn. (Oxford: Blackwell, 1982). Winkler, K.P.: Berkeley: An Interpretation (Oxford: Clarendon Press, 1989). kenneth p. winkler Blanshard, Brand (1892–1987) Blanshard integrated metaphysics with epistemology and the other branches of philosophy in a systematic whole. His The Nature of Thought (1939) moves from psychological and epistemological investigations to the more abstruse topics of metaphysics. Metaphysically, he advocated the theory of the concrete universal, the doctrine that all relations are internal, and the thesis of cosmic necessity. For Blanshard, the things which are the objects of perception are not the ultimately real things. When analyzed into their properties and relations, they are found to be interconnected with all other things, their very natures being affected by their relations. Based on the doctrine of internal relations, Blanshard’s theory envisages a network of relatedness among things tantamount to the entire universe. Moreover, on Blanshard’s account, universals, the objects of ideas, are more real than things. The kinds of universals are abstract, generic, qualitative and specific. Rejecting abstract universals, he favored instead the theory of the concrete universal. The generic universal he viewed as a being in thought requiring further determination, while he considered the qualitative universal, such as whiteness or sweetness, to be subsumable under the generic universal. Specific universals, such as a specific color or odor or taste, he admitted into nature, contending that individual things are congeries of such universals. Because every thing is composed of specific universals and is internally related to all other things, its particularity as spatio-temporal is exposed as unreal, and must be reconceived as its being part of a whole. Hence “the only true 151

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 152

body particular is the absolute” (Blanshard, 1939, vol. I, p. 639). Blanshard’s theory of the world as a system of necessarily related parts is, in sum, his conception of cosmic necessity, a necessity both causal and logical. Since the aim of thinking is to express this cosmic necessity, reason is the human faculty which seeks the relations that bind all things together in a necessary whole. In mid-career Blanshard undertook to prepare a trilogy in defense of reason, construed as sovereign, against what he despised as its detractors in contemporary thought. At this juncture Blanshard preferred that his philosophy be called “rationalist” rather than “idealist” (see idealism; rationalism). The volumes of Blanshard’s trilogy are Reason and Goodness (1961), Reason and Analysis (1962), and Reason and Belief (1974). The absolute for Blanshard is the only true individual. In principle intelligible throughout, it embraces all things in the network of necessary relations. Yet Blanshard could find no reason for, and many reasons against, the attributions of rightness and goodness, mind or consciousness, to the absolute. Ultimate reality, the universe, the absolute (he used these terms interchangeably), is to be found “in no part of it, however great, but only in the whole. It is the universe itself, not indeed as a scattered litter of items but as the one comprehensive and necessary order that a full understanding would find in it” (Blanshard, 1974, p. 523). w r i t i n gs The Nature of Thought (1939); (New York: Macmillan, 1940). Reason and Analysis (New York: Macmillan, 1962). Reason and Belief (New York: Macmillan, 1974). Reason and Goodness (New York: Macmillan, 1961). andrew j. reck body John Locke (Locke, 1690) distinguishes between two kinds of bodies: mere masses of matter and living bodies. These 152

kinds are distinguished from each other and from persons by way of their persistence conditions. What it takes for a person to survive, for a person at one time to be identical to some person that exists at a later time, is for there to be a continuity of consciousness between the earlier person and the later one. In contrast, a living body can survive a change of parts without any consciousness being present, so long as that change is in accordance with the kind of diachronic organization that would be considered a single life. Masses, on the other hand, cannot survive any gain or loss of parts. What it is to be a given mass is just to be that collection of that stuff. A collection of different stuff is, therefore, a different mass. So says Locke. The intuition behind Locke’s identity conditions for masses is that they are mere masses; all there is to such an object is the matter that composes it. If we are to respect this “mereness”, there seems no principled reason for placing any restrictions on which collections of matter should count as objects. Any collection of matter, no matter how arbitrarily grouped, has just as much claim to being a mass as any other collection. If we view the world four-dimensionally, as I prefer, the grouping of matter along the temporal dimension is equally arbitrary. Any collection of matter at one time and any collection of matter at a later time compose, along the temporal dimension, a single persisting mass. If the earlier collection and the latter one contain different parts, the persisting mass that is composed of them has survived a change of parts. Whereas the mereness of masses led Locke to hold very strict conditions for their continued existence, so strict as to prohibit change of parts, the mereness leads me in just the opposite direction. I call fourdimensional collections of matter “hunks” and hold that any filled region of spacetime, no matter how arbitrarily delineated, contains such a body. Many philosophers would insist that a body must at least be spatio-temporally continuous. Even a mere collection of matter must be collected, must have some internal unity. However, spatio-temporal continuity itself seems an inadequate condition for

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 153

boet hius unity. Along the spatial dimension, contact alone does not seem a significant enough connection to make two objects compose one. And along the temporal dimension, it seems possible for one object to be replaced by an extremely similar object in a way that makes for spatio-temporal continuity without the two objects either being or composing a single object. Thus many philosophers are led to require causal connectedness between the temporal stages of an object and between the parts of any one stage. It is not easy to specify which kinds of causal connections are the right kinds to generate a single body. Perhaps one plausible candidate is provided by Locke: the various interactions between objects are of the right sort to make those objects compose a single object if the interactions together constitute a life. But this is no clearer than the term “life”. Furthermore, like Locke’s persistence conditions for masses, his conditions for living bodies will exclude many ordinary objects. So, if we want to count ordinary objects as bodies we need some criteria in addition to, or instead of, Locke’s. How are we to choose the right persistence conditions for bodies from among all these alternatives? Perhaps we should simply accept many different kinds of bodies, a different kind for each persistence condition that has been mentioned. Locke himself accepted both masses and living bodies. This suggestion becomes less plausible once we realize that a great multitude of persistence conditions might be recommended and that many of these different kinds of bodies would be existing simultaneously in a single location. To avoid this problem, we should select one persistence condition. I accept the non-restrictive condition: any collection of matter, no matter how arbitrarily delineated, is a body. What there are are mere hunks of matter. Some of these hunks are counted as mountains, some as dogs and some as people. A single hunk might count as a living thing, a person, an athlete, an adult and a woman, but this does not require that there be several different things in one place simultaneously. It is one body that plays different roles. In spite of the many arguments in the literature defending

more restrictive conditions the selection of any of those conditions, I believe, would be fundamentally arbitrary. Positing mere hunks recognizes the arbitrariness involved in any distinction between bodies and nonbodies and respects that arbitrariness by rejecting the distinction. See also Aristotle; bundle theory; change; chisholm; continuant; identity; matter; ontology; part/whole; persons and personal identity; substance; temporal parts, stages; see also the extended essay on modality and possible worlds. bibl i ography Cartwright, R.: “Scattered Objects,” in Analysis and Metaphysics, ed. K. Lehrer (Dordrecht: Reidel, 1975), 153–71. Chisholm, R.M.: “Parts as Essential to Their Wholes,” Review of Metaphysics 26 (1973), 581–603. Heller, M.: The Ontology of Physical Objects (Cambridge: Cambridge University Press, 1990). Locke, J.: An Essay Concerning Human Understanding (London, 1690); ed. P.H. Nidditch (Oxford: Oxford University Press, 1975), 328–48. van Inwagen, P.: Material Beings (Ithaca, NY: Cornell University Press, 1990). mark heller Boethius (c.480–524) Roman philosopher. Boethius’s work that had the greatest influence on the history of metaphysics in the Latin world was not his famous De consolatione philosophiae, but a treatise that was referred to in the Middle Ages as De hebdomadibus. Its real title is a question submitted to Boethius by a friend: “How can substances be good in virtue of the fact that they have being when they are not substantial goods?” (Quomodo substantiae in eo quod sint bonae sint cum non sint substantialia bona). The most striking thing about this work is the way in which Boethius approaches the problem. He will solve this question according to the method “that is usual in mathematics”. His exposition starts therefore with eight 153

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 154

b o l z a n o, b er nar d propositions from which the remainder of the argument can be deduced. Boethius presents the model of an axiomatic metaphysics that proceeds more geometrico. The second axiom reads: “Being (esse) and that which is (quod est) are different.” For being itself does not exist, but that which actually exists is “that which is”. The precise meaning of this difference is controversial. It is usually interpreted as the distinction between a concrete thing and its substantial form, that by which a thing is (quo est). Boethius uses this distinction for the explanation of the ontological difference between created being and the highest being. The mark of created being is that it is composed: its being and quod est are not identical. Through this composition it is distinguished from the highest being, which is simple. “Every simple has its being and that which is as one” (Axiom VIII). De hebdomadibus was intensively commented upon from the Carolingian times. One of the most important commentaries was that of Aquinas in the thirteenth century. In that century Boethius’s axioms were frequently cited in the debates about the distinction between essence and being (existence) in things. w r i t i n gs The Latin text of De hebdomadibus with an English translation in H.F. Steward, E.K. Rand, and S.J. Tester, ed. and trans.: Boethius, The Theological Tractates and the Consolation of Philosophy (Cambridge, MA: Harvard University Press, 1978). b i b l i og rap hy McInerny, R.: Boethius and Aquinas (Washington, DC: The Catholic University of America Press, 1990). Schrimpf, G.: Die Axiomenschrift des Boethius (De hebdomadibus) als philosophisches Lehrbuch des Mittelalters (Leiden: Brill, 1966). jan a. aertsen Bolzano, Bernard (1781–1848) Mathematician, logician and metaphysician, 154

Bolzano was professor of theology in Prague 1805–19, when he was dismissed because of his liberal views. Bolzano’s main work, the Wissenschaftslehre (1837) (Theory of Science), postulates propositions “in themselves”, defined as “assertions . . . [which] may or may not have been put in words or even formulated in thought” (1837, sect. 19) (see proposition, state of affairs). Propositions are the “matter” of human judgments; they, not the acts of thinking are the concern of logic. Properties and relations of these Platonic (see plato; platonism) entities, like analyticity, implication and probability, are defined for the first time with the aid of variables. When a truth implies and explains a second, it is its ground. This relation orders truths in themselves so that “from the smallest number of simple premisses [follows] the largest possible number of the remaining truths” (1837, sect. 221). In opposition to Kant he took mathematics to be grounded not in intuition, but in pure concepts. He strove to provide adequate arithmetic definitions of real number, function, continuity, limit, point, etc., recast the foundations of geometry in terms of structured point sets, and showed that infinite sets have no paradoxical properties (1851). Bolzano took matter to be a continuous array of monads (see monad, monadology), and argued that the soul is simple and indestructible (1827). He offered a new cosmological argument for the existence of God, claiming that mutable substances require a “constant and unchangeable source of force” (1827, p. 296). In agreement with Bentham he measured all actions by the standard of public utility; even religion is “the sum of doctrines or opinions that have a beneficial or detrimental effect upon the virtue or happiness of a person”. Bolzano, a lone forerunner of analytic philosophy, endeavored to clarify basic philosophical and mathematical concepts which “everyone knows and does not know”. He advanced no claim without circumspect argument; his exposition is a model of clarity and precision. Regrettably, his attempt to rescue philosophy from the epigones of idealism failed because of the growing rift

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 155

bradley,f ranc i s herbert between philosophy and the exact sciences, and because of his persecution. Interest was revived through Husserl’s Logical Investigations (1900–1), and is now reflected in numerous articles and monographs (see the bibliographies in the Gesamtausgabe (1969ff.) ). writings Athanasia (Sulzbach, 1827); (Frankfurt: Minerva, 1970). Gesamtausgabe; founded by E. Winter, ed. J. Berg, with B. van Rootselaar, A. van der Lugt, J. Lougil, et al. (Stuttgart: Fromann, 1969ff.), 36 vols. in 1994. With introductions, comprehensive bibliographies, and a biography by E. Winter. Paradoxien des Unendlichen (Leipzig, 1851); trans. D.A. Steele, Paradoxes of the Infinite (London: Routledge, 1950). Wissenschaftslehre (Sulzbach, 1837); ed. W. Schultz (Leipzig: Meiner, 1929); (Aachen: Scientia, 1970); trans. R. George Theory of Science (Oxford: Blackwell, and Berkeley: University of California Press, 1972). Another edition ed. J. Berg, trans. B. Terrell (Dordrecht: Reidel, 1973).

without duration. There are also inner boundaries, like the half-way point in the flight of an arrow. Boundaries in this sense raise many ontological questions. Do they really exist or are they mathematical fictions? Are they parts of their objects, or of the surroundings, or neither? Alternatively, boundaries are simply “thin” parts of the same dimensionality as their wholes. At stake is whether the highly successful mathematics of continuous structures, like the real numbers, which treat extents as composed of extensionless points, truly depict reality. bi bliography Aristotle: Metaphysics, ed. W. Jaeger (Oxford: Clarendon Press, 1957), Bk. 5, ch. 17, 1022; Bk. 11, ch. 3, 1061. Stroll, A.: Surfaces (Minneapolis: University of Minnesota Press, 1988). Whitehead, A.N.: An Enquiry Concerning the Principles of Natural Knowledge (Cambridge: Cambridge University Press, 1919); 2nd edn. (Cambridge: Cambridge University Press, 1925), esp. Part III, “The Method of Extensive Abstraction.” peter simons

b i b l i og rap hy Berg, J.: Bolzano’s Logik (Stockholm: Almqvist and Wiksell, 1962). Coffa, J.A.: The Semantic Tradition from Kant to Carnap (Cambridge: Cambridge University Press, 1991). Morscher, E.: Das logische An-Sich bei Bernard Bolzano (Salzburg and Munich: Anton Pustet, 1973). rolf george paul rusnock

Bradley, Francis Herbert (1846–1924) British philosopher. A convenient way of placing Bradiey’s monism (see monism/ pluralism) and idealism in context is to see him as finding logical and epistemological grounds for rejecting (1) an ultimate ontology of externally related facts (see logical atomism); and (2) a physicalism (see physicalism, materialism) of the influential physics-based kind (cf. Quine). the rejec t i on of ul t i mat e f acts

boundary The boundaries of extended objects may be thought of in two ways: as limits or as thin parts. Limits of the object have fewer dimensions than it has itself: a three-dimensional brick has surfaces without thickness; the edge where two faces meet is a one-dimensional line; the corner where three faces meet is a point. An enduring event like a kiss has a beginning and an end

This rejection rests on Bradley’s often misunderstood argument to the effect that unconditional predication is incoherent. The argument is here stated for monadic predications in respect of a single individual as ultimate subject but it applies equally to relational predications with respect to pairs, etc. (Essays on Truth and Reality (1914) pp. 225–33; Appearance and Reality (1893) 155

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 156

b r a d l e y,fr ancis her b er t chs. II, III; Principles of Logic (1883) vol. I, pp. 99–100). Assume “R” is the proper name of an individual. If “Ra”, “Rb”, “Rc”, etc. express genuinely unconditional predications then the only condition under which, for example, “Ra” could be true would be R’s being a, and the only condition under which “Ra” could be false would be R’s not being a. The truth value of “Ra” could not depend on the truth values of any other of the propositions “Rb”, “Rc”, etc. Hence such propositions would all have to be logically independent of one another and any one of them could, as a matter of logical possibility, be the only one true, i.e., be a complete description of and give “perfect” knowledge of reality. But if, for example, “Ra” alone could be a complete description of reality then “Ra” must be construed as asserting not merely that R is a but in effect as asserting that R is merely a; and likewise for “Rb”, “Rc”, etc. However, if this is so, as Bradley argues, a contradiction arises, since if “Ra” were true then “Rb” and “Rc” etc. would have to be false (given that a is neither b nor c, etc.). Bradley concludes that the unconditional verbal form which we must in fact employ to express the predications involved in our thinking must be seen as misleading with respect to the form of the judgments we make. Any judgment must always be radically conditional in form and might more properly be expressed by, for example, the formula “R(x)a”. This formula must be read not as indicating that the truth value of any conceivable predication will be conditioned, at the ultimate limit of analysis, by conditions that will in fact be unknown. It must be read as indicating that there cannot conceivably be an absolute, unconditional, determination of the truth value of any predication. Hence our knowledge cannot be a superstructure underpinned by an ultimate ontology of externally related. or atomistic, perceptible facts (Essays, pp. 209–10). t r u t h s , d egr ees o f tr uth and “ i d e a l co nstr uctio ns” Bradley therefore concludes that with respect to any linguistically communicable truth 156

whatsoever it can only make sense to describe it as true, or false, relative to a specified ideal construction. Any such a construction will be a more or less comprehensive logically interrelated system of ideal contents predicated as a connected whole of reality (i.e., of the genuinely individual, ultimate subject of predication). Thus for Bradley the primary repository of truth comes to be, not the proposition construed as a content capable of possessing in isolation a determinate eternal truth value, but the ideal construction viewed diachronically. Hence an ideal construction is not (except for special purposes) to be construed as a determinate system of contents subject to the operations of a Fregean propositional or predicate calculus (see Frege). In its primary sense it will be a system of ostensibly coherent ideal contents which is continuously confrontable in practice with a given and which is prone to modification and supplantation by extended, and sometimes logically incompatible, versions of itself (e.g., Essays, pp. 75–9). Such extended versions will, so far as they are internally coherent, allow a person to have knowledge which will be nearer – but on a path which of necessity is only asymptotic – to absolute truth with respect to the given reality. t he ideal c onst ructions of physi c i sts as not privi l eged met aphysical l y If the argument against absolute truth and falsehood is valid it follows that no system of objects knowable through any ideal construction (e.g., the physical world as identified through the ideal contents internal to ultimate particle theories) can be taken as such, for the purpose of metaphysics, to be identical with reality. In Bradley’s alternative terminology, no such system can constitute anything other than a more or less partial and inadequate appearance of reality. However, equally, it follows that reality will be present, albeit as a partial and inadequate appearance, in the contents of even the most fragmentary ideal constructions (Essays, pp. 28– 42; Appearance and Reality, pp. 323–7; Principles, vol. I, pp. 110–11, fn. 40). Hence for Bradley

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 157

bradley,f ranc i s herbert there can be no question but that the theories of physics, at any time in their history, will contain truth and allow us to have more or less extensive knowledge of reality. But this knowledge will have no special significance for the metaphysician, despite its enormous practical significance in our real world (Essays, p. 123; Appearance and Reality, pp. 231–6, 434–5). e p i s t e m olo gical p r io r ity o f t h e i d e al co nstr uctio n o f ou r r eal wo r ld The system of objects, according to Bradley, that we ordinarily call “our real world” will be the object of our knowledge when, at any waking moment, we think of the unlimited totality of particulars of humanly perceptible kinds and sentient subjects that presently exist, or have existed, or will exist, at some time, somewhere in space, relative to the egocentrically demonstrable objects of our present perceptions (i.e., relative to our bodies) (Essays, pp. 28– 49, ch. XVI; Appearance and Reality, pp. 187–8). It is by contrast with this world alone that we can give primary application to distinctions like those between the genuinely historical and the fictional, the real and the merely imagined, the existent and the non-existent, what is true and what is false, what is actual and what is merely possible, etc. And it is within terms of the more or less fragmentary constructions from which the ideal construction of our real world will have developed in the course of our lives that we initially come to have knowledge of ourselves as opposed to others, of the inner as opposed to the outer and so on (Essays, pp. 356–7). Hence, it is from within the ideal construction of our real world, the material for which is fundamentally given in our waking sense perceptions, that any of the indefinitely various, more or less discrete, ideal constructions that we frame will be predicated of reality (e.g., Essays, p. 210). t i m e a n d sp ace no t “p r incip les o f i n d i viduatio n” At this juncture it is essential to appreciate the significance of Bradley’s contention that

time and space are not “principles of individuation” (Principles, vol. I, pp. 63– 4; see also Appearance and Reality, Appendix C, pp. 527–33). Any ideal content that we can exercise in thinking will be universal (e.g., Appearance and Reality, p. 34) hence, Bradley maintains, there can be nothing in the idea of a temporal and/or spatial series that logically guarantees that there is not an indefinite multiplicity of spatio-temporally unrelated spatio-temporal series. We can have no reason, therefore, to hold that the spatio-temporal series of our real world is uniquely real. In fact Bradley maintains that there is an indefinite multiplicity of such series that can be objects of our knowledge (e.g., Appearance and Reality, pp. 186–7). For example, the spatio-temporal series of our real world is distinguishable in our thinking from those spatio-temporally unrelated series which are the intentional objects of dream experiences, of works of fiction, of the ideal constructions of ultimate particle physicists and so on. The reality of such series, Bradley maintains, cannot be reduced to the datable psychological acts occurring in the spatio-temporal series of our real world but on the other hand they clearly cannot be thought of as identical with any part of that series. We can distinguish our so-called real spatio-temporal series from indefinitely many less real series by reference to the idea that it is the one that contains the intentional objects of these perceptions we are having now. However, given that time and space are not principles of individuation, it follows that the uniqueness of these perceptions that we are having (e.g., in reading this now) cannot, according to Bradley’s metaphysics, be consistently thought to derive their uniqueness from being datable states attributable to particulars of spatio-temporally locatable kinds existing in a given uniquely real spatio-temporal series. bradley’s f i nite centers of immedi ate experience Such experiences are, Bradley maintains, to be construed ultimately (i.e., within metaphysics) as experiences in a plurality of 157

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 158

b r e n t a no , fr anz finite centers of immediate experience. It is only within and via the representative activities of such centers that systems of intentional objects of any discernible kinds whatsoever (mental as opposed to physical, human as opposed to non-human, self as opposed to not-self, temporal as opposed to eternal, etc.) can be distinguished and known. However, these centers are not construed as Leibniz’s monads are (see monad, monadology): they are not contingently existing individuals. Their mode of being, Bradley holds, must be taken to be adjectival on that which is truly individual or real. Nevertheless Bradley maintains that in the experiences and activities of the plurality of finite centers that which is truly individual (the Absolute) can be coherently taken (1) to have its whole being and self-realization; (2) to be immediately but non-relationally present; and (3) to be knowable propositionally with increasing – but of necessity never complete – adequacy through increasingly coherent and comprehensive ideal constructions (Appearance and Reality, chs. XIII, XIV, XXVI; Essays, chs. XIV and XI, p. 350, fn. 1; Principles, vol. II, Bk. III, pp. 590–1, p. 595, fn. 25). w r i t i n gs Appearance and Reality. A Metaphysical Essay (London, 1893); (Oxford: Clarendon Press, 1968). Essays on Truth and Reality (Oxford: Clarendon Press, 1914); (Oxford: Clarendon Press, 1962). Principles of Logic (London, 1883); 2nd rev. ed. with commentary and terminal essays, corrected, 2 vols. (Oxford: Oxford University Press, 1967). b i b l i og rap hy Ingardia, Richard: Bradley: A Research Bibliography (Bowling Green, OH: The Philosophy Documentation Center, 1991). Manser, A.R. and Stock, G., eds.: The Philosophy of F.H. Bradley (Oxford: Clarendon Press, 1984). Wollheim, R.: F.H. Bradley (Harmondsworth: Penguin Books, 1959). guy stock 158

Brentano, Franz (1838–1917) German philosopher-psychologist. Brentano taught for much of his life in the University of Vienna, where his students included Husserl, Christian von Ehrenfels (1859–1932), Carl Stumpf (1848–1936), Kasimir Twardowski (1866–1938) and Meinong. Of these Husserl, notoriously, was the founder of phenomenology, Ehrenfels and Stumpf were instrumental in the formation of the Gestalt-psychological movement in Berlin, Twardowski was almost single-handedly responsible for the founding of modern Polish philosophy, and Meinong established what has come to be known as the “theory of objects”. Common to all of these thinkers is the use of psychology, following the example of Brentano himself, as the basis for the development of new and original ideas in ontology. Brentano’s rigorous and analytic style of teaching and his doctrine of the unity of scientific method (see unity of science) formed part of the background also of the logical positivism of the Vienna circle. Brentano’s early works concern the metaphysics and psychology of Aristotle. For Aristotle, as seen through Brentano’s eyes, the two realms of thinking and of corporeal substance are, as it were, attuned to each other. Perceiving and thinking amount to something like a taking in of form from the one into the other. Forms or universals exist, accordingly, in two different ways: within corporeal substance and (as “inexistent”) within the soul. They exist only as immanent to individual substances in one or other of these two different ways. When I see a red object, then I see something that is composed of matter and form. What I take in is the form alone, but this form is in fact still connected to (and thus individuated by) its matter. What I know intellectually is this form itself, for example the redness. And this is not a transcendent redness subsisting in some Platonic realm, but rather a redness here on earth (see Plato; Platonism). Only one sort of essence is, as far as Aristotle is concerned, free of materiality in this sense: the essence mind or intellect. Of this essence, and of the concepts abstracted therefrom, we can have knowledge other than via sensory images. Mind or intellect is,

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 159

brentano, franz as Brentano puts it, “with the highest intelligibility completely intelligible” (1867, p. 136, trans, p. 90). Psychology, accordingly, enjoys a peculiarly noble status within the system of the sciences, and our knowledge of psychological phenomena (for example of mental causality, of the relations of part, whole and dependence among mental phenomena) can provide a firm foundation for our knowledge of corresponding concepts as these are applied also to entities of other sorts. It will be clear from the above how one has properly to interpret Brentano’s thesis in Psychology from an Empirical Standpoint (1924) to the effect that “Every mental phenomenon is characterized by what the Scholastics of the Middle Ages called the intentional (or mental) inexistence of an object” (1924, p. 124, trans. p. 88). As Brentano himself puts it in the very next sentence: “Every mental phenomenon includes something as object within itself.” This thesis is to be taken literally – against the grain of a seemingly unshakeable tendency to twist Brentano’s words at this point. Only in the writings of Husserl, Meinong and other students of Brentano do we find a systematic treatment of intentionality as a matter of the mind’s directedness to transcendent objects in the world. By the time of his lectures on descriptive psychology given in Vienna University in 1889–90, Brentano has developed a rich ontological theory of parts and of unity. As Brentano himself puts it, he seeks to construct a psychological characteristica universalis, whose letters and words would reflect the different mental constituents or elements of the mind, and whose syntax would reflect the relations between these constituents in different sorts of mental wholes. His ideas here can be seen to stand at the beginning of a tradition which results inter alia in Husserl’s development of the formal ontology of parts and wholes in the Logical Investigations, as also in Leiniewskian mereology and categorial grammar (see Le1niewski). In the theory of substance and accident put forward toward the end of his life (see his Theory of Categories (1933)), Brentano

adopts a new sort of mono-categorial ontology, seeking once more to develop and refine an original Aristotelian theory. Where Brentano had earlier held that mental acts have an inferior being in relation to their subjects, he gradually came to believe that all entities exist in the same way, that “existence” has only a strict and proper sense (that all uses of this term which depart therefrom are illegitimate). Everything that exists, he now says, is a concretum, a “real thing”. Hence he has to find some way of coping with what Aristotle wants to say about the relation between accident and substance – and with what he himself wants to say about mental acts and their subjects – without appealing to special, inferior, “dependent” entities. Brentano solves this problem by turning Aristotle’s theory on its head: it is not, for Brentano, that the accident is an inferior entity existing in or on its substance. Rather, the substance itself is included within the accident as its proper part. That is, Brentano conceives the accident as the substance itself augmented in a certain way. Thus when one has a mental act, then the subject of this act (one’s self) is present as a part of the act. The act, according to Brentano, is not some extra entity attached to the self; it is the self momentarily augmenting itself, mentally, in a certain way. This gives Brentano a means of explaining how it is, when one is seeing and hearing, that it is the same self that is subject in both acts. That is, it gives him a means of accounting for the unity of consciousness, which is to say, for the fact that experience does not resolve itself into a bundle or multiplicity of scattered bits. It is crucial to the Brentanian theory that there be no extra entity which would make up the difference between substance and accident. For this third entity would be precisely an “inferior existent” of the sort he is now determined to get rid of. An accident is a thing, no less than its substance. There are no jumps and runs, on this new dispensation, but only jumpers and runners; no thinkings and perceivings, but only thinkers and perceivers. In this way, as Chisholm has noted, Brentano anticipated 159

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 160

b r o a d , char lie d unb ar contemporary developments in the direction of an Adverbial theory of perception. What, then, are the ultimate substances of Brentano’s ontology? One group of ultimate substances we have met already: they are the mental substances or souls which become augmented to form those half-way familiar things we call hearers, thinkers, and so on. It is natural, now, to suppose that the remaining ultimate substances in the Brentanian ontology are just material or concrete things, and Brentano’s philosophy has indeed often been interpreted along these lines, particularly by those who would see him as having anticipated a reist or concretist doctrine of the sort propounded by Leiniewski or Kotarbinski (1886–1981). In fact, however, Brentano takes as nonmental substances – as ultimate individuators – the places which material things occupy. Things in the normal sense are accidents of such places. The totality of places is itself a substance, a certain spatial continuum (see space and time). Movement within this continuum is not, as we normally suppose, a matter of the perseveration of one thing through a continuum of places which it successively occupies. Rather, it is a matter of neighboring parts of the unitary substance experiencing in succession a chain or ripple of similar accidental determinations. Here, therefore, Brentano anticipates later substantival interpretations of the space–time continuum which were formulated in the wake of the Special Theory of Relativity. w r i t i n gs Deskriptive Psychologie; ed. R.M. Chisholm and W. Baumgartner (Hamburg: Meiner, 1982). Kategorienlehre (Hamburg: Meiner, 1933); trans. R.M. Chisholm and N. Guterman, The Theory of Categories (The Hague, Boston, MA, and London: Martinus Nijhoff, 1981). Psychologie des Aristoteles (Mainz, 1867); trans. R. George, The Psychology of Aristotle (Berkeley: University of California Press, 1977).

160

Psychologie vom empirischen Standpunkt, vol. I (Leipzig: Meiner, 1874), 2nd edn. (Leipzig: Meiner, 1924); trans. A.C. Rancurello, D.B. Terrell, and L.L. McAlister, Psychology from an Empirical Standpoint. bibl i ography Chisholm, R.M.: Brentano and Meinong Studies (Amsterdam: Rodopi, 1982). Smith, B.: Austrian Philosophy: The Legacy of Franz Brentano (La Salle, IL: Open Court, 1994), ch. 3. barry smith Broad, Charlie Dunbar (1887–1971) British philosopher. Broad wrote extensively about a wide variety of traditional metaphysical topics, including existence, substance, qualities, relations, things, processes, events, change, time, space, causation, objects, mind, self and consciousness. His views were formed before the impact of “linguistic method” made philosophers more cautious about regarding metaphysics as a search for general truths about reality. He did not engage in speculative system building, but in “critical philosophy”, analyzing and clarifying fundamental concepts, drawing distinctions overlooked by common sense and science alike, and examining dispassionately the evidence for our basic beliefs. Few could match his ability to distinguish subtly different theses or his patience in marshalling arguments for and against each. Often he conceded that he was not sure which thesis won out, and, as one might expect of someone originally trained in the sciences, he hoped that empirical evidence would eventually provide answers. Scientific Thought (1923) is notable for his discussion of the analogies and disanalogies between space and time, as well as for a critique of the argument against the reality of time by his teacher and predecessor at Cambridge, McTaggart (1923, ch. 2). Broad offered final views on this subject in volume 2, part 1, of Examination of McTaggart’s Philosophy (1933–8). Discussing causation (see the extended essay), he expressed doubts

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 161

bundl e t heory about the orthodox analysis of singular causal propositions in terms of general laws (1933, pp. 241–5). In Scientific Thought (ch. 8) and The Mind and Its Place in Nature (1925, ch. 4), he followed Locke in arguing that physical objects have shape, size, position and mass, the “primary qualities” recognized by physics, but no “secondary qualities”, such as color or temperature (see quality, primary/secondary). Public, persistent, physical objects act causally on our perceptual systems to produce private, short-lived, nonphysical, and mind-dependent existents he called “sensa”. These sensa have the familiar visual, auditory, and tactual qualities that constitute perceptual experience, which is the basis for our judgments about reality. Broad was unmoved by the objection that introducing sensa makes the existence of physical objects a speculative hypothesis. The ontology of human beings presented in The Mind and Its Place in Nature is akin to but not identical with classical dualism. Substantial vitalism and the more modern view that organisms are biological mechanisms are rejected in favor of “emergent vitalism”, the belief that their behavior is due to properties of matter which first appear at the organic level (1925, ch. 2). Reductive materialism (see physicalism, materialism), which identifies mental processes with molecular movements in the brain, he judged to be false because the two types of events have different properties (1925, p. 622). To analytic behaviorism he objected that, however completely a body answers to behavioristic tests for intelligence, it makes sense to ask “Has it a mind or is it an automaton?” (1925, p. 614). His own suggestion was that a mind is a “compound” comprising a living brain and nervous system together with a “psychic factor”, which interacts with appropriate living organisms to produce mental activity and which might even survive the death of the body (1925, p. 651). He favored this theory because it accommodated evidence for paranormal psychical phenomena that he thought should be taken seriously.

wri t i ngs Examination of McTaggart’s Philosophy, 2 vols., Vol. 1 (Cambridge: Cambridge University Press, 1933), Vol. 2 in two parts (Cambridge: Cambridge University Press, 1938). Kant: An Introduction, ed. C. Lewy (Cambridge: Cambridge University Press, 1978). This contains Broad’s Cambridge lectures (1950–1, 1951–2) on the philosophy of Kant, esp. as found in the Critique of Pure Reason. The Mind and Its Place in Nature (London: Kegan Paul, Trench, Trubner & Co. Ltd., 1925); (London: Routledge and Kegan Paul, 1951; New York: The Humanities Press, 1951). Perception, Physics, and Reality; An Enquiry into the Information that Physical Science Can Supply about the Real (Cambridge: Cambridge University Press, 1914). Religion, Philosophy and Psychical Research (London: Routledge and Kegan Paul, 1953; New York: Harcourt, Brace and Company, 1953). Scientific Thought (London: Kegan Paul, Trench, Trubner and Co., 1923). bibl i ography Schilpp, P.A., ed.: The Philosophy of C.D. Broad (New York: Tudor Publishing Company, 1959). This volume includes Broad’s reply to critics and a complete bibliography of his writings through July 1959. douglas c. long

bundle theory The view that an individual thing is nothing more than a bundle of properties. It is opposed to the view that an individual thing is a substance or substratum. Berkeley voices preference for a bundle theory over a substance theory (at least in the case of unthinking things) in the following passage: In this proposition “a die is hard, extended, and square,” [some] will have it that the word “die” denotes a subject or substance

161

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 162

b u n d l e theo r y distinct from the hardness, extension, and figure which are predicated of it, and in which they exist. This I cannot comprehend; to me a die seems to be nothing distinct from those things which are termed its modes or accidents. (Principles of Human Knowledge, para. 49) Bundle theories are often motivated by the fear that a substance would be (in Locke’s phrase) “something I know not what”, or worse yet, a bare something, devoid of features (see bare particular). The fear is misplaced, however, since from the fact that a substance is something distinct from its properties, it does not follow that it does not have any properties; nor does it follow that its nature cannot be known. In the discussion that follows, it will be assumed that a bundle of properties is a set of properties, but what is said should hold equally well if a bundle is any other sort of complex entity (e.g., a whole) of which properties are the sole constituents. If a thing were really nothing more than a set of properties, then any set of properties would constitute a thing. That is absurd – there is no individual constituted by the set of properties (being an alligator, being purple). To avoid this objection, sophisticated bundle theorists (such as Russell and Goodman) typically say that a thing is not just any set of properties, but a set of properties united by the relation of co-instantiation. Intuitively speaking, co-instantiation is the relation that holds among a number of properties just in case they are all possessed by the same individual. For purposes of the bundle theory, however, it must be assumed that co-instantiation is a relation relating properties alone, not a relation relating properties to an already constituted individual. We are to explain individuals in terms of co-instantiation rather than vice versa. The version of the bundle theory that identifies things with bundles of coinstantiated properties is open to two notable objections. First, if a thing were a set of properties, how could anything ever change its properties? For a thing to have a property, according to the bundle theory, is for that property to be a member of it. 162

So a thing could change its properties only if the set identical with it could change its members. But that is impossible; a set is defined by its members. Bundle theorists might seek to avoid this objection by identifying a thing with a sequence of sets of properties – for example, the sequence containing FGH on Monday, FGK on Tuesday, and so on. They could then say that a thing changes its properties by having different properties as members of successive elements in the sequence. Whether this is an adequate account of change or not, it invites the objection that a thing’s entire career is now made essential to it. The individual identical with the sequence above changes from having H to having K; but since sequences are defined by their elements just as much as sets by their members, the individual could have had no other history than FGH followed by FGK. The second objection to the bundle theory is that it implies a dubious version of Leibniz’s principle of the identity of indiscernibles. By the standard principle of individuation for sets, a set x and a set y are distinct if, and only if, one has a member that the other lacks. If individuals are sets of properties, it follows that two individuals are distinct only if one has a property that the other lacks. But is it not conceivable that there could be two individuals that were perfectly alike in all their properties – the same in color, shape, mass, and so on – yet distinct for all that? There is no guarantee that individuals can always be differentiated by their properties unless we have recourse to impure properties – properties such as being identical with Socrates or being six feet to the north of Plymouth Rock. Such properties presuppose already constituted individuals and so could not be the ultimate materials from which individuals are assembled. One can imagine a third version of the bundle theory that escapes the objections so far mentioned. This version would be analogous to linguistic versions of phenomenalism, which decline to identify material objects with systems of sense data, but maintain none the less that material-object discourse is translatable into sense-datum

