Georges Dicker-Descartes - An Analytical and Historical Introduction-Oxford University Press (2013)

Descartes
This page intentionally left blank

Descartes
An Analytical and Historical Introduction
SECOND EDITION
GEORGES DICKER
1
Oxford University Press is a department of the University of Oxford.
It furthers the Universitys objective of excellence in research,
scholarship, and education by publishing worldwide.
Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With oces in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press

in the UK and certain other countries.
Published in the United States of America by
Oxford University Press
198 Madison Avenue, New York, NY 10016
Georges Dicker 2013
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means, without the prior permission in
writing of Oxford University Press, or as expressly permitted by law, by license, or under
terms agreed with the appropriate reproduction rights organization. Inquiries concerning
reproduction outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data

Dicker, Georges, 1942
Descartes: an analytical and historical introduction / Georges Dicker. -2nd ed.
p. cm.
Includes bibliographical references.
ISBN 978-0-19-538032-3 (pbk : alk. paper)ISBN 978-0-19-970160-5 (updf)
1. Descartes, Ren, 15961650. Meditationes de prima philosophia. 2. First philosophy.
3. God-Proof, Ontological. 4. Methodology. 5. Knowledge, Theory of. I. Title.
B1854.D53 2013
194dc23 2012030580
1 3 5 7 9 8 6 4 2
Printed in the United States of America
on acid-free paper
To Alvina,
Beloved wife and best friend
P R E FA C E TO T H E S ECO N D E D I T I O N
In the two decades since the rst edition of this book appeared, I have
continued to think about and regularly to teach Descartess Meditations.
I have also had the benet of discussing my ideas about the Meditations
and related Cartesian texts with many colleagues and students, and of
reading some ne recent work on Descartes. My continued engagement
with that ever-fascinating thinker has led to an expansion of knowledge,
to some changes of judgment, and to a deeper understanding. In this new
edition, I share these fruits of my journey.
The most signicant expansion is a new chapter on the Fourth Medita-
tion. In the rst edition, I only summarized very briey that Meditations
main theme before moving on to the Fifth Meditation. Here, I oer a full-
scale treatment, targeting especially the issue of how assent to a clearly
and distinctly perceived truth can be, as Descartes maintains, a free act
of will if, as he also maintains, such assent is unavoidable. In light of the
theodicean concerns of the Fourth Meditation, and of the partly peda-
gogical aims of the book, I also provide a substantial discussion of the
traditional problem of evil.
The most signicant change of judgment comes in my treatment
of Descartess cogito. In the rst edition, I argued that this most-famous
of Descartess demonstrations could in the end not avoid the dilemma of
being either question-begging or invalid. Here, partly as a result of an
extended correspondence with Gary Iseminger, I defend a more sympa-
thetic assessment of the cogito, while still incorporating important ele-
ments of my earlier discussion such as Descartess grounding of the
cogito (or at least one version of it) in the theory of substance.
The main exegetical, interpretative, and evaluative themes of other
parts of the book are in general the same as in the rst edition. But
vii
viii Preface to the Second Edition
throughout, I have tried to pass the text through the lter of my current
thinking and to make needed revisions without comprising the integrity
of the original work. The result is that in virtually every section, some
material has been reworked or updated, usually with a view to currency,
accuracy, clarication, or completeness. For example, for currencys sake
I have eliminated the account of the now-discredited memory defense
against the accusation of the Cartesian Circle, and substituted for it a
more plausible counterpart that sees the mere pastness of a clear and
distinct perception as the source of its doubtfulness absent the divine
guarantee of the truth of clear and distinct perceptions; for accuracys
sake, I have provided what I regard as more faithful analyses of Des-
cartess rationale for dismissing the insanity hypothesis in the First Med-
itation and of his notion of eminent containment in the Third and Sixth
Meditations; for claritys sake, I have streamlined the presentation of the
problem of the Circle and sharpened the defense of the solution to it that
I favor; for completenesss sake, I have enriched my discussion of the sub-
stance theory and of the alternatives to it in chapter , discussed the
question of whether the unreconstructed cogito needs an additional,
general premise, added a more text-based discussion of Kants objection
to the Ontological Argument and some discussion of Descartess modal
version of that argument in chapter , and enriched the discussion of
primary and secondary qualities and of mind-body issues in chapter .
Another dierence is that this edition no longer includes the text of
the bulk of the Meditations. Given the availability of numerous inex-
pensive editions of that work, and the augmented length and expense
involved in reprinting the premier English translation of it by John
Cottingham, this change seemed advisable.
The passage of time has not erased my debt to the many teachers,
mentors, colleagues, students, and friends, some of whom are now sadly
gone, who gave me advice, support, or inspiration in writing the rst edi-
tion. They included William H. Hay and Marjorie H. Stewart, to whom
the rst edition was dedicated, as well as Jonathan Bennett, Jos Bernar-
dete, Arthur Bierman, Roland P. Blum, Roderick M. Chisholm, Fred
Dretske, Richard Feldman, Robert Gemmett, Jack Glickman, Eli Hirsch,
Brian ONeil, Ingmar Persson, William L. Rowe, Marcus G. Singer,
Ellen Suckiel, George J. Stack, James Syfers, James Van Cleve, Rudolph
H. Weingartner, and Paul Zi. To these I must now add people who have
given me valuable comments in writing or in conversation since the ap-
pearance of the rst edition, including Jean-Marie Beyssade, Krasimira
Preface to the Second Edition ix
Filcheva, Richard Glauser, Gary Iseminger, Marie Jayasekera, and Peter

Millican, and people whose kind remarks encouraged me to work on a
new edition, including Gordon Barnes, Andrew Chignell, Gary Hateld,
James Mahon, Catherine McKeen, Keith McPartland, Peter Ohlin, Elliott
Sober, and Margaret Wilson. I also thank my beloved wife, Alvina Green-
berg, for her unwavering support and for preparing the line drawings
that grace several pages of this book. Finally I am grateful to my caring
and loyal son, Keith.
Pittsford, New York G. D.
October
P R E FA C E TO T H E F I R ST E D I T I O N
Descartess Meditations speaks to the philosophical novice as well as the

sophisticate; it introduces basic issues of philosophy in way that is brief,
compelling, and penetrating, and it develops them with a subtlety that
remains exhilarating to us, Descartess philosophical descendents. No
wonder, then, that the Meditations continues to be read and analyzed at
all levels of the philosophy curriculum, from the introductory course to
the graduate seminar.
Like a number of other books on Descartess philosophy, this work is
essentially a commentary on his masterpiece. But unlike many of those
books, it is addressed to students of the Meditations at virtually all levels,
and to general readers interested in philosophical issues and their history.
In order to address the introductory student, I have sought not only to
provide some historical background, but also and especially to elicit a
number of basic issues and concepts from the Meditationsto milk that
great text for central philosophical ideas. For example, in analyzing the
cogito in chapter , I try to relate it in a systematic way to the issue of
substance and identity through change and even to the problem of uni-
versals, no less than to the dualistic view of persons to which it serves as
the point of entry. In further discussing that view of persons in chapter
, I not only analyze Descartess case for dualism but also present the
problem of interaction and some historical and contemporary responses
to it. Again, in discussing Descartess views about the material world in
chapter , I not only expound his proof of the material world as an at-
tempt to answer the skeptical doubts generated in Meditation I and
against the background of his arguments for Gods existence but also dis-
cuss his views about the nature of material things, including the theory
of primary and secondary qualities (where I draw some comparisons with
xi
xii Preface to the First Edition
Locke). I have tried to do these things in a language and style accessible

to todays college students, yet without sacricing rigor. For the intro-
ductory students sake, I have also tried, especially at strategic points in
chapter , to explain briey some matters that would be taken for granted
in a book addressed solely to advanced readers, including some elemen-
tary points of logic and such things as the a prioria posteriori distinction.
For students who are studying Descartes at the next higher level
typically in a survey of modern philosophy courseI have sought to
cover, in a balanced way, the main themes and arguments of the Medita-
tions, as well as the main criticisms that they have evoked. For example,
chapter presents a detailed reconstruction and a critique of Descartess
main causal argument for Gods existence, as well as a critical survey of
the main positions on the vexed problem of the Cartesian Circle; chapter
oers an extensive analysis of the Meditation V Ontological Argument
for Gods existence and the main objections that have been raised against
it; and chapter provides a reconstruction of Descartess proof of the
real distinction between mind and body and an analysis of his exchange
with Arnauld over that argument.
To address upper-level undergraduates studying Descartes in courses
on Rationalism (or on Descartes), I have tried to do a number of dif-
ferent things. First, I have sought to provide reconstructions of Des-
cartess central arguments that neither oversimplify them nor become
unnecessarily technical; second, to discuss relevant episodes from Des-
cartess other writings, including the Objections and Replies, Principles of
Philosophy, Passions of the Soul, Discourse on the Method, and a few of the
letters; and nally, to draw upon, and sometimes discusses critically,
recent English-language Descartes scholarship. Thus, upper-level under-
graduates should nd this book a useful research tool and gain from it
some sense of the nature of contemporary Descartes scholarship.
I venture to hope that these latter attributes will also make this book
useful to graduate students and to some of my peers. To such other readers
of Descartes, I oer here my own reections on several key Cartesian is-
sues. For example, chapter inquires whether Descartess skeptical argu-
ments in Meditation I are self-refuting; chapter oers a reconstruction of
the cogito from the substance theory and an assessment of the cogitos
force; chapter provides a detailed examination of Descartess various
causal principles and their interrelations, and advances a critique of the
view (rst proposed by Anthony Kenny and later adopted by James Van
Cleve and Bernard Williams) that Descartess Meditation III argument
Preface to the First Edition xiii
avoids circularity because its purpose is to vindicate only the general rule
that all clear and distinct perceptions are true. It also proposes a solution
to the problem of the Cartesian Circle that builds on the work of Alan
Gewirth and Harry Frankfurt. Chapter connects Kants objection to the
Ontological Argument to problems about negative existential statements,
and explicates Cateruss objection in terms of the distinction between the
formal and material modes of speech; and chapter explores the implica-
tions of Descartess view of matter as a purely extended substance for the
individuation and identity-conditions of bodies and oers an overall as-
sessment of Cartesian Dualism focusing on issues of logical versus causal
independence of mind and body not usually discussed in treatments of
the topic.
NOTE ON THE REFERENCES
A N D A B B R E V I AT I O N S
All references to the works of Descartes are given in parentheses within

the text. All quotations from Descartes are from Ren Descartes, The
Philosophical Writings of Descartes, Volumes I and II, translated by John
Cottingham, Robert Stootho, and Dugald Murdoch (Cambridge: Cam-
bridge University Press, ) and from The Philosophical Writings
of Descartes, Volume III: The Correspondence, translated by John Cotting-
ham, Robert Stootho, Dugald Murdoch, and Anthony Kenny (Cam-
bridge: Cambridge University Press, ). References to Volume I are
abbreviated as CSM I, references to Volume II as CSM II, and refer-
ences to Volume III as CSMK.
Whenever possible and useful, page references have also been given to
the following two one-volume selections from the same translation:
Ren Descartes, Meditations on First Philosophy with Selections

from the Objections and Replies (Cambridge: Cambridge Univer-
sity Press, 1986), abbreviated as M;
Ren Descartes, Selected Philosophical Writings (Cambridge: Cam-
bridge University Press, 1988), abbreviated as SPW.
References to M have not been given for the Meditations, but only for
the Objections and Replies, because the page numbers for the Medita-
tions in M are identical with the page numbers for the Meditations in
CSM II.
Page references are given, as well, to the complete, original-language
edition of Descartess works, Ren Descartes, Oeuvres de Descartes,
xv
xvi Note on the References and Abbreviations
volumes, edited by C. Adam and P. Tannery (Revised edition, Paris:

J. Vrin, ). References to each volume are given by roman nu-
meral and, where necessary, letter, for example, AT X for Volume X and
AT VIIIA for Volume VIIIA.
All references to works other than Descartess are given in the notes.
All works cited in the notes are listed in the bibliography.
CONTENTS
Preface to the Second Edition vii

Preface to the First Edition xi
Note on the References and Abbreviations xv
1. Meditation I and the Method of Doubt 3

1. DE S C A RT E S S G OA L 3
2 . T HE C A RT E SI A N D OUBT 10
3 . I S T H E C A RT E S I A N D O U B T S E L F R E F U T I N G ? 27
3.1 The Deceptiveness of the Senses 27
3.2 The Dream Argument 30
3.3 The Deceiver Argument 32
2. Meditation II: The Cogito and the Self 39

1 . DE S C A RT E S S I A M T H I N K I N G , T HE R E F OR E I E X I S T 39
2 . T H E C E RTA IN T Y OF ON E S O W N T HO U G H T S 42
3 . A PROB L E M F OR T HE CO G I T O 47
4 . T HE S U B S TA NC E T HEORY 49
5 . A R E C ON S T RU C T ION OF T H E CO G I T O B A S E D ON T H E
S U B S TA NC E T HEORY 54
6 . C R I T IC A L DI S CU S S ION OF T H E R E C ON S T RU C T E D CO G I T O 58
6.1 The Substance Theory and the Argument from Change 58
6.2 The Corollary 64
6.3 The Assumption That Thoughts Are Properties 68
6.4 The Inference to I Exist 69
7 . A DEF E N S E OF T H E U N R E C ON S T RU C T E D CO G I T O 71
xvii
xviii Contents
8 . D OE S T H E U N R E C ON S T RU C T E D CO G I T O R EQ U I R E
A N A DDI T IONA L PR E MI S E? 75
9 . DE S C A RT E S S C ON C E P T ION OF T H E S E L F 80
1 0 . C A RT E SI A N D UA L I S M 86
3. Meditation III: The Criterion of Truth and the Existence of God 91

1. DE S C A RT E S S C R I T E R ION OF T RU T H 91
2. T HE PROJE CT OF ME DI TAT ION III 98
3. F ROM T HE IDE A OF GOD TO GOD 100
3.1 The Nature of Ideas 101
3.2 Objective Reality and Formal Reality 105
3.3 The Core Argument 110
3.4 The Central Argument of Meditation III : The Subargument,
the Core Argument, and the Sequel 114
4 . C R I T IC I SMS OF DE S C A RT E S S C E N T R A L A RG U M E N T I N
ME DI TAT ION III 1 2 9
4.1 The Subargument 130
4.1.1 The Precontainment Principle 130
4.1.2 Degrees of Reality 137
4.1.3 Justifying the Causal Maxim 139
4.2 The Problem of the Cartesian Circle 144
4.2.1 The Restriction of the Doubt to Past Clear and Distinct
Perceptions Defense 147
4.2.2 The General Rule Defense 153
4.2.3 The Radical Doubt of Reason and the Creation of the
Eternal Truths 164
4.2.4 The Validation of Reason 170
4.3 A Final Criticism of the Core Argument 176
4. Meditation IV: Error, Freedom, and Evil 181

1 . T HE I S S U E S OF T HE F O URT H ME DI TAT ION 181
2. E R ROR A ND T H E W IL L 181
3 . S OM E P O S S I B L E OB J E C T ION S 184
3.1 Assenting and Deciding to Believe 184
3.2 Irresistibility and Freedom 186
4 . T H E C OH E R E N C E OF C A RT E SI A N F R E E D O M 190
5 . DE S C A RT E S S T ROUBL ING L E T T ER T O ME S L A ND 197
Contents xix
6 . E R R OR A N D E V I L 205
6.1 The Problem of Evil 205
6.2 Cartesian Theodicy 214
6.3 Some Critical Reflections 217
5. Meditation V: The Ontological Argument for the Existence

of God 221
1. DE S C A RT E S S ON TOLOGIC A L A RGUME N T 221
2. CR I T IQ UE OF T HE ON TOLOGIC A L A RGUME N T 228
2.1 Gaunilos Objection 229
2.2 Kants Objection 231
2.3 Further Consideration of Kants Objection 237
2.4 Cateruss Objection 242
3. S OME IMPL IC AT ION S F OR DE S C A RT E S S SYS T E M 254
6 : Meditation VI: Dualism and the Material World 257

1. T HE S C OPE OF ME DI TAT ION VI 257
2 . DE S C A RT E S S PRO OF OF T H E R E A L DI S T I N C T ION B E T W E E N
M IND A ND B O DY 258
3 . DE S C A RT E S S PRO OF OF T H E M AT E R I A L WOR L D 274
4 . DE S C A RT E S ON T HE NAT UR E OF T HE M AT E R I A L WORLD 282
4.1 Primary and Secondary Qualities 283
4.2 Matter, Space, and Solidity 292
4.3 Bodies as Substances versus Bodies as Modes of Substance 295
5 . DUA L I SM A ND T HE PROBL E M OF IN T E R ACT ION 301
6 . A N A S SE S SME N T OF C A RT E S I A N DUA L I S M 313
Bibliography 329
Index 335
Descartes
| 1 |
Meditation I and the Method of Doubt
1. Descartess Goal
If one had to describe in a single word what Descartes does in his First
Meditation, that word would have to be the verb doubt. Throughout
Meditation I, Descartes doubts, or calls into question, his previous be-
liefs. From a logical point of view, however, Descartess famous and
dramatic decision to doubt all his previous beliefs is not his point of
departure. Rather, his logical point of departure is a statement of the
purpose for which he will doubt them. In the very rst sentence of Med-
itation I, Descartes declares that he must question his beliefs if I [he]
wanted to establish anything at all in the sciences that was stable and
likely to last, and near the end he repeats that he must withhold assent
from his previous beliefs if I [he] want[s] to discover any certainty
(CSM II , SPW , AT VII ). The two phrases in quotation marks are
crucial, because in them Descartes states his purpose: to discover what,
if anything, is really certain. It is for the sake of this goal that Descartes
resolves to doubt his previous beliefs. An analogy of Descartess own
nicely illustrates his basic strategy. Suppose that you had a basket full
of apples, that you feared some of them might be rotten, and that you
wanted to nd the good ones. How might you proceed? Well, the easiest
way would be to turn all the apples out of the basket, inspect them, and
put back into the basket only the unspoiled ones (CSM II , SPW ,
M , AT VII ). Likewise, by trying to doubt all of his beliefs, Des-
cartes hopes to nd some beliefs that he cannot doubt, that is, that are
genuinely certain. Indeed, this quest for certainty (to borrow a phrase
from the American philosopher John Dewey) is the engine that drives
Descartess Meditations as a whole.
Descartess goal does not come as the conclusion of an argument, so it
would be futile to look for some line of reasoning leading up to it. As one

Descartes
contemporary commentator on Descartes points out, it is proper for the

course of an inquiry to be guided by its goal and, so far as logic is con-
cerned, the goal of the inquiry may be postulated as a matter of free
choice... . Descartes states his purpose in the First Meditationss opening
sentence. Nevertheless, before proceeding with Meditation I, we should
ask why this goal seemed important to Descartes, and why his attempt to
attain it remains of interest today.
One reason is simply that the question What, if anything, is really
certain? is an intrinsically interesting one. We live in a world where
there are diverse opinions, views, and theories about many matters.
Often, these views are held with great condence; yet what passes for
knowledge is continually changing, as new scientic discoveries are made
and new theories devised. Experts disagree on many important matters.
Have you ever asked yourself, then, whether anything is really certain?
Or are humans bound to live in a sea of uncertainty, taking their guid-
ance from the prevailing expert opinion of the dayan opinion that
may change by tomorrow? Descartes, for one, was profoundly dissatis-
ed with such a prospect. He was convinced that by a careful application
of thought, genuine certainty could be found, even about matters of the
greatest moment. Does the thought that perhaps nothing is certain leave
you dissatised? Can you sympathize with Descartess hope that careful
thinking can lead to genuine certainty about important matters? If so,
then you can appreciate the most basic reason for asking what, if any-
thing, is really certain: the intrinsic interest of the question.
Another reason pertains to the nature of knowledge. There seems to
be a close connection between knowing something and being certain of
it. Suppose I were to tell you, I know that Paris is the capital of France,
but I am not completely certain that it is. I admit that its a little doubtful
in my mind whether Paris really is the capital of France. You may well
feel that there would be something very wrong (and not merely an igno-
rance of geography) with those statements. If I really know that Paris is
the capital of France, you may want to insist, then I am also (completely)
certain that it is. Now Descartes, and many other philosophers both
before and after him, have maintained that knowledge and certainty
indeed go hand in hand. This provides a further reason for asking what,
if anything, is genuinely certain.

Harry Frankfurt, Demons, Dreamers, and Madmen, p. . The rst half of this book is
a very helpful and readable analysis of Meditation I.
There are also reasons of a historical kind (though even these have
modern counterparts). Let us consider two that exercised a powerful in-
uence on Descartes: the rise of science in the seventeenth century, and
the revival of philosophical skepticism.
In the seventeenth century, there occurred a series of scientic discov-
eries that challenged and eventually destroyed the medieval conception of
the universea conception that had endured for nearly two thousand
years. According to this conception, which was rooted in the physics of
Aristotle ( b.c.) and the astronomy of Ptolemy ( a.d.), the
universe is a sphere with the earth at its center. The moon, planets, sun,
and stars all revolve around the earth in xed, circular orbits. The universe
contains two regions, the sublunar and the supralunar. The sublunar region
is the interval between the moon and the earth; the supralunar region
comprises everything from the moon to the outermost circumference of
the universe. The dierence between the two regions pertains to the kinds
of changes that occur in each of them. In the sublunar region, there are
several dierent sorts of changes: coming into being and passing out of
being (what Aristotle called generation and corruption), changes in things
qualities, changes of position (locomotion), changes in the number of
things that exist. But in the supralunar region, there is only one kind of
change: perfect, circular motion of the heavenly bodies around the earth
(and in epicycles around certain points in their own orbits around earth).
Furthermore, while all locomotion in the supralunar region is circular, loco-
motion in the sublunar region is always rectilinear motion toward what
Aristotle termed a things natural place. Heavy things like earth move
downward toward their natural places; light ones like re move upward
toward theirs. The upshot is that the two regions have completely dierent
principles of motioncompletely dierent physics.
An additional but related element in the medieval cosmos is that of
hierarchy. There is a genuine bestbetterworse scale built into the uni-
verse. This hierarchy is no mere subjective human value judgment.
Rather, it is built into the very fabric of things; for it stems from the
types of changes found in the two regions. Since the sublunar region con-
tains such changes as decline, death and decay (corruption), it is not as
admirable as the supralunar region, where nothing ever dies or passes
out of being, but rather everything exhibits only perfect, circular motion.
Thus the supralunar region is better than, or superior to, the sublunar.
A nal, crucial element in the medieval conception of the universe is
teleology. Teleology comes from the Greek word telos, which means
Descartes
purpose. According to the medieval conception, everything in the uni-

verse serves a purpose. Thus, when Galileo claimed to have seen the
moons of Jupiter through his newly discovered telescope, it was argued
against himas a serious scientic argumentthat Jupiter could not
have any moons, since they would serve no purpose. The purpose of each
thing throughout all of nature is, roughly, to be as like God as it is pos-
sible for that kind of thing to be. Thus, the perfect, circular motion of the
stars approximates the immutability (unchangeability) of God much
more closely than the more varied and chaotic changes in the sublunar
region. But even in the sublunar region, where individual creatures are
born, grow, decay, and die, some resemblance to Gods immutability is
preserved in the xity of species (which was not to be seriously chal-
lenged until Darwins The Origin of Species in the nineteenth century).
This conception of the universe as an earth-centered sphere exhibiting
built-in hierarchy and teleology endured for almost two thousand years.
It had the authority of Aristotle (who was revered as the Philosopher
throughout the Middle Ages), of the Church, and of the universities
behind it. But the scientic discoveries of the seventeenth centuryrst
in astronomy and then in physicschallenged its details and eventually
ruined its ensemble. Let us glance at a few examples. The astronomer
Tycho Brahe showed that comets, which were assigned to the supralunar
region, are not indestructible (they burn up), thereby providing an ex-
ample of corruption where no such change was supposed to occur. He
also discovered sunspotsanother example in the supralunar region of
changes (qualitative ones) that were supposed to occur only in the sub-
lunar region. Kepler, building on observations made by Brahe, showed
that the planets do not move with an even circular motion, but an uneven
elliptical one. Galileos experiments refuted Aristotles view that all
objects in the sublunar region move in a straight line toward their nat-
ural places. These and other discoveries undermined the hierarchical
conception of the universe, since this conception was rooted in the dier-
ence between the types of changes in the inferior sublunar and the su-
perior supralunar regions, that is, in the idea that the physics of the two
regions are entirely dierent. Finally, as Copernicuss theorythat the
earth is not static and at the center of the universe, but instead revolves
around the sungained ever greater acceptance, it ruined the entire
medieval cosmos.
The teleological conception of nature did not escape the onslaught of
the new science either. Consider just the implications of Newtons First
Law, which had been anticipated by Galileo. According to that law, a body
remains in its state of motion or rest unless some force acts upon it. The
implication is that in order to explain why a body accelerates or deceler-
ates, no reference to purpose is required or relevant. Putting the matter
crudely, all that is required is a reference to the push, pull, or gravita-
tional attraction of some other body. By generalizing from this example,
one can gain some appreciation of the transformation that resulted once
teleology was expelled from nature. Now, all physical changes were to be
explained in terms of mathematically formulable laws that made no ref-
erence to purpose, rather than in terms of things striving to realize a
purpose inherent in nature. Formerly, the universe could be conceived on
the analogy of a giant living organism striving toward a goal. Now, it
would be conceived on the analogy of a huge machine operating in accor-
dance with purely mechanical principles.
While the conception of the universe that had reigned virtually uncon-
tested for centuries was being undermined by the new science, the old
certainties were being eroded from another direction as well. In the late
sixteenth century, there occurred a revival of philosophical skepticism
led by the French essayist Michel de Montaigne (). Skepticism
has been an important tradition in philosophy since antiquity, and con-
tinues to have adherents to this day. As a philosophical position, skepti-
cism calls into question the possibility of knowledge. Skeptics typically
use certain arguments intended to show that our cognitive faculties (our
senses, reason, and memory) are not adequate to enable us to distinguish
between truth and falsehood, and so not adequate to enable us to obtain
knowledge. Montaigne revived these arguments, which date back at least
to the Greek skeptic Pyrrho ( b.c.).
Most of Montaignes arguments, like those of Pyrrho and other early
skeptics, were directed against the senses. Since we shall have occasion to
examine such arguments with care later, we shall not discuss them in
detail now. But their general tenor can be gleaned from a few passages
from Montaignes Essays:
We no longer know what things are in truth; for nothing comes to

us except falsied and altered by our senses. When the compass,
the square, and the ruler are o, all the proportions drawn from
them, all the building erected by their measure, are also neces-
sarily imperfect and defective. The uncertainty of our senses
makes everything they produce uncertain... . Furthermore, who
Descartes
shall be t to judge these dierences?... . If he is old, he cannot

judge the perceptions of old age, being himself a party in the dis-
pute; if he is young, likewise; healthy, likewise; likewise sick,
asleep, or awake. We would need someone exempt from all these
qualities ... and by that score we would need a judge that never
was... . To judge the appearances that we receive of objects, we
would need a judicatory instrument; to verify this instrument,
we need a demonstration; to verify the demonstration, an instru-
ment: there we are in a circle... . Now if anyone should want to
judge by appearances anyway, to judge by all appearances is im-
possible; for they clash with one another by their contradictions
and discrepancies, as we see by experience. Shall some selected
appearances rule the others? We shall have to verify this selection
by another selection, the second by a third; and thus it will never
be nished.
Descartes was very familiar with these and other skeptical arguments.
They provided an additional incentive for him to inquire what, if any-
thing, is really certain. As we shall see, he attempted to refute skeptical
arguments once and for all, by rst carrying them much further than
anyone had previously done, and then showing that even his radicalized
versions of the arguments could be answered.
To understand Descartess quest for certainty, one further factor needs
to be mentioned: Descartes was a mathematical genius. He discovered
analytical geometry and invented the Cartesian coordinates (which are
named after Cartesius, the Latinized version of Descartes). While Des-
cartes was still in secondary school (he attended a Jesuit college named
La Flche), he came to feel that most of what he was being taught was not
genuine knowledge. At the same time, he was impressed and delighted
with the clarity and certainty that he found in mathematics. Accordingly,
he conceived the idea that all genuine knowledge ought to be as clear and
certain as mathematical knowledge. This became, for Descartes, the fun-
damental requirement for knowledge: it must be as certain as geometry
or algebra. Only so would it be immune to the skeptics attacks.
In order to fulll this requirement, Descartes devised a method, which
he elaborated at length in an early work entitled Rules for the Direction of
the Mind and in his famous Discourse on the Method. We need not go into

Michel de Montaigne, Essays and Selected Writings, pp. .
the details of this method, but shall only state its most basic rule. This
rule directs us to accept no propositions as true except (a) those which
are so obvious and clear that they cannot be doubted so long as one is
thinking of them attentively and (b) those which logically follow from
propositions of kind (a). In other words, certainty is to be attained by
making sure, as in mathematical proof, that knowledge has the pattern
of a deductively valid argument starting from self-evident, unshakeable
premises. As the following passage from the Discourse on the Method
shows, Descartes had high hopes for this method:
Those long chains composed of very simple and easy reasonings,

which geometers customarily use to arrive at their most dicult
demonstrations, had given me occasion to suppose that all the
things which can fall under human knowledge are intercon-
nected in the same way. And I thought that, provided we refrain
from accepting anything as true which is not, and always keep to
the order required for deducing one thing from another, there
can be nothing too remote to be reached in the end or too well
hidden to be discovered. (CSM I , SPW , AT VI )
The purpose of the doubt that Descartes adopts in Meditation I can

now be claried. Its purpose is to determine what propositions, if any,
cannot be doubted. For if Descartes can nd such indubitable proposi-
tions, then he will have the propositions of kind (a) mentioned above,
from which he can then hope to deduce propositions of kind (b). By his
doubt, then, Descartes does not mean to reject permanently all of his
former beliefs. Some of them may well be true. But if they are, then Des-
cartes wants to rediscover them, in the sense of showing that they follow
logically from basic, indubitable propositions. The main purpose of the
doubt is to nd these indubitable propositions, so that Descartes can use
them as foundations upon which to rebuild his knowledge. The doubt is
a way of rethinking everything from the beginning, so as to achieve the
certainty that Descartes is seeking.
In this section, we have sought to understand why Descartes embarks
on his famous and seminal quest for certainty. We have seen that the
question What, if anything, is certain? is an intrinsically interesting
one and that it seems to be closely related to the question What, if any-
thing, do we really know? We have also come to an appreciation of why
Descartes, a mathematical genius living at a time when old certainties
Descartes
were being shaken by the new science and skeptical philosophers were
renewing their corrosive attacks on the very possibility of knowledge,
should have adopted the quest for certainty as his basic goal. Finally, we
have briey sketched the method that his mathematical pursuits inspired
him to devise for attaining that goal. At this point, then, we turn to an
examination of the text of Meditation I, where Descartes puts this method
to work.
2. The Cartesian Doubt

In this section, we shall follow closely the movement of Descartess thought
in Meditation I, by constructing a step-by-step summary of the Meditation.
As we have already seen, Descartes begins with a statement of his purpose.
We may accordingly enter the following statement as the rst step of our
summary:
. I want to discover what, if anything, is absolutely certain.
Shortly after stating this goal, Descartes declares that reason now
leads me to think that I must hold back my assent from opinions which
are not completely certain and indubitable just as carefully as I do from
those which are patently false (CSM II , SPW , AT VII ). To say
that reason leads me to think that ... is to imply that one is reasoning
from some premise. Then from what premise is Descartes here reasoning?
The answer is clear: from his statement of purpose. For given that ones
purpose is to nd absolute certainty, one has an excellent reason not to
accept things that are uncertain! Descartes is here reasoning directly
from his goal to what he must do in order to attain it. Thus, we may enter
the following statement as the second step in our summary:
. For this purpose, I must withhold belief from things that are
not entirely certain and indubitable just as carefully as from
those which are obviously false.
Descartes here declares that he will be just as wary of what is even

slightly doubtful as of what is obviously false. This attitude can be clari-
ed by contrasting it with the attitude of a person concerned with prac-
tical matterssay with making money. In order to succeed, such a person
must adopt an attitude which is precisely not the one Descartes here
adopts. He cannot be just as suspicious of what is only somewhat doubtful
as of what is obviously false. Instead, he must be willing to take risks, to
act on the probability that his beliefs are correct. Otherwise, he will
remain paralyzed, so to speak, and never achieve his end. Suppose, how-
ever, that your only purpose is to discover what is certain. Suppose that
at least for the moment, you are not concerned with any practical ends,
but only with attaining absolute certainty. Then Descartess policy of
withholding belief even from matters that are only slightly doubtful is
perfectly reasonable.
But what exactly does it mean to withhold belief (or assent)? Well, it
means the same thing as does the more commonly used phrase, to sus-
pend judgment. To clarify this concept, notice that although there are
only two possibilities regarding the truth of a statement (the statement
is either true or false), there are three dierent postures regarding belief
of a statement. These three belief-postures, or doxastic attitudes (to
borrow a term from a contemporary American philosopher Keith Leh-
rer), are as follows: () one can believe the statementaccept it as true
(); one can disbelieve the statementreject it as false; () or one can
withhold the statementneither believe nor disbelieve it. To give an
example, we can compare the doxastic attitudes of a theist, an atheist,
and an agnostic toward the statement God exists. A theist is someone
who believes God exists. An atheist is someone who disbelieves God
exists. An agnostic is someone who withholds (belief in) God exists
(neither believes nor disbelieves it). As this example illustrates, there is a
big dierence between disbelieving and withholding. Disbelieving a
statement, p, is the same thing as believing its denial or negation, not-
p: the atheist, who disbelieves God exists, thereby believes God does
not exist. On the other hand, withholding (belief in) a statement com-
mits one neither to the statement itself nor to its negation: the agnostic
believes neither God exists nor God does not exist. Withholding,
then, is a neutral, noncommittal attitude, by which one avoids commit-
ting oneself to the truth of either a statement or its denial.
We can now see more clearly why Descartes withholds (belief in) state-
ments that are uncertain. He decides to adopt a policy that will never
allow him to accept any statement that is uncertain. So which of the three
doxastic attitudes must he take toward statements that are uncertain?
Well, obviously, he must not believe them. Should he then disbelieve
them? No. For then he would believe their negations, which would go
Descartes
against his policy; for often the negation of an uncertain statement is

itself uncertain (often both p and not-p are uncertain). For example,
there will be a major earthquake in California next year is uncertain,
but so is there will not be a major earthquake in California next year.
Therefore, the only policy toward uncertain statements that can never
lead to accepting an uncertain statement is that of withholding belief.
One can even give a plausible argument showing that when a statement
is uncertain, its negation is never certain. This argument (which, inciden-
tally, is not given by Descartes himself, and which you can skip over with-
out losing track of this exposition of the First Meditation) depends on the
following classication of statements. Statements may be classied as
either certain or not certain, and those that are not certain may be further
subdivided into statements that are uncertain and statements that have
no credibility whatsoever. These last two classes do not overlap, for to call
a statement uncertain is to concede that it has some degree of credibility.
Only as a joke could one say that a statement that has no credibility what-
soever, such as + = , is uncertain; for this statement is not (merely)
uncertain, it is absurd or obviously false. So, although it belongs in the
class of statements that are not certain (since otherwise it would have to
be classed as certain), it does not belong to the subclass of that class con-
sisting of uncertain statements. Rather, it belongs in the subclass (of the
class of statements that are not certain) consisting of statements that
have no credibility. If this classication of statements is correct, then one
can construct a simple argument showing that if a statement p is uncer-
tain, then its negation, not-p, is not certain. The rst premise is:
() If not-p is certain, then p has no credibility.
The only other premise is:
() If p has no credibility, then p is not uncertain.
To see the plausibility of these two premises, let p stand for some obvious
falsehood, such as + = . Then () seems indisputably true. Further, in
light of the distinction between statements that are altogether lacking in
credibility and statements that are merely uncertain, () also seems to be
true. But from () and (), it follows that:
() If not-p is certain, then p is not uncertain.

The step from () and () to () is valid, because it has the following obvi-
ously valid form (called Hypothetical Syllogism):
If P, then Q
If Q, then R
___________
? If P, then R
To see that the step has this form, substitute not-p is certain for P, p
has no credibility for Q, and p is not uncertain for R. Finally, it follows
from () alone that:
() If p is uncertain, then not-p is not certain.
The step from () to () is valid, because it has the following form:
If P, then not-Q
_______________
? If Q, then not-P
To see that the step from () to () has this form, substitute the state-
ment not-p is certain for P and the statement p is uncertain for Q. To
see that the form itself is valid, consider a simple example:
(a) If this is a triangle, then this does not have four sides.
________________________________________________
? (a) If this has four sides, then this is not a triangle.
Another way to see why the step from () to () is valid is to apply to it

two rules of logic. One is Contraposition, which says that
If P, then Q.
___________________
? If not-Q, then not-P.
is a valid form of argument. The other is Double Negation, which says that
not not-P is equivalent to P (e.g., today is not not Wednesday is equiva-
lent to today is Wednesday). Substitute not-p is certain for P and p is
not uncertain for Q. Then applying Contraposition to () yields if p is not
not uncertain, then not-p is not certain; and applying Double Negation to
Descartes
this last statement yields (). The above argument, then, appears to be
sound. If that is right, then the argument provides an additional reason
why Descartes must adopt a policy of withholding rather than disbeliev-
ing uncertain statements: their negations are never certain.
But how is Descartess policy to be implemented? Obviously, he cannot
examine all of his beliefs individually: that would be an endless task.
Accordingly, he proposes to examine the basic principles on which his
beliefs rest. For if these principles are uncertain, then so are any beliefs
resting on them. We may enter the following statement, then, as the
third step in our summary:
. To do this, I need not examine all my beliefs individually, but

only the basic principles on which they rest.
In order to understand what comes next, it is important to realize that

in his Meditations Descartes means to be speaking not only for himself
but for anyone who is seeking to determine the certainty of his or her
beliefs. Despite the fact that Descartess entire Meditations is written in
the rst person singular, the work is certainly not intended as a report of
one mans idiosyncratic musings. Rather, Descartes means to speak for
all of us. He is convinced that anyone who embarks upon the quest for
certainty methodically and without becoming confused will travel the
same route as he himself does in his Meditations. Indeed, some recent
commentators seek to convey the impersonality of the I in the Medita-
tions by always referring to Descartes in the third person, as the medi-
tator, and even to convey its gender neutrality by referring back to the
meditator as she. We shall not follow this practice, but simply note
here that the I of the Meditations is as much an invitation to the reader
to put himself or herself in Descartess place, as it is a way for Descartes
to report his own progress.
Accordingly, suppose that we each ask ourselves, on what basic prin-
ciples are my beliefs based? Moreover, in asking this question, let us not
target our more theoretical or esoteric beliefs, like beliefs about sub-
atomic particles, or outer space, or the distant past. Let us target,
instead, those beliefs that we hold most condently and unhesitatingly.
What sorts of beliefs are these? Well, they are beliefs about our pre-
sent, immediate physical surroundingssuch as your belief that there

See, for example, Gary Hatelds excellent Descartes and the Meditations.
is a book on your desk or that someone is speaking to you. What are

such beliefs based on? The answer is obvious: they are based on your
present perceptions of senseon what you now see, hear, feel, and so
on. If someone asked why you believe there is a book on your desk, the
answer would be Because I see it or Because I feel it; if someone
asked why you believe someone is speaking to you, the answer would be
Because I hear him or Because I see him. The beliefs that we accept
as most obvious and certain, then, are based on our present percep-
tions of sense. The basic principle on which they rest is that the senses
provide us with highly reliable information about our physical sur-
roundings. This is exactly what Descartes, speaking not idiosyncrati-
cally but for all of us, goes on to say: Whatever I have up till now
accepted as most true I have acquired either from the senses or through
the senses (CSM II , SPW , AT VII ). Let us enter this step into
our summary, as follows:
. Until now, everything Ive accepted as most obvious and cer-

tain has been based on the senses.
It is one thing, however, to say that something is accepted as certain,

and another to say that it really is certain. Can we assert, then, that
beliefs acquired by using our senses really are certain? No, answers
Descartesat least not without qualication; for sometimes the senses
are deceptive. Descartes gives no examples of this deceptiveness in Med-
itation I, perhaps because the skeptics had made such examples so famil-
iar to Descartess contemporaries that he judged it unnecessary to do so.
When he reviews this point in Meditation VI, however, he does give an
example, which we may usefully mention now:
Sometimes towers which had looked round from a distance

appeared square from close up; and enormous statues standing
on their pediments did not seem large when observed from the
ground. In these and countless other such cases, I found that the
judgements of the external senses were mistaken. (CSM II ,
SPW , AT VII )
The occasional deceptiveness of the senses, then, provides a reason not

to trust them uncritically. As Descartes puts it: But from time to time
I have found that the senses deceive, and it is prudent never to trust
Descartes
completely those who have deceived us even once (CSM II , SPW ,

AT VII ). Let us enter this point as the next step of our summary:
. But the senses have sometimes deceived me; Id better not

trust them completely.
With this step, Descartes introduces a theme that occupies him

throughout the rest of Meditation I: a critique of the senses. By this cri-
tique, Descartes wants to determine to what extent, if any, the senses
provide certainty. He will argue, as we shall see, that they provide no cer-
tainty. Thus while the primary purpose of the Cartesian doubt is to nd
certainty, it also has a secondary purpose: to show that certainty is not
provided by the sensesthereby preparing the way for Descartess own
views about how certainty can be attained, which he will present in the
subsequent Meditations.
In his critique of the senses, however, Descartes wishes to be scrupu-
lously fair. He knows that despite what skeptics may say, we are all very
strongly inclined to believe that our senses provide us with highly
reliableindeed certaininformation. As the saying goes, Seeing is be-
lieving. To challenge this deep-seated conviction in a way that will lead
us to reconsider it seriously, Descartes takes into account what can be
said on behalf of the senses. He tries, so to speak, to give the senses as
much credit as they are due. This desire to anticipate whatever can be
reasonably said in defense of the senses lies behind his next point:
Yet although the senses occasionally deceive us with respect to

objects which are very small or in the distance, there are many
other beliefs about which doubt is quite impossible, even though
they are derived from the sensesfor example, that I am here,
sitting by the re, wearing a winter dressing-gown, holding this
piece of paper in my hands, and so on. Again, how could it be
denied that these hands or this whole body are mine? (CSM II
, SPW , AT VII )
Descartess point here is one that naturally occurs to anyone thought-

fully trying to weigh the impact of the fact that the senses are sometimes
deceptive. This is that while the senses can indeed lead us astray in cases
where the object perceived is very small or far away (as in the case of
seeing a tower in the distance), it does not follow that the senses are
unreliable when the object is sizable, close by, or the like (as in the case of
seeing a piece of paper in my own hands). To generalize the point, Des-
cartes is saying that just because objects are sometimes misperceived
because the conditions of observation are poor, it does not follow that the
senses are unreliable even when the conditions of observation are good.
Let us enter this important point into our summary:
. Perceptions occurring in poor conditions of observation (e.g.,

when the object is minute or distant) are suspect, but cant I
trust perceptions occurring in favorable conditions (e.g., of a
piece of paper in my own hands as I sit by the re)?
The position reached in step would seem to be simply intelligent

common sense. The general principle it embodiesthat the senses may
be trusted so long as the conditions of observation are goodseems em-
inently reasonable. Yet, doubts might be raised. For example, one might
ask whether it is really enough that the conditions of observation be fa-
vorable: Is it not also required that the perceiver know that they are favor-
able? But how is one to know this except by other perceptions that are
also potentially deceptive? And if we require that these latter perceptions
also be known to occur in favorable conditions, then we shall have to
appeal to still further perceptions for knowledge of those conditions, and
so on without endthereby launching ourselves into what philosophers
call a vicious innite regress.
This line of thought, however, is not one that Descartes pursues. First,
he briey mentions, but quickly dismisses, the possibility that he might be
insane, like madmen whose brains are so damaged that ... they rmly
maintain that they are kings when they are paupers, or say they are dressed
in purple when they are naked, or that their heads are made of earthen-
ware, or that they are pumpkins, or made of glass (CSM II , AT VII ).
Instead of exploring that alarming possibility, he invokes one of his most
famous arguments: the Dream Argument. But before looking at that argu-
ment, let us pause to ask: Why does Descartes so easily dismiss the in-
sanity hypothesis? The reason, we suggest, stems from the nature of his
project. He is not trying to determine whether a person who is demented

This line of argument is developed in my Is There A Problem About Perception and
Knowledge? , and in my Perceptual Knowledge: An Analytical and Historical Study,
chap. .
Descartes
or insane can discover any certainty, or much less, as he also puts it (re-
vealing the fuller scope of his ambitions), whether such a person can
establish anything in the sciences that [is] stable and likely to last (CSM
II , SPW , AT VII ). Nor, despite the fact that Descartes was without
doubt one of the greatest geniuses who ever lived, is he asking whether a
person endowed with a superior mind can discover any certainty or estab-
lish anything stable and lasting in the sciences. Rather, he is asking whether
a rational mind, endowed with normal human intelligence, can, by careful
and methodical thinking, achieve these ends. So, the hypothesis that he
might be insane is simply irrelevant to his inquiry. Of course, someone
may say that he ought to have considered it more seriously, but the ques-
tion would then be: why ought he? Nothing in his project commits him to
doing so, since his purpose is to determine whether a rational mind can
attain certainty. Furthermore, the question of whether a demented per-
son could attain certainty, or establish lasting results that would remain
xed and stable in her mind, seems almost to answer itself.

In this regard, consider the very beginning of Descartess Discourse on the Method:
Good sense is the best distributed thing in the world: for everyone thinks him-
self so well endowed with it that even those who are hardest to please in ev-
erything else do not usually desire more good of it sense than they possess. In
this it is unlikely that everyone is mistaken. It indicates rather that the power
of judging well and of distinguishing the true from the falsewhich is really
what we properly call good sense or reasonis naturally equal in all men, and
consequently that the diversity of our opinions does not arise because some of
us are more reasonable than others but solely because we direct our thoughts
along dierent paths and do not attend to the same things. For it is not enough
to have a good mind; the main thing is to apply it well. The greatest souls are
capable of the greatest vices as well as the greatest virtues; and those who pro-
ceed but very slowly can make much greater progress, if they always follow the
right path, than those who hurry and stray from it.
For my part, I have never presumed my mind to be in any way more perfect
than that of the ordinary man; indeed, I have often wished to have as quick
a wit, or as sharp an imagination, or as ample and prompt a memory as some
others. And apart from these, I know of no other qualities which serve to per-
fect the mind; for as regards reason or sense, since it is the only thing that
makes us men and distinguishes us from the beasts, I am inclined to think that
it exists whole and complete in each of us. (CSM I , SPW , AT VI )
The rst sentence is not a sarcastic remark. Descartes thinks that all sane human beings
possess the power to reason well and to judge correctly (what he here calls good sense),
and that part of possessing that power is recognizing that one possesses it. The passage
as a whole reects very well the standpoint from which Descartes wrote the Meditations.
On the other hand, Descartes cannot just dismiss the possibility that
he might be dreaming, since it is a plain fact that even ordinary, rational
humans have dreams. Thus he exclaims,
As if I were not a man who sleeps at night, and regularly has all
the same experiences while asleep as madmen do when awake
indeed sometimes even more improbable ones. (CSM II , SPW
, AT VII )
The fact of dreaming, then, provides a reason to doubt even perceptions

occurring under the best conditions. This is that such perceptions can be
exactly duplicated in a vivid dream. Of course, not all dreams are that
life-like: some dreams have a dream-like quality. But all that is needed to
provide some reason to doubt even our best perceptions is that some
dreams be so realistic, so apparently authentic, as to be indistinguishable
(during the time of their occurrence) from waking experience. Now it
seems dicult to deny that this is the case. As Descartes puts it:
As I think about this more carefully, I see plainly that there are
never any sure signs by means of which being awake can be dis-
tinguished from being asleep. The result is that I begin to feel
dazed, and this very feeling only reinforces the notion that I may
be asleep. (CSM II , SPW , AT VII )
Let us accordingly enter, as the next step in our summary, the following
response to step :
. No, for even perceptions occurring in ideal conditions of ob-

servation may be indistinguishable from vivid dreams.
In step , Descartes has reached a skeptical position with regard to the

senses. This can be appreciated by comparing () with (). Step only
implied that the senses should not be trusted completely or uncritically,
since some perceptionsthose that occur under adverse conditions
are deceptive. This is hardly a skeptical result; it is, as already noted, just
common sense. But the Dream Argument of step , as far as Descartes is
concerned, shows that not even our best perceptions yield any cer-
tainty. For instance, one cannot be absolutely certain, merely on the basis
of seeing a sheet of paper in good light from a few feet away, that there
Descartes
really is a sheet of paper there; since one could have the same conscious
experience in a dream. This is a skeptical position.
To understand Descartess next step, we need to bring into focus the
pattern of development that he began in step (). This pattern is a pro-
con, or dialectical one. Step (), where Descartes began his critique of
the senses, was a rst con pointone directed against the senses. Step
() was a pro pointa point in favor of the senses. And step ()the
Dream argumentis a new con point. This dialectical structure, in which
each new point is a response to the previous one, allows Descartes to de-
velop his critique of the senses without overlooking what can be said on
their behalf.
Descartess next point, accordingly, is a new pro point. But since
Descartes has already reached a very skeptical position regarding the
senses, this new point is a last-ditch eort to salvage something from the
senses. Descartes now suggests that even if we can never be certain that
we are perceiving reality rather than having a vivid dream, we can at least
be sure that the images we have in our dreams are derived from reality
are like paintings, which must have been fashioned in the likeness of
things that are real (CSM II , SPW , AT VII ). This suggestion is
essentially a hypothesis concerning the origin of dreams: their contents,
though illusory, must be based upon something real that we previously
perceived and thus must in some degree correspond to reality.
The details here are interesting. First, Descartes suggests that at least
dream images of heads, hands, eyes, and whole bodies must be derived
from those very things, somewhat as a painters depiction of imaginary
animals can only represent parts of animals jumbled up in various ways.
Then, he asks himself, in eect, what if a painter comes up with some-
thing wholly ctitious and unreal? He responds that even so, at least the
colors used in the painting must be real ones. He then says that likewise,
even if heads, hands, eyes and so on are imaginary, the simplest and most
universal elements of dream imagescorporeality, extension (= three-
dimensional spread-outness, or three dimensionality), shape, quantity,
size, number, place, durationmust be derived from real counterparts.
Descartes thus suggests that the colors used in a painting are to images
of ctitious creatures in the painting as the simplest elements in dream
images are to real counterparts of them. This is a very abstract analogy,
since it abstracts from the fact that colors are components of or ingredients
in the paintings, whereas the real things from which the simplest ele-
ments in dreams are supposed to be derived are counterparts of those
elements. But the point of the analogy is that both the colors and the
counterparts have to be real: they must really exist.
From his analogy, Descartes then infers that sciences that deal with
these simplest things, especially mathematics (arithmetic and geometry)
may be more secure than those that deal with composite things. Here he
is not only highlighting the apparent certainty of mathematics, but also
preparing the ground for his own geometricized physics, to be discussed
later in this book. In his early work, Rules for the Direction of the Mind, he
calls the simple elements just listed simple natures and says that the
whole of human knowledge consists uniquely in our achieving a distinct
perception of how all these simple natures contribute to the composition
of other things (CSM I , AT X ).
At this point, Descartes also declares:
[W]hether I am awake or asleep, two and three added together

are ve, and a square has no more than four sides. It seems im-
possible that such transparent truths should incur any suspicion
of being false. (CSM II , SPW , AT VII )
We shall postpone comment on this remark for a few moments, since it

raises a special puzzle. So we may now enter, as the next step in our sum-
mary, the following:
. Mustnt the images in a dream at least be derived from some-

thing real? Further, isnt mathematics secure even in sleep?
As already mentioned, this suggestion is a last-ditch attempt to sal-

vage something from the senses. For it concedes that we can never be
certain (on the basis of our senses) that we are perceiving reality, rather
than dreaming, insisting only on a very minimal and indirect link between
perceptual experience and reality.
Nevertheless, Descartes nds that not even this tenuous link is absolutely
certain. To show why it is not, he invokes a nal, extraordinary skeptical
argumentperhaps the most radical skeptical consideration imaginable:
And yet rmly rooted in my mind is the long-standing opinion

that there is an omnipotent God who made me the kind of crea-
ture that I am. How do I know he has not brought it about that
there is no earth, no sky, no extended thing, no shape, no size, no
Descartes
place, while at the same time ensuring that all these things ap-
pear to me to exist just as they do now? What is more, since I
sometimes believe that others go astray in cases where they
think they have the most perfect knowledge, may I not similarly
go wrong every time I add two and three or count the sides of a
square, or even in some simpler matter, if that is imaginable?
(CSM II , SPW , AT VII )
Descartes is here saying that, contrary to (), the images in ones dreams
need not even correspond to anything real. There need be no connection at
all, not even the most tenuous one, between my perceptual experience and
physical reality; for perhaps there is no physical world at all! Perhaps, instead,
an omnipotent (all-powerful) God has created me such that I have experi-
ences exactly like the ones I would have if there were such a world, that is,
visual, tactile, auditory, gustatory, olfactory and kinesthetic experiences so
vivid and orderly that it seems to me that I am perceiving physical things
even though there really are no such things. In short, perhaps God has so
created me that I hallucinate the entire physical world! How can I possibly
know that this is not the case, since by hypothesis all of my perceptual ex-
periences would be exactly the same if it were?
For the rst time in the Meditations, then, Descartes is here calling
into question the very existence of the physical world. His Deceiver Argu-
ment, as we may call it, goes far beyond the Dream Argument, which
questioned only whether we can tell when we are perceiving physical
things, not whether such things exist. It also goes beyond the arguments
of Pyrrho, Montaigne, and all other earlier skeptics, who had suggested
that the senses can deceive us about the nature of the physical world but
never that they can deceive us about its very existence. It is with the De-
ceiver Argument, then, that Descartes implements his strategy of carrying
skepticism even further than the skeptics themselves as a preparation for
showing, in his subsequent Meditations, that even this radicalized skepti-
cism can be refuted.
In giving his Deceiver Argument, Descartes even raises the possibility
that God deceives him about simple mathematics. This is puzzling, for
two reasons. First, if Descartes is going to call even simple mathematics
into doubt, then his project of nding certainty seems doomed from the
start; for to carry out that project, Descartes intends to use philosophical
reasoning: he intends to use logic. But if even simple arithmetic can be
doubted, then why cant logic be doubted too? Consider, for example, two
of the most obvious rules of logic, called Modus Ponens and Modus Tollens,
respectively. Modus Ponens says that any argument of the form
If P, then Q.
P.
__________
? Q.
is valid. Modus Tollens says that any argument of the form
If P, then Q.
not-Q.
__________
? not-P.
is valid. Now these logical rules are not any simpler or easier to grasp
than + = . So if Descartes really means to doubt simple arithmetic, it
seems that he must also doubt the simplest rules of logic. But then how
can he legitimately use logic to overcome his doubt? This problem reap-
pears in an urgent way in Meditation III, as we shall see in chapter .
The second reason why Descartess calling mathematics into doubt in
the First Meditation is puzzling is that throughout that Meditation, Des-
cartes is examining beliefs based on the senses (as seen in point () of
our summary). But Descartes did not believe that mathematical beliefs
are based on the senses; he took them to based on the use of reason. So
why does he even mention mathematics at all in Meditation I? This ques-
tion can also be expressed in a different terminology, namely, the a
priori/a posteriori terminology that was made famous by Immanuel
Kant () and is now commonplace in philosophy. An a priori
statement is dened as one that can be known just by thinking, and
such statements are said to be items of a priori knowledge. By contrast,
an a posteriori statement (also called an empirical statement) is dened
as one that can be known only by experience, that is, by sense percep-
tion or by introspection of ones own feelings and moods; such state-
ments are said to be items of a posteriori, or empirical, knowledge.
Using this terminology, the key point is that Descartes believed that
mathematical statements are a priori; he did not believe that they are a
posteriori, or empirical. For example, he would have said (as would most
philosophers today) that + = can be known to be true just by using
Descartes
ones mindjust by thinking about what the statement says. One need
not consult ones experience (i.e., make any observations or perform
any experiments or introspect ones own feelings or moods) to be sure
that + = is true. This mathematical statement, like other mathe-
matical truths, is an item of a priori knowledge, not of a posteriori, or
empirical, knowledge. So our question can be put this way: Since Des-
cartes in Meditation I is examining a posteriori, or empirical, knowledge,
why does he even mention the a priori statements of mathematics?
A plausible solution to these puzzles has been oered by Harry Frankfurt.
Drawing on remarks that Descartes made when he was questioned about
the Meditations by a Dutch scholar named Burman, who recorded those
remarks in a volume titled Conversation with Burman, Frankfurt suggests
that throughout Meditation I, Descartes deliberately adopts a stance which is
not really his own considered or nal position, namely, that of a person who
believes that all knowledge rests on the senses. In other words, Descartes
poses as a philosophical beginner or novice, who naturally assumes that
all knowledge is a posteriori and starts for the rst time to reect critically on
this belief. As the argument of the Meditation unfolds, it becomes increas-
ingly evident to this novice that his stance is untenable, because of the
weaknesses of the senses that the Meditation itself brings to light. Thus, by
the end of Meditation I, the novice is prepared to give up his faith in the
senses and to receive the more authentic epistemology (= theory of knowl-
edge) that Descartes will oer in his subsequent Meditations. The relevance
of this point is that when Descartes mentions mathematics in Meditation I,
he is still thinking of it through the eyes of the novice, who believes that
mathematical knowledge, like all other knowledge, is empirical. Thus, nei-
ther the certainty regarding mathematics that Descartes expresses in his
response to the Dream Argument (point in our summary) nor the doubt of
mathematics that he expresses in the Deceiver Argument reect a correct
understanding of mathematics as a science not based on sense perception
but on reason. The upshot is that the doubt concerning mathematics that
Descartes expresses in Meditation I is possible only on what will turn out to
be the false assumption that mathematics is empirical.
Descartes anticipates two possible objections to his Deceiver Argument.
First, it might be said that God would not allow him to be always deceived,
since God is supposed to be supremely good. Descartes responds that if it

Frankfurt, Demons, Dreamers, and Madmen, p. .

Frankfurt, Demons, Dreamers, and Madmen, pp. .
were contrary to Gods goodness to allow him to be always deceived, then

it would also be contrary to Gods goodness to allow him to be sometimes
deceived. In order to avoid misinterpreting Descartes, it is important to
understand that he is not here (or anywhere while presenting the Deceiver
Argument) assuming that God exists. At this point in his Meditations,
where he is calling his beliefs into doubt, it would obviously be illegitimate
for him to make such an assumption. His point is that if God exists, then
there seems to be no reason why constant deception should be inconsis-
tent with his supreme goodness if occasional deception is not. This point
raises a diculty for anyone who believes in a supremely good God, but it
does not commit Descartes himself to assuming the existence of God.
Later in the Meditations (as we shall see), Descartes tries to prove that a
supremely good God exists, and after doing so, he himself tries to solve the
diculty just raised, by arguing that human error is due to our misusing
our own free-will, rather than to Gods deceiving us. But at this point, he is
merely showing that a simple appeal to Gods goodnessunsupplemented
by philosophical reection that would explain why he allows occasional
errorcannot answer the Deceiver Argument. The second possible objec-
tion that Descartes considers takes a tack almost contrary to the rst:
some people, Descartes says, might deny that there is a God powerful
enough to always deceive. Descartes responds that if he was created by some
source less powerful than God, then this only makes it more likely that he
is an imperfect being who is always deceived!
Let us enter the Deceiver Argument into our summary, as a response
to step ():
. No. For perhaps an omnipotent God has so created me that I hal-

lucinate the entire physical world, and even go wrong in doing
simple math. If you say that a supremely good God wouldnt
always deceive me, my reply is that sometimes he does allow de-
ception: so why not always? If you deny that I was created by an
omnipotent God, I answer that the less powerful was my cre-
ator, the more likely I am imperfect and so always deceived.
The argument of Meditation I is now substantially complete. The posi-

tion Descartes has reached is an utterly skeptical one:
I have no answer to these arguments, but am nally compelled to

admit that there is not one of my former beliefs about which a
Descartes
doubt may not properly be raised; and this is not a ippant or

ill-considered conclusion, but is based on powerful and well
thought-out reasons. (CSM II , SPW , AT VII )
Referring to all of the beliefs that he has so far surveyed (i.e., all beliefs
that, at least from the point of view of a philosophical novice, rest on the
senses), he adds: So in future I must withhold my assent from these
former beliefs just as carefully as I would from obvious falsehoods, if I want
to discover any certainty (CSM II , SPW , AT VII ). This statement
simultaneously rearms Descartess goal, reiterates his method of doubt,
and indicates (by the phrase, these former beliefs) the full, sweeping
range of beliefs that this method has now led him to withhold. Let us enter
Descartess skeptical conclusion into our summary:
. So far, I have found nothing that there isnt some reason to

doubt.
Before closing Meditation I, Descartes does one more thing. He notes that
it will be dicult to stick with his decision to withhold belief from all the
things that have now been found doubtful, especially since he has so long
been accustomed to taking them for granted. Accordingly, in order to coun-
terbalance his tendency to accept them, he adopts a special methodological
device. He deliberately pretends that there is a malicious demon of the
utmost power and cunning [who] has employed all his energies in order to
deceive me (CSM II , SPW , AT VII ). This evil demon does the most
striking thing mentioned in the Deceiver Argument of (): he causes Des-
cartes to hallucinate the entire physical world, including even Descartess
own body. We may enter this nal step into our summary, as follows:
. To be sure that I remain faithful to my resolution not to ac-

cept as true anything which isnt absolutely certain, I shall
deliberately assume that a powerful, evil demon is continually
deceiving me about the existence of the entire physical world,
including even my own body.
It is very important to understand that Descartes is not asserting that

there actually is such a deceiver. He does not know that to be true any-
more than he at this point knows that there is a (good) God. Rather, he is
using the possibility that there might be such a deceiver as a way of enforc-
ing his methodological doubt. Furthermore, the deceiver hypothesis, as
we shall call this device, will soon serve an additional, related function. In
Meditation II it will become a kind of litmus test for certainty. When
any given proposition, p, presents itself as being possibly certain (as a
candidate for certainty, so to speak), Descartes will ask: could the evil
deceiver fool me about p (make me falsely believe that p)? If the answer
is yes, then p is not certain and indubitable. But, should there be any
case where the answer is no, then Descartes will at last have discovered
something which is absolutely certain.
3. Is the Cartesian Doubt Self-Refuting?

Although the First Meditation is only a few pages long, it is a seminal philo-
sophic text that has evoked thousands of pages of commentary. Entire
books have been written about particular themes in it, such as the dream
argument and the deceiver argument. Obviously, we cannot possibly do
justice to all of this critical commentary in a work like this one. In this sec-
tion, however, we shall try to provide some assessment of the First Medita-
tion by considering one possible objection to Descartess doubt of the
senses. We shall consider this objection in relation successively to each of
the three reasons that Descartes gives for doubting the sensesthe decep-
tiveness of the senses, the Dream Argument, and the Deceiver Argument.
3.1 The Deceptiveness of the Senses

In step () of our summary of Meditation I, Descartes asserts that his
senses are sometimes deceptive and concludes that he should not trust
them completely. But how does Descartes know that his senses have
sometimes deceived him? The only possible answer seems to be: by using
his senses. For the only way to discover a perceptual error is by using the
senses themselves. For example, how does Descartes know that the
tower, which he perceived from afar to be round, is really square? Well,
by subsequently getting a better view of it, from a nearer distance. So,
he uses his sense of vision in order to discover and to correct his earlier
error. This point has led some critics of Descartes to raise an objection.
The only way a person can know that his or her senses are sometimes
deceptive is by using those senses themselves. Therefore, one cannot
use the premise that ones senses are sometimes deceptive to support
the conclusion that they may always be deceptive; for that very premise
could never be known if the conclusion drawn from it were true. One
Descartes
twentieth-century American philosopher, Maurice Mandelbaum, stated

this objection as follows:
[A]n epistemologist might ... argue that once we admit that the
same tower can look round from a distance and square when seen
close at hand, or that the same mountain can look one color and
then another, we are no longer able to maintain that the testimony
of the senses is a reliable guide to the nature of objects. His argu-
ment would be that if the senses can sometimes deceive us by vir-
tue of giving us diering reports, it is at least theoretically possible
that they always do so; or, at the least, he can challenge us to pro-
duce any clear criterion by means of which we can in every case
know when our senses deceive us, and when they do not.
However, ... one cannot prove that the senses actually do some-
times deceive us without assuming that they sometimes do not.
I would therefore contend that this skeptical argument is
self-refuting. It consists in drawing the conclusion that we can
never know whether our senses are deceiving us from the fact
that they actually do deceive us; however, ... this premisethe
statement that they do sometimes deceive uscould not itself
be known to be true if the conclusion of the argument, that we
can never know they are deceiving us, were itself taken as true.
Although Mandelbaum does not mention Descartes by name (and per-

haps would not have wished to be interpreted as specically criticizing
Descartes), this passage is an especially clear and forceful statement of
an objection that is commonly made against Descartes. Drawing on the
passage, we can break the objection down into three parts. First, a self-
refuting argument is dened as an argument whose premise(s) cannot be
known if its conclusion is true. Second, Descartess reasoning in step of
the First Meditation is interpreted as follows:
() If my senses sometimes deceive me, then it is possible that

my senses always deceive me.
() My senses sometimes deceive me.
_____________________________________________________
? It is possible that my senses always deceive me.

Maurice Mandelbaum, Philosophy, Science, and Sense Perception, p. .
Finally, it is pointed out that this argumentalthough a perfectly valid

case of Modus Ponensis self-refuting; for if its conclusion were true,
then its second premise could not be known. This is because, as we have
seen, the only way to know that ones senses are sometimes deceptive is
by using those senses themselves.
Before we consider whether this objection is damaging to Des-
cartess position, we should pause to note that skeptical arguments
are, in general, vulnerable to the charge of self-refutation. Suppose
that a skeptic were to advance an argument for the conclusion that
nobody knows anything. Then the obvious rebuttal would be: If no-
body knows anything, then you yourself do not know anything, and so
do not know the premises (or the conclusion) of your skeptical argu-
ment. This suggests that skeptics must not formulate their position
dogmatically as Nobody knows anything; they must formulate it
more cautiously. Indeed, it has sometimes been said that the only
rational course for a skeptic is to remain silentto refrain from arm-
ing or denying anything. But this goes too far, for there are ways that
skeptics can formulate their position that are not self-refuting. For ex-
ample, they can say that there are reasons for suspending judgment on
many matters normally assumed to be known. They can back this up by
advancing arguments on both sides of a question and just leaving mat-
ters there. Or they can argue that we lack certain kinds of knowledge,
such as perceptual knowledge, memory knowledge, or knowledge of
the future. But skeptics must not try to prove that we dont know
anything; and in giving their skeptical arguments, they must be careful
not to assume that they possess the very kind of knowledge those ar-
guments are supposed to impugn. A skeptic, then, must always walk a
ne line to avoid self-refutation.
So is Descartess reasoning in step () self-refuting? No, it is not; for
although Descartes is frequently misinterpreted on this point, he does
not make use of the argument (displayed on page above) that is often
attributed to him. He does not argue from the premise that his senses are
sometimes deceptive to the conclusion that they may always be decep-
tive. Rather, he concludes that he should not trust his senses com-
pletely. But this only means that some of his perceptions (i.e., those that
occur under poor conditions of observation) should not be trusted; for as
we have seen, Descartes explicitly indicates in his next steppoint in
our summarythat he does not mean to extend this conclusion to per-
ceptions occurring in good conditions of observation. Thus, critics who
Descartes
attribute this argument to Descartes have simply not read him carefully
enough. We can conclude that his reasoning in step is not self-refuting.
Of course, by the time Descartes gets to the end of Meditation I, he
does reach a thoroughly skeptical position with regard to sense percep-
tion. As one commentator has pointed out, Descartess overall argument
in Meditation I moves from saying that some perceptions are deceptive
(step ), to saying that any perception may be deceptive (because of the
Dream Argument, step ), to saying that every perception may be decep-
tive (because of the Deceiver Argument, step ). Does this mean that his
skepticism about the senses is self-refuting after all? To see whether it
does, we must consider the Dream Argument and the Deceiver Argument.
3.2 The Dream Argument

Some Descartes scholars have oered subtle, complex reconstructions of
the dream argument. But for our purposes, we can work with a fairly
simple version of it:
. I sometimes have vivid dreams that are qualitatively just like my best
(waking) perceptions.
. If I sometimes have vivid dreams that are qualitatively just like my
best perceptions, then I cannot distinguish with certainty between
my best perceptions and vivid dreams.
. I cannot distinguish with certainty between my best perceptions
and vivid dreams. [from () and ()]
. If I cannot distinguish with certainty between my best perceptions
and vivid dreams, then even my best perceptions provide no certainty.
. Even my best perceptions provide no certainty. [from() and ()]
A paraphrase of () could be: I can never be certain whether I am per-

ceiving objects under ideal conditions of observation or having a vivid
dream. A paraphrase of () could be: Even when I am perceiving objects
under ideal conditions of observation, I cannot be certain that I am.
Although this argument is valid, it is vulnerable to the charge of
self-refutation. That charge could be developed as follows. How does

Bernard Williams, Descartess Use of Skepticism, p. .

See, e.g., Edwin M. Curley, Descartes Against the Skeptics, pp. , and Margaret
Wilson, Descartes, pp. and pp. .
Descartes know that () is true? In order to know that vivid dreams are
just like his best waking perceptions, he would have to compare dreams
with waking perceptions and note their similarity. But in order to make
this comparison, he must identify his vivid dreams as dreams and his
best waking perceptions as waking perceptions and then note their
similarity. Now if he cannot tell his vivid dreams and his best waking
perceptions apart in the rst place, then he cannot identify the former
as dreams and the latter as waking perceptions; so he cannot make the
comparison needed to know whether they are alike or unalike. Con-
sider an analogy. Suppose someone showed you a drawer containing
one hundred one-dollar bills, informed you that fty of the bills were
counterfeit, and asked you to compare the real bills with the counter-
feit ones. Further, suppose that there were absolutely no detectable
dierence between the genuine and the counterfeit bills (that they
were indistinguishable). Then, isnt it obvious that you could not even
begin the comparison? But if () is true, then Descartes cannot tell
vivid dreams apart from waking perceptions anymore than you could
tell the real bills apart from the counterfeit ones. So he cannot make
the comparison needed to know that () is true. Thus if () is true, then
Descartes cannot know that () is true. Therefore, the rst step in the
argumentfrom () and () to ()is self-refuting; and we need not
even examine the second step, from () and () to ().
It seems that we must be concede that the Dream Argument is self-
refuting, for it seems to fully satisfy the denition of a self-refuting argu-
ment. However, it can be argued on Descartess behalf that this result is
not really damaging to his position; for Descartes uses the dream argu-
ment for a special, limited purpose, namely, to show that he cannot dis-
tinguish with certainty between vivid dreams and his best perceptions
and therefore that even his best perceptions can be doubted. But this
use of the Dream Argument does not require that Descartes know that
premise () is true, or that he be certain of its truth. It only requires that
he genuinely believe that the premise is true. Descartes, so to speak, nds
himself believing that he sometimes has vivid dreams that duplicate
waking perceptions occurring under even the best conditions of observa-
tion. And this mere belief gives him a legitimate reason to doubt whether
he can ever distinguish with certainty between vivid dreams and waking
perceptions.
The point underlying this defense of Descartes is that a ground for
doubt need not itself be something that you know or are certain of; it
Descartes
need only be something that you genuinely believe. To see this, consider
the following imaginary dialogue:
q: What time is it?

a: (looking at his watch) Its : P.M.
q: Are you absolutely certain of that?
a: No.
q: Why not?
a: My watch isnt always correct.
q: Are you absolutely certain of that?
Notice that even if A were to answer no to Qs nal question, A would still

have his grounds for being uncertain that the time is : P.M. It would
be absurd for Q to challenge A by saying: Well then, since you are not
certain that your watch is sometimes incorrect, you ought to be certain
that its : P.M. For such a challenge rests on a false assumption,
namely, that a ground or reason for doubt must itself be something of
which one is certain. For A to have some ground to doubt that it is :
P.M., he need not be certain that his watch is sometimes incorrect. The
mere fact that he genuinely believes that it is sometimes incorrect is a
legitimate reason for him to be less than certain that it is now : P.M.
Likewise, the mere fact that Descartes genuinely believes that he some-
times has dreams that are qualitatively just like his best perceptions is
a legitimate reason for him to be less than certain that he can distinguish
such perceptions from vivid dreams.
3.3 The Deceiver Argument

Even if what has been said so far in defense of Descartess doubt were not
wholly satisfactory, this would not matter very much; for Descartess
skepticism about the senses in Meditation I does not ultimately rest
either on the occasional deceptiveness of the senses (step ) or the Dream
Argument (step ). Those arguments are only preliminaries. They are
merely intended to shake our faith in the senses a little, so as to prepare
the way for Descartess main skeptical argument: the Deceiver Argument.
Whatever may be said about the other skeptical arguments of the First
Meditation, the Deceiver Argument is not vulnerable to the charge of self-
refutation. The basic reason for this is that the Deceiver Argument, like
those other skeptical arguments, is intended to show that the senses pro-
vide no certainty. So if its premises could only be known to be true by
using the senses, then the argument would be self-refuting. However, the
arguments premises are in fact all a priori statements, which can be
known just by thinking. So even if its conclusionthat the senses pro-
vide no certaintyis true, this does not prevent its premises from being
known. Therefore, it is not a self-refuting argument. Let us spell all this
out in more detail.
To do this, we need to use the concept of an analytic statement. An
analytic statement can be dened as a statement that is true solely in
virtue of the meanings of its constituent terms. A common example is
the statement, All bachelors are unmarried. Although this statement is
not worded as a denitionit does not say, the term bachelor means
the same as the term unmarriedit is obviously true by denition
(i.e., true solely by virtue of the meanings of the terms it contains).
Sometimes analytic statements are also called conceptual truths, since
they are true solely because of the relationships between the concepts
they involve. Analytic statements contrast with synthetic statements,
which are statements whose truth or falsity does not depend solely on
the meanings of terms. For example, the statement that All bachelors
are taxpayers is synthetic.
Now an important point about analytic statements is this: if a state-
ment is analytic, then it is a prioriit can be known to be true just by
thinking; for to know that such a statement is true, one need only under-
stand what it says. One need not consult experience (make any observa-
tions or perform any experiments or introspect ones feelings). Of course,
in order to learn the meanings of words, beings constituted, or wired,
like humans do need to have various sorts of experiences. At rst, this
fact might seem to conict with saying that analytic statementsor
indeed any statementsare a priori. But to see better why a statement
like All bachelors are unmarried is a priori, compare it again with All
bachelors are taxpayers. Even after one knows what the latter state-
ment means, one can still be totally in the dark as to whether it is true or
false; for its truth or falsity depends on facts about law and society that
can only be known by experience. On the other hand, once one has
learned what All bachelors are unmarried meanswhich admittedly
requires various sorts of experienceno further experience is needed to

Some contemporary philosophers, notably the American thinker W. V. Quine (
), reject the distinction between analytic and synthetic statements. This discussion
assumes that the distinction is a tenable one.
Descartes
know that this statement is true. Thus the statement can be known to
be true independently of any experience, except for the experience(s)
needed to learn the meanings of its constituent terms. And this is all
that is meant by saying that it can be known just by thinking, or is an a
priori statement.
Having dened the concept of an analytic statement, and noted the
principle that analytic statements are a priori, we can begin to see why
the premises of the Deceiver Argument are a priori. It is because the ar-
gument turns mainly on a purely analytic proposition, knowledge of
which is accordingly a priori. This proposition is the causal conception of
perception (CCP):
(CCP) For any person S and material object M, S perceives M at

time t only if M is a cause of Ss perceptual experience at t.
According to this conception, which underlies Descartess discussion of

perception throughout the Meditations and which is accepted by many
contemporary philosophers, as well, perception by denition involves a
causal element: the object perceived must be one of the causes of the
perceivers experience. For example, if S now sees a pen, then the pen
must be one of the causes of Ss visual experience. The pen is, of course,
not the only cause: the causes of Ss visual experience also include light
striking Ss retinae, processes in Ss nerves and eyes, and so forth. But the
pen must be a cause, or one of the causes, of Ss present visual experi-
ence. This is why CCP says that M is a cause, rather than the cause, of Ss
present perceptual experience.
Before showing exactly how the Deceiver Argument rests on CCP, we
should support our claim that CCP is an analytic truth. One way to do so
is to suppose that S were to say
(A) I see a pen, but it is not the case that a pen is one of the causes
of my present visual experience.
Surely, A is a contradiction: if the pen is not even one among the causes
of my present visual experience, then it is absurd to say that I see it. The
rst four words of A assert that I see the pen; and the remaining words
deny, or take back, this assertion. But the negation of a contradiction is
always an analytic statement. For example, the negation of Some bach-
elors are married (which is a contradiction since it means that at least
one individual is both married and not married) is All bachelors are un-
married, which, as we saw, is analytic. Now the negation of A is
(B) If I see a pen, then a pen is a cause of my present visual

experience.
One way to see why the negation of A is B is to note that the logical form
of A is p and not-q (e.g., John is wearing his right shoe and not wearing
his left shoe). The denial of this form is not (p and not-q) (e.g., it is not
the case that John is wearing his right shoe and not wearing his left
shoe). But the latter is equivalent to if p, then q (e.g., If John is wearing
his right shoe, then he is wearing his left shoe), which is the logical form
of B.] Thus, B is analytic. Now B means exactly the same thing as
(C) I see a pen only if a pen is a cause of my present visual

experience.
This is because the logical form of B, if p then q, is equivalent to the log-

ical form of C, p only if q. (To see this, note that If my car runs, then
there is gas in the tank says exactly the same thing as My car runs
only if there is gas in the tank.) So, since B is analytic and says the
same thing as C, C is analytic. But if C is analytic, then it seems evident
that analogues of C involving senses other than vision and persons
other than I (e.g., Mary smells a rose only if a rose is one of the causes
of her present olfactory experience, John tastes a potato chip only if a
potato chip is one of the causes of his present gustatory experience,
Henry hears a bell only if a bell is one of the causes of his present
auditory experience, etc.) are also analytic. But then CCP, which is
nothing but the general statement that summarizes all such particular
ones, is analytic.
A dierent demonstration that CCP is analytic has been oered by
the philosopher Paul Grice (). Grices demonstration goes as
follows. Suppose that S is having a visual experience exactly like seeing a
clock on the shelfor, as Grice puts it, that it looks to S as if there is a
clock on the shelf. Furthermore, suppose that there really is a clock on
the shelf within Ss eld of view, before his eyes. Is this sucient for it

A fuller presentation of this argument can be found in my Perceptual Knowledge,
pp. .
Descartes
to be true that S sees the clock? No, it is not; for suppose, as is logically
possible, that Ss visual experience is being produced by an expert di-
rectly stimulating Ss cortex or by some kind of post-hypnotic sugges-
tion, so that even if the clocks position on the shelf were altered or
the clock were entirely removed, Ss visual experience would remain
unchanged: it would continue to look to S as if there is a clock on the
shelf. In that case S does not see the clock, even though it is there
before his very eyes. And the reason he does not see it is that it plays
no part in causing his visual experience. Grices example shows very
clearly that in order for S to see the clock, it is not enough that () S
has a visual experience in which it looks to him as if there is a clock on
the shelf and () there actually is a clock there in front of Ss eyes.
Rather, S does not see the clock unless () the clock is also a cause of Ss
present visual experience. The concept or denition of seeing thus con-
tains an inexpungible causal element. It seems safe to generalize from
what is here true of vision to perception in general, and so to conclude
that CCP is a conceptual, analytic truth about the nature of perception.
(More precisely, CCP expresses a logically necessary condition, and so a
partial denition, of perception.)
How does CCP relate to the deceiver argument? Well, the basic point
made by the argument is that any perceptual experience that M causes S
to have might be exactly duplicated by God, or (to switch to the possi-
bility envisioned in Descartess deceiver hypothesis) by some powerful,
evil demon. Therefore, S can never be certain that M is causing the expe-
rience, and hence, given CCP, can never be certain that she is perceiving
M. Let us make this argument even more explicit. From CCP, it follows
that:
() I can sometimes be certain that I perceive a material object M

only if I can sometimes be certain that M is causing my percep-
tual experience.
Now we may also assert that
() I can sometimes be certain that M is causing my perceptual

experience only if it is not the case that any (every) perceptual
experience caused by M could be caused in some other manner.

H. P. Grice, The Causal Theory of Perception, p. .
But the key idea of the deceiver argument is that
() Any (every) perceptual experience caused by M could be

caused in some other manner (e.g., by an evil demon).
It follows from ()-(), however, that
() I can never be certain that I perceive M.
The conclusion, (), follows from premises ()-(), because the argument
has the valid form
p only if q.
q only if not-r.
r.
___________
? not-p.
Probably, you can see by reecting for a moment that this form is valid.
Another way to see this is to note again that p only if q says the same thing
as if p, then q. So, the above form is equivalent to this one:
If p then q.
If q then not-r.
r.
___________
? not-p.
But this last form is valid: from if q then not-r and r, one can deduce not-q
by Modus Tollens (and Double Negation); and then from not-q and if p
then q one can deduce the conclusion, not-p, by Modus Tollens.
Not only is the Deceiver Argument valid, but we can now see more
fully that it is not vulnerable to the charge of self-refutation. For premise
, following as it does from CCP, is itself analytic and so a priori. (This is
because if P is analytic and Q follows logically from P, then Q is analytic
too: analyticity is hereditary with respect to entailment. Strictly speaking,
() does not follow from CCP alone. Rather, it follows by Modus Ponens
from CCP together with If CCP, then (). But this last statement is itself a
long analytic statement, as can be seen by substituting for CCP and for
() the clauses that these labels abbreviate.) Premise , we may assert,
Descartes
is also analytic and so a priori: it depends for its truth solely on the mean-
ings of its constituent terms, notably the term certain. Furthermore,
premise expresses merely a logical possibility; and so our knowledge of
it does not depend on the senses either. Finally, it would be dicult to
maintain that the sorts of experiences required to learn the meanings of
the terms in (), (), and () must be genuine perceptions of reality. It
seems that even if, as in the evil-demon scenario, all sense experiences
were hallucinatory, there would be no reason in principle why one could
not learn the meanings of these terms. Consequently, the truth of the
arguments conclusion would not prevent the premises from being both
understood and known. Therefore, the argument is not self-refuting.
The deceiver argument still haunts the pages of contemporary books
and articles on epistemology, though nowadays it is usually put in a more
scientic, modern-sounding way. Any perceptual experience caused by
a material object stimulating ones sense-receptors (eyes, ears, nose,
etc.), it is argued, might instead be caused by a very advanced neurophys-
iologist (or team of neurophysiologists) directly stimulating ones brain
with painless electrodes. Perhaps, it is then suggested, all of our percep-
tual experience is caused in some such way, so that we never really per-
ceive material objects at all, but only hallucinate them. How can we
possibly know that this is not so, since our perceptual experience would
be exactly the same if it were so? It is not hard to recognize this line of
reasoning as being Descartess Deceiver Argument in modern dress.
As we shall see when we come to the Sixth Meditation, Descartes him-
self tried to refute the deceiver argument. He thereby hoped to answer
skepticism once and for all by refuting the most radical argument in its
favoran argument he had himself invented. To measure his success,
however, we must rst turn to Meditation II, where Descartes takes the
rst step toward answering the deep, unsettling skepticism generated in
Meditation I.