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 163

bundl e t heory discourse (see sensa). Similarly, the third version of the bundle theory would refuse to identify an individual with any set of properties, but would offer instead to translate any statement about individuals into a statement exclusively about properties. For example, it might translate “There is a red, round thing here” as “Redness and roundness are here co-instantiated”; but it would not identify the red, round thing with the complex of properties at the place in question, or indeed with anything at all. The sentence “There is a red, round thing here” would get counted true as a whole sentence, even though the phrase “red, round thing” lacks a referent. Note how the third version of the bundle theory avoids our two objections. By refusing to identify things with items (such as sets) that are defined by their constituents, it avoids the objection about change. It also accommodates the possibility of a world in which (intuitively speaking) two things are exactly alike, since it admits the possibility of a world in which the same total set of properties is co-instantiated twice over. Since the co-instantiated sets are not allowed to crystallize into things, there is no question about what makes the two things two. The third version of the bundle theory is not without its costs. It avoids the objections to previous versions by refusing to find within one’s ontology any elements or complexes of elements with which individuals may be identified. But if individuals are not identical with anything, then strictly speaking, they do not exist. Any individual who wishes to believe in his or her own existence must therefore reject the third form of the bundle theory. Another philosophical position that goes by the name “bundle theory” is the view, most famously espoused by Hume, that a self is nothing but a bundle of thoughts and experiences. This view bears obvious analogies to the more general bundle theory discussed above, and it must deal with similar objections – for example, how to allow for the logical possibility that two selves might have exactly the same thoughts and experiences. It, too, is more plausibly developed in a translational than in an identificational

direction, and it may take in its stride (or, as in Buddhist philosophy, gladly embrace) the consequence that there are no selves. But it faces stiff challenges. For example, how are we to translate negative judgments, such as “I am not now feeling pain”, into an idiom free of reference to the I? The passive construction “Pain is not now felt” is too sweeping – perhaps pain is felt by someone, even if not by the speaker. “Pain is not now felt here” similarly oversteps what is known to be the case – however small here is, perhaps some tiny creature shares that space with the speaker and feels pain. “This sensation is now occurring and is not coinstantiated with pain” is a good try, but still fails to be equivalent with the original; it implies, as “I am not in pain” does not, that a particular sensation is occurring. Perhaps the best strategy for bundle theorists would be to try to steer a course between the second two versions of the theory considered here. This would involve holding that individuals are entities in their own right that come into being when certain sets of properties are co-instantiated. They are not identical with sets of properties (as in version 2), nor is talk of them merely a way of speaking about the patterns of co-instantiation among properties (as in version 3). Instead, individuals are ontological emergents; they emerge from bundles of properties, but are not identical with them. bi bliography Castañeda, H.-N.: “Thinking and the Structure of the World,” Critica 6 (1972), 43–81. Chisholm, R.M.: “On the Observability of the Self,” Philosophy and Phenomenological Research 30 (1969), 7–21. Goodman, N.: The Structure of Appearance, 2nd edn. (Indianapolis, IN: Bobbs-Merrill, 1966), 200–11. Loux, M.: Substance and Attribute (Dordrecht: Reidel, 1978), 115–39. Russell, B.: An Inquiry into Meaning and Truth (London: George Allen & Unwin, 1940); (Baltimore, MD: Penguin Books, 1967), 89–101, 121–3. 163

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 164

b u r i d a n, jean Van Cleve, J.: “Three Versions of the Bundle Theory,” Philosophical Studies 47 (1985), 95–107. james van cleve

Buridan, Jean (c.1295–1358) Philosopher and scientist, born at Béthune, France. Buridan studied at the University of Paris under Ockham and also taught there, serving as rector in 1328 and 1340. He also served as ambassador for the university at the papal court in 1345. Buridan’s main philosophical works are the Summulae de dialectica (1487) and various commentaries on works of Aristotle. Buridan’s overall philosophical tendency was nominalistic (see nominalism) and skeptical. He is best known for his work in logic and his doctrine of free will (see the extended essay). In logic he developed theories of the modality of propositions (see proposition, state of affairs) and the syllogism, and he appears to have been the first to provide a deductive derivation of the laws of deduction. The means he developed to find the middle term of a syllogism came to be known as “the bridge of asses” (pons asinorum), because it allowed dull students to pass from the premises to the conclusion of a syllogism. Buridan’s doctrine of free will may be characterized as a form of intellectual determinism. The will chooses what reason represents to it as best, although there is no particular time frame within which the will must choose. This view has the extraordinary consequence that if reason presents two alternative choices as equally good, the will cannot make a choice. This difficulty is usually illustrated by what has come to be called “Buridan’s ass”. According to this example, a hungry donkey given equal bales of hay would starve to death; since its intellect could not represent one as better than the other, its will could not make a choice. It is not known who used this example first, but there are some antecedents in Ghazali (1058–1111) and Aristotle. Buridan himself speaks of a dog starving when confronted with equal portions of food. 164

wri t i ngs Consequentiae, in Iohannis Buridani Tractatus de consequentiis ed. H. Hubien (Louvain: Publications Universitaires de Louvain, 1976). In Metaphysicam Aristotelis quaestiones (Paris, 1518); (Frankfurt am Main: Minerva, 1964). Perutile compendium totius logicae Joannis Buridani cum praeclarissima solertissimi viri Joannis Dorp expositione (Venice: 1499); (Frankfurt/Main: Minerva, 1965). Quaestiones on Aristotle’s Ethics (Paris, 1498), Physics (Paris, 1509), De anima and Parva naturalia (Paris, 1516), and Politics (Paris, 1530). Sophismata, in T.K. Scott, Johannes Buridanus: Sophismata (Stuttgart-Bad Cannstatt: Frommann-Holzboog, 1977). Summulae de dialectica (Paris, 1487). bibl i ography Hughes, G.E.: John Buridan on Self-Reference: Chapter Eight of Buridan’s “Sophismata” (Cambridge and New York: Cambridge University Press, 1982). Pinborg, J., ed.: The Logic of John Buridan (Copenhagen: Museum Tusculanum, 1976). jorge j.e. gracia Butler, Joseph (1692–1752) British moral philosopher and natural theologian, best known today for the ethical theories in his Sermons (1726), but is also noteworthy for the substantial metaphysical treatise, The Analogy of Religion (1736). This is the most influential work in the tradition of empirical, or experimental, theism. Butler’s metaphysical arguments, like his ethical ones, have a deeply practical motive. He seeks to persuade his readers to turn to a Christian way of life. To do this he combats the most fashionable antiChristian arguments of his day. These were offered not by atheists, but by deists, who accepted philosophical demonstrations of the existence and governance of God (particularly the argument from design) but rejected belief in divine intervention and the

9780631199991_4_P2002.qxd 1/12/09 3:04 PM Page 165

butler, j oseph claims of revelation, holding that a rational deity could sustain and guide us without such special devices. Butler assumes at the outset that God exists and governs the world, since his opponents also did. In the first part of the Analogy he argues that someone who accepts divine governance can find evidence in nature that we live in what he calls a “state of probation”: that is, that we inhabit a world in which we are given the opportunity (and obligation) to choose a path toward moral maturity that will fit us for entry into another life. In the second part he argues that if this is accepted, there is no good reason to turn aside from the signs of revelation that such a providential scheme would give us reason to expect. Butler’s arguments are inductive, or, in his language, analogical, not demonstrative. He begins by making a case for the reality of an afterlife (one that does not require prior agreement on the existence of God). He stresses the frequency of radical transformation in nature (e.g., from caterpillars to butterflies), and argues that death may consist, not in the destruction of persons and the powers they have, but the mere destruction of those means the body provides for the exercise of those powers. Hence death may, for all we know, entail transformation rather than destruction. He then invokes the theistic teleology that he shares with his opponents, and argues that just as we are able, within this life, to learn lessons from both the good and the bad experiences of youth, that fit us for adult life, so the experiences of earthly life as a whole may be intended by God to prepare us for another. He therefore draws an analogy between the earlier and later stages of this life on the one hand, and the pre-mortem and post-mortem lives of human agents on the other. He insists at length that the order we discern in this life is a moral one, in which our creator “declares for virtue”; the moral structure of human nature, described in the Sermons, is part of the evidence for this. Once the probationary character of human life is accepted, Butler maintains, it is foolish to ignore the likelihood that God will have made revelatory signs available to

us. Given his prior commitment to divine governance, the deist has no good reason to hold he would not have done so. Butler’s case is throughout prudential as well as inductive: that even if the probability of life’s being probationary is not as high as he holds, as long as there is some reasonable degree of likelihood of this, it is foolish to ignore the demands of virtue, or reject the claims of revelation without careful examination. While many of Butler’s detailed arguments are shrewd and ingenious, his metaphysical claims depend in large measure on assuming the prior proof of God’s existence. The Analogy is still of great value, but more as a mine of apologetic defenses than as a body of metaphysical argument. In an era where most apologists, wisely or not, seek to defend Christianity without natural theology, the probability of life being a state of probation, and the probability of the Christian revelation being true, have to be judged together rather than in sequence. Butler remains the finest classical advocate of the view that, given the depth of our ignorance of divine purposes, probabilities are all most of us have available, and that our religious decisions should therefore be taken with prudence. wri t i ngs The Analogy of Religion, Natural and Revealed, to the Constitution and Course of Nature (1736); Vol. II of The Works of Bishop Butler, ed. J.H. Bernard (London: Macmillan, 1900). bibl i ography Broad, C.D.: “Bishop Butler as Theologian,” in his Religion, Philosophy and Psychical Research (London: Routledge and Kegan Paul, 1953), 202–19. Jeffner, A.: Butler and Hume on Religion (Stockholm: Diakonistyrelsens Bokforlag, 1966). Mossner, E.C.: Bishop Butler and the Age of Reason (New York: B. Blom, 1971). Penelhum, T.: Butler (London: Routledge and Kegan Paul, 1985). terence penelhum 165

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 166

Cantor, Georg (1845–1918) One of a group of late nineteenth-century mathematicians and philosophers (with Frege, Dedekind (1831–1916), Peano (1858– 1932), Hilbert (1862–1943) and Russell) who, in their different ways, transformed both mathematics and the study of its philosophical foundations. The philosophical import of Cantor’s work is threefold. First, it was primarily Cantor who turned arbitrary collections into objects of mathematical study, sets (see class, collection, set), thereby reshaping the conceptual structure of basic mathematics. Second, and in connection with this, he created a coherent mathematical theory of the infinite, in particular a theory of transfinite numbers. Third, he was the first to indicate that it might be possible to present mathematics as nothing but the theory of sets, or at least to push in this direction, among other things making set theory (thus, in fact, the theory of the infinite) the study of the basis on which mathematics is founded. This has had a profound effect on the philosophy of mathematics, not least because it contributes substantially to the view that the foundations of mathematics should itself become an object of mathematical study, and because it emphasizes that classical mathematics involves infinity in an essential and variegated way. Indeed, Cantor’s work renders study of such philosophically important matters as continua and the infinite vacuous without some knowledge of the mathematical developments he wrought (see finite/infinite). Cantor’s main, direct achievement is his theory of transfinite numbers and infinity during 1880–95. This necessitates extensive revision of traditional doctrines going back, via the scholastics, to Aristotle, which reject the actual-infinite as a subject of 166

rational treatment. Cantor introduced a division of the actual-infinite into the transfinite (or increasable) infinite, and the absolute infinite. According to Cantor, only the latter is beyond rational (thus mathematical) treatment. He argued convincingly that the transfinite is in fact implicitly present in ordinary mathematics, which should therefore take the infinite seriously. Like Frege, Cantor characterized sameness of size (cardinal equivalence) in terms of one-to-one onto correspondence, thus accepted the various paradoxical results known to Galileo and others (e.g., that the collection of all natural numbers has the same cardinality as that of all even numbers). He added to the stock of these surprising results by showing in 1874 that there are only as many algebraic (and thus rational) numbers as there are natural numbers, and in 1878 that there are more points on a line than there are natural (or rational or algebraic) numbers, thus showing for the first time that there are at least two different kinds of infinity present in ordinary mathematics, so exposing the need for a coherent theory of these infinities. Cantor’s theorem of 1892 (that the set of all subsets, the power set, of a given set must be cardinally greater than that set) goes further, for it shows that ordinary mathematics must accept indefinitely many different kinds of infinity. Cantor’s work in mathematical analysis in the period 1878–82 also showed a pressing need for an extension into the infinite of the indexing (counting) function of the natural numbers. Cantor identified the fundamental property of the natural numbers as counting numbers (their discreteness) in their being well-ordered, thus that, in addition to being linearly ordered, there is a first element, and every element with a successor has a unique successor (see

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 167

carnap, rudol f continuous/discrete). There being nothing to restrict well-ordering to finite collections, Cantor introduced a scale of general ordinal or counting numbers (the first infinite number was eventually called ω by Cantor) to reflect well-orderings in general. Cantor’s radical idea for a theory of infinite size was to base the notion of cardinality on that of ordinality. Although cardinal and ordinal number coincide in the finite case, this is false of infinite collections. Cantor shows how to circumvent this obstacle by dividing the ordinal numbers into the number classes, each consisting of those ordinals representing sets of the same size. Cantor then introduced a scale of cardinal numbers of well-ordered sets (the ℵ–numbers) standing for these classes. If one assumes, not only that all sets have a size, but that all sets can be well-ordered (as Cantor did, and as has later been accepted by modern set theory, not without controversy), then all infinite sizes are represented in the scale of alephs. The continuum problem, of enormous importance in the development of Cantor’s work, is the problem of which aleph represents the cardinality of the continuum. Cantor’s famous continuum hypothesis (CH) is the conjecture that this cardinality is ℵ1 the second infinite aleph, represented by Cantor’s second number-class. The continuum problem was the first in Hilbert’s list of 24 central mathematical problems in a celebrated address in 1900, and is now widely thought to be insoluble. CH was shown to be independent of the standard axioms of modern set theory in two steps, the first due to Gödel (the consistency, 1938) and the second Paul Cohen (1934– ) (the consistency of the negation, 1964). Moreover, work stemming from that of Cohen shows that it is consistent to assume that the cardinality of the continuum can be represented by almost any of the vast sequence of ℵ-numbers. Cantor’s conception of set has given rise to some dispute. It is often thought that it is wide enough to admit the universe of sets itself as a set, thus giving rise to what has become known as Cantor’s paradox. If the universe were a set, Cantor’s theorem would say that its power set must be larger than it; but since this latter is a set of sets, it

must be contained in the universal set, and thus be smaller. Cantor’s statements about the nature of sets are too vague to allow of decisive judgment, but it seems to follow from his earlier considerations of the absolute infinite that the collections involved in the paradoxes cannot be proper sets. Other indications also point in this direction (see Hallett, 1984). Moreover, correspondence with Hilbert in the late 1890s and Dedekind in 1899 (see Cantor, 1991) shows clearly that Cantor was well aware that contradictions will arise if such collections are treated as ordinary sets. Indeed, in this correspondence Cantor suggested explicitly that consistency be taken as a criterion of set existence, thus presaging a doctrine central to Hilbert’s work. wri t i ngs Briefe, ed. H. Meschkowski and W. Nilson (Berlin, Heidelberg, and New York: Springer-Verlag, 1991). Gesammelte Abhandlungen mathematischen und philosophischen Inhalts (Berlin: SpringerVerlag, 1932; Berlin, Heidelberg, and New York: Springer-Verlag, 1980). bibl i ography Hallett, M.: Cantorian Set Theory and Limitation of Size (Oxford: Clarendon Press. 1984). Purkert, W. and Ilgauds. H.J.: Georg Cantor (Basel: Birkhäuser, 1987). michael hallett Carnap, Rudolf (1891–1970) A leading German logical positivist (see logical positivism). At university. Carnap studied both physics and philosophy. From 1925–36, he was an important participant in the Vienna Circle. In 1936 Carnap emigrated to the United States, where his work in philosophy of science, philosophy of language, modal logic and inductive logic shaped and promoted the absorption of logical positivist ideas into the American philosophical mainstream. In his autobiography (in Schilpp, 1963, p. 45), Carnap laments the vague, inconclusive character of traditional metaphysics: “most of the controversies in traditional metaphysics appeared to me sterile 167

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 168

c a r n a p , r udo lf and useless . . . I was depressed by disputations in which the opponents talked at cross purposes; there seemed hardly any change of mutual understanding, let alone of agreement, because there was not even a common criterion for deciding the controversy.” This anti-metaphysical animus informs Carnap’s two most important books, Der logische Aufbau der Welt (1928) and Logische Syntax der Sprache (1934). In both these works Carnap proposes to replace philosophy with a successor discipline devoted to the application of modern logic to the clarification of the concepts of science. In Der logische Aufbau der Welt Carnap advocates the development of constitution systems as the successor to philosophy. A constitution system is an ordered system or definitions of scientific concepts. Constitution systems are to be comprehensive, embracing all the concepts of all the formal and empirical sciences. By “definition”, Carnap means explicit definition. Carnap assumes, in effect, a version of the simple theory of types as the background language and logic for constitution systems; and he takes Whitehead and Russell to have demonstrated that mathematics can be unproblematically developed in this framework. Each constitution system has a basis: a domain of individuals and primitive relations over the domain. Carnap believes that there are alternative bases for constitution systems, and hence distinct, though equivalent, systems; he holds forth the prospect of systems with a physical basis of fundamental particles and fundamental magnitudes. However, to illustrate constitution systems, Carnap sketches the development of an epistemologically oriented system with an autopsychological basis: the individuals are the total momentary experiences of a person; the single undefined relation is that of recollected similarity. The order of definition in this system is to reflect epistemological priority in the application of concepts. A constitution system contains definitions for the concepts employed by the existing formal and empirical sciences. Carnap relies on this feature of these systems to motivate his claim that any rational statement can be

168

formulated within a constructional system. Statements drawing on concepts not so definable are non-rational. Carnap believes his autopsychological constitution system expresses the core common to various epistemological positions, capturing the insights each emphasizes. This achievement makes evident that these epistemological positions differ only in their non-rational, metaphysical assertions. So, Carnap holds that his sample system captures the Kantian insight that objective knowledge requires the synthesis of something given to the form of the unity of the object (see Kant). However, for Carnap, synthesis is not understood in terms of an ordering that a transcendental subject imposes on a given manifold in accordance with the immutably valid forms of thought, as neo-Kantians held (see transcendental ego). Instead, type theory replaces transcendental logic; formal definitions replace synthesis. Talk of transcendental subjects as well as unconceptualizable thingsin-themselves disappears (see noumenal/ phenomenal). Similarly, in defining scientific concepts from an autopsychological basis, this constitution system captures the Machian empiricist insight that all empirical knowledge arises from experience, that every scientific statement is reducible to an equivalent one concerning elementary experiences (see empiricism; Mach). But this reducibility does not in any way ontologically privilege the elementary experiences over objects defined at later stages of the system. In Logische Syntax der Sprache Carnap urges that philosophy be replaced by the logic of science, by the logical syntax of languages for science. Central to Carnap’s view of logical syntax is his notion of a language or linguistic framework, a notion that, with modifications, Carnap held for the rest of his career. Languages are to be described in purely formal terms via formation rules defining sentencehood and transformation rules defining a consequence relation for the language. The logical pluralism voiced in Carnap’s Principle of Tolerance gives logical syntax its significance: “In logic, there are no morals. Everyone is at liberty to build up his own logic . . . All that is required of him is that

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 169

c ast añeda, hec t or- neri . . . he must state his methods clearly, and give syntactical rules instead of philosophical arguments” (Carnap, 1934, p. 52). In denying that there is a right or wrong in logic, Carnap rejects any framework transcendent notion of fact or truth. He accordingly comes to distinguish sharply between the decision to adopt a linguistic framework and the epistemic evaluation of sentences of a particular framework. The former is a matter of practical decision, ultimately of preference. The latter evaluations are constrained by the defining rules of a particular language. Carnap adopts the Principle of Tolerance in response to disputes in the foundations of mathematics, disputes that struck him as sterile as traditional philosophical debates. In Logische Syntax der Sprache Carnap explicates analyticity in terms of consequence. He exhibits specifications of languages of differing logical strengths all of whose mathematical truths are analytic. Carnap thus seeks to persuade us that mathematics flows from the defining rules of a language. He maintains that foundational debates arise from the adoption of different languages. Tolerance counsels that these quarrels should be replaced by the metamathematical investigation of languages formalizing various foundational approaches. It should be noted, however, that the mathematics required in Carnap’s metalanguage for the crucial definition of consequence makes his understanding of the analyticity of mathematics vulnerable to the charge of vicious circularity. Carnap applies logical syntax to the problem of explicating empirical testability (Carnap, 1936, 1937). For Carnap, empiricism is not a thesis; it is rather a recommendation that investigators restrict themselves to certain languages as the frameworks for formalizing scientific theories. Here again we see Carnap’s distinctive approach to philosophy: the explication (that is, the replacement) of a hitherto philosophical notion by a precise, formal notion and the corresponding construal of a philosophical thesis as the recommendation of a linguistic framework.

writings Der logische Aufbau der Welt (Berlin: WeltkreisVerlag. 1928); trans. R. George, The Logical Structure of the World (Berkeley: University of California Press. 1969). Foundations of Logic and Mathematics, in Foundations of the Unity of Science, vol. 1, ed. O. Neurath et al. (Chicago: University of Chicago Press. 1939), 139–213. Logische Syntax der Sprache (Vienna: J. Springer, 1934); trans. A. Smeaton, The Logical Syntax of Language (London: Routledge and Kegan Paul, 1937). Meaning and Necessity (Chicago: University of Chicago Press, 1947). “Testability and Meaning,” Philosophy of Science 3 (1936), 419–71; 4 (1937), 1–40. bi bliography Schilpp, P.: The Philosophy of Rudolf Carnap (La Salle, IL: Open Court, 1963). thomas ricketts Castañeda, Hector-Neri (1924–91) One of America’s leading analytical metaphysicians. His principal contribution in this area is guise theory, first expounded in Castañeda (1974) and subsequently developed and refined in numerous essays leading to his (1989) volume. Guise theory is at once a complex and global view of language, mind, ontology, and predication. To launch guise theory, Castañeda directs us to triads such as the following: (1)

Before the pestilence Oedipus believed the previous King of Thebes was dead. (2) It is false that before the pestilence Oedipus believed that Antigone’s paternal grandfather was dead. (3) Antigone’s paternal grandfather was the same as the previous King of Thebes. Evidently (1)–(3) are all true (and hence mutually consistent). But, Castañeda queries, how can this be so, given that tenets (T1)– (T3) are each theoretically plausible? (T1) For any individuals x and y, if x is (genuinely or strictly) identical with y,

169

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 170

c a s t a ñ eda, hecto r -ner i then, for any property P, x has P if, and only if, y has P. (T2) The sameness relation expressed by statement (3) is genuine identity. (T3) In so far as a statement like (1) says something of or about the previous King of Thebes, the sentential matrix “Before the pestilence Oedipus believed was dead” predicates a property of the denotation (if any) of the singular term filling the blank. As a closer inspection reveals, however, it is quite impossible that all of (1)–(3) and (T1)–(T3) are true. For (1) and (3), taken in conjunction with the trio (T1)–(T3), entail: (4) Before the pestilence Oedipus believed that Antigone’s paternal grandfather was dead. And (4) flatly contradicts (2). In guise theory the above puzzle is resolved as follows. Holding fast to the truth of (1)–(3), Castañeda also accepts (T1) and (T3), dictating thereby a rejection of (T2). Theorizing that the sameness relation expressed by statement (3) is weaker than strict identity – he dubs the relation in question consubstantiation – guise theory finds that the previous King of Thebes and Antigone’s paternal grandfather are genuinely different individuals: Castañeda’s guises. As he puts it, the network of triads such as (1)–(3) “is like a huge prism: it breaks down ordinary objects into a system of (infinitely) many guises” (1978, p. 195). Owing to this theoretical stance toward (1)–(3), notice, Castañeda’s ontological prism proves unrelenting: since there are perfectly analogous triads as regards, say properties and propositions, each of these join ordinary concrete individuals in separating into their component guises, where the latter are one and all genuinely different items in the ontological inventory. According to Castañeda, guises are complex entities generated as follows. Given any number of properties, say {F1 . . . Fn}, the set forming operator { . . . } generates the set {F1 . . . Fn}. Next, for any such set, the concretizer operator, c, generates the guise c {F1 . . . Fn}. Guises are adjudged identical exactly on the condition that their cores 170

have exactly the same members, where the core of c {F1 . . . Fn} is the set {F1. . . . Fn}. For every set of properties, observe, there is a corresponding guise, and each guise enjoys a bona fide ontological status. Because some of these individuals actually exist, e.g., the current president of the United States, whereas others do not. e.g., the round square, the ontology and semantics of guise theory is Meinogian (see Meinong). In addition to strict identity and consubstantiation, guise theory includes at least these additional sameness relations: conflation, consociation, transsubstantation, and transconsociation. Properties are predicated of guises either internally or externally, where P is had internally by a guise g just in case P belongs to g’s core, and P is had externally by g if, and only if. there is a sameness relation R and a guise g having P in its core, and g bears R to g. So fortified, Castañeda documents that guise theory provides a unified account of a wide range of problems concerning reference to non-existents, negative existentials, referential opacity, names and rigid designation, indexicals, and other matters (Castañeda, 1989, pp. 235– 61). writings “Direct Reference, the Semantics of Thinking, and Guise Theory,” in Themes From Kaplan, ed. J. Almog, J. Perry, and H. Wettstein (New York and Oxford: Oxford University Press, 1989). “Philosophic Method and the Theory of Predication and Identity,” Noûs 12 (1978), 189–210. “Reference, Reality, and Perceptual Fields,” Proceedings and Addresses of the American Philosophical Association 53 (1980), 763– 823. “Thinking and the Structure of the World,” Philosophia 4 (1974), 3–40. Thinking, Language, and Experience (Minneapolis: University of Minnesota Press, 1989). bi bliography Tomberlin, J.E., ed.: Agent, Language, and the Structure of the World (Indianapolis, IN: Hackett, 1983).

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 171

c ategories Tomberlin, J.E.: Hector-Neri (Dordrecht: Reidel, 1986).

Castañeda

james e. tomberlin categories The philosophically most useful and reasonably precise notion of categories was introduced by Aristotle, whose Categories is the locus classicus for discussions of the topic. Categories are the most general kinds of things (“thing being used here as applicable to anything whatever), the highest, summa, genera. The ideal theory of categories must satisfy at least two conditions: (1) it must be exhaustive, i.e., every thing must fall under one of the theory’s categories; and (2) its categories must be mutually exclusive, i.e., no thing may fall under more than one category. To know what categories of things there are would thus be to know the most general constitution of reality, which is a defining goal of metaphysics. The summa genera are ordinarily the highest ends of complex hierarchies of subordinate genera, reaching at their other ends the so-called infimae species, i.e., kinds no longer divisible into subordinate kinds. The genus/species relationship (approximated in twentieth century philosophy, with respect to qualities, by the determinable/ determinate distinction, e.g., between color and red) must not be confused with any kind of composition (see determinate/ determinable). For example, human being is a species of animal, but its being an animal is not a characteristic simply added to, say, being two-footed: rather, the latter is a way in which the former is determinate. A real definition states the genus (e.g., animal) under which the species defined (e.g., human being) falls, and the letter’s differentia (e.g., two-footed) distinguishes it from other species falling under that genus. Therefore, a category, being a summum genus, i.e., a genus that falls under no other genus, cannot be given a real definition. Of course, it can be given a nominal definition (an explanation of the usual meaning of the name of what is defined), but a nominal definition, unlike a real definition, may tell us little if anything about the nature of what is defined.

The above is the classical, Aristotelian theory of categories. In his Categories Aristotle lists ten categories: substance (e.g., a horse), quantity (e.g., two yards long), quality (e.g., white), relation (e.g., double), place (e.g., in the Lyceum), time (e.g., yesterday), position (e.g., sitting), possession (e.g., armed), action (e.g., cutting) and being acted upon (e.g., being cut). The Aristotelian theory dominated classical and medieval philosophy. It has been generally rejected in modern philosophy, largely because the Aristotelian theory of substance has been rejected. According to Aristotle, a substance is an individual, particular thing that endures through change. The items in the other categories, called accidents, owe their being solely to their presence in substances. If, as modern philosophers generally have done, one thinks of the items Aristotle regarded as substances as mere bundles of qualities, or volumes of extension, or aggregates of corpuscles, the distinction between substance and accident appears to lose its significance (see aristotle; bundle theory). Hence, there has been little use of the notion of a category in modern philosophy, except in the vague sense of any fundamental or basic class or concept or even word. A major exception to this was Kant, who applied it to what he regarded as the twelve fundamental pure (non-empirical) concepts of the understanding, which correspond to what he took to be the twelve most general forms of judgment and hence make possible our knowledge of objects; for this reason they are called by him transcendental concepts. All judgments are (1) universal or particular or singular; (2) affirmative or negative or “infinite” (e.g., “The soul is non-mortal”); (3) categorical or hypothetical or disjunctive; and (4) problematic or assertoric or apodeictic. The respective pure concepts are (1) unity, plurality, totality; (2) reality, negation, limitation; (3) inherence (of a predicate in a subject), causality, reciprocity (interaction); and (4) possibility, existence, necessity. Kant’s categories are better thought of as epistemological in nature, and thus quite different from Aristotle’s, although the distinction between metaphysics and 171

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 172

c a t e g or ies epistemology does not have a straightforward application to Kant’s philosophy. In the philosophy of language, the notion of a category has been used rather loosely for types of words or uses of words, and the phrase “category mistake” was introduced by Ryle (1949) for statements or views that involve confusion of such types, A non-philosophical example he gave was that of someone’s asking “But where is the University?” after having been shown the various buildings, offices and departments constituting the University. Ryle’s wellknown chief philosophical example was the view that certain psychological words, such as “thinks”, signify the occurrence of nonphysical, mental events, since ex hypothesi they do not signify the occurrence of physical events; while the truth is, according to Ryle, that they do not signify the occurrence of events at all. We shall limit ourselves here to a discussion of the metaphysical topic of categories. Such a discussion may be helped by considerations about knowledge or about language, but must not be confused with them. It is fairly clear that Aristotle’s list of categories is not exhaustive, and that some of them overlap, or are reducible to others, or are not sufficiently basic to qualify as categories. For example, it can be argued that place and time should be understood as reducible to spatial and temporal relations, and then they would belong to the category of relations. The rest of Aristotle’s categories, except that of substance, seem to be subordinate genera of the more general kind that today is called a property and arguably is more properly classified as a category. As a result of this line of reasoning, we reach the most common contemporary list of categories, namely, that of individual things (which may include Aristotelian substances but also momentary particulars such as sense data (see sensa), properties and relations. But in order to avoid paradoxes (which seem to arise if we speak of all properties, thus implying that all properties have a common property, the property of being a property) we may need to accept Russell’s theory of types, according to which a sharp distinction 172

must be made between properties of individual things, properties of such properties, properties of these properties, and so on, each level constituting a distinct category for the precise reason that the items on it can have no common property with the items on any of the others and thus cannot belong with any of them to the same genus and so to the same category. But many contemporary philosophers believe that further categories, not rooted in Aristotle’s list, should be acknowledged. The most familiar example is states of affairs, which if actual are called facts (see fact; proposition, state of affairs). These would be the entities that supposedly correspond to (usually indicative) sentences, somewhat as individual things correspond to (some) proper names and properties correspond to (some) predicates. Another example is sets, e.g., the set containing as its members this article, the Parthenon, and Alpha Centauri. Many believe that sets are needed for understanding the nature of mathematics. Even if we revise the theory of categories in these ways, it faces severe difficulties. First, there is the challenging and often discouraging task of determining with respect to each alleged category whether it is not reducible to some of the other categories. What causes the difficulty is the vagueness of the notion of reduction (see reduction, reductionism). The mere fact that something involves, or even consists of, certain things, does not mean that it is reducible to them in the philosophically interesting sense of no longer needing to be taken into account as a distinct entity. An example is the category of states of affairs. The state of affairs Jones being white seems to consist of nothing more than Jones, the color white, and perhaps the so-called “nexus” of exemplification. But knowing this provides us with no adequate account of what it is for these constituents of the state of affairs to “hang together” and make up a single entity. A related problem is whether an alleged category is in fact not subsumable under another category, presumably as a subordinate genus, even though one of a very high order. Often, attempts at such

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 173

c ategories a subsumption do not cast much light on any metaphysical issue. An example is the category of events. Is it subsumable under the category of individual (particular) things? Indeed, like the latter, events have spatiotemporal location, but the fact remains that they are too fundamentally different from paradigmatic individual things (e.g., a horse) to make illuminating the claim that they belong to the same genus, and thus to the same category. A second difficulty is the possibility of cross-classification. A familiar view is that there are only two categories: the mental and the material. If so, mental properties and mental individuals would be subordinate genera of the mental, and material properties and material individuals would be subordinate genera of the material. But it seems obvious that mental and material properties have something in common, namely, being properties, and therefore cannot belong to different categories, and mutatis mutandis for mental and material individuals. Yet is it not just as obvious that mental properties and mental individuals have something in common, namely, being mental, and therefore cannot be assigned to different categories, and mutatis mutandis for material properties and individuals? A third difficulty is presented by the existence of concepts, and thus possibly of corresponding entities, that do not seem to fit in any of the usual categories. The following would be examples. (1) Existence and identity (whether of individuals or of properties, the latter being either specific, “exact similarity”, or generic, “inexact similarity”, and thus admitting of degree); it is surely at least hasty to classify existence as a property and identity as a relation. (2) Actuality and possibility, understood as characteristics of states of affairs: to classify them as properties would be to ignore the categorial status of states of affairs by allowing that they have something in common with individual things, namely, properties. (3) Simplicity and complexity; these cannot be properties of individual things

if an individual thing consists at least in part of its properties, i.e., is not a bare particular. (4) The logical connectives (negation, disjunction, etc.), exemplification (what is expressed by “is” in its sense of predication, as in sentences of the form “x is F”), and the quantifiers (all. some), the ontological status of all of which has been vigorously defended by Bergmann (1992) and Grossmann (1983). Some of the concepts occasioning this third difficulty may be definable in terms of others, but it is certain that some must be taken as undefined, as primitive. A fourth difficulty, related to the third, was recognized by Aristotle himself and was accorded much attention by the mediaeval philosophers. That is the existence of concepts even more general than those represented by Aristotle’s list of categories and ranging over the things subsumed under the latter; they cannot be Aristotelian categories precisely because they range across Aristotle’s categories. The concepts in question are what the mediaeval philosophers called transcendentals, the most common examples being Being, One, True, and Good. Could it be that they are themselves the categories, i.e., the summa genera, Aristotle’s categories being subordinate genera? This might seem plausible especially in the case of Being, which may be thought of as the genus under which all things must fall. But in the Metaphysics (998b15–28) Aristotle argued that being is not a genus, on the (questionable) grounds that if it were the differentiae dividing it into subordinate genera and species would not be, since it is impossible for a genus to be a predicate of the differentiae of its subordinate genera or its species. This argument was in effect the basis of the doctrine of the transcendentals. Aristotle and the medieval philosophers who were influenced by him held that such transcendental concepts apply to everything, but at most by analogy, not by representing common properties, whether generic or specific. Is there a way out of these difficulties for a traditional theory of categories, whether 173

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 174

c a u s a t io n Aristotelian or modern, a way to allow metaphysics the goal of giving the most general account of the constitution of reality? We may attempt to do so by regarding the concepts that do not fit in the hierarchy of the categories, and are not reducible to any that do, as principles or concepts that allow us to think and speak of reality, which does have a categorial structure, but do not themselves correspond to parts of reality, and thus can be properly called syncategorematic. In this way we could resolve, for example, the difficulties concerning subsumption and cross-classification, since they (indeed, the theory of categories as a whole) have to do with classification and therefore arise out of our relying on the concept of identity (especially generic identity of properties, or similarity.) They concern ultimately questions about what is more like what, and if nothing in reality corresponds to the concept of identity, then its applications cannot be judged as true or false, but at most as normal or idiosyncratic, and may be expected sometimes to be in conflict yet remain equally legitimate. To reach such a conclusion is, of course, to take what has been called a transcendental turn, though one less extreme than Kant’s or that (quite different from Kant’s) taken by some recent linguistic philosophers, and perhaps closer to the mediaeval doctrine of the transcendentals. To take such a turn is to recognize that our conception of reality necessarily involves elements that do not themselves correspond to items in reality. But how exactly it does so requires detailed examination of the role of each of the relevant concepts, which, following the tradition, we can now call transcendental. It is quite insufficient and indeed misleading just to speak vaguely of “our making the world”. The categorial structure of the world is fixed, it is in no way subject to personal whim or cultural or linguistic custom or convention, and the adequacy of our statements, thoughts, and theories can be judged by comparing them with the world, once the primary (paradigmatic) applications of the transcendental concepts, especially that of generic identity (inexact similarity of properties), are them174

selves fixed. And even though these applications are not fixed by corresponding to something in reality, to suppose that they are fixed by personal whim or cultural or linguistic custom or convention would be to ignore their fundamental status as the determinants of our whole conception of reality, and therefore as presupposed by any attempted explanation or justification of them. bibl i ography Aristotle: Categories (many versions). Aristotle: Metaphysics (many versions). Bergmann, G.: New Foundations of Ontology (Minneapolis, MN: University of Minnesota Press, 1992). Brentano, F.: The Theory of Categories (Hamburg: Meiner, 1933); trans. R.M. Chisholm and N. Guterman (The Hague: Martinus Nijhoff, 1981). Butchvarov, P.: Being Qua Being: A Theory of Identity, Existence, and Predication (Bloomington, IN and London: Indiana University Press, 1979). Chisholm, R.M.: On Metaphysics (Minneapolis: University of Minnesota Press, 1989). Grossmann, R.: The Categorial Structure of the World (Bloomington, IN: Indiana University Press, 1983). Kant, I.: Critique of Pure Reason (many versions). Ryle, G.: The Concept of Mind (London: Hutchinson’s University Library, 1949). Whitehead, A.N. and Russell, B.: Principia Mathematica (Cambridge: Cambridge University Press, 1910–13). panayot butchvarov causation causation

see the extended essay on

change An object undergoes a change if, and only if, it possesses a property at one time and does not possess this property at an earlier or later time. The explication of this definition depends on one’s theory of objects, one’s theory of time and one’s theory of properties.

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 175

change Objects can be conceived in one of two ways, as substances or as wholes of temporal parts (see substance; temporal parts, stages). If an object x is a substance, then x is a particular that exists at each time x is said to exist and that exemplifies each property that x is said to have. If an object x is a whole of temporal parts, then x is composed of distinct particulars, each of which exists at one instant only, such that whatever property x is said to have at a certain time is exemplified by the particular (temporal part) that exists at that time. If x is a substance, then “x possesses a property at one time and does not possess this property at an earlier or later time” implies that the particular that possesses the property at one time is identical with the particular that does not possess this property at another time. If x is a whole of temporal parts, then this definition implies that one temporal part of x possesses a certain property F at one time and that another temporal part of x does not possess F at another time. A further explication of our original definition of change is possible if we introduce two different theories of time, the tenseless theory and the tensed theory. According to the tenseless theory, temporal determinations consist only of the relations of earlier, later and simultaneity. On this theory, changes are described by permanently true tenseless sentences of such forms as “The object x possesses (tenseless) F at the time t and does not possess (tenseless) F at the later time t”. The tensed theory has several versions, but one of them holds that in addition to the temporal relations there are temporal properties of pastness, presentness and futurity (even though these are not properties of the ordinary sort, such as redness). The tensed theorist would describe changes by using transiently true tensed sentences of the form “The object x now possesses F-ness but will soon not possess Fness”. A central difference is that the tensed theorist supplements our original definition of change (which explained change solely in terms of temporal relations) by an account in terms of temporal properties. An object changes from being F to not being F if, and only if, the F-ness of x first possesses

presentness and later possesses pastness. On this account, change may be viewed as the acquiring and losing of temporal properties by the states of an object. A second version of the tensed theory of time holds that only what is present possesses properties and that there are no properties of futurity or pastness. Change would be described by such sentences of the form “It is (now) the case that the object x possesses F-ness and it will be the case that x does not possess F-ness”. But proponents of this view have not yet succeeded in offering an adequate semantic and metaphysical analysis of such sentences. For example, what is the semantic content of “it will be the case that”? If this phrase does not ascribe the property of futurity to the obtaining (truth) of a certain state of affairs (proposition), what is its semantic content? A third way to explicate our original definition is in terms of the theory of properties. There are at least two theories of properties, the causal theory and the consistency theory. According to the causal theory, something is a property if, and only if, it bestows upon its possessor a causal power, i.e., the capacity to affect something else or be affected by something else. For example, a ball satisfies the grammatical predicate “is moving” and since the ball’s motion bestows upon the ball the power to impinge upon and move some other thing, this predicate expresses a property of the ball. By contrast, the ball satisfies the grammatical predicate “is being remembered by John” but since being remembered by John does not bestow upon the ball any causal power, this is not a property of the ball. Thus, if the ball is moving at time t1 and is resting at time t2, it undergoes a change, but if it is being remembered by John at t1 but not at t2 it does not (in this respect) undergo a change. According to the consistency theory, something F is a property if, and only if, it is possible to predicate “F” of something consistently. Since it is consistent to predicate “being remembered by John” of the ball, this predicate expresses a property of the ball. But it is not consistent to predicate of the ball or of anything else the predicate 175

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 176

c h i s h olm, r o der ick milto n “has no properties”, since if something satisfied this predicate it would have at least one property, namely, the property of satisfying this predicate. According to this view, the ball does change from t1 to t2 by virtue of the fact that it is being remembered by John at the first time but not at the second time. See also the extended essay on persistence. b i b l i og rap hy Armstrong, D.M.: Universals and Scientific Realism (Cambridge: Cambridge University Press, 1978). Mellor, D.H.: Real Time (Cambridge: Cambridge University Press, 1981). Oaklander, L.N.: “Delmas Lewis on Persons and Responsibility: A Critique,” Philosophy Research Archives 13 (1987–8), 181–7. Smith, Q.: “A New Typology of Temporal and Atemporal Permanence,” Noûs 23 (1989), 307–30. Smith, Q.: “Problems With the New Tenseless Theory of Time,” Philosophical Studies 52 (1987), 77–98. quentin smith Chisholm, Roderick Milton (1916–99) American epistemologist and metaphysician who has been a seminal figure in contemporary philosophy, Chisholm has helped renew interest in metaphysics during the last third of the twentieth century. Raising the art of philosophical analysis to new heights, his Socratic searches for analyses are legendary (see analysis). Chisholm challenged influential antimetaphysical analytical movements such as logical positivism and linguistic philosophy. His brand of analytical metaphysics incorporates the Aristotelian notion that metaphysics is a “first science” which studies fundamental ontological categories, and utilizes techniques and theories of modern logic to construct philosophical analyses and metaphysical theories (see Aristotle; categories; ontology). Chisholm defends an ontology of concreta and abstracta which includes as fundamental 176

entities substances (material objects and persons) and attributes (which have necessary existence) (see concrete/abstract; substance; universals). Thus, he combines two traditional forms of realism: Aristotelianism about substances and Platonism about attributes. Chisholm strives for a parsimonious account of the intuitive data (see simplicity, parsimony): (1) sense data are eliminated in favor of persons and ways of appearing, for example, “I see a blue sense datum” is paraphrased as “I am appeared to bluely” (see adverbial theory; sensa); (2) any material thing which can persist through mereological change is eliminated in favor of a sequence of material objects such that none of them can survive the loss of a part and each of them differs in some part from its predecessor and successor in the sequence (see part/whole); (3) Chisholm intimates that persons are persisting physical substances which do not undergo mereological change (see persons and personal identity); (4) eliminating places and times, he finds a need for boundaries (see boundary; space and time). In two other cases, Chisholm has changed his ontology. In these cases, difficulties appear to arise due to a need for entities which seem to be a hybrid of concreta and abstracta. (a) Typically, events are changes in spatiotemporally located things, yet some such events recur (see event theory). Chisholm held that concrete events are eliminable in favor of repeatable states of affairs, abstracta such as A person’s walking or Jones’s talking (see proposition, state of affairs). He then switched to the view that an event is a state of an individual, where such a state, x’s being F, is a repeatable concretum that exists if, and only if, x is F. (b) Consider a belief, b, a person, S, has about himself and which is expressible in first-person language. Chisholm shifted from the view that b is directed upon an abstract state of affairs entailing S’s non-qualitative haecceity to the view that an abstractum cannot be non-qualitative and b is S’s self-attribution of a qualitative attribute. Accordingly, Chisholm holds that individuals do not exist in possible worlds: all such worlds are qualitative

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 177

c l ass, collec t i on, set abstracta (see the extended essay on modality and possible worlds). Chisholm maintains that abstracta are individuated by their cognitive content, that a self-presenting psychological state is not identical to a physical state (see mental/ physical; the extended essay on the mind/ body problem) and that there are synthetic a priori propositions. The latter two points relate to Chisholm’s important foundationalist epistemological theories. Chisholm also holds that intentionality uniquely characterizes the psychological, and that reference is a function of the psychological (not of linguistic or causal phenomena). Lastly, Chisholm argues that an adequate account of human freedom and power entails causal indeterminism (see the extended essay on free will). Here (and elsewhere) Chisholm makes important contributions to ethical theory. writings The First Person (Minneapolis: University of Minnesota Press, 1981). “Objects and Persons: Revisions and Replies,” Grazer Philosophische Studien 7/8 (1979), 317–88. On Metaphysics (Minneapolis: University of Minnesota Press, 1989). Person and Object: A Metaphysical Study (La Salle, IL: Open Court, 1976). A Realistic Theory of Categories (Cambridge and New York: Cambridge University Press, 1996). Theory of Knowledge (Englewood Cliffs, NJ: Prentice Hall, 1966). gary s. rosenkrantz class, collection, set These three distinct notions are usually taken to represent totalities made up of elements, which are said to belong to them. There are intuitively clear differences between the three, which however are less clear than is apparent at first sight. A class is often thought of as the extension of a property (or concept), the collection of all those things (of whatever realm one is talking about) which have that property

or fall under that concept (see extension/ intension). The elements in the class are thus “unified” by the property whose extension they make up. A collection is intuitively thought of simply as an agglomeration or aggregate of objects not necessarily united by any specific property. In practice, though, it is hard to conceive of any collections that do not have a unifying property, for the very description of apparently amorphous collections specifies a unifying property, even if this only amounts to a simple enumeration. Still, one might say that such properties are accidental, perhaps due more to artefacts of our language than anything else. (We will come back to this.) The notion of set is different again, for the central difference between sets on the one hand and classes and collections on the other is that sets are assumed to be themselves single objects of the same (logical) type as the elements that compose them, this is meant in the sense that they are themselves available for being taken as elements of further sets (or collections or classes). The relation of sets to properties is again often taken to be accidental, and it is assumed that there is a primitive notion of collection prior to any sophisticated consideration of classes. But note that the description of collections, thus the specification of an underlying property, is indispensable once we start considering infinite collections, as mathematics has to. With both classes and sets, the correct identity principle is that of extensionality. This says that any two classes (sets) with the same elements are identical, even if they have been described on the basis of two different properties. The philosophical interest of the theory of sets (founded by Cantor and others) stems from both its mathematical and its logical importance. There are three reasons for this. First, concentration on sets allowed the collection of numbers or points without there being any obvious form, geometric or otherwise, to hold the elements together, and then, despite this, stressed that such a collection be treated as a self-subsistent mathematical object, i.e., as an object as legitimate and justified as numbers or wellknown, even intuitable, functions or forms. 177

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 178

c l a s s , co llectio n, set Second, this reinforced the claim that mathematics is necessarily based on some kind of privileged intuition, particularly spatiotemporal intuition. In particular, it was allowed that the sets so produced might be infinite as well as finite, and that the mathematical properties they can possess will include those of being ordinally and cardinally numerable. In accordance with this, Cantor developed a widely accepted and precise theory of infinity, including a theory of infinite number. Third, the theory of sets (and classes) was intimately connected with the development of logic through the work first of Frege and then of Russell, since it was what they intended as the basic of logicism, the thesis that mathematics ultimately reduces to pure logic. Can sets be completely arbitrary? Or, more precisely, can every class (i.e., the extension of every property) be treated as a set? Both Frege and Russell assumed at first that the answer is yes, as is implicit in Frege’s Basic Law V (Frege, 1893) and Russell’s Principle of Comprehension (CP) (Russell, 1903). This gives some credence to logicism, if one assumes (1) that mathematics operates with the extensions of concepts; and (2) that logic should (among other things) give the basic laws governing the behavior of concepts and their extensions. However, the set-theoretic antinomies show that Basic Law V or CP, when taken as principles about sets, cannot be right; the extensions of some properties cannot be “simple” objects (thus sets) on pain of contradiction. This is shown clearly by Russell’s paradox, discovered independently by Russell (in 1900) and by Zermelo (before 1902). Other more complicated antinomies were discovered before this (e.g., that of the greatest cardinal, Cantor’s paradox), but all involved other assumptions which could be, and were, challenged. Russell’s paradox has the merit of isolating CP as the principle at fault, thus destroying the assumption of a neat connection between classes and sets, and (despite Russell’s later efforts) between logic and sets. The twentieth century has seen the development of a very important axiomatic theory of sets based on a more complicated 178

connection between extensions of properties (classes) and sets, regulated by Zermelo’s axiom of separation or the stronger axiom of replacement. This axiomatization began with Zermelo (1871–1953) in 1908, and was followed by the contributions of Fraenkel (1891–1965), Skolem (1887–1963), von Neumann (1903–1957), Gödel and Bernays (1888–1977) through the 1920s and 1930s. Together with the development of precise logical frameworks, this led to formal systems such as that known as the ZermeloFraenkel system (ZF). This is loosely based either on the so-called “iterative” conception or on some idea of limiting size, the idea that classes can be sets provided they are not “too big” (see Hallett, 1984). Such frameworks preserve the spirit of Cantorian set theory, and (providing one maintains Zermelo’s Axiom of Choice to provide a proof of the well-ordering theorem) enables in particular a faithful representation of the theories of infinite number that Cantor had developed. A general consequence of this development is the discretization of mathematics (see Continuous/discrete). More particularly, the axiomatized theory (together with the recognition of the importance of metamathematical problems) enables a precise formulation to be given to the question of whether set theory can solve certain of its central problems, in particular the continuum problem. Zermelo style set theory is a theory of pure sets, with the more general notion of class left out altogether. But there are theories which introduce classes in something like the original sense alongside sets, thus as a second sort of object in the sense that they are quantified over, but not in the sense that they can be members (of sets or classes) in the way that sets themselves can. (Some classes are sets, but clearly not all can be.) Which predicates of the language give rise to classes? There is a conservative view, expressed in the system known as Gödel-Bernays set/class theory, which says that any predicate which does not contain quantifiers over classes determines a class. This system is essentially no stronger than ZF set theory, and is in effect a convenient way of allowing reference to the properties of ZF sets in the

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 179

clear and dist i nct theory. A more liberal view is that any predicate determines a class, and this system is indeed stronger. (There are others. See Fraenkel, Bar-Hillel and Levy, 1973, ch. 2, sect. 7.) In both systems, full complementation (denied for sets) is restored; the complement of the class determined by the property ψ is the class determined by ψ (the “opposite” of ψ), and the two classes always exhaust the whole domain. The two simplest complementary classes are the empty class (set), the extension of the property “not identical with itself” (or any other contradictory property), and the universal class, the extension of the property “identical with itself”. The restriction of membership, however, is enough to block the derivation of the known paradoxes. Set theories in the tradition of Zermelo allow the extension of as many properties as possible to be sets while staying free of the known contradictions. But there is a more restrictive view going back to Russell (the vicious circle principle, and the Ramified Theory of Types), Poincaré (1854–1912) and Weyl (1885–1955). According to this view, the paradoxes are not due to the assumption that “overly large” classes are sets, but rather to use of impredicative specifications to pick out sets. (As a first approximation, the specification of a set is impredicative if it contains quantifiers whose range includes the object being specified.) This position faces difficulties. It has to explain why it is wrong to accept as sets collections which, when so taken, do not give rise to paradoxes like Russell’s, collections such as the classical continuum. Second, if important sets such as these are to be excluded, then a replacement for the classical theory of real numbers and real functions has to be developed. However, the difficulty is to arrive at a natural predicative system which achieves this (see Beeson, 1985). b i b l i og rap hy Beeson, M.J.: Foundations of Constructive Mathematics (Berlin: Springer-Verlag, 1985). Fraenkel, A., Bar-Hillel, Y., and Levy, A.: Foundations of Set Theory (Amsterdam: North-Holland, 1973).

Frege, G.: Grundgesetze der Arithmetik, vol. I (Jena, 1893); (Hildesheim: Olms, 1966). Hallett, M.: Cantorian Set Theory and Limitation of Size (Oxford: Clarendon Press, 1984). Russell, B.A.W.: The Principles of Mathematics (Cambridge, 1903); 2nd edn. (London: George Allen and Unwin, 1937). mictael tallett clear and distinct A central concept in Cartesian epistemology, where it provides a criterion of truth (“whatever we perceive very clearly and distinctly is true” – cf. Œuvres de Descartes, vol. VI, p. 33, vol. VII, p. 35) and a rule for making judgments (“include nothing more in your judgments than what presents itself to your mind so clearly and distinctly that you have no occasion to doubt it” – cf. ibid., vol. VI, p. 18, vol. VII, p. 59). Descartes concedes that it is difficult (particularly in metaphysics) to distinguish clear and distinct perceptions or ideas from those which are not clear and distinct (ibid. vol. VII, p. 157, vol. VIIIB, p. 352). He offers no formal definition of these notions until the Principles of Philosophy (1644) (vol. I, p. 45), where his account, though widely cited, is unhelpful. Descartes believes the distinction is best explained by examples (Œuvres, vol. VII, p. 164), and uses the wax example (ibid., vol. VII, pp. 30–1) for this purpose. Our perception of this body becomes clear and distinct when we eliminate from it whatever we are not compelled to ascribe to the body, namely, everything except extension, flexibility and changeability. Similarly, our conception of the mind becomes clearer and more distinct when we recognize that we cannot but ascribe thought to it, but can deny it sensation (when sensation is conceived as involving the body). Systematic doubt provides a technique for achieving clarity and distinctness. These concepts do not, for Descartes, imply adequacy. We can form a concept of God which is clearer and more distinct than any other idea we have. When we conceive him as supremely perfect, this enables us to identify many properties we must, and 179

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 180

c on c e p t many properties we must not, ascribe to him; but though we do have that much knowledge of his essence, he has countless attributes we cannot grasp (ibid., vol. VII, p. 46). Clarity and distinctness does imply conceptual possibility, and hence (since the existence of God guarantees that there is a power capable of creating whatever we conceive clearly and distinctly), real possibility (cf. ibid., vol. VII, pp. 71, 78). b i b l i og rap hy Descartes, R.: Œuvres de Descartes, 12 vols., ed. C. Adams and P. Tannery (Paris: Leopold Cerf, 1897–1913). edwin curley

concept In the history of philosophy the term “concept” and kindred expressions have been used in a variety of technical senses (e.g., by Aquinas, Kant, Frege). The majority of contemporary philosophers, however, use the term in its central nontechnical sense, which is exhibited in complex gerundive phrases of the form “the concept of being F”. In what follows “concept” will be used in this way. Concepts are intensional entities in the sense that two concepts can apply to exactly the same objects and nevertheless be distinct. For example, the concept of being a triangle is not identical with the concept of being a trilateral. This example shows that concepts are indeed hyperintensional in the sense that they can be distinct even if they necessarily apply to the same objects. Because concepts are hyperintensional, they are ideally suited to serve as the senses (meanings) of predicates. For example, “is a triangle” expresses the concept of being a triangle; “is a trilateral” expresses the concept of being a trilateral. Since these concepts are not identical, we have a neat explanation of why the indicated predicates are not exact synonyms. Concepts are a kind of universal, so each of the standard views on the ontological status of universals has been applied to

180

concepts as a special case. nominalism: only particulars (and perhaps collections of particulars) exist; therefore, either concepts do not exist or they are reducible (in the spirit of Carnap) to collections of particulars (including perhaps particulars that are not actual but only possible). conceptualism: concepts exist but are dependent on the mind. Realism: concepts exist independently of the mind. Realism has two main versions: in rebus realism – a concept exists only if it has instances; ante rem realism – a concept can exist even if it has no instances. For example, the concept of being a man weighing over a ton has no instances; however, it is plausible to hold that this concept does exist. After all, this concept would seem to be what is expressed by the predicate “is a man weighing over a ton”. Perhaps the most perplexing question about concepts is how they succeed in being about objects. On one view, there is a primitive, unanalyzable relation of representation that holds between concepts and objects. This view has the disadvantage of making representation an unexplained mystery. A second view is that the relation of representation is analyzable in terms of resemblance, causation, or some other naturalistic notion. While not mysterious, none of these analyses has, thus far, succeeded in avoiding clear-cut counterexamples. A third view is that what is needed is a certain sort of logical theory, specifically, an intensional logic. An intensional logic promises to provide a systematic account of the logical behavior of intensional entities – properties, relations, states of affairs, propositions and concepts (see proposition, state of affairs). The idea is that concepts are logical constructs whose ultimate “constituents” are the real properties and relations of things in the world. A concept is about those objects that have the properties and relations required by the correct logical analysis of the concept. On this approach, the need for a primitive relation of representation thus disappears; at the same time, the easy counterexamples that beset naturalistic analyses (see naturalism) can evidently be avoided.

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 181

c oncret e / abstrac t b i b l i og rap hy Bealer, G.: Quality and Concept (Oxford: Clarendon Press, 1982). Carnap, R.: Meaning and Necessity (Chicago: University of Chicago Press, 1947). Frege, G.: “On Sense and Reference,” Translations of the Philosophical Writings of Gottlob Frege, ed. P. Geach and M. Black (Oxford: Basil Blackwell, 1952). Peacock, C.: A Study of Concepts (Cambridge, MA: MIT Press, 1992). george bealer concrete/abstract Realists and antirealists presuppose an intuitive distinction between abstracta and concreta in their debates about the problem of universals (see nominalism; Platonism). Evidently, every entity is either concrete or abstract, and no entity is both. Plausibly, the division between concreta and abstracta is a basic categorial division; on some views, it is the most basic categorial division. Examples of abstracta are squareness (a property); betweenness (a relation); there being horses (a proposition); the null set; and the number 7 (see class, collection, set; proposition, state of affairs; relations). Examples of concreta are a stone (a material substance); God (a non-physical substance or soul); events such as hurricanes (see event theory); instants and seconds (times); points and expanses of space (places, see space and time); the particular wisdom of Socrates (a trope); the [arbitrary] sum of Earth and Mars; the Earth’s surface (a boundary); and shadows and holes (privations). It is desirable that a philosophical analysis of the concrete/ abstract distinction be ontologically neutral, that is, allow for the possibility of entities of any intelligible sort, given some plausible view about the nature, existence conditions and interrelationships of entities of those sorts. This desideratum seems to require allowing for the possibility of entities of the aforementioned kinds. Ten attempts have been made to analyze the concrete/abstract distinction. (1) Unlike abstracta, concreta are in space.

Observe that for the purposes of (1), “in space” means occupying a place, which is different than standing in spatial relationships. A place stands in spatial relationships, but a place does not occupy a place. If a place occupied a place, then an absurd infinite regress of places would be generated. It might be thought that, trivially, a place occupies itself, but this seems to confuse the relation of Identity with the relation of Occupation. (Note that, arguably, a parallel line of reasoning applies to times, where being in time is occurring at a time (or times) or existing at a time (or times)). We can now see that (1) is inadequate because places are concrete, but are not themselves in space; (1) is also inadequate because although a Cartesian soul or spirit would be a concrete entity, such a being would not be in space. (2) Unlike abstracta, concreta are in space or in time. If absolute time is a necessary being, then (2) avoids the problems pertaining to Cartesian souls and places. For on that condition, necessarily, a Cartesian soul or a place is in time. On the other hand, it might be assumed that, possibly, time is relational. But, on that assumption, (2) does seem to have difficulties with Cartesian souls. Specifically, it appears that if it is possible that time is relational, then there could be a static world containing a Cartesian soul engaged in an atemporal contemplation of necessary truths. Although such an atemporal, non-spatial, thinking substance would be a concrete entity, it would not stand in any spatial or temporal relationship. Moreover, (2) implies that an abstract entity does not exist in time. As we shall see below, this claim is highly problematic; all other things being equal, an analysis of the concrete/abstract distinction that does not rely upon this assertion is better than one which does. (3) Unlike abstracta, concreta are in space– time. There are three objections to this analysis. First, like (1), (3) falsely implies that Cartesian souls are abstract. Second, just as places are not in space, so they are not in, that is,

181

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 182

c on c r e t e / ab str act do not occupy, space–time. Third, (3) incorrectly implies the impossibility of there existing a three-dimensional space and a separate temporal dimension; if such a threedimensional space is possible, then various concrete entities could be in space but not in space–time. (4) Unlike abstracta, concreta stand in spatial and temporal relationships. There are two difficulties with (4). First, although a Cartesian soul does not stand in spatial relationships, a Cartesian soul is a concrete entity. Second, although a place does stand in spatial and temporal relationships, a time does not stand in spatial relationships. Thus, (4) fails to provide a logically necessary condition for concreteness. (5) Unlike abstracta, concreta stand in spatial or temporal relationships. (5) is not subject to the earlier difficulties with places and times, since places stand in spatial relationships, and times stand in temporal relationships. But the problem with Cartesian souls that affects (2) also affects (5). In addition, some abstract entities stand in temporal relationships. For instance, at midnight there are various people who exemplify Sleepiness, a sharable property. And at 3 p.m. a certain metaphysician believes that the null set is peculiar, in which case, at 3 p.m. the null set is believed by that metaphysician to be peculiar. Moreover, if abstract entities exist in time, then they enter into temporal relationships, for instance, the relationship of existing at the same time as George Washington. Another reason to think that abstract entities stand in temporal relationships is that properties undergo relational change. Consider, for example, an abstract entity such as Sleepiness. Clearly, this property cannot undergo intrinsic change, unlike, for instance, Plato, who intrinsically changes when he awakens. But something can stand in temporal relationships by undergoing relational change, even if the thing in question does not undergo intrinsic change. So, something’s being immutable is compatible with its undergoing relational change. Necessarily, whatever changes its relationships to other entities stands in temporal 182

relationships. For instance, a fundamental particle which never undergoes intrinsic change, and which is circled by other particles over some interval of time, stands in temporal relationships to these circling particles. Similarly, insofar as Sleepiness is exemplified by Plato at one time and not at another, this property undergoes relational change. Therefore, Sleepiness does stand in temporal relationships. Thus, (5) has the mistaken implication that Sleepiness is a concrete entity. Furthermore, according to some Aristotelian or neo-Aristotelian theories of universals, an abstract universal has spatial location or locations. For instance, according to such theories, Roundness is located wherever there is a round object. It follows that (5) fails to provide a logically sufficient condition for concreteness. (6) Unlike abstracta, concreta are capable of moving or undergoing intrinsic change. Because a Cartesian soul can undergo intrinsic change, (6) has the desirable implication that a Cartesian soul is a concretum. Nevertheless, (6) is inadequate because points and instants are concrete but incapable of either moving or undergoing intrinsic change. (7) Concreta have contingent existence, whereas abstracta have necessary existence. There are four difficulties affecting (7). First, (7) implies that a necessary God would be abstract, when such a God would be concrete. Second, if space, time, space–time, or mass–energy are necessary beings, then (7) implies that they are abstract, even though they are concrete. So, (7) has the undesirable characteristic of not being neutral about the modal status of such beings. Third, an Aristotelian universal such as Horseness has contingent existence (because its existence depends upon the existence of horses), but a universal of this kind (and every universal) is abstract. Fourth, a set of contingent, concrete beings, for instance, the set of horses, seems to be an abstract entity which has contingent existence. So, (7) fails to provide either a logically necessary

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 183

c oncret e / abstrac t or a logically sufficient condition for concreteness. (8) Unlike concreta, abstracta are exemplifiable. Although it appears that all universals are abstract entities, the problem with (8) is that there are abstract entities of other kinds which are not exemplifiable, for instance, sets. Thus, (8) fails to provide a logically necessary condition for abstractness. (9) Unlike concreta, abstracta are (intellectually) graspable. Although it is plausible that no concrete entity is graspable, (9) is unsatisfactory because it seems that abstracta of certain kinds could not be grasped, e.g., sets of concreta, as well as haecceities of necessarily non-conscious material substances (see haecceity). (10) Unlike abstracta, concreta can be causes or effects. (10) is unsatisfactory for the following reasons. According to one camp, all causes and effects are concrete events (see the extended essay on causation). On this view, (10) has the absurd implication that concreta of other kinds, e.g., substances, are non-concrete. One possible reply is that substances, but not abstract entities, e.g., properties, can be involved in causal relationships. But, if all causes and effects are concrete events, then it is hard to fathom the sense of “involvement” intended. For according to such an event ontology, an event’s occurring does not entail that a substance exists, and an event cannot be identified with a substance’s exemplifying a property at a time or the like. Moreover, since causal relationships hold in virtue of laws correlating properties of things, there is a fairly clear sense in which abstracta are involved in causal relations (see law of nature). Finally, there is some reason to think that abstract facts (see fact) or the like can be causes or effects (cf. Kim 1981). What follows is an attempt to provide an adequate analysis of the concrete/abstract distinction in the face of the foregoing difficulties. The analysis that will be proposed utilizes the intuitive metaphysical concept of ontological categories. It should be

observed that many predicates do not express categories in the relevant sense, e.g., the predicates “red,” “square,” “bachelor,” and (the disjunctive predicate) “substance or surface.” This notion of a category will not be defined here; however, any comprehensive understanding of the world presupposes their use. A crucial part of this proposed analysis is an intuitive conception of a class of categories of a certain level of generality. Intuitively, the categories on the following list, L, appear to be at the same level of generality. (L) Event, Place, Time, Limit, Collection, Trope, Privation, Property, Relation, and Proposition. Other examples of ontological categories which appear to be at this level of generality are Substance and Set. Intuitively, a category’s being at this level of generality is its being at the level of generality of the categories on L. Let the level of generality in question be called level C (as explained below). A category is at level C if and only if (i) it is neither a species nor a genus of a category on L, and (ii) it is not a species of a category which is not on L and which satisfies (i). Observe that each one of the categories on L has as its genus either (the category of) Concrete or (the category of) Abstract. These genera are the level B categories. The genus of the two level B categories is Entity. As the summum genus, Entity is the sole (and universally applicable) level A category. Examples of species of level C categories are Surface, Shadow, and Soul. These species are level D categories (Rosenkrantz and Hoffman, 1991; Hoffman and Rosenkrantz, 2003). I propose to analyze the concrete/ abstract distinction as follows. An entity is concrete just in case it belongs to a level C category which is possibly instantiated by something which has spatial or temporal parts; an entity is abstract just in case it is not concrete. The general concept of a part (see part/whole) is taken as undefined. Examples of spatial parts are the halves of a material object, and examples of temporal parts are the halves of a day. Such proper parts are concrete. There may also be logical parts, for example, the conjuncts of the 183

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 184

c on s c i o usness molecular proposition that the sky is blue and the Moon is made of green cheese. Any proper parts of this kind are abstract. The basic idea of the proposed analysis is that an entity’s being concrete depends upon “the kind of company it keeps” within an ontological category at level C. Entities that belong to such a common level C category essentially have an ontologically significant resemblance that suffices for their belonging to that category, even if they differ in some ontologically significant respect. The proposed analysis of the concrete/ abstract distinction gives the desired result in the problem cases described earlier. First, even if a Cartesian soul lacks spatial and temporal parts, a Cartesian soul belongs (or at least would belong) to a level C category, namely, Substance, which could have some (other) instance with spatial parts, for example, a horse. Thus, the proposed analysis of the concrete/abstract has the desired consequence that Cartesian souls are concreta. Moreover, because some times and places have temporal or spatial parts, this proposal has the desirable implication that respectively, times and places are concrete entities. Furthermore, even if there could be an abstract entity, e.g., Sleepiness, which stands in spatial or temporal relationships, it is impossible that an abstract entity has spatial or temporal parts. Thus, the proposed analysis has the desired consequence that every instance of the level C category of Property is an abstract entity; likewise for every other level C category of abstract entity. Also note that the proposed analysis has the important advantage of being neutral about many ontological issues that other attempted analyses are not neutral about. For instance, regardless of whether universals are Platonic or Aristotelian, the proposed analysis implies that universals are abstracta. Furthermore, on the proposed analysis, Cartesian souls would be concreta. Moreover, the proposed analysis does not presuppose that all necessary beings are abstract, for whether or not time, space, space–time, or a substance are contingent or necessary, the proposed analysis implies that they are concrete entities. Lastly, the proposed analysis has the desirable implication that sets 184

are abstracta, even if some sets have spatiotemporal location in virtue of having spatiotemporally located elements. For sets belong to the level C category, Set, and it is impossible that an entity which belongs to that category has spatial or temporal parts. Observe that because being a part is a transitive relation, whereas being an element is not, elements of sets are not parts of them. (The notion that a set has spatial or temporal parts seems to confuse a set with a mereological sum, that is, a concrete collection which has parts.) See also the extended essay on realism and antirealism about abstract entities. bibl i ography Aristotle: Categoriae, trans. J.L. Ackrill, Aristotle’s Categories and De Interpretatione (Oxford: Oxford University Press, 1963). Campbell, K.: Metaphysics: An Introduction (Encino, CA: Dickenson, 1976). Hoffman, J. and Rosenkrantz, G.: “Platonistic Theories of Universals,” in The Oxford Handbook of Metaphysics, ed. M.J. Loux and D.W. Zimmerman (Oxford: Oxford University Press, 2003), 46–74. Kim, J.: “The Role of Perception in A Priori Knowledge: Some Remarks,” Philosophical Studies 40 (1981), 339–54. Loux, M.L., ed.: Universals and Particulars: Readings in Ontology (Garden City, NY: Doubleday, 1970). Rosenkrantz, G.: Haecceity: An Ontological Essay (Dordrecht: Kluwer, 1993), 56–68. Rosenkrantz, G. and Hoffman, J.: “The Independence Criterion of Substance,” Philosophy and Phenomenological Research 51 (1991), 835–53. gary s. rosenkrantz Consciousness The terms “conscious” and “consciousness” apply to a number of phenomena, all of them central to our mental lives. Though closely related, these phenomena are distinct, and require independent discussion. One phenomenon pertains roughly to being awake. A person or other creature is

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 185

consciousness conscious when it’s awake and mentally responsive to sensory input; otherwise it’s unconscious. This kind of consciousness figures most often in everyday discourse. A second phenomenon called consciousness occurs when a person or other creature is aware, or conscious, of something. One is conscious of something when one perceives it. One is also conscious of something when one thinks about it as being present; thinking about something as distant in space or time, like Saturn or Caesar, does not intuitively result in our being conscious of it. Most important theoretically, we describe thoughts, desires, perceptions, feelings, and other mental states as being conscious when we are aware of those states in a subjectively unmediated way. Thus Locke (1975/1700) wrote that “[c]onsciousness is the perception of what passes in a Man’s own Mind” (II, i, 19). This kind of consciousness is a property of mental states themselves, rather than of individuals that are in such states. Every conscious individual apprehends things in a characteristic way and from a distinct point of view. And an individual’s point of view brings a kind of unity to its conscious states. It is controversial whether this apparent unity is due to a connection among an individual’s conscious states or to some tie those states have to some underlying aspect of the individual’s mental make-up (see persons and personal identity). Conscious states are seldom the focus of attention; they simply occur within our stream of consciousness. States we deliberately attend to are introspectively conscious. Such introspective consciousness results in our awareness of ourselves as mental beings and as centers of consciousness (see self-consciousness). Locke, like Descartes, held that mental functioning is always conscious; as Descartes (1984/1641) put it, “we cannot have any thought of which we are not aware at the very moment when it is in us” (p. 171). Descartes and Locke thereby identify mind with consciousness; mental states are all conscious because being conscious is essential to being mental. Not everybody accepts this identification. Freud (1966–74/1915) famously maintained that many mental states occur

without being conscious and, indeed, that mental processes are not in themselves conscious. Most cognitive psychologists today agree, for reasons similar to, but not the same as, Freud’s. And even common sense countenances mental states that aren’t conscious; we sometimes know that somebody else wants or feels something, despite that person’s being wholly unaware of the desire or feeling. Mental states fall into two broad categories: intentional states, such as thoughts and desires (see intentionality), and qualitative states, such as pains and perceptual sensations (see sensa). So all the states in one category might be conscious even if not all of those in the other category are. Thus Descartes held that all intentional states are conscious. By contrast, many today who acknowledge that intentional states are not always conscious nonetheless insist that all qualitative states are. How, they ask, could a mental quality, such as redness or painfulness, occur without one’s being immediately conscious of it? What would it be like for one to be in a qualitative state if that state were not conscious? The idea that qualitative states are invariably conscious is inviting, and leads some to apply “consciousness” simply to conscious qualitative states. Even Freud (1966–74/1915) denied that emotions, which have a qualitative feel, can strictly speaking be unconscious; we loosely call emotions unconscious, he held, when the individual that has them is unaware of their true representational character. Some theorists hold that consciousness marks an unbridgeable gulf separating people from the rest of reality. This reflects the odd cognitive disorientation that Wittgenstein (1953, I, §412) called attention to when we reflect about how, as conscious beings, we might fit into the natural order. Others have argued that we can accommodate and explain all the phenomena we call consciousness within a scientific framework (see the extended essay on the mind/body problem; mental/physical). Much of the sense of mystery that surrounds consciousness results from assuming that all mental states are conscious. Since 185