Some contemporary philosophers would object to this claim, on the ground that
it assumes the possibility of a private language and that this assumption was proved
wrong by Ludwig Wittgenstein () in his Philosophical Investigations. There is
no consensus among philosophers, however, whether Wittgensteins argument against
the possibility of a private language is sound; nor is there even agreement concerning the
exact nature of that argument.
| 2 |
Meditation II
The Cogito and the Self
1. Descartess I am thinking, therefore I exist

In the opening paragraph of Meditation II, Descartes reminds himself of
the deeply troubling doubts he raised in Meditation I. To convey how un-
settling these doubts are, he uses an image. He compares himself to a
man caught in a deep whirlpool, who is so disoriented that he can neither
stand on the bottom nor swim to the surface. Nevertheless, he says that
he will persist in his quest for certainty, by continuing to withhold assent
from everything that is doubtful, in the hope of nding just one thing
that is certain. Shifting to a different image, he compares himself to
Archimedes, the Greek mathematician who boasted that if he had a long-
enough lever mounted on one absolutely xed and immovable point,
he would be able to move the entire earth. Likewise, Descartes hopes
to nd just one absolutely certain piece of knowledge, so as to build an
entire system of knowledge upon it. The shift from the image of the man
tumbled around in the whirlpool to the image of Archimedes symbolizes
very well the path that Descartes will traverse in the Meditations, from
total doubt to certainty about many things. We may begin a summary of
Meditation II, then, as follows:
. I shall continue to withhold assent from whatever is even

slightly doubtful, just as though it were false, until I nd just
one thing that is certain.
Accordingly, Descartes reminds himself of what is in doubt. Perhaps

everything that he seems to see or remember is unreal. Perhaps body,

Descartes
shape, extension, movement and place (i.e., the entire material world)
are only illusions. Perhaps the sole certainty is that there is no certainty
to be had about the world. Nor can Descartes be certain that a God, or
any other being, has put these very doubts or thoughts into his mind. For
perhaps he produces them himself. But then, how about this self? Can its
existence be doubted too? To be sure, Descartes can and does doubt the
existence of his body. But, he now asks, does this mean that he can doubt
that he exists? This question has brought Descartes to the very brink of
discovery. So let us record it in our summary:
. I have convinced myself that perhaps the entire physical world,

including my own body, does not exist. Does it follow that per-
haps I do not exist?
Now comes Descartess answer: No: if I convinced myself of some-

thing (or thought anything at all [added by Descartes in the French
translation]), then I certainly existed (CSM II , SPW , AT VII ).
Here then, at last, is the rst certainty that he was seeking: it is his own
existence, as revealed in his very attempt to doubt it. Let us record this
moment of discovery in our summary:
. No: if I convinced myself of something, or even thought any-

thing at all, then, certainly, I existed.
Descartes immediately applies to this rst discovery his litmus test

for certainty, namely the evil demon hypothesis of Meditation IBut
there is a deceiver of supreme power and cunning, who is constantly de-
ceiving me. Of course, the words there is here mean suppose there is.
As before, Descartes is not asserting that there is a deceiver; rather, he is
using the possibility that there might be one to test the certainty of I
exist. And the result is that this statement unquestionably meets the
test; for how can the demon deceive Descartes unless Descartes exists?
So Descartes concludes that I exist is absolutely certain. To paraphrase
the test and its result:
. But what if there is an evil demon constantly deceiving me?

Then I certainly exist, if he is deceiving me; and he can never
cause me not to exist so long as I think that I exist. So I am, I
exist must be true whenever I assert it or think it.
Meditation II
Steps and contain Descartess famous proof of his own existence,

called the cogito. This Latin name comes from the formulation Cogito,
ergo sum, which means I am thinking, therefore I exist. This argument,
which was also expressed by Descartes in French as Je pense, donc je suis,
is usually rendered in English as I think, therefore I am (or exist). But I
am thinking, therefore I exist is arguably a better translation, for a rea-
son oered by John Cottingham:
[T]he correct English translation of cogito/je pense, when these

words occur in Descartes discussion of the certainty of his exis-
tence, should employ the so-called continuous presentI am
thinkingrather than the simple present, I think. For what
makes me certain of my existence is not some static or timeless
fact about methat I am one who thinks; rather, it is the fact
that I am at this moment engaged in thinking. And so long as I
continue to be so engaged, my existence is guaranteed.
Interestingly enough, Descartes never uses the sentence I am

thinking, therefore I exist in the Meditations. Furthermore, not every-
thing that Descartes says in steps and turns on the idea that he exists
just because he is thinking. For example, in step we nd the idea that to
be deceived he would have to exist. This is based on the general principle
that to be the object of some act (such as an act of deception) one must
exist. This principle has nothing special to do with the connection
between thinking and existing: another application of it might be that in
order to be punched in the nose, one must exist. Still, most of what Des-
cartes says about the certainty of his own existence, both in the Medita-
tions and in his other works, does turn on the idea that his thinking
proves his existence. Thus, in step , we can nd the arguments I con-
vinced myself of something, therefore I existed and I thought of some-
thing, therefore I existed; while in step we can nd the arguments I
now think that I exist, therefore I cannot now be caused not to exist and
I think at time t that I exist, therefore I exist is true at time t. Surely
these arguments are variations on Descartess fundamental idea: that his
thinking proves his existence. As for the famous sentence itself, one
place where Descartes does use it is in his Discourse on the Method, where
he says:

John Cottingham, Descartes, p. .
Descartes
I noticed that while I was thus trying to think everything false, it

was necessary that I, who was thinking this, was something. And
observing that this truth I am thinking, therefore I exist was so
rm and sure that all the most extravagant suppositions of the
sceptics were incapable of shaking it, I decided that I could accept
it without scruple as the rst principle of the philosophy I was
seeking. (CSM I , SPW , AT VI ; see also CSM I ,
SPW , AT VIIIA )
Since most of what Descartes says to show that his existence is certain
does turn on the connection between his thinking and his existence, we
shall restrict our attention from now on to the classic formulation of the
cogito as I am thinking, therefore I exist.
2. The Certainty of Ones Own Thoughts

Lets begin with a question that was raised by Pierre Gassendi, the author
of the fth set of Objections to the Meditations. (Descartes published the
Meditations together with seven sets of objections written by other phi-
losophers and theologians of his day, and with Descartess own replies to
these. This entire debate, called the Objections and Replies, is a fascinating
one, which we shall draw upon frequently in what follows.) Gassendi
said: But I do not see that you needed all this apparatus, when on other
grounds you were certain ... that you existed. You could have made the
same inference from any one of your other actions, since it is known by
the natural light that whatever acts exists (CSM II , M , SPW ,
AT VII ).
Descartess reply was this:
When you say that I could have made the same inference from
any one of my other actions you are far from the truth, since I
am not wholly certain of any of my actions, with the sole excep-
tion of thought... . I may not, for example, make the inference I
am walking, therefore I exist, except in so far as the awareness
of walking is a thought. The inference is certain only if applied to
this awareness, and not to the movement of the body which
sometimesin the case of dreamsis not occurring at all,
despite the fact that I seem to myself to be walking. Hence from
Meditation II
the fact that I think I am walking I can very well infer the exis-
tence of a mind which has this thought, but not the existence of
a body that walks. And the same applies in other cases. (CSM II
, M , SPW , AT VII )
Descartes is here making at least two points: (a) I am walking, and other
reports of my physical actions, are not certain; (b) I am thinking, and
more specic reports of my own thoughts (such as I think that I am
walking), are certain. This is why I am thinking, but not I am walking,
can be used as a premise from which to prove my existence.
Point (a) is easy enough to understand, in light of the doubt about every-
thing physical raised in Meditation I. Point (b), on the other hand, takes us
beyond the skeptical doubts of Meditation I and introduces us to one of
Descartess most important and inuential positive doctrines. This is that
each of us has absolutely certain, indubitable knowledge of his/her own
present thoughts. Descartess doctrine is not merely that I am thinking is
certain for each of us; it is much broader than that. The doctrine is that all
beliefs, assertions or judgments about ones own present thoughts enjoy a
special certainty that makes them immune to the skeptical doubts of Med-
itation I. We may formulate it like this. Let p be any statement that you
believe. Then even if p itself is uncertain, I am thinking about p is certain,
as is I believe that p. For example, even if there is a horse in the eld is
uncertain, I am thinking about there being a horse in the eld and I
believe that there is a horse in the eld are certain. Thus for every state-
ment that you think about or believe, whether it is certain or not, there is a
corresponding one (or more) which is certain. Also, let p be any statement
of the type: I perceive X (where perceive means see, touch, hear,
taste, smell). Then even if p is uncertain, I seem to perceive X is certain.
This will yield as many certain statements as there are things you seem to
perceive. For example, I seem to see a horse, I seem to touch a hand, I
seem to hear a car, I seem to smell a rose, I seem to taste a pear, and so
on are all statements that can be certain. Still other types of statements
that can be certain, according to Descartess doctrine, are:
I doubt that p. (e.g., I doubt that bodies exist.)

I judge that p. (e.g., I judge that this wax is hard.)
I want X. (e.g., I want to know more.)
I dont want X. (e.g., I dont want to be deceived.)
I imagine O. (e.g., I imagine a unicorn.)
Descartes
What all these and similar statements have in common is that they
describe only ones own present state of mind; they do not make any
claim about anything existing independently of ones own thinking.
(Notice that this is what distinguishes I seem to see a horse from I see
a horse: I seem to see a horse is to I see a horse as I think I am
walking is to I am walking.) Descartes calls the subject matter of such
statements cogitationes (Latin for thoughts, sg. cogitatio); his view is that
cogitationes constitute an easily overlooked but crucially important area
of certainty. For this reason, he is prepared to substitute any cogitatio-
statement (i.e., any statement about one of his own present thoughts,
like the statements just listed) for the premise I am thinking in the
cogito. For example, toward the end of Meditation II he uses the argu-
ment, I judge that a piece of wax exists, therefore I exist (CSM II ,
SPW , AT VII ); and in The Search for Truth he uses I doubt, there-
fore I exist. (CSM II , AT X ). Descartess view that all cogitatio-
statements are certain explains why he formulates the cogitos premise
in such a large variety of dierent ways, both in his Meditations and in
his other works.
How plausible is Descartess view? To answer this question, let us eval-
uate Descartess view by using his own test of certainty. To do this, we
can ask the following sequence of questions:
Could an evil deceiver make it the case that
. I think that there is a physical world, but there is no physical

world?
. I think that I have a body, but I do not have a body?
. I think that am I walking, but I am not walking?
. I think that I am thinking, but I am not thinking?
. I think that I see a horse, but I do not see a horse?
. I think that I seem to see a horse, but I do not seem to see a horse?
It should be obvious, in light of previous discussion, that the answer to

questions (), (), (), and () is yes. Notice, however, that the answer to
question () has to be no. The deceiver cannot possibly make me think
that I am thinking when I am not thinking, for to think that one is
thinking is to be thinking! Thus, I am thinking passes Descartess lit-
mus test for certainty. Indeed, even without using the device of the de-
ceiver, we can see that Descartes has made a valid point: One cannot
possibly think falsely that one is thinking, since to think that one is
Meditation II
thinking is already to be thinking. Thus, I am thinking is absolutely

indubitable. The point is exceptionally well put by Fred Feldman:
Lets consider the ... premise, I think ... Why should we view
this as being . . . certain for Descartes? The answer, I believe,
turns on a remarkable feature of thought. If you think you are
walking, you might be mistaken. Maybe you are only dreaming
that you are walking. However, if you think you are thinking, you
must be right. You cannot think that you are thinking, if you are
not thinking. So if an evil demon causes you to think that you are
thinking, he wont cause you to make a mistake. You will be right.
You will be thinking. If, in a dream, you think that you are
thinking, you will be right. There is no possibility of error here.
You cannot make a mistake, if you think you are thinking.
Notice that the point here is not that the evil deceiver could not make
you think that you are not thinking; no doubt he could do thathe could
make you gullible enough to think that you are not thinking when in fact
you are thinking. What he could not do is make it false that you are
thinking when you think that you are thinking.
What about question ()? Here matters are more complicated. To see
this, suppose that the evil deceiver confuses me about the dierence
between horses and zebras: he makes me think that horses, rather than
zebras, have stripes. Now, suppose that at a certain time, I have an expe-
rience as of seeing a striped, equine animal. (The phrase experience as of
seeing is here meant to indicate that the experience need not be an ac-
tual seeing: it might be an actual seeing, but it might also be a dream or a
hallucination, it does not matter which.) Is this a case where I wrongly
think that I seem to see a horse, so that the answer to () is yes? To answer
this question, we need to note that () is ambiguous, because seems has
more than one meaning. It can signify belief, as when the doctor says,
you seem to have an infection. Here the doctor is using seems to
express her belief that the patient has an infection. The word is also used
to express belief in such locutions as It seems to me that ..., It would
seem that ..., and It seems that ... But seem can also be used in a
very dierent way; namely, to signify the quality of an experience; as in
The moon seems at and yellow tonight. Here the speaker is not

Fred Feldman, A Cartesian Introduction to Philosophy.
Descartes
expressing a belief that the moon really is at and yellow. Rather, the
speaker is only describing how the moon looks. This ambiguity of seem
carries over to the statement I seem to see a horse. This statement could
mean (a) I believe that I see a horse; or it could mean (b) I have an ex-
perience as of seeing a horse. So question () could be a way of asking,
(a) Could it be true that I think that I believe that I see a horse,
though I dont believe that I see a horse?
Or it could be a way of asking,
(b) Could it be true that I think that I have an experience as of

seeing a horse, though I dont have an experience as of seeing a
horse?
Now it seems that the answer to (a) is no; for even if the animal that I
think I believe I see is striped, it is still true that I believe it to be a horse.
Just because I falsely believe that horses are striped, it does not follow
that I am not in a state of believing that I see a horse. On the other hand,
the answer to (b) seems to be yes; for if I am having an experience as of
seeing a striped animal, then I am wrong in thinking that it is an experi-
ence as of seeing a horse. It is really an experience as of seeing a zebra.
Perhaps, however, we can revise (b) so that the answer to it will again be
no, as follows:
(c) Could it be true that I think that I have an experience as of

seeing what I take to be a horse, though I do not have an experi-
ence as of seeing what I take to be a horse?
At this point, someone might object that if the deceiver were powerful
enough, he could cause me to go wrong in a dierent way. He might make
it true that I think I believe I see a horse, even though I really disbelieve
that I see a horse; or that I think I have an experience as of seeing what I
take to be a horse, even though I really have an experience as of seeing
what I take not to be a horse. Now it is not clear that such states of mind
as these are possible. But even if they are, we can defend Descartes
against this objection by using a point made in the previous chapter.
There we argued, you will recall, that it was legitimate for Descartes to
dismiss the possibility that he might be insane. But a person who thought
he believed he saw a horse while really disbelieving this, or who thought
Meditation II
he had an experience as of seeing what he took to be horse while having

an experience as of seeing what he took not to be a horse, would be so
radically confused about his own conscious beliefs as not even to be sane.
So, Descartes can legitimately assume that he is not in such a condition.
His basic question is not, How can an insane person use philosophical
reasoning to discover certainty? Rather, he is asking how a rational per-
son can use philosophical reasoning to arrive at certainty. It is already
built into this question that the person is not radically confused about
her own conscious beliefs. Descartess test for certaintycould the de-
ceiver fool me about p?must be understood to include the stipulation
that the person signied by me is sane.
It appears, then, that Descartess view about the certainty of cogitatio-
statements is plausible. The statement I am thinking can be absolutely
certain. And although this is a delicate matter about which more could
be said, it appears that many other (though perhaps not all) cogitatio-
statements can be certain too, provided they are carefully formulated.
3. A Problem for the Cogito

So far we have conned our discussion of the cogito to its premise, I am
thinking. The next matter to be considered is the inference from this
premise to the conclusion, I exist. At rst this inference looks obvious
and unproblematic. But it will prove to be as perplexing upon reection
as it was compelling at rst glance.
The basic diculty can be put very simply. What entitles Descartes to
use the rst-person pronoun I in the premise of his proof? Descartess
use of this pronoun seems to assume or presuppose the very thing he is
supposedly proving; namely, that he exists. Thus, the cogito seems to be
what is called a question-begging argument, that is, according to the stan-
dard denition of question-begging, an argument that takes for granted
or assumes the very conclusion that it is supposed to prove.
Bertrand Russell (), one of the twentieth centurys greatest
philosophers, makes a suggestion that might seem to solve the diculty.
He suggests that the I in I am thinking ought to be regarded only as a
grammatical convenience, rather than as referring to a self or person who
exists as something distinct from the thinking. After all, it would not
have been grammatical for Descartes to express the cogitos premise
merely as thinking or as am thinking: the I is needed to obtain a
Descartes
grammatical sentence. But the I need not therefore refer to something

that exists over and above the thinking any more than the pronoun it
in the sentence it is raining refers to something distinct from the rain.
Indeed, that it could be eliminated without changing the sentences
meaning, by saying Rain is falling. Likewise, Russell suggests, the pre-
mise of the cogito, strictly speaking, asserts only There is a thought or
Thinking is occurring now. If this premise is substituted for I am
thinking, then the argument no longer begs the question.
Unfortunately, Russells suggestion does not solve the diculty; for it
is certainly not valid to argue, there is a thought, therefore I exist. For,
supposing I had ceased to exist but someone else were thinking, the pre-
mise would be true and the conclusion false. His suggestion, if taken as
an attempt to defend the cogito, goes from the frying pan into the re.
Nor would it help to add the premise If there is a thought, then I exist,
for the same supposition shows that this premise is false.
The problem that the cogito seems to run into, then, is this. If we stick
with Descartess famous formulation, I am thinking, therefore I exist,
then the argument appears to be question-begging, whereas if we sub-
stitute the Russellian formulation, There is a thought, therefore I exist,
then the argument is invalid. Thus, the cogito appears to be either
question-begging or invalid.
Some philosophers, motivated by this problem or variations on it,
have tried to interpret the cogito as something other than an argument
for ones existence. For example, in the Finnish philosopher Jaakko
Hintikka, in a famous article in The Philosophical Review entitled Cogito,
ergo sum: Inference or Performance? argued for what he called the per-
formatory rather than inferential interpretation of the cogito. Hintikkas
basic idea was this. Suppose that you try to doubt your own existence.
You will immediately discover that this is impossible, because in the very
attempt to think that perhaps you do not exist, your existence is mani-
fest. Furthermore, your certainty that you exist is not then based on any
argument; rather, it is based solely on the act or performance of trying to
doubt your existencean act that causes you to see that you exist, almost
as playing music causes sound. The relation between the thinking in the
cogito and the realization that I certainly exist is not one between premise

Bertrand Russell, A History of Western Philosophy, p. . Russells exact formulation
is There are thoughts.

I do not mean to imply that Russell himself oers his suggestion as a defense of the
cogito, but only to fault the idea that it might successfully serve as such a defense.
Meditation II
and conclusion, but rather between process and product, so that the no-
tion of question-beggingness cannot even be applied.
This novel interpretation of the cogito, however, has been generally
rejected by Descartes scholars, for two dierent reasons. First, it makes the
certainty that I exist depend narrowly on only one specic thought; namely
the attempt to doubt my own existence. But, as we have seen, Descartes
believed that any of his thoughts, regardless of its content, established his
existence. Second, when one asks exactly why trying to doubt my existence
causes me to be certain that I exist, the only clear answer seems to be that I
accept the argument: If I try to doubt my existence, then I exist; I am trying
to doubt my existence; therefore I exist. But then the performatory cogito
reduces to an inference or argument after all. Accordingly, it will be best for
us to pursue the topic of the cogito by considering why Descartes thought
that I am thinking, therefore I exist is a successful argument for, or proof
of, ones own existence. Such an approach has the further advantage that it
will bring out several underlying principles of Descartess thought.
4. The Substance Theory

A key underlying principle that Descartes appeals to when explaining the
cogito to his critics is what I shall call the substance theory. This section
will temporarily digress from the cogito in order to expound the sub-
stance theory and the main argument for it, so that we can then see, in
the following section, how this theory can be used to support the cogito.
The substance theory is an answer to the philosophical question: What
is a thing? In order to understand this rather strange-sounding question,
we must rst understand a distinction that philosophers make, between
a thing and a property. To appreciate the need for this distinction, reect
on the following, fallacious argument.
() Socrates is Socrates.
() Socrates is snub-nosed.
() Socrates is fat.
() Socrates is wise.
________________________________
? One thing (Socrates) is many things.
The conclusion is absurd: a single thing cannot possibly be identical with

many different things. What then is wrong with the argument? The
Descartes
answer is that the is in (), (), and () is being understood in the same
sense as the is in (). But this is an error. For the is in () is what logi-
cians call an is of identity; while the is in ()-() is an is of predica-
tion. The is of identity is used to assert that a designated item is one
and the same entity as some item, as for example in The Morning Star
(Venus seen in the morning) is the Evening Star (Venus seen in the
evening). The is of predication is used to attribute a characteristic or
property to a thing, as in Venus is round. To see this distinction more
clearly, notice that we could substitute the sign = for the is in ();
while it would be wrong to do so in ()().
Corresponding to the logical distinction between these two iss, is a
fundamental distinction within reality: that between a thing and a char-
acteristic or property. Socrates designates a thingan animate, living
one. But snub-nosed, fat, and wise do not designate things; they
designate properties. What the fallacious argument about Socrates shows
is that in order to avoid the one-many paradox generated in its conclu-
sion, we must include properties as well as things in our ontology. (The
word ontology means an account of what there is. It is derived from
logos, Greek for account, and onto, Greek for being.)
Having made the ontological distinction between a thing and a prop-
erty, we can return to the question: what is a thing? There are two tradi-
tional, opposed answers to this question. The bundle theory holds that a
thing is merely a collection of coexisting properties. For example an
apple, according to the bundle theory, is nothing but roundness, redness,
tartness, squashiness, and so on coexisting at a certain place and time.
The substance theory holds that a thing is composed of various properties
plus an underlying substance to which these properties belong. The apple,
on this view, is composed not just of the properties just mentioned, but
also of an underlying substance in which all these properties are said to
inhere. The bundle theory is favored by Empiricist philosophers, such
as Berkeley and Hume in the eighteenth century and Bertrand Russell in
the twentieth century. The substance theory was upheld by Aristotle and
most medieval thinkers, and in the Modern period by the Rationalists
Descartes, Spinoza, and Leibniz. Since the substance theory is the richer
onepostulating more in a thing than does the bundle theorylet
us inquire into the rationale for it. Why not adopt the bundle theorys
simpler view that a thing is just a collection of coexisting properties?
The most salient answer is provided by an argument that we shall call the
argument from change. This argument, which dates back at least to Aristotle,
Meditation II
can be discerned in Descartess famous illustration in Meditation II of the

piece of wax. Suppose, Descartes says, that I heat a piece of wax that has
been freshly brought from a beehive. As I heat the wax, its properties change:
its hardness is replaced by a soft, gooey texture, its lumpish shape by an
elongated one, its brown color by a translucid tint, its fragrant aroma by a
smoky smell. Even its capacity to make a tapping noise when struck with a
nger is lost. Yet one and the same piece of wax still exists despite all these
changes. Why is this so? Why is it not the case, instead, that the wax ceases
to exist, and that another, new object begins to exist? The answer proposed
by the argument from change is that although the properties of the wax
have changed, the underlying substance has not: one and the same substance
still exists, and continued to exist throughout the process of change.
The argument from change is an important one, so let us try to formu-
late it carefully. In order to do this, we must rst make a distinction
between a determinate and a determinable property. A determinate prop-
erty is one that is absolutely specic, and a determinable property one
that is not. Consider, for example, the property of being elliptical (ellip-
ticalness). This is a determinable, rather than a determinate, property,
because there are many dierent elliptical shapes: highly elongated ones,
moderately elongated ones, and so on. A determinate property must be
absolutely specic. So, only an ellipse satisfying a particular equation
would be a determinate property. Color is another example of a deter-
minable property, since there are many dierent shades of color. Even
particular colors like red, blue, and yellow are determinables rather than
determinates, since there are many shades of each of these colors. Only
absolutely specic shades of a color are determinate properties. (It does
not matter that we may not have a name for each such shade.)
Having distinguished between determinate and determinable prop-
erties, we can state the argument from change in proper form:
() We can distinguish between (a) all of a things determinate

properties changing without the things ceasing to exist, and (b)
a things ceasing to exist.
() We can distinguish between (a) and (b) only if a thing is com-
posed, in addition to its properties, of a permanent, underlying
substance.
____________________________________________________
? A thing is composed, in addition to its properties, of a perma-
nent, underlying substance.
Descartes
To grasp this argument, think again of Descartess example of the melting

wax. The example is supposed to illustrate a case of (a), in contrast to (b).
Premise () can be seen as a reply to the challenge, What justies us in
taking it to be a case of (a) rather than (b)that is, in thinking that the
wax, all of whose determinate properties have changed, hasnt therefore
ceased to exist? The conclusion, which follows validly from the premises,
is a statement of the substance theory itself. Thus, the argument pur-
ports to demonstrate that the substance theory is correct.
You may be wondering why the term determinate is needed in pre-
mise . The answer has to do with the nature of change. We say, of course,
that when a thing alters, its properties or characteristics have changed.
But what does this mean? Not that the properties themselves have
changed, because a property itself can never change: red, for example, is
just red, and to say that it had changed would really mean that it had
been replaced by another property, say by the property blue. Thus, to say
that a things properties have changed is to say that it has lost certain
properties and acquired others. Suppose, then, that the word determi-
nate were left out of premise . Then the premise would say that a thing
could lose all of its properties, including all its determinable properties,
without ceasing to exist. But this would be false; for a things determin-
able properties include shape and sizenot this or that specic shape
and size, but just having some shape and size or other, or having what
Descartes called extensionand a (physical) thing cannot lose all shape
and size whatsoever without thereby also ceasing to exist. Thus, the term
determinate must be included in premise for that premise to have a
chance of being true.
You may now ask: why then isnt the word determinate also included
in premise ? The answer is that to preserve the arguments validity, the
conclusion would then also have to include that word, and so would have
to read:
A thing is composed, in addition to its determinate properties, of

a permanent, underlying substance.
But this is false; for it means that a thing is composed only of its deter-
minate properties and its underlying substance. But a determinate
property cannot possibly be present unless a corresponding determin-
able property is also present; for example, squareness cannot be pre-
sent unless shape is present, and redness cannot be present unless color
Meditation II
is present. Thus, in order for the arguments conclusion to stand a

chance of being true, the word determinate must not be included in
premise .
Before seeing how the substance theory can be used to support
Descartess cogito, we must be sure to understand two further points.
First, in his discussion of the piece of wax, Descartes is not suddenly
assuming that physical objects exist. Not until the Sixth and last Med-
itation will Descartes overcome his doubt concerning the existence of
the physical world. So, the entire discussion of the wax in Meditation
II is hypothetical: Descartes is considering the question, if physical
things did exist, what would they be? Second, although the argument
from change and the substance theory are more easily grasped by ref-
erence to examples of physical objects such as the wax, Descartes
would apply exactly the same reasoning to the purely mental or spiri-
tual thing that he will call the mind or soul. At this point in the Medi-
tations, Descartes has not yet established what he later will try to
provethat the mind is a purely mental or immaterial substance. But
to avoid a likely, yet radical, misunderstanding of Descartes, it is cru-
cial to realize that for Descartes the argument from change and the
substance theory apply to such a nonmaterial thing no less than to
a material one. Indeed, his use of wax as an example in giving the ar-
gument from change is, in one way, misleading: the thing whose exis-
tence the cogito is supposed to prove is not a physical substance but a
nonphysical, purely mental one. It is because we (Descartess readers)
are so familiar with physical objects and because Descartes has, in any
case, not yet shown that the mind is not physical, that he uses wax as
an example. But from the point of view of the doctrine about the
mind toward which he is building, it would have been more accurate
for him to apply the argument from change directly to the mind. And
in the Synopsis of the following six Meditations that he prefaced to the
Meditations, where he is not trying to work toward his doctrines in a
step-by-step manner but only summarizing some of his main results,
he does just that:
[T]he human mind ... is a pure substance. For even if all the ac-
cidents of the mind change, so that it has dierent objects of the
understanding and dierent desires and sensations, it does not
on that account become a dierent mind. (CSM II , SPW , AT
VII )
Descartes
The term accidents here refers to properties of a certain sortnamely,

accidental properties. (Accidental properties are those that a thing hap-
pens to have but need not have in order to be what it is, e.g., a certain
triangles property of being blue. Accidental properties contrast with es-
sential properties, which are those that a thing must have in order to be
what it is, e.g., the triangles property of being three-sided.) Further, Des-
cartess phrase dierent objects of the understanding and dierent
desires and sensations here refers to specic, determinate thoughts. So
Descartes identies the accidental properties of a mind with its specic,
determinate cogitationes. But if a minds accidental properties are specic
or determinate thoughts, then those properties are also determinate
properties. Therefore, Descartess argument in the above passage could
be briey summarized as a human mind is a substance, since even if all
its determinate properties change, it is still the same mind. This is simply
the argument from change, now applied directly to the mind.
5. A Reconstruction of the Cogito Based

on the Substance Theory
We are now in a position to show how the substance theory relates to the
cogito. We shall do this by giving a step-by-step reconstruction of the
cogito that uses the substance theory as its very rst premise. The rst
premise of our reconstruction, then, is
() A thing is composed of its properties or characteristics plus an

underlying substance to which they belong.
To understand this premise, it is important to grasp what the word

thing means in it. After all, thing is the vaguest noun in the English
language. Depending on the context, it can be made to stand for virtually
anything, including shadows, surfaces, edges, black holes, symphonies,
ideasthe list could be prolonged indenitely. Now in (), thing is
being given a specic meaning: it signies an entity that could conceivably
exist independently of all other entities. The reason for assigning this
meaning to thing in () is that it is in this sense that the word is under-
stood in the philosophical question that the substance theory claims to
answer, namely, What is a thing? This question is not about such
things as shadows, surfaces, borders, or other entities that could not
Meditation II
conceivably exist apart from some other entity. It is a question about

what some philosophers call concrete thingsentities that could at
least conceivably exist on their own, apart from any other entities.
What premise () says, then, is that the minimum that can exist on its own
is always a substance and its properties. There could not be a free-oating
property, that is, a property existing without a substance to which it belongs;
nor could there be a substance without any property. So there follows di-
rectly from premise (i.e., as a corollary of the substance theory),
() If there is a property or characteristic, then there must be a

substance to which it belongs.
This corollary, which rules out free-oating properties and is some-

times called the substance-property principle, is explicitly put forward
by Descartes in several places. For example, in his Reply to the fourth set
of Objections, he says,
[A]ttributes [= properties] ... must inhere in something if they

are to exist; and we call the thing in which they inhere a sub-
stance. (CSM II , AT VII )
Likewise, in his Principles of Philosophy (a work that presents, in a more

didactic manner, the same system of thought as the Meditations and also
includes much of Descartess physics), Descartes says:
[W]e should notice something very well known by the natural

light: nothingness possesses no attributes or qualities. It follows
that, wherever we nd some attributes or qualities, there is nec-
essarily some thing or substance to be found for them to belong
to. (Part I, # [CSM I , SPW , AT VIIIA ])
[N]othingness possesses no attributes, that is to say, no prop-
erties or qualities. Thus, if we perceive the presence of some at-
tribute, we can infer that there must also be present an existing
thing or substance to which it may be attributed. (Part I, #
[CSM I , SPW , AT VIIIA ])
These quotations call for a short digression. Notice that in them, Des-
cartes tries to deduce () from the principle that nothingness possesses
Descartes
no propertiesa principle which was a commonplace of medieval phi-

losophy. His argument seems to be that
(i) Nothingness has no properties.

(ii) If there were a property that did not belong to anything, then
nothingness would have a property.
_____________________________________________________
? (iii) Every property belongs to something.
This, however, is a weak argument. For even if we grant (i), (ii) is not
obviously true. Why should the fact that there is some free-oating
propertyone that does not belong to anythingbe thought to imply
that this property belongs to nothingness (i.e., to nonbeing)? Further-
more, (iii) does not imply that the something to which a property
belongs is a substance. Why could it not just belong to a thing, conceived
as a bundle or cluster of properties? So it seems that () cannot really be
established in the way Descartes here suggests. On the other hand,
() does not need to be established in that way, since, as we have already
seen, () does follow directly from (). So the failure of the argument
from (i) and (ii) to (iii) does no damage to our reconstruction.
Let us return to that reconstruction. Its next step is an assumption
that Descartes makes. This is
() A thought is a property.
This assumption is quite explicit in the passage quoted earlier from the
Synopsis of the Meditations, where Descartes characterizes thinking,
willing, and perceiving certain things as accidents (i.e., accidental prop-
erties, as we saw) of the mind. For Descartes, then, a persons thought of
an apple is related to the person in the same way as an apples roundness
is related to the apple: just as an apples roundness is a property of the
apple, a persons thought is a property of the person. This assumption
seemed obvious to Descartes; let us reserve comment on it until the next
section.
An important consequence of the Assumption and the Corollary (pre-
mises and , respectively) is
() If there is a thought, then there is a substance to which it

belongs.
Meditation II
Descartes himself, in responding to Thomas Hobbess Objections to Med-

itation II, invokes this principle, deducing it from ():
[I]t is certain that a thought cannot exist without a thing that is

thinking; and in general no act or accident can exist without a
substance for it to belong to. (CSM II , AT VII )
The general statement after the semicolon is () and is used to support

the statement before the semicolon, which is (). (Thing must here
mean substance; otherwise the rst statement would not follow from,
or be a particular application of, the second.) The other premise required
to deduce () from (), namely (), is left unstated or understood by
Descartes.
Next, remember that the premise of the cogito, once the apparently
question-begging I is expurgated from it, becomes simply:
() There is a thought.
This much at least, Descartes claims to know indubitably: recall his doc-
trine that each of us can have absolutely certain knowledge of our own
present thoughts, and remember also that () is the premise that Ber-
trand Russell was prepared to give to Descartes.
Finally, Descartes moves from () and () to the conclusion
() There is a substance to which this thought belongs: I.
This completes the reconstruction of Descartess argument.

This reconstruction, we suggest, faithfully captures at least one way in
which Descartes explains his cogito. The least that can be said for it is that
no other construal (certainly none that treats the cogito as something
other than an inference) shows as well how the cogito relates to meta-
physical principles that Descartes accepted, and that he himself invoked
when defending the cogito against objections. The reconstruction also
shows how Descartes might have defended his cogito against the objec-
tion that it is either question-begging or invalid. For it shows how he
could have thought it possible to prove his existence from the premise
There is a thought, with the help of certain further premises that he
accepted. Anthony Kenny, whose discussion of the cogito can be highly
recommended, summarizes the argument in essentially the same way as
our steps ()():
Descartes
When he thinks, he is aware of a thoughtno matter, yet, to

whom or what the thought belongs [= ()]. Since, by the light of
nature [by the light of nature means, for Descartes, by the
light of reasoni.e., by reason employed in the very best way],
he knows what a thought is, he knows that it is an attribute and
not a substance [= ()]. Again, by the light of nature, he knows
that every such attribute must belong to a substance [= () and
()]. So he concludes to the existence of the substance of which
the thought he perceives is an attribute. This he calls ego; or, if
you like, he concludes that the I in I am thinking does refer to
a substance and is not just a grammatical convenience. [= ()]
Kenny does not point out that () derives from () and is thus a corollary
of the substance theory; the logical starting point of his reconstruction is
the second step of ours. Aside from this dierence, the two reconstruc-
tions are the same.
6. Critical Discussion of the Reconstructed Cogito

Let us now inquire whether the cogito, as we have reconstructed it, is
successful. We shall proceed by commenting separately on each of its
various steps.
6.1 The Substance Theory and the Argument from Change

The substance theory is not without its problems. The main diculty with
the theory is that substance is unperceivable. Imagine, for example, that
you want to see the substance of a block of wood. So, you obtain a carpen-
ters plane and plane away a layer of wood. What do you then see? Well,
you see a new set of propertiesa new size (slightly smaller), shape, and
color. It is obvious that planing away still another layer will not get you any
closer to seeing the blocks underlying substance. No matter how many
layers you plane away, you will see only more propertiesuntil, in the end,
all the wood is gone. What this kind of thought-experiment shows is that
substance is just not something that can be perceived: nothing would even
count as perceiving it. Substance is, as philosophers say, unperceivable in

Anthony Kenny, Descartes, p. .
Meditation II
principle. For this reason, Empiricist philosophers in the seventeenth and

eighteenth centuries became increasingly critical of the substance theory;
and many contemporary philosophers reject substance altogether.
In order to reject the substance theory, however, one must refute the
argument for it, namely the argument from change. This is not easy to
do, and we shall not attempt to do it in this work. But we can protably
make some remarks about the prospects of refuting the argument.
Premise can be challenged, in the following way. Can we really distin-
guish between all of a things determinate properties changing and the
thing ceasing to exist? In order for a thing to continue to exist, must it
not retain at least some of its determinate propertiesif only a single
one? And isnt this condition in fact satised in the case of Descartess
piece of wax?for although the waxs observable properties (e.g., its
size, shape, color, texture) have all changed, doesnt the wax still retain a
certain chemical compositionor perhaps a certain atomic structure
in virtue of which it continues to exist despite the change?
A possible defense against this line of attack can be given by appealing
to the following thought-experiment. Suppose that chemists or atomic
physicists informed us that when a piece of wax is melted, its color,
shape, hardness, and other observable properties are not the only ones
that change. Rather, its chemical composition and/or atomic structure
are also altered, perhaps only very slightly. Would we then feel bound to
agree that the piece of wax Descartes describes ceases to exist when we
melt it? It may well seem that we would notat least not merely on the
strength of what we have supposed so far, namely, that both the waxs
more obvious properties and its scientic ones have altered; for if, by
cooling and molding the melted stu, we could easily get it back to its
former shape, size, color, texture, and consistency, use it as a candle or a
water-repellent, and so forth, then would we not still consider it to be
wax? If so, then it seems that what accounts for the waxs continuing to
exist when melted is not that it retains scientic properties such as its
chemical composition or atomic structure. Those could change, as well as
the more obvious properties of shape, size, hardness, color, and so on.
Also, if the wax, when exposed to heat, kept its more obvious properties
and only its scientic properties changed, would it not still be wax? So,
premise may well seem to withstand the challenge we raised against it.