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 186

c on s c i o usness consciousness involves mental functioning, we cannot explain it by appeal to anything that isn’t mental. But if being mental implies being conscious, perhaps any explanation of consciousness in terms of the mental will be circular. These considerations help bolster the sense that we cannot bridge the gap between mind and non-mental reality, but they result simply from assuming that all mental states are conscious. Various writers have pressed other difficulties in explaining consciousness. Levine (2001) acknowledges that specific brain events very likely result in specific conscious mental qualities. But he urges, as does Locke (1975/1700), that there may be no way to explain why each brain event results in the conscious qualities it does, or indeed why it results in any at all. Chalmers (1996) concurs, arguing that this is the Hard Problem in explaining consciousness. Others have argued that such explanation is possible. For one thing, Locke, Levine, and Chalmers all assume that qualitative properties cannot occur without being conscious, which restricts the range of possible explanations. Moreover, the sense that conscious qualities aren’t susceptible to scientific explanation may be largely due to our having at present no developed scientific theory about the connection between brain function and mental qualities. Levine contrasts our lack of understanding of that connection with the understanding we do have of water and its chemical composition. So perhaps when we come to have a theory of mental qualities and their connection with brain function which is as developed as current chemistry, the two cases will then be intuitively on a par. Skepticism about explaining qualitative consciousness is also due in part to a widespread view, advanced by Locke (1975/ 1700) among others, that consciousness is the only way we can know about mental qualities. Even if we’re occasionally mistaken about what qualitative state we are in, consciousness nonetheless provides our only access to mental qualities. This view reflects the conviction, common since Descartes, that consciousness yields infallible, incorrigible access to our mental states, 186

and indeed that consciousness reveals everything about their mental nature. And if we can learn about mental qualities only from consciousness, scientific explanation is precluded. This view again presupposes that mental qualities only occur consciously, since if they also occur non-consciously, we would know about them independently of consciousness. But subliminal perceiving is non-conscious, and it discerns the same qualitative similarities and differences that conscious perceiving does. So we have compelling reason to describe non-conscious, subliminal perceiving in the same qualitative terms we use for conscious perceiving. These similarities and differences, moreover, provide a way to describe and explain mental qualities independently of the way we are conscious of them (Rosenthal, 2005, Part II), thereby casting doubt on traditional claims of infallible or incorrigible access. Nagel (1974) has argued that qualitative consciousness must be understood in terms of what it’s like for one to have a conscious experience, which may not seem susceptible to scientific explanation. But the very notion of what it’s like for one arguably runs together two independent aspects of conscious experiences. As G.E. Moore (1922) emphasized, the qualitative character of conscious experiences, in respect of which they differ, is a different property from their consciousness, which they have in common. Even if qualitative states were always conscious, qualitative character would be a distinct property from the property of being conscious. So the two properties may well require independent explanations, which is obscured by just focusing on what it’s like for one. Block (1995) has argued that the kind of consciousness that is special to qualitative character is distinct from the kind in virtue of which states figure in rational thought, action, and speech. Block calls the first phenomenal consciousness and the second access consciousness. And he argues that the two require distinct accounts. Block’s distinction has been influential in philosophy and among scientific investigators, since qualitative consciousness is

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 187

consciousness plainly a special phenomenon. But it’s unclear whether phenomenal consciousness, as Block conceives of it, occurs in subliminal perception, which is not conscious in any commonsense, intuitive way. If it does, Block is simply distinguishing between mental qualities, which need not occur consciously, and consciousness ordinarily so called. Jackson (1986) has urged that, when an individual first has a novel qualitative experience, that individual learns something new, namely, what it’s like for one to have that experience. Moreover, no amount of physical information, he argues, would result in one’s knowing what it’s like for one to have that experience. Jackson concludes that what one knows in such a case is something non-physical. If so, conscious experience involves some nonphysical aspect. But it may be that knowing what it’s like for one to have a particular experience is distinct from the kind of knowing in which one has information, physical or not. Perhaps such knowledge consists simply in being acquainted with the experience, that is, simply in being conscious of it. If so, physical information would fail to help not because the knowledge is about something nonphysical, but because information by itself never results in one’s being acquainted with something. There is reason to think that knowing what it’s like for one to have a particular experience does consist simply in being acquainted with experiences of that sort. When one knows by having information what something is, one can say that it is such-and-such. But simply knowing what it’s like for one to experience red, for example, does not enable any such informative statement about experiencing red. A mental state’s being conscious consists in one’s being conscious of that state in a subjectively immediate way. This pivotal idea underlies a cluster of theories, according to which a state’s being conscious is a matter of one’s having some awareness of that state. Because of the appeal to such higher-order awareness, these theories are known as higher-order theories. The higher-order theory that has dominated traditional thinking about

consciousness is the inner-sense theory, advanced by Locke (1975/1700) among others, and today by Armstrong (1980) and Lycan (1996). On this view, the higherorder states in virtue of which we are aware of our conscious states are akin to perceptions. But perceiving always involves some qualitative character, and the relevant higher-order states do not. So the comparison with perceiving is arguably misleading. It’s tempting to see our higher-order awareness as serving to monitor our mental states, much as actual perceiving monitors external and bodily conditions. But it sometimes happens that we are conscious of ourselves as being in states that we are not actually in, so as to build a picture of our mental lives that makes sense to ourselves or to others (Nisbett and Wilson, 1977). Such confabulatory consciousness goes against the monitoring model, and hence against the inner-sense theory. An alternative higher-order theory holds that a state is conscious if one has a suitable thought that one is in that state (Rosenthal, 2005). Our awareness of our conscious states will be subjectively unmediated if the relevant higher-order thoughts do not rely on any inference or observation that one is conscious of. This theory avoids the difficulties of the inner-sense model, and has various additional advantages. Brentano (1973/1874) argued that our higher-order awareness of conscious states is intrinsic to those states. But this view is hard to sustain. The higher-order awareness must involve a mental assertion, since doubting or wondering whether one is in some state does not result in one’s being conscious of that state. But no mental state is both a mental assertion and a case of doubting or wondering. So, when a case of doubting or wondering is conscious, the relevant higher-order awareness will be distinct from the doubting or wondering itself, and hence not intrinsic to it. Some have challenged the basic principle on which higher-order theories rely, that a state’s being conscious consists in one’s being conscious of that state in a subjectively immediate way. Thus Searle (1992) argues that we never observe our mental 187

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 188

c on t e n t states. But observing things is not the only way of being conscious of them. Dretske (1995) urges that a state’s being conscious consists not in one’s being conscious of it, but in its being a state in virtue of which one is conscious of something. But subliminal perception also results in our being conscious of things; we are aware of the things we subliminally perceive, though not consciously aware of them. Indeed, subliminal perceiving would not affect our behavior and mental functioning if it did not make us in some way conscious of things. So Dretske’s theory has difficulty accommodating non-conscious perceiving. Dennett (1991) has advanced a different challenge to higher-order theories, arguing that the relevant hierarchy of states is not psychologically realistic. There is no difference, he argues, between how things seem to one and how they seem to seem. But collapsing that distinction again leads to difficulty with non-conscious states, such as subliminal perceptions. Some form of higher-order theory very likely offers the best way to accommodate the difference between conscious and non-conscious mental states. b i b l i og rap hy Armstrong, D.M.: “What Is Consciousness?,” in D.M. Armstrong, The Nature of Mind (St. Lucia, Qld: University of Queensland Press, 1980), 55–67. Block, N.: “On a Confusion About a Function of Consciousness,” The Behavioral and Brain Sciences 18:2 ( June 1995), 227– 47. Brentano, F.: Psychology from an Empirical Standpoint, ed. Oskar Kraus, English edn. ed. Linda L. McAlister, trans. Antos C. Rancurello, D.B. Terrell, and Linda L. McAlister (London: Routledge and Kegan Paul, 1973; originally published 1874). Chalmers, D. J.: The Conscious Mind: In Search of a Fundamental Theory (New York: Oxford University Press, 1996). Dennett, D. C.: Consciousness Explained (New York: Little Brown, 1991). Descartes, R.: The Philosophical Writings of Descartes, ed. John Cottingham, Robert Stoothoff, and Dugald Murdoch (Cam188

bridge: Cambridge University Press, 1984), vol. II; originally published 1641. Dretske, F.: Naturalizing the Mind (MIT Press/Bradford Books, 1995). Freud, S.: “The Unconscious,” in The Complete Psychological Works of Sigmund Freud, trans. and ed. James Strachey, 24 vols. (London: The Hogarth Press, 1966– 74; originally published 1915), Vol. I, 166–215. Jackson, F.: “What Mary Didn’t Know,” The Journal of Philosophy 83:5 (May 1986), 291–5. Levine, J.: Purple Haze: The Puzzle of Consciousness (New York: Oxford University Press, 2001). Locke, J.: An Essay Concerning Human Understanding, ed. Peter H. Nidditch (Oxford: Clarendon Press, 1975; originally published 1700). Lycan, W.: Consciousness and Experience (Cambridge, MA: MIT Press/Bradford Books, 1996). Moore, G.E.: “A Refutation of Idealism.” In Philosophical Studies, by G.E. Moore (London: Routledge and Kegan Paul, 1922), 1–30. Nagel, T.: “What Is It Like to Be a Bat?,” The Philosophical Review 83:4 (October 1974), 435–50. Nisbett, R. E. and Wilson, T. DeCamp: “Telling More Than We Can Know: Verbal Reports on Mental Processes,” Psychological Review 84:3 (May 1977), 231–59. Rosenthal, D. M.: Consciousness and Mind (Oxford: Clarendon Press, 2005). Searle, J. R.: The Rediscovery of the Mind (Cambridge, MA: MIT/Bradford Books, 1992). Wittgenstein, L.: Philosophical Investigations, ed. G.E.M. Anscombe and R. Rhees, trans. G.E.M. Anscombe (Oxford: Basil Blackwell, 1953). david m. rosenthal Content Mental states appear to come in two distinct kinds. On the one hand, there are states, like pains or tickles, whose nature is exhausted by what it feels like to have them, by their individuative phenomeno-

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 189

cont ent logies. Such states appear not to be “about” anything or to “mean” anything. On the other hand, there are states, like believing that snow is white, or desiring that the cat not scratch the furniture, which appear to have no interesting phenomenologies whatever, but which do seem to be about things, to mean something. For these latter sorts of state – states which Russell dubbed “propositional attitudes” – what they mean is referred to as their propositional content, or content for short. (The other part, the part designated by such psychological verbs as “believe” and “desire,” is the attitude adopted toward the propositional content.) The content of a propositional attitude is typically specified, in language, through the use of a “thatclause” – Jane desires that the cat not scratch the furniture, John believes that snow is white. The notion of propositional content raises a number of vexed questions in metaphysics, about which there is nothing but controversy. On the face of it, a belief attribution like the one mentioned in the preceding paragraph (mutatis mutandis for the other psychological states) appears to relate John by way of belief to some thing – the proposition that snow is white (see proposition, state of affairs). Thus, it seems correct to infer from John believes that snow is white to There is something that John believes. This seems to show that propositional contents are objects of some sort, to which persons can bear various psychological relations. But what sorts of objects are propositional contents, what sorts of thing are things believed? They seem to be abstract: that snow is white is not in Manhattan or in my car. They seem to be language-independent: that snow is white looks as if it might have been true even if no one had devised a language in which to express it. They seem to be independent of the existence of any particular mind: two people can share the thought that snow is white. They seem even to be independent of the existence of any mind whatever: that snow is white looks as if it might have been true even if no one had, or

even if no one could have, thought about it. Furthermore, and as the examples illustrate, propositional contents have conditions of truth (and falsity) and appear, indeed, to have their truth conditions essentially: no proposition could be the proposition that snow is white unless it were true if and only if snow is white. All of the preceding points are accommodated by the view that a propositional content is a set of possible worlds, namely, the set of all the worlds at which the proposition is true. Such a view has been quite popular in recent philosophy. But there are problems with it. Consider the belief that either snow is white or it is not white and the belief that 2 + 2 = 4. These appear to be distinct beliefs: it seems possible to believe the one without thereby believing the other. Yet since they are both necessarily true, they are both true in all possible worlds. A possible worlds conception of propositional content would appear, therefore, not to be able to discriminate between them. It would appear to have to conclude that anyone who believes one necessary truth believes them all. And that does not seem right. (For further discussion, see Stalnaker, 1984, 1999.) These considerations give one reason to hold that propositional contents are not merely sets, but more like structured complexes of objects and properties. The content of the belief that snow is white is the structured complex made up out of the substance snow and the property of being white (along with the property of exemplification or instantiation). This gets around the problem of believing necessary truths: the difference between the belief that 2 + 2 = 4 and the belief that either snow is white or it isn’t consists, in part, in the fact that the former involves the property of addition, whereas the latter does not. Unfortunately, a famous set of considerations due to Frege (1892/1980) seems to indicate that it cannot be right either. Consider the belief that water is potable and the belief that H2O is potable. These appear not to be the same belief, for it seems as if someone may have the one without thereby having the other. Indeed, it seems as 189

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 190

c on t e n t if a person may believe that water is potable and not only fail to believe that H2O is potable, but in fact actively believe, without contradiction, that H2O is not potable. The property of being water, however, just is the property of being H2O – or so science appears to teach us. So it seems as if belief contents must be made up out of constituents that are even more fine-grained than objects and properties. Such more finegrained constituents are normally referred to as modes of presentations of objects and properties. One of the large unresolved questions in the metaphysics of content concerns the nature of modes of presentation. (For further discussion, see Salmon, 1986; Schiffer, 1990; Soames, 2002.) Another important class of metaphysical problems raised by the topic of propositional content concerns the content relation. By virtue of what sort of fact is some token neural state the belief that p? (See the extended essay on the mind/body problem.) This question may be broken up into two others: By virtue of what sort of fact is a token state a belief (as opposed to, say, a desire)? And, by virtue of what sort of fact does it express that proposition that p? Concentrating on the second question, many philosophers are inclined to believe that the fact in question must be naturalistic (see naturalism), probably causal. There are many reasons for this conviction. Some are purely ontological: philosophers are loath to countenance properties that are not either identical with, or supervenient upon, the properties described by physics (see physicalism, materialism; reduction, reductionism; supervenience). Others are of a more explanatory character: it is hard to see how to give the content properties of beliefs a causal role in the explanation of behavior, on the assumption that they are not fundamentally naturalistic in nature. A non-reductive naturalism about content properties seems committed, implausibly, either to a peculiar sort of double causation or to the essential incompleteness of physics (see Kim, 1979, 2005; Yablo, 1992). It seems, then, that there is much to be said for a reductive naturalism about the content properties of beliefs. Unfortunately, however, 190

attempts to articulate a reductive naturalism of the required kind have met with very little success. Indeed, important arguments are available to the effect that content properties cannot be naturalized. Many of these highlight the allegedly normative character of the notion of content (see Davidson, 1980; Kripke, 1982; Boghossian, 1989, 1990a). The current impasse over the metaphysics of content has had a predictable effect – it has encouraged a growing skepticism about content. A significant number of contemporary philosophers are inclined to think that perhaps there are no mental states with content at all, that the idea of a contentful mental state is simply part of a bad and false ordinary psychological theory (see Churchland, 1981). It is unclear whether their skcepticism is justified; indeed, it is unclear whether it is even coherent (see Boghossian, 1990b; Wright, 2002). bibl i ography Boghossian, P.A.: “Naturalizing Content,” in Barry Loewer and Georges Rey, ed., Meaning in Mind: Fodor and His Critics (Oxford: Basil Blackwell, 1990), 65–86. Boghossian, P.A.: “The Rule Following Considerations,” Mind 98:392 (October 1989), 507–49. Boghossian, P.A.: “The Status of Content,” The Philosophical Review 99 (1990), 157– 84. Churchland, P.M.: “Eliminative Materialism and the Propositional Attitudes,” Journal of Philosophy 78 (1981), 67– 90. Davidson, D.: “Mental Events,” in his Essays on Actions and Events (Oxford: Clarendon Press, 1980). Frege, G.: “On Sense and Meaning” (1892), in Translations from the Philosophical Writings of Gottlob Frege, ed. P. Geach and M. Black (Totowa, NJ: Rowman and Littlefield, 1980), 56–78. Kim, J.: “Causality Identity and Supervenience in the Mind–Body Problem,” Midwest Studies in Philosophy 4 (1979), 31– 49. Kim, J.: Physicalism, or Something Near Enough (Princeton, NJ: Princeton University Press, 2005).

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 191

cont i nui t y Kripke. S.: Wittgenstein on Rules and Private Language (Cambridge, MA: Harvard University Press. 1982). Salmon, N.: Frege’s Puzzle (Cambridge, MA: Mit Press, 1986). Schiffer, S.: “The Mode-of-Presentation Problem,” in Propositional Attitudes, ed. C.A. Anderson and J. Owens (Stanford, CA: Csli, 1990), 56–78. Soames, S: Beyond Rigidity: The Unfinished Semantic Agenda of Naming and Necessity (Oxford: Oxford University Press, 2002). Stalnaker, R.: Inquiry (Cambridge. MA: Mit Press, 1984). Stalnaker, R.: Context and Content (Oxford: Clarendon Press, 1999). Yablo, S: “Mental Causation,” The Philosophical Review 101 (1992), 245–80. Wright, C: “What Could Anti-Realism About Ordinary Psychology Possibly Be?” The Philosophical Review 111:2 (April 2002), 205–33.

See also body; broad; change; event theory; temporal parts/stages; the extended essay on persistence.

paul a. boghossian

continuity A concept which now, strictly speaking, applies to a mathematical function, and not primarily to a domain. Initially, continuity was thought of as a notion that applies to the whole function (as in, for instance, a continuous line), with exceptional points specified where a “break” is. However, in the early nineteenth century, the property of continuity at a point was defined (by Cauchy (1789–1857) and Bolzano), which means intuitively that the correlates under the function of points which are “close” to the given point are also “close”. One is then free to say that a function is continuous at no points or at one point, or over a range of points, and there is then no need to assume the intuitive “unity” of this range. This frees the notion of the continuity of a function from any assumption that the underlying domain must be continuous (see class, collection, set). What it depends on instead is the underlying neighborhood structure which is used to make the notion of “close” precise. (For instance, in the modern conception, which generalizes Bolzano’s definition to the context of topological spaces, the continuity of a function does not depend on either the domain or range themselves being continua.)

contingent identity

see identity

continuant Continuants continue through time. They persist. In contrast, occurrents occur. Paradigm continuants are people, tables and rocks. Paradigm occurrents are events, such as an avalanche or a birth. Continuants change. The changes themselves are occurrents. A continuant persists if its temporal parts are connected in the right way. The right connection might be mere mereological summation, spatio-temporal continuity, causal dependence or, for people, continuity of consciousness. Many philosophers would object to this appeal to temporal parts. First, some hold that in order for a continuant to persist it must be one and the same thing that exists at different times and that the account here has two different things, two distinct temporal parts, existing at the different times. Second, some hold that, by definition, continuants cannot have temporal parts. The occurrent/ continuant distinction, they say, just is the distinction between having temporal parts and not having them.

bi bliography Broad, C.D.: An Examination of McTaggart’s Philosophy (London, 1933); (Cambridge: Cambridge University Press, 1976), 138ff. Hirsch, E.: The Concept of Identity (Oxford: Oxford University Press, 1982). Shoemaker, S.: Identity, Cause, and Mind (Cambridge: Cambridge University Press, 1984). Swoyer, C.: “Causation and Identity,” Midwest Studies in Philosophy 9 (1984), 593–622. Wiggins, D.: Sameness and Substance (Cambridge, MA: Harvard University Press, 1980). mark heller

191

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 192

c on t i n uo us / d iscr ete The continuity of things other than functions is either derivative on this latter, as with motion (the assumption that there is a continuous function from time to position), or is a loose way of saying that one is actually dealing with a continuum, as in the “continuity of space”. See also continuous/discrete. michael hallett continuous/discrete The notions discrete and continuous apply to two different kinds of quantity corresponding to the two different kinds of question “How many?” and “How much?” For instance, one can ask “How many apples are there on the tree?” and “How much water is there in the lake?”, whereas the questions “How much apples are there?” and “How many water is there?” make no sense. The first kind of quantity (numerical quantity) applies to concepts under which fall differentiated individual objects (the concept itself specifies some unit), and the second (continuous quantity) to concepts under which fall undifferentiated material (see mass terms) and where it can be sensibly asked in what measure or magnitude the substance is present, for instance, about its volume, area, length or weight. (Of course, it makes sense to ask quite different questions like “What is the weight of the apple?”, or “How many cups of water are there?”, since individual objects are often composed of continuous material, and undifferentiated material can be separated into units.) The distinction in this form goes back at least to Aristotle, and the two different kinds of quantity in later mathematical guise become the two notions whole number and real number. In modern physics, the distinction is alive in the two notions of quantum and field. Note that it is the paradigmatically continuous quantities that are taken as basic in classical physics, thus mass, position, time, velocity, momentum and acceleration. Connections between the two notions have been of deep philosophical and mathematical interest since the Pythagoreans’ claim that “All is like number”, a claim complicated by the discovery of irrational 192

quantities (see Presocratics). For instance, a descendant of this view is the position (the subject of Zeno’s celebrated paradoxes of motion) that space and time, while both continua, are actually composed of discrete elements, points and instants. The most profound ancient attempt to bridge the gap between discrete and continuous was that of Eudoxus (as given in Book V of Euclid’s Elements: see Heath, 1925), which attempts reduction of irrational measured quantity ultimately to ratios between numbered quantities, an attempt in which one can recognize Dedekind’s nineteenth-century theory of real number (see Stein, 1990). Why should one think that numerical and continuous quantity are connected? First, there is clearly an arithmetical component to continuous quantity (when divided into units) which is closely related to the algebra of whole numbers, for we combine the two types of quantity, for instance in saying that we have a total weight of 9 x w. This intermixing increases when one extends the algebra of whole numbers to that of the positive and negative integers and then to the rational numbers. Moreover. it is exploited and underlined in achievements like the Archimedean method of exhaustion, for this shows that one can get arbitrarily good approximations to measured quantities such as area, by using discrete sums. Here the order structure of the rationals, particularly their denseness (i.e., the fact that between any two rationals there is a third, in effect the assumption of infinite divisibility) plays an important role, as does the socalled “Archimedean axiom”, which says that whenever we have two quantities x and y with x < y, then there will always be an n such that y < nx (see Stein, 1990). Consequently, the structure of the discrete quantities, the whole numbers, and that of the rational numbers, ought to be embedded in the continuous quantities. A further question is then whether it is possible to generate the structure of the continuous out of the discrete. It was often mistakenly thought (e.g., by Kant) that infinite divisibility is actually what characterizes continuity and therefore continuous magnitude. In fact, this is not enough. The situation was

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 193

cont i nuous / di scret e only finally clarified by the development of the theories of real number and sets through the combined work of Bolzano, Dedekind (1831–1916), Cantor and Hilbert (1862–1943) in the nineteenth century. Bolzano’s work was important in two respects. First, Bolzano saw that what is important in real quantity is the combination of the algebraic and the order structures. Indeed, he isolated one fundamental property, namely that every bounded increasing sequence has a least upper bound (the l.u.b. property). Second, Bolzano thought it a mistake to try to base the general notion of real magnitude in the particular types of continuous magnitude referred to in the theories of space, time or motion, thus in effect eschewing any intuitions thought central to these. What Bolzano lacked was any precise account of how the algebraic and order structures of real number (what we would now call the system of complete, ordered fields) can be constructed from that of natural number. This was provided independently by Dedekind and Cantor. Dedekind gave a profound analysis of the notion of continuity, for which the l.u.b. property holds, and a demonstration of how the ordered field of real numbers (continuous quantity) can be defined in terms of the field of rational numbers, thus in effect in terms of the whole numbers (numerical quantities). Cantor and Dedekind were even clearer than Bolzano in rejecting the idea that direct intuition of space and time is what underlies the theoretical concept of the continuum. (This position is consciously anti-Kant, since Kant argued that mathematics must be based on “pure intuition” of space and time.) They argued, on the contrary, that the theoretical notion of continuity is itself needed in order to give a precise account of the nature of space and time. Indeed, it then becomes an hypothesis (or an “axiom”) that space is actually continuous. The reduction effected by Cantor and Dedekind only works when one appeals to the concept of infinity and the modern theory of sets (see class, collection, set). The settheoretic continuum has a very rich structure, and as a result, particularly with Cantor’s distinction between countable and

uncountable infinities, the modern theory of Lebesgue measure shows that some of the paradoxes presented by several of the Zeno arguments can be overcome. It is consistent to assume both that the continuum is made up of points, and that, while the points themselves have no size, and while no “small” (countably infinite) sub-collection of points has a positive size either, the whole continuum does have a size (see Grönbaum, 1968). The Cantor and Dedekind definitions take as their basis the notion of whole number, the measure of discrete quantity. It was Aristotle’s conception that what distinguishes discrete collections from a continuous mass is that the collection itself can be divided into elements “with no common boundary”. In the modern analysis, due to Frege, Cantor, and Dedekind, this requirement on the objects being counted is replaced by a requirement on the ordering in the counting numbers, namely that there is a first element, and that each element in the order has a unique successor. The smallest infinite such collection is, in effect, the natural number sequence. (Part of Frege’s and Dedekind’s achievement was to show how to characterize this sequence. Frege also argued further that, properly speaking, number specifies a property of the concept itself, thus something necessarily abstract, and does not reflect a property of the objects that fall under it.) Cantor generalized this requirement to the notion of well-ordering, thus extending the notion of counting number, and thus of discreteness, to the infinite. The acceptance by modern set theory that all sets (infinite as well as finite) can be wellordered is therefore tantamount to the claim that all mathematics can be based on discrete collections. On the other hand, the continuum problem (see cantor) perhaps shows that indeed there are still mysteries in the assumption that the continuum is made up of discrete points. bi bliography Frege, G.: Grundlagen der Arithmetik (Breslau, 1884); trans. J.L. Austin, The Foundations of Arithmetic (Oxford: Black-well, 1953). 193

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 194

c on t i n uum Grünbaum. A.: Modern Science and Zeno’s Paradoxes (London: Allen and Unwin, 1968). Heath, T.H.: The Thirteen Books of Euclid’s Elements, 3 vols., 2nd edn. (Cambridge: Cambridge University Press, 1925: New York: Dover Publications, 1956). Stein, H.: “Eudoxus and Dedekind: on the Ancient Greek Theory of Ratios and Its Relation to Modern Mathematics,” Synthese 84 (1990), 163–211. michael hallett continuum Any mathematical domain that possesses the property of being continuous. Since ancient times, continua have been of perennial concern to philosophers, e.g., through the questions of whether space, time, matter and motion, etc., are continuous, and what this might mean. The question reached genuine precision in the nineteenth century with the clarifications wrought by Bolzano, Dedekind (1831–1916), Cantor and Hilbert (1862–1943), leading to a categorical characterization of the continuum of real numbers (e.g., through Dedekind cuts) as a complete, ordered field. This is generally taken to be the paradigm of a continuum, although it is possible to give definitions of continuity, and therefore of continua, in frameworks which do not presuppose classical set theory. See also class, collection, set; continuity; continuous/discrete. michael hallett convergence If the methods of science were applied to ever widening data sets over a hypothetical infinite long run, would they have to converge on the truth? The pragmatic theory of truth says “Yes” (see pragmatism; theories of truth). Pragmatists define truth as that upon which inquiry converges. Realists disagree; they hold that it is a priori possible that evidence might be misleading, even in the limit. Even if truth cannot be defined in terms of convergence, the question remains of whether successive theories in fact take us 194

closer to the truth. This is one gloss of the idea that science is “progressive”. The difficulty is to define the idea of closeness to the truth. When two theories are both false, what makes one more truth-like? If Jane is 6 feet tall, the hypothesis that she is 5 feet 10 inches is clearly closer to the truth than the hypothesis that she is 5 feet 7 inches. The problem is to extend this idea to other sorts of hypotheses. bi bliography Laudan, L.: “A Confutation of Convergent Realism,” in Scientific Realism, ed. J. Leplin (Berkeley: University of California Press, 1984), 218–49. Peirce, C.S.: “How to Make Our Ideas Clear,” in Essays in Philosophy of Science, ed. V. Thomas (Indianapolis, IN: BobbsMerrill, 1957), 31–56. Russell, B.: “Pragmatism,” in his Philosophical Essays (New York: Simon and Schuster, 1966), 79–111. Russell, B.: “William James’ Conception of Truth,” in his Philosophical Essays (New York: Simon and Schuster, 1966), 112–30. Sober, E.: “Likelihood and Convergence,” Philosophy of Science 55 (1988), 228–37. elliott sober copula In traditional formal logic the statement “Socrates is wise” is analyzed into subject, predicate and copula (“is”). In metaphysics, however, the term “copula” is understood to refer to that fundamental connection that some philosophers suppose to hold between universals and particulars, and which may also be spoken of as “instantiation” or “exemplification”. The latter two terms are, indeed, more usual. More generally, the term refers to the supposed fundamental connection between substance and attribute, whether or not the attribute is taken as a universal. Johnson (1964) and Strawson (1959, ch. 5, pp.8– 9) spoke instead of a “non-relational tie”, Bergmann (1967, Bk. 1, Pt. 1) of a “tie”or “nexus”. The idea is always that here is a connection that is deeper, and stands behind,

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 195

cosmology mere relation. Frege’s idea that functions (very roughly, attributes) are “unsaturated”, that they call for completion by objects (1891), may reflect the same idea. Alternatively, it may reflect the opposite idea, that a copula is redundant in metaphysics and may be dispensed with. b i b l i og rap hy Bergmann, G.: Realism, Milwaukee, WI and London: University of Milwaukee Press, 1967). Frege. G.: “Function and Concept” (1891); in Translations from the Writings of Gottlob Frege, ed. and trans. P. Geach and M. Black (Oxford: Blackwell. 1960), 21–41. Johnson, W.E.: Logic, 3 vols. (New York: Dover, 1964). Strawson, P.F.: Individuals (London: Methuen, 1959). d.m. armstrong

cosmology In a wide sense, cosmology equals metaphysics: reality studied philosophically. This entry will instead discuss the study of the cosmos at large scales, using data from astronomy and physics. Cosmology involves severe verificational difficulties. Since light takes time to travel, far distant objects are seen as they were billions of years ago: how, then, shall we distinguish spatial from temporal variations in our universe’s properties? Again, if current cosmological models are even roughly correct then very much is so distant that light from it cannot yet have reached us, while on many models most of it will never be visible. Cosmologists must observe things very indirectly – i.e., with the aid of much theory – and must grant that much will never be knowable. Their science makes nonsense of neo-verificationism’s equation of the true with what would, in the long run, have warranted assertability (see principle of verifiability). How could a cosmologist ever be warranted in saying, for instance, exactly which path a particular particle followed after falling into a region from which light could not escape?

Trust in simplicity has carried cosmology far, however. Present-day observations are all consistent with Newton’s principle that the same basic laws operate everywhere. General relativity’s elegant equations, which associate gravity with spatial curvature, have likewise withstood all tests. True, the beautifully simple Perfect Cosmological Principle, that our universe is much the same in all big spatio-temporal regions, now seems erroneous: the associated Steady State models have given way to Big Bang ones which more straightforwardly explain why galaxies, like ink spots on an inflating balloon, rush apart at speeds proportional to the distances between them. Evidence for a Bang, a universe-wide explosion in which space itself began expanding, includes cosmic background radiation greatly uniform over the sky, and the observed amounts of hydrogen, helium, lithium and deuterium. All this would be explicable by immense early pressures and temperatures; and telescopes, as they probe further back in time, do indeed reveal more and more density and violence. Still, the almost equally beautiful Cosmological Principle, that our universe is much the same in all big spatial regions at any one time, has proved more successful than anyone dared hope. Starting life as a mere simplifying hypothesis, it is now a seldomquestioned dogma. It might nevertheless turn out that the Bang started off cold, which could help explain how galaxies managed to form in a highly uniform universe. (The cosmic background radiation which suggests a very hot Bang could come instead from early, massive stars.) Or the picture of near-total uniformity might need to be replaced by one of a space split into topologically ill-integrated domains. Cosmic strings, walls and other defects at domain boundaries could influence the matter distribution importantly: strings might be seeds for galaxies, for instance. There is actually a problem of why our universe is not crammed with knot-like defects, “monopoles” so massive and so numerous that their gravity would re-collapse it at once. And there is the more general Horizon or Smoothness Problem: the problem of how it could be in the least uniform, granted 195

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 196

c os m o l o gy that at early times it could seem to have been split into vastly many parts which had never interacted. (How could ships over one another’s horizons behave in coordinated fashion without benefit of signalling?) A currently popular solution is Inflation, the theory that everything exploded extremely rapidly before switching to much more leisurely expansion. A tiny, well coordinated region would in this case have come to include much more than is now visible to us: monopoles would be pushed far apart: and space’s resultant flatness – think of the surface of a gigantically inflated balloon – would yield the leisurely expansion in question, an expansion just fast enough to prevent or defer gravitational re-collapse and just slow enough to encourage galaxies to form. In Inflation’s absence, the early expansion speed would need enormously accurate tuning to permit their formation. If Inflation occurred then we can see only a minuscule fragment of the cosmos: perhaps as little as one part in one-followed-bya-million-zeros. Everything else lost any causal tie with us during the inflationary process, and has not since had time to link up with us. Now, the cosmos is by definition Absolutely Everything; there cannot be two “cosmoses”; but contemporary cosmologists often speak of multiple universes, meaning gigantic domains having few or no causal ties with one another. Belief in several universes is clearly “metaphysical” in that only one of them could be directly known. Yet an “antimetaphysical” opponent of them will have to reject Inflation, and Inflation is the dominant cosmological hypothesis nowadays because it solves so many puzzles so straightforwardly. It can even say how all parts of our huge universe managed to spring into existence and begin expanding nearly or completely simultaneously. Starting from a tiny seed, perhaps a mere quantum fluctuation, inflationary expansion would be driven by gravitational processes which could create more and more at no cost in energy because gravitational binding energy, like all binding energies in physics, is negative energy. It could precisely cancel the mass-energy of newly created matter. 196

Multiple universes can be obtained in other ways as well, Instead of being “closed”, curved round upon itself (thanks to gravity) like the surface of a sphere, space may be “open” – as it must be unless the gravitating matter visible to us is supplemented by much “dark matter” (massive neutrinos, perhaps). It then extends infinitely far, so we get infinitely many universes in the sense explained above: infinitely many regions causally distinct from one another because of light’s finite speed. Or, if closed, space might oscillate. Bangs and Squeezes succeeding one another for ever; now, each new oscillation might be called a new universe. Or again, if one universe could spring into existence through a quantum fluctuation then would it not be simpler to believe that greatly many had done so? Or might there not be an eternally inflating situation in which universes appear as bubbles inside which Inflation has ended? These various scenarios are all defensible (and attackable, even refutable) by physical arguments. There is little reason to suppose that all universes would seem to possess the same properties. Simplicity may demand that their most basic laws and properties be the same, yet contemporary physics suggests how derived laws and overt properties could be different. Gravity, electromagnetism, and the nuclear strong and weak forces, were probably unified into a single force at earlyBig-Bang temperatures, all particles perhaps then being massless. As things cooled, forces and particles could well have become differentiated in ways varying from place to place. Rather as a freezing pond becomes covered by ice-crystal domains with random orientations, the cooling products of a Hot Big Bang could split into regions differently oriented in an abstract “space”: the space in which overt properties become fixed during the “symmetry breaking” of which today’s physicists talk. Force strengths and particle masses could be determined by scalar fields which differed from region to region in a largely random way. It might thus be that in some regions – “other universes” pushed by Inflation far beyond the reach of our telescopes –the gravitational force between two protons was about as strong as the

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 197

cosmology electromagnetic instead of being many trillion times weaker, while protons and electrons were about equally massive. As Hume and Kant suspected, the situation visible to us may be very untypical of the cosmos as a whole. It could actually be that in many regions the overt geometry of space–time was, say, five- or six-dimensional. “Super-strings theory” suggests that the large scale fourdimensionality of the space–time familiar to us results from “compactification” in which further dimensions become tightly curled up. We cannot be confident that in absolutely all regions the same number of dimensions would compactify. Another ground for accepting greatly many universes with very varied (overt) properties is that this could help us understand the fact that there is a universe – ours – with life-permitting properties. Recent findings suggest that our universe is “finetuned for life” in the sense that slight changes in force strengths and particle masses would have ruled out organisms of any plausible kind. This could be an illusion, no doubt. Very odd life forms might be possible. Life in the sun might be based on plasma fields. Organisms inside neutron stars might use the nuclear strong force much as our bodies use the electromagnetism which underlies chemistry. However, such oddities can seem implausible. It can further be argued that much fine tuning is needed for there to be neutron stars, or suns, or atoms, or a universe lasting more than a microsecond and containing more than light rays and black holes. Considerable interest therefore attaches to the Anthropic Principle that the cosmic region which we observe must (selfevidently) possess properties compatible with the existence of observers. It is altogether plausible that most regions do not. Note (1) that the Anthropic Principle – as stated above, on the basis of a careful reading of B. Carter who enunciated it – is as tautologous as that villains are knaves; (2) that it can none the less give us an important reminder that the life-permitting situation which we see may very well not be typical of the cosmos, since even if lifepermitting situations were rare we should still (self-evidently) find ourselves inside

one; (3) that use of the Principle to help make our cosmic placement unmysterious is not a suggestion that the cosmos was planned for intelligent observers, let alone for humankind; or that we caused our cosmic region to have life-permitting properties; or that the existence of many other regions had somehow made it more probable that our particular region would develop life-permitting properties when the dice of random symmetry-breaking were tossed. Users of the Anthropic Principle need say nothing more controversial than (a) that in a cosmos large enough and varied enough it would be likely that life would appear somewhere or other even if it demanded very precise tuning of such things as force strengths and particle masses; and (b) that any region which living beings observed would be (believe it or not) one in which life was possible. Much confusion has been caused by distinguishing a “weak” from a “strong” Anthropic Principle. As intended by Carter, the weak Principle says that our spatiotemporal surroundings are (self-evidently) life-permitting; the strong, that (equally evidently) our universe is so. Alas, what one cosmologist calls a universe, another may describe as spatio-temporal surroundings. Again, Carter’s remark that our universe must be life-permitting has been widely misunderstood. It does not say that it was deterministically fated to become life-permitting, let alone that it had to become life-containing or that God fine-tuned it. If our universe is just one region of a very varied cosmos, a region perhaps very unusual in the fact that living beings can observe it, then it might easily look exactly as if it had been fine-tuned by God without having been so in fact. Theologians, however, could comment that the existence of multiple universes is no less conjectural than God’s. The reasons for believing in multiple universes are ultimately reasons of simplicity, and it is by no means clear that God would be non-simple. An infinite person, a divine Fine Tuner, might be in important respects simpler than any finite being. Or God might not be a person at all. Conceived Neoplatonically, God is the unconditionally real ethical requirement that there be a cosmos, a requirement 197

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 198

c os m o s which is itself creatively effective. Its cosmosproducing power would not, of course, follow from its sheer definition, yet linguistic analysis can seem to show that it would be the sort of reality which might have such power – and that the having of it would be exactly as simple as the lack of it. Theologians might also claim that God provides the best or the only answer to why there is any cosmos: any realities beyond mere possibilities and truths about them, such as the truth that two apples and two apples would be four apples or that good possibilities ought to be actualized. Yet is the sheer presence of a cosmos, a world of existing things, a genuine puzzle? Cosmologists have differed widely over this. G. Gamow (1904–68) thought it would indeed be one if the cosmos had a temporal starting point. He therefore proposed that the Bang was preceded by an infinitely prolonged contraction. F. Hoyle’s fight for a Steady State was largely motivated by a wish to avoid a beginning of things. W.B. Bonnor and J.A. Wheeler much preferred an infinity of cosmic oscillations to any infinitely dense state in which everything originated. S.W. Hawking prided himself on removing all need for a creator by making time progressively more space-like at ever-earlier moments in the Bang. Others have seen similar advantages in C.W. Misner’s idea that earlier and earlier processes ran progressively faster so that by the “clocks” of those processes themselves our universe could stretch backwards infinitely, or in A.D. Linde’s eternal Inflation (with bubble universes), or in the notions of E.P. Tryon and A. Vilenkin, that universes appear as quantum fluctuations in an eternally existing superspace or in a “foam” lacking clear distinctions of space and time. On the other hand, philosophers such as Hume and A. Grünbaum have said there would be nothing problematic in a beginning of things even if one granted (which Grünbaum would not) that time could exist before things did. And while some theologians – most notably Pius XII – have treated the Bang as proving God’s existence, others have insisted that God’s creative action is in no way specially associated with a first cosmic instant. 198

See also finite/infinite; why there is something; world. bibl i ography Barrow, J.D. and Tipler, F.J.: The Anthropic Cosmological Principle (Oxford: Clarendon Press, 1986). Davies, P.C.W.: The Accidental Universe (Cambridge: Cambridge University Press, 1982). Feinberg, G. and Shapiro, R.: Life Beyond Earth (New York: William Morrow, 1980). Harrison, E.R.: Cosmology (Cambridge: Cambridge University Press, 1981). Hawking, S.W. and Israel, W., eds.: Three Hundred Years of Gravitation (Cambridge: Cambridge University Press, 1987). Leslie, J., ed.: Physical Cosmology and Philosophy (New York: Macmillan, 1990). Leslie, J.: Universes (London and New York: Routledge, 1989). Munitz, M., ed.: Theories of the Universe (New York: Macmillan, 1957). Rozental, I.L.: Big Bang, Big Bounce (Berlin: Springer-Verlag, 1988). Weinberg, S.: The First Three Minutes, 2nd and rev. edn. (London: Fontana, 1983). john leslie cosmos On one widely accepted definition, the cosmos is the totality of existing things and events: it thus contrasts with any timeless realm of mere possibilities or Platonic truths. On this definition. God might be held to be part of the cosmos. Often, however, as in theistic “cosmological arguments”, God is viewed as its creator and orderer, existing outside it. Again, the cosmos may be contrasted with chaos, for to the Greeks kosmos meant “order” as well as “world”. In this case it might exist side by side with chaos, or it might have replaced chaos as in Hesiod and Milton (Paradise Lost, Bk. III: “at His Word, the formless mass . . . came to a heap . . . and wild Uproar stood ruled, stood vast Infinitude confined”). Such points all provoke philosophical disputes. Some urge that existents and their causal orderliness are brute facts, not Godproduced. Others think, with Spinoza, that

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 199

cosmos the cosmos has properties such that it can itself be called “God”. Among Neoplatonists (see Neoplatonism), a first might describe God as the creatively powerful ethical requirement that there exist a good cosmos; a second, as the ethical requiredness of such a cosmos, a requiredness which acts creatively; a third, as the cosmos, considered as having this requiredness. The first would then say God was distinct from the cosmos; the second, that God was an aspect of the cosmos; the third, that God and the cosmos were identical. These would be merely verbal disagreements. Other philosophers, again, have argued that an absence of all existents would be logical nonsense, or that events cannot fail to have orderliness. And still others have doubted the meaningfulness of talk about existence as a whole, or about any order beyond what minds impose on their experiences. It would seem, however, that an absence of all existents involves no actual contradiction, and that events have a causal orderliness which is neither logically inevitable nor created by our minds. Modal realism, the view that logical possibilities must all exist somewhere, is treated with suspicion by most logicians and may even render induction untrustworthy (since logically possible worlds which were orderly only in part could seem to form a wider range than ones orderly throughout; see the extended essay on modality and possible worlds). And the view, often attributed to Kant, that events have no causal order in themselves, would appear to make our unconscious minds into super-geniuses able to impose patterns whose complexity modern science only just begins to grasp. Fichte and F.W.J. Schelling (1775–1854), indeed, described the structure of the cosmos as a product of imagination, but it is a structure too complicated for this to be plausible. It also survives attempts to trivialize it by saying that absolutely any objects or events, like any dots scattered on graph paper, must obey some formula or other. It survives them because the vast majority of such formulas would be too messy for scientists to grasp at all. The cosmos, though indeed complicated, is not a chaos and can be understood.

Let us reject, too, the idea that it is meaningless to talk of existence in its totality. It is one thing to insist with Hume and Kant that we cannot know the entire cosmos, and quite another to deny meaning to all existing things, a phrase a child might understand. Intelligibility outruns verifiability. Modern cosmologists recognize that vastly much material lies beyond the horizon set by how far light can have traveled toward us since the start of the Big Bang, and that most of it may well never be visible from our cosmic region. The same applies to material nearer by but inside black holes. Things surely do not drop out of existence as soon as they fall through black-hole horizons. The sum of all existents must be singular, obviously – and “cosmos” has no plural so one cannot so much as speak of “many possible cosmoses”. However, modern cosmologists often talk of many possible universes and even propose their actual existence side by side or in succession. To them, “a universe” may mean only a huge cosmic domain, perhaps entirely separate from all other such domains or perhaps linked to them spatially or temporally. One ground for believing in multiple universes is that this may be simplest; any mechanism able to generate one universe might be expected to generate many. A second is that various physical theories suggest that a cosmos would soon split into huge domains with different properties. (It would thus cease to be “a cosmos” if this had to mean a unity in which such matters as the relative strengths of gravity and electromagnetism, or the relative masses of the electron, the proton and the neutron, were the same everywhere.) A third is that the existence of greatly many such domains might help explain why at least one domain is life permitting. Nowadays a popular theory is that our universe was born from the chaos of a “space–time foam”, then quickly inflated enormously before settling down to more leisurely expansion. Another possibility is that chaotic cosmic inflation continues eternally: our universe is just one of countless bubbles inside which it has ended. 199

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 200

c ou n t e r factuals See also cosmology; finite/infinite; pantheism; why there is something; world.

b i b l i o g r ap hy Diamandopoulos, P.: “Chaos and Cosmos,” in The Encyclopedia of Philosophy, ed. P. Edwards, 8 vols. (New York: Macmillan, 1967), vol. 2, 80–1. Laird, J.: Theism and Cosmology (London: Allen and Unwin, 1940). Leslie, J.: “Demons, Vats and the Cosmos,” Philosophical Papers 18 (1989), 169–88. Linde, A.D.: Inflation and Quantum Cosmology (San Diego, CA: Academic Press, 1990). Munitz, M.K.: “One Universe or Many?,” Journal of the History of Ideas 12 (1951), 231–55. john leslie

counterfactuals We distinguish how things actually were, are or will be from how things would have been or would be in this, that or the next eventuality. Nixon was re-elected, but had the American public known the truth about Watergate during the campaign he would not have been reelected. My lawn is green, but had it not been watered during the week it would be brown. I will be alive tomorrow, but were I to jump from the Empire State building today I would not be alive tomorrow, Conditionals like: “If the American public had known the truth about Watergate, then Nixon would not have been re-elected, “If my lawn had not been watered during the week, then it would (now) be brown”, and “If I were to jump from the Empire State building today, I would not be alive tomorrow” are called “counterfactual conditionals”, or simply “counterfactuals”. They are often symbolized as “p → q” (read “If p had been the case, q would have been the case” or “If p were the case, q would be the case”). p is the antecedent and q is the consequent of the conditional, and “ →” is used to distinguish counterfactual conditionals from indicative conditionals like “If my lawn was not watered last week, it is now brown” and “If I jump from the Empire State building 200

today, I will not be alive tomorrow” (sometimes symbolized as “p → q”), and the material conditional: (usually symbolized as “p ⊃ q”) which is true if, and only if, either p is false or q is true (or both). It has sometimes been argued that “p → q” and “p ⊃ q” are logically equivalent. It is obvious though that: “p → q” and “p ⊃ q” are not logically equivalent: “If I were to jump from the Empire State building today. I would be alive tomorrow” is false, and yet it has a false antecedent and a true consequent. The term “counterfactual conditional comes from the fact that use of the constructions “If p had been the case, q would have been the case”, and “If p were the case q would be the case” typically indicates that the speaker, or writer, takes p to be false: but there are exceptions, witness: “I realized that Smith was the murderer when I realized that had the 6:40 a.m. train been late his alibi would have been worthless.” Also although counterfactuals typically have consequents the speaker or writer takes to be at least doubtful, they can be used in contexts where it is known that the consequent is true, witness: “Smith failed, and he still would have failed had he worked. Counterfactuals are central to the discussion of at least three major topics in metaphysics: dispositional properties, laws, and causation (see the extended essay; see also disposition; law of nature). What makes solubility in water a dispositional property is the fact that for x to be soluble in water is, roughly, for x to be of a nature such that were it put in water, it would dissolve. What makes “Metals expand on heating” a law (approximately speaking) is the fact that not only is it the case that heated metals expand, it is in addition true that those not in fact heated would have expanded had they been heated. It is this latter fact which reflects the fact that it is no accident that heated metals expand. Finally, the sense in which a cause c brings about its effect e is connected to the fact that typically had c not occurred, e either would not have occurred or would have occurred in some significantly different way. Had CFCs not been released into the atmosphere, there would either be no hole

9780631199991_4_P2003.qxd 1/12/09 3:05 PM Page 201

c ount erfac t ual s in the ozone layer or a much less serious one. In all three cases there is dispute about how exactly to tease out the connection with counterfactuals, but not whether there is a conceptual connection to be teased out. It is natural to think of a counterfactual conditional “p → q” as about a possible (but typically non-actual) state of affairs where p obtains but which is otherwise much as things actually are (see the extended essay on modality and possible worlds; see also possible worlds). When we say that had the American public known the truth about Watergate, Nixon would not have been re-elected, we are saying that in the possible but non-actual state of affairs where the American public knew the truth about Watergate but where things otherwise were much as they actually were – the truth about Watergate was as it actually was, the voting system was as it actually was, the values of the public were much as they actually

were, and so on – Nixon was not re-elected. Thus the appeal of accounts (e.g., Stalnaker (1984) or Lewis (1973)) of the truth conditions of counterfactuals in terms of possible worlds of the following general shape: “p → q” is true (that is, is true at the actual world) if, and only if, the possible worlds most similar to the actual world where p is true are worlds where q is true. bi bliography Lewis, D.: Counterfactuals (Oxford: Blackwell.1973). Jackson, F., ed.: Conditionals (Oxford: Oxford University Press, 1991). Sosa, E., ed.: Causation and Conditionals (Oxford: Oxford University Press. 1975). Stalnaker, R.: Inquiry (Cambridge. MA: MIT Press, 1984). frank jackson

201

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 202

Davidson, Donald (1917–2003) Davidson’s general approach to metaphysics follows a long-standing tradition in trying to derive the basic features of reality from the structure of language. The particular angle he introduces comes from his suggestion that the structure of language and its revelations about the large features of reality are refracted in the effort to formulate a comprehensive, formal theory of truth for a canonically regimented version of natural language (see Davidson, 1984, essay 13; for some criticisms of this approach to metaphysics, see Rovane, 1986). The most original specific application of this method to metaphysics is to be found in Davidson’s claim that such a theory for that fragment of language which contains sentences such as “The boiler exploded”, yields an ontology of events. That is to say, it yields the idea that events such as explosions exist as particulars, in the way that boilers do. There is something very natural and appealing about the suggestion because, among other things, it allows for our redescription of the same event in different ways. Thus one can say of the same event that it was an explosion and that it was a domestic disaster (see Davidson, 1980, essay 9; for a different, more universalist view of events, see Chisholm, 1970. For a different view of events as particulars, see Kim, 1976. Davidson discusses the criterion for individuation of events in 1980, essay 8, and in Davidson, 1985.) Davidson extends the point to sentences such as “Jones buttered the bread”, arguing that these sentences about human agency too are to be treated as quantifying over events which get intentional descriptions. He points out that this affords a satisfying theoretical treatment of sentences with 202

adverbial modifiers. In the standard predicate calculus, adverbially modified predicates are usually represented as distinct predicates, but that representation fails to capture the seeming validity of the argument which goes from, say, “Jones buttered the toast with a knife” to “Jones buttered the toast.” There is simply no sanction for Ga from Fa. But the intuitive validity of the argument is preserved if we take events to exist as particulars, and treat the canonical representation or “logical form” of such sentences as quantifying over them. Thus we may go from ∃x (x was a buttering of the toast by Jones, and x was with a knife) to ∃x (x was a buttering of the toast by Jones) (see Davidson, 1980, essay 6. There are, of course other suggestions in the literature for handling adverbial modification, such as, for instance, that one should introduce not quantification but modifiers of predicates.) He further exploits the point to give an analysis of sentences citing causes, such as. “His pressing the button caused the explosion.” The ontology of events allows him to make a distinction between two different aspects of what these sentences convey: causal relations which hold between events and which are purely extensional, and causal explanations which are intensional in the sense that they, unlike causal relations, depend upon how the events are described (Davidson, 1980, essay 7). Davidson, then, develops this distinction to provide a solution to the traditional mind/ body problem (see the extended essay). Mental events are identical with physical events, but when we gather these mental events into types, there is a principled objection to their being identical with the types we gather physical events into. This is because unlike the particulars (the tokens of mental

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 203

davi dson, donal d events), the types are essentially dependent on the concepts we employ in describing the events; and there are no lawlike correlations between these mental concepts and the concepts we employ in physical descriptions of events. So, at the level of concepts there is no reduction of the mental to the physical, but, as far as ontology goes this does not imply a dualism since any mental event is identical with some physical event or other. He calls this hybrid position, “anomalous monism”, and the view of identity it proposes, “token-identity” (Davidson, 1980, essay 11). In addition to token-identity, Davidson’s metaphysics of mind posits a dependency relation between the mental and the physical, which he calls supervenience. This is the idea that a psychological predicate will not distinguish anything that is not also distinguished by some physical predicate(s). The underlying motivation for such a dependency was to make a claim in the philosophy of mind parallel to a position in the study of values which both denied a reduction of values to facts of nature at the same time as it did not make values mysteriously autonomous. In positing supervenience, Davidson was able to claim that a denial of a definitional as well as a nomological reduction of mental properties to physical properties was compatible with a dependency relation between them which disallowed one from saying that two things were indistinguishable physically but were different in some mental respect. This allowed the scientific study of physical nature maximum comprehensiveness in its dominion without any concession to mind/body reductionism. More recently he has also invoked supervenience to quell a worry which has loomed at least since Descartes – that the mental is epiphenomenal. In contemporary discussions this worry has sometimes been expressed as the worry that mentality makes no difference to causal relations. (See Kim, 1984 and Sosa, 1984 who raise this worry for Davidson’s anomalous monism, in particular. For a response on behalf of Davidson, see LePore and Loewer, 1987.) So expressed, supervenience is not needed to deal with it. The worry is handled

in Davidson’s metaphysical framework by simply appealing to the extensional nature of causal relations. Since events which enter causal relations are often described in mental terms, mental events (uncontroversially) enter causal relations. But for him it makes no sense to go on to ask whether mental events make a difference to causal relations in the sense that they cause other events in virtue of being mental. Causal relations being purely extensional do not hold or occur in virtue of anything, mental or physical. However the worry about epiphenomenalism is sometimes expressed as not being about causal relations in particular, but more generally as, say, the idea that we may alter (in the limit, even strip all) mental properties without at all affecting the physical properties of things. It is to this worry that Davidson responds by pointing out that if it were true it would contradict the weak dependency relation of supervenience as characterized above. However this subject throws up many questions issuing from modal intuitions about identity, intuitions whose relevance Davidson has always been suspicious of; these questions are at present the subject of much controversy in metaphysics and the philosophy of mind. There are other aspects of Davidson’s philosophy – such as his views on realism, objectivity and the nature of truth – which may be treated as being part of metaphysics, but because of their integral relation with epistemological themes, are best discussed within epistemology. See also event theory; reduction, reductionism; theories of truth. writings Essays on Actions and Events (Oxford: Oxford University Press, 1980). Inquiries into Truth and Interpretation (Oxford: Oxford University Press, 1984). “Reply to Quine on Events” (1985), in Actions and Events: Perspectives on the Philosophy of Donald Davidson, ed. E. LePore and B. McLaughlin (Oxford: Blackwell, 1986). 203

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 204

death b i b l i og rap hy Chisholm, R.M.: “Events and Propositions,” Noûs 4 (1970), 15–24. Kim, J.: “Epiphenomenal and Supervenient Causation,” Midwest Studies in Philosophy 9 (1984), 257–70. Kim, J.: “Events as Property Exemplifications,” in Action Theory, ed. M. Brand and D. Walton (Dordrecht: Reidel, 1976). LePore, E., ed.: Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson (Oxford: Blackwell, 1986). LePore, E. and Loewer, B.: “Mind Matters,” Journal of Philosophy 84 (1987), pp. 630– 642. LePore, E. and McLaughlin, B., ed.: Actions and Events: Perspectives on the Philosophy of Donald Davidson (Oxford: Blackwell, 1986). Rovane, C.: “The Metaphysics of Interpretation,” in Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, ed. E. LePore (Oxford: Blackwell, 1986). Sosa, E.: “Mind–Body Interaction and Supervenient Causation,” Midwest Studies in Philosophy 9 (1984), 271–81. akeel bilgrami

death Death provokes a wide variety of philosophical questions. There is the fundamental conceptual question about the nature of death itself. There is the metaphysical question about whether we continue to exist after death. There is the epistemic question whether death is in some distinctive way unknowable. Finally, there are ethical and value-theoretic questions about death: is death an evil for the one who dies? Is it irrational to fear death? Why is it wrong to kill people? Specialists in medical ethics have written extensively on the problem of formulating an acceptable criterion of death for human beings. While this is an important public-policy issue to which philosophers may make useful contributions, it seems not to be as fundamentally metaphysical in nature as the question about the analysis of the concept of death. 204

Some of the philosophical literature on death seems to presuppose that there is a concept of death uniquely applicable to people. However, it seems more natural to suppose that the central concept of death applies uniformly to things of every biological sort and that the word “dies” expresses this “biological concept of death” whenever we say (using the word literally) that some organism dies. the standard anal ysi s According to the most popular analyses, “x dies at t” means roughly the same as “x ceases to live at t”. Reflection on facts about suspended animation suggest that this analysis fails to capture precisely what we mean when we say that something dies. When an organism enters suspended animation (as for example a microscopic laboratory specimen does when placed in liquid nitrogen), it ceases to be alive. Yet it does not die – this is especially clear if the organism is going to be revived later. Further reason to doubt that death can be defined as the cessation of life is provided by organisms that reproduce by division. When such an organism divides, it apparently ceases to exist. Hence, it ceases to be alive. Yet, once again, it is inappropriate to say that it has died. the survival of death In the Phaedo, Plato presents a dualistic conception of persons according to which each person is composed of two main parts, a body and a soul. Death occurs when body and soul are separated. The body is purely physical and begins to deteriorate at death. Unless mummified, it will soon disintegrate and go out of existence. On the other hand, since the soul is that in virtue of which the organism lives, it must be immortal and imperishable. It existed prior to birth and will continue to exist after death. Thus, on this view, though the person does not survive death, the body may survive for a short time, and the soul survives eternally. This view has obvious affinities to traditional Christian doctrines.

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 205

deat h Materialists hold that a person is a living human body. Some hold that life is essential to persons, so that at the moment of death the person ceases to exist, and is replaced by a corpse. Other materialists deny that life is essential to the things that are persons. They claim that death marks a change in state of a continuing entity – the body. On this view, most people continue to exist as corpses for a few months or years after death. They cease to exist when they disintegrate. Continued existence of this sort would almost certainly be devoid of experience, and if so would be of no value to the dead (former) person. t h e m y s ter y o f death A number of philosophers have maintained that death is a mysterious or unknowable phenomenon. Some who maintain this view apparently do so because they believe that in order to understand death, one must understand how the experience of being dead presents itself to those who are dead. If one also believes that the dead have no psychological experiences, one will face this paradox: in order to understand death, one must understand how it feels to feel nothing at all. It seems, however, that the underlying epistemic requirement is unreasonable. Being dead is not an experience; it does not “feel like” anything at all to those who are dead; hence, the understanding of death cannot call for an understanding of what death feels like to those who are dead. Perhaps we understand death well enough when we take note of its occurrence in other organisms. Assuming that one is also a biological organism, one can conclude that one’s own death will be relevantly like those other deaths. t h e e v i l o f death It is natural to fear death, and to think that death is ordinarily a great evil for the one who dies. Epicurus argued that the fear of death is irrational since death is never evil for the one who dies. His argument was based on two main premises: (1) the notion

(derived from his dualistic conception of persons) that each person ceases to exist at death; and (2) the claim that nothing bad can happen to a person at a time when he or she does not exist. Defenders of the Deprivation Approach acknowledge that people cannot undergo painful experiences once they are dead. Nevertheless, they insist, death may be bad for the deceased inasmuch as it deprives them of the good things they would have experienced if they had not died. Murder is generally taken to be the paradigm of morally impermissible action, yet it is not easy to explain precisely why murder is wrong. If the Epicurean view were true, then (provided it were done painlessly) murder would never harm its victim. The Deprivation Approach seems to imply that whenever continued life would be of overall negative value to the victim, he or she is not harmed by painless murder. This also seems wrong. Traditional forms of utilitarianism imply that murder is morally required whenever the intended victim is “dragging down” the worldwide utility total. This seems to imply (absurdly) that we ought to kill everyone who would lead a life that is, on the whole, unhappy. See also life; persons and personal identity; physicalism/materialism; vitalism; the extended essay on the mind/body problem. bibl i ography Donnelly, J., ed.: Language, Metaphysics, and Death (New York: Fordham University Press, 1978). Epicurus: “Letter to Menoeceus,” trans. C. Bailey, in The Stoic and Epicurean Philosophers, ed. with an introduction by W.J. Oates (New York: The Modern Library, 1940), 30– 4. Feldman, F.: Confrontations with the Reaper: A Philosophical Study of the Nature and Value of Death (New York: Oxford University Press, 1991). Fischer, J.M.: Essays on Death (Stanford, CA: Stanford University Press, 1993). Nagel, T.: “Death,” Noûs 4 (1970), 73–80; repr. in his Mortal Questions (Cambridge: 205

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 206

d e s c a r tes, r ené Cambridge University Press, 1979), 1– 10. Plato: Phaedo, in The Dialogues of Plato, trans. B. Jowett (New York: Random House, 1937), vol. 1, 441–501. fred feldman Descartes, René (1596–1650) French philosopher and mathematician. It would be hard to overestimate the philosophical influence of Descartes. Often called the “father of modern philosophy”, his arguments on doubt, the foundations of knowledge, and the nature of the human mind, are familiar to countless students. But while Cartesian ideas almost inevitably form the point of departure for our understanding of how epistemology and philosophy of mind developed from the early modern period to the present day, the situation with respect to metaphysics is not so simple. There is some evidence that Descartes’s own motivating interests in philosophy were not primarily metaphysical. Non adeo incumbendum esse meditationibus (“You should not give such obsessive attention to metaphysical meditations”) he told the young student Frans Burman (Œuvres de Descartes, vol. V, p. 165; The Correspondence, p. 346); he gave similar advice to that keen amateur metaphysician Princess Elizabeth of Bohemia (Œuvres de Descartes, vol. III, pp. 692ff.; The Correspondence, pp. 227ff.). Most of Descartes’s time as a young man was occupied with mathematical and scientific concerns, including detailed work in specific areas such as geometry and optics (on both of which subjects he published essays in 1637) as well as grand theorizing about cosmology and the nature of matter (developed in his unpublished treatise Le Monde (1633)). Even when he came to publish the Discourse on the Method (1637), he devoted only one short section (Part IV) to metaphysics; the rest of the work is concerned with his early education and intellectual development, current scientific interests, and plans for future research. In general, there is a considerable amount of evidence to support the thesis of Charles Adam that metaphysics was of merely subsidiary interest to the historical 206

Descartes, and that he embarked on metaphysical inquiries for one reason alone – to provide solid foundations for his scientific system (Œuvres de Descartes, vol. XII, p. 143). But whatever Descartes’s own personal priorities may have been, metaphysics nonetheless forms an integral part of his conception of philosophy. In the celebrated simile which he deploys in the 1647 Preface to the French Edition of the Principles of Philosophy, philosophy is compared to a tree of which “the roots are metaphysics, the trunk is physics, and the branches emerging from the trunk are the other sciences” (ibid., vol. IXB, p. 14; The Philosophical Writings of Descartes, vol. I, p. 186). Even here, it is the fruit to be collected from the extremities of the branches which Descartes goes on to stress: the value of the system lies in the practical benefits it can bring to mankind (cf. Œuvres de Descartes, vol. VI, p. 62; Philosophical Writings, vol. I, p. 142). But it is also made clear that only a soundly rooted tree can bear such fruit. One of Descartes’s frequent criticisms of the scholastic philosophy in which he had been trained as a young man is that it often started from principles which were either obscure or doubtful or both: “nothing solid could have been built on such shaky foundations” (Œuvres de Descartes, vol. VI, p. 8; Philosophical Writings, vol. I, p. 115). We know from Descartes’s correspondence with his friend Marin Mersenne that as early as 1629 he had begun to compose a “little treatise” on metaphysics which aimed to prove “the existence of God and of our souls when they are separated from the body” (Œuvres de Descartes, vol. I, p. 182; The Correspondence, p. 29). The treatise was, however, laid aside, and by the time he came to write his metaphysical masterpiece, the Meditations (1641), Descartes had broadened his conception of metaphysical inquiry; he wrote to Mersenne that he had chosen the title “Meditations on First Philosophy” to show that “the discussion is not confined to God and the soul, but treats in general of all the first things to be discovered by philosophizing” (Œuvres de Descartes, vol. III, p. 235; The Correspondence, p. 157). In the order of discovery unfolded in the Meditations, what

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 207

descart es, rené the meditator reaches first of all is the indubitable knowledge of his own existence (Second Meditation). This result suggests (at the start of the Third Meditation) a general rule for the development of further knowledge, namely that “whatever I perceive very clearly and distinctly is true” (Œuvres de Descartes, vol. VII, p. 35; Philosophical Writings, vol. II, p. 24); however, since the doubts of the First Meditation have still left open the possibility that we might go astray even in our clearest and simplest perceptions, the meditator rapidly realizes that no further progress can be made “until I examine whether there is a God, and if there is, whether he can be a deceiver” (Œuvres de Descartes, vol. VII, p. 36; Philosophical Writings, vol. II, p. 25). The remainder of the Third Meditation is spent establishing the existence of a perfect, non-deceiving God: the idea of such a being, which I find in my mind, could not have been generated from my own resources, but must have as its cause an actually existing God. “By the word ‘God’ I understand a substance that is infinite, eternal, immutable, independent, supremely intelligent, supremely powerful . . . All these attributes are such that the more carefully I examine them, the less possible it seems that they could have originated from me alone. So it must be concluded that God necessarily exists” (Œuvres de Descartes, vol. VII, p. 45; Philosophical Writings, vol. II, p. 31). The existence of God, once established, is used to set up a sound method for humans to seek the truth, namely restraining their will so as to assent only to what is clearly perceived: God, though he has given man a limited intellect, guarantees none the less that it is, in principle, a reliable instrument for the pursuit of truth, and that, when carefully used, it will not lead us fundamentally astray (Fourth Meditation). Once this principle is established, the meditator can proceed to lay down the metaphysical foundations for a secure philosophical system: these are, on the one hand, my perception of matter as an “extended thing” – whatever can be quantitatively defined, and is the “subject matter of pure mathematics” (Fifth Meditation), and, on the other hand, my

perception of myself as a “thinking, nonextended thing” which is entirely distinct from the body (Sixth Meditation). This last result is of course the famous thesis of socalled “Cartesian dualism” – the conception of mind and body as separate and incompatible substances. It is significant that when Descartes presents the thesis, he provides direct metaphysical underpinning for it, in the shape of an appeal to the deity: “the fact that I can clearly and distinctly understand one thing [mind] apart from another [body] is enough to make me certain that the two things are really distinct, since they are capable of being separated, at least by God” (Œuvres de Descartes, vol. VII, p. 78; Philosophical Writings, vol. II. p. 54). It may be seen from this brief summary that the role of God in Cartesian metaphysics is absolutely central. But Descartes’s reliance on the deity in developing the foundations of his philosophy is problematic in at least two ways. The first is the famous puzzle of the “Cartesian circle”: if God is to be invoked to underwrite the reliability of the human mind, how can we be sure of the reliability of those perceptions we need to establish the existence of God in the first place? (cf. Œuvres de Descartes, vol. VII, p. 246; Philosophical Writings, vol. II, p. 171). The second problem concerns the details of Descartes’s proof of God’s existence. Despite his professed aim of sweeping away all preconceived opinions and basing his “first philosophy” on completely clear and transparent premises, the proof of God in the Third Meditation relies on what are (to the modern ear at least) highly questionable assumptions about causation (see the extended essay). According to Descartes, the cause of my idea of God must actually contain all the perfection represented in the idea. It is “manifest by the natural light”, claims Descartes, that “there must be at least as much reality in the cause as in the effect”, and hence “that what is more perfect cannot arise from what is less perfect” (Œuvres de Descartes, vol. VII, p. 40; Philosophical Writings, vol. II, p. 28). What Descartes is in effect presupposing here is a theory of causation that is deeply indebted to the scholastic philosophical apparatus which 207

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 208

d e s c a r tes, r ené it is his official aim to supplant. According to the scholastic conception, causality is generally understood in terms of some kind of property-transmission: causes pass on or transmit properties to effects, which are then said to derive their features from the causes. This traditional conception of causality is largely bypassed in Descartes’s mathematically based physical science; but here in his metaphysics he appears to accept it all on trust. This type of problem, indeed, is not confined to presuppositions about causation. Throughout the argument for God’s existence, the reader is faced with a positive barrage of traditional technical terms (“substance” and “mode”, and terms denoting various grades of reality – “formal”, “objective”, “eminent” and the like), whose application the reader is asked to take as self-evident. In short, when endeavoring to establish the metaphysical foundations for his new science, Descartes seems unable to free himself from the explanatory framework of his scholastic predecessors. (Similar strictures are applicable to Descartes’s other strategy for proving God’s existence, the so-called “ontological argument” which Descartes puts forward in the Fifth Meditation: Œuvres de Descartes, vol. VII, p. 66; Philosophical Writings, vol. II, p. 46). The structure of Cartesian metaphysics is often described as “rationalist” in character. The term is an awkward and often ambiguous one. Sometimes it is used to denote a purely a prioristic conception of knowledge; but Descartes’s conception is certainly not of this kind. It is true that his version of the ontological argument does try to prove God’s existence simply from the definition or essence of God, but many other elements of his metaphysical system (the Cogito, the causal proof of God’s existence, and the proof of the external world in the Sixth Meditation) proceed a posteriori, and rely on existential premises of various kinds. What makes the term “rationalist”, in a broader sense, seem appropriate, is Descartes’s belief that the human mind is innately endowed with a God-given “light of reason” or “natural light”, on the basis of which it has the power to discern the nature of reality. In Descartes’s early work, 208

the Rules for the Direction of our Native Intelligence (c.1628), it is the light of reason that enables us to intuit the “simple natures” – the fundamental building blocks for systematic knowledge of God, mind and matter (see Rule Four and Rule Twelve). This broadly “rationalistic” aspect of Descartes’s metaphysics is complicated by one of his most perplexing doctrines – that of the divine creation of the eternal truths. This doctrine is not found in the Meditations, but it is explicitly asserted in Descartes’s correspondence, as early as 1630, and it surfaces again in the Replies to the Objections: “God did not will that the three angles of a triangle should be equal to two right angles because he recognized that it could not be otherwise; . . . it is because he wills that the three angles of a triangle should necessarily equal two right angles that this is true and cannot be otherwise” (Œuvres de Descartes, vol. VII, p. 432; Philosophical Writings, vol. II. p. 291; cf. Letter to Mersenne of April 15, 1630, Œuvres de Descartes, vol. I, p. 145; The Correspondence, p. 23). Descartes thus departs from the traditional theological notion that God’s omnipotence extends only to what is logically possible. For Descartes, God is not only the creator of all actually existing things, but he is the author of necessity and possibility; he was “just as free to make it not true that the radii of a circle were equal as he was free not to create the world” (Œuvres de Descartes, vol. I, p. 152; The Correspondence, p. 25). Some of Descartes’s critics objected that this was incoherent, but Descartes replied that just because we humans cannot grasp something, this is no reason to conclude that it is beyond the power of God. God thus turns out, on Descartes’s conception, to be in a real sense incomprehensible: our soul, being finite, cannot fully grasp (Fr. comprendre, Latin comprehendere) or conceive him (ibid.). The doctrines of the divine creation of the eternal truths and the incomprehensibility of God make the character of Descartes’s metaphysics very much less “transparent” than the rationalist label implies. If the structure of the fundamental principles of logic is not ultimately accessible to human reason, but depends on the inscrutable will

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 209

determinate / determinabl e of God, then the human mind is not, after all, able to uncover their fundamental rationale. Indeed, if the principles of logic are arbitrary fiats of the divine will, which could be otherwise (though in a sense not accessible to our intellect), then there appear to be elements of opacity and contingency at the very heart of Cartesian metaphysics. If this is right, then the contrast between Descartes’s metaphysical “rationalism”, with its alleged optimism about the powers of human reason, and Hume’s later skepticism about our ability to discern the ultimate basis for the way things are, turns out not to be as stark as is often supposed. See also clear and distinct; light of nature; rationalism; the extended essay on the mind/body problem. writings Descartes’s main metaphysical works are the Meditations on First Philosophy (1641) and Part I of the Principles of Philosophy (1644). See also Part IV of the Discourse on the Method (1637). Descartes’s correspondence is also a valuable source for his views on metaphysics. All these materials are contained in the following editions. The Correspondence, ed. and trans. J. Cottingham, R. Stoothoff, D. Murdoch, and A. Kenny, The Philosophical Writings of Descartes, vol. 3 (Cambridge: Cambridge University Press, 1991). Œuvres de Descartes, ed. C. Adam and P. Tannery, rev. edn., 12 vols. (Paris: Vrin/CNRS, 1964–76). The Philosophical Writings of Descartes, ed. and trans. J. Cottingham, R. Stoothoff, and D. Murdoch, 2 vols. (Cambridge: Cambridge University Press, 1985). b i b l i og rap hy Beyssade, J.-M.: La philosophie première de Descartes (Paris: Flammarion, 1979). Cottingham, J., ed.: The Cambridge Companion to Descartes (Cambridge and New York: Cambridge University Press, 1992). Cottingham, J.: Descartes (Oxford: Blackwell, 1986).