There is a fuller critical discussion of the argument from change in Georges Dicker,
Humes Epistemology and Metaphysics: An Introduction, pp. .
Descartes
Some philosophers, whom we may call scientic realists, would not

be impressed by this possible response. They would insist that retaining
some determinate scientic property is what enables a thing to persist
through change. They would support their position by saying that if, for
example, a quantity of liquid lost the molecular structure HO, then it
would cease to be water, no matter how much it might still be like water,
quench our thirst, and so forth; or, to take another example, that if a
piece of yellow shiny metal ceased to have atomic number , it would
cease to be goldit would become fools gold. If this line of thought is
correct, then it seems to show that premise () of the argument from
change is false.
However, we shall not further pursue the question of whether that pre-
mise is true or false, because even if scientic realists are right in saying that
a things retaining a certain determinate scientic property is a necessary
condition for it to persist through change, it would not follow that substance
is not needed to give a satisfactory account of what is required for a thing to
retain its identity through change. For although scientic realists may be
right in saying that a things retaining a certain determinate scientic prop-
erty is a necessary condition for it to persist through change, it cannot be a
sucient condition. To see why, consider the following thought-experiment.
Suppose that a gold statue occupies a place P at a time t, that at time t it
vanishes from that place, and that at time t, an exactly similar gold statue
begins to occupy a place P. Further, suppose that it is not the case that the
statue traveled or was carried from P to P; rather, the gold statue simply
ceased to occupy P and, later, an exactly similar gold statue began to occupy
P. Further, the statue at P has the same atomic number, , as the one at
P. Now let us ask: was the statue at P numerically identical with the statue
at P? It is important to understand what this question means. It is not
asking whether the statue at P was qualitatively the same as the one at P.
By hypothesis, it was: we just said that the statue at P was exactly like the
one at P, meaning that it had the same weight, color, shape, size, workman-
ship, and so forth. Furthermoreand this is importantthe statue at P
had exactly the same atomic structure, atomic number , as the statue at
P. Thus, the statue at P had exactly the same properties, including the sci-
entic ones, as the one at P. The question, then, is only whether the statue
at P and the one at P were numerically the same: was there just one statue
which existed at P at the earlier time and also at P at the later time, or were
there two statues, one at the earlier time and another at the later time?
(Was the situation like that of one person at two dierent times, or like
Meditation II
that of two identical twins? Was the statue at P the same one as the statue
at P, or only the same as the one at P?) The answer, it would seem, is that
the statue at P and the one at P were not numerically identical. The rea-
son is that when the statue at P ceased to occupy P, it ceased to exist,
since it did not then travel to P (or, we may also imagine, to any other
place). Further, when the exactly similar gold statue began to occupy P at
t, it began to exist, since it had not traveled from P (or from any other
place). But one thing cannot have two beginnings of existence. So, the
statue at P and the statue at P were qualitatively identical, but numeri-
cally distinct (like the twins). If this is right, then it shows that retaining
the same scientic properties it not a sucient condition for a thing to
endure through change, since the statue at P and the statue at P both had
atomic number .
In light of this point, a proponent of the substance theory might revise
the argument from change to go as follows:
(R) We can distinguish between (a) all of a things determinate

properties except for scientic property P (say, the things chemi-
cal composition or its atomic structure) changing without the
things ceasing to exist and (b) a thing ceasing to exist and an-
other thing with the same scientic property P beginning to exist.
(R) We can distinguish between (a) and (b) only if a thing is
composed, in addition to its properties, of a permanent, under-
lying substance.
____________________________________________________
? A thing is composed, in addition to its properties, of a perma-
nent, underlying substance.
Many contemporary philosophers, however, would reject this version of

the argument from change as well, for they would reject premise R.
Drawing on ideas in John Lockes An Essay Concerning Human Understanding
(), they would say that what is needed for a thing to continue existing
through time and change is not some sort of mysterious and unperceivable
substance but, rather, that it must meet the requirement of spatio-tempo-
ral continuity. Roughly speaking, the idea is that a thing continues to
exist, or retains its identity through time, only if its career in space and

See John Locke, An Essay Concerning Human Understanding, Book II, chapter .
Descartes
time is not interrupted, but is continuous. In other words, it must not be

the case that, like the vanishing statue in our example, the thing ceases to
occupy its place, and that a thing with the same properties begins to occupy
a dierent place (or the same place later), without the things moving from
one place to the other (or returning to the same place later).
Although Locke endorses the requirement of spatio-temporal conti-
nuity, he argues, in eect, that it too is only a necessary condition for
identity through time. Locke argues that a thing must also remain the
same sort of thing; it cannot, for example, go from being a piece of paper
to being a mound of ashes, even if the paper is spatiotemporally contin-
uous with the asheseven if, say, the paper is ignited and the burning
paper is carried from place to place , at which point it has become just
ashes. This additional requirement yields the now-popular idea that iden-
tity through change may be dened, without any appeal to substance, as
spatio-temporal continuity under a sortal.
There is reason to think, however, that Descartes would not agree that
the above considerations refute the argument from change. To see why,
we need to look rst at a second possible objection to the original version
of the argument from change, and then to see how Descartes would
answer it. This possible objection is directed against premise ()that
we can distinguish between (a) all of a things determinate properties
changing without the things ceasing to exist, and (b) a things ceasing to
exist only if a thing is composed, in addition to its properties, of a perma-
nent, underlying substance. The objection is that this premise should be
replaced with this one: we can distinguish between all of a things deter-
minate properties changing without the things ceasing to exist, and (b)
a things ceasing to exist only if a thing is composed, in addition to its
properties, of at least one determinable property. The entire argument
would then go this way:
() We can distinguish between (a) all of a things determinate

properties changing without the things ceasing to exist, and (b)
a things ceasing to exist.
(R*) We can distinguish between (a) and (b) only if a thing is
composed, in addition to its properties, of at least one determin-
able property.
____________________________________________________
? A thing is composed, in addition to its determinate properties,
of at least one determinable property.
Meditation II
Now this possible objection will not do, because (R*) is false. For it does
not seem that merely retaining some determinable propertyfor ex-
ample some shape or other, or some size or other, or some color or other,
and so forthis a sucient condition for a thing to retain its identity
through time though all of its determinate properties have changed,
even if the requirement of spatio-temporal continuity is also satised.
Suppose, for example, that your car were put into a powerful compactor
and reduced to a six-by-six foot solid cube of metal: that cube would still
have a shape and size and a color, and it would be spatio-temporally
continuous with what had been your car, but surely your car would no
longer exist.
Nevertheless, as we shall see more fully in chapter , there is good ev-
idence that Descartes would not only have accepted the reformulated ar-
gument that uses R*, but would also have seen it as actually establishing
his version of the substance theory. For ultimately he understands
thing in such a way that a physical thing could lose of all of its determi-
nate properties and yet remain the same thing. It could do so by virtue of
retaining a single determinable property, namely, the property of shape-
and-size, which he calls extension. According to the view that Descartes
seems ultimately committed to, an ordinary object like a piece of wax or
a rock or a car is not a thing in its own right, that is, an entity that could
conceivably exist apart from all other entities. Rather, it is only a cluster
of properties belonging to a single, all-encompassing physical thing. This
all-encompassing physical thing is nothing less than the entire physical
worldthe entire physical universe. The piece of wax discussed in
Meditation II must then be seen as only a miniature model that Descartes
uses to represent the entire physical world. And the point of the argu-
ment from change, as he there presents it, is that so long as the physical
world retains the one determinable property of extension, it continues to
exist as one and the same physical world. On this view, furthermore,
physical substance is identical with the determinable property of exten-
sion. More carefully stated, it is identical with the one instance or occur-
rence of (the determinable property of) extension that exists. Since, on
this view, the determinable property of extension is identical with (phys-
ical) substance, the reformulated argument from change still establishes
the substance theory. Furthermore, premise of the argument now
seems to be on solid ground: it is no longer even relevant to object, as we
did against the original version of the argument, that in order to persist
through change, a thing must retain at least one determinate scientic
Descartes
property such its chemical composition or its atomic structure. For all
that is needed, instead, is that it retain the determinable property of ex-
tension (three-dimensionality), which it can do even if all of its determi-
nate properties, including both its readily perceivable ones and its
scientic ones, have altered. Finally, substance must no longer be thought
of as a mysterious something underlying all of a things properties.
Instead, it is just the things most basic, dening property: extension in
the case of matter, and thought in the case of mind. So, the substance
theory is no longer vulnerable to the standard objection that substance is
in principle unperceivable.
As will be shown in chapter , this way of conceiving the matter is
arguably the most accurate interpretation of Descartess view of sub-
stance. But philosophically, there is a high price to be paid for it. This is
that one must give up the view that ordinary objects are things in their
own right, in favor of the view that they are just clusters of properties
of the one all-encompassing physical substance. Not many philosophers
would be willing to go that far away from common sense. So, unless
philosophers who reject such a one-substance view of the physical
world want to accept the theory of an unperceivable substance under-
lying a things properties (which most do not), they have to nd some
other way of dealing with the argument from change, by giving some
account of identity through change that does not rely on substance. As
we have seen, some contemporary philosophers have in eect tried to
do this, by appealing to the concept of spatio-temporal continuity
under a sortal.
6.2 The Corollary

The Corollary of the Substance theory expresses a certain view about
the status of properties: it says that a property cannot exist on its
own, but rather can exist only in a substance. Although, at rst, this
view may seem unproblematic, it is not the only possible view about
the status of properties. In fact, it is one of three competing views. To
understand these views, notice rst that what is distinctive about a
property is that it can be had by several dierent things. The property
of white or whiteness, for example, can be had by a sheet of paper, a

For a defense of this kind of view see Dicker, Humes Epistemology and Metaphysics,
pp. .
Meditation II
piece of chalk, a snowake, and indenitely many other things. By con-

trast, the paper, chalk or snowake cannot be similarly had by dif-
ferent things (though they might be used, or even legally owned, by
several dierent persons). Thus, what is special about properties is just
this: unlike particular, individual things, a property can be shared, or
had in common, by many dierent things. This raises a puzzling ques-
tion: How is it possible for many dierent things, which may be spa-
tially far removed from each other, to share one property (e.g., for a
snowake, a piece of paper, and a piece of chalk all to be white)? Each
of the three views just mentioned attempts to answer this question,
and the problem of deciding between the three views is called the
problem of universals. (In presenting this problem, the noun univer-
sal means the same as the word property.) The problem of universals
is a hard one, which we shall not try to solve. We shall only summarize
the three competing solutions, and point out some problems associ-
ated with each of them.
. According to Platonic Realism, universals (properties) exist indepen-

dently of particular things. To say that two or more things share a
common property is to say that they stand in a special relationship
usually called exemplicationto the universal. Thus the snowake,
paper, and chalk each exemplify whiteness, which exists indepen-
dently of them and of other particular white things. According to Pla-
tonic Realism, then, properties are free-oating; and the corollary of
the substance theory (as well as the theory itself) is false. The term
Platonic Realism derives from the name of this views originator and
principal exponent, the Greek philosopher Plato ( b.c.).
. According to Moderate Realism, universals exist only in particular
things. To say that two or more things have a common property is to
say that the property somehow exists in those things. Whiteness, for
example, does not exist on its own apart from white objects; it exists
in the snowake, the paper, the chalk, and other white things. Mod-
erate Realism is no doubt closer to common sense than Platonic Re-
alism. It is also the view implicit in the substance theory and reected
in its corollary, according to which properties cannot be free-oating
but must exist in a substance. Historically, Moderate Realism
derives mainly from Aristotle ( b.c.), who was Platos stu-
dent and who broke from his master precisely on the fundamental
question of the status of universals.
Descartes
. According to Nominalism, only particular things exist. Universals do

not really exist, either independently of particular things or in those
things. To say that two or more things have a common property
(e.g., are white) is only to say that the things resemble each other in
a certain way, and that we therefore apply the same name to them
(Nominalism derives from name). Thus, our ontology need not
include universals; particulars that resemble each other in various
ways are all that really exists. Some classical defenders of Nomi-
nalism are the English philosophers Hobbes (), Berkeley
(), and Hume ().
Each of the above theories faces diculties. The most obvious di-
culty for Platonic Realism is simply that independently existing univer-
sals are very mysterious entities; for, unlike particular objects, they
cannot be located in space. For example, whiteness itselfunlike the
snowake, the paper, and the chalkdoes not have a spatial location and
does not occupy a volume of space; for if it did, how would it dier from
particular white things? So with respect to independently existing uni-
versals, questions such as Where is it? or How large is it? simply lack
answers.
A less obvious diculty, which Plato himself anticipated (in the Par-
menides, one of Platos many Dialogues, which contain the bulk of his
philosophy) is known as the problem of the Third Man. It can be put
this way. Suppose we ask, Is the universal, whiteness, itself white? If we
say that whiteness is white, then in addition to particular white things
such as the snowake, the paper, and the chalk, we have another white
object, namely, whiteness. But then, the same question arises about the
particulars plus the universal as arose about the particulars alone,
namely, What makes it true that these things (including whiteness itself)
are all white? If we answer that they all exemplify yet another universal,
say whiteness, then the very same problem recurs with respect to white-
ness: Is it white? If the answer is yes, then for the same reason we are
forced to introduce yet another universal, whiteness, and so on ad inni-
tum. On the other hand, if we try to avoid this innite regress (as phi-
losophers call such a series) of whitenesses by saying that whiteness
itself is not white, then it is extremely hard to see how particular white
things can truly be said to exemplify whiteness or how the theory pro-
vides any answer to the question, What makes it true that white things
are all white?
Meditation II
Faced with such diculties, it is very natural to give up Platonic Re-

alism in favor of Moderate Realism (as did Aristotle, who was thoroughly
familiar with the obscurities and problems involved in postulating inde-
pendently existing universals). But Moderate Realism also faces di-
culties. One such diculty is: How can one thing, such as whiteness, be
in several dierent things? Here you may object that whiteness isnt a
thing but a quality, so that the diculty is not a genuine one. But
saying that whiteness is a quality, rather than a thing, does not eliminate
the basic diculty; for a quality is at least something that existsan
existentand it is puzzling how any single existent can be in several
dierent things. Although at rst it seems just common sense to say
that the color white is literally in the snowake, the paper, and the chalk,
the minute we think about the matter seriously, we are bound to nd it
perplexing.
There is another diculty for Moderate Realism. Suppose that all par-
ticular white things were destroyed. Would whiteness itself cease to be?
It seems not; rather, it seems that whiteness would continue to exist no
matter what happened to particular white things like snowakes and bits
of chalk. But if so, then we seem forced back to Platos view that white-
ness exists independently of white things.
The diculties faced by both Platonic and Moderate Realism may well
make Nominalism attractive. For according to Nominalism, universals do
not really exist at all; so we need not worry about answering the puzzling
questions just raised regarding them. Unfortunately, however, Nomi-
nalism faces its own diculties. One problem for Nominalism is that any
two or more things resemble each other in some wayif only in that they
all exist, or are in space and/or time. Therefore, in order to explain how a
piece of chalk and a snowake can both be white, it is not sucient
merely to say that the snowake and the chalk resemble each other. We
must say, instead, that the snowake and the chalk resemble each other
in respect of. But the only way to ll in this blank, it seems, is by refer-
ring to a property: in respect of being white. But then, we have reintro-
duced the universal that it was the purpose of Nominalism to avoid.
Another diculty for Nominalism, as Bertrand Russell pointed out, is
that even if we succeeded in avoiding commitment to universals in favor
of resembling particulars, it seems that we would still be committed to
the existence of at least one universal; namely, resemblance itself. But if
we allow even one universal into our ontology, then the puzzling ques-
tions about universals arise with respect to that universal (the camels
Descartes
nose is let into the tent, so to speak), and so the advantages of Nomi-
nalism are lost.
It can be seen, then, that Platonic Realism, Moderate Realism, and
Nominalism each face certain diculties. Yet, it would seem that one of
the three views must be correct; for they seem to exhaust the possibil-
ities. To solve the problem of universals, one would have to show that
one of the views can be formulated in such a way as to meet the di-
culties it faces.
6.3 The Assumption That Thoughts Are Properties

Given an ontology according to which whatever exists is either a sub-
stance or a property, it seems natural to classify thoughts as properties.
But this naturalness is somewhat deceptive. For it is not easy to say why
a thought could not be a substance rather than a property. Of course, if
one assumes Descartess dualistic view of the self, according to which
the self is a mental substance that could conceivably exist apart from
anything physical, then it becomes possible to argue that since () this
substance must have some properties (by the substance theory), ()
these properties must be mental rather than physical ones (by the
theory of dualism), and () the only mental items available to ll this
role are thoughts, it follows that thoughts are properties. But this line
of argument is not available when reconstructing the cogito; for Des-
cartes, who at that point is still trying to prove his existence, has not
yet establishedand is indeed only starting to build towardhis du-
alism. So some other argument is needed to show that thoughts are
properties.
Perhaps the following argument (which is not Descartess) will do. A
substance is something that can acquire and lose properties while remain-
ing the same substance, that is, without losing its identity. But could a
thought acquire and lose properties while remaining the same thought
without losing its identity? It seems not: if, for example, I think of a ship
with one chimney and then of an exactly similar ship except that it has
two chimneys (or if I think of $,. and then think of $,.), am
I not thinking a dierent thought? If so, then it would seem that thoughts
cannot be substances and must, therefore, be properties.

There is an excellent discussion of the issue, featuring a defense of nominalism, in
chapter of H. H. Price, Thinking and Experience.
Meditation II
Descartes, in any case, took it as obvious that thoughts are properties,

rather than substances. We shall not comment further on this assump-
tion, except only to note that it is an assumption Descartes makes. Thus,
the reconstructed cogito assumes not only the truth of the substance
theory but also the correctness of a particular application of it. It has
become apparent, then, that despite Descartess wish to doubt every-
thing that is not absolutely certain, even in expounding the cogito he
makes assumptions that might be questioned. Concerning such assump-
tions, it has been well said by Richard Schacht that:
Descartes did not question these assumptions, not because he

was deliberately trying to cheat, but rather because they were
such fundamental assumptions of medieval thought that it sim-
ply did not occur to him that they were assumptions standing in
need of justication, and perhaps untenable ones. Descartes
broke rather signicantly with medieval philosophical thought
in his program and proposed method; but he still took for
granted a number of very basic axioms of medieval reasoning.
(One encounters more of this sort of thing in his proofs of the
existence of God.) It remained for subsequent philosophers to
notice that the method he proposed, and the standards of knowl-
edge he set, would undercut not only the claim of many of our
commonsense opinions to be genuine knowledge, but also many
of the basic axioms or assumptions which he still retains. This
does not establish that they are actually wrong; it only shows
that his proposed method and standard had consequences he
did not foresee.
6.4 The Inference to I Exist

Suppose that we accept all the premises from which, according to our re-
construction of Descartess reasoning, he derives his existence. Does the
argument then succeed in proving Descartess existence? No, it does not.
For nothing in the argument entitles Descartes to call the substance
whose existence is derived in the arguments last step, I. At best the ar-
gument only proves the existence of a thinking substance; it does not

Richard Schacht, Classical Modern Philosophers, p. .
Descartes
prove that this substance is myself. The rst-person pronoun, I, so far

as the logic of the argument goes, is a completely gratuitous addition to
the conclusion drawn in step (). The criticism that we are here making
has been incisively put by Anthony Kenny. He rst notes, in the same
spirit but rather more pointedly than Schacht, that:
[T]he principle that where there are attributes there must be a

substance does not seem as unquestionable since the writings of
Berkeley and Hume as it did to Descartes. Too often, when Des-
cartes tells us that something is taught by the natural light in our
souls, he produces a doctrine taught by the Jesuits at La Flche
[La Flche was the school Descartes attended as a youth].
Kenny then unleashes the telling objection to Descartess conclusion:
But even if we accept the principle, there seems some doubt

whether the conclusion it licenses is in fact sum. [Sum is Latin
for I am.] Is not Descartes rash in christening the substance in
which the doubts of the Meditations inhere ego? To be sure, he
explains that he is not yet committing himself to any doctrine
about the nature of the ego; not until the Sixth Meditation, for
instance, will he prove that it is incorporeal. But what I refers
to must at least be distinct from what you refers to; otherwise
the argument might as well run cogitatur, ergo es [Latin for
thinking is occurring, therefore you exist] as cogito ergo sum.
Has Descartes any right to make such an assumption about the
substance in which these thoughts inhere? In Hyperaspistes
wrote: You do not know whether it is you yourself who think, or
whether the world-soul in you thinks, as the Platonists believe.
To this pertinent criticism Descartes had no real reply.
To elaborate (but also to make one small critical observation) on Kennys

criticism, the reconstruction of the cogito from the substance theory at-
tempts to derive the conclusion I exist from the premise there is a
thought, taken in conjunction with several metaphysical principles, ()
through (), all of which are stated impersonally, that is, without any ex-
plicit or implicit reference to the self or I. Kennys key point is that from

Kenny, Descartes, pp. .
Meditation II
such premises, the conclusion I exist does not logically follow; at best,
what follows is that there exists some thinking substance to which said
thought belongs. He illustrates this point by asserting that the I of the
conclusion might just as well be you, so that Descartes is rash in chris-
tening the substance in which the doubts of the Meditations inhere ego.
He also quotes Hyperaspistess remark that you do not know whether is it
you yourself who think, or whether the world-soul in you thinks. It would
seem, however, that the case of the world-soul is not a good illustration of
Kennys point: he should have stuck with his rst illustration only; for the
statement that the world-soul is thinking in me does entail that I exist
(due to the in me). Still, this does not aect the basic point that I exist
doesnt follow from the premises of the reconstruction, anymore that you
exist or he/she exists. From those premises, nothing at all follows about
the identity of the thinking substance in which the thinking occurs.
7. A Defense of the Unreconstructed Cogito

In light of the failure of the reconstructed cogito, let us reexamine the
original charge that the basic, unreconstructed argument, I am thinking,
therefore I exist, is question-begging. The reason given above for this
claim went as follows:
What entitles Descartes to use the rst-person pronoun I in the

premise of his proof? Descartess use of this pronoun presupposes
the very thing he is supposedly proving, namely, that he exists.
Thus the cogito appears to be what is called a question-begging ar-
gument, that is, an argument that takes for granted or assumes
the very conclusion that it is supposed to prove.
The apparent question-beggingness of Descartess proof stems from the

special semantics of the pronoun I. This pronoun is what we may call a
uniquely referring term. To see what is meant by this, compare I with
you, he, or she. The referents of you, he, or she may be the same
when uttered by dierent persons; for example, you and I could both refer
to Hillary Clinton by means of the pronoun she (if we were both talking
about her) or by means of the pronoun you (if we were both talking to
her). But the referent of I must be dierent when uttered by dierent
persons; you and I cannot both use I to refer to me; Hillary Clinton and
Descartes
Michelle Obama cannot both use I to refer to Hillary, and so on. Fur-
thermore, the referents of you, he, and she, when uttered by the
same person, can obviously vary, and can be determined only when some
contextual factor other than the mere utterance of those pronouns serves
to clarify who is being referred to. By contrast, no such clarication is
needed to determine who I refers to: its referent is always the very per-
son who utters it. Even if two dierent actors both utter I while playing
Hamlet, their uses of I refer to one and the same character, namely
Hamlet. Simply put, the pronoun I has only one useit refers to or calls
attention to oneself. These points, or more briey the point that I is a
uniquely referring term, seem to be all that lies behind the charge that in
saying I am thinking, Descartes is already referring to himself, and so
already assuming or taking it for granted that he exists.
It is arguable, however, that this objection is too hard on Descartes.
First, note that the denition of a question-begging argument as one
that takes for granted or assumes the very conclusion that it is supposed
to prove, although it is the standard denition, is quite loose. An argu-
ment is only a set of statements such that some of those statements are
given as reasons or grounds for another statement in the set. (The con-
text, or the presence of such words as therefore, hence, consequently,
since, because, for the reason that, and so forth, indicate that the
statements are being so treated.) As such, an argument cannot literally
take for granted or assume anything: only an arguera person who
produces an argumentcan do that. It seems, then, that a more careful
denition of a question-begging argument is needed. We need to recog-
nize that the notion of begging the question is essentially an epistemo-
logical one. We propose the following denition:
a question-begging argument = df an argument such that it is

impossible to know the premise(s) of the argument without
already knowing its conclusion.
In light of this denition, and of our points about the special semantics
of I, we can give a plausible defense of the cogito against the charge that
it begs the question, for it seems possible for a person to know I am
thinking without already knowing I exist. This could come about if the
following two conditions were both satised:

I owe this insight to Gary Iseminger.
Meditation II
() A person might know that I always refers to the self without

grasping that the self exists. For there are uniquely referring
terms, notably proper names, that refer to nonexistent things,
e.g., Superman, Batwoman, Hamlet (when used to refer to
Shakespeares ctional prince), Shangri-La, and the like, as well
as pronouns like he, she, it, and even you, that refer to
nonexistent things, as when they are used to refer to Superman,
Batwoman, Hamlet, Shangri-La, or to ones imaginary muse.
There is nothing about the pronoun I that demonstrates that, unlike
a proper name or a singular pronoun, it can only refer to something
that actually exists. The pronoun I does not magically create its
own referent. But in that case, just as Descartes insists, it is the
actual thinking in the cogito that proves ones own existence: the
thinking certainly and really existsit is manifestly occurring
now, but it is my thinking, so I must exist. If someone complains
that the thinking may not be mine, Descartes has an intuitively
compelling answer: the fact that it is I who am doubting and
understanding and willing is so evident that I see no way of
making it any clearer (CSM II , AT VII , SPW ). His point
here could be put by saying that the mineness or rst-personal
nature of my own thinkingthe fact that it is my thinking, not
anyone elses, or no onesis given, that it is a datum.
This point does not mean that I might not know that I existthat
is a separate issue, on which we need not take a position. Rather, it
only means that I can know how the pronoun I functions, and use
that pronoun, without thereby knowing that I exist. Of course, I
cannot assert that I exist unless I do in fact exist, but neither can I
assert anything else unless I exist: this is just a general point about
the necessary conditions of my making any assertion whatsoever
that has no special relevance to the cogito; indeed it is just an illus-
tration of Gassendis remark that you could have made the same
inference [that you exist] from any one of your ... actions, since it is
known by the natural light that whatever acts exists (CSM II ,
M , SPW , AT VII )a remark which, as we saw earlier,
does not count against the cogito since it ignores the special cer-
tainty of I am thinking and of other cogitatio-statements.

There is a dicult discussion that may present a challenge to this claim in Bernard
Williams, Descartes: The Project of Pure Enquiry, pp. .
Descartes
() A person might know I am thinking without ever having

had the thought that this entails I exist, or ever having had the
thought that if I am thinking, then I exist. Prior to drawing the
conclusion I exist from the premise I am thinking, it might
never have occurred to her that her thinking entailed her exis-
tence. Of course, after going through the cogito just once, a per-
son might never again have the thought I am thinking without
also realizing that having this thought necessitates that she ex-
ists. But there is no reason why that should have bothered Des-
cartes; quite the contrary, it harmonizes well with his view that
the cogito can serve as a stable foundation for other knowledge.
It might now be objected, however, that I am thinking begs the ques-

tion because the use of I covertly asserts that I exist. Now it seems that
the point already made about names and pronouns that refer to nonex-
isting things is enough to refute this objection too. But the objection also
runs into another problem, because it seems reasonable to say that in any
valid argument, the premises covertly assert the conclusion. Take for ex-
ample the classic syllogism:
All humans are mortal

Socrates is human
___________________
? Socrates is mortal.
Then if we complain that I am thinking covertly asserts I exist, should

we not also complain that all humans are mortal and Socrates is human
covertly assert Socrates is mortal? Or take the modus ponens argument:
If it rains, then the ground gets wet

It rains
_____________________________
? The ground gets wet.
Again, if we complain that I am thinking covertly asserts I exist, then

should we not also complain that the premises of this argument covertly
assert its conclusion? But if the answers to these questions are yes, then
we seem to be committed to the repugnant conclusion that all logically
valid arguments are question-begging.
Meditation II
We can say more on behalf of the cogito. Even if you are not convinced
by the above considerations that it eludes the charge of being question-
begging, you should not conclude that the entailment from I am thinking
to I exist is unimportant. On the contrary, there are at least two dif-
ferent reasons why this entailment would remain important even if the
cogito were question-begging. To appreciate the rst reason, try to put
yourself in Descartess frame of mind at the beginning of Meditation II.
There Descartes was in a state of radical, disorienting uncertainty, as was
conveyed by his image of the man caught in a deep whirlpool who can
neither touch the bottom nor swim to the surface; for he was uncertain of
the very existence of the entire physical world, including even his own
body. Now what the entailment of I exist by I am thinking shows is
that even if all my beliefs about the material world, including even those
about my own embodiment, are uncertain, my existence remains certain;
for it still remains certain that I am thinking. But just from this one very
meager certainty, it already follows that I exist. Thus, even in the midst of
the most extreme uncertainty, one can become perfectly certain of ones
own existence. So even if one thinks that the entailment of I exist by I
am thinking does not, strictly speaking, amount to a proof that one ex-
ists due to the issue of question-begging, it still shows something impor-
tant: even in the face of the extreme, disorienting doubt generated by the
arguments of Meditation I, ones own existence remains unshakably cer-
tain. Or, to put it another way, even doubting the existence of my own
body is not tantamount to doubting my existence.
The second reason why the entailment is important is this. If I am
thinking entails I exist but does not entail I have a body, then I
exist does not entail I have a body. Now this suggests (though it does
not by itself prove) Descartess view that a person could exist without a
body, as a mere thinking thing or disembodied mind. Thus the cogito, as
Descartes intimates in many places, serves as a springboard to his mind-
body dualismthe famous doctrine about the self to which we shall turn
our attention in section .
8. Does the Unreconstructed Cogito Require

an Additional Premise?
Some of Descartess contemporaries criticized the cogito from an angle
that we have not yet discussed. They claimed that the argument
Descartes
I am thinking
___________
? I exist
is simply invalid as it stands. They suggested that Descartes should have

added the premise whatever thinks exists, or some similar premise, to
make the argument valid. Their point can be put by noting that, from a
strictly logical point of view, the cogito is invalid because it has the logical
form
p
___
?q
which, of course, is invalid. Now the obvious way to x the argument

(assuming that it needs xing) would be to add the premise if I am
thinking, then I exist, thereby turning the argument into one with the
form
If p, then q
p
_________
?q
which, of course, is valid (Modus Ponens); or to add some similar premise,

such as whatever thinks exists, that would make the argument valid.

In modern quanticational logic, which did not exist in Descartess day, the cogito
could be symbolized as
Ba
________________
? (x)[Bx. (x = a)]
But, as Hintikka points out, in systems of logic in which this pattern of inference is valid
(or in which the principle {Ba (x)[Tx. (x = a)]} is provable), it is assumed that the
term which replaces a . . . must not be empty. It turns out, therefore, that we in fact
decided that the sentence I exist is true when we decided that the sentence I think is
of the form B(a). Jakko Hintikka, Cogito, Ergo Sum: Inference or Performance? p. .
Thus, if the cogito is symbolized in this way, then it is valid but question-begging.
Meditation II
So long as we were supposing that the I in I am thinking begs the

question, the issue of whether the cogito needs another premise was moot.
When an argument is question-begging, it cannot possibly require another
premise to make it valid, whether we use the standard denition of beg-
ging the question in terms of the premises taking the conclusion for
granted, or our proposed, improved denition in terms of the impossibility
of knowing the premises without already knowing the conclusion. For part
of any explanation of why an argument is question-begging is that its pre-
mises entail its conclusion: only that fact could explain why the premises
take the conclusion for granted, or why it is impossible to know the pre-
mises without knowing the conclusion. But if the premises of an argument
entail its conclusionif the argument is logically valid as it standsthen
it obviously cannot require another premise to make it valid.
However, now that we are no longer supposing that the cogito begs the
question, we can sensibly raise the question of whether it requires an
additional premise. Descartess own reply to this question was, it seems,
not wholly consistent. On the one hand, in the Principles of Philosophy,
Part I, section , he seems to concede that the cogito presupposes an
additional premise:
And when I said that the proposition I am thinking, therefore I exist

is the rst and most certain to occur to anyone who philosophizes
in an orderly way, I did not in saying that deny that one must rst
know what thought, existence, and certainty are, and that it is im-
possible that what thinks should not exist, and so forth. But
because these are very simple notions, and ones which on their
own provide us with no knowledge of anything that exists, I did not
think they needed to be listed. (CSM I , SPW . AT VIIIA )
Here Descartes seems to admit that the cogito requires, as an additional

premise, the principle that it is impossible that what thinks should not
exist. But on the other hand, in a number of places Descartes insists that
the cogito is completely compelling without the addition of any additional
premise. Thus in replying to the second set of Objections, he says:
And when we become aware that we are thinking things, this

is a primary notion that is not derived by means of any syllo-
gism. When someone says, I am thinking, therefore I am, or
exist, he does not deduce existence from thought by a syllogism,
Descartes
but recognizes it as something self-evident by a simple intuition

of the mind. This is clear from the fact that if he were deducing it
by means of a syllogism, he would have to have had previous
knowledge of the major premise Everything which thinks is, or
exists; yet in fact he learns it from experiencing in his own case
that it is impossible that he should think without existing. (CSM
II , M , SPW , AT VII )
How are we to reconcile these two passages? We may interpret Descartes as

saying, in the second one, that I exist is known by an immediate (one-
step) inference from I am thinking. He adds that one only becomes aware
of an additional premise as a result of rst grasping the necessary connec-
tion between thinking and existing in ones own case. The added premise is
only recognized in retrospect; it is not needed to grasp what Descartes sees
as the blinding force of the cogito. In other words, the additional, general
premise everything which thinks is, or exists is grasped only as a by-product
of grasping the necessity of the move from I am thinking to I exist in
ones own individual case. Descartes would presumably say the same thing
about the conditional if I am thinking, then I exist: it is not needed in the
cogito, but instead is known only as a result of grasping in ones own case
that one must exist since one is thinking. While this reading of the passage
may not render it perfectly consistent with the passage from the Principles,
it seems to harmonize best with Descartess repeatedly-stated claim that I
am thinking, therefore I exist is his very rst piece of knowledge.
Note that there is certainly a reason why it is important for Descartes
to insist that the cogito does not need an additional premise. This is that
it seems possible that an evil deceiver could fool him about any such
additional premise as whatever thinks exists or if I am thinking, then
I exist. The cogito needs to be, so to speak, deceiver-proof (using the
term proof as in waterproof), but it cannot be deceiver-proof if it
requires a premise other than I am thinking.
It remains true, of course, that the cogito is formally valid only if
the additional premise is added. But Descartess position is that the
additional premise need not be already known in order for one to
grasp the necessity of the connection between ones own thought

This claim is still not quite correct, since Descartes admits that one must rst know
what thought, existence, and certainty are. But perhaps, as the passage from the Princi-
ples suggests, his real point is that the cogito yields his rst knowledge about what exists.
Meditation II
and ones existence. That connection is grasped immediately, and the

knowledge of the additional premise depends on that grasp, rather
than the other way around. Here, so to speak, epistemology is prior to
(formal) logic.
Before drawing our discussion of the cogito to a close, let us consider a
nal question about it: Is there a way to integrate our defense of the un-
reconstructed cogito into the argument that is supposed to show that I
am a substance? Indeed there is, and it can be summarized as follows:
. I am thinking.
. If I am thinking, then I exist [needed for the inference to () to be
formally valid, but not needed to know that () on the basis of ()].
. I exist. [from () and (), though () is known only in retrospect]
. If I am thinking, then I have a thought (call it) T.
. I have a thought T. [from () and ()]
. If I have a thought T, then there is a thought T.
. There is a thought T. [from () & ()]
. A thought is a property. (Descartess assumption)
. If there is a thought T, then T is a property. [from ()]
. If x is a property, then there is a substance to which x belongs. (cor-
ollary of the substance theory)
. If there is a thought T, then there is a substance to which T belongs.
[from () & ()]
. There is a substance to which T belongs. [from () & ()]
. If I have a thought T and there is a substance to which T belongs,
then I am the substance to which T belongs.
. I am the substance to which T belongs. [from (), (), and ()]
. If I am the substance to which T belongs, then I am a thinking sub-
stance.
. I am a thinking substance. [from () & ()]
Here the rst three steps constitute the (unreconstructed) cogito, while
the remaining steps draw on ideas contained in the reconstructed one.
The only premise that goes beyond those ideas is (), and that premise
seems very plausible.

Compare with the excellent discussion of this point in Gary Hateld, Descartes
and the Meditations, pp. . See also the ne but somewhat dierent discussion in
Descartes
9. Descartess Conception of the Self

So far, we have examined only the three opening paragraphs of Medita-
tion II, in which Descartes advances his famous proof that he exists. We
now turn to the rest of Meditation II, in which Descartes introduces and
claries his conception of the self. Lets begin with a summary of the
main steps Descartes takes in the rest of the Meditation (the numbers
continue the sequence begun at the start of this chapter):
. I now know that I am, but not yet what I am.

. I shall therefore review my former beliefs about myself to see
if any of them are certain and indubitable. I believed that I was
(a) a man;
(b) a being with a face, hands, arms, etc.,i.e., with a body; and
(c) a being who was nourished, who moved about, and who engaged
in sense perception and thinkingactions that I assigned to the
soul.
. Only thinking indubitably belongs to me; for everything else
Ive just mentioned depends on my body, which may not exist
at all (since there may be an evil deceiver).
. I am only a thing that thinks (doubts, understands, arms,
denies, wills, refuses, imagines, feels).
Satised that he has proved his existence, Descartes turns in step to

the question What am I? As he puts it, But I do not yet have a sucient
understanding of what this I is, that now necessarily exists (CSM II ,
SPW , AT VII ). His treatment of this question takes up the rest of
Meditation II and a substantial portion of Meditation VI, as well.
In order to make sure that he will accept no beliefs about himself that
are not absolutely indubitable, Descartes proposes, in step , to apply the
same method to beliefs specically about himself that he applied more
generally in Meditation Ithe method of doubt. This requires him to
examine his former beliefs about himself, in order to see which, if any, of
them can withstand the test of doubt. Thus, he says:
I will therefore go back and meditate on what I originally believed

myself to be, before I embarked on this present train of thought.
I will then subtract anything capable of being weakened, even
minimally, by the arguments now introduced, so that what is left
Meditation II
at the end may be exactly and only what is certain and unshake-
able. (CSM II , SPW , AT VII )
Accordingly, Descartes now reviews his former beliefs about himself.

Some of the points he makes about those beliefs require explanation.
Regarding (a), he says:
But what is a man? Shall I say a rational animal? No; for then I
should have to inquire what an animal is, what rationality is, and
in this way one question would lead me down the slope to other
harder ones, and I do not now have the time to waste on sub-
tleties of this kind. (CSM II , SPW , AT VII )
This is an ironical remark, in which Descartes is expressing his contempt

for the Aristotelian method of denition by genus and specieswhich
was standard at Descartess timeaccording to which man is dened,
famously, as a rational [species] animal [genus]. Descartes believed that
this method of denition only leads to further questions about how to
dene the genus and the species and that there is a better method, in-
volving clear and distinct ideas, for grasping the meaning of important
notions. We shall see Descartes putting this method to work in a moment.
Having expressed his dissatisfaction with the prevailing method of def-
inition, Descartes turns his attention in (b) to what he spontaneously
believed about himself, that is, to beliefs he had acquired quite apart from
supposedly learned denitions of man, such as the belief that he had a
face, limbs, and so on. But in (c) he returns to beliefs that stem from
Aristotelian conceptions prevailing at his time. To us, it sounds very
strange to say that eating, moving around, and sense perception (here
conceived not as a type of cogitatio but as an activity requiring sense-
organs and therefore dependent on the body) are activities of the soul. But
this description would not have puzzled Descartess seventeenth-century
readers, because it is based on the Aristotelian conception of the soul
with which they were familiar.
According to Aristotle, the soul is what makes a thing alive; it is the
principle of life. So plants and nonhuman animals, as well as humans,
have souls. However, there is a dierence of degree between the souls of
plants, nonhuman animals, and humans; for a soul can have several dif-
ferent parts or faculties, and the lower-grade souls lack some of these.
Specically, the souls of plants have only the nutritive facultythe
Descartes
part of the soul that allows a thing to take in nourishment. The souls of
animals, in addition to the nutritive faculty, also possess the sensory,
appetitive, and locomotive (motion-originating) faculties, which allow
the animal to detect, desire, and move toward food or nourishment. Fi-
nally, human souls, in addition to having all of the faculties just men-
tioned, also have the faculty of rational thought (hence the denition of
a human being as a rational animal). In light of this theory of the soul,
which Aristotle expounded in a work titled De Anima (Latin for On the
Soul), we can understand Descartess talk of assigning eating, moving,
and sense perception to the soul: these are the activities of the nutritive,
locomotive, and sensory faculties of the human soul, respectively.
In step (), Descartes gives the fundamental reason why he can accept
as certain almost none of the beliefs that he has just reviewed. The hypo-
thesis that there may be an evil deceiver, who fools him about the exis-
tence of the entire physical world, is still in force. As he puts it,
But what shall I now say that I am, when I am supposing that
there is some supremely powerful and . . . malicious deceiver,
who is deliberately trying to trick me in every way he can? (CSM
II , SPW , AT VII )
The deceiver hypothesis shows that no belief that implies that Des-
cartes has a body can be certain; thus, it shows at one stroke that none
of the beliefs listed in (a)-(c) can be certainwith one exception. The
exception is Descartess belief that he is thinking. On the ground that
there may be a deceiver who fools him about the existence of the whole
physical universe, Descartes can doubt that he possesses a body and so
that he takes in food or moves around or has sense-perceptions (in the
sense involving physical stimuli and sense organs). But the deceiver hy-
pothesis cannot shake Descartess belief that he is thinking. For, as we
saw in section above, while it would be possible for a very powerful
deceiver to fool Descartes into thinking that he had a body even if he
didnt have one, it would be impossible for any deceiver, no matter how
powerful, to fool Descartes into thinking that he was thinking when he
wasnt thinkingfor thinking that one is thinking is thinking. Thus the
deceiver hypothesis here works, so to speak, to set a boundary or limit
to what can be doubted. It is as if Descartes were saying, I can be

I owe the idea of boundary-setting to Jonathan Bennett.
Meditation II
fooled up to this point but not beyond it. To be sure, this boundary-
setting aspect of the deceiver hypothesis is not a proof that I am
thinking is certain, because one could not sensibly argue, it is certain
that if Im deceived into thinking that Im thinking, then I am thinking;
it is certain that I am deceived into thinking that Im thinking; there-
fore it is certain that I am thinking. Rather, the boundary-setting
aspect of the deceiver hypothesis works negatively, by showing that
even this most radical skeptical hypothesis cannot disprove the cer-
tainty of I am thinking.
Having applied the method of doubt to his former beliefs about him-
self, Descartes introduces, in step , his revolutionary conception of the
self as essentially a spiritual (mental) substance, mind, or soul. It is easy
to miss the revolutionary nature of Descartess conception; for, whether
or not you personally believe in the existence of a soul that could con-
tinue to exist without the body, you can surely recognize this notion of
the soul as central to the Judeo-Christian view of human beings. The
conception of a human being as one composed not only of a physical
part (the body) subject to all the laws of biology and physics but also a
nonphysical part (the soul) not limited by these laws, whether or not
you personally accept it, is at least very familiar to you; for it is the mod-
ern version of the Judeo-Christian view of humans. But when Descartes
wrote his Meditations, the prevailing conception of the soul was the
Aristotelian one just sketched, according to which the soul and the body
form a single substance. This is not to say that Descartess conception of
the self was totally unprecedented in the history of philosophy. On the
contrary, Plato had conceived the soul as an immaterial entity that out-
lives the body, as is clear, for instance, from his Dialogue Phaedo. And
there are important vestiges of Platos view in Aristotleand even
more so in Aquinas and other medieval Christian thinkers. But what
Descartes did was renovate Platos dualistic conception of the self by
giving an extraordinarily sharp, clear account of itone that he
intended to be satisfactory even in light of the scientic revolution of
the seventeenth centuryas well as forceful new arguments for it. His
account, then, provides the modern philosophical underpinnings for
the Judeo-Christian view of human beings.
Descartes was aware that his conception of the self was bound to seem
barren, uninformative, and overly abstract to his readers. So in Medita-
tion II, he does two dierent things to combat this impression. Let us
look at them in turn.
Descartes
First, Descartes tries to give content to his conception of a purely

thinking substance by reminding us of what thinking coversnamely,
all conscious states. These include doubting, understanding, asserting,
denying, willing, refusing, imagining, and seeming to perceive. As Des-
cartes puts it, ingeniously listing the very thoughts he has been report-
ing in the Meditations:
This is a considerable list, if everything on it belongs to me. But

does it? Is it not one and the same I who is now doubting almost
everything, who nonetheless understands some things, who af-
rms that this one thing is true, denies everything else, desires
to know more, is unwilling to be deceived, imagines many things
even involuntarily, and is aware of many things which apparently
come from the senses? (CSM II , SPW , AT VII )
Second, Descartes tries to show that the conception of a physical

object, though it initially seems easier to grasp than the conception of a
pure mind, is in fact just as abstract. As one writer insightful notes, this
is the purpose (or, rather, one of the purposes) of the passage about the
wax. This can be seen by attending to the way Descartes introduces that
passage:
From all this I am beginning to have a rather better under-

standing of what I am. But it still appearsand I cannot stop
thinking thisthat the corporeal things of which images are
formed in my thought [by imagination], and which the senses
investigate, are known with much more distinctness than this
puzzling I which cannot be pictured in the imagination. And yet
it is surely surprising that I should have a more distinct grasp of
things which I realize are doubtful, unknown and foreign to me,
than I have of that which is true and knownmy own self. But I
see what it is: my mind enjoys wandering o and will not yet
submit to being restrained within the bounds of truth. Very well
then; just this once let us give it a completely free rein, so that
after a while, when it is time to tighten the reins, it may more
easily submit to being curbed.