Gaukroger, S.: Cartesian Logic (Oxford: Clarendon Press, 1989). Kenny, A.: Descartes (New York: Random House, 1968). Marion, J.-L.: Sur la théologie blanche de Descartes (Paris: Presses Universitaires de France, 1981, rev. ed. 1991). Wilson, M.D.: Descartes (London: Routledge, 1978) john cottingham determinable minable

see

determinate/deter-

determinate/determinable In 1921, the Cambridge logician W.E. Johnson introduced the contemporary use of the terms “determinate” and “determinable” into the philosophical lexicon by citing a pair of exemplars: I propose to call such terms as colour and shape determinables in relation to such terms as red and circular which will be called determinates. (1921, p. 174) The determinable–determinate relation thus holds between pairs of predicables or properties, and Johnson proceeded to characterize it by citing four of its characteristic marks or features. First, determinate properties come in families, and to each such family of determinates corresponds one and only one determinable from which it “emanates”. [Any] one determinable such as colour is distinctly other than such a determinable as shape or tone: i.e. colour is not adequately described as indeterminate, since it is, metaphorically speaking, that from which the specific determinates, red, yellow, green, etc., emanate; while from shape emanates another completely different series of determinates. (1921, pp. 174–5) Second, determinables and determinates plainly differ in scope. Determinable properties are broader or more general than their corresponding determinates; determinate properties, narrower or more specific than their superordinate determinables. Determinables and determinates of a given family thus form a hierarchy of scope-inclusions, after the 209

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 210

d e t e r m inate / d eter minab le manner of colored and red, red and crimson, and crimson and Harvard crimson. In Johnson’s terminology, pairs of predicables belonging to such an hierarchy are related as (narrower) sub-determinates to (broader) super-determinates. Third, such determinables as shape, pitch, and colour are ultimately different, in the important sense that they cannot be subsumed under some one higher determinable, with the result that they are incomparable with one another. (1921, p. 175) Highest-level determinables in consequence mark out ultimate dimensions of comparison. They are qualitative respects in which objects are deemed to resemble and differ. The fourth and final mark of the determinable–determinate relation, however, is decisive. The relation is non-conjunctive. That is, the subset relationship obtaining between the extension of a pair of predicables related as super- to sub-determinate does not admit of explanation or analysis in terms of a third, differentiating property’s being conjoined with the super-determinate to restrict the resultant extension to that of the sub-determinate, as equal-sided may be conjoined with parallelogram to single out the narrower extension of rhombus or, classically, the differentium rational is conjoined with the genus animal to pick out the species man. In contrast, the only property that could be conjoined with, e.g., colored to yield a predicable having the extension of red is red itself. An item is not red by virtue of being both colored and F, for some property F distinct from red. Being red is not being something in addition to being colored; being red is rather a way of being colored. Essentially these features are singled out in the discussion of the determinable– determinate relation offered by Searle as well. Searle’s focus is on specificity. His chief concern is to distinguish the specificity relationship obtaining between super- and sub-determinate properties from those obtaining between genus and species, between a conjunction of diverse determinates and the determinable of one its members, between an arbitrary disjunction and one of its 210

disjuncts, and between a disjunction that includes a determinable and one of that determinable’s determinates. He concludes that a is a determinate of b just in case a is both non-conjunctively more specific than b and logically related to every c which is nonconjunctively more specific than b, where two terms are logically related if either entails the other or the negation of the other. Formally a is more specific than b provided that a entails b, but not conversely, and a is nonconjunctively more specific than b provided that, in addition, there are no properties c and d such that a is equivalent to (c & d ) and c, but neither d nor its negation, entails b. Historically, the notion of a family of determinate qualities falling under a common single determinable forms the framework of the eighteenth-century dispute between Locke and Berkeley on the topic of “general ideas”. Locke was interpreted as holding that one could form ideas of determinables that were not at the same time ideas of lowest-level (narrowest) determinates of those determinables for example, the (determinable) idea of a triangle that was not the idea of any determinate sort of triangle (equilateral, isosceles, right, scalene, etc.). Berkeley, in contrast, argued that even the most “general” ideas resembled particular images in necessarily being completely determinate, their “generality” deriving entirely from the manner in which they were considered and applied in comparative judgments of similarity and difference. On the contemporary scene, the general decline of interest in traditional questions of Platonistic metaphysics has led to a correlative scarcity of work specifically addressed to the relationships of inclusion, exclusion, subordination, and incompatibility obtaining among determinable and determinate qualities. Recognition of the special characteristics of those relationships, however, potentially carries with it interesting consequences for a variety of metaphysical concerns. The non-conjunctivity of the relationship of sub- to super-determinate qualities (at every level of generality), for example, is prima facie difficult to reconcile with the thesis of elementarism, that descriptive predicates of the second and higher types are

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 211

det ermi nism in principle eliminable in favor of predicates of the first type, a traditional stop on the road toward some forms of nominalism. Such higher-order determinable predicates as “is a color” and “is a shade of red” (true of, e.g., scarlet, crimson, carmine and maroon) at least do not readily lend themselves to any obvious reductive analysis in terms of first-order predicates. The “ultimate difference” of such highestlevel determinables as color and shape, on the other hand, suggests that they may play the role of descriptive (as opposed to pure or metaphysical) categories, thereby functioning to limit negative predications as much as positive ones. In particular, one can hold that both the positive and the negative predication of a determinate quality presuppose the correct ascription of its corresponding highest-level determinable, so that neither “red” nor “not-red”, for example, can be truly predicated of any item – e.g., an electron, the number seventeen – of which “colored” is not also correctly predicable. Finally, we may note the fact that the family of determinates under a common determinable at each level of specificity appears to consist of pairwise incompatible qualities. No spatially extended particular, for example, can co-instantiate both red and green or (more determinately) both crimson and scarlet at every point of its surface. Same-level determinates under a single determinable, that is, evidently (necessarily) exclude one another. If this is correct, however, it may ultimately prove illuminating to treat such a higher-order relationship of “quality exclusion” as the fundamental form of “negative fact”, underlying both predicate and propositional negation and supplying a basis in terms of which they can be analyzed and understood. See also Platonism. b i b l i og rap hy Johnson, W.E.: Logic, vol. 1 (Cambridge: Cambridge University Press, 1921: repr. New York: Dover Publications, 1964). Searle, J.: “On Determinables and Resemblance,” with S. Körner, Proceedings of the

Aristotelian Society, suppl. vol. 33 (1959), 141–58. jay f. rosenberg determinism The thesis that the world is deterministic is the thesis that the state of the world at one time “fixes” or “determines” the state of the world at any future time (or at both future and past times in some stronger versions of the claim). As such the thesis must not be confused either with the thesis of fatalism (that what will happen at a time is “destined” to happen, irrespective of what happens at some earlier time, in particular irrespective of an agent’s choices at that earlier time), nor with the claim sometimes made against those who deny “determinate reality” to the future (or to the past and the future) that reality is “timeless” in the sense that even what is not present in time and even what is future still has full reality. The notion of “fixation” or “determination” that is usually had in mind is this: from a total description of the state of the world at one time, and a specification of all of the laws of nature, a total description of the world at any other time can be derived by a purely logical, deductive, inference. Naturally, given the richness of magnitudes in the world, no claim that such descriptions could be given in any reasonably finitistic language is intended. This general idea of determinism is quite problematic, however, as many issues concerning what is to count as a state of the world, and of its full specification, and many issues concerning what is to count as a law of nature, arise. As Russell pointed out, a too liberal reading of what counts as a law of nature, one that lets any true general correlation between states at different times have lawlike status, would make the deterministic nature of the world trivially true, for there would, then, always be a law connecting the state at one time to the states at all other times so that the latter were fully fixed by the former, no matter what the world was like. Similarly, a too generous stance with what can count as a state at a time can trivialize the notion of determinism. If we let states at one time include reference to features of 211

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 212

d e t e r m inism some other time (by, for example, letting “is such that five minutes later the following magnitude holds of the following object” count as a state feature) then one would be able to infer the later state from the earlier even without references to laws of nature, no matter what the world was like. The necessity, familiar from other problem areas of philosophy, of restricting the notion of lawlike generalization to a proper subset of all true generalizations and of delimiting the appropriate features to be considered genuine occurrent physical properties of the world appears here as a necessary condition to avoid trivializing the notion of determinism. The idea that the world is deterministic if, and only if, it is “predictable in principle” has been common since the time of Laplace (1749–1827). But it is far from clear that such a tight association of determinism with predictability, even in principle, should be drawn. One can, for example, imagine worlds that are, intuitively, deterministic, but where some in-principle limitation on the ability of any cognizer to know the state of the world at one time is posited. Under these circumstances while future states would be fully determined by present states, the in-principle block on complete knowledge of the present state would make in-principle prediction impossible. On the other hand when the attempt is made to delimit the class of generalizations that are to be counted as laws of nature, various suggestions for necessary conditions for lawlikeness do have an epistemic tinge to them (that, for example, the functional connections of states to states be computable, for example). The questions involved when one asks if an idealized world described by some specified fundamental physical theory is one in which determinism holds or does not hold are fraught with great complexity. Even for such “simple” cases as Newtonian particle mechanics, issues involved in projecting motion through multiple particle collisions and in the possibility of particles entering an interaction by “coming in from infinity” in a finite time serve to make naive claims that a Newtonian particle world is deterministic less than straightforwardly acceptable. Much attention has been directed recently to 212

“chaotic systems”. Here while the future behavior of a system is, in some idealized sense, fully fixed by its initial state, there exist, arbitrarily close to that initial state, states that would lead to a radically different future evolution for the system. One could argue that such systems are deterministic, but not predictable, as noted above, or one could, as some theorists have, argue that the existence of such systems casts doubt on the reality of “exact” initial states as genuine features of the world as opposed to idealizations of theory that go beyond reality. Here, again, we see how important issues of what counts as legitimate physical state intertwine with issues of determinism. The variety of space–times allowed by General Relativity make issues of determinism even more complex (see space and time). Singularities in the space–time, space–times in which the global partitioning of the space– time into spaces at a time is impossible, and space–times that have closed causal loops, are all new possibilities that complicate the issue of whether or not a specified world ought to be characterized as deterministic. Finally there is the current quantumtheoretic picture of the world. The state attributed to systems by the theory, the system’s quantum state, is generally taken to observe an equation of evolution that is deterministic in nature. But, it is argued, the real states of the world, values observed upon measurement, are determined from the quantum state only probabilistically. The correlations between observed values at one time and those at some other time are then, apparently, non-deterministic. Further, there are a number of important theorems, socalled “proofs of the non-existence of hidden variable”, that are designed to show that the non-deterministic relation among observed values at different times cannot be generated out of a deterministic relation between the values of some “deeper” “hidden” parameter values at those times. Since that status of measurement in quantum mechanics is itself very problematic, as is the physical interpretation of the quantum states, it is impossible to argue without controversy that quantum theory, if true, shows, once and for all, the world to be non-deterministic. But the

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 213

dewey, john theory clearly describes a world in which determinism, if it does hold, will be a subtler matter than previously imagined. Traditionally philosophy has often contended with the issue of the alleged incompatibility of determinism and “free will” (see the extended essay). It is, however, far from clear that any understanding of the place of the notions of free agency, choice and will in human action will really hinge upon a showing that the world is non-deterministic in the sense intended above. At least it seems clear that indeterminacy in the world will not, by itself, provide a “place” for free will. An act generated by a spontaneous or non-determined physical happening seems as remote from our idea of an act generated out of free will as does one generated out of a determined physical happening. See also newton; space and time. b i b l i og rap hy Earman, J.: A Primer of Determinism (Dordrecht: Reidel, 1986). lawrence sklar

Dewey, John (1859–1952) American pragmatist. In Experience and Nature (1925). Dewey applies the empirical method to a rich notion of experience in order to develop an ontoloby that undercuts the many dualisms that plague philosophy and are his targets in his numerous other writings. Dewey objects to the dichotomies of classical and modern philosophy – of being and becoming, mind and matter, theory and practice, facts and values – not only because they give rise to sterile philosophical puzzles but because they reflect and perpetuate class distinctions between those who enjoy the life of the mind and those who must engage in physical labor. He speaks of the “hateful irony” (1925, p. 99) of a philosophy that exalts the life of reason while paying no attention to the conditions that make such a life possible. Experience is, for example, setting a watch, being pushed by the wind, listening to the

fourth Brandenburg Concerto, understanding how watches work. Only in reflection do we distinguish experiencing and that which is experienced, noting that experiencing is not restricted to knowing and that what is experienced is the rich world of our everyday lives. Metaphysics, which deals with the most general traits of existence, finds that existence is both stable and precarious, that it shows both recurrent features and individuality. There is for Dewey only one realm of Being, every existent is an event. Events have immediate, final qualities that are not known, though they stimulate the inquiry that leads to knowledge and the deliberate production of values. Events have beginnings and endings, they have and are causes and effects. Just as it is a mistake to regard (as did the Greeks) our world of becomings as inferior to a realm of Being (or of finalities), so it is a mistake to regard (as do the moderns) causes or the earlier parts of a history as more real than effects or the later parts of that history. Thus life, self-sustaining interactions between a thing and its environment, appeared later than inanimate matter (stable, recurrent orders of events), and minded behavior (interactions of an organism with others of its kind through speech) occurred even later; but ontologically they are on the same level. Though Dewey regards the life of the mind as the most conspicuous of nature’s ends, and as one of our highest goods, he warns against identifying ends (endings) with goods; qua endings they may be good, bad, or indifferent. Consciousness presupposes communication. A baby’s cry, an organic response, becomes a signal when it elicits a useful response from an adult but it is not yet language; language is present when one language user counts on another’s understanding and cooperation. Thus speech with others precedes speech with ourselves (thought), meanings are not psychic existences but primarily properties of cooperative behavior and secondarily of objects. Objects (in the first instance the things of our ordinary lives, but also the objects of the various sciences) are events with meanings. 213

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 214

d e w e y , jo hn “Meaning” is, however, in Dewey’s use multivocal. Things have meanings or even “essences” when we take them as signalling consequences that are important for us; things have meaning when they make sense. The existence of error shows that meanings are objective, they indicate possible interactions, hence mistakes are possible. Meanings are values; literature, ceremony, etc., provide the meanings in terms of which life is judged. What all these meanings share is that they are deliberately constructed means to our various ends, including that greatest human good, shared experience (1925, p. 159). Communication is thus both instrumental and final. There is no mind/body problem in the traditional sense, there is instead the distinction between routine behavior and intelligent behavior, i.e., ultimately a social problem. Thinking occurs when a situation is indeterminate, when the outcome depends on what we do. In a closed deterministic world, there would be no consciousness. It follows from the priority of communication to thought that human individuals are products of their society who bring to the simplest experience the habits and meanings they have been taught. Yet, when a situation is indeterminate, when it provokes thought, the result will be reconstruction. Though Dewey is not an idealist (see idealism) – there is an existence antecedent to knowledge – the object of knowledge is not that antecedent existence, it is what the knower makes of it, how it is taken, what meaning it is given. The same event may be known as a piece of paper, as a valuable historical document, or as something which will aid in making a fire. All individuals are products of their society, yet “every thinker puts some portion of an apparently stable world in peril and no one can wholly predict what will emerge in its place” (1925, p. 172). Dewey is profoundly aware of the fact that all change in social relations, for better or for worse, is due to individuals who question and challenge the existing order. In nature as a whole, in the arts, and in the life of the community, there is a tension between stability and spontaneity, between that which is predictable and 214

that which is original. Without either there would be no human flourishing. Experiences are either useful labor or consummatory, and in the best cases, the former are also the latter. One of Dewey’s great contributions to modern thought is the notion of a consummatory experience. Unless some enjoyments had come to us fortuitously, we would never inquire into their conditions, we would never attempt to secure or reproduce them. Values are the result of the critical evaluation and deliberate production of enjoyments that come to us naturally, i.e., values are natural events. Philosophical criticism heightens our appreciation of the goods of art, science and social companionship and makes us aware of their arbitrary distribution which prevents most human being from having “the richest and fullest experience possible” (1925, p. 308). Although Dewey distinguishes between metaphysics and morality, the motivation for his unitary ontology is his moral commitment to a humane and liberal society. See also identity; pragmatism. wri t i ngs The collected works of Dewey have been published in three series (The Early Works, The Middle Works, The Later Works), ed. J.A. Boydston (Carbondale, IL: Southern Illinois University Press). Experience and Nature (1925) is Vol. I of The Later Works (1981). bi bliography Rorty, R.: “Dewey’s Metaphysics,” in his Consequences of Pragmatism (Minneapolis: University of Minnesota Press, 1982), 72–89. Santayana, G.: “Dewey’s Naturalistic Metaphysics,” in Dewey and His Critics, ed. S. Morgenbesser (New York: The Journal of Philosophy, 1977), 343–58. Sleeper, R.W.: The Necessity of Pragmatism (New Haven, CT and London: Yale University Press, 1980). ruth anna putnam

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 215

disposi t i on discrete

see continuous/discrete

disposition A tendency to be or to do something. Fragility, solubility, elasticity, ductibility and combustibility are all dispositions. Fragile things tend to break when struck; water-soluble things tend to dissolve when immersed in water. A type of thing a disposition is a tendency to be or to do is a manifestation of the disposition. Thus, breaking is a manifestation of the disposition of fragility; dissolving is a manifestation of the disposition of solubility. Dispositions have activating conditions. Striking a fragile object can activate the disposition of fragility; immersing a water-soluble object in water can activate the disposition of water-solubility. A disposition can have many types of manifestations and many types of activating conditions. Both cracking and shattering, for example, can manifest the disposition of fragility, and both being struck and being dropped can activate it. To activate a disposition, an activating occurrence must cause a manifestation of the disposition in “the right sort of way”, where that way varies from disposition to disposition (Prior, 1985, pp. 9–10; Smith, 1985). For example, for a striking to activate the disposition of fragility in an object, it must break the object. To do that, it must cause the object to break. But not just any way of causing an object to break counts as breaking the object. If one person’s striking an object causes another person to become angry and kick the object in such a way as to break it, then the first striking of the object did not break the object, even though it was a cause of the object’s breaking; the kick, not the first strike, broke the object. The question arises as to how a striking must cause something to break to make it the case that the striking broke the thing. This is a problem of “deviant” or “wayward” causal chains, a problem that remains unresolved. It is fairly widely held that dispositional properties are counterfactual properties (Prior, 1985, pp. 5–10; Ryle, 1949, p. 43). Possession of a disposition is satisfaction of a (perhaps complex) counterfactual condition. The view is typically formulated as a thesis

about dispositional terms: dispositional terms can be defined by counterfactual sentences. On this view, the property of water-solubility is (roughly) expressed by the counterfactual “were x immersed in water, x’s immersion would (at least begin to) dissolve x”. Thus, something is water-soluble if, and only if (roughly) it is such that were it immersed in water, its immersion in water would (at least begin to) dissolve it. Those who hold this view divide into the phenomenalist camp and the realist camp (Mackie, 1973, p. 142). The phenomenalist camp denies that to possess a disposition an object or substance must possess some other property in virtue of which it has the disposition, that is, in virtue of which it satisfies the counterfactual condition (Ryle, 1949, p. 43). The phenomenalist denies that water-soluble things must have some property in virtue of which they are such that were they immersed in water, their immersion would (at least begin to) dissolve them. On the phenomenalist view, it is possible for two things be exactly alike except that one does and the other does not satisfy this counterfactual condition. There need not be any other difference between the two things in virtue of which the one does and the other does not satisfy the condition. The realist view of dispositions claims that things have dispositions in virtue of having other properties. The properties in virtue of which things possess dispositions are bases or grounds for the disposition. Labelling it the “realist” view is, however, somewhat misleading. For proponents of phenomenalism hold that there really are dispositions. Moreover, proponents can allow that a disposition has a basis; they need deny only that dispositions logically or metaphysically require bases (Prior, 1985, p. 29). According to the realist view, dispositions must have bases (Armstrong, 1986, pp. 87–8). But realists divide over whether dispositions must have bases which are intrinsic properties of the things that possess the disposition (Prior, 1985, ch. 4). It has been claimed that it is at least logically possible for the only basis for a disposition to be an historical property, such as, for example, the property of having been produced by a 215

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 216

d i s p os i tio n certain process (Mackie, 1973, p. 131; Tooley, 1972, p. 287). This claim implies that it is logically possible for two things to have exactly the same intrinsic properties yet for the one to possess a certain dispositional property and the other to lack that property, solely in virtue of differences in their causal histories. At best, however, this would show only that it is logically possible for a disposition to fail to have an intrinsic basis. But a reason that might be given for the thesis that dispositions typically do not have intrinsic bases is this: whether an object or substance will manifest a disposition typically depends not only on the thing’s intrinsic properties and whether an activating event occurs, but also on surrounding circumstances. Glass, for example, would not shatter were it struck when encased in protective covering; steel would shatter were it struck under extremely cold temperatures (Smith, 1977, pp. 441–3), Activating occurrences for dispositions will activate the dispositions only under certain circumstances, circumstances that are standing conditions for the manifestations of the disposition. Thus, it might be held that the bases for dispositions typically include extrinsic properties. This reason for holding that dispositions typically lack purely intrinsic bases can, however, be rejected. Nothing is, for example, fragile period; things are only fragile under certain circumstances. For example, steel is fragile under extremely cold temperatures, though it is not fragile under ordinary room temperatures (Prior, 1985, pp. 46–9). We ordinarily say simply that glass is fragile and that steel is not because it is understood that we are talking about fragility under certain ordinary conditions. On this view, dispositional predicates such as “is fragile” are incomplete predicates. The complete predicates for “is fragile” will be ones of the form “is fragile under conditions C” (Prior, 1985, pp. 8–9). An intrinsic property of a certain kind of object may be a basis for fragility under C but not a basis for fragility under C*, a distinct standing condition from C. Realists also divide over whether dispositions must ultimately have categorical bases (Armstrong, 1986, pp. 87–8; Prior, 1985, ch. 5). Categorical bases are properties (or 216

states) that are not themselves dispositional properties (or states). The view that dispositions must have categorical bases implies that dispositional properties are not fundamental properties: they are invariably possessed in virtue of the possession of nondispositional, non-counterfactual properties. Dispositions, as noted above, are often taken to be counterfactual properties. The counterfactual properties in question are, some realists claim, “potentialities” that must be ultimately grounded in “actualities”, namely categorical bases. Categorical properties are often understood to be intrinsic properties. Thus, the strongest version of the thesis that dispositions must have categorical bases is that it is metaphysically necessary that if something has a disposition, then it has some non-dispositional, non-counterfactual, intrinsic basis for the disposition. This strongest version of the thesis is, however, controversial. Finally, realists also divide over the relationship dispositions bear to their bases (Prior, 1985, ch. 6). One view has it that when something possesses a disposition, the disposition is the basis for it in the thing in question. One problem for this view is that a disposition can have multiple bases. The basis of fragility can vary with the kind of thing in question; the basis may be a certain crystalline structure in one kind of object and a different crystalline structure in another. Indeed, a given disposition may have more than one basis within the same substance. A certain piece of cloth may have two bases for being water-absorbent: its threads may be made of water-absorbent material, and the cloth may be weaved in such a way that it absorbs water between its inner threads (Mackie, 1973, p. 148). Since the bases in question are not identical, they cannot both be identical with the property of water-absorbency. Moreover, concerns about multiple bases aside, a property will be a basis for a disposition only relative to the laws of a world. But sameness of extension in every possible world is at least a necessary condition for property identity. Thus, a property may be a basis for a disposition in our world and yet fail to be a basis for the disposition in a logically possible world in

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 217

dretske, fred which different laws are operative. The disjunction of the bases (in our world) for a disposition will not be coextensive with the disposition in every possible world. The leading theory of dispositions today is the functionalist theory, a realist theory according to which a disposition is a secondorder state, a state of having a state with a certain causal role (Lewis, 1986a, pp. 223– 4; Prior, 1985, ch. 7; Prior et al., 1982). The causal role will consist of the first-order state’s (appropriately) causing the manifestation(s) of the disposition in response to appropriate activating conditions under certain standing conditions. The bases are realizations of the dispositions: they are the first-order states with the relevant causal roles. On this view, a state can be dispositional relative to certain states and non-dispositional relative to others since it can be secondorder relative to certain states and firstorder relative to others. Dispositions will ultimately have non-dispositional bases only if some states are not second-order relative to any others. The basis of a disposition is a cause of the manifestation(s) of the disposition. But some functionalists deny that types of disposition states are causally relevant to the causation of manifestation(s) of the dispositions (Lewis, 1986a, pp. 223–4; Lewis, 1986b, p. 268; Prior, 1985; Prior et al., 1982). However, be this as it may, it seems that we can at least causally explain a manifestation of a disposition by citing the disposition itself. We say that the substance dissolved because it is soluble, or that it shattered because it is fragile. The explanation in the case of solubility, for example, seems to amount to this: the thing dissolved because it is in a state that makes it dissolve when immersed in water. This explanation excludes other hypotheses about why the thing in question dissolved (Block, 1990, pp. 162–3). See also counterfactuals; law of nature; possible worlds. b i b l i og rap hy Armstrong, D.M.: A Materialist Theory of Mind (London: Routledge, 1968).

Block, N.: “Can the Mind Change the World?,” in Meaning and Method: Essays In Honor of Hilary Putnam, ed. G. Boolos (Cambridge: Cambridge University Press, 1990), 137–70. Lewis, D.: “Causal Explanation,” in his Philosophical Papers, vol. II (Oxford: Oxford University Press, 1986[a]), 214– 40. Lewis, D.: “Events,” in his Philosophical Papers, vol. II (Oxford: Oxford University Press, 1986[b] ), 241–69. Mackie, J.L.: Truth, Probability and Paradox (Oxford: Oxford University Press, 1973). Prior, E.: Dispositions, Scots Philosophical Monographs (Aberdeen: Aberdeen University Press, 1985). Prior, E., Pargetter, R., and Jackson, F.: “Three Theses About Dispositions,” American Philosophical Quarterly 19 (1982), 251–7. Ryle, G.: The Concept of Mind (London: Hutchinson, 1949). Smith, A.D.: “Dispositional Properties,” Mind 86 (1977), 439–45. Tooley, M.: “Armstrong’s Proof of the Realist Account of Dispositional Properties,” Australasian Journal of Philosophy 50 (1972), 283–7. brian p. mclaughlin Dretske, Fred (1932– ) is an American philosopher who has made significant contributions to philosophy of mind and epistemology among other areas. He has taught at the University of Wisconsin, Stanford University, and Duke University. Dretske’s work is unified by a desire to provide naturalistic accounts of perception, knowledge, and meaning. Philosophical naturalism is multi-faceted, including epistemological, ontological, and conceptual varieties among others. In epistemology, naturalism has been identified with everything from Quine’s radical claim that we should replace traditional epistemology with psychology to the more modest idea that epistemology is not autonomous and requires substantive help from the sciences. Ontological naturalism is the view that only natural entities (and properties) exist and is often indistinguishable from materialism or physicalism. Conceptual 217

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 218

d u m m e tt, michael naturalism is the view that philosophical theories should not contain any unreduced intentional or normative concepts in their formulation. Dretske’s naturalism emphasizes external features such as informationcarrying/sustaining causal connections and functional roles. To this end, Dretske has developed the distinction between epistemic and non-epistemic seeing, proposed an information-theoretic account of propositional knowledge, and defended an externalist functional role theory of semantic content and mind. He has also made important contributions to the debate over laws of nature, causation, and epistemic closure principles. See the extended essay on causation; see also naturalism; physicalism, materialism; law of nature; intentionality. w r i t i n gs Explaining Behavior: Reasons in a World of Causes (Cambridge, MA: MIT Press, 1988). Knowledge and the Flow of Information (Cambridge, MA: MIT Press, 1981). Naturalizing the Mind (Cambridge, MA: MIT Press, 1995). Perception, Knowledge and Belief (Cambridge: Cambridge University Press, 2000). Seeing and Knowing (Chicago: University of Chicago Press, 1969). richard gallimore Dummett, Michael (1925– ) Michael Dummett has been influential in recent Anglo-American analytical philosophy by putting the philosophy of language at center stage, and advocating a revisionist philosophy of logic. He has established himself as perhaps the foremost commentator and interpreter of Frege and the later Wittgenstein. His most important contribution is the thesis of semantic antirealism. On Dummett’s formulation, this thesis, for any given region of discourse, is that we do not have to regard every declarative statement of that region as determinately true or false independently of our means of coming to know what its truth value is. That is, the semantic antirealist refuses to accept the principle of bivalence for the region of 218

discourse in question; and accordingly has to revise classical logic, which embodies that principle. Dummett emphasizes the manifestation requirement, to the effect that one’s grasp of any aspect of meaning should in principle be capable of being manifested in one’s observable behavior. This behaviorist principle, which he draws from the work of the later Wittgenstein, is then used to argue for the legitimacy of intuitionistic, as opposed to classical, meanings of the logical operators. The thesis of semantic antirealism had its origins in his classic paper “Truth” (1959). There it concerned the determinacy of truth value of claims about other minds. The thesis has been extended to affect statements about the past, the future, counterfactual conditionals and, most importantly, mathematical statements. In the case of mathematics, Dummett’s major contribution has been a new meaningtheoretic foundation for intuitionistic logic and mathematics, based on considerations of reducibility and harmony drawn from modern proof theory. The loci classici here are his papers “The Philosophical Basis of Intuitionistic Logic” (1973c) and “The Justification of Deduction” (1973b), and his monograph Elements of Intuitionism (especially the section “Concluding Philosophical Remarks”). Other important ideas in contemporary theory of meaning derive from Dummett’s work, which represents a synthesis of the lasting insights of Frege and Wittgenstein. In his monumental studies on Frege and his highly influential two-part essay “What Is a Theory of Meaning?” (1975, 1976) he argued for molecular, as opposed to atomistic or holistic, theories of meaning; and for full-blooded, as opposed to modest, theories. A molecular theory takes the meanings of sentences as primary. A full-blooded theory aims at an explanatory reduction of semantic notions. In a theory of meaning, a theory of force will form a shell around a theory of sense. The aim of the theory of sense is to characterize the assertability conditions of sentences, which are the primary bearers of meaning. Although sentences express thoughts, and thought would be impossible

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 219

duns scotus, john without language, thoughts do not, according to Dummett, subsist in a Fregean third realm. The characterization of the senses of sentences will nevertheless proceed in a compositional manner, as first articulated by Frege. From Wittgenstein we draw the lesson that meaning is public; language is essentially social. This is what grounds the norms of reasoning and logic. Dummett has made an enduring contribution in pursuing these lines of thought to revisionist conclusions in metaphysics and the philosophy of logic and mathematics. See also intuitionism mathematics.

logic

and

writings Elements of Intuitionism, with the assistance of R. Minio (Oxford: Oxford University Press, 1977). Frege: Philosophy of Language (London: Duckworth, 1973[a] ). Frege: Philosophy of Mathematics (Cambridge, MA: Harvard University Press, 1991). The Interpretation of Frege’s Philosophy (London: Duckworth, 1981). “The Justification of Deduction” (1973[b]); repr. in Truth and Other Enigmas (London: Duckworth, 1978), 290–318. The Logical Basis of Metaphysics (Cambridge, MA: Harvard University Press, 1991). “The Philosophical Basis of Intuitionistic Logic” (1973[c] ); repr. in Truth and Other Enigmas (London: Duckworth, 1978), 215–47. “Truth” (1959); repr. in Truth and Other Enigmas (London: Duckworth, 1978), 1– 24. Truth and Other Enigmas (London: Duckworth, 1978). “What Is a Theory of Meaning (I)?,” in Mind and Language, ed. S. Guttenplan (Oxford: Oxford University Press, 1975), 97– 138. “What Is a Theory of Meaning (II)?,” in Truth and Meaning, ed. G. Evans and J. McDowell (Oxford: Clarendon Press, 1976), 67–137. neil tennant

Duns Scotus, John (c.1265–1308) One of the most influential and respected of the medieval scholastic theologian/philosophers (known to subsequent generations of scholastics as the “Subtle Doctor”), was born probably in the town of Duns in Scotland. He is known to have studied at Oxford before going to Paris, the chief center of learning in Europe at the time, where he encountered the radical Augustinian Henry of Ghent (d. 1293), Godfrey of Fontaines (thirteenth century), and others. Around 1300 he was back lecturing in Oxford for a spell, and on his return to Paris he became involved on the side of Pope Boniface (1235–1303) in his quarrel with Philip the Fair. Political troubles eventually forced him to desert Paris for Cologne, where he died. Given the shortness of his life the amount of written work he produced is remarkable. There is the usual commentary on the Sentences of Peter Lombard in several versions, Questions on Logic, Questions on Aristotle’s De Anima, On the First Principle (perhaps the most elaborate of the many scholastic attempts to prove the existence of God), Most Subtle Questions on Aristotle’s Metaphysics, and a set of Quodlibetal Questions, as well as some less significant treatises. Scotus was in the Franciscan Order from an early age and participated in the late scholastic effort to reinterpret the then dominant Aristotelian–Arab philosophical tradition in a way that made room for the truth of the basic dogmas of orthodox Christian theology. The Franciscans, influenced as they were by the Augustinian tradition, were more willing than most to deviate from the philosophic norms of the day in order to save theology, and Scotus’s work is perhaps best viewed as a reworking of the Aristotelian and Avicennian materials he inherited to produce a philosophic framework within which the difficult doctrines of Christian dogma, such as the Trinity, the Incarnation, divine foreknowledge and providence, etc., could be understood. Although Scotus always thought of himself as preserving the basics of Aristotelian philosophy, in metaphysics the result of his work was a radical revision of the Aristotelian program. 219

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 220

d u n s s c o tus, jo hn Where Aristotle had argued that being was not univocally predicated of things in different categories (e.g., substance, quantity, quality, etc.), Scotus, after some hesitations, claimed in his later works that there is a univocal concept of being applicable across the Aristotelian categories and applicable both to God and creatures. This doctrine allowed Scotus to treat metaphysics as a science of “being qua being” in a much stricter sense than Aristotle had. He developed a theory of “transcendental” terms, i.e., concepts which like “being” applied cross-categorically and treated metaphysics as the study of these. Scotus also took Aristotle’s notions of distinctness in number and distinctness in being or form and developed out of them several notions of identity and distinctness. Entities that are at least by divine power separable, i.e., at least one of them can exist even if the other does not, are “absolutely really” distinct, and those that cannot be so separated are “absolutely really” the same. But within the class of entities that are “absolutely really” the same we find pairs of entities which are “qualifiedly” distinct. For example, where a and b are absolutely really the same but each is definable independently of the other, a and b are “formally” distinct. This distinction is, as Scotus says, “on the side of the thing”; it is not a mere conceptual distinction. Since formal distinction is compatible with “real” sameness, formal sameness turns out to be more restrictive than “real” sameness. Even among entities that are formally the same Scotus allows for certain kinds of distinction, including “mental” distinctions which cannot occur prior to the entities’ being thought of. The principle that whatever is true of some subject is true of anything the same as that subject, holds, in Scotus’s view, unrestrictedly only of items which are in no way distinct, a kind of logical sameness which Scotus does not think of as a real relation. Certainly items which are only absolutely really the same can differ in respect of what is true of them. On the vexed topic of universals Scotus improved on the solution offered by Avicenna. He believed that there were entities that did not of themselves have 220

numerical unity but only a unity less than that, i.e., either specific unity or generic unity. These entities, sometimes called common natures, could be considered in three ways: (1) in themselves where each of them was neither existent nor non-existent, neither one nor many; (2) as they exist in particular realities and where each of them is numerically many; and (3) as objects of cognitive acts and states where they have “objective” existence and numerical unity (see objectivity). It is only in the last way that these natures are subjects for logical predicates such as universality. Scotus is particularly emphatic in his belief that whatever actually exists is in fact either one or many individual things, although this oneness or manyness belongs to common natures only accidentally (see universals and particulars). The individuation of a common nature is accomplished not by its enmatterment, as had been the view of many Aristotelians, but by an ultimately unknowable entity which determines a specific common nature to a particular individual in the way a specific difference determines a genus to one of its species. This individuating difference is in the individual in question really the same as the common nature it determines but nevertheless formally distinct from it. His theory of individuators enabled Scotus to treat Aristotelian natural forms as individuated prior to their belonging to matter and thus as much better candidates for human souls in Christian theology. Finally, Scotus made a fundamental break with the classical tradition in claiming that free, contingent causes, i.e., those that were not determined to bring about what they in fact do, could be superior in ontological perfection to the necessary causes with which Greek philosophy had populated the divine and eternal realm. The result was an acceptance of the basic contingency of the world and of the indeterminacy of causal chains in it which was quite foreign to Aristotelians. wri t i ngs Duns Scotus, Philosophical Writings, trans. and ed. A. Wolter, OFM (New York: Nelson, 1962).

9780631199991_4_P2004.qxd 1/12/09 3:06 PM Page 221

duns scotus, john God and Creatures: The Quodlibetal Questions, trans. F. Alluntis, OFM and A. Wolter, OFM (Princeton, NJ: Princeton University Press, 1975). Opera omnia, ed. P.C. Balic (Vatican City: Vatican City Press, 1950– ); far from complete but includes most of the commentary on the sentences in two versions known as the Ordinatio and the Lectura respectively. Opera omnia, 26 vols. (Paris: Vives, 1891–5); reproduces most of the Lyons edition (ed. Wadding) of 1639; most of Scotus’s known works are included but also some spurious ones. A Treatise on God as First Principle, trans. and ed. A. Wolter, OFM (Chicago: Franciscan Herald Press, 1966).

bibl i ography Balic, P.C., OFM: “The Life and Works of John Duns Scotus,” in John Duns Scotus, 1265– 1965, ed. J.K. Ryan and B.M. Bonansea (Washington, DC: The Catholic University of America Press, 1965), 1–27. Bettoni, E., OFM: Duns Scotus: The Basic Principles of his Philosophy, trans. B. Bonansea, OFM (Washington, DC: The Catholic University of America Press, 1961). Clatterbaugh, K.C.: “Individuation in the ontology of Duns Scotus,” Franciscan Studies 32 (1972), 65–73. Jordan, M.J.: Duns Scotus on the Formal Distinction (Ann Arbor, MI: University Microfilms International, 1984). Wolter, A., OFM: The Transcendentals and Their Function in the Metaphysics of Duns Scotus (Washington, DC: The Catholic University of America Press, 1946). martin m. tweedale

221

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 222

element One constant throughout the early history of philosophy and the later history of natural science has been a simple procedural one: when faced with a complex whole, resolve it into its constituent parts, its “elements”. The natural philosophers of Ionia sought to determine the “elements” (stoicheia) of which all bodies are composed or from which these bodies originally derived. Empedocles (c.495–c.435 bc) suggested that the elements are fire, air, earth, and water, each with its own natural place and characteristic set of qualities; this fourfold division was widely accepted by later philosophers. Euclid (fl. c.300 bc) titled his pioneering work on geometry, The Elements; he had shown how all of the multiplicity of propositions regarding plane figures could be derived from a small set of simple definitions and axioms. Medieval theories of method spoke of analysis and synthesis, or of resolution and composition, and recommended the breaking down into elements and subsequent reconstituting of the original complexes as the primary mode of understanding. Boyle built his new chemistry in the 1660s around the distinction between elements and compounds, each having distinctive chemical properties. Chemical elements are simple and unmixed, incapable of resolution into other bodies. It proved much more difficult than Boyle had expected to determine which was element and which was compound. Only a century later did the analytical methods of Lavoisier (1743–94), based on precise weight measurement, yield the first table of chemical elements. Dalton’s (1766–1844) atomic theory linked elements with atoms possessing distinctive chemical properties; Avogadro (1776–1856) explained compounds in terms of molecules; Berzelius 222

(1779–1848) successfully distinguished elements on the grounds of relative atomic weight. A chemical element is defined today in terms of the number of protons in the nucleus of its atom; its chemical properties depend on the normal number of electrons in its outer shell. Since the number of neutrons in the nucleus of an element may vary, the element may have different “isotopes” depending on the number of such neutrons, or (which amounts to the same thing) on the total atomic weight. See also Presocratics. bibl i ography Hintikka, J. and Remes, U.: The Method of Analysis: Its Geometrical Origin and Its General Significance (Dordrecht: Reidel, 1974). Kirk, G.S., Raven, J.E., and Schofield, M.: The Presocratic Philosophers, 2nd edn. (Cambridge: Cambridge University Press, 1983). Knight, D.: Ideas in Chemistry: A History of the Science (New Brunswick, NJ: Rutgers University, 1992). ernan mcmullin empiricism Empiricism is a broad tendency in the theory of knowledge. Each of its many forms lays stress on experience (typically sensory experience) as a source of knowledge or belief. The Greek equivalent of “empiricism” was first used nearly 2,000 years ago by Galen (ad 129–199), who argued that medical knowledge was solely a matter of experience. Empiricism has played an important role in philosophy ever since, but the influence of the “British empiricists”

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 223

empiricism of the seventeenth and eighteenth centuries has been especially enduring. Empiricism, on their account, makes two complementary claims, one concerning the content of thought (here called content-empiricism), and another concerning the justification of belief (justification-empiricism). According to the claim concerning content, experience is the ultimate source of all of our conceptions. As Locke, for example, put it, all of our ideas – the only “immediate object[s] of Perception, Thought, or Understanding”, according to him (Essay II viii 8) – spring from experience. “Our Observation employ’d either about external, sensible Objects; or about the internal Operations of our Minds”, Locke held, “is that, which supplies our Understandings with all the materials of thinking” (Essay II i 2). According to the claim concerning justification, experience is the only source of evidence for our beliefs, or the only source for those beliefs that are nonanalytic, “factual” or informative. As Hume for example, wrote, “it is only experience, which . . . enables us to infer the existence of one object from that of another” (Enquiries, p. 164) and the existence of the other, if it is not inferred, can be known only by sensation or reflection. Neither content-empiricism nor justification-empiricism is, on its face, a metaphysical claim, but each has metaphysical bearing. Content-empiricism tends to limit the extent of legitimate metaphysical thinking. We can genuinely think of x, it tells us, only if we can derive a conception of x from experience. If a conception of substance, for example, cannot be so derived, then we cannot really think of it. On the assumption that an expression is meaningful only if a conception lies behind it (an assumption endorsed by some empiricists but denied by others), it follows that we cannot meaningfully speak of substance. Justification-empiricism tends to limit the extent of justified (or justifiable) metaphysical belief. Even if we can conceive (and meaningfully speak) of substance, we can justifiably believe in it only if experience warrants. Empiricist constraints have been applied not only to the conceptions of professed

metaphysicians (substance, universals, the external world), but to the conceptions of religion (God, the soul, freedom of the will), science (atoms or corpuscles, physical forces, absolute space and time), and everyday life. Empiricist metaphysicians have responded to these constraints in three broad ways: by refusing to endorse certain metaphysical claims, at times disavowing the very possibility of “metaphysics”; by positively denying certain claims, and banning (from their systems of the world) the entities those claims presuppose; and by reconstructing suspect conceptions or beliefs, so as to protect them from empiricist criticism. The third response calls not only for a clear sense of what is “given” in experience (as do the first two), but for an identification of those constructions or routines that add to the stock of available conceptions without violating contentempiricism. Locke and Hume spoke, for example, of combining ideas into larger wholes; combination and recombination were, they thought, acceptable routines. Empiricists in the early twentieth century spoke not of combination but of logical construction, claiming that in certain cases, sentences expressing initially suspect conceptions could be derived, by deductive means (logical principles together with definitions), from more basic sentences free of suspicion. Their successors were more liberal, allowing a conception to be genuine (or an expression to be meaningful) if claims in which it figured could be confirmed, inductively, by observation-reports. There is a fourth and perhaps more elusive response to empiricist constraints, in which a suspect claim or conception is not reconstructed but reclassified. An apparent claim of fact becomes a tool for coping with experience; a word appearing to designate a property becomes (part of ) an expression of emotion or allegiance (see objectivism and projectivism). As Berkeley wrote, “the communicating of ideas marked by words is not the chief and only end of langauge, as is commonly supposed. There are other ends, as the raising of some passion, the exciting to, or deterring from an action, the putting the mind in some particular disposition” (Works, vol. 2, p. 37). The history of empiricism illustrates all four of the 223

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 224

e m p i r i c ism responses sketched here: skeptical unbelief, outright denial, reconstruction and reclassification. l oc k e According to Locke’s influential statement of content-empiricism, all ideas are derived from sensation (the mind’s perception of external objects) or reflection (the mind’s perception of its own operations). The mind is originally “void of all Characters” (Essay II i 1); hence no idea or notion is innate. But the mind can transform what experience supplies: it can break ideas down into their simple parts; it can arrange the parts in new ways; it can arrive at ideas of infinite space and duration by enlarging ideas of finite intervals (and by seeing, in a manner Locke does not elaborate, that the process of enlargement needn’t end); and it can frame ideas of universals by abstracting traits from resembling particulars. We have, Locke argued, no clear idea of substance, because no such idea can be derived by any of these means. Our only idea of substance is obscure and relative: a supposition of “we know not what” supporting qualities. Locke, as Berkeley later observed, “banter’d” – mocked or ridiculed – the idea of substance. “Here,” Locke wrote, “as in all other cases, where we use Words without having clear and distinct Ideas, we talk like Children” (Essay II xxiii 2). But Locke held none the less that there are both corporeal and incorporeal substances, leaving his successors to wonder whether a consistent empiricist can do so. The obscurity of the idea of substance convinced Locke that some metaphysical topics could not be usefully investigated. He therefore tried to detach certain practical questions from their metaphysical underpinnings, proposing, for example, that personal identity depends not on identity of substance, but on continuity of consciousness (see persons and personal identity). Leibniz objected that “for all its apparent thinness”, the idea of substance is “less empty and sterile” than Locke supposed. Leibniz’s objection was encouraged by his belief that the idea of substance is innate. This belief 224

admits of many interpretations, but on one rather modest reading (suggested by Leibniz himself), it means merely that reflection enables us to find, within ourselves, a metaphysically fertile idea of substance. Locke denied even this, because reflection, he thought, reveals no more about substance than sensation does. Leibniz sometimes portrayed Locke as believing that justification always rests on experience (on “induction and instances”, as Leibniz put it), but Locke’s views are more complex. Locke believed that mathematical knowledge is a priori, even though it is not uninformative or “trifling”. “We may be certain in Propositions, which affirm something of another”, Locke wrote, “which is a necessary consequence of its precise complex Idea, but not contained in it” (Essay IV viii 8). But our beliefs about nature rest entirely on experience. This renders them uncertain, and because knowledge (for Locke as for Descartes) requires certainty, our knowledge of nature is very slight. We can, however, have more or less well-supported beliefs about a wide range of things. It is likely, for example, that bodies are systems of tiny corpuscles, too small to be seen or felt. The “corpuscular hypothesis” provides the best explanation of the changing world and our perception of it. There is, Locke believed, a necessary connection between a body’s “real essence” (its corpuscular constitution) and its observed or manifest qualities. If we had insight into that essence – if we had, for example, microscopical eyes – we could predict those qualities without trial. But in our present state such insight is beyond us. berkeley Berkeley used content-empiricism to argue against the very possibility of matter or material substance. Talk of matter (or of existence “without the mind”) is, he argued, either contradictory or empty. “As to what is said of the absolute existence of unthinking things without any relation to their being perceived,” he wrote, “that seems perfectly unintelligible” (Works, vol. 2, p. 42). Berkeley retained belief in spiritual substance; we have, he insisted, a “notion” of

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 225

empiricism it, derived from reflection on our own selves. Bodies undoubtedly exist, Berkeley argued, because they are nothing more than collections of ideas of sense. These ideas are “more strong, orderly, and coherent than the creatures of the mind” (ibid., p. 54). The distinction between reality and illusion is not merely preserved, but rendered empirical. Early modern empiricists, like their rationalist contemporaries, denied the existence of abstract universals. “All things that exist”, as Locke proclaimed, “are only particulars” (Essay III iii 6). How then do we think in universal terms? According to Locke we form abstract ideas. We make ideas general, he explained, “by separating from them the circumstances of Time, and Place, and any other Ideas, that may determine them to this or that particular Existence” (ibid.). Berkeley, followed by Hume, denied that abstraction is possible. It is a received axiom, Berkeley wrote, “that an impossibility cannot be conceiv’d”. Because “nothing abstract or general” can exist in the world, “it should seem to follow, that it cannot have so much as an ideal existence in the understanding” (Works, vol. 2, p. 125). We think in universal terms not by forming abstract or indeterminate ideas, but by attending to selected parts or aspects of determinate ones.

to see how we could come to have one. “Philosophers begin to be reconcil’d”, Hume wrote, “to the principle, that we have no idea of external substance, distinct from the ideas of particular qualities. This must pave the way for a like principle with regard to the mind, that we have no notion of it, distinct from particular perceptions” (Treatise, p. 635). Hume denied the suggestion (made among others by Locke and Berkeley) that an active mind presides over the ideas that pass within it. Ideas succeed one another according to impersonal associative principles (see associationism). This is an aspect of Hume’s Naturalism, his conviction that human nature is continuous with nature as a whole. And like the rest of nature it should, he urged, be studied empirically. According to Hume’s justification-empiricism, the objects of human reason or inquiry are either relations of ideas or matters of fact. Relations of ideas are discoverable a priori, “without dependence on what is anywhere existent in the universe” (Enquiries, p. 25), because it is a contradiction to deny them. Matters of fact, including all existence claims, can be known only by experience. God’s existence, therefore, cannot be established a priori. Empirical arguments for God’s existence cannot be dismissed so readily, but they too are unsuccessful.

hume

kant and his infl uence

The mind’s perceptions, Hume held, are either impressions or ideas. Impressions are the forceful perceptions we have when we sense and feel; ideas the more feeble perceptions we have when we think and reason. According to Hume’s statement of contentempiricism, all ideas are copies of impressions. “When we entertain . . . any suspicion that a philosophical term is employed without a meaning or idea”, Hume wrote, “we need but enquire, from what impression is that supposed idea derived?” “If it be impossible to assign any,” he concluded, “this will serve to confirm our suspicion” (Enquiries, 8, 22). Substance as the support of qualities is an “unintelligible chimera”. We have no impression of substance, and it is difficult

Kant rejected the content-empiricism of his predecessors: “though all knowledge begins with experience”, he wrote, “it does not follow that it all arises out of experience” (Critique, p. 41). Certain concepts or categories, among them substance and cause and effect, are necessary conditions for the possibility of experience. It follows that experience cannot be their source. Kant also rejected justification-empiricism. The categories enter into truths which are, be held, both synthetic and a priori. (“Every event has a cause” is one example.) But these truths apply, he insisted, only to the objects of possible experience. Kant was largely responsible for an important shift in the way content-empiricism 225

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 226

e m p i r i c ism was understood. Locke, Berkeley and Hume were concerned with what might be called the genesis of content. The human mind, they suggested, is at first content-free; it is later stocked with conceptions by experience. In the basic case the mind is passive, a receptacle waiting for whatever experience deposits. Kant questioned whether any idea or concept could be acquired in this way. Post-Kantian empiricists, persuaded by his arguments, turned their attention from genesis to analysis. They worried less about the coming-to-be of conceptions, and more about their empirical import. At about the same time, and for related reasons, philosophers came to believe that whole thoughts or propositions are more revealing, as objects of analysis, than isolated concepts or expressions. Mill is one nineteenth-century empiricist who shows the influence of these developments. When he defined “matter” as “a permanent possibility of sensation”, he did not describe how to generate the concept from sensations. He tried instead to show that for every statement or thought about material objects, there is “an equivalent meaning in terms of Sensations and Possibilities of Sensation alone”. The analytic emphasis of empiricism grew more pronounced in the twentieth century. james James, who dedicated Pragmatism to Mill’s memory, described his pragmatic method as “a method of settling metaphysical disputes that otherwise might be interminable” (Pragmatism, p. 28). “To attain perfect clearness in our thoughts of an object”, he claimed, “we need only consider what conceivable effects of a practical kind the object may involve – what sensations we are to expect from it, and what reactions we must prepare” (ibid., p. 29). The pragmatic method helped to support radical empiricism, the view that relations, like things themselves, are matters of direct experience. Relations, James believed, are all that is required to hold experience together; “the directly apprehended universe needs no extraneous transempirical connective support”. From this James inferred that the dualism of mental 226

and physical (see mental/physical; the extended essay on the mind/body problem) is not fundamental; one experience can be mental or physical – subject or object – depending on its relations to other experiences. (See pragmatism.) russell According to Russell’s principle of acquaintance, “every proposition which we can understand must be composed wholly of constituents with which we are acquainted” (Problems, p. 58). This criterion of content is not itself empiricist, because acquaintance, defined by Russell as direct awareness, may take non-empirical forms. But Russell went on to exclude this possibility. The objects of acquaintance are either particulars we experience (sense data or objects of introspection), or universals abstracted from them (see sensa). physical objects, Russell held, are not among the objects of acquaintance. In The Problems of Philosophy (1912) he viewed them as the causes of sense data. He defended the instinctive belief in their existence as a hypothesis accounting for “the facts of our own life”. Elsewhere, obeying the maxim that “wherever possible, logical constructions are to be substituted for inferred entities”, Russell tried, like Mill, to reduce propositions about physical objects to propositions about actual and possible sense data. logi c al posi t i vism According to the logical positivists (or “logical empiricists”), traditional metaphysics lacks cognitive meaning or significance (see logical positivism). A metaphysical utterance about God or the absolute may express an emotion or attitude, but it is incapable of being true or false, and it can make no contribution to knowledge. Cognitively meaningful propositions are either analytic, true or false as a matter of logic and definition, or empirical, confirmable or disconfirmable by observation. Hence no truths are synthetic and a priori; metaphysics, dogmatic or Kantian, is impossible. If philosophy is

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 227

empiricism more than a purely analytical enterprise (a “department of logic”, as Ayer called it), it is an attempt to achieve “a synthesis and generalization of the results of the various sciences” (as Carnap suggested in Ayer, 1959, p. 80). This attempt can (as Carnap allowed) be called metaphysics, but in depending on science it differs strikingly from both the “completely isolated speculative science of reason” whose “groping” Kant had lamented (Critique, p. 21), and the metaphysics of experience he had sought to put in its place. In “Empiricism, semantics, and ontology” (1950), Carnap introduced the notion of a linguistic framework, a system of linguistic forms governed by fixed rules. He suggested that questions of existence are either internal to a framework, in which case they can be answered by empirical or analytic means, or external, in which case they are practical questions about the value of adopting a framework. This suggests that traditional metaphysics is more than emotive: it is a response to questions we cannot avoid if we hope to theorize, even though it misconstrues those questions as theoretical. quine Quine’s repudiation of analytic truth led to an extreme justification-empiricism: “no statement”, he argued, “is immune to revision” – to reasonable revision, he seemed to say – in response to “recalcitrant experience” (1980, p. 43). Quine remained a content-empiricist, quoting with approval the empiricist slogan nihil in mente quod non prius in sensu (“[there is] nothing in the mind which [is] not first in sense”) (1992, p. 19), but he argued against the assumption that single sentences, or even whole theories, are independent vehicles of empirical meaning. (According to Quine’s holism, single sentences or theories are generally too small to bear meaning in isolation. Reductionist projects such as Russell’s are therefore bound to fail (see reduction, reductionism).) “Ontological questions”, Quine wrote, “are on a par with questions of natural science.” (ibid., p. 45). Physical objects, for example, are “irreducible

posits”, justified because they “expedite our dealings” with experience (ibid., pp. 44, 45). empi ri c i sm and met aphysics Empiricism continues to influence the practice of metaphysics. Recent metaphysical writing is, for example, deeply informed by the results of science. Writers on the metaphysics of color pay careful attention to work in color science; writers on space, time, and causality look closely at theories in physics. No one suggests that science can dictate metaphysical conclusions, if only because “science” does not speak with a single voice. Modest metaphysicians may want to clarify and systematize what scientists have to tell us. Bolder metaphysicians may want to argue (for example) that the success of one theory tells against the existence claims of another, or that existence claims in one domain can be reduced to those in a second. Recent metaphysical writing is also informed by the method of science. Metaphysical theories are often tested by their conformity to “data”; if the theory and the data can be brought into reflective equilibrium, the theory has a claim on our allegiance. Even when the data are not themselves “empirical” – when they include, for example, the pre-analytically plausible judgments sometimes known as “intuitions” – both the method and the theory have some right to be called “empiricist”. (Strawson’s descriptive metaphysics is empiricist in this broad sense.) Whether data qualify as “empirical” is, in any case, a contested question. For many metaphysicians, the “facts of experience” include far more than they did for Locke or Hume. Recent metaphysics has also been shaped by changes in empiricism. Some recent philosophers of mind, for example, have denied the existence of qualia or conscious states (see consciousness). Their case is broadly empirical: we can best account for the “facts of experience”, they argue, without invoking conscious states. For empiricists such as Hume and Russell, the facts of experience are those conscious states. They are more certain than anything else, providing 227

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 228

e p i c u r us the material for our conceptions and the evidence for our beliefs. For the critics of qualia, empirical facts must be intersubjectively available. They may include one’s saying that one is in a conscious state, or even one’s believing that one is, but if we can best account for such facts without invoking the state itself, then metaphysics can simply do without it. See also rationalism. b i b l i og rap hy Ayer, A.J., ed.: Logical Positivism (New York: Free Press, 1959). Berkeley, G.: The Works of George Berkeley, ed. A.A. Luce and T.E. Jessop (London: Nelson, 1949), vol. 2. Carnap, R.: “Empiricism, Semantics, and Ontology” (1950); repr. Meaning and Necessity, enlarged edn. (Chicago: University of Chicago Press, 1956). Carnap, R.: Meaning and Necessity, enlarged edn. (Chicago: University of Chicago Press, 1956). Hume, D.: Enquiries Concerning Human Understanding and Concerning the Principles of Morals (London, 1748); ed. L.A. SelbyBigge and P.H. Nidditch (Oxford: Clarendon Press, 1975). Hume, D.: A Treatise of Human Nature (London, 1739–40); ed. L.A. Selby-Bigge and P.H. Nidditch (Oxford: Clarendon Press, 1978). James, W.: Pragmatism (New York, 1907); (Cambridge, MA: Harvard University Press, 1975). Kant, I.: Critique of Pure Reason (Riga, 1781, 1787); trans. N. Kemp Smith (New York: St. Martin’s, 1965). Locke, J.: An Essay Concerning Human Understanding (London, 1690); ed. P.H. Nidditch (Oxford: Clarendon Press, 1975). Quine, W.V.: From a Logical Point of View, 2nd edn. rev. (Cambridge, MA: Harvard University Press, 1980). Quine, W.V.: Pursuit of Truth, rev. edn. (Cambridge, MA: Harvard University Press, 1992). Russell, B.: The Problems of Philosophy (London: Williams and Norgate, 1912); 228

(Oxford: 1959).

Oxford

University

Press,

kenneth p. winkler Epicurus (341–271 bc) Greek philosopher and founder of the Epicurean school. Epicurus adopts many key metaphysical doctrines from the atomic theory of Democritus (mid- to late fifth century bc), but he attempts to eliminate from atomism its tendencies toward reductionism (see reduction, reductionism), skepticism and determinism. Like Democritus, he holds that reflection on the nature of being and not being shows that the universe consists of unchanging, indivisible material bodies and void. Beginning with the Parmenidean principle that nothing comes into being out of non-being or perishes into non-being. Epicurus argues that the totality of what exists can never vary, since something existent can neither perish into the nonexistent nor be generated from it. If, moreover, only things that are spatially extended can exist, and only tangible things are spatially extended, what exists must be tangible – hence a material body. That motion and change exist, he believes, is a self-evident empirical fact that demonstrates the existence of void: bodies can move only if there is something intangible or void which allows them passage by giving way and offering no resistance. Epicureans sometimes speak of void as either place or unoccupied space, which some have thought to be a confusion. But the confusion is only apparent. Aristotle had argued that Democritean atomism treats void as a place that is either empty or filled; thus when an object moves into a void, body and void will be coextensive in the same place (Physics, 216a 26ff.). But, whereas the existence of occupied place is hardly controversial, atomism also needs a stronger conception of void that treats bodies and void as mutually exclusive; void, that is, must serve as the interval between bodies. In attempting to meet such challenges to Democritean atomism, Epicurus arguably develops the first conception of space in antiquity broad enough to encompass both the location of

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 229

epi c urus bodies and the intervals between them. That tensions in his theory remain is not surprising, given the long history of debates between proponents of absolute and relational theories of space. The dualism of body and void explains the continual change we see in the world and also the permanence and conservation of being. It also shows, he believes, that there must be atoms. At the macroscopic level, we see material bodies changing, breaking or wearing away. Such bodies are divisible, hence penetrated by void. But the divisibility of matter, he argues, must be finite, since infinite physical divisibility would lead to bodies being destroyed into non-being. When processes of physical division arrive at bodies with no admixture of void, what remains are “atomic” (meaning “uncuttable”) bodies. Because they are eternal and indestructible, they account for the conservation of being while underwriting the processes of loss and repair visible in macroscopic bodies. To meet the objection that physical atoms might be theoretically or conceptually divisible to infinity, Epicurus offers a theory of minimal parts. He argues that a finite magnitude cannot contain an infinite number of smaller magnitudes (as Zeno had claimed); rather, atoms contain but a finite number of theoretical minima. Size, weight, shape, and tangibility are the basic properties of atoms. Accidental properties, which occur at the phenomenal level, include such secondary properties as color, and time, which Epicurus takes to be an accident associated with the motions of bodies (see quality, primary/secondary). However, unlike Democritus, he takes phenomenal properties to be real and he nowhere suggests that they enjoy a diminished epistemological or ontological status. Nor does he believe that the right kind of bridging laws would enable us to reduce macroscopic properties to more primary ontological components. To be sure, he thinks that all states and events are rooted in the movements of atoms. But he rejects the eliminative materialism of Democritus along with his skepticism about macroscopic properties. These general aspects of Epicurus” metaphysics are most readily seen in his philo-

sophy of mind, where he is at great pains to make room for the data of folk psychology. He thinks that both reductionism and determinism eliminate the concepts of belief, desire, and rationality needed to describe and understand mental life. For example, he claims that reductionism and determinism are self-refuting since they abolish all grounds for distinguishing rational from non-rational argument. Reductionism, in effect, requires us to view ourselves in a way that we cannot; it demands that we ultimately adopt a view of the world that eliminates from it our own point of view as rational agents. Similarly, when one tries to argue for determinism, one must engage in a rational action (i.e., arguing) that presupposes exactly what one is attempting to deny – that one’s action is not predetermined by antecedent causes. These objections to reductionism and determinism have no doubt received more sophisticated formulations since, but Epicurus is the first to have given these issues their particular shape. So too, he is the first philosopher in antiquity to hold an explicitly incompatibilist theory of action. Unfortunately, almost all the details of his positive account have been lost. It is reasonably certain that he postulates unpredictable, random motions in atoms – “swerves” – which are supposed to prevent our actions from being wholly determined by our genetic make-up and environment. However, the precise connections he sees between microscopic indeterminacy and our macroscopic intentions, settled habits, and character remain largely a matter of speculation. For the Epicurean, the study of metaphysics has instrumental value at best. Its chief goal is to alleviate human suffering by freeing individuals from fears about the gods and death. Only atomism, he believes, provides the requisite metaphysical solace. It shows that the present configuration of the cosmos is merely one of innumerable rearrangements of the infinite store of atoms. Our world has taken its form by purely natural, non-teleological mechanisms, and it will be destroyed by those same mechanisms. Atomism therefore frees us from the fear of the gods, since they play no role in nature. It also eliminates our fear of death. 229

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 230

e s s e n c e / accid ent When we die, our atoms merely disperse; we are therefore annihilated and no longer vulnerable to harm. Nor can death harm us when we are alive, since “when we exist death is not present, and when it is present, we do not exist”. For the Epicurean, demonstrating that death is nothing to us is the highest achievement of metaphysics and its greatest consolation. w r i t i n gs Epicurea, ed. H. Usener (Leipzig, 1887); (Rome: L’Erma di Bretschneider, 1963). The Hellenistic Philosophers, 2 vols., ed. A.A. Long and D. Sedley (Cambridge: Cambridge University Press, 1987); contains extensive bibliography. b i b l i og rap hy Annas, J.: Hellenistic Philosophy of Mind (Berkeley: University of California Press, 1992). Furley, D.: Two Studies in the Greek Atomists (Princeton, NJ: Princeton University Press, 1967). Nussbaum, M.C.: The Therapy of Desire: Theory and Practice in Hellenistic Ethics (Princeton, NJ: Princeton University Press, 1994). phillip mitsis essence/accident The essential properties of a thing, collectively called its essence, are those of its properties that it must have so long as it exists at all; they are the properties that the thing would have in any possible world. The accidental properties of a thing, by contrast, are those of its properties that it could exist without. If you heat a piece of wax (to use a famous example from Descartes’s Meditations), it loses its hardness and its previous shape, but continues to exist, thereby showing that those properties were accidental to it. But the property of being extended or spread out in space is, according to Descartes, essential to the wax: to think of the wax as no longer being extended is to think of it as no longer existing at all. 230

The notion of essence as just introduced must not be confused with another that sometimes misleadingly goes by the same name. When it is a necessary truth that all Ks are Ls, people sometimes speak of Ks as being essentially Ls, or of L-hood as being an essential property of Ks. This is only to say that any K must be an L if it is to remain a K; it is not necessarily to say that any K must be an L if it is to continue to exist at all. For example, it is a necessary truth that all sprinters are two-legged; but this does not imply that a given sprinter could not lose a leg. Unfortunately, he could – in which event he would no longer be a sprinter, but would still exist and be the same individual as before. So being twolegged is not part of the essence of any individual sprinter, even if it is part of the essence of the kind sprinter (in the sense that it is a necessary condition for belonging to the kind). Confusion on this score would be avoided if the term “essential property” were always reserved for those properties that are necessary to a thing’s very existence and not merely to its membership in some kind. In logical symbols, the notion of an essential property may be defined as follows: for any individual x and property P, P is an essential property of x if, and only if, Nec(x exists → x has P). Note that this is a formula involving what logicians call necessity de re: a free variable occurs within the scope of a modal operator. By contrast, statements saying what is necessary for membership in a kind involve only necessity de dicto: in “Nec ∀x(Kx → Lx)” the only variables occurring within the scope of modal operators are bound variables. Which properties of a thing are essential to it and which merely accidental? There are two extreme answers to this question: pan-essentialism, or the view that everything has all of its properties essentially, and anti-essentialism, or the view that nothing has any of its properties essentially (except perhaps for “universal” properties, such as existing and being such that 1 + 1 = 2). The first answer is often thought to be implied by Leibniz’s doctrine that all the properties of an individual are contained in

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 231

essenc e / accident its “complete concept”; it was also advocated by some of the Absolute Idealists under the slogan “All relations are internal” (i.e., essential to their relata) (see idealism). For criticisms of some of the arguments of the panessentialists, see Moore (1919–20). The other extreme answer has been advocated by many empiricists, including Locke, Mill, Ayer and Quine (see empiricism). Their view is that the necessity of a thing’s having one property is always conditional upon its having some other property; in the terminology above, there are no properties essential to individuals, but only properties essential to kinds. There are also many views that fall between the extremes. A famous in-between view is that of Descartes, who held that there are two kinds of things in the universe, each with its own essence: material things, whose essence is extension, and thinking things, whose essence is thought. (Here it is important to note that “All thinking things are essentially thinking things” is not the triviality it may appear to be. Its proper logical symbolism is not the de dicto “Nec ∀x(Tx → Tx)”, but the de re “∀x(Tx → Nec(x exists → Tx)”. Another in-between position, one that recognizes more essential properties than Descartes’s famous two, is that of Aristotle, who held that the species to which any individual belongs (e.g., man or horse) is essential to the individual. For Aristotelians, one cites the essence of a thing whenever one correctly answers the question “What is it?” Anti-essentialists (e.g., Quine, 1963, and Putnam, 1983) have sometimes argued that essentialism breeds contradictions. Consider the following three statements about a statue of a swan made out of a lump of clay: (1) The statue (s) is identical with the clay (c). (2) Being swan-shaped is essential to the statue. (3) Being swan-shaped is not essential to the clay. By Leibniz’s Law, if s and c are identical, whatever is true of s must be true of c;

hence (1)–(3) form an inconsistent set. Foes of essence have claimed that believers, in essential properties would have to accept all three. In fact, however, there is no reason why believers in essence should accept all three statements. Those who recognize relatively few essential properties would reject (2), perhaps voicing the suspicion that some who accept it have confused the uncontroversial de dicto statement “Necessarily, all statues of swans are swan-shaped” with the disputable de re statement “All statues of swans are necessarily swan-shaped.” Those who are more liberal in the number of properties they count as essential might accept (2), but would in that case no doubt deny (1). The statue is one thing and the lump of clay another, they might say, even if both occupy the same place and contain exactly the same molecules. The two differ precisely in that being swan-shaped is essential to the statue, but not to the clay. Multiplication of essences thus leads to a corresponding multiplication of entities. See also possible worlds; the extended essay on modality and possible worlds. bibl i ography Cohen, S.M.: “Essentialism in Aristotle,” Review of Metaphysics 31 (1977–8), 387–405. Hughes, G. and Cresswell, M.: An Introduction to Modal Logic (London: Methuen, 1968). Moore, G.E.: “External and Internal Relations,” Proceedings of the Aristotelian Society XX (1919–20), 40–62; repr. in his Philosophical Studies (Totowa, NJ: Littlefield, Adams and Co., 1968), 276–309. Plantinga, A.: The Nature of Necessity (Oxford: Clarendon Press, 1974). Putnam, H.: “Why There Is No Ready-Made World,” in his Realism and Reason (Cambridge: Cambridge University Press, 1983), 205–28. Quine, W.V.: “Reference and Modality,” in his From a Logical Point of View (New York: Harper and Row, 1963), 139–59. 231

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 232

e s s e n c e and essentialism Wiggins, D.: Sameness and Substance (Cambridge, MA: Harvard University Press, 1980). james van cleve essence and essentialism Essentialism is fundamentally an idea about property possession: an object has a property essentially if it has it in such a way that it is not even possible that it exist but fail to have it. The clearest examples seem to involve abstract objects: the number two has the property of being even, for example, and it is certainly hard to see how it could exist but lack that property. The null set (if indeed there is such a thing) has the property of having no members; it seems clear that it could not have had, say, four or five members, or, indeed, any members at all. But it is not only abstract objects that can plausibly be thought to have essential properties. According to traditional theology, God has his most important properties (wisdom, knowledge, power, benevolence, being the creator of everything distinct from himself ) essentially. It is also plausible to think that human persons have essentially such properties as possibly being conscious, possibly knowing that 7 + 5 = 12, possibly being able to act, possibly having goals, and the like. Somewhat stronger essential properties of human beings would be such conditional properties as being conscious if functioning properly (if not subject to dysfunction); being able to act if functioning properly, and the like. And all things have trivially essential properties essentially: such properties as being self identical, and not being a married bachelor. The idea of essential properties is connected with a medieval distinction between modality de dicto and modality de re. Roughly speaking, a statement of modality de dicto predicates possibility or necessity of a proposition: for example, possibly, the number of planets is greater than 9. A statement of modality de re, on the other hand, predicates of some object the property of having some property essentially or accidentally: for example, the number of planets (i.e., nine)

232

is essentially greater than 7, or Socrates was accidentally snubnosed. A de dicto proposition may be true when the corresponding de re proposition is false: possibly, the number of planets is greater than 9 is true, but the number of planets is possibly greater than 9 is false. (It is also possible that the de dicto proposition be false when the corresponding de re proposition is true.) Although the distinction between modality de re and modality de dicto was the stock in trade of every medieval graduate student in philosophy, it was disastrously lost in the modern repudiation of all things medieval; it was painfully re-won during the present century. An object has a property essentially if and only if it has it and could not possibly have lacked it. Another way to put the same thing is to say that an object x has the property P essentially if and only if x has it in every possible world in which x exists. It would be incorrect (although not uncommon) to say that an object has a property essentially if and only if it has it in every possible world; the problem with this is that if it were correct, then either contingent objects such as you and I and the rest of us have no essential properties (because we do not exist in every possible world) or we have properties in worlds in which we do not exist. (That is, there are possible worlds in which we do not exist but nevertheless have properties.) Neither of these alternatives is at all palatable. A consequence of this account of essential property possession is that the property existence is essential to whatever has it: for clearly whatever exists is such that it could not have existed but lacked existence. If actualism (the view that there neither are nor could be things that do not exist) is true, then everything has existence and furthermore has it essentially. A special case of a property essential to an object is its essence (or essences): an essence E of an object x is a property it has essentially which is furthermore such that it is not possible that there be something distinct from x that has E. Some think the idea that there are individual essences goes back to Aristotle; it is clearly present in Boethius,

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 233

evans, garet h and was discussed in some detail by Duns Scotus. Haecceities are a special kind of individual essence; the haecceity of an object is the property of being that very object. (Clearly an object x has essentially the property of being that very object; and clearly nothing else could have had the property of being x.) Essentialism has had something of a checkered career in twentieth-century philosophy. At the beginning of the century, those who were at all willing to think about essentialism and essences were inclined to think these notions belonged to a period of philosophy which was long past. The second third of the century saw the rise and floruit of logical positivism; the positivists were inclined to think not just that essences and essentialism belong to an outmoded past, but that the very idea of an essential property is incoherent. (Indeed, on this way of thinking, there really is not any such idea; such terms as “essence” and “essentialism”, while they look as if they mean something sensible, are in fact cognitively meaningless.) The middle third of the century also saw determined attacks on the notion of essential properties by Quine (1953). Calling the idea that there are such things “Aristotelian Essentialism”, he mounted a tenacious attack on the coherence of the notion of a thing’s having an essential property. It is widely conceded at present, however, that this attack is question begging in that it takes for granted (contrary to what the essentialist thinks) that alleged examples of modality de re are really disguised examples of modality de dicto. Through a delicious historical irony, however, essences and essentialism received a new lease on life partly through the efforts of the logical positivists. The positivists very commendably emphasized the importance of logic for philosophy, or at least for certain areas of philosophy; by virtue of this emphasis there arose a renewed interest in modal logic and the semantics of modal logic; and it is but a short step from the semantics of modal logic to an appreciation of the notions of essential properties and essences.