Arthur Danto, What Philosophy Is, p. .
Meditation II
Let us consider the things which people commonly think they

understand most distinctly of all; that is, the bodies which we
touch and see. I do not mean bodies in generalfor general per-
ceptions are apt to be somewhat more confusedbut one partic-
ular body. Let us take, for example, this piece of wax. (CSM II ,
SPW , AT VII )
Having carefully prepared us for a comparison of his novel conception

of the self (which can be neither perceived by the senses nor pictured by
imagination) with the seemingly easier conception of an ordinary physical
object (which can be examined by the senses and pictured by imagination),
Descartes asks, So what was it in the wax that I understood with such
distinctness? (CSM II , SPW , AT VII ) This question concerns our
conception of the wax; it can be paraphrased as, What constitutes our con-
ception of a body, for example, this piece of wax? Descartes arrives at his
answer by a process of elimination. First eliminated are the waxs observ-
able propertiesits shape, size, texture, color, smell, and so on. These do
not constitute our conception of the wax, because even if they all change,
the wax remains. So our conception must be an abstract one: Let us con-
centrate, take away everything which does not belong to the wax, and see
what is left: merely something extended, exible, and changeable (CSM II
, SPW , AT VII ). What Descartes says next makes the conception
even more abstract. He asks, Since we conceive the wax as something ex-
ible and changeable, is not our conception of the wax composed of the
specic sizes and shapes that we can imagine the wax taking on? No, he
answers; for we conceive that the wax can take on innitely many dierent
shapes and sizes, but we can only imagine (i.e., picture or visualize) a nite
number of these. At last, Descartes turns his attention to extended:
And what is meant by extended? Is the extension of the wax

also unknown? For it increases if the wax melts, increases again
it boils, and is greater still if the heat is increased. I would not be
making a correct judgement about the nature of wax unless I
believed it capable of being extended in many more dierent
ways that I will ever encompass in my imagination. (CSM II ,
SPW , AT VII )
In this important sentence, we may interpret Descartes as saying that to

conceive the wax as something extended is to conceive it as something
Descartes
that can take on a great many dierent shapes and sizes (something three-
dimensional). Our conception of the wax, then, is merely the conception
of something that can take on various (three-dimensional) shapes and
sizes. Descartes adds that we have this conception neither by the senses
nor by imagination but only by reason. Moreover, it is now a clear and
distinct conception, unlike the imperfect and confused one he began
with; for, unlike a conception involving the properties perceived by the
senses or involving the specic shapes and sizes pictured by imagination,
his conception now contains all thatand only whatthe wax must
have to remain the same wax. It is a puried conception, arrived at by a
careful process of reasoning.
Finally, notice how Descartes has fullled his purpose of showing that
our conception of a body is as abstract as his conception of a mind. The pu-
ried, clear and distinct conception of a body, as weve just seen, is the
conception of something that can take on various shapes and sizes. But what is
the conception of a purely thinking substance but the conception of some-
thing that can take on various thoughts (cogitationes)doubts, desires, beliefs,
sensations, and so on? Moreover, neither conception comes from the senses
or imagination; both are purely intellectual. The upshot is that Descartess
radically novel conception of the self as a purely thinking substance is no
more abstract or dicult than the true conception of a material body.
10. Cartesian Dualism

By the time Descartes has nished discussing the piece of wax, he has, in
eect, expounded his single most characteristic doctrine, namely, Carte-
sian Dualism. Cartesian Dualism is the view that the universe contains
two radically dierent kinds of substance: () mind, dened as a thinking,
unextended substance, () body (i.e., matter), dened as an extended, un-
thinking substance. Notice, then, that mind and matter are dened as
opposites: mind is thinking, while matter is unthinking; matter is ex-
tended (three-dimensional) while mind is unextended. Matter occupies
space but doesnt think, and a mind thinks but doesnt occupy space.
When trying to focus on Descartess distinctive concepts of mind and
matter, it is helpful to bear in mind his Latin names for them: res cogitans
(thinking thing) and res extensa (extended thing).
Descartess dualism is perhaps his most important doctrine, for it had
sweeping and far-ranging implications. On the one hand, it helped to
Meditation II
clear the way for modern physical science. As we saw in the last chapter,
the prevailing, Aristotelian physics of Descartess day held that the uni-
verse is inherently purposeful or teleological. In other words, everything
that happened, whether it was the motion of the stars in the supralunar
region or the growth of a tree in the sublunar region, was supposed to be
explained by certain purposes, goals or ends working themselves out
within nature. Aristotle called such purposes nal causes, and nal
causes were considered to be indispensable for explaining the ways nature
operated. A battle was taking shape between the scholastic defenders of
this traditional science and the proponents of the new science of Kepler
and Galileo, which denied the relevance of nal causes for explaining
nature. Descartess dualism provided a powerful philosophical rationale
for the newer conception, for one implication of Descartess dualism is
that nal causes are expelled from the physical universe, or res extensa.
The only place left for nal causes is the mind, or res cogitans. Thus, Des-
cartess Dualism helped prepare the way for modern physics, which does
not explain nature by reference to purposes. On the other hand, Carte-
sian dualism also ensured the possibility of immortality; for if the mind
or soul is really a dierent substance from the body, then the destruction
of the body does not entail the extinction of the mind. Thus, Cartesian
Dualism simultaneously helped to clear the way for modern physics and
held the door open for religious beliefs about the immortality of the soul.
Before leaving Meditation II, we need to raise a question: How far has
Descartes really come toward establishing his dualism at this point in his
Meditations?for although Descartes, by the end of Meditation II, has
explained his dualism by expounding his conceptions of both res cogitans
and res extensa, it doesnt follow that he has shown this dualism to be
true, that is, shown that mind and matter really are two dierent sub-
stances, one thinking and unextended, the other extended and un-
thinking. How close has Descartes really come to proving this at the end
of Meditation II?
To begin with an obvious point, he certainly has not fully established
his dualism, because he has not shown that there is such a thing as matter.
Remember that Descartes still doubts the existence of the material world.
So he certainly hasnt shown (and isnt claiming to have shown) that the
universe actually contains any res extensa. Not until the sixth (and last)
Meditation does he try to show this.
The signicant question, rather, is whether Descartes has already
shown that mind is a dierent substance from any matter that may exist,
Descartes
if any exists. In other words, has he shown that there is such a thing as a
purely thinking substance? This is a more delicate issue, which Des-
cartess contemporaries repeatedly asked him to clarify.
To focus the question better, lets leave the Meditations for a moment
and consider a passage from Descartess Discourse on the Method, Part
Four, paragraph :
Next I examined attentively what I was. I saw that while I could

pretend that I had no body and that there was no world and no
place for me to be in, I could not for all that pretend that I did not
exist. I saw on the contrary that from the mere fact that I thought
of doubting the truth of other things, it followed quite evidently
and certainly that I existed; whereas if I had merely ceased
thinking, even if everything else I had ever imagined had been
true, I should have had no reason to believe that I existed. From
this I knew I was a substance whose whole essence or nature is
simply to think, and which does not require any place, or depend
on any material thing, in order to exist. Accordingly this I
that is, the soul by which I am what I amis entirely distinct
from the body ... and would not fail to be whatever it is, even if
the body did not exist. (CSM I , SPW , AT VI )
Here, Descartes appears to be saying that, merely from the fact that he can
doubt the existence of matter but cannot doubt his own existence, it follows
that he is a purely thinking substance. There is a similar passage in his Search
After Truth: I ... am not a body. Otherwise, if I had doubts about my body,
I would also have doubts about myself, and I cannot have doubts about that
(CSM II , AT X ). Descartess line of reasoning in such passages has
been called the Argument from Doubt. It may be summarized this way:
() I can doubt that (my) body exists.

() I cannot doubt that I exist.
_____________________________
? I am not a body.
Is this argument valid? No, it isnt. Just because I can doubt that my
body exists but not that I exist, it doesnt follow that I am not a body; for
I might very well be a body but not know it. The Argument from Doubt is
no better than this argument (which we can imagine Louis XVI giving
Meditation II
before the French revolution that ended his reign): I can doubt that the
last King of France exists; I cannot doubt that I exist; therefore I am not
the last King of France. Sometimes the fallacy in such arguments is
called the masked man fallacy: I know who my father is; I do not know
who the masked man before me is; therefore this masked man is not my
father.
Does Descartes rely on the Argument from Doubt in his Meditations,
as he appears to have done in his Discourse on the Method and Search After
Truth? It would be unfortunate if he did, for the Meditations are Des-
cartess most careful and rigorous presentation of his philosophy.
Initially, it may look as if Descartes uses the Argument from Doubt
even in his Meditations. For in Meditation II he does say:
I am not that structure of limbs which is called the human body.

I am not even some thin vapor which permeates the limbsa
wind, re, air, breath, or whatever I depict in my imagination; for
these are the things which I have supposed to be nothing. With-
out changing this supposition, I nd that I am still certain that I
am something. (CSM II with n. , SPW with n. , AT VII )
Matters of detail apart, this is again the Argument from Doubt. The tran-
sition from step to step in our summary of the Meditation also looks
much like the Argument from Doubt.
The very next sentence after the passage just quoted, however, shows
that Descartes does not wish to rely on that argument in his Meditations,
for he says:
And yet may it not perhaps be the case that these very things
which I am supposing to be nothing, because they are unknown
to me, are in reality identical with the I of which I am aware? I
do not know, and for the moment I shall not argue the point,
since I can make judgements only about things which are known
to me.
Here Descartes pulls back from asserting that he is not a body; he admits
that for all he knows at this point, he may be one. All he knows is that he
is at least a thinking being.
Accordingly, step in our summary of Meditation IIthe claim that I
am only a thing that thinksshould not be understood to mean I know
Descartes
that I am only a thing that thinks. Rather, step means I know only
that I am a thing that thinks. The position of the word only is crucial.
This word serves to limit what Descartes is claiming to know about what
he is, not to exclude body from what he may, in reality, be. The upshot is
that the weakness of the Argument from Doubt does not vitiate Des-
cartess reasoning in Meditation II, simply because Descartes does not
here rely on that argument. Not until Meditation VI does Descartes give
an argument intended to prove denitively that the mind is a dierent
substance from the body. We must wait until we reach that argument to
decide whether it succeeds in establishing Cartesian Dualism.
| 3 |
Meditation III
The Criterion of Truth and the Existence of God
1. Descartess Criterion of Truth

At the outset of Meditation III, Descartes summarizes the knowledge that
he has attained in the previous two Meditations.
I am a thing that thinks: that is, a thing that doubts, arms, denies,
understands a few things, is ignorant of many things, is willing, is
unwilling, and also which imagines and has sensory perceptions;
for, as I have noted before, even though the objects of my sensory
experience and imagination may have no existence outside me,
nonetheless the modes of thinking which I refer to as cases of
sensory perception and imagination, insofar as they are merely
modes of thinking, do exist within meof that I am certain.
In this brief list I have gone through everything that I truly know,
or at least everything that I have so far discovered that I know.
So far, this is all that Descartes claims to know; everything else is still
subject to doubt. Only what Descartes takes himself to have established
by the cogito, that he thinks (i.e., doubts, arms, denies, etc.) and there-
fore exists, is secure. So we may summarize his rst step in Meditation III
simply like this:
. So far I know only that I am a thing that thinks.
In his remaining Meditations, Descartes will build upon this one unshake-
able piece of knowledge, which is itself based solely on the cogito.

Descartes
We have seen, of course, that the cogito is by no means as simple or

unproblematic as it seems. But from this point on, let us put aside any
worries about the cogito. For even if, despite the defense of the unre-
constructed cogito that was oered in . and ., you reject the cogito
as a proof of ones own existence (on the ground that it is either
question-begging or invalid), the certainty of I am thinking and of
other cogitationes, the certainty of the inference from I am thinking
to I exist, and the certainty of I exist, all remain secure. And these
three points are sucient for the further uses to which Descartes will
put the cogito.
Having armed his existence as a thinking thing, Descartes now asks
himself, Can I learn anything more from this? As he puts it:
Now I will cast around more carefully to see whether there may
be other things within me which I have not yet noticed. I am cer-
tain that I am a thinking thing. Do I not therefore also know
what is required for my being certain about anything? (CSM II
, SPW , AT VII )
Descartess idea here is this. He has one instance of absolutely certain,

unshakeable knowledge. So, by examining this single, shining example of
genuine knowledge, he should be able to discover the feature of it that
makes it so unshakeable. He can then consider whether any further prop-
ositions also have this feature: if they do, he will be able to extend his
knowledge to them, as well. We may summarize his idea this way:
. Can I learn anything more from this? I am certain that I am a

thinking thing. Dont I therefore also know what is required
for my being certain about anything?
Accordingly, Descartes asks himself: what assures me that I am a

thinking thing? What is the characteristic of this piece of knowledge that
makes it so certain? He answers,
In this rst item of knowledge there is simply a clear and distinct

perception of what I am asserting. (CSM II , SPW , AT VII )
The feature of his rst item of knowledge that renders it so certain,

Descartes declares, is simply that it is a clear and distinct perception.
Meditation III
So, he reasons, perhaps he can now safely generalize that whatever he

perceives clearly and distinctly is true; for, as he continues:
[T]his would not be enough to make me certain of the truth of

the matter if it could ever turn out that something which I per-
ceived with such clarity and distinctness was false. So I now seem
to be able to lay it down as a general rule that whatever I perceive
very clearly and distinctly is true. (CSM II , SPW , AT VII )
Here, Descartes has extracted from his rst item of knowledge his
famous criterion of truththat whatever he perceives clearly and dis-
tinctly is true. Let us more closely examine both (i) what he extracts this
criterion from and (ii) the content or meaning of the criterion itself.
Descartes speaks as if the rst item of knowledge from which he
extracts his criterion is just the one clearly and distinctly perceived prop-
osition, I am a thinking thing. Descartess knowledge that he is a
thinking thing, however, is really a complex piece of information. Its ele-
ments include at least the various components of the basic, unrecon-
structed cogito, namely, the knowledge (a) that he is thinking, (b) that his
thinking necessitates his existence, and (c) that he exists. We may inter-
pret him, therefore, as saying that what assures him that he is a thinking
thing is that he very clearly and distinctly perceives the cogitohere seen
as a tight package of certainties composed of (a), (b), and (c). In other
words, we can understand him as deriving or extracting his criterion of
truth from the cogito, by means of the following argument: If something
could be clearly and distinctly perceived yet false, then this would cast a
shadow of doubt on the cogito itself. But the cogito is absolutely indubi-
table. Therefore, what is clearly and distinctly perceived cannot be false;
so it must be true. Accordingly, we may summarize Descartess third
stephis extraction of the clarity and distinctness criterion of truth
from the cogitoas follows:
. What assures me that I am a thinking thing? Only that I per-

ceive the cogito very clearly and distinctly. So, it seems that I
can already infer that whatever I perceive clearly and dis-
tinctly is true; for
() If my clear and distinct perceptions could be false, then

the cogito would not be certain.
Descartes
() The cogito is certain.

____________________________________________________
?() My clear and distinct perceptions cannot be false;
that is, whatever I perceive clearly and distinctly is true.
Adopting a recent commentators usage, we can call this argument from

() and () to (), the extraction argument. That Descartes advances
the extraction argument is conrmed by the following passage from the
Discourse on the Method, Part Four, paragraph , where he explicitly
derives his criterion of truth from the cogito:
After this I considered in general what is required of a proposi-

tion in order for it to be true and certain; for since I had just
found one that I knew to be such, I thought that I ought also to
know what this certainty consists in. I observed that there was
nothing at all in the proposition I am thinking, therefore I exist
to assure me that I am speaking the truth, except that I see very
clearly that in order to think it is necessary to exist. So I decided
that I could take it as a general rule that the things we conceive
very clearly and distinctly are all true. (CSM I , SPW , AT
VI )
But what exactly does Descartes mean by a clear and distinct percep-
tion? In his Principles of Philosophy, under the caption What is meant by
a clear perception, and by a distinct perception, he oers the following
denition:
I call a perception clear when it is present and accessible to the

attentive mindjust as we say that we see something clearly
when it is present to the eyes gaze and stimulates it with a su-
cient degree of strength and accessibility. I call a perception dis-
tinct if, as well as being clear, it is so sharply separated from all
other perceptions that it contains within itself only what is clear.
(CSM I , SPW , AT VIIIA )
This denition is not as helpful as one might like, for it is basically just an
analogy: a clear and distinct perception is an intellectual perception like

Gary Hateld, Descartes and the Meditations, p. .
Meditation III
the visual perception of an object in good conditions, when we can dis-

criminate the object from its environment and make out each of its var-
ious parts. This does not give us a denite way of telling when a perception
is clear and distinct; and, indeed, Descartes admits in a number of places
that this is not always easy to do.
One standard interpretation of Descartess notion of clarity and dis-
tinctness sees it as being inspired by mathematics. The idea is that a
simple mathematical proposition, like + = , is so clear and obvious
that it cannot be doubted and that it is propositions of this kind that
Descartes calls clear and distinct. As Frederick Copleston puts it in his
well-known, multi-volume History of Philosophy:
This criterion of truth was doubtless suggested to Descartes by

mathematics. A true mathematical proposition imposes itself, as
it were, on the mind: when it is seen clearly and distinctly, the
mind cannot help assenting to it. Similarly, I arm the proposi-
tion, I think, therefore I am, not because I apply some extrinsic
criterion of truth, but simply because I see clearly and distinctly
that so it is.
Although this interpretation of Descartess criterion is no doubt partly

right, it is not completely satisfactory; for Descartes, as we have seen,
derives the criterion from the cogito. But even if we regard the cogito in its
simplest, classic, unreconstructed form (as we are now doing), it has a
certain complexitya complexity that gets masked when the cogito is
referred to as one proposition (as it is by Descartes himself in the pas-
sage we quoted above from his Discourse on the Method) or as the propo-
sition, I think, therefore I am (as it is by Copleston in the passage just
cited); for the unreconstructed cogito has three components: I am thin-
king, If I am thinking, then I exist (or it is impossible for what thinks
not to exist), and I exist. (The generalization it is impossible for what
thinks not to exist, as we suggested in ., is grasped only by rst
grasping the necessary connection between thinking and existing in
ones own case. This premise, though it is needed for the arguments for-
mal validity, is recognized in retrospect; it is not needed to grasp the
blinding force of the cogito. Descartes would presumably say the same

Frederick Copleston, A History of Philosophy, p. .
Descartes
thing about the conditional if I am thinking, then I exist: it is known

only as a result of grasping in ones own case that one must exist since
one is thinking.) Furthermore, the certainty of each of these components
stems from a dierent source. I am thinking is certain because it reports
only my own present conscious state. If I am thinking, then I exist is
certain because it is obviously impossible for I am thinking to be true
and I exist to be false. I exist is certain because it is the conclusion of
an argument which has the features that (a) one of its premises reports
only my own present conscious state and self-evidently necessitates the
arguments conclusion, and (b) it becomes formally valid with the addi-
tion of a premise that expresses the necessity referred to in (a).
Now, in order for clear and distinct perceptions all to be akin to math-
ematical ones, Descartess criterion would have to be derived exclusively
from the second componentfrom the proposition if I am thinking,
then I exist (or it is impossible for what thinks not to exist); for only
this proposition shares with mathematical propositions the feature
that primarily explains their certainty, namely, that a true mathemat-
ical proposition is necessarily true. The proposition that + = ,
for example, doesnt just happen to be true. Rather, it must be true; it
could not possibly be false. As philosophers usually put it today, there
are no possible worlds in which one plus one does not equal two. The
same goes for the proposition if I am thinking, then I exist: this prop-
osition couldnt possibly be false. There are no possible worlds in which
I am thinking but I do not exist. By contrast, the propositions I am
thinking and I exist could have been false, for I might not have
existed. There are possible worlds in which I do not exist; and if any of
those worlds had been actual, then I exist, as well as I am thinking,
would have been false. Of course, if I am thinking or I exist are
uttered or entertained in a given possible world, call it W, then those
propositions must be true in W. But this does not mean that I am
thinking or I exist are true in all possible worlds, for they need not be
uttered or entertained in all possible worlds. So these propositions are
not necessary; instead, they are contingent. (The term contingent, as it
is used in philosophy, applies to all propositions that are neither ne-
cessary nor impossible.) Yet, Descartes evidently regards I am thinking
and I exist as clear and distinct, too. So his criterion can be satised
by propositions of fundamentally dierent typescontingent ones as
well as necessary ones. Furthermore, Descartes surely regards the
validity of the argument
Meditation III
If I am thinking, then I exist

I am thinking
_______________________
? I exist
as clear and distinct. So his criterion can also be satised by the logical
step from the premise(s) of a valid argument to its conclusion. Finally,
Descartes regards I exist as clear and distinct; so his criterion can be
satised by the conclusion of a valid argument whose premises describe
only ones own present thoughts or are obvious necessary truths.
The upshot is that, in addition to being satised in () the special case
of a proposition (I exist) that is self-evidently necessitated by a propo-
sition that describes only ones own present thoughts (I am thinking,
or any other cogitatio-statement), Descartess criterion of clarity and dis-
tinctness can be satised by items of four dierent kinds: () contingent
propositions describing only ones own present thoughts, () obvious
necessary propositions, () the step, transition, or inference from the
premise(s) to the conclusion of a valid argument, and () the logical con-
sequences of ()s and/or ()s. No doubt, this is what Descartes intended;
for, as we shall see, he goes on to use his criterion throughout his attempt
to rebuild his knowledge: the criterion has, so to speak, a lot of work to
do. Perhaps the best way to interpret Descartess criterion, then, is to see
it as a kind of pass, ticket, or license, saying that in rebuilding ones
knowledge, it is permissible to build on the ve types of items just listed.
Although Descartess criterion of truth may now seem quite complex
and perhaps even a bit slippery, this need not worry us further; for
nothing will stop us from asking whether the specic propositions that
Descartes will put forward as clear and distinct are really as unques-
tionable as he takes them to bewhether they are as obviously certain as
the cogito and its component parts.
Let us conclude this section, then, by summarizing how Descartes will
use his criterion of truth in his subsequent Meditations. Basically, he will
use it to overcome his doubt concerning all matters beyond his own exis-
tence as a thinking thing. Specically, he will use it to show:
. that a perfect God exists (Meditation V),

. that mind is really a dierent substance from any matter that may
exist (Meditation VI), and
. that the material world exists (Meditation VI).
Descartes
Descartess basic strategy, then, is to derive his criterion of truth from

the cogito and then to apply it in successively proving Gods existence,
mind-body dualism, and the existence of the physical world.
2. The Project of Meditation III

The three-point strategy just outlined highlights the three main the-
ses that Descartes will seek to establish in his subsequent Meditations:
the existence of God, mind-body dualism, and the existence of the
physical world. However, as you may surmise from the fact that it
skips over the rest of Meditation III and Meditation IV, it is an oversim-
plied account of Descartess strategy; for even after extracting his
criterion of truth from the cogito, Descartes is not yet willing to use it
in arguing for his remaining three main theses. Instead, he embarks
on a fascinating, famous, and (as we shall see) problematic attempt to
justify or vindicate that very criterion. The key element in this at-
tempt, as we shall see, is a complex argument for the existence of a
perfect God who guarantees the truth of clear and distinct perceptions
(thus the argument for Gods existence in Meditation V is not the rst
such argument that Descartes gives). It is to Descartess quest for a
vindication of his criterion of truth, then, that we must next turn our
attention. We shall initially proceed, as before, by summarizing his
reasoning step-by-step.
In the paragraph immediately following the extraction argument, Des-
cartes reminds himself that much remains doubtful. We may paraphrase
him this way:
. I must remember my doubt concerning many things whose

existence once seemed obvious, that is, physical objects that I
perceived by my senses. Only the fact that I have ideas of these
objects is clearly and distinctly perceived. The origin of these
ideas and whether they resemble anything existing outside me
are still unknown.
The next paragraph is an extremely crucial one. In its rst segment

(But ... minds eye), Descartes turns his attention to the most obvious
propositions he can identify, such as simple mathematical ones. We can
paraphrase the segment this way:
Meditation III
. But cant I now at least say that I perceive simple propositions of

mathematics, such as + = , clearly enough to arm their
truth? The only reason Ive found for doubting such things is that
perhaps God gave me a nature such that I am deceived even about
what seems most evident. Whenever I think of an all-powerful
God, I must admit that if he wishes, he can easily make me go
wrong even about things that I think I perceive utterly clearly.
Here, then, Descartes is expressing a doubt even about propositions that

he most clearly and distinctly perceivesa doubt based on the possi-
bility of a deceiving God that he raised in Meditation I. But now, consider
what he says in the next segment (Yet ... contradiction). We may para-
phrase it this way:
. But every time I actually attend to these things, I am so con-

vinced by them that Im impelled to say: No one can cause me
not to exist so long as I am thinking that I exist, or make it
true that Ive never existed since I now exist, or even that two
plus three are not equal to ve; or that any other proposition
in which I see an obvious contradiction is true.
Here, Descartes asserts that at the actual time that he is clearly and dis-
tinctly perceiving the cogito, or a simple instance of the law of noncontra-
diction (i.e., I cannot both never exist and exist now), or even a simple
mathematical proposition, he cannot doubt it. Such propositions as I am
thinking, therefore I exist, or not both p and not-p, or even + = are, to
borrow a term coined by E. M. Curley, assent-compelling. They cannot be
doubted during the time that one is clearly and distinctly perceiving them.
But doesnt this contradict what Descartes said in the preceding seg-
ment, where he admitted that an omnipotent God could deceive him even
about things he most clearly and distinctly perceives? No, for there Des-
cartes did not admit that he could doubt a proposition while clearly and
distinctly perceiving it. Rather, he admitted that while thinking about an

The need for the word even, which does not appear in John Cottinghams transla-
tion of the Latin text of the segment that is here being closely paraphrased, was pointed
out to me in correspondence by Jean-Marie Beyssade, as the correct translation of Des-
cartess vel forte etiam, which Cottingham translates, like vel, simply as or.

Edwin M. Curley, Descartes Against the Skeptics, p. .
Descartes
omnipotent God, he had to concede that such a God would be able to deceive
him even about the most obvious things. From this concession, it does not
follow that Descartes can doubt any one of those things while actually paying
attention to it. It only follows that even though he cannot doubt a proposi-
tion so long as he clearly and distinctly perceives it, he can doubt whether
the fact that he clearly and distinctly perceives a proposition guarantees
that it is true. As James Van Cleve succinctly puts it: [Descartes] might be
uncertain of the general connection between clear and distinct perception
and truth, yet certain of every proposition [he] ... clearly and distinctly
perceive[s]. Thus, in the rst two segments of the paragraph (points and
of our paraphrases), Descartes is weighing his certainty about particular,
occurrent clear and distinct perceptions, against a general doubt concern-
ing the reliability of his cognitive faculties (notably his faculty for clear and
distinct perception), based on the possibility of a deceiving God.
In the third and nal segment of the paragraph (And since ... else), Des-
cartes declares that there is only one way he can emerge from this oscillation
between doubt and certainty. We can paraphrase the segment this way:
. Although my reason for doubting clear and distinct perception

is very slight and metaphysical, to remove it I must determine
whether (a) God exists, and (b) God can be a deceiver; for with-
out knowing this, I cannot be perfectly certain of anything else.
Thus, in order to resolve the tension between his certainty about partic-
ular, occurrent clear and distinct perceptions and his doubt about the
reliability of clear and distinct perception in general, Descartes believes
he must eliminate the possibility of a deceiving God, by establishing the
existence of a nondeceiving God. He must show that his clear and dis-
tinct perceptions are guaranteed to be true by God himself.
3. From the Idea of God to God

Accordingly, Descartes now turns to the question of Gods existence,
which becomes the main topic of Meditation III. In that Meditation, Des-
cartes advances two related proofs of Gods existence. (He gives a third,

James Van Cleve, Foundationalism, Epistemic Principles, and the Cartesian Circle,
p. .
Meditation III
quite dierent proofthe famous Ontological Argumentin Meditation V).

Both Meditation III proofs are Descartess own special versions of what is
called the Cosmological Argument for the Existence of God. The Cosmo-
logical Argument, which comes in several dierent versions, attempts to
prove that God exists by showing that the existence of the world, or of
things in the world, requires an original cause or an ultimate explanation.
(By contrast, the Ontological Argument given in Meditation V attempts
to prove that God exists because of the very concept or denition of God
as an absolutely unsurpassable being.) For example, Saint Thomas Aqui-
nass seminal versions of the Cosmological Argument in the thirteenth
century reason from the existence of certain eects in the world to God
as the rst cause of those eects. Like Aquinass arguments, Descartess
Meditation III proofs reason from certain eects to God as the cause of
those eects. However, unlike Aquinass arguments, Descartess proofs
cannot appeal to any of Gods eects in the physical world; for remem-
ber that at this point in the Meditations, the existence of the entire phys-
ical world is still in doubt. Accordingly, Descartess strategy is to argue
from the idea of God that he nds in his mind to God as the cause of that
idea. His rst proof starts just from the idea of God and attempts to
show that God himself is the only possible cause of that idea. His second
proof builds on the rst by trying to show that only God could cause the
existence of a thinking thing that has the idea of God. The second proof
depends upon the rst, so we shall concentrate our attention on the
rst.
3.1 The Nature of Ideas

Since Descartess argument starts from the idea of God, he prepares the
way for the argument by discussing the nature of ideas. We shall simply
list the points he makes, and comment briey on each of them.
. An idea represents something; it is like a picture.
This is Descartess most fundamental and inuential point about ideas.

An idea, according to him, is essentially a mental representation of its
objectof the thing of which it is an idea. This is true whether or not that
object really exists, since an idea must have a content, must be an idea of
something. The same view of ideas can be found in many major philoso-
phers that Descartes inuenced, including Locke, Hume, and Kant.
Descartes
. An idea itself cannot be false.
Descartess point can be put this way: even if a picture doesnt depict ac-
curately or if what it depicts doesnt exist, the picture itselfconsidered
merely as an imagecannot be false (or true). Rather, the picture is just
something that exists in its own right, whether or not what it depicts
also exists. Likewise, an idea, considered purely in terms of its content,
that is, without regard to whether that content corresponds to anything
else, cannot be either false or true.
. Falsity (and truth) becomes possible only when I make a judg-

ment, especially when I judge that an idea in my mind corre-
sponds to or resembles something outside my mind.
Given that an idea itself cannot be false (or true), truth and falsity become
possible only when some judgment or assertion is made with respect to
ideas. Descartess basic point here, which is still a commonplace in phi-
losophy, is that truth and falsity pertain not to concepts or ideas but,
rather, to assertions, statements, judgments, propositions, and the like.
For example, the concept or idea horse is neither true nor false. Only an
assertion or proposition that uses this concept (e.g., Some horses are
thoroughbreds) can be true or false. Since Descartess chief purpose
after establishing his own existence is to attain knowledge of things
existing outside his own mind (rst God, then other things), he here

In a brief but dicult passage in the Third Meditation, Descartes complicates mat-
ters by saying that some ideas, specically those of light and colours, sounds, tastes,
heat and cold, and the other tactile qualities (i.e., of secondary qualities, see chapter ,
section .) are materially false. A rough but useful characterization of material falsity
is this: In the Third Meditation, some sensory ideas were said to be materially false,
which means that they provide material for false judgment (AT VII [see also CSM
II , AT VII ]). (Such judgments occur when one arms that ... external
objects have properties in them that resemble our sensations of color, sounds, and other
so-called secondary qualities) (Gary Hateld, Descartes and the Meditations, p. ). In
the paragraph where Descartes introduces his view that ideas of color, heat and cold,
sound and so forth are materially false, he contrasts these with ideas that, as when he
examined the idea of the wax, he clearly and distinctly perceived to represent properties
of material things (if such things exist, which is unknown at this point), including size,
or extension in length, breadth, and depth; shape, which is a function of the boundaries
of this extension; position, which is a relation between various items possessing shape;
Meditation III
emphasizes the type of error he is most anxious to avoidjudging that

some idea in his mind corresponds to a reality outside his mind when it
does not.
Having touched on the basic issue still before him (i.e., the correspon-
dence or noncorrespondence of the ideas in his mind to things existing
outside his mind), Descartes now focuses this issue more sharply, by pre-
senting a possible classication of ideas.
. My ideas seem to fall into three classes:

(a) innate (i.e., inborn),
(b) adventitious (caused by objects located outside me),
(c) ctitious (invented by me).
It is important to understand that Descartes is not at this point arming

that his ideas actually fall into these three classes, but only that they ap-
pear to do so; for, remember, he does not yet knowbut is only begin-
ning to investigatethe causes of his ideas. Thus, he immediately adds,
in eect,
. I cannot yet be sure how my ideas really divide up; perhaps

they all fall into just one of these three classes.
As he puts it: But perhaps all my ideas may be thought of as adventi-

tious, or they may all be innate, or all made up; for as yet I have not
clearly perceived their true origin (CSM II , SPW , AT VII ).
Next, Descartes examines the reasons he formerly had for thinking
that some of his ideas fall into class (b) and, further, that those ideas re-
sembled the objects that caused them. In eect, he here puts up for ex-
amination a standard seventeenth-century view of perception that has
adherents to this day, called the representational theory of perception.
and motion, or change in position (CSM II , SPW , AT VII ). This suggests that
part of what Descartes also means by an ideas being materially false is that its content
exhibits features that are dierent from anything that belongs to material objects as they
would be described (assuming they existed at all) in a correct (Cartesian) physics, and
in that sense falsely portrays, or misleads us about, the nature of the world. Descartes
scholars have proposed dierent interpretations of his dicult notion of material falsity,
which will not be discussed in further detail in this book, since it does not play any role in
his central arguments. More will be said in chapter , however, about Descartess physics.
Descartes
According this to view, when a person perceives a material object, the

object causes an idea in the persons mind, the person has absolutely cer-
tain knowledge of the idea, and the idea gives the person knowledge of
the object by representing it, which it does by resembling at least some of
its properties (see Figure .).
Figure .
As we shall see later, Descartes himself eventually endorses such a

view, but what he examines in Meditation III is his previous reasons for
accepting it. We can paraphrase him this way:
. My reasons for thinking that some of my ideas come from, and

resemble, things existing outside me are () that nature has
taught me to think this and () that the ideas occur indepen-
dently of my will.
Descartes now nds these reasons to be very weak; for the rst one only
means that he has a natural, spontaneous inclination to believe that
some ideas proceed from and resemble external objects. Descartes con-
trasts this inclination with the natural light. This is none other than the
capacity for clear and distinct perception. He makes the strong claim that
whatever is revealed to me by the natural light ... cannot in any way be
open to doubt (CSM II , SPW , AT VII )a claim that should be
seen in the context of the oscillation between doubt and certainty dis-
cussed earlier. But whatever may be said in favor of the natural light, the
same cannot be said for nature; for just as natural impulses can drive
one to choose evil over good, so they can lead one to choose error rather
than truth. As for the fact that some ideas come independently of ones
will, it proves nothing; for perhaps some unknown faculty within the self
produces those ideas anyway, much as happens in dreams. And even if
the ideas did come from external objects, it would not follow that they
must resemble those objects. Thus, Descartes concludes that,
. These reasons are very weak; I must nd another way of in-

vestigating whether some of the things of which I possess
ideas exist outside me (CSM II , SPW , AT VII ).
Meditation III
3.2 Objective Reality and Formal Reality

To grasp Descartess way of investigating this question, we need to under-
stand a metaphysical framework that he presents in the rest of the short
but important paragraph from whose rst sentence we have just quoted.
The paragraph continues as follows:
In so far as the ideas are <considered> simply <as> modes of

thought, there is no recognizable inequality among them: they
all appear to come from within me in the same fashion. But in so
far as dierent ideas <are considered as images which> represent
dierent things, it is clear that they dier widely. Undoubtedly,
the ideas which represent substances to me amount to more and,
so to speak, contain within themselves more objective reality
than the ideas which merely represent modes or accidents. Again,
the idea that gives me my understanding of a supreme God, eter-
nal, innite, <immutable>, omniscient, omnipotent and the cre-
ator of all things that exist apart from him, certainly has in it
more objective reality than the ideas that represent nite sub-
stances. (CSM II , SPW , AT VII )
Here Descartes begins by adding a further point to those already

made about the nature of ideas. This new point is that ideas can be
regarded in two dierent ways: (a) as states of the thinker or modes of
thought and (b) as representations of their objects. An analogy may be
helpful. Consider a fresco painted on a wall. The fresco can be regarded
in two quite dierent ways. It can be regarded simply as an array of
colors and shapes on the wall. Or it can be regarded as a representation
of, say, Julius Caesar. Likewise an idea, since it is a mental representa-
tion of something, has two aspects. On the one hand, it is simply a
state of the thinkeran episode or occurrence in the thinkers mental
history. On the other hand, it is a representation of its object. This dis-
tinction is nicely explained in the following passage by A. S. Pringle-
Pattison:
It is important to remember . . . the distinction signalized by

Descartes between an idea as a mental state, a psychical occur-
rence, and the same idea functioning in knowledge and con-
veying a certain meaning... . [I]n [the former] respect all ideas
stand upon the same footing. . . . The treatment of ideas so
regarded belongs to psychology. But ideas not only exist as facts
Descartes
in the mental history of this or that individual; they have also ...
a content or meaning; they signify something other than them-
selves. We regard them, in Descartess words, as images, of
which one represents one thing and another a dierent thing,
and this is [an] important aspect of ideas.
Having distinguished these two ways of regarding ideas, or two aspects

of ideas, Descartes makes a further observation. Considered merely as
states of a thinker, all ideas have the same status: they are just modes or
(accidental) properties (of the thinker). But considered as representations
of their objects, they do not all have the same status; for some of them
represent other modes or properties; some represent nite substances,
and one represents an absolutely innite substance, namely, God. For
example, my idea of squareness represents squareness, which is a mode
or property of physical things, while my idea of fear represents fear,
which is a mode or property of thinking things. Again, my idea of myself
represents me, a nite thinking substance; while my idea of a stone rep-
resents a stone, which is a nite extended thing. Finally, my idea of God
represents an innite substance, namely, God. Diagrammatically, we can
represent the metaphysical framework that Descartes has just intro-
duced as shown in Figure ..
Of course, at the beginning of Meditation III, Descartes is not asserting
or assuming that all the objects of his ideas actually exist. On the con-
trary, at this point he is claiming to know only that one nite thinking
substance and some modes of that substance existnamely, himself and
his own thoughts. But he is asserting that at least his idea of God, and his
ideas of both thinking and extended nite substances and of their modes
all exist. He is asserting that all the items on the left-hand side of our
diagram exist and that a very few of the items on its right-hand side
existnamely, himself as a nite thinking substance and his own tho-
ughts as modes or properties of that substance. Whether any of the other
items on the right-hand side exist is still unknown, and is, indeed, the
very issue Descartes is beginning to investigate.

John Locke, An Essay Concerning Human Understanding, p. .

We here abstract from the point, made in chapter , section ., and to be revisited
in chapter , that on Descartess nal view a stone would not be a substance, but only a
mode of an all-encompassing extended substance (which Descartes would still presum-
ably regard as nite even if it were innitely extended, because it would lack some of the
attributes that belong only to the one truly innite substance, namely God).
Meditation III
Figure .
There is something else, however, that Descartes is asserting even at

the beginning of Meditation III. He is asserting that the objects repre-
sented by his ideasthe things on the right-hand side of our diagram
have dierent degrees of reality: nite substances have more reality than
modes, and an innite substance has more reality than nite substances.
You will ask how he can say this if he does not yet even know that any of
those things (except himself and his thoughts) exist. The answer is that
he is talking about the degree of reality that the things would have if they
existed. Thus, Descartess notion of reality must not be confused with
actual existence. An analogy may be useful: even if all the money in the
world were destroyed, it would still be true that $, is more than $.
Likewise, Descartes is saying that even if no innite substance exists,
innite substance has more reality than nite substance, and that even if
no nite substance exists, nite substance has more reality than a mode.
Notice also that Descartes frequently uses the term more perfection in
place of more reality: for him these terms are interchangeable. Histori-
cally speaking, this interchangeability, as well as the notion of degrees of
reality, derive largely from Plato and the Neo-Platonists, for whom being
(reality) and goodness are the same, and the highest good is also that
which is most real.
Now on the one hand, the concept of degrees of reality may strike you
as roughly right even though a bit fuzzy: you may be inclined to agree
that in some sense, God has more reality than a mere nite substance
and that a nite substance has more reality than a transient mode or
Descartes
property. On the other hand, the concept of degrees of reality may seem
obscure and questionable: you may want to protest that it makes no
sense to talk about degrees of realitythat reality is an all-or-nothing
matter, not one that admits of more or less. More will be said later
about the basis for Descartess concept of degrees of reality. But for now,
let us accept this concept, at least provisionally, for the sake of under-
standing how Descartess argument for the existence of God is supposed
to work; for the concept of degrees of reality lies behind two key ideas in
the argument, without which the argument cannot even be formulated.
The rst key idea is that some ideas represent their objects as having more
reality than other ideas represent their objects as having. Thus, the idea of a
nite substance represents its object as having more reality than the idea
of a mode represents its object as having, and the idea of God represents
its objectGodas having more reality than the idea of a nite sub-
stance represents its object as having. The second key idea, which actu-
ally provides the basis for the rst, is that the degree of reality that an idea
represents its object as having depends on the degree of reality possessed by
the object itself. In other words, if X has more reality than Y does, then the
idea of X represents X as having more reality than the idea of Y repre-
sents Y as having. So, for example, since God has more reality than any
nite substance, the idea of God represents (portrays) him as having
more reality than the idea of a nite substance represents it as having.
To express these two key ideas, Descartes uses a technical terminology
that he borrowed from medieval scholasticism and adapted for his own
purpose. He puts the rst key idea this way: Some ideas contain more objec-
tive reality than others. This claim means exactly the same thing as Some
ideas represent their objects as having more reality than other ideas represent
their objects as having. Thus, the term objective has here a completely dif-
ferent meaning from the modern one, where it has to do with objectivity
with what is actually the case independently of our beliefs and prejudices.
You must erase all such connotations from your mind in order to under-
stand Descartess meaning for this term. Instead, try to link the term ob-
jective reality with the notion of an object of thought; for an ideas objective
reality depends strictly on (the degree of reality possessed by) the object of
the idea, on what the idea is about. It pertains to the second of the two ways
of regarding ideas mentioned aboveto their nature as representations of

The notion of an ideas representing its object as having a certain amount of reality
comes from Curley, Descartes Against the Skeptics, p. .
Meditation III
their objects. This has nothing to do with objectivity in the modern sense.
It is also helpful to paraphrase Descartess term in various ways. For ex-
ample, in addition to Some ideas represent their objects as having more
reality than others, one could say, Some ideas exhibit more reality in their
contents than others. Some translators of the Meditations have chosen to
substitute a completely dierent term, less misleading to the modern ear
than objective reality. For example, the British philosophers Elizabeth
Anscombe and Peter Geach translate objective reality, very aptly, as rep-
resentative reality. Using this translation, Descartess point would be put
this way: some ideas contain more representative reality than others.
Descartess second key idea, we saw, is that the degree of reality an idea
represents its object as having depends on the degree of reality possessed by
the object itself (i.e., the degree of reality the object would have if it existed).
To use the new term just introduced, this would be put by saying that an
ideas degree of objective reality depends on the degree of reality had by the
ideas object. To express this second key idea, however, Descartes introduces
a second technical term, formal reality. His idea, expressed using this
term, in addition to objective reality, is that an ideas degree of objective re-
ality depends on its objects degree of formal reality. The term formal reality
is, of course, just as new to you as was the term objective reality. But at
least the term formal does not have the misleading connotations that ob-
jective has. Think of formal reality as the kind of reality that a thing has,
not in virtue of what it represents (that would be objective reality again),
but, rather, in terms of its status as either a mode or property, a nite sub-
stance, or an innite substance. Formal reality, then, is not too far removed
from what people usually mean simply by reality. It refers to a things ac-
tual status in the worldor at least to the status it would have if it existed.
(Anscombe and Geach translate it as actual or inherent reality.) Notice,
then, that while objective reality is a special type of reality that belongs only
to ideas in virtue of their representational function, everything, including
ideas, has some degree of formal reality. Indeed ideas, being modes or prop-
erties of a thinker, have (along with modes or properties of other sub-
stances) the lowest degree of formal reality in Descartess three-level
hierarchy (modes, nite substances, and innite substance).
We can now represent this hierarchy diagrammatically as shown in
Figure ..