See also essence/accident; possible worlds; the extended essay on modality and possible worlds. bi bliography Adams, R.: “Actualism and Thisness,” Synthese 57 (1981), 3– 42. Kripke, S.: Naming and Necessity, in Semantics of Natural Language, ed. Donald Davidson and Gilbert Harman (Dordrecht: Reidel, 1972); published separately (Cambridge, MA: Harvard University Press, 1980). Menzel, C.: “The True Modal Logic,” Journal of Philosophical Logic 20 (1991). Plantinga, A.: The Nature of Necessity (Oxford: Clarendon Press, 1974). Quine, W.V.: “Three Grades of Modal Involvement” (1953), in his The Ways of Paradox, and Other Essays (Cambridge, MA: Harvard University Press, 1976). Salmon, N.: Reference and Essence (Princeton, NJ: Princeton University Press, 1981). alvin plantinga Evans, Gareth (1946–1980). Philosopher of language and mind who held the Wylde Reader in Mental Philosophy at Oxford until his tragic death at the early age of 34 in 1980. Evans’ research was aimed at understanding semantics, and he produced seminal work on proper names, pronouns, indexicals, demonstratives, and vagueness. One distinctive feature of Evans’ approach was that his accounts of the semantics of expressions of some of these kinds of expression made appeal to the perceptual and cognitive operations of users of the language. It was for this reason – that in some cases he took the analysis of thought to precede the analysis of language – that Michael Dummett (1993, p. 4) described Evans as the first post-analytic philosopher of language. As for demonstratives and indexicals, Evans followed Frege in recognizing the need for an account of the sense of these expressions, in addition to their reference. The

233

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 234

e v a n s , gar eth sense of a demonstrative or indexical expression is to be understood, for Evans, as the way in which the hearer of the expression must think about the object referred to if she is to correctly understand the expression. Thus, for Evans, some kinds of referring expression have a Fregean sense, meaning that in order for a hearer to understand an utterance employing that expression they must think of the referent in a particular way. But not all referring expressions have this feature. Evans’ theory of the semantics of proper names (first published in Evans 1973, but given vastly improved form in Chapter 11 of Evans 1982) is a nuanced and under-appreciated alternative to the descriptivist and causal theories that dominated the twentieth century. On his view, proper names do not have anything corresponding to a Fregean Sense, meaning that the understanding of an utterance employing that expression does not require the hearer to think of the referent in some particular way. Like causal theories and unlike descriptivist theories, on Evans’ account the content of any information associated with the name does not determine the name’s referent. But unlike both causal theories and descriptivist theories, however, Evans’ theory – to a very rough first approximation – holds that the referent is the object that is the causal source of the information, accurate or not, associated with the name. Because those who name something or someone are typically providers of information about that object, the importance of the circumstance of the naming event is thus secured in Evans’ theory. But though important, it is not sacrosanct. Evans’ theory is constructed to accommodate the fact that sometimes the reference of a name changes (Evans provides a number of examples, including “Madagascar,” which at one time referred to part of the coast of the African continent). As is often noted, there are many topics on which Evans made seminal contributions. But what is less often noted is that the bulk of these topic-specific contributions were made in the service of an unusually complex, sophisticated, and nuanced pro234

gram. Many of the deeper aspects of Evans’ genius come into view only through a grappling with the program as a whole. writings “Can There Be Vague Objects?,” Analysis 4 (1978), 208. “The Causal Theory of Names,” Proceedings of the Aristotelian Society, suppl. vol. (1973) 47, 187–208. Collected Papers (Oxford: Oxford University Press, 1996). This collection includes the 12 articles cited above and below, as well as two unpublished papers, “Does Tense Logic Rest Upon a Mistake?” and “Molyneux’s Question.” “Commentary upon Jerry A. Fodor’s ‘Methodological Solipsism Considered as a Research Strategy in Cognitive Psychology’,” The Behavioral and Brain Sciences 3:1 (1980), 79–80. “Identity and Predication,” Journal of Philosophy 72:13 (1975), 343–63. “Pronouns,” Linguistic Inquiry 11:2 (1980), 337–62. “Pronouns, Quantifiers, and Relative Clauses (I),” The Canadian Journal of Philosophy 7:4 (1977), 777–97. “Pronouns, Quantifiers, and Relative Clauses (II),” The Canadian Journal of Philosophy 7:3 (1977), 467–536. “Reference and Contingency,” The Monist 62:2 (1979), 161–89. “Semantic Structure and Logical Form,” in Truth and Meaning: Essays In Semantics, ed. Gareth Evans and John McDowell (Oxford: Clarendon Press, 1976). “Semantic Theory and Tacit Knowledge,” in Wittgenstein: To Follow a Rule, ed. S. Holzman and C. Leich (London: Routledge and Kegan Paul, 1981). “Things Without the Mind: A Commentary upon Chapter Two of Strawson’s Individuals,” Philosophical Subjects: Essays Presented to P.F. Strawson, ed. Zak van Straaten (Oxford: Clarendon Press, 1980). “Understanding Demonstratives,” in Meaning and Understanding, ed. H. Parret and Jacques Bouveresse (Berlin: W. de Gruyter, 1981).

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 235

event theory The Varieties of Reference, ed. John McDowell (Oxford: Oxford University Press, 1982). b i b l i og rap hy Bermúdez, José Luis, ed.: Thought, Reference, and Experience: Themes from the Philosophy of Gareth Evans (Oxford: Oxford University Press, 2005). Davies, Martin: “Gareth Evans (12 May 1946–10 August 1980),” in Donald M. Borchert, ed., The Encyclopedia of Philosophy, 2nd edn., Detroit, MI: Macmillan Reference. Dummett, Michael: Origins of Analytical Philosophy (Cambridge, MA: Harvard University Press, 1994). Grush, Rick, ed.: Special issue on “The Philosophy of Evans,” Electronic Journal of Analytic Philosophy 6 (1998). rick grush

event theory An event is something that happens, an occurrence, something that occurs in a certain place during a particular interval of time. Although the concept of change has a philosophical history that is coeval with western philosophy itself, and although the concept of an event seems inextricably tied to that of change, the concept of an event seems not to have been the focus of sustained philosophical treatment until fairly recently. Due no doubt to a re-emergence of interest in the concept of change and to the growing use of the concept of an event in scientific writing and in theorizing about science, the idea of an event began, in the twentieth century, to take on a philosophical life of its own in the work of mcTaggart, A.N. whitehead, and C.D. broad. In addition, interest in the nature of events has been sparked by versions of the Mind–Body Identity Thesis (see the extended essay on the mind/body problem) formulated explicitly in terms of events (e.g., every mental event is a physical event) and by the idea that getting a clear picture of the nature of events would facilitate discussion of other

philosophical issues (e.g., see the extended essay on causation). Discussions of events have tended to focus on two fundamental questions: Are there events?, and, if so, What is the nature of these entities? These two questions have usually been treated together, since whether or not there are events depends, at least in part, on what events would be like if there were any. While some philosophers have simply assumed that there are events, others have argued explicitly for the claim that there are such entities. Such arguments have typically been concerned with the finding of semantic theories for certain ordinary claims that apparently have to do with the fact that some agent has done something or that some thing has changed. This semantic focus is correct. A metaphysically appropriate reason for thinking that there are entities belonging to some kind or other consists of two arguments. First, there is a deductive argument, whose premise is some commonsensical claim (e.g., “Vesuvius erupted,” “Jack fell down”) and whose conclusion is that there are entities belonging to the kind in question (there are eruptions and fallings, which are events). Second, there is an inductive argument, an inference to the best explanation of the fact that the commonsensical premise means what it in fact does, where what it means is at least in part revealed by the logical relations it bears to other claims. And that best explanation will show, that the premise does indeed entail that there are entities belonging to the kind in question. Thus, the deductive validity of the inference, for example, from “Vesuvius erupted” to “there are eruptions,” is supported by an induction. It is in this way that Donald Davidson argued that there are events and actions. He argued that, to explain the entailment of “Jones killed Smith” by “Jones killed Smith in the kitchen” (and other claims involving adverbial modifiers) and the entailment of “there was a short circuit” and “there was a fire” by the singular causal claim, “the short circuit caused the fire,” we should suppose that such claims implicitly quantify over killings, short circuits, and fires, which 235

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 236

e v e n t theo r y are events. Thus, for example, the best analysis of “Jones killed Smith in the kitchen” has been argued to be “there was a killing of Smith by Jones and it was performed in the kitchen” (in symbols, (∃x)(Killing(Smith, Jones, x) & In(the kitchen, x) ) ), which entails “(∃x)(Killing(Smith, Jones, x).” But “(∃x)(Killing(Smith, Jones, x)” is the analysis of “there is a killing of Smith by Jones.” So, if it is true that Jones killed Smith in the kitchen, then there is an action (which Davidson takes to be a species of event) which is a killing of Smith by Jones. Thus, there would be evidence for the claim that the truth of some ordinary claim deductively implies the claim that there are events and actions, if the inductive argument for a certain semantic analysis of that ordinary claim is a good one. Opponents of Davidson’s analysis (e.g., Romane Clark and Terence Horgan) have argued that alternative semantic theories are better able to explain the semantic features of Davidson’s target sentences without supposing that they entail that there are events (and actions). More recently, Terence Parsons has shown how to extend (with modifications) Davidson’s semantic theory to a much broader class of sentences; and Lombard has shown how to modify Davidson’s analysis of action (and change) sentences so that entailments not captured by that analysis (e.g., the entailment of “Jones did something,” “something happened to Smith,” and “there was a killing” by “Jones killed Smith”) can be explained; the modified analysis of “Jones killed Smith” should look something like this: there was an action, and it was a killing, and its agent was Jones, and its patient was Smith. While the idea that the meaning of certain sentences appears to require supposing that those sentences speak of events has been an idea particularly associated with Davidson, the semantics of singular terms for events has been extensively studied by Philip Peterson, Zeno Vendler, and Jonathan Bennett. Of particular interest is the distinction between perfect nominals, like “Jones’s killing of Smith,” which behave semantically as if they refer to events (or actions, or sometimes, states), because they take quantifers 236

(some killing of Smith), plurals (twelve stabbings of Smith), and adjectives (the gruesome stabbing of Smith), and imperfect nominals, like “Jones’s killing Smith,” which behave semantically as if they refer to fact-like entities (such terms do not take quantifiers, plurals, or adjectives). Bennett has argued that much of what is wrong in some theories of events (e.g., Kim’s) can be traced to confusions involving these two sorts of nominals and to expressions, e.g., “the killing,” that are ambiguous between and an event- and a fact-interpretation. Most philosophers presume that the events whose existence is established by such arguments are abstract particulars, “particulars” (see universals and particulars) in the sense that they are non-repeatable and spatially locatable, “abstract” (see concrete/abstract) in the sense that more than one event can occur simultaneously in the same place. Some philosophers who think this way associate (however inexplicitly) the concept of an event with the concept of change; an event is a change in some object or other. (Some philosophers, like Bennett, have doubts about this, while others, like Jaegwon Kim and David Lewis, deny it outright, holding that the category of events should include, not only changes (e.g., the barn’s turning red), but also states (e.g., the barn’s being red).) Thus, the time at which an event occurs can be associated with the shortest period of time during which the object, which is the subject of that event, changes from the having of one to the having of another, contrary property. Since no object can have both a property and one of its contraries at the same time, there can be no instantaneous events, and every event occurs at some interval of time. Some philosophers (e.g., Roderick M. Chisholm) take events to be literally capable of recurrence, and thus to be universals. However, for an event to recur is for there to be distinct times, t and t′, such that it occurs at t and also occurs at t ′. But, while an event can be occurring at a time that is only part of the whole period during which it occurs (e.g., the movement of the planet Jupiter during July was occurring at noon on

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 237

event theory July 4th), no event occurs at any time that does not include all the times during which it is occurring. But, for an event to recur, it must occur at a time, e.g., t, that does not include a time, namely t′ (when it presumably occurs again), at which it is occurring. Therefore, events cannot recur. Events, if they are changes in objects, inherit whatever spatial locations they have from the spatial locations, if any, of the objects that those events are changes in. Thus, an event that is a change in an object, x, from being F to being G, is located wherever x is at the time it changes from being F to being G. Thus, events do not get their spatial locations by occupying them; if they did, the way physical objects are often said to do, then distinct events could not occur in the same place simultaneously (just as distinct physical objects cannot occupy the same place at the same time). But it does seem that more than one event can occur at the same time and place. However, some philosophers (e.g., Quine, and Davidson in his later views) hold that events, like physical objects, are concrete, and hold that events and physical objects are not to be thought of as belonging to distinct metaphysical kinds. It seems clear that some events are those of which another event is composed; for example, the sinking of a ship seems to be composed of the sinkings of its parts, in much the way that the ship is composed of its parts. However, it also appears to be the case that not every group of events are those of which another is composed; there just seems to be no event composed of a certain explosion on Venus and the death of Caesar. What is not clear is what the principles are that determine when events compose more complex events. The question of what the conditions are under which a number of events compose another seems just as vexed as the question of what the conditions are under which a number of physical objects compose another. On the other hand, some, e.g., Thomson, have denied that not every group of events are those of which another is composed, and have embraced a principle of unrestricted event fusion.

Some views of events seem compatible with there being subjectless events, events that are not the changes in anything whatsoever. Perhaps Whitehead’s is such a view. Such a view is sometimes connected with an attempt to “construct” ordinary objects, continuants, out of events. Whether such a view is a possible one is an unsettled issue. What seems clear, however, is that subjectless events could not be changes, for it seems absurd to suppose that there could be change that was not a change in or of anything whatsoever. And it is not clear what to make of a concept of an event that was detached from that of change. Any serious theory about the nature of entities belonging to some metaphysically interesting kind must address the issue of what properties, if any, such entities have essentially (see essence/accident; essence and essentialism). In the case of events, the issue is made more pressing, for example, by the fact that certain theories of event causation (e.g., David Lewis’s) require that reasoned judgments be made with regard to whether certain events would have occurred under certain, counterfactual circumstances (see counterfactuals). To deal with such issues, the essential features of events must be determined. In the recent literature on events, attention has been given to four essentialist issues. One such issue is whether or not the causes (or effects) of events are essential to the events that have them; Peter van Inwagen has suggested that an event’s causes (but not its effects) are essential to the event that has them, while Lombard has argued that neither the causes nor the effects of events are essential to them. Another issue concerns whether or not it is essential to each event that it be a change in the entity it is in fact a change in. Bennett and Lewis appear to have suggested that the subjects of events are not essential, while Lombard and Kim have argued that they are. A third essentialist thesis concerning events is that it is essential to each event that it occur at the time at which it in fact occurs. Lombard has argued in favor of this proposal, while Bennett and Lewis have argued against it. And the fourth is that it 237

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 238

e v e n t theo r y is essential that each event be a change with respect to the properties it is in fact a change with respect to. Though the first three essentialist issues have received some attention, by far the issue attracting the most has been the last. This is due to the prominence given to debates between the defenders of Kim’s (“multiplier”) view, which claims, for example, that no stabbing can be a killing and that no signing of a check can be a paying of a bill, and the defenders of Davidson’s (“unifier”) view on the identity of events, according to which stabbings can be killings and a signing of a check can be a paying of a bill. Once the question of whether there are events is settled (either by argument or assumption) in the affirmative, philosophers turn to the construction of theories about events. Often, the theory has, as a chief component, a “criterion of identity” for events, a principle giving conditions necessary and sufficient for an event e and an event e′ to be one and the same event. Though there is no general agreement on this, such a principle, it appears, is sought because, when it satisfies certain constraints, it is a vehicle for the articulation of a view about what it is to be an event and how events are related to objects belonging to other kinds (in the way that the usual criterion of identity for physical objects – sameness of spatio-temporal location – embodies the idea that to be a physical object is to be a entity that is made of matter that fills up the space it occupies). Current in the literature are several general types of theory about events, all of which have their supporters and their opponents. W.V. Quine held that physical objects, like events, have temporal parts, and that events may be identified with those temporal parts, and are thus concrete particulars (see temporal parts, stages). Events and physical objects would thus share the same condition of identity: sameness of spatiotemporal location. (Whitehead at one time expressed the view that events are the most fundamental particulars and that they are more basic than physical objects in that the latter are constructions out of events. Variations on this idea seem to be found in 238

more contemporary writers, such as Quine.) It would appear that such a view could not be correct, however, if it were the case that the very idea of an event is the idea of a change in some physical object. Jaegwon Kim’s interest in events centered, in part, on the fact that they seem to figure as the objects of empirical explanations. Since what is typically explained is an object’s having a property at a certain time, Kim takes an event to be the exemplification of a property (or relation) by an object (or objects) at a time. This idea, combined with some other views that Kim holds, lead him to the view that an event e is the same as an event e′ if and only if e and e′ are the exemplifications of the same property by the same object(s) at the same time. Kim’s view has been criticized, principally by Lombard and Bennett, on the grounds that what it says about events is more plausibly seen as truths about facts. In addition, the view has been subjected to criticism from those whose intuitions concerning the identity of events more closely match those of Davidson. Some philosophers, however, have insisted that events are to be understood as facts (e.g., Wilson) or as, along with facts, a species of states of affairs (e.g., Chisholm). Views such as these would, however, have to deal with what seem to be important differences between events and facts, differences rooted in the idea that facts don’t, while events do, occur, and the idea that events do, while facts do not, have spatial and temporal locations. In addition, Davidson argued that any theory of events that construes them as fact- or proposition-like will imply that there is only one event that occurs (see “The Logical Form of Actions Sentences” in his Essays on Actions and Events). Davidson was interested in finding a “coordinate system” in which to “locate” events, in the way that spatio-temporal coordinates specify the locations of physical objects; as a result, Davidson proposed that the network of causes and effects provided such a framework and that events, being essentially the things that cause and are caused, are identical just in case they occupy the same place in that framework,

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 239

event theory that is, just in case they have the same causes and effects. Myles Brand objected to Davidson’s view on the grounds that (a) there are counterexamples, involving the fission of particles followed by a fusion of their parts, to the proposed criterion of identity, and (b) it implies that it is impossible for there to be more than one event which lacks both causes and effects. In his later work, Davidson abandoned this position in favor of Quine’s. Another view, one which places the concept of causation at the heart of the idea of an event, is David Lewis’s. Lewis apparently thinks that events are of philosophical interest only insofar as they bear on other philosophical topics, and that what one should say about events should be driven by the demands of these other issues. Lewis holds that events are causes and effects and tried to construct a theory of events whose features would make events fit neatly into his counterfactual analysis of event causation. In some respects, Lewis’s view is like Myles Brand’s, in that both are moved by the idea that more than one event can occur simultaneously in the same place. Lewis takes an event to be a property-in-intension (a function from possible worlds to sets of things that have that property) of a spatiotemporal region, so that while two distinct events can occur in the same place at the same time, they are such that one could have had a spatio-temporal location different from that of the other. Jonathan Bennett, like Lewis, thinks that not much of a theory about events is forthcoming if one thinks about events on their own; their nature should be supposed to be whatever (and only whatever) it needs to be in order to make constructive use of them in the discussion of other philosophical issues. Also, like Lewis, Bennett takes an event to be a property. But, for Bennett, such a property seems to be a propertyin-extension (one related to the set of things that actually have that property) and is a particular. That is, Bennett thinks that events are tropes. Lawrence Lombard’s view is, like Kim’s, a variation on a property exemplification account. According to Lombard, this idea

is derived from the idea of events as the (non-relational) changes that physical objects undergo when they change. Such changes are construed as exemplifyings of dynamic properties (properties the possession of which implies change), that is, as “movements” by objects from the having of one to the having of another property through densely populated quality spaces, where each quality space is a class of contrary properties the mere having of any member of which by an object does not imply change. Events can then be divided into atomic events and events composed of atomic events, where an event is atomic just in case (roughly) it is a continuous change in a single partless thing with respect to certain (atomic) quality spaces. Non-atomic events are identical just in case they are composed of the same atomic events; and atomic events are identical just in case they are simultaneous movements by the same atomic object through the same portion of the same atomic quality space. bi bliography Bennett, J.: Events and Their Names (Indianapolis, IN: Hackett Publishing Company, 1988). Brand, M.: “Identity Conditions for Events,” American Philosophical Quarterly 14 (1977), 329–37. Brand, M.: “Particulars, Events, and Actions,” in Myles Brand and Douglas Walton, ed., Action Theory (Dordrecht: D. Reidel, 1976), 133– 57. Brand, M. and Douglas W., ed.: Action Theory (Dordrecht: D. Reidel, 1976). Broad, C.D.: An Examination of McTaggart’s Philosophy (Cambridge: Cambridge University Press, 1933 and 1938). Casati, R. and Achille C. V., ed.: Events (Aldershot, Hants: Dartmouth Publishing, 1996). Casati, R. and Achille C. V., ed.: 50 Years of Events: An Annotated Bibliography, 1947–1997 (Bowling Green, OH: The Philosophy Documentation Center, 1997). Chisholm, R. M.: “Events and Propositions,” Noûs 4 (1970), 15–24. 239

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 240

e v e n t theo r y Chisholm, R. M., “States of Affairs Again,” Noûs 5 (1971), 179–89. Clark, R.: “Concerning the Logic of Predicate Modifiers,” in Davidson and Harman (1972). Davidson, D.: Essays on Actions and Events (New York: Oxford University Press, 1980). Davidson, D.: “Reply to Quine on events,” in LePore and McLaughlin, Actions and Events, 172–6. Davidson, D. and Harman, G., ed.: Semantics of Natural Language (Dordrecht: D. Reidel, 1972). Dretske, F.: “Can Events Move?,” Mind 76 (1967), 479–92. Hacker, P.M.S.: “Events and Objects in Space and Time,” Mind 91 (1982), 1– 19. Higgenbotham, J., Pianesi, F., and Varzi, A., ed.: Speaking of Events (New York: Oxford University Press, 2000). Horgan, T.: “The Case Against Events,” The Philosophical Review 87 (1978), 28–47. Kim, J.: “Causation, Nomic Subsumption, and the Concept of Event,” Journal of Philosophy 70 (1973), 217–36. Kim, J.: “Events as Property Exemplifications,” in Brand and Walton (1976), 159–77. Kim, J.: “Events and Their Descriptions: Some Considerations,” in N. Rescher et al., ed., Essays in Honor of Carl. G. Hempel (Dordrecht: D. Reidel, 1969). LePore, E. and McLaughlin, B., ed.: Actions and Events: Perspectives on the Philosophy of Donald Davidson (Oxford: Basil Blackwell, 1985). Lewis, D.: “Events,” in his Philosophical Papers, Vol. II (New York: Oxford University Press, 1986), 241–69. Lombard, L. B.: “Causes, Enablers, and the Counterfactual Analysis,” Philosophical Studies 59 (1990), 195–211. Lombard, L. B.: “Events, Counter-factuals, and Speed,” Australasian Journal of Philosophy 70:2 (June 1992), 187–97. Lombard, L. B.: Events: A Metaphysical Study (London: Routledge and Kegan Paul, 1986). Lombard, L. B.: “How Not To Flip the Prowler: Transitive Verbs of Action and 240

the Identity of Actions,” in LePore and McLaughlin (1985), 268–81. Lombard, L. B.: “Sooner or Later,” Noûs 29:3 (1995), 343–59. McTaggart, J.M.E.: The Nature of Existence, Vol. II (Cambridge: Cambridge University Press, 1927). Parsons, T., Events in the Semantics of English: A Study in Subatomic Semantics (Cambridge, MA: MIT Press, 1990). Quine, W.V.: “Events and Reification,” in LePore and McLaughlin (1985), 162– 71. Quine, W.V.: “Things and Their Place in Theories,” in his Theories and Things (Cambridge, MA: Harvard University Press, 1981). Rescher, N. et al., ed.: Essays in Honor of Carl. G. Hempel (Dordrecht: D. Reidel, 1969). Taylor, B.: Modes of Occurrence (Oxford: Basil Blackwell, 1985). Thalberg, I.: “A World Without Events?,” in B. Vermazen and M. Hintikka, ed., Essays on Davidson: Actions and Events (Oxford: Clarendon Press, 1985), 137– 55. Thomson, J. J.: Acts and Other Events (Ithaca, NY: Cornell University Press, 1977). Tiles, J.E.: Things that Happen (Aberdeen: Aberdeen University Press, 1981). van Inwagen, P.: “Ability and Responsibility,” The Philosophical Review 87 (1978), 201–24, esp. 207–9. Vendler, Z.: “Facts and Events,” in his Linguistics and Philosophy (Ithaca, NY: Cornell University Press, 1967), 122– 46. Vermazen, B. and M. Hintikka, ed.: Essays on Davidson: Actions and Events (Oxford: Clarendon Press, 1985). Whitehead, A.N.: The Principles of Natural Knowledge (Cambridge: Cambridge University Press, 1919). Whitehead, A.N.: Process and Reality (Cambridge: Cambridge University Press, 1929). Wilson, N.: “Facts, Events, and Their Identity Conditions,” Philosophical Studies 25 (1974), 303–21. lawrence b. lombard

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 241

exi stence evolution In ordinary language, evolution means change. In biology, the term has a narrower sense, according to which only a population of organisms can evolve: this happens precisely when there is change in the population’s genetic composition. Biological use of the term emphasizes the idea of common descent. Current theory says that all terrestrial life is genealogically related. Evolutionary biology also seeks to explain patterns of similarity and difference. Natural selection is widely taken to be important; the extent of its importance is a matter of continuing debate. Despite the gap between biological and vernacular usage of the term “evolution”, the idea of extending the biological concept has exercised a continuing allure. Herbert Spencer (1820–1903) thought that Darwin’s theory could be generalized into an allencompassing theory of change; more recently, theories of cultural evolution have attempted a more modest extrapolation from biological ideas. It remains to be seen whether the biological metaphor is a fruitful one.

b i b l i og rap hy Bowler, P.: Evolution: The History of an Idea (Berkeley: University of California Press, 1984). Futuyma, D.: Evolutionary Biology (Sunderland, MA: Sinauer Publishers, 1986). Sober, E.: Philosophy of Biology (Boulder, CO: Westview Press, 1993). elliott sober

existence Perhaps the most fundamental questions about the concept of existence are what sort of concept it is, whether it can be analyzed or elucidated, and what falls under it. We shall here concentrate on the first two of these questions, with primary emphasis being placed on the first. In so doing, we shall inevitably pay some attention to the third question; but it will only be considered in detail elsewhere in the volume (for detailed discussion, see fictional

truth, objects and characters; hypostasis, reification). what sort of c oncept i s exi stence? Here attention has been focused chiefly on the question whether existence is a property. Since Kant many philosophers have thought this question crucial to a proper assessment of the so-called Ontological Argument for the existence of God. But in order to avoid the question whether properties themselves exist, and to help clarify the logical grammar of existence claims, it is often held to be convenient to concentrate instead on the linguistic analogue of this question. This is whether the word “exists” is a predicate, i.e., whether it is an expression which is true or false of things. Examples of predicates are expressions like “is an author” or “is triangular” as they occur in such sentences as “Goethe is an author” and “France is not triangular”. So why should “exists”, as it occurs in such sentences as “Kangaroos exist”, “Dodos don’t exist”, “Goethe exists” and “Holmes doesn’t exist”, not be similar? One thought – traceable to Hume and Kant – is that predicating “exists” of an Individual, unlike, “is an author” (say), appears to be redundant, “Goethe exists and is German” does not tell us anything more than does “Goethe is German”. But predicating “is an author” of an individual certainly can add something. “Goethe is an author and is German” is considerably more informative than “Goethe is German”. In itself, however, this does not show that “exists” is not a predicate. At best, it shows that “exists” is a predicate true of everything. Admittedly, this would make “exists” an odd kind of predicate, a limiting case like “is either triangular or not triangular”; but it would remain a predicate none the less. However, the apparent fact that, if “exists” is a predicate, it must be true of everything, does indicate a difficulty. For it seems hard to reconcile its being a predicate true of everything with the truth of present-tense, singular negative existential claims such as “Holmes doesn’t exist”. To see this, consider the true sentence “France is not triangular” 241

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 242

e x i s t e nce which contains the complex predicate “is not triangular”. This is true since the predicate is true of the referent of “France”, i.e., France. But nothing similar can be said of the expression “doesn’t exist” For if “exists” is true of everything, “doesn’t exist” is true of nothing. And so the equally true sentence “Holmes doesn’t exist” cannot be true by virtue of the phrase “doesn’t exist” being true of anything. Thus (the argument goes) the expression “doesn’t exist” cannot be a predicate, and so neither can “exists” itself. One response – associated in the first instance with Frege – is to treat “exists” not as what is called a first-level predicate, a predicate true of individuals, but as a secondlevel (or more generally an n + 1th-level) predicate, a predicate of first-level (or nth-level) concepts. This seems to sit well with such claims as that kangaroos exist and that dodos do not. For these claims can easily be construed as saying that the concept kangaroo is instantiated or that the concept dodo is not. Exists, then, becomes like the concept numerous. To say, for example, that cockroaches are numerous is to say that the concept cockroach has numerous instantiations. But how is this to go over to singular claims? Here advocates of the view that “exists” expresses a second- or higher-level concept are apt to claim at least initially that names like “Goethe” and “Holmes” really do express concepts in some way. The most familiar way of elaborating this suggestion is via a treatment of definite descriptions – expressions paradigmatically of the form “the F” in English – suggested by Russell (1905). Russell argued that the best way to understand apparent subjectpredicate sentences containing them – sentences of the form “The F is G” – is to treat them as a conjunction, something equivalent to: the concept F is uniquely instantiated and the thing that instantiates it is G. But then “The F exists” amounts to the claim that the concept F is uniquely instantiated and the thing that instantiates it exists, or more simply, since the last conjunct is redundant, that the concept F is uniquely instantiated. So sentences of the form “The F exists” can be seen to involve “exists” 242

functioning as a second-level predicate, the predicate “is instantiated”. The trick now is to interpret names in some way as definite descriptions. Thus if “Goethe” is interpreted as meaning the same as “the author of Faust”, then “Goethe exists” will amount to the claim that the author of Faust exists, i.e., the concept author of Faust is uniquely instantiated. And if “Holmes” is interpreted as “the greatest detective”, then “Holmes doesn’t exist” will mean that the greatest detective does not exist, i.e., that the concept greatest detective is not uniquely instantiated. However, even leaving aside problems with Russell’s particular theory of definite descriptions, the idea that all apparent singular referring expressions should be thought of as covert definite descriptions is not very plausible. It is possible of course to define a particular name (say) so that it is synonymous with some definite description; such names are often called “descriptive names”. But to model all singular referring expressions on descriptive names is to disregard two facts. First, it is possible to understand sentences involving demonstratives – words such as “this” and “that” used, for example, to refer to objects manifestly occurring within one’s perceptual field – without that understanding’s being mediated by descriptive information. And second, the understanding of a definite description does not in itself presuppose knowledge of who or what satisfies the description. (There may be no such satisfier, for example.) Whereas no one could understand referring demonstratives, and probably most names, without knowing which objects or people they referred to. Or so it is claimed. If this is right, then the view that “exists” is invariably a second- or higher-level predicate looks difficult to defend. It is perhaps conceivable that for the most part genuine referring expressions do not function as covert descriptions, but that in existential claims they do. But names do not seem to vary in function in this way. “Goethe” surely means the same thing in “Goethe is an author” as in “Goethe exists.” And in any case there is much direct evidence that “exists” does sometimes function as a firstlevel predicate.

9780631199991_4_P2005.qxd 1/12/09 3:06 PM Page 243

exi stence One piece of evidence is that “exists” does not function in exactly the same way as “numerous” does. For “numerous”, which is undoubtedly a predicate of concepts, does not sensibly form singular claims: “Goethe is numerous” does not make sense. But “Goethe exists” certainly does. Second, tensed and modal expressions of existence seem much less problematically to involve predications of individuals. Thus “France did not exist” is true by virtue of the predicate “did not exist” being true of the referent of “France”. Similarly with “France might not have existed”. But if these forms of words use complex predicates of individuals, it seems implausible, without further explanation, to suppose that ordinary present tense, singular negative existentials do not. To be sure, something has to be said about those existence claims in which “exists” is most naturally treated as a predicate of concepts (“Kangaroos exist”, “Dodos do not exist”, and so on). But it is arguable that “exists” can function as both a first- and a higher-level predicate without being ambiguous. Compare the case of “disappears”. In “Dodos have disappeared” it is plainly functioning as a second-level predicate, while in “Lord Lucan has disappeared” it is plainly functioning as a first-level predicate. Of course, both “exists” and “disappears” would be being used in slightly different ways in the different types of sentence, but this is a far cry from saying that they are ambiguous. The ordinary conception of ambiguity is not that fine-grained. This still leaves the problem of present tense, singular negative existentials, which appear to show that “exists” cannot be a first-level predicate. If it were (so the argument goes), then “Holmes doesn’t exist” ’s would be true by virtue of “doesn’t exist’s being true of the referent of “Holmes”; and this is surely false, since Holmes does not exist. But now contrast the ontologically expansionist suggestion of Meinong. According to this suggestion, we should take such claims as “Holmes doesn’t exist” at face value. We should take them to be saying of some individual – in this case, the fictional detective Holmes – that he does not exist. So the claim “Holmes doesn’t exist” is

true precisely because the complex predicate “doesn’t exist” is true of the referent of “Holmes”, i.e., Holmes himself. And “exists” can happily function as a first-level predicate in these claims. On the face of it, of course, this seems to entail a contradiction: that Holmes exists and that he does not. But (says the Meinongian) this would be a mistake. All it entails is that there are individuals who do not exist and that one of them is Holmes. Being a non-existent object is not the same as being a non-existent existent. According to the Meinongi

blackwell-Companion-to-Metaphysics.pdf - PDFCOFFEE.COM (2024)