Ren Descartes, Descartes: Philosophical Writings, pp. , .
Descartes
Figure .
Note that the diagram positions ideas only in terms of their objective
reality. In terms of formal reality, their position is at the right-hand
bottom of the diagram: like fear or squareness, an idea is a mode or prop-
erty of a substance.
3.3 The Core Argument

The hierarchical metaphysical framework just presented provides nearly
all the materials needed in order to formulate Descartess basic argument
for Gods existence in Meditation III, which we shall call his core argu-
ment. Indeed, the only further element we need to add is a general prin-
ciple concerning the causes of ideas. It should come as no surprise that
Descartes uses such a principle, for as previously noted, his argument is
essentially a causal one, which reasons from the idea of God to God him-
self as the only possible cause of this idea. The principle that he uses is:
The cause of an idea must have as much formal reality as the idea
contains objective reality.
It is crucial to understand that this principle does not say that an idea
must have as much objective reality as it has formal reality: it does not
compare ideas with ideas. It compares the degree of formal reality had by
the cause that produced an idea with the degree of objective reality con-
tained in that idea. Thus the principle says two dierent things: () an
idea requires a cause, and () this cause must be an adequate one. In
other words, not only must an idea have a cause, but the more objective
Meditation III
reality the idea hasthe more reality it represents its object as having
the more formal reality its cause must have. Despite the abstract nature
of Descartess principle and the technical terminology he uses to express
it, the principle is rather plausible. Descartes brings this out in his Princi-
ples of Philosophy, Part I, no. , by applying it to a concrete example:
[T]he greater the amount of objective perfection [= objective re-

ality] they [= ideas] contain within themselves, the more perfect
their cause must be. For example, if someone has within himself
the idea of a highly intricate machine, it would be fair to ask what
was the cause of his possession of the idea: did he somewhere see
such a machine made by someone else; or did he make such a
close study of mechanics, or is his own ingenuity so great, that he
was able to think it up on his own, although he never saw it any-
where? All the intricacy which is contained in the idea merely
objectivelyas in a picturemust be contained in the cause,
whatever kind of cause it turns out to be; and it must be con-
tained not merely objectively or representatively, but in actual
reality, either formally or eminently, at least in the case of the
rst and principal cause. (CSM I , SPW , AT VIIIA )
To see still better the plausibility of Descartess principle, consider also

the following elaboration of his example, by John Cottingham:
Recall Descartes example of the highly intricate machine; and

for the sake of simplicity let us follow Descartes own compar-
ison between ideas and pictures, and consider the case of a
drawing or a picture rather than an idea. Suppose a ve-year-old
child produces a highly complicated design for a computera
design which we know could only be produced by a highly skilled
mathematician with a mental age vastly superior to the childs.
The fact that the design is only a drawing and not an actual com-
puter does not block the causal question: the representative or
objective intricacy of the design still has to be accounted for. Of
course, the child might simply have copied down the drawing
from a book. But this simply pushes the argument one stage fur-
ther back. We are, it seems, justied in asserting that somewhere
along the line of causation there must be an actual entity or being
that really does possess sucient complexity to account for the
Descartes
complexities which are to be found in the design. And what goes

for the drawing goes equally for an idea: complex representa-
tional content requires a complex cause.
To arrive at Descartess basic or core argument for Gods existence,

we need only relate the principle just presented to the metaphysical
framework described in the previous section. Specically, we need to ask
whether the principle shows how any of our ideasthose of modes,
nite substances, or innite substanceare caused. Can it take us from
knowledge of items on the left-hand side of our diagrams to knowledge
of the existence of any of the items on the right-hand side? Let us con-
sider rst the ideas of modes or propertiesthose that contain the
lowest degree of objective reality. These ideas may, in conformity with
Descartess principle, be caused by anything that has at least as much
formal reality as they contain objective reality. So ideas of modes can be
caused by modes, which have exactly as much formal reality as those
ideas contain objective reality. But they can also be caused by nite sub-
stances, which have more formal reality than those ideas contain objec-
tive reality, or by God, who has even more formal reality. The mere fact
that I have an idea of a mode, then, does not prove that any mode (other
than that idea itself, which is a mode of me) actually exists, since this
idea could be caused instead by a nite substance or by God. Consider,
next, the ideas of nite substances. These ideas cannot, in conformity
with Descartess principle, be caused by modes. For a mode has less for-
mal reality than the idea of a substance contains objective reality. So if a
mode were to cause the idea of a nite substance, this would violate the
principle that the cause of an idea must have as much formal reality as
the idea contains objective reality. On the other hand, the idea of a nite
substance can, in conformity with Descartess principle, be caused in
either of two dierent ways: by a nite substance, which has exactly as
much formal reality as the idea contains objective reality, or by God,
who has even more formal reality. So again, the mere fact that I have an
idea of a nite substance cannot prove that such a substance exists,
since this idea might be caused by an innite substance (or by myself,
since I am a nite thinking substance). So at this point, Descartess prin-
ciple has still not enabled him to know the existence of anything beyond

John Cottingham, Descartes, pp. . For another helpful explanation of the same
point, see Bernard Williams, Descartes: The Project of Pure Enquiry, pp. .
Meditation III
himself (and that knowledge, of course, stems only from the cogito, not
from any principle about the causes of ideas). However, there still
remains one idea to be consideredthe idea of an innite substance, or
God. This is the idea of a substance that is innite, <eternal, immu-
table>, independent, supremely intelligent, supremely powerful, and
which created both myself and everything else (if anything else there
be) that exists (CSM II , SPW , AT VII ). How is this idea
produced? It cannot, in conformity with Descartess principle, be caused
by a mode; for a mode has far less formal reality than the idea of God, as
just described, contains objective reality. But neither can it be caused by
a nite substance; for such a substance still has less (innitely less) for-
mal reality than this innitely rich idea of God contains objective reality.
In fact, it is now obvious that there is only one way the idea of God could
be caused. It could only be caused by God himself, for only God himself
has as much formal reality as the idea of God contains objective reality.
Thus, the idea of God diers from all other ideas. It is a uniquely privi-
leged idea; for, alone among all ideas, the idea of God is such that from
the mere fact that I have the idea, it follows that the object of that very
ideaGod himselfis also the cause of that idea and must, therefore,
really exist.
This core argument for Gods existence can be formulated as follows:
. The cause of an idea must have as much formal reality as the idea con-
tains objective reality.
. Only a perfect God has as much formal reality as my idea of God con-
tains objective reality.
. The cause of my idea of God is a perfect God (from propositions and ).
. A perfect God really exists (from proposition ).
We saw earlier that objective reality and formal reality are tech-
nical terms that Descartes uses in expressing the two key ideas involved
in his argument: () that some ideas represent their objects as having
more reality than other ideas represent their objects as having, and ()
that the degree of reality an idea represents its object as having depends
on the degree of reality possessed by the object itself. Notice, then, that
these two key ideas can be expressed without even using the terms ob-
jective reality and formal reality, since we in fact introduced those
ideas before reexpressing them in those terms. Likewise, Descartess core
Descartes
argument can be formulated without the terms objective reality and

formal reality. The formulation would go this way:
. The cause of an idea must have as much reality as the idea represents its object
as having.
. Only a perfect God has as much reality as my idea of God represents God as
having.
. The cause of my idea of God is a perfect God (from propositions and ).
. A perfect God really exists (from proposition ).
So if you are having trouble grasping Descartess argument, this trouble

should not stem from his technical terminology. Rather, it should stem from
the substantive claims the argument makesthat there are dierent degrees
of reality, that ideas represent their objects as having these dierent degrees
of reality, that the idea of God represents God as having innite reality, and
that the only adequate cause of an ideas representing its object as having a
given degree of reality is a thing which actually has at least that degree of
reality. We shall critically examine some of these claims in section .
3.4 The Central Argument of Meditation III:

The Subargument, the Core Argument, and the Sequel
In Meditation III, Descartes does not formulate his core argument as con-
cisely as we have just done. Rather, he embeds it within a longer argu-
ment that starts from certain very general principles about causality and
ends with the vindication of his criterion of truth. His discussion can be
broken down into three episodes: () an argument, which we will call
the subargument, intended to establish the core arguments rst pre-
mise; () a survey of ideas that is designed to exhibit the uniquely pri-
vileged nature of the idea of God, and that culminates in Descartess
discovery of the core arguments second premise; and () an argument,
which we will call the sequel, that goes from the core arguments con-
clusion to the vindication of the clarity and distinctness criterion of
truth. The subargument argument, core argument, and sequel constitute
one continuous extended argument that is usually considered to be the

The idea of restating Descartess argument without using his technical terms is due
to Curley, Descartes Against the Skeptics, p. (though Curley himself paraphrases out
only the term objective reality).
Meditation III
central metaphysical argument of the Meditations. The purpose of this

section is to present this extended argument.
The subargument argument begins this way:
Now it is manifest by the natural light that there must be at least

as much <reality> in the ecient and total cause as in the eect
of that cause. For where, I ask, could the eect get its reality
from, if not from the cause? And how could the cause give it to
the eect unless it possessed it? It follows from this both that
something cannot arise from nothing, and also that what is more
perfectthat is, contains in itself more realitycannot arise
from what is less perfect. (CSM II , SPW , AT VII )
Descartess reasoning here starts from a premise, that can be extracted

from his two rhetorical questions, and which we shall call the precon-
tainment principle. It may be put this way:
() A cause must precontain the reality of its eect.
An alternative way to read the passage is to interpret Descartess two

rhetorical questions as containing the following argument for ():
(-) An eect can get its reality only from its cause.
() A cause can give its reality to its eect only if this cause pos-
sesses that reality.
____________________________________________________
? () A cause must precontain the reality of its eect.
In fact, this argument is maximally faithful to the text. However, it is not

stronger than the version that starts with () as a basic premise. For (-)
means either that
(-a) An eect must take its reality from its causeits reality
must be transferred to it from its cause
or
(-b) An eect can be produced only by its cause.

Descartes
But if (-) means (-a), then (-) comes too close to () to avoid begging
the question. On the other hand, if (-) means (-b), then () must be
replaced by
(*) A cause can produce an eect only by precontaining the

reality of that eect
which again comes so close to () that it begs the question.

Another reason not to favor the alternative reading over the one that
starts with () is that there are several passages where Descartes says that
() is the true starting-point of his argument. Thus, in one of his letters, he
writes: I proved this [i.e., that a mind could not have the idea of a perfect
God unless such a God really existed] from the principle that there can be
nothing in an eect which is not previously present in the cause (to Hyper-
aspistes, August , CSMK , AT III ). That he regards the precon-
tainment principle as absolutely basic is also conrmed by his reply to one
of Gassendis objections, where Descartes refers to the axiom There is
nothing in the eect which did not previously exist in the cause (CSM II
, AT VII ; my emphasis), and by a remark in his reply to the Second
Set of Objections, where he gives, as the reason why nothing cannot be the
cause of a thing, that such a cause would not contain the same features as
are found in the eect (CSM II , AT VII ). Of course, it can hardly be
said that the meaning of the precontainment principle is clear, and later we
shall have to ask what it means. Our present point is simply that the prin-
ciple may fairly be taken as the opening premise of Descartess argument.
Descartess second premise, which he leaves unstated in the above
passage, is
() There are degrees of reality.
As we have already seen, this idea, which can be traced back at least to
Plato, is essential to Descartess argument.
From these two premises, Descartes says that two thing follow (It
follows both that ... and that ...). First, he says it follows that
() Something cannot come from nothing.
This claim, which might be called the nothing-comes-from-nothing

principle, can be traced back to ancient Greek philosophy and in medieval
Meditation III
times was expressed in the Latin dictum Ex nihilo, nihil t (From nothing,
nothing comes). It means that something cannot be caused to exist or
occur by nothing(ness) or nonbeing. Descartes derives it from (). His
reasoning is, presumably, that since a cause must precontain the reality
of its eect, and since nothing(ness) cannot possibly contain anything,
nothing(ness) cannot be the cause of anything. Second, Descartes says
it follows that what is more perfectthat is, contains in itself more
realitycannot arise from what is less perfect, in other words,
() A cause must contain at least as much reality (perfection) as

its eect.
This claim, which reasserts Descartess paragraphs opening statement

that there must be at least as much <reality> in the ecient and total
cause as in the eect of that cause, is a general principle that says that
a cause must be adequate to its eect. We may call it, following John
Cottingham, the Causal Adequacy Principle. Descartes derives it from
() and (). His reasoning, presumably, is that since the cause must pre-
contain the reality of its eect and given that the cause and the eect
each have a certain degree of reality, the eect cannot have more reality
than the cause; for if it did, then its surplus reality could not have been
precontained in the cause.
The Causal Adequacy Principle was a commonplace of Scholastic phi-
losophy. Descartess innovation was to apply this principle to ideas, and
in particular, to the objective reality or representational content of
ideas. He does this in the continuation of the passage quoted above:
And this is transparently true not only in the case of eects

which possess <what the philosophers call> actual or formal re-
ality, but also in the case of ideas, where one is considering only
<what they call> objective reality. A stone, for example, which
previously did not exist, cannot begin to exist unless it is pro-
duced by something which contains, either formally or eminently
everything to be found in the stone... . But it is also true that the
idea ... of a stone, cannot exist in me unless it is put there by
some cause which contains at least as much reality as I conceive
to be ... in the stone... . [I]n order for a given idea to contain

Cottingham, Descartes, p. .
Descartes
such and such objective reality, it must surely derive it from some
cause which contains at least as much formal reality as there is
objective reality in the idea. (CSM II , SPW , AT VII )
Descartess language here is very compact. In the sentence about the

stone, he is asserting two dierent things rather than only one. First, he
is asserting that a stone cannot begin to exist unless it is produced by
somethingunless it has some cause. He then says the same thing
about the idea of the stone: it, too, cannot exist in me unless it is put
there by some cause. Second, he is asserting that the causes of the stone
and of the idea of the stone must be adequate ones: they must be causes
which contain enough formal reality to produce the eect. He seems
not to distinguish these two points; his sentence structure fuses them
together. Yet, as Bernard Williams has insightfully pointed out, there are
two importantly dierent points here. One is the Causal Adequacy Prin-
ciple, calling for a certain type of causean adequate one. The other is a
principle that simply calls for a cause, period. So, in addition to the Causal
Adequacy Principle, Descartes is here also invoking the principle that
() Everything must have a cause.
This principle, which is commonly called the causal principle or (fol-

lowing David Hume) the causal maxim, is dierent fromin a sense
more basic thanthe Causal Adequacy Principle; for the latter only says
that if something has a cause, then that cause must have as much reality
as its eect; it does not say whether a thing must have a cause. By contrast,
the causal maxim asserts the universal need for a cause. (Hume actually
formulates the maxim as whatever has a beginning [emphasis added] of
existence must have a cause of existence. This has the advantage of not
entailing that God must have a cause, since God is supposed to exist eter-
nally. But some philosophers, including Descartes, would say that God is
his own cause; not because he existed before he existed, which would be
absurd, but because he exists necessarily, in virtue of his own nature.)
We already know the basis of the Causal Adequacy Principlethat is,
of (). It is derived from () and (). But what is the basis for ()? One
possibility would be to say, as have some philosophers, that () needs no
defensethat it is simply self-evident that everything must have a cause.

Williams, Descartes: the Project of Pure Enquiry, p. .
Meditation III
But Descartes does not treat () as self-evident, at least not in this pas-
sage; for he presents the need for a cause of the stone and the need for a
cause of the idea of the stone as specic illustrations of, or as applica-
tions of, () and/or (). So ()the general principle calling for those
causesis derived from one or both of those propositions. But () does
not follow from (): () calls for an adequate cause in cases where we
admit that there is a cause but says nothing about whether there needs
to be a cause in the rst place. It seems, then, that Descartes takes () to
follow from (), or perhaps simply to be equivalent to (). His reasoning
seems to be that since something cannot come from nothing, everything
must have a cause; he seems to derive the causal maxim from the noth-
ing-comes-from-nothing principle. In the next section, we will see that
there is a hidden diculty (rst spotted by Hume) in this reasoning.
Before continuing with the argument, we should address a delicate
terminological matter. In the passage we have just discussed and in
others that we shall encounter, Descartes uses the adverbial locution
that the cause must formally or eminently contain all the reality of the
eect. An examination of the relevant texts shows that for Descartes,
whenever a cause formally or eminently contains all the reality of its ef-
fect, it has at least as much formal reality as its eect contains either
formal or objective reality, and that is really the only point one needs to
know to follow Descartess argument. But for readers who are interested
(others can skip the next two paragraphs without losing track of the ar-
gument), we shall here oer an explanation of Descartess terminology.
There are three passages in the Meditations that are especially relevant
to the question of what formally or eminently contain means. Here
they are:
[A] A stone, for example, which previously did not exist, cannot
begin to exist unless it is produced by something which contains,
either formally or eminently everything to be found in the stone;
similarly, heat cannot be produced by an object which previously
was not hot, except by something of at least the same order
<degree or kind> of perfection as heat, and so on. (CSM II , AT
VII , SPW )
[B] As for all the other elements which make up my ideas of cor-
poreal things, namely extension, shape, position, and move-
ment, these are not formally contained in me, since I am nothing
Descartes
but a thinking thing; but since they are merely modes of a sub-
stance, and I am a substance, it seems possible that they are con-
tained in me eminently. (CSM II , AT VII , SPW )
[C] So the only alternative is that it [i.e., the cause of my ideas of
sensible objects] is that it is in another substance distinct from
mea substance which contains either formally or eminently all
the reality which exists objectively in the ideas produced by this
faculty... . This substance is either a body, that is, a corporeal
nature, in which case it will contain formally <and in fact> every-
thing which is to be found objectively <or representatively> in
the ideas; or else it is God; or some creature more noble than
body, in which case it will contain eminently whatever is to be
found in the ideas. (CSM II , AT VII , SPW )
These passages show that Descartess distinction between formal and

eminent containment does not introduce some third type of reality, em-
inent reality, in addition to formal and objective reality; rather, as the
adverbial locution formally or eminently contain implies, the distinc-
tion refers to the way in which reality, whether formal or objective, is
contained in anything. Passages A and C show that one object can for-
mally or eminently contain the reality contained in another object, such
as a stone or a quantity of heat; and also that an object can formally or
eminently contain the objective reality of an idea. Passage C shows that
at least one way in which an object can eminently contain the formal re-
ality of another object, or the objective reality of an idea, is by containing
more (or a higher degree of) formal reality than the other object contains
formal reality, or more formal reality than the idea contains objective re-
ality. This point accords with John Cottinghams translators note to the
paragraph where Descartes expounds the subargument, in which Cot-
tingham writes that in scholastic terminology, to possess a property
formally is to possess it literally, in accordance with its denition; to
possess it eminently is to possess it in some higher form (CSM II ,
note ; SPW , note ). Passage B strongly suggests, and Cottinghams
note implies, a special requirement for it to be true that one thing, X,
formally contains either the formal reality of another object Y or the ob-
jective reality of an idea I; namely, that the formal reality of X be not only
equal to but also of the same kind as the formal reality of Y or the objective
reality of I, where kind means either (a) mental or (b) physical. The rea-
son why B strongly suggests this requirement is that Descartes gives, as
Meditation III
the reason why I do not formally contain the other elements which make
up my ideas of corporeal things, namely extension, shape, position, and
movement is that I am nothing but a thinking thing. So, the reason
why I do not formally contain the reality of these elements but contain
that reality only eminently is not, contrary to what one might think,
merely that I contain more formal reality than they contain objective re-
ality (or possibly formal reality: it isnt clear whether elements which
make up my ideas of corporeal things refers directly to the properties of
those things, or to the representational content of the ideas of those
properties, or to both of these), but rather that as a thinking thing I
contain a dierent kind of reality than they do, by virtue of their invol-
ving extension and its modes. The requirement that formal contain-
ment involve sameness of ontological kind also seems to be implied by
passage C, where Descartes says that only corporeal (i.e., bodily) nature
formally contains the objective reality contained in ideas of bodies, and
the gloss of formally as in fact (added in the French translation as en
eet) reinforces this requirement as well.
Against the background of these these textual points, we can see the
purpose of Descartess distinction between formal and eminent contain-
ment. It is to allow for four dierent ways in which a cause can contain at
least as much reality as its eectfour ways in which the Causal Ade-
quacy Principle can be satised. First, the cause may contain exactly the
same degree and the same kind (i.e., physical or else mental) of formal
reality as the eect contains either objective reality (if it is an idea) or
formal reality. In that case, Descartes says that the cause formally con-
tains all the reality that the eect contains, or formally contains every-
thing to be found in the eect. Second, the cause can contain more than
but the same kind of formal reality as the eect contains objective reality
or formal reality. Third, the cause can contain at least as much as but a
dierent kind of formal reality than the eect contains formal or objec-
tive reality. Fourth, the cause can contain both more than and a dierent
kind of formal reality than the eect contains objective or formal reality.
In the second, third, and fourth cases, Descartes says that the cause emi-
nently contains all the reality that the eect contains. A simpler way to
put all of this is to say that a cause formally contains the reality of its ef-
fect when it has exactly the same degree and kind of formal reality as its
eect has objective or formal reality, and that a cause eminently contains
the reality of its eect in all other cases where it contains at least as much
formal reality as its eect contains formal or objective reality, that is, in
Descartes
all cases where it does not formally contain that reality. The most impor-
tant point for Descartess argument, as previously noted, is that when-
ever a cause either formally or eminently contain contains the objective
reality of an idea, the rst premise of his core argument is true: the cause
of the idea contains at least as much formal reality as the idea contains
objective reality.
To return to the argumentassuming that () is established, Desca-
rtes now applies () to the objective reality of ideas. The objective re-
ality (representational or informational content) of an idea of a stone
(i.e., the fact that this idea represents a stone as having a certain degree
of reality) must have a cause, no less than the fact that a stone exists.
Descartess next step, then, is to derive from () the following:
() The objective reality of an idea must have a cause.
To arrive at the core arguments rst premise, Descartes uses one fur-
ther premise. That premise is contained in the following passage:
And although the reality which I am considering in my ideas is

merely objective reality, I must not on that account suppose
that the same reality need not exist formally in the causes of my
ideas, but that it is enough for it to be present in them objec-
tively. For just as the objective mode of being belongs to ideas by
their very nature, so the formal mode of being belongs to the
causes of ideasor at least to the rst and most important
onesby their nature. And although one idea may perhaps
originate from another, there cannot be an innite regress here;
eventually one must reach a primary idea, the cause of which
will be like an archetype which contains formally <and in fact>
all the reality <or perfection> which is present only objectively
<or representatively> in the idea. (CSM II , SPW , AT VII
)
Here Descartes admits that the objective reality of an idea is sometimes

borrowed or derived from the objective reality of one or more other ideas.
A modern example might be that an engineers idea of a rocket engine
(one that carries its own air supply as well as its fuel) could be derived
from the engineers idea of a jet engine (one that takes its air supply from
outside): perhaps the engineer constructed his idea of a rocket engine
Meditation III
from elements in his idea of a jet engine. But Descartes would insist that
such an explanation of how the engineers idea of a rocket engine origi-
nated is incomplete. For now we need an explanation of how the engi-
neer acquired the idea of a jet engine. In the end, Descartes would say,
this explanation cannot be just the fact that the engineer constructed it
from yet another idea. Rather, it must be either that the engineer has
such a brilliant and inventive mind that he created the idea of a jet engine
or the idea(s) from which it is constructed, or else that he got that idea
from observing a jet engine itself or from something having the same as
or a higher level of complexity than a jet engine. In other words, ulti-
mately, the cause of an ideas representational content cannot be just the
representational content of another idea. It must be some nonrepresen-
tational fact about the world. Representational content is, in the end,
parasitic on nonrepresentational facts.
Perhaps the following analogy can help to bring out Descartess point.
Imagine a mirror image. This mirror image could itself be a reection
of an image reected by a second mirror. And the image in the second
mirror could be the reection of an image reected by a third mirror.
Indeed, the mirrors could be so arranged that there might be a very long
series of mirrors, each reecting a mirror image reected by its prede-
cessor in the series. But this series could not go on innitely. It must
terminate in a mirror image which is not itself the image of an image, but
which is caused by something other than a mirror image. The same holds
for the representational content of an idea: it may be derived from an-
other ideas contentand that from yet another ideas content. But such
a series cannot continue innitely; in the long run it must terminate in a
content that is caused by something other than an ideas content. In Des-
cartess terminology, then, the objective reality of an idea must ulti-
mately be caused by the formal reality of something, not just by the
objective reality of another idea.
Descartess premise, then, could be put this way:
Although this cause may be the objective reality of another idea

or ideas, ultimately (or in the long run) it must be the formal
reality of something.
To simplify matters a little, however, let us formulate the premise as:
() This cause must be the formal reality of something.

Descartes
We have now assembled all the materials Descartes oers in support

of the core arguments rst premise: () says that a cause must contain as
much reality as its eect, () says that the objective reality of an idea
must have a cause, and () says that this cause must be the formal real-
ity of something. From these three statements, Descartes derives the
premise
() The cause of an idea must have as much formal reality as the

idea contains objective reality.
The next episode of the Meditation is not, strictly speaking, part of the
extended argument (subargument, core argument, and sequel). Rather,
its function is to bring out the unique nature of the idea of God, thereby
preparing the way for the core arguments second premise. To some extent,
we have already anticipated this material in our explanation of the core
argument in the previous section. But it is worth seeing briey how Des-
cartes himself presents the matter. He begins dramatically, by announcing
an implication of () and of the weakness of his previous reasons for
thinking that some of his ideas are caused by things outside himself:
If the objective reality of any of my ideas turns out to be so great

that I am sure the same reality does not reside in me, either for-
mally or eminently, and hence that I myself cannot be its cause,
then it will necessarily follow that I am not alone in the world,
but that some other thing which is the cause of this idea also
exists. But if no such idea is to be found in me, I shall have no
argument to convince me of the existence of anything apart from
myself. For despite a most careful and comprehensive survey,
this is the only argument I have so far been able to nd. (CSM II
, SPW , AT VII )
In other words, () now oers the only possible way of escaping from
solipsism, that is, from the extraordinary view that only I and my own
thoughts exist.
Descartes now proceeds, in eect, by raising the following causal ques-
tion: is the objective reality of any of my ideas such that the ideas object
(what the idea is of or about) must also be the ideas cause? If the answer
is yes, then this will show that the object of the idea cannot be nonexis-
tent or merely ctitious, but must really exist, in order to cause the idea.
Meditation III
So Descartes makes an inventory of his ideas, with a view to answering

his causal question. He nds that, in addition to his idea of himself as a
nite, thinking substance, he has ideas of
(a) God,
(b) inanimate physical objects,
(c) other humans, animals, and angels.
My ideas of other humans, animals and angels, says Descartes, could

easily be constructed by combining elements taken from my ideas of
myself, of inanimate physical objects, and of God even if no other hu-
mans, animals or angels existed. So my having ideas in class (c) does not
show that other humans, animals, or angels exist. My ideas of inani-
mate physical objects properties could originate entirely from myself
if only because they are ideas of modes, which contain less objective
reality than the formal reality I myself contain as a nite substance. My
ideas of those objects themselves could also originate from myself, since
they are ideas of nite substances and I, as a nite thinking substance,
possess sucient formal reality to cause such ideas. So my having ideas
in class (b) does not show that physical objects really exist. So, says
Descartes,
[T]here remains only the idea of God; and I must consider

whether there is anything in the idea which could not have orig-
inated from myself. By the word God I understand a substance
that is innite, <eternal, immutable>, independent, supremely
intelligent, supremely powerful, and which created both myself
and everything else (if anything else there be) that exists. All
these attributes are such that, the more carefully I concentrate
on them the less possible it seems that the idea I have of them
could have originated from me alone. So from what has been said
it must be concluded that God necessarily exists. (CSM II ,
SPW , AT VII )
Here Descartes nds, in the rst place, that the idea of God contains
more objective reality than Descartes possesses formal reality. So, his
having this idea does show, nally, that something other (and greater)
than himself existsthat he is not alone in the world, that solipsism is
false. But further, Descartes nds that only one entity possesses enough
Descartes
formal reality to cause this great idea, namely, God himself. So God must
really exist.
Descartess survey of ideas, then, has culminated in the completion of
his core argument; for the survey has led to his discovering the core argu-
ments second premise,
() Only a perfect God has as much formal reality as my idea of

God contains objective reality.
But from () and (), there follows
() The cause of my idea of God is a perfect God.
Finally, it obviously follows from () that
() A perfect God exists.
As important as is this conclusion, it is not yet the nal conclusion of

Descartess extended argument; for remember what Descartess project in
Meditation III wasto vindicate his criterion of truth. To achieve this pur-
pose, Descartes declared that he needed to show both that (a) God exists
and (b) God is not a deceiver. We have yet to see how he establishes (b)
and exactly how (b) is supposed to vindicate the clarity and distinctness
criterion of truth. To see this, we must look at the third and nal stage of
Descartess extended argumentthe part we are calling the sequel.
The sequels key premise is this:
() To deceive is an imperfection.
From this premise, together with (), Descartes concludes,
() God is not a deceiver.

It is also possible to read the episode just discussed as providing an alternative proof
of this conclusion, as follows:
i. The object of an idea must also be its cause just in case only that object has enough
formal reality to cause this idea.
ii. Only a perfect God has enough formal reality to cause my idea of God.
iii. The cause of my idea of God is a perfect God. [from (i) and (ii)]
iv. A perfect God exists. [from (iii)]
Meditation III
Descartes presents this brief piece of reasoning in a number of places,

saying that its premise is known by the natural light. For example, near
the end of Meditation III, he writes:
By God I mean ... the possessor of all ... perfections ... who is
subject to no defects whatsoever. It is clear enough from this
that he cannot be a deceiver, since it is manifest by the natural
light that all fraud and deception depend on some defect. (CSM
II , SPW , AT VII )
Again, at the beginning of Meditation IV, he says:
To begin with, I recognize that it is impossible that God should

ever deceive me. For in every case of trickery or deception some
imperfection is to be found; and although the ability to deceive
appears to be an indication of cleverness or power, the will to
deceive is undoubtedly evidence of malice or weakness, and so
cannot apply to God. (CSM II , SPW , AT VII )
Finally, Descartes maintains, () vindicates the clarity and distinct-

ness criterion of truth. This can be seen in the Synopsis of the Medita-
tions, where he announces.
In the Fourth Meditation it is proved that everything that we

clearly and distinctly perceive is true ... (CSM II , SPW , AT
VII )
The passage in Meditation IV to which he is referring is this:
[E]very clear and distinct perception is undoubtedly something,

and hence cannot come from nothing, but must necessarily have
God for its author. Its author, I say, is God, who is supremely
perfect, and who cannot be a deceiver on pain of contradiction;
hence the perception is undoubtedly true. (CSM II , SPW ,
AT VII )
In Meditation V, he makes the point this way:

Descartes
Now, however, I have perceived that God exists, and at the same
time I have understood that everything depends on him, and
that he is no deceiver; and I have drawn the conclusion that
everything which I clearly and distinctly perceive is of necessity
true. (CSM II , SPW , AT VII )
These passages show that Descartes believes he can go from God is

not a deceiver to whatever I clearly and distinctly perceive is true. But
they do not show exactly how the one proposition is supposed to lead to
the other. So let us try to ll in the missing step(s) in the reasoning. The
key premise that we need is this:
() If my clear and distinct perceptions could be false, then God

would be a deceiver.
This premise rests on the assent-compelling nature of clear and distinct

perceptions. Recall that Descartes says that while one is having a clear
and distinct perception, one cannot doubt it: one is compelled to assent
to it. Thus, if one of my clear and distinct perceptions could nevertheless
be false, then I would be making an error that I was powerless to correct:
I would be irremediably deceived. And since God created me and my cog-
nitive equipment, this would mean that he was a deceiver.
With the help of (), Descartes can at last derive his nal conclusion;
for it follows from () and () that
() My clear and distinct perceptions cannot be false, that is,

whatever I clearly and distinctly perceive is true.
This completes Descartess extended argument. Descartess criterion of

truth, originally extracted from the cogito, is now nally vindicated by
Gods veracity (truthfulness).
Since the extended argument is quite lengthy, it may be helpful to
summarize it, by listing all of its steps in one place. Here, then, is the
entire argument. Steps ()() are the subargument; steps ()() are
the core argument; steps ()() are the sequel.
. A cause must precontain the reality of its eect. (premise).

. There are degrees of reality (premise).
. Something cannot come from nothing (from step ).
Meditation III
. A cause must contain at least as much reality as its eect (from steps
and ).
. Everything must have a cause (from step ).
. The objective reality of an idea must have a cause (from step ).
. This cause must be the formal reality of something (premise).
. The cause of an idea must have as much formal reality as the idea
contains objective reality (from steps , , and ).
tains objective reality (premise).
. The cause of my idea of God is a perfect God (from steps and ).
. A perfect God exists (from step ).
. To deceive is an imperfection (premise).
. God is not a deceiver (from steps and ).
. If my clear and distinct perceptions could be false, then God would
be a deceiver (premise).
. My clear and distinct perceptions cannot be false, that is, whatever I
clearly and distinctly perceive is true (from steps and ).
In concluding this section, we should note that Descartes thought that

his proof that a nondeceiving God exists raises a problem of its own. This
problem, which is the main topic of Meditation IV, is the following: How
does it happen that I sometimes make errors (hold false beliefs), if God is
not a deceiver? By proving that God is not a deceiver, hasnt Descartes
proved too much, that is, proved that error is impossible? The gist of Des-
cartess answer to this question is that error arises from misusing my free
will, by adopting beliefs even about matters that my intellect does not clearly
and distinctly perceive. So long as I restrict myself to matters that I do clearly
and distinctly perceive, error is impossible (CSM II , SPW , AT VII
). In the following chapter, we shall analyze Meditation IV and con-
sider more closely Descartess theory of error. But in the rest of the present
chapter, we shall oer an assessment of Descartess extended argument.
4. Criticisms of Descartess Central

Argument in Meditation III
Descartess central argument in Meditation III raises many issues and can
be criticized in a number of dierent ways. We shall consider three pos-
sible lines of criticism: one focusing on the subargument, a second on
Descartes
Descartess overall strategy, and a third on the core arguments second

premise.
4.1 The Subargument

Our critical discussion of the subargument will focus on its rst ve steps
(see p. ). We shall discuss three diculties in those ve steps.
.. The Precontainment Principle

The rst diculty is that premise (), the precontainment principle,
involves a highly problematic conception of causation (cause-and-eect).
Exactly what does it mean to say that a cause must precontain the reality
of its eect, or that there can be nothing in an eect which is not previ-
ously present in the cause (letter to Hyperaspistes, August , CSMK
, AT III )? In this subsection, we shall examine some possible
answers to this question.
The simplest interpretation of the precontainment principle would be
that the cause must actually precontain its eect, much as a fetus is pre-
contained in the womb. In an illuminating article, the distinguished his-
torian of ideas Arthur O. Lovejoy () discusses what he calls the
preformationist assumption about causality. He writes,
That there cannot be more in the eect than there is in the

cause is one of the propositions that men have been readiest to
accept as axiomatic; a cause, it has been supposed, does not ac-
count for its eect, unless the eect is a thing which the eye of
reason could somehow discern in the cause, upon a suciently
thorough analysis.
Lovejoy goes on to show that this preformationist assumption can

be traced back to antiquity, was pervasive both in medieval and in sev-
enteenth- and eighteenth-century philosophy, and continued to be put
forward by some twentieth-century thinkers.
If the precontainment principle is interpreted in this way, however,
then it is open to a serious objection: it seems to be simply false. This was

Arthur O. Lovejoy, The Meanings of Emergence and its Modes, p. .
Meditation III
powerfully argued by David Hume (), whose views about cau-

sation remain enormously inuential today. Hume wrote:
The mind can never possibly nd the eect in the supposed

cause, by the most accurate scrutiny and examination. For the
eect is totally dierent from the cause, and consequently can
never be discovered in it. Motion in the second Billiard-ball is a
quite distinct event from motion in the rst; nor is there any-
thing in the one to suggest the smallest hint of the other . . .
[E]very eect is a distinct event from its cause. It could not,
therefore, be discovered in the cause.
Humes basic point is that if the eect were contained in the cause, then
it would be possible for us to nd or discern the eect by carefully exam-
ining the cause. But we cannot do this; rather, we must wait for experi-
ence to show us what eect will follow from any given cause.
In making this point, Hume explicitly refers to the cause and the ef-
fect as being each an event. This reects an important insight of Humes,
namely, that the true members of a cause-eect relationship are events,
rather than objects. Often, our ordinary speech masks this fact. For ex-
ample, we say that the rock broke the window. Here it almost sounds
as if the cause is one object (the rock), and the eect another (the bro-
ken window). But, of course, what really happened is that the rocks hit-
ting the window caused the windows breaking. Now the rocks hitting the
window and the windows breaking are not objects or things; they are
events or occurrences. Once we understand this point, the idea of the
causes containing the eect immediately looks suspect. For it makes
little if any sense to say that the rocks hitting the window contained
the windows breaking: it certainly did not do so in the literal sense in
which, for example, a box of chocolates contains the chocolates or the
chocolates contain their caramel llings. Of course, there are cases of
causality that seem to t the preformationist assumption much better,
such as the case of conception and birth. But in such a case, it is true
only in a general, rough sense that the mother caused the baby. No
biologist studying the process of reproduction would describe what hap-
pened in such an inexact way. Rather, what really happened is that a

David Hume, An Enquiry Concerning Human Understanding, sec. , part , para.
and .
Descartes
complex sequence of events involving the mother caused the event of

the babys birth. And again, it makes little if any sense to say that each
of those events contained their eects. It would seem, then, that if
Descartess precontainment premise is interpreted to mean that the
cause must literally precontain its eect, then the points made by Hume
show the premise to be erroneous.
In all probability, however, Descartess principle should not be inter-
preted so literally; for after all, his subarguments premise says that the
cause must precontain the reality of the eect, not that it must precon-
tain the eect itself. So perhaps his premise means that in cases where
the cause is a change in or an act of some object(s) (loosely called the
cause) and the eect is the coming into existence of another object(s)
(loosely called the eect), the latter object(s) cannot possess any prop-
erty that was not also possessed by the former. John Cottingham, in his
Descartes, interprets Descartess premise along these lines, calling it the
heirloom view of causation:
To say that whatever produces a stone must itself have all the
features found in the stone seems to imply a kind of heirloom
view of causationthat the only way an eect can have come to
possess some property is by inheriting it, heirloom fashion, from
its causes.
As Cottingham goes on to argue, however, the heirloom principle is

highly questionable, for it seems vulnerable to counterexamples. Thus,
the author(s) of the Second set of Objections to the Meditations wrote to
Descartes:
You say ... that an eect cannot possess any degree of reality or
perfection that was not previously contained in the cause. But we
see that ies and other animals, and also plants, are produced
from sun and rain and earth, which lack life... . [H]ence it does
happen that an eect may derive from its cause some reality which
is nevertheless not present in the cause. (CSM II , AT VII )
To this Descartes replied that if animals have perfections that are not
present in the sun, rain and earth, then this only shows that the sun, rain

Cottingham, Descartes, pp. .
Meditation III
and earth are not the total causes of animal life (CSM II , AT VII ).
As Cottingham points out, however, this reply commits Descartes to the
view that there can never be genuinely emergent propertiesthat is,
properties that were never possessed, in any previous state of the world,
by the things whose operations caused those properties to occur. But this
conicts with the view, held by evolutionary biologists, that conscious-
ness evolved from nonconscious forces and elements. Of course, even
today, there are people who deny that consciousness evolved from inani-
mate elements. But the damaging point, so far as Descartess argument
is concerned, is that he could not even allow that there is a legitimate
controversy here; for according to him, the precontainment principle is
supposed to be obvious to any rational mind, quite apart from any empir-
ical evidence provided by sciences like chemistry and biology.
Perhaps, however, the precontainment principle can be interpreted
even more charitably. In his book The Miracle of Theism, John Mackie
writes:
Though we ordinarily admit that great eects can be brought

about by very small causes, these can only be partial causes, not
the whole cause of the great eects. Large trees can grow from
small seeds, but only by taking in a lot of nourishment as they
grow. Trivial accidents can precipitate revolutions, but only
where there are great repressed forces waiting to be triggered or
released. And so on. We commonly assume that there are conser-
vation principlesthe conservation of mass, or of energy, or of
the sum of the two, or of momentumwhich operate as con-
straints on possible processes of causation or production or
growth. Descartess dictum that there must be as much reality in
the total cause as in the eect can be understood as an attempt
to capture the general form of which such specic conservation
principles are instantiations.
Mackies remarks are directly addressed to (), the causal adequacy prin-
ciple, rather than to (), the precontainment principle, from which Des-
cartes derives (). But they can be seen as oering a possible defense of
() itself, no less than of (). This defense would be that (), like (), is a
very general principle of conservation, sanctioned by science. The basic

John Mackie, The Miracle of Theism, p. .
Descartes
idea would that if the reality of an eect were not somehow preconta-
ined in the cause of that eect, this would violate a general conservation
principle.
It is not necessary for us to try to spell out this idea more carefully; for
if Descartess premise is interpreted as a general conservation principle,
then it cannot serve his purposes, since, as Mackie goes on to point out,
[Such a principle] is not known or knowable a priori. We have

no rational guarantee, apart from experience, and apart from
scientic theories developed from and conrmed by such experi-
ence, that it will hold ... It cannot do the work that Descartes
intends it to do in an absolutely secure rebuilding of human
knowledge, since it is itself supported only by a wide range of
interpreted observation, and its precise scope and implications
are uncertain.
The key point here is that the premises of Descartess argument must be
knowable without dependence on what could be learned only from ob-
serving the operations of the physical world; for the argument is sup-
posed to show that Gods existence can be known without any use of the
senses and even if the existence of the physical world is still in doubt. But
as Mackie indicates, () and/or (), construed as general conservation
principles, can be known only by widely based observations of the phys-
ical world. So they cannot, construed in that way, play the role demanded
of them by Descartess argument.
Finally, it can be argued that even if one were to assume knowledge
of the physical world, conservation principles of the kind Mackie has in
mind cannot legitimately be used to argue for Gods existence. For such
principles, being ultimately based on our observations of physical
things operations and interactions, concern what we might call the
internal structure of the physical universethe regular qualitative
and, especially, quantitativerelationships among dierent spatial
and temporal parts of the universe. But, as Immanuel Kant argued in
his Critique of Pure Reason (), we are not justied in assuming that
these principles therefore also provide reliable information about how
the universe as a whole is related to a source that is supposed to have
created it.

Mackie, The Miracle of Theism, pp. .
Meditation III
We have considered three possible interpretations of Descartess pre-

containment principle. It can be interpreted as a statement of what
Lovejoy calls the preformationist assumption about causality, as a
statement of what Cottingham calls the heirloom view of causation,
and as a statement of what Mackie calls the general form of specic
conservation principles. We have seen that on the rst two interpreta-
tions, the principle is highly questionable and that on the third interpre-
tation, it cannot serve Descartess purposes. This result, of course, does
not show that the conclusion of the subargument, namely () (which is
the core arguments rst premise, and which it is the whole purpose of
the subargument to justify), is false. But it calls into serious question
Descartess attempt to justify or support that premise by appealing to
the precontainment principle.
Before concluding this section, therefore, we should note that in one
place, Descartes tries to establish his premise without relying on the
precontainment principle. In his Reply to the Second set of Objections,
Descartes says:
The fact that there is nothing in the eect which was not previ-
ously present in the cause, either in a similar or in a higher form is
a primary notion which is as clear as any that we have; it is just the
same as the common notion that Nothing comes from nothing.
For if we admit that there is something in the eect that was not
previously present in the cause, we shall also have to admit that
this something was produced by nothing. (CSM II , AT VII )
Here, Descartes seems to be defending a key premise needed to secure

(), namely (), the Causal Adequacy Principle, which says that a cause
must contain as much reality as its eect, without relying on (), the pre-
containment principle. The rst three steps of his subargument would
then be these:
() There are degrees of reality.

() Something cannot come from nothing.
() A cause must contain at least as much reality as its eect
(from steps and ).
In the passage just cited, Descartes can be read as asserting that () fol-
lows from () and (), because if an eect did contain more reality than
Descartes
its cause, then it would have to get some of its reality from nothing,
which violates (). Since this way of supporting () makes no use at all
of the precontainment principle, but only of the degrees of reality pre-
mise and the nothing-comes-from-nothing principle, it might seem to
show that Descartess subargument can proceed without using the pre-
containment principle. It must be said that in reading Descartes this
way, we are bending over backward to be charitable; for in the very next
sentence after the quoted passage, he defends ()the nothing-comes
from nothing principleby appealing to back to the precontainment
principle:
And the reason why nothing cannot be the cause of a thing is

simply that such a cause would not contain the same features as
are found in the eect. (CSM II , AT VII )
But this appeal to the precontainment principle is quite unnecessary, for

there is a much better reason why nothing cannot be a cause of a thing:
nothingbeing just nothing, or nothingness, or nonbeingobviously
cannot cause anything. So despite Descartess remark, this version of
his subargument could proceed without relying on the precontainment
principle.
Unfortunately, however, this alternative version of the subargument
suers from a fatal aw; for, as James Van Cleve has shown, it subtly
begs the question. For why should one accept Descartess claim that if
an eect e contains more reality than its cause c, then the surplus re-
ality must have come from nothing? Why not maintain, instead, that this
surplus reality all comes from the lesser cause, c? Descartes would no
doubt reply that the lesser cause c cannot be the cause of the greater ef-
fect ethat c spent itself in producing the nonsurplus part of es re-
ality. But this assumes exactly the point to be proved, namely, that a
cause must contain at least as much reality as its eect.
To conclude, it seems clear, in light of Van Cleves criticism, that the
alternative version of the subargument Descartes oers in his Reply to
the Second Set of Objections is unsuccessful. Descartes must therefore
rely on the version he gives in Meditation III, which is built on the highly
problematic precontainment principle.

James Van Cleve, in personal correspondence and in On a Little-noticed Fallacy in
Descartes, manuscript.
Meditation III
.. Degrees of Reality
A second diculty in the subargument concerns premise , that there are
degrees of reality. Thomas Hobbes, author of the third Set of Objections,
crisply challenged this premise:
Moreover, M. Descartes should consider afresh what more re-

ality means. Does reality admit of more and less? Or does he
think one thing can be more of a thing than another? If so, he
should consider how this can be explained to us with that degree
of clarity that every demonstration calls for, and which he has
himself employed elsewhere. (CSM II , AT VII )
Descartess reply was this:
I have also made it quite clear how reality admits of more and
less. A substance is more of a thing than a mode; if there are real
qualities or incomplete substances, they are things to a greater
extent than modes, but to a lesser extent than complete sub-
stances; and, nally, if there is an innite and independent
substance, it is more of a thing than a nite and dependent sub-
stance. All this is completely self-evident. (CSM II , AT VII
)
Here, Descartes seems to complicate his three-level hierarchy (innite

substance, nite substance, and modes) by adding an intermediate level
between nite substances and modes: real qualities or incomplete sub-
stances. In his Reply to the fourth Set of Objections, he gives an example
of an incomplete substance: a hand (CSM II , AT VII ). It is a sub-
stance because it is a material thing like a rock or a stick but an incom-
plete one because its function cannot be understood apart from the
whole body of which it forms a part. As Bernard Williams points out,
however, these points are basically unassimilated relics in Descartess
metaphysics of Aristotelian views that play no role within Descartess
own scheme. And as Anthony Kenny points out, Descartes himself else-
where strongly rejects the theory of real accidents. So the real import
of the above passage is that Descartes is reasserting, as completely

Williams, Descartes: The Project of Pure Enquiry, p. .

Anthony Kenny, Descartes, p. .
Descartes
self-evident, his three-level hierarchy. In the same vein, in his Reply to

the second set of Objections (where he complies with a request that he
present his whole system in the geometric format of denitions, postu-
lates, axioms, propositions, and corollaries), he treats it as an axiom (a
self-evident starting point) that
There are various degrees of reality or being: a substance has

more reality than a mode; an innite substance has more reality
than a nite substance. Hence there is more objective reality in
the idea of a substance than in the idea of a mode; and there is
more objective reality in the idea of an innite substance than in
the idea of a nite substance. (CSM II , AT VII )
Descartes, then, seems to have considered his theory of degrees of re-

ality so obvious as not to require any defense or explanation. Yet one can
certainly question whether the basis of his hierarchy is clear. The stan-
dard reading of the hierarchy is that it turns on the idea of dependence
and independence: if X can exist independently of Y but not vice-versa,
then X has more (formal) reality than Y does. It can then be suggested
that since nite substances depend on innite substance, or God, for
their existence but God does not depend on them (or on anything else)
for his existence, innite substance has more reality than nite sub-
stances and that since modes or properties cannot be free-oating but
must exist in a substance, nite substances have more reality than modes
or properties. This suggestion works well for the case of God and nite
substances: the latter depend causally on God for their existence, but not
vice-versa. Indeed, Descartes holds that nite substances depend caus-
ally on God in a very radical way: God not only creates them, but also
preserves them in existence from moment to moment, or continually
recreates them; for Descartes thinks that it is evident by the natural
light ... to anyone who attentively considers the nature of time that the
same power and action are needed to preserve anything at each indi-
vidual moment as would be required to create that thing anew if it were
not yet in existence (CSM II , SPW , AT VII ). But the suggestion
does not work for nite substances and modes or properties. To see why,
recall the account of the substance theory in chapter , section .,
according to which the minimum that can exist on its own is a substance
plus a property: there can be no property without a substance, and there
can be no substance without a property. Thus, there is a two-way, rather
Meditation III
than (as is sometimes thought) only a one-way, relation of dependence

between substance and property. As Bishop Berkeley (), the
great Irish philosopher, put it, It seems no less absurd to suppose a sub-
stance without accidents, than it is to suppose accidents without a sub-
stance. It remains unclear, then, why properties are supposed to have
less reality than substances. Is it perhaps because a substance can un-
dergo change without losing its identity, whereas a property cannot? Per-
haps, but then Descartess hierarchy is based on two dierent notions of
degrees of reality, rather than only one. In any case, his theory of
degrees of reality can hardly be said to be clear and distinct, or to be as
unquestionable as the inference from I am thinking to I exist.
.. Justifying the Causal Maxim

A third problem with the subargument is that Descartess justication
for (), the causal maxim that everything must have a cause, is faulty. As
we saw, Descartes evidently believed that () follows logically from (),
the nothing-comes-from-nothing principle, that something cannot come
from nothing. Indeed, he may have thought, as did many philosophers
before him, that () and () are logically equivalent, that is, that each
statement follows logically from the other, so that they make the same
assertion in two dierent ways. But there is a hidden mistake here. To
expose this mistake, we need to see that () is ambiguous. It can mean
(a) Nothing (= nothingness or nonbeing) cannot be a cause.
Or it can mean
(b) Something cannot exist without a cause.

George Berkeley, A Treatise Concerning the Principles of Human Knowledge, Part I,
section .

Ms. Chelsie Mack, a student at the College at Brockport, SUNY, wrote an insightful
remark that may blunt the force of this criticism: Finite substances such as the Medita-
tor, or the lump of wax the Meditator mused about in [Meditation II], are more real than
modes or properties in the sense that nite substances have a degree of independence
from their modes and properties. For instance, the waxs shape, size, temperature and
consistency (i.e., its properties) changed, but it continued to be the same piece of wax.
Similarly, the Meditator has dierent thoughts and ideas, but remains the same thinking
thing.
Descartes
The rst philosopher to notice the ambiguity of () and to distinguish

between (a) and (b) was David Hume, in a famous section of his A Trea-
tise of Human Nature (). Hume went on to point out, in eect, that
even if (a) is self-evident, this does not mean that (b) is self-evident
and to argue that (b) is not self-evident. Furthermore, since () follows
only fromand is indeed equivalent to(b) and does not follow from
(a), Hume concluded that () is not self-evident either.
Humes distinction raises a problem for Descartess subargument.
Suppose that () means (a). Then () follows from (), because

__________________________________________
? (a) Nothing cannot be a cause.
is a valid argument, since nothing (nothingness, nonbeing) cannot possibly

precontain anything. However, () does not then follow from (), because
(a) Nothing cannot be a cause.

______________________________
? () Everything must have a cause.
is an invalid argument, since the fact that nothing(ness) cannot serve as

a cause does not prove that a thing could not just spring into existence
without any cause whatsoever, including nothing(ness) or nonbeing.
On the other hand, suppose that () means (b). Then () does follow
from (), because
(b) Something cannot exist without any cause.

_______________________________________
? () Everything must have a cause.
is valid (indeed, statements b and are logically equivalent). However,

() then does not follow from (), because

__________________________________________
? (b) Something cannot exist without any cause.

David Hume, A Treatise of Human Nature, book I, Part , sec. .
Meditation III
Figure .
is an invalid argument. It is invalid because, as Bernard Williams has

emphasized, () only covers cases where we admit that there is a cause; it
says nothing at all about the possibility of something existing without
any cause.
Using the statement numbers already assigned to represent the state-
ments themselves, and an arrow to represent therefore, we can summa-
rize the diculty diagrammatically as shown in Figure .. In brief:
Descartes can go validly from () to () or from () to (), but he cannot
go validly from () to () to ().
To conclude this section, we shall suggest two possible solutions to
the diculty just discussed. Each of these solutions is designed to put
(), which it is the whole purpose of the subargument to justify, in the
strongest possible light. The rst possible solution would be to treat
()and also (b), since these are equivalentas a basic premise or
axiom. Using numbers and arrows the same way as in Figure . above,
and using + to show that the statements linked by the + are intended
to support the statement below them not individually but jointly, the
entire subarguments structure could then be diagrammed as shown in
Figure ..
The advantage of this solution is that the subargument would then
not rely on the awed method of justifying () just criticized. On the
other hand, the argument would now rely on four basic premises (i.e.,
premises that are not themselves supported by arguments but used as


The method of diagramming arguments used here is explained, among other places,
in Stephen Thomas, Practical Reasoning in Natural Language, pp. and pp. .
Descartes
Figure .
starting points of the argument): (), (), () = (b), and (). If the argu-
ment is to provide the absolute certainty Descartes sought, then each of
those four premises must be absolutely certain. But we have seen that
premise , far from being absolutely certain, is highly problematic. Pre-
mise involves the rather fuzzy notion of degrees of reality. Premise ,
which says that the cause of an ideas objective or representational re-
ality must (ultimately) be the formal or nonrepresentational reality of
something, is somewhat dicult to grasp, though it does seem plau-
sible on reection, as the analogy of mirror images given in the pre-
vious section may have brought out. As for (), its indubitability was
famously challenged by Hume on the grounds that one can conceive of
something springing into existence without any cause, that is, of an
uncaused beginning of existence. E. M. Curley responds to Humes chal-
lenge this way:
Admittedly I can conceive of something springing into existence

ex nihilo. But I cannot believe that this ever happens.
This may well be a persuasive defense of (): Can you seriously believe
that something has or will ever come into existence, with no cause or
explanation whatsoever?
Perhaps, then, it would not be unreasonable to take () and () as basic
premises; and perhaps () can pass muster, as well, despite its unclarity.
But one can surely wonder whether those premises are absolutely cer-
tain. Furthermore, the problematic status of (), the precontainment
principle, remains a major drawback for this solution.

Curley, Descartes Against the Skeptics, p. .
Meditation III
The other possible solution would be to scrap the subargument and to

take () as itself a basic premise. Then the entire argument, with its steps
renumbered to reect its shorter length, would reduce to this:
. The cause of an idea must have as much formal reality as the idea
contains objective reality (basic premise).
tains objective reality (basic premise).
. The cause of my idea of God is a perfect God (from and ).
. A perfect God exists (from ).
. To deceive is an imperfection (basic premise).
. God is not a deceiver (from and ).
. If my clear and distinct perceptions could be false, then God would be
a deceiver (premise, based on the assent-compellingness of clear and
distinct perceptions).
. My clear and distinct perceptions cannot be false, i.e., whatever I per-
ceive clearly and distinctly must be true (from and ).
To see the strength of this approach, it is worth quoting once again a pas-
sage from Cottinghams Descartes, including this time an important ob-
servation that Cottingham makes at the very beginning of the passage:
[T]he Causal Adequacy Principle actually seems on rmer ground

when applied to ideas than when applied to the physical world
[emphasis in the original]. Recall Descartes example of the
highly intricate machine; and for the sake of simplicity let us
follow Descartes own comparison between ideas and pictures,
and consider the case of a drawing or a picture rather than an idea.
Suppose a ve-year-old child produces a highly complicated design
for a computera design which we know could only be produced
by a highly skilled mathematician with a mental age vastly supe-
rior to the childs. The fact that the design is only a drawing and
not an actual computer does not block the causal question: the
representative or objective intricacy of the design still has to be
accounted for. Of course, the child might simply have copied down
the drawing from a book. But this simply pushes the argument
one stage further back. We are, it seems, justied in asserting that
somewhere along the line of causation there must be an actual
entity or being that really does possess sucient complexity to
account for the complexities which are to be found in the design.
Descartes
And what goes for the drawing goes equally for an idea: complex
representational content requires a complex cause.
What conclusion should we draw from our exploration of the grounds

for the core arguments rst premise? Perhaps the reasonable conclusion
is that the premise is plausible, especially if it is taken as a basic premise
rather than defended by the subargument Descartes himself gives for it.
But one may still wonder whether the premise is as unshakably certain as
Descartes took it to be.
4.2 The Problem of the Cartesian Circle

In this section, we turn to one of the most controversial issues in Des-
cartes scholarship: the problem of the Cartesian Circle. This problem is
not an objection to Descartess specic argument for Gods existence in
the Third Meditation; it is an objection to his overall strategy in that Med-
itation. Remember what his purpose was: he wanted to establish or vindi-
cate his principle that whatever I perceive clearly and distinctly is true;
that is, to show that clear and distinct perceptions are trustworthy. For
on the one hand, he found them to be assent-compelling, but on the
other hand, he found that whenever he thought about an omnipotent
God, he had to admit that such a God could easily deceive him even about
his clearest and most distinct perceptions. To escape from this oscilla-
tion, he declared that he must nd out whether God exists and can be a
deceiver. So he gave an argument, meant to show that God exists and is
not a deceiver, and that clear and distinct perceptions are therefore trust-
worthy. But suppose we ask Descartes: what assures you that the pre-
mises of your argument are true and that its steps are valid? The only
possible answer for Descartes (and the one he in fact gave) is that he
clearly and distinctly perceives the premises to be true and the steps to be
valid. But this uses his criterion of truth in the very argument that is
supposed to establish that he can safely use it. He is assuming the very
thing he set out to prove, namely that his clear and distinct perceptions
are true. So the overall argument of Meditation III is viciously circular.
In Descartess own lifetime, this now-famous problem was raised in
both the Second and Fourth sets of Objections to the Meditations. The
clearest statement of the problem was the one in the fourth set, written
by Antoine Arnauld. Arnauld wrote:

Cottingham, Descartes, pp. .
Meditation III
I have one further worry, namely how the author avoids rea-
soning in a circle when he says that we are sure that what we
clearly and distinctly perceive is true only because God exists.
But we can be sure that God exists only because we clearly and
distinctly perceive this. Hence, before we can be sure that God ex-
ists, we ought to be able to be sure that whatever we perceive clearly
and evidently is true. (CSM II , SPW , M , AT VII )
Here Arnauld is saying that in order for Descartes to be in a position to

give a proof of Gods existence, he must already know that whatever he
perceives clearly and distinctly is true; for if Descartes does not already
know that he can trust his clear and distinct perceptions, then he cannot
be sure that the premises of his proof of Gods existence are true, or that
its steps are logically valid. Hence, Arnauld concludes, Descartes ought
not to have held that his criterion of truth requires a divine vindication.
He should have held that this criterion is known to be fully reliable even
before Gods existence is known. Arnauld does not mention another op-
tion, which would have been to hold, contrary to what Descartess text
seems to say, that Gods existence can be known before knowing that
whatever one perceives clearly and distinctly is true.
Recent Descartes scholars have formulated the problem in the fol-
lowing way. Descartes seems to hold both:
() I can know that whatever I clearly and distinctly perceive is true

only if I rst know that God exists and is no deceiver
and
() I can know that God exists and is no deceiver only if I rst know
that whatever I clearly and distinctly perceive is true.
Statement () is implied by what Descartes says is the purpose of the

Meditation III argument for Gods existence, namely, to prove that what-
ever I clearly and distinctly perceive is true. To see why Descartes seems also

This formulation was rst suggested by Willis Doney in The Cartesian Circle,
, and subsequently adopted by James Van Cleve in his masterful Foundational-
ism, Epistemic Principles, and the Cartesian Circle in . Both of these essays are
reprinted in Willis Doney ed., Eternal Truths and the Cartesian Circle: A Collection of Studies
(New York: Garland, ).
Descartes
to be committed to (), we need only ask: what entitles him to be sure

that the premises of his argument for Gods existence are true, and to be
sure that the inferences drawn from those premises are valid? The only
possible answer seems to be this: that he clearly and distinctly perceives
those premises to be true and those inferences to be valid. Furthermore,
the texts show that this answer is the one Descartes would give. In the
Third Meditation, he says that the premises of his argument for Gods
existencespecically, the Causal Adequacy Principle and the principle
that to deceive is an imperfectionare known by the natural light
(CSM II , ; SPW , ; AT VII , ). But in the Principles of Philos-
ophy, he says that
[T]he light of nature or faculty of knowledge which God gave us

can never encompass any object which is not true in so far as it is
indeed encompassed by this faculty, that is, in so far as it is clearly
and distinctly perceived. (CSM I , SPW , AT VIIIA )
This passage implies that the light of nature or natural light is the very
power or faculty of the mind by which clear and distinct perceptions are
obtained. Thus, when Descartes says that the premises of his theological
argument are known by the natural light, he is saying that they are
known by being clearly and distinctly perceived, and so, it seems, commit-
ting himself to (). But now, if () and () above are both true, then Des-
cartes can never know either that whatever he clearly and distinctly perceives
is true or that God exists and is no deceiver; for to know either of these prop-
ositions, he would have to know the other one rst: but he cannot know
them both rst! Yet, by the end of the Third Meditation, Descartes does
claim to know both propositions. The problem is, How is this possible,
given the evidence that Descartes is committed to both () and ()?
To solve this problem, one would have to show either that Descartes is
not really committed to (), or that he is not really committed to (). Two
dierent ways of doing this would be to show that appearances to the
contrary notwithstanding, Descartes really holds either that:
A. I can know that whatever I clearly and distinctly perceive is true

before knowing that God exists and is no deceiver

This analysis is from Van Cleve, Foundationalism, Epistemic Principles, and the
Cartesian Circle, p. .
Meditation III
or that
B. I can know that God exists and is no deceiver before knowing

that whatever I clearly and distinctly perceive is true.
Since strategy A says, in eect, that no divine vindication of the criterion

of truth is needed, let us call it the vindication-not-needed strategy.
Since strategy B says, in eect, that the criterion of truth is not needed
to prove Gods existence and veracity, let us call it the criterion-not-
needed strategy. In the next two subsections, we shall examine recent
defenses of each strategy.
.. The Restriction of the Doubt to Past

Clear and Distinct Perceptions Defense
This defense, which for brevitys sake we shall henceforth refer to as the
restriction defense, claims that Descartes doesnt really accept thesis (),
that I can know that whatever I clearly and distinctly perceive is true only if
I rst know that God exists and is no deceiver. For the proof of Gods exis-
tence and veracity is not required to guarantee that present clear and
distinct perceptions are true, but only in cases where I am relying on past,
remembered clear and distinct perceptions. In other words, what we may
call the divine guarantee is not needed to guarantee the proposition
that whatever I clearly and distinctly perceive is true, but only to guarantee
this proposition: Whatever I remember having clearly and distinctly
perceived is true. The restriction defense should not be confused with
the now-discredited, so-called memory defense, which turns on the idea
that memory is fallible but that God guarantees that memories of (only)
clear and distinct perceptions are infallible; for it does not turn on any
doubt about the reliability of memory, but only on the pastness of the
perceptions, which are assumed to be correctly remembered. The basic
assumption of the restriction defense is, as Gary Hateld puts it, that

The main proponent of the memory thesis was Willis Doney in The Cartesian Cir-
cle. Later, however, Doney advanced a dierent solution to the problem of the circle, in
his Descartess Conception of Perfect Knowledge, (also reprinted in his (ed.)
Eternal Truths and the Cartesian Circle). The most powerful critique of the memory de-
fense is Harry G. Frankfurt, Memory and the Cartesian Circle, . The memory
defense and the reasons for its failure were discussed in the rst edition of this book, but
will not be covered in this edition.
Descartes
the reliability of clear and distinct perception was never itself really
placed in doubt, merely our ability to remain convinced of its reliability
when we arent having such perceptions.
To see how the divine guarantee is supposed to work according to the
restriction defense, let us use Descartess own illustration in Meditation
V. Suppose I have worked through a proof that the three angles of a tri-
angle equal two right angles. Schematically, the proof looks as shown in
Figure ..
Figure .
So long as I hold the entire proof before my mind and clearly and dis-
tinctly perceive each premise and step, I do not need the divine guar-
antee, because I am not relying on past clear and distinct perceptions.
The two cases where I need it are () when I remember that I proved that
a + b + c = but I am no longer attending to the proof at all and so not
currently clearly and distinctly perceiving any part of it, and () when the
proof is too long or complicated for me to grasp all of it at once, so that
by the time I get to (say) step n, I am no longer clearly and distinctly per-
ceiving any of the earlier premises or steps. Since the restriction defense
denies ()that is, says that I can know that whatever I clearly and dis-
tinctly perceive is true before (without) knowing that God exists and is no
deceiverit conforms to the vindication-not-needed strategy.
There are several passages in Descartess works that suggest that he
means to restrict the doubt to only past clear and distinct perceptions. In
the passage from Meditation V where he oers the above illustration, he
says,

Hateld, Descartes and the Meditations, pp. .
Meditation III
Admittedly my nature is such that so long as I perceive some-

thing very clearly and distinctly I cannot but believe it to be
true. But my nature is also such that I cannot x my mental vi-
sion continually on the same thing, so as to keep perceiving it
clearly; and often the memory of a previously made judgment
may come back, when I am no longer attending to the argu-
ments which led me to make it. And so other arguments can
now occur to me which might easily undermine my opinion, if I
were unaware of God; and I should thus never have true and
certain knowledge about anything, but only shifting and ch-
angeable opinions. For example, when I consider the nature of a
triangle, it appears most evident to me, steeped as I am in the
principles of geometry, that its three angles are equal to two
right angles; and so long as I attend to the proof, I cannot but
believe this to be true. But as soon as I turn my minds eye away
from the proof, then in spite of still remembering that I perceived
it very clearly, I can easily fall into doubt about its truth, if I am
unaware of God. (CSM II , SPW , AT VII )
Now, however, I have perceived that God exists, and at the same
time I have understood that everything else depends on him, and
that he is no deceiver; and I have drawn the conclusion that
everything which I clearly and distinctly perceive is of necessity
true. Accordingly, even if I am no longer attending to the argu-
ments which led me to judge that this is true, as long as I remem-
ber that I clearly and distinctly perceived it, there are no
counter-arguments which can be adduced to make me doubt it,
but on the contrary I have true and certain knowledge of it. And
I have knowledge not just of this matter, but of all matters which
I remember ever having demonstrated, in geometry and so on.
Furthermore, when Descartes tries to refute the charge that his rea-
soning is circular, he seems again to resort to the restriction defense.
Thus, the author(s) of the Second set of Objections had said:
You are not yet [at the start of Meditation III] certain of the exis-
tence of God, and you say that you are not certain of anything,
and cannot know anything clearly and distinctly until you have
Descartes
achieved clear and certain knowledge of the existence of God. It

follows from this that you do not yet clearly and distinctly know
that you are a thinking thing, since, on your own admission, that
knowledge depends on the clear knowledge of an existing God;
and this you have not yet proved in the passage where you draw
the conclusion that you clearly know what you are. (CSM II ,
SPW , M , AT VII )
Descartess reply to this was:
When I said that we can know nothing for certain until we are
aware that God exists, I expressly declared that I was speaking only
of knowledge of those conclusions which can be recalled when we
are no longer attending to the arguments by means of which we
deduced them. (CSM II , SPW , M , AT VII )
Later, in replying to Arnaulds previously-quoted version of the circu-

larity charge, Descartes refers back to this response, saying:
Lastly, as to the fact that I was not guilty of circularity when I

said that the only reason we have for being sure that what we
clearly and distinctly perceive is true is the fact that God exists,
but that we are sure that God exists only because we perceive this
clearly: I have already given an adequate explanation of this point
in my reply to the Second Objections ... where I made a disti-
nction between what we in fact perceive clearly and what we
remember having perceived clearly on a previous occasion. To
begin with, we are sure that God exists because we attend to the
arguments which prove this; but subsequently it is enough for us
to remember that we perceived something clearly in order for us
to be certain that it is true. This would not be sucient if we did
not know that God exists and is not a deceiver. (CSM II , SPW
, M , AT VII )
Although the above passages suggest that Descartes adopted the re-
striction defense, that defense faces serious objections, both philoso-
phical and textual. The main philosophical objection is that it seems to
imply that every time one wishes to use the divine guarantee, one must
Meditation III
go through the proof of Gods existence and hold it before ones mind. To
see this, notice rst that on the restriction defense, the proof of Gods
existence must, of course, itself not rely on past clear and distinct percep-
tions. Rather, it must be possible to grasp the entire proof at once, instead
of having to rely on premises or steps that were clearly and distinctly per-
ceived only in the past; for if such premises or steps were required, then
one could never be certain of the conclusion, since all past clear and dis-
tinct perceptions are suspect until that conclusion has been established.
Now Descartes did, in fact, hold that his proof of Gods existence can,
after some practice, be grasped all at once, in a single mental intuition.
The diculty, however, concerns proofs that are too long to be grasped all
at once, such as lengthy mathematical proofs. It is precisely in those cases,
according to the restriction defense, that the divine guarantee is needed.
But how would one then use the guarantee? Would it be enough to appeal
to the fact that one previously proved Gods existence from clearly and
distinctly perceived premises and steps? No; for then one would be
defending the reliability of some past clear and distinct perceptions (e.g.,
those of the rst several steps of a mathematical proof) by means of an
appeal to other past clear and distinct perceptions (i.e., those of all the
steps of the theological argument)which is circular if the reliability of
past clear and distinct perceptions is in doubt. To use an example of Lud-
wig Wittgensteins (), it would be like trying to establish the
truth of a newspaper story by consulting another copy of the same news-
paper. So to use the divine guarantee, one must go through the proof of
Gods existence all over again. One must do this each time one uses the
guarantee, which is, at least, an awkward result. Furthermore, Descartes
himself denies that such repeated rehearsals of the proof are needed: in
the second passage quoted above from Meditation V, he says,
[So] long as I remember that I clearly perceived it [i.e., that God

exists and is no deceiver] ... I have true and certain knowledge of
it. And I have knowledge not just of this matter, but of all mat-
ters which I remember ever having demonstrated, in geometry
and so on. (CSM II , SPW , AT VII )
Here Descartes explicitly says that, to use the divine guarantee, one need
only remember having clearly and distinctly perceived (a proof of) Gods
existence. As we have just seen, this would be manifestly circular if the
Descartes
function of the divine guarantee were to insure the reliability past clear
and distinct perceptions.
The other diculty for the restriction defense is a textual one: it sim-
ply does not square with what Descartes says in the Meditations. Recall
the important passage, near the beginning of the Third Meditation, where
Descartes was oscillating between certainty about present clear and dis-
tinct perceptions, and doubt about the reliability of his cognitive fac-
ulties. There Descartes said, referring to things that are very simple and
straightforward in arithmetic or geometry, for example that two and
three added together make ve, and so on:
[T]he only reason for my later judgment that they were open to
doubt was that it occurred to me that perhaps some God could
have given me a nature such that I was deceived even in matters
which seemed most evident. And whenever my preconceived
belief in the supreme power of God comes to mind, I cannot but
admit that it would be easy for him, if he so desired, to bring it
about that I go wrong even in those matters which I think I see
utterly clearly with my minds eye. (CSM II , SPW , AT
VII )
Here Descartes does not even mention past clear and distinct percep-
tions. Rather, his point is that an omnipotent God could easily make him
go wrong even about the things that he currently perceives with the
utmost clarity and distinctness and cannot doubt while he is attending to
them. So, Descartes here carries his doubt much further than the restric-
tion defense allows. Of course it is true that at any particular time, only
past clear and distinct perceptions can be doubted, since present clear
and distinct perceptions are assent-compelling. But this does not solve
the problem; for it does not mean that at any time after having clearly
and distinctly perceived that God exists and is not a deceiver, I cannot
doubt that proposition. So the fact that I earlier clearly and distinctly
perceived it does not allow me to escape from the oscillation between
being certain of this proposition and doubting it (and all the ones that
depend on it).
The basic problem for the restriction defense can be put like this. To
say that past clear and distinct perceptions can be doubted is to say
that even though I clearly and distinctly perceived some proposition in
the past, that proposition may have been false. This is to grant that the
Meditation III
fact that a proposition is clearly and distinctly perceived does not entail
that the proposition is true; it is to open a logical gap between clarity
and distinctness and truth. But once that gap is opened, how can it be
closed? The restriction defenses only answer is to point to the fact that
we are incapable of doubting a present clear and distinct perception
to the assent-compellingness of present clear and distinct perceptions.
But this assent-compellingness is merely a psychological fact about us;
it does not in itself go any way toward showing that clear and distinct
perceptions must be true. Therefore, it cannot possibly close the gap:
once it is granted that a proposition that was clearly and distinctly per-
ceived in the past might be false, there is no way to prevent doubt
about that proposition from recurring whenever it is not being clearly
and distinctly perceived, regardless of how strongly we were convinced
of its truth at the moment that it was being clearly and distinctly
perceived.
.. The General Rule Defense

According to the criterion-not-needed-strategy, Descartes holds that his
criterion of truth does require vindication, and the purpose of his theo-
logical argument is to supply that vindication. However, the criterion of
truth is not necessary for the theological argument itself. Note, then,
that the point of the label criterion-not-needed is not that Descartes
doesnt need his criterion of truth to complete his overall program but
only that the criterion is not needed for his proof of Gods existence and
veracity.
The obvious question for the criterion-not-needed-strategy is, How
can Descartes legitimately appeal to the clarity and distinctness of the
premises and steps of his proof of Gods existence and veracity if he
admits that his criterion of truth does need a vindicationone that is
not available until after the proof is complete and a nondeceiving God is
known to exist? In other words, how can Descartes both appeal to the
clarity and distinctness of his proof of Gods existences premises and
steps and yet hold that the clarity and distinctness criterion is not needed
for that proof?
Let us examine an answer rst proposed by Anthony Kenny and sub-
sequently adopted, with minor modications, by James Van Cleve and by
Bernard Williams. This answer is that the divine guarantee that the proof
of Gods existence provides is needed only for knowledge of the general
Descartes
rule that whatever I perceive clearly and distinctly is true; it is not needed
for the individual, particular clear and distinct perceptions appealed to in
the proof (or for any other individual or particular clear and distinct per-
ceptions). Let us call this answer the general rule defense. Our thesis
in this subsection will be that although the general rule defense is inge-
nious, it is ultimately unsuccessful.
Notice that both the restriction defense and the general rule defense
hold (albeit for very dierent reasons) that the divine guarantee is not
required for presently occurring clear and distinct perceptionsthe re-
striction defense because the divine guarantee is needed only for past
clear and distinct perceptions, the general rule defense because it is ne-
eded only for the generalization that whatever I perceive clearly and dis-
tinctly is true. Thus, on both defenses, the clear and distinct perceptions
that I have when I go through the proof of Gods existence and veracity
are immune from all doubt: there is no circle.
What textual support is there for the general rule defense? Mainly
that Descartes never says that he can doubt particular clear and distinct
perceptions, such as his perceptions that + = , cogito ergo sum, or not-
(p and not-p). On the contrary, he steadfastly maintains that such percep-
tions are assent-compellingthat is, he cannot doubt them while they
actually occur. Thus, near the beginning of Meditation III, immediately
after admitting that an omnipotent God could easily make him err even
in those matters which I think I see utterly clearly with my minds eye,
he insists that he nevertheless cannot doubt such matters while actually
attending to them:
Yet when I turn to the things themselves which I think I perceive

clearly, I am so convinced by them that I spontaneously declare:
let whoever can do so deceive me, he will never bring it about
that I am nothing, so long as I continue to think that I am some-
thing; or make it true at some future time that I have never
existed, since it is now true that I exist; or even bring it about
that two and three added together are more or less than ve, or

The general rule defense was rst suggested by Kenny, Descartes, chapter , and
more perspicuously presented in part I of his The Cartesian Circle and the Eternal
Truths, . Van Cleve endorses it (with some modications and elaborations) in
his Foundationalism, Epistemic Principles, and the Cartesian Circle, pp. . See Van
Cleves n. for one indispensable clarication of Kennys position. Williams takes a sim-
ilar position in his Descartes: The Project of Pure Enquiry, p. .
Meditation III
anything of this kind in which I see a manifest contradiction.

In a passage in Meditation V where Descartes is referring to mathematical

truths whose proofs he clearly and distinctly perceives, he makes the
same point this way:
I have already amply demonstrated that everything of which I

am clearly aware is true. And even if I had not demonstrated this,
the nature of my mind is such that I cannot but assent to these
things, at least so long as I clearly perceive them. (CSM II ,
SPW , AT VII )
Notice that Descartes here explicitly says that his clear and distinct per-
ceptions would be assent-compelling, even if he had not established his
general rule that clear and distinct perceptions must all be true. A couple
of pages later, he puts the point very concisely:
[M]y nature is such that so long as I perceive something very

clearly and distinctly I cannot but believe it to be true. (CSM II
, SPW , AT VII )
Then he gives an example:
For example, when I consider the nature of a triangle, steeped as

I am in the principles of geometry, it appears most evident to me
that its three angles are equal to two right angles; and so long as
I attend to the proof, I cannot but believe this to be true. (CSM II
, SPW , AT VII )
Finally, there is an exceptionally strong statement of the point in Des-

cartess Reply to the Second Set of Objections. Referring to the clear per-
ceptions of the intellect, Descartes says:
Now some of these perceptions are so transparently clear and at

the same time so simple that we cannot ever think of them with-
out believing them to be true. The fact that I exist so long as I am
thinking, or that what is done cannot be undone, are examples of
truths in respect of which we manifestly possess this kind of
Descartes
certainty. For we cannot doubt them unless we think of them;

but we cannot think of them without at the same time believing
that they are true ... Hence we cannot doubt them without at
the same time believing that they are true; that is, we can never
doubt them. (CSM II , M , SPW , AT VII )
Here Descartes claims that some truths, such as the cogito and the law of
noncontradiction, are so simple in their content that one cannot even
think of them without clearly and distinctly perceiving them and therefore
being compelled to assent to them. So, since one cannot doubt a proposi-
tion without thinking of it, these truths can never be doubted at all.
What Descartes does admit, however, is that when he thinks about an
omnipotent God, he has to concede that such a God could make him go
wrong even about those things that seem most evident to him. Now the
key claim of the general rule defense is that this admission does not con-
tradict Descartess claim that particular clear and distinct perceptions
cannot be doubted while they occur; for according to the general rule
defense, Descartess admission has a specic meaning. It means that,
when Descartes thinks of Gods omnipotence, he can doubt the general-
ization that whatever he most clearly and distinctly perceives is true.
From this, it does not follow that he can doubt any particular clear and
distinct perception that he is having. From
I can doubt that for every proposition p, if I clearly and distinctly

perceive that p, then it is true that p
it does not follow that
There is a proposition p such that I clearly and distinctly perceive

that p, and I can doubt that p.
As Van Cleve puts it, [Descartes] might be uncertain of the general con-
nection between clear and distinct perception and truth, yet certain of
every proposition [he] ... clearly and distinctly perceive[s].
Anthony Kenny puts the matter this way. Before proving Gods exis-
tence and veracity, Descartes is prepared to admit

Van Cleve, Foundationalism, Epistemic Principles, and the Cartesian Circle, p. .
See also p. , above.
Meditation III
() For some proposition p, I clearly and distinctly perceive that

p, but maybe not-p.
However, Descartes is not prepared to admit any particular case of ().

For example, he is not prepared to admit
(a) I clearly and distinctly perceive that + is , but maybe +

is not
or
(b) I clearly and distinctly perceive that if I am thinking then I exist,

but maybe I am thinking yet I do not exist.
As Kenny points out, it is perfectly possible for a person to accept ()

without accepting any specic case of (). This is much like admitting that
some of ones beliefs are false without admitting (paradoxically) that any
particular one of them is false. Almost all people, unless they are ex-
tremely arrogant, would probably admit
() For some p, I believe p; but not-p.
But hardly anyone would admit any particular case of (), such as I
believe that it is raining, but it is not raining; for as the British philoso-
pher G. E. Moore () pointed out, this sentenceor any other
sentence of the form I believe that p, but not-pis highly paradoxical.
(Such sentences are all cases of what philosophers call Moores Par-
adox.) Likewise, then, Descartes can admit () without accepting para-
doxical statements like (a) and (b).
After proving Gods existence and veracity, however, Descartes no
longer admits (). For now he has established the opposite of (), namely
the general rule that
() For all p, if I clearly and distinctly perceive that p, then p is

true.

The material in this and the next two paragraphs closely follows Kenny, The Carte-
sian Circle and the Eternal Truths, pp. .
Descartes
Kenny also formulates the matter this way. Even before proving Gods
existence and veracity, the assent-compellingness of clear and distinct
perceptions means that the following statement is true of me:
() For all p, if I clearly and distinctly perceive that p, then I

cannot doubt that p.
On the other hand, because of the general doubt about the reliability of
clear and distinct perception based on the possibility of a deceiving God,
it is false (before proving Gods existence and veracity) to say
() I cannot doubt that: for all p, if I clearly and distinctly per-

ceive that p, then p is true.
But after the proof of a perfect God and the consequent removal of this
generalized doubt, () also becomes true of me.
Despite what Kenny says, however, you might think that Descartess
admission that before knowing Gods existence and veracity he can doubt
even those things that he most clearly and distinctly perceives must con-
tradict at least his assertion in the second set of Replies that some very
simple truths, like the cogito and the law of noncontradiction, can never
be doubted at all. On one interpretation of Descartess admission, it
would seem to contradict that assertion. If the admission is interpreted
to mean that I can doubt whether any of the propositions that I clearly
and distinctly perceive are true, then, given also that the cogito and the
law of noncontradiction are propositions that I clearly and distinctly per-
ceive, it seems to follow that I can doubt themwhich contradicts the
assertion that I can never doubt them. But there is another way to inter-
pret Descartess admission, on which it does not lead to contradiction.
The admission can be interpreted to mean that I can doubt whether all
the propositions that I clearly and distinctly perceive are true, that is, that
I can admit that some or most of them might be false. This certainly does
not mean that I can doubt the truth of every proposition that I clearly and
distinctly perceive: there may be some exceptional ones that I can never
doubt. So the admission does not contradict the claim that I cannot doubt
the cogito or the law of noncontradiction. Now according to the general
rule defense, what can be doubted before proving Gods existence and
veracity is just this general rule that all clearly and distinctly perceived
propositions are true; and this is quite compatible with saying that some
Meditation III
specic clearly and distinctly perceived propositions can never be doubted.

Thus, it appears that Descartess admission, if interpreted in accordance
with the general rule defense, is consistent even with his assertion that
some propositions can never be doubted.
Now that we have presented the general rule defense, it is time to
examine it critically. We shall raise two dierent diculties for this
defense, and argue that while it can deal with the rst diculty, it suc-
cumbs to the second one.
The rst diculty is that if knowledge of the general rule that whatever I
perceive clearly and distinctly is true is not needed for the proof of Gods
existence and veracity, then presumably it is not needed for any other argu-
ments whose premises and steps are clear and distinct, such as mathematical
proofs. So what is the use of the general rule? Can Descartes avoid circularity
only at the price of rendering the rule (and its divine vindication) useless?
An illuminating answer to this question has been oered by Bernard
Williams in his Descartes. Essentially, Williamss answer is that without
the general rule, one would be limited to momentary episodes of cer-
tainty: one would be certain of only those propositions that one was cur-
rently perceiving clearly and distinctly and so currently unable to doubt.
But as soon as one stopped attending to those propositions, one could
wonder whether what was previously perceived clearly and distinctly is
true. But this means that one would never really know or, Williams sug-
gests, even believe anything; for both knowledge and belief are on-going
states rather than momentary episodes. As Williams puts it, Descartes
nds that he has encountered a number of propositions which are

irresistible [i.e., believed if even thought of, like the cogito and the
law of noncontradiction]. These encounters, however, have not
given [Descartes] any knowledge... . There is even a sense, and an
important one, in which they have not given him any beliefs. The
perceptions of these propositions which have occurred so far, and
in which they revealed their irresistibility, do not satisfy what is
virtually a formal requirement on knowledge or (in a full sense)
belief, that it should be an on-going dispositional state... . The
clear and distinct perceptions ... which [Descartes] has had are all
time-bound, in the sense that he may at one time clearly and dis-
tinctly perceive that p is true, but not at another.

Descartes
Williams goes on to say that, faced with the prospect of having only such
momentary episodes of certainty, one has three choices: () just give
up the search for stable, certain beliefs, () freeze ones attention per-
manently on just one clear and distinct perception, or () adopt some
acceptance-rule that would, as it were, promote momentary episodes
of certainty into full-edged beliefs. He then interprets Descartes as
taking the third option, as admitt[ing] some acceptance-rule for beliefs
which are on-going and not time-bound as the clear and distinct percep-
tions are. The rule that Descartes adopts, he says, is Accept as on-going
beliefs just those propositions which are at any time clearly and distinctly
perceived to be true.
In his Descartes, John Cottingham adopts a view very similar to Wil-
liamss. Before quoting a passage from Cottinghams book, however, we
should note that he interprets Descartess claim that present clear and
distinct perceptions cannot be doubted dierently than we have done.
He takes it to mean not merely that present clear and distinct percep-
tions are assent-compelling (i.e., cannot be doubted while they occur),
but also that they are self-guaranteeing (i.e., known to be true just by
being had). There is, indeed, a deep issue at stake in this dierence of in-
terpretation, to which we shall come shortly. But the point to note for
now is that having made the claim that clear and distinct perceptions are
self-guaranteeing, Cottingham faces, all the more urgently, the question,
What is the use of the divine guarantee? As he puts it, How do we con-
strue Descartes frequent assertions that God is the source and guarantor
of all knowledge? The answer, he says,
lies in the very temporary nature of the self-guaranteeing ashes of

intuition which the meditator enjoys. The guaranteed recognition
of truth lasts, for a given proposition, only so long as the meditator
holds that proposition in front of his mind; as soon as his attention
wanders, even for a moment, the guarantee vanishes ... Once Gods
existence is established, however, then we have the possibility of
progressing beyond such isolated ashes of cognition and building
up a systematic body of knowledge. As Descartes puts it, we can
move from cognitio (mere cognition) to scienta (stable knowledge).


Meditation III
This is essentially the same way of accounting for the use of the general
rule that Williams proposes.
Let us turn to the second diculty for the general rule defense. This
diculty concerns the most crucial issue for that defense: Does it re-
ally provide a way of showing that I can know that God exists and is no
deceiver before knowing that whatever I clearly and distinctly perceive
is true, thereby solving the problem of the circle? Despite the inge-
nuity of the general rule defense, we may question whether it really
does solve the problem. The proponents of this defense would presum-
ably say (as, in fact, Kenny does say) that before Gods existence and
veracity are proved, the general principle that whatever I perceive
clearly and distinctly is true can be doubted. Now what is it to doubt
this principle? It is not to doubt an isolated and perhaps insignicant
proposition. Rather, doubting the general principle must consist in
thinking that even when I am having a clear and distinct perception,
which admittedly I cannot doubt while I am having it, I may neverthe-
less be mistaken: the proposition that I clearly and distinctly perceive
may actually be false. Or, to put it in a way that uses verb tenses so as
to highlight the fact that I cannot doubt a proposition at the time I am
clearly and distinctly perceiving it, doubting the general principle must
consist in thinking, Even when I was having a clear and distinct per-
ception, which admittedly I could not doubt at the time I was having
it, I may nevertheless have been mistaken: the proposition that I was
then clearly and distinctly perceiving may actually have been false.
There seems to be no other way to understand what doubting the prin-
ciple that whatever I perceive clearly and distinctly is true amounts to.
Notice, then, that the general rule defense is incompatible with the
view that clear and distinct perceptions are not only assent-compel-
ling, but also self-guaranteeing. For the doubt of the general rule to
have any content, it must at least allow for doubt of past clear and
distinct perceptions and so for the possibility that a proposition that
was clearly and distinctly perceived was nevertheless false and, there-
fore, not known to be true. To put it dierently, the general rule
defense must allow a doubt concerning the faculty for clear and dis-
tinct perceptionfor a doubt concerning the very reliability of human
reason.
Now according to the general rule defense, Descartes purports to
remove this doubt by giving a proof of the general rule itself, turning on
Descartes
Gods existence and veracity. But if, prior to knowing Gods existence and
veracity, I must admit that I could be mistaken even about my clearest
and most distinct perceptions, then how can such a proof provide any
lasting assurance? Admittedly, I will be unable to doubt the proof so long
as I attend to it. If I can grasp the entire proof at once, including the step
leading to the nal conclusion that whatever I perceive clearly and dis-
tinctly is true, then there will be a time or times when I cannot doubt that
general principle. But the moment I turn away from the proof, the doubt
can recur. I can then say to myself:
I remember constructing an argument showing that whatever I

perceive clearly and distinctly is true, using certain premises and
making certain logical steps that I perceived most clearly and dis-
tinctly and could not doubt at the time. One of those premises
was even the premise that God is not a deceiver, from which I
inferred that he would not allow my clear and distinct percep-
tions to be false. But when I stop attending to this argument and
think instead of the supreme power of God, I must admit that he
could easily cause me to go wrong even about those things that I
perceive utterly clearly, including of course the premises and the
steps of the very argument which was supposed to show that this
is not so. So, I cannot be certain even now that there is a God
who would not deceive me, or that whatever I clearly and dis-
tinctly perceive is true.
Thus, it seems that Descartes cannot emerge from the oscillation between
his certainty about presently occurring clear and distinct perceptions,
and his general doubt about the reliability of even his clearest and most
distinct perceptions, from which his theological argument was supposed
to free him.
Consider an analogy. Suppose there were a drug that works as follows.
While its eect lasts, various propositions are indubitable for the drug-
taker: he cannot possibly doubt them. As long as the drugs eect lasts,
these propositions are assent-compelling. Furthermore, one of the prop-
ositions which is thus indubitable while the drug is working is the prop-
osition that whatever I perceive while under the drugs inuence is true.
So while I am under the drugs inuence, I cannot doubt that perceiving
a proposition while under the drugs inuence guarantees truth: I am
certain that it does. Now, how could this temporary certainty permanently
Meditation III
remove any antecedent doubts I might have had about whether being
under the drugs inuence guarantees truth? And why should it prevent
me from doubting this after the drugs eect has worn o ? You may want
to complain that its unfair to compare clear and distinct perception to a
drug; for the drug could well be a cause of error, illusion or delusion,
whereas the clear and distinct perceptions of the intellect are our best or
most reliable way of accessing truth. But this complaint would miss the
point of the analogy. This is that if we assume that the most that can be
said in favor of clear and distinct perceptions (before proving Gods exis-
tence) is that they are assent-compelling, then we are by that very as-
sumption granting that their deliverances might be no better than the
eects of a drug. So these deliverances, no matter how assent-compelling
they may be when clear and distinct perceptions are occurring, cannot
provide any guarantee that clear and distinct perceptions are true, even
if one of the deliverances of clear and distinct perception is that clear and
distinct perceptions themselves are true.
Of course, if Descartes is right to hold, in the Second set of Replies,
that some propositions are so simple that they can never be thought of
without being grasped and so obvious that they cannot be grasped with-
out being believed, then there will be a few, exceptional propositions
(e.g., the cogito and the law of noncontradiction) that are completely in-
dubitable. But these propositions do not include the propositions God
exists and is no deceiver and Whatever I clearly and distinctly perceive
is true. So Descartes will be unable to emerge from his vacillation about
the truth of all things beyond the few exceptional propositions.
Indeed one philosopher, Harry Frankfurt, has in eect suggested that
without the divine guarantee of clear and distinct perceptions, Descartes
would not claim to know that even those few exceptional propositions
are true. Frankfurt maintains that the most basic problem of the Medita-
tions is the relationship between indubitability and truth. So, he insists
that when Descartes claims that a proposition is indubitable, this must
never be confused with his claiming that the proposition is true. On
Frankfurts reading, then, we are to see Descartes as countenancing the
thought There are some propositions which I can never doubt, but which

See Harry G. Frankfurt, Descartes Validation of Reason, (reprinted in
Doney, ed., Descartes: A Collection of Critical Essays (Garden City, Doubleday/Anchor
Books, ), idem, Demons, Dreamers, and Madmen (New York: Bobbs-Merill, ), pp.
.
Descartes
may nevertheless be false. As we shall see in the next subsection, there

is indeed a case for saying that Descartes would countenance such a
thought. But one may question whether the thought is a coherent one,
for does it not amount to saying that there are some propositions that I
both can and cannot doubt? Be that as it may, the problem that we have
described also concerns the relationship between indubitability and
truth; for the problem is, basically, that although some propositions
cannot be doubted at certain times (while one is clearly and distinctly
perceiving them or their derivation from other currently clearly and dis-
tinctly perceived propositions), they can be doubted at other times, and
so admitted, at least at those times, to be possibly false. This problem
remains even if one holds, against Frankfurt, that Descartess claim that
a few exceptional propositions can never be doubted conicts with saying
that those propositions might be false.
.. The Radical Doubt of Reason and the Creation

of the Eternal Truths
Our critique of the general rule defense has brought us to the heart of the
problem of the Cartesian Circle. The goal that Descartes sets for himself
at the beginning of the Third Meditation is, at bottom, to overcome a doubt
about the reliability of human reason (whether it be called the natural
light, the power of clear and distinct perception, or whatever) by means
of an argument or chain of reasoning. To borrow Harry Frankfurts
phrase, Descartess goal is to provide a validation of reason. Now, to
overcome a doubt about the reliability of reason by means of a chain of
reasoning, one must obviously use reason. But if the reliability of reason
is itself doubtful, then how can the results of this (or any other) use of
reason be trusted? Isnt Descartess project of overcoming a doubt about
reason by using reason doomed from the start? As David Hume put it:
There is a species of skepticism, antecedent to all study and phi-

losophy, which is much inculcated by Des Cartes and others, as a
sovereign preservative against error and precipitate judgement.
It recommends an universal doubt, not only of all our former
opinions and principles, but also of our very faculties; of whose

Frankfurt, Dreamers, Demons, and Madmen, chap. ; idem, Descartes Validation
of Reason.
Meditation III
veracity, say they, we must assure ourselves, by a chain of rea-

soning, deduced from some original principle, which cannot pos-
sibly be fallacious or deceitful. But neither is there any such
original principle, which has a prerogative above others, that are
self-evident and convincing: or if there were, could we advance a
step beyond it, but by the use of those very faculties, of which we
are supposed to be already dident. The Cartesian doubt, there-
fore, were it ever possible to be attained by any human creature
(as it plainly is not) would be entirely incurable; and no reasoning
could ever bring us to a state of assurance and conviction upon
any subject.
The twentieth-century Oxford philosopher, H. A. Prichard, put the same

point this way:
Descartes idea of how the doubt has to be allayed, if at all, is

obviously mistaken; and we can see that it is mistaken without
even considering the actual way in which Descartes considers
that he allays it, viz. by discovering what he considers to be a
proof that there exists a deity who is no deceiver. For as Des-
cartes is representing the matter, such a proof could only consist
in perceiving clearly and distinctly that certain things which he
perceives clearly involve a necessity that such a God exists; and
as precisely what he is doubtful of is whether even when he per-
ceives something clearly and distinctly he is not mistaken, he will
become doubtful, when he reects on his having this proof,
whether here too, he is not mistaken... . Descartes, therefore, it
is obvious, is setting himself an impossible task; and we can
discover it to be impossible without considering his attempt to
execute it.
In light of this powerful objection, it may be tempting to regard Des-

cartess assertions that an omnipotent God could deceive him about even
his clear and distinct perceptions as aberrations, that is, as temporary,
ill-considered departures from his own main line of thought. It may be
tempting to say, with Cottingham, that for Descartes clear and distinct

Hume, An Enquiry Concerning Human Understanding, Section , part I, para. .

H. A. Prichard, Descartess Meditations, pp. .
Descartes
perceptions are self-guaranteeing (known to be true just by being had).

The Cartesian doubt would then be restricted, as it was in the rst and
second Meditations, to sense-perception, leaving the clear and distinct
perceptions of reason as safe resources upon which to rebuild ones
knowledge.
Although it might be reasonable to contend that this is the course
Descartes should have taken, it seems factually incorrect to say that it is
the course he actually did take. There are two dierent reasons for this.
First, as we have already seen, there are a number of passages in the
Meditations and in Descartess Replies to the Second Set of Objections where
he says that God might have given him a nature such that he goes wrong
even about matters that seem most evident to him. Nor are the passages
already discussed the only ones of this kind. In his Replies to the Second Set
of Objections, he also says that an atheist ... cannot be certain that he is
not being deceived on matters which seem to him to be very evident (CSM
II , SPW , AT VII ). In the Principles of Philosophy, he says that
prior to knowing the existence of God, the mind is still ignorant as to
whether it may have been created with the kind of nature that makes it go
wrong even in matters which appear most evident (CSM I , SPW ,
AT VIIIA ), and that the proof of Gods existence and veracity dis-
poses of the most serious doubt which arose from our ignorance about
whether our nature might not be such as to make us go wrong even in mat-
ters which seemed to us utterly evident (CSM I , SPW , AT VIIIA
). In short, there is compelling textual evidence that Descartes was pre-
pared to extend his doubt even to reason itself.
The second reason for denying that Descartes held clear and distinct
perceptions to be self-guaranteeing stems from a doctrine of his that up
to now, we have not mentioned. This is his view that God created the
principles of logic and mathematics. According to this doctrine, which
commentators call the creation of the eternal truths, the principles of
mathematics and logic depend on God, and he could have made those
principles dierent from what they are. Although Descartes never men-
tions this extraordinary doctrine in the Meditations, he asserts it several
times in his correspondence. Here are some representative passages:
The mathematical truths which you call eternal have been laid
down by God and depend on Him entirely no less than the rest of
His creatures. Indeed to say that these truths are independent
of God is to talk of Him as if he were Jupiter or Saturn and to
Meditation III
subject Him to the Styx and the Fates. Please do not hesitate to
assert and proclaim everywhere that it is God who has laid down
these laws in nature just as a kind lays down laws in his kingdom.
(To Mersenne, April , , CSMK , AT I )
As for the eternal truths ... we must not say that if God did not
exist nevertheless these truths would be true; for the existence
of God is the rst and most eternal of all possible truths and the
one form which alone all others proceed... . [People] should ...
take the . . . view, that since God is a cause whose power sur-
passes the bounds of human understanding ... these truths are
therefore something less than, and subject to, the incomprehen-
sible power of God. (To Mersenne, May , , CSMK , AT
I )
You ask me by what kind of causality God established the eternal
truths. I reply: by the same kind of causality as he created all
things, that is to say, as their ecient and total cause... . [God]
was free to make it not true that all the radii of the circle are
equaljust as free as he was not to create the world. (To Mer-
senne, May , , CSMK , AT I )
[E]ven those truths which are called eternalas that the whole
is greater than its partswould not be truths if God had not so
established, as I think I wrote you once before. (To Mersenne,
May , , CSMK , AT II )
God cannot have been determined to make it true that contra-
dictories cannot be true together, and therefore . . . he could
have done the opposite. (To Mesland May , , CSMK ,
AT IV )
Descartess doctrine of the creation of the eternal truths is powerful

evidence against the view that he held clear and distinct perception to be
self-guaranteeing; for this doctrine implies a possible scenario on which,
according to Descartes, the very principles that we take to be the most
clear and obvious could be false. According to this scenario, (a) God could
have made it false that + = , or that not-(p and not-p), or that a whole
is greater than any of its parts, or the like, and yet (b) have made our minds
such that we cannot see the sum of two and two as being anything other
than four, or the truth of p as being anything other than incompatible
Descartes
with the truth of not-p, or a whole as being less than, or equal to, any of
its parts, and so forth. As Descartes himself says to Arnauld:
[S]ince everything in truth and goodness depends on [Gods]

omnipotence, I would not dare to say that God cannot make ...
[it] that one and two should not be three. I merely say that He
has given me such a mind that I cannot conceive ... an aggregate
of one and two which is not three, and that such things involve a
contradiction in my conception. (To Arnauld, July , ,
CSMK , AT V )
In his Replies to the Sixth set of Objections, Descartes admits that we

cannot understand how God could have made the eternal truths dif-
ferent than we take them to be. But just as in his letters, he insists that
God could have done this:
[T]here is no need to ask how God could have brought it about

from eternity that it was not true that twice four make eight, and
so on; for I admit that this is unintelligible to us. Yet on the other
hand I do understand, quite correctly, that there cannot be any
class of entity that does not depend on God; I also understand
that it would have been easy for God to ordain certain things
such that we men cannot understand the possibility of their
being otherwise than they are. (CSM II , AT VII )
Notice that Descartess doctrine seems to clash with his own claim, in
the Second Set of Replies, that at least some propositions can never be
doubted at all. For to hold that God could (a) have made it false that not-
(p and not-p) or that + = , but (b) created our minds in such a way that
we cannot see those propositions except as being true, seems tanta-
mount to doubting those propositions, at least in some minimal way.
Here the consistency of Descartess position is under maximum strain.
Most philosophers today, including the present writer, would reject
Descartess doctrine of the creation of the eternal truths. They would
hold, as Leibniz held, that the truths of logic and pure mathematics are
true in all the possible worlds that God could have created, so that not
even an omnipotent God could dictate or alter them. They would also
agree with the view, put forward by Aquinas, that omnipotence does not
require the power to do logically impossible things but only the power to
Meditation III
do whatever is logically possible. Nevertheless, Descartess doctrine

should not be dismissed as a mere anomaly or eccentricity in his thought.
For the doctrine is in a certain sense deeply Cartesian; it is a striking
manifestation of what might be called Descartess philosophical radi-
calism. Other manifestations of this radicalism are, for example,
Descartess carrying traditional skeptical arguments to the extreme of
doubting the existence of the entire physical world; his attempt to rebuild
all knowledge from one single point of absolute certainty; his sharp du-
alism of thinking and unextended, versus extended and unthinking, sub-
stance; and his view that nite substances depend on God not only for
their original creation but also for their continuance in existence from
one moment to the next. Descartess notions that omnipotence entails
even the power to do logically impossible things and that God legislated
the principles of logic and mathematics bespeak the same radical cast of
mind. One need not agree with these notions to recognize that they are
inherent parts of Descartess thought.
It might now seem, however, that Descartess radical doubt of reason
depends on his doctrine of the creation of the eternal truths. The idea
would be that such propositions as + = or both p and not p are so crazy
that they could be true only if God made them true. Such a dependence
of the doubt of reason on the doctrine of the creation of the eternal
truths would greatly reduce the interest of Descartess doubt of reason,
just because that doctrine is so mind-boggling. So, it is important to see
that the radical doubt of reason does not really have to depend on the
doctrine of the creation of the eternal truths. For if it makes any sense to
suppose that God could have made + = true, then it also makes sense
to suppose that + = could be unalterably true; and if it makes any
sense to suppose that God could have made both p and not p true, then it
also makes sense to suppose that p and not p could both be true whether
God made them true or not. But then the doubt of reason can still be
generated: why couldnt + = be unalterably true, yet God have so
made our minds that we cant see the sum of and as being anything
other than ; and why couldnt both p and not p be true even without
Gods having made them so, yet God have so made our minds that we
cant see both p and not p as being anything but false? To allow the possi-
bility that God gave us radically defective minds is to allow that it could
be that (a) plus is unalterably , that for some p, p and not p is unalter-
ably true, that the radii of a circle are unalterably unequal, and so on, and
yet that (b) God has so made us that + = is assent-compelling for us,
Descartes
not both p and not p is assent-compelling for us, and the radii of a circle are
equal is assent-compelling for us, and so forth. Briey put, the point is
that as implausible as Descartess doctrine of the creation of the eternal
truths may be, that doctrine is not necessary to generate his doubt of
reason; for suppose that the doctrine is wrong and that not even God
could alter the principles of logic and mathematics. Then Descartes can
still maintain that an omnipotent God could deceive us about those prin-
ciples, simply by creating our minds in such a way that the principles that
seem necessarily true and indubitable to us are dierent from the ones
that actually and unalterably hold.
.. The Validation of Reason

In light of Descartess doubt of reason, it appears that the only way to
solve the problem of the circle that would be consistent with Descartess
own position would be to show that, appearances to the contrary not-
withstanding, it is possible to use reason to overcome a doubt about the
reliability of reason itself. Interestingly enough, two contemporary Des-
cartes scholars, Alan Gewirth and Harry Frankfurt, have argued that
Descartes undertakes exactly this project and actually succeeds in car-
rying it out. Both of these writers oer far-reaching and complex inter-
pretations of Descartess theory of knowledge, and we cannot do justice
to their work here. Rather, we shall briey outline their defense of Des-
cartes. Then, we shall argue that their defenseor, more precisely, a
defense of our own built partly on a key point in their defenseis,
indeed, viable.
The key point common to both Gewirths and Frankfurts defense of
Descartes is that his doubt concerning human reason is not arbitrary or

I am indebted to James Van Cleve for this point. Margaret Wilson makes the same
point in her Descartes, pp. .

The original source for this approach to the problem of the circle is Alan Gewirth,
The Cartesian Circle, . Gewirth also expounds his view in two more recent arti-
cles: The Cartesian Circle Reconsidered, , and Descartes: Two Disputed Ques-
tions, . All three of these articles are reprinted in Willis Doney ed., Eternal Truths
and the Cartesian Circle (New York: Garland ). The latter two of the articles are part
of a very useful exchange with Anthony Kenny, to which Kennys contribution is his
The Cartesian Circle and the Eternal Truths. Frankfurts contributions, in which he ac-
knowledges Gewirths inuence, are in his Descartess Validation of Reason and in his
Demons, Dreamers, and Madmen (esp. chap. ).
Meditation III
willful. Rather, as Descartes himself says in several places, his doubt is

based on reasons. Thus, Gewirth writes:
At the outset of his enterprise, Descartes had laid it down that it

will be sucient to reject all my opinions, if I nd some reason
for doubting in each one; at the end of the First Meditation,
accordingly, the doubt is said to be based upon valid and medi-
tated reasons, and in the Third, when it is stated in its most
extreme form, Descartes calls it a Metaphysica dubitandi ratio [a
metaphysical reason for doubt] which depends upon his opin-
ion concerning his deception by an omnipotent God. The very
nature of the doubt as conceived by Descartes thus involves an
ineluctable rational element: the must be some reason for
doubting before one determines oneself thereto. (Quoting from
Meditation I [CSM II , SPW , AT VII ; CSM II , SPW
, AT VII ] and Meditation III [CSM II , SPW , AT
VII ], respectively; and from Descartess letter to Clerselier of
January , [CSM II , AT IXA ].)
In John Cottinghams translation, the passage that Gewirth quotes from

the letter to Clerselier says, before we can decide to doubt, we need some
reason for doubting and continues and that is why in my First Medita-
tion I put forward the principal reasons for doubt. As one recent com-
mentator puts it, the First Meditations method of universal doubt clearly
indicates that doubting, far from being an arbitrary enterprise, requires
us to attend to reasons, even such that we may later nd to be question-
able or false. The key point, then, is that the deceiving-God hypothesis
is supposed to provide a reason or argument for doubting, and that at least
by the time we get to Meditation III, it is supposed to provide a reason for
doubting even the reliability of reason. Therefore, argue Gewirth and
Frankfurt, Descartes can refute this reason for doubt (i.e., disqualify it as
a reason for doubt) by showing that the deceiving God hypothesis is not a
good reason for doubting the reliability of reason. But that is exactly what
Descartess proof of a veracious God shows; for it shows that the deceiving
God hypothesis is ruled out by, or is incompatible with, a chain of
argument built entirely on clear and distinct perceptions. Notice that this

Gewirth, The Cartesian Circle, p. .

Noa Naaman-Zauderer, Descartes Deontological Turn, p. .
Descartes
defense of Descartes does not assume that clear and distinct perceptions
must be true, which would, of course, beg the question. Rather, it turns on
the point that the most careful and conscientious application of reason
(i.e., one that involves only clear and distinct perceptions and is not sub-
ject to the weaknesses of sense perception, memory, and muddled or care-
less thinking) leads to the conclusion that the reason Descartes had for
doubting the reliability of reason is a bad one, thereby nullifying it as a
reason for doubt.
Nevertheless, it may initially seem that this defense of Descartes is
open to a fatal objection; for the content of Descartess doubt is precisely
whether the best possible use of reason is reliable or trustworthy. If his
doubt pertained to anything elsesay, only to the reliability of the
senses, or of memory, or of other peoples testimony, or the likethen
showing that the reasons given for such a doubt are defeated by the best
use of reason could refute the doubt. But, given that Descartess doubt
extends to whether the best use of reason is itself reliable, the logic of the
situation may seem to be as follows. Gewirth and Frankfurts Descartes
argues that since the best use of reason leads to the conclusion that rea-
son itself is reliable, it is unreasonable to doubt whether reason is reli-
able. The critic, however, can reply that Descartess radical doubt was
precisely whether the best use of reason is reliable, so the argument pur-
porting to show that it is reliable cannot be trusted unless one assumes
that this doubt has already been dispelled before giving the argument.
Therefore, it may seem, the argument cannot refute the radical doubt.
We shall now argue, however, that this powerful-looking objection can
be answered. As already noted, the key point in the Gewirth-Frankfurt
defense of Descartes is the correct insistence that Descartess doubt of
reason is itself supposed to be based on reasons. Now, what exactly is the
reason that generates Descartess radical doubt? It is, as Descartes says in
Meditation III, the idea that an omnipotent God could make him go wrong
even about the things he perceives with the utmost clarity: Whenever
my preconceived belief in the supreme power of God comes to mind, I
cannot but admit that it would be easy for him, if he so desired, to bring
it about that I go wrong even in those matters which I think I see utterly
clearly with my minds eye (CSM II , SPW , AT VII ). Descartess
reason for doubting reason, then, is this: Surely an omnipotent God
could easily have given a mere creature like poor little me a radically de-
fective intellectone that goes wrong even in matters that seem utterly
clear and evident.
Meditation III
Notice, however, the following crucial point about this reason for
doubting reason: it relies on an inference. This is the inference from
() It is possible that there exists an omnipotent God
to
() It is possible that there exists a God who makes me go wrong

even about the things I perceive clearly and distinctly.
Indeed, it seems that the inference from () to () must go through at

least one intermediate step, namely:
(a) It is possible that there exists a God who can do anything.

(derived from step and leading to step )
Or perhaps the inference should be seen as involving yet another inter-

mediate step, namely,
(b) It is possible that there exists a God who can make me go

wrong even about the things I perceive clearly and distinctly.
(derived from step a and leading to step )
The inference from () to (), regardless of how exactly it is construed,

is essential to Descartess radical doubt; for without this inference, Des-
cartess doubt of reason would be merely willful and arbitrary. It would
reduce to the bald assertion that he might be mistaken even about what
he perceives most clearly. Furthermore, the inference has to bear an enor-
mous weight. It must be solid enough to convince Descartes that it is at
least possible that the things he perceives with the utmost clarity could be
false. Indeed, as the passages we cited in connection with Descartess doc-
trine of the creation of the eternal truths suggest, the inference must bear
the weight of convincing him that it could actually be false that + = ,
that not both p and not-p, that a whole is greater than any of its parts, and so
on; for whether or not omnipotence requires the power to make these

In an early version of this section, I attempted, in eect, to deny this point. Roland
P. Blums objection to this denial led me to see the importance of the point. The material
that follows is also inuenced by Gewirths The Cartesian Circle Reconsidered, sec. .
Descartes
principles false, the inference commits one to admitting that they might
in any case be false. Descartess insistence that such principles cannot be
doubted while one is attending to them, and his claim in the second set of
Replies that they can never be doubted at all, can serve to impress on us
how heavy a weight his inference from Gods omnipotence must bear.
But if Descartes can relyand indeed rely so heavilyon an inference
from the possibility that an omnipotent God exists to the possibility that
this God deceives him even about what seems utterly clear to him, then
why can he not rely on the chain of reasoning which leads to the conclusion
that although an omnipotent God who could deceive him does indeed exist,
this God is also an absolutely perfect being who, therefore, would not deceive
him about his clear and distinct perceptions? You may object that Des-
cartess proof of a nondeceiving God is neither as simple nor as obvious as
is the inference from Gods omnipotence to his possibly deceiving us about
even the simplest things. But this objection, even if correct, is here besides
the point; for the question posed by the problem of the circle is not whether
Descartess specic argument for Gods existence and veracity actually suc-
ceeds in dispelling the doubt of reason but whether any such argument
could succeed. And our suggestion is that it could. For once we grant the
legitimacy of the use of reason required to infer the possibility of our going
wrong about the simplest things from the possibility that there is an om-
nipotent God, there is no reason in principle to deny the legitimacy of the
use of reason which leads to the conclusion that the omnipotent God who
actually exists is a perfect being who, while still fully able to deceive his
creatures, would not wish to do so. Indeed, consistency requires that if we
allow the former use of reason to be legitimate, then we must also allow the
latter to be legitimate. As James Van Cleve incisively put it: if reason can
be used to attack reason, then it can also be used to defend reason.
This defense of Descartes can also be put another way. If we look at the
inference from the possibility that there is an omnipotent God to his pos-
sibly deceiving us about the simplest things in complete isolation from all
other considerations bearing on the existence and nature of God, then
we may suppose that we might be deceived about the things we perceive
most clearly; for since an omnipotent God would surely have the power
to deceive us, nothing, so far, prevents us from suspecting that he
might or does deceive us. So, the inference provides a reasoned basis
Descartess very slight and, so to speak, metaphysical reasonfor

In conversation.
Meditation III
doubting the reliability of reason. However, once we have come to a fuller

view of the considerations bearing on Gods existence and nature and
have deployed the chain of reasoning which leads to the conclusion that
a God exists who is not only omnipotent but also supremely good, the
suspicion is removed and the doubt is dispelled; for although the most
careful reasoning we are capable of leads us to believe that God could, if
he wished, deceive us even about the simplest matters and although this
reasoning thereby initially provides some reason for doubting reason
itself, the same kind of reasoning, further and more deeply pursued,
shows that God does not in fact (wish to) deceive us. Thus, the original
reason for doubt can operate only on the condition that we accept the
argument that eventually dispels it. As some philosophers might put it,
the reason for doubt is only a prima facie or defeasible reason, which
can be disposed of by further reection.
To put it yet one more way: although a perfect God exists does entail
()that it is possible that there exists a God who makes me go wrong even
about the things I perceive clearly and distinctlyit also entails that God
does not (wish to) make me go wrong about the things I perceive clearly and
distinctly. Thus, Descartes does not show that it is impossible that there is a
God who deceives him about the things he perceives clearly and distinctly;
he does not show that () abovethat it is possible that there exists a God
who makes me go wrong even about the things I perceive clearly and
distinctlyis false. Rather, he shows that it is false that there is a God who
actually does deceive him even about the things he perceives clearly and
distinctly. And this is enough to refute the radical doubt of reason.
It might be asked, Which of the two strategies earlier discussed does
our defense of Descartes employ? Are we saying that Descartes can know
that whatever he clearly and distinctly perceives is true before knowing
that a nondeceiving God exists (the vindication-not-needed strategy)
or that he can know that God exists and is no deceiver before knowing
that whatever he perceives clearly and distinctly is true (the criterion-not-
needed strategy)? The answer is that according to the defense we have
proposed (and assuming for the moment that Descartess theological ar-
gument is sound), Descartes comes to know those two propositions simul-
taneously. Before the argument for Gods existence and veracity is complete,
the inference from Gods omnipotence to his possibly deceiving us provides
a reason to doubt clear and distinct perceptions. But the moment the
argument is completed, that reason is nullied, and Descartes comes to
know both that a nondeceiving God exists and that his clear and distinct
Descartes
perceptions are true. The entire argument is, as it were, on probation

until its nal conclusion is reached. Thus, our defense relates to the two
strategies as follows. In common with both strategies, it recognizes that
Descartess theological argument uses clear and distinct perceptions. But
in contrast to both strategies, it does not assume that the perceptions
used in the argument are known to be reliable before the argument is com-
plete. Rather, they are used only because they constitute the most careful
use of the intellect we are capable of. However, in common with the criteri-
on-not-needed strategy, our defense holds that once the theological
argument is complete, reason is validated (since the reason that was given
for doubting reason is nullied). And in contrast to the vindication-
not-needed strategy, our defense holds that the purpose of the theological
argument is not merely to justify appealing to the results of past uses of
reason; rather, its purpose is to provide a validation of reason itself.
4.3 A Final Criticism of the Core Argument

If we assume that the problem of the Cartesian Circle can be disposed of
(either in the manner we have proposed, or, failing that, by restricting
the Cartesian doubt to sense-perception), then Descartess Meditation III
argument for Gods existence may, at this point, still seem worthy of ac-
ceptance; for, as we saw in section .., the core arguments rst
premisethat the cause of an idea must have as much formal reality as
the idea contains objective realitycan be reasonably maintained, espe-
cially if it is taken as a basic premise, rather than supported by the subar-
gument Descartes himself oers for it. May we conclude, then, that
Descartess argument is successful?
There is reason to doubt that such an evaluation would be correct. For
consider the core arguments second premise, namely, Only a perfect
God has as much formal reality as my idea of God contains objective re-
ality. This premise means that only a perfect God has as much reality as
my idea of God represents him as havingin other words, that my idea
of God is so rich in its informational content that only a supremely per-
fect God himself has as much reality as that content displays. This claim
is a basic premise in the argument; it is not derived from any other

Cf. Van Cleve, Foundationalism, Epistemic Principles, and the Cartesian Circle,
p. , n. .
Meditation III
premise of the argument and it is supposed to be known just by contem-

plating ones idea of God. But given the limits of the human mind, is such
a claim plausible? In the Fifth set of Objections, Gassendi wrote:
But ... the idea, or its objective reality, is not to be measured by the
total formal reality of the thing (i.e., the reality which the thing has
in itself) but merely by that part of the thing of which the intellect
has acquired knowledge (i.e., by the knowledge that the intellect
has of the thing). Thus you will be said to have a perfect idea of a
man if you have looked at him carefully and often from all sides;
but your idea will be imperfect if you have merely seen him in
passing and on one occasion and from one side ... You claim that
there is in the idea of an innite God more objective reality than in
the idea of a nite thing. But rst of all, the human intellect is not
capable of conceiving innity, and hence it neither has nor can
contemplate any idea representing an innite thing. Hence if
someone calls something innite he attributes to a thing which
he does not grasp a label which he does not understand. (CSM II
, AT VII ; see also CSM II , AT VII )
Gassendis objection might be put this way. The degree of objective reality
contained in an idea depends on the formal reality of only that part of the
ideas object that we understand, not on the full degree or amount of for-
mal reality possessed by the object itself. [Notice that this goes directly
against the second key idea behind the core argument; namely that the
degree of reality that an idea represents its object as having (its degree of
objective reality) depends (solely) on the degree of reality possessed by the
object itself (on its objects degree of formal reality)]. But the human mind
cannot grasp the innite degree of formal reality had by God. So Des-
cartess premise that only a perfect God has as much formal reality as his
idea of a perfect God contains objective reality is simply false. Many things
that are far less perfect than God have as much formal reality as Des-
cartess (necessarily inadequate) idea of God contains objective reality.
Replying to a similar objection, which had also been raised in the First
set of Objections, Descartes wrote:
At this point, however, he shrewdly asks whether I am clearly

and distinctly aware of the innite ... Let me say rst of all that
Descartes
the innite, qua innite, can in no way be grasped. But it can still
be understood, in so far as we can clearly and distinctly under-
stand that something is such that no limitations can be found in
it, and this amounts to understanding clearly that it is innite.
(CSM II , AT VII )
Replying to Gassendi himself, Descartes makes the same point at greater

length:
I shall make one point about the idea of the infinite. This, you
say, cannot be a true idea unless I grasp the infinite; you say
that I can be said, at most, to know part of the infinite, and a
very small part at that, which does not correspond to the infi-
nite any better than a picture of one tiny hair represents the
whole person to whom it belongs. My point is that, on the
contrary, if I can grasp something, it would be a total contra-
diction for that which I grasp to be infinite. For the idea of
the infinite, if it is to be a true idea, cannot be grasped at all,
since the impossibility of being grasped is contained in the
formal definition of the infinite. Nonetheless, it is evident
that the idea which we have of the infinite does not merely
represent one part of it, but does really represent the infinite
in its entirety. The manner of representation, however, is the
manner appropriate to a human idea; and undoubtedly God,
or some other intelligent nature more perfect than a human
mind, could have a much more perfect, i.e., more accurate and
distinct, idea. Similarly we do not doubt that a novice at
geometry has an idea of a whole triangle when he understands
that it is a figure bounded by three lines, even though geome-
ters are capable of knowing and recognizing in this idea many
more properties belonging to the same triangle, of which the
novice is ignorant. Just as it suffices for the possession of an
idea of the whole triangle to understand that it is a figure
contained within three lines, so it suffices for a true and com-
plete idea of the infinite in its entirety if we understand that
it is a thing which is bounded by no limits. (CSM II , AT
VII )
Again, in one of his letters, Descartes says:

Meditation III
[I]t is possible to know that God is innite and all-powerful

although our soul, being nite, cannot comprehend or conceive
Him. In the same way we can touch a mountain with our hands
but we cannot put our arms around it as we could put them
around a tree or something else not too large for them. To com-
prehend something is to embrace it in ones thought; to know
something it is sucient to touch it with ones thought. (To Mer-
senne May , , CSMK , AT I )
In these passages, Descartes can be seen as answering in reply to Gas-

sendis objection, in eect, In order for the core arguments second pre-
mise to be true, I need not have a completely adequate understanding of
Gods nature. I need only know that his power, goodness, and knowledge
have no limits.
In his Descartes, Bernard Williams interprets this reply as follows: We
can clearly and distinctly conceive that God is innite, but not how he is.
Williams then argues that this reply is unsuccessful:
Descartess course ... is, in eect, that he can clearly and dis-
tinctly conceive that God is actually innite, but not how he is.
But that this is an unsatisfactory line of defence can be seen if
one reverts to Descartess own helpful analogy of the man who
had the idea of the very complex machine. From the fact that a
man has this idea ... it could be inferred that either he had seen
such a machine (or . . . had been told about it) or that he was
clever enough to invent it. But clearly such inferences will hold
only if the man has a quite determinate idea of the machine. If a
man comes up and says that he has an idea of a marvelous
machine which will feed the hungry by making proteins out of
sand, I shall be impressed neither by his experience nor by his
powers of invention if it turns out that that is all there is to the
idea, and that he has no conception, or only the haziest concep-
tion, of how such a machine might work.
Williams is not here questioning Descartess principle that the cause of

an idea must have as much formal reality as the idea contains objective


Descartes
reality; he allows that one could use such a principle to infer a very im-
pressive cause from a very impressive idea. Rather, Williams is questi-
oning whether, given this principle, our idea of God is suciently
impressive to warrant the inference to God as its cause. His point is that
since, for example, we do not understand how Gods omnipotence works,
or how he can possess knowledge of all past, present and future events,
or how he can allow evil if he is supremely good, our idea of God is not
suciently rich or detailed in its content to warrant an inference to God
as its cause. For it is simply not the case that only a perfect (omnipotent,
omniscient, omnibenevolent) God has as much formal reality as our idea
of him contains objective reality.
To see Williamss point more clearly, suppose that the content of a
mans idea of a machine that can turn sand into protein were just this:
machine that can turn sand into protein. Is it true that only a machine
that really can turn sand into protein would have as much complexity
(formal reality) as does that ideas content (objective reality)? No, for
such a machine would have to have many extraordinary components, ca-
pable of executing many extraordinary chemical reactions, that are not
represented in this mans idea. So its false that only a machine that can
really turn sand into protein has as much complexity as does the content
of this mans idea. Likewise, suppose that the content of a humans idea
of God is just: being who is all-powerful, all-knowing, and supremely
good. Is it true that only a being that really has these properties has as
much formal reality as that ideas content or objective reality? No,
because such a being would have to have many properties (including
detailed knowledge of every past, present, and future truth, knowledge
of how to create something out of nothing, knowledge of how apparent
evils all lead to a greater good) that are not represented in a human idea.
As Daniel Dennett neatly says in a recent article, Descartes seems to be
confusing an idea of a wonderful thing with a wonderful idea of a wonderful
thing. It seems just false, then, that only a perfect or innite God has as
much formal reality as a human beings idea of God has objective reality.
This rebuttal of Descartess reply to Gassendi seems very powerful. In
the end, then, it is doubtful that Descartess Meditation III argument, for
all its profundity, provides a compelling case for the existence of God. As
we shall see in chapter , however, Descartes oers us yet another proof
of Gods existencethe famous Ontological Argument.

Daniel C. Dennett, Descartess Argument from Design, , p. .
| 4 |
Meditation IV
Error, Freedom, and Evil
1. The Issues of the Fourth Meditation

In his Fourth Meditation, Descartes addresses two questions that natu-
rally arise from his previous Meditations: () How can I make errors if
God is not a deceiver? () How can God be a supremely perfect being if
he has so created me that I am capable of making errors? The rst prob-
lem must be addressed for the Cartesian doubt to be completely and
permanently cured, for so long as we do not know what the cause of our
errors is, there is room for a residual doubt about the reliability of our
faculties. The second problem is a special case of the classic problem of
evil: how can we reconcile the belief in a perfect God with the existence
of evil (in this case the existence of error, seen as a particular kind of
evil) in the world that he created? The two problems are intertwined
in Descartess exposition, but they will be discussed separately in this
chapter.
2. Error and the Will

Near the beginning of the Fourth Meditation, Descartes writes:
I know by experience that there is in me a faculty of judgement

which, like everything else which is in me, I certainly received
from God. And since God does not wish to deceive me, he surely
did not give me the kind of faculty which would ever enable me
to go wrong while using it correctly. (CSM II , SPW , AT VII
)

Descartes
In other words, since God is not a deceiver, it must be that I cannot go

wrong provided I use the intellectual powers that he has given me cor-
rectly. So my errors must come from some misuse of my faculties. What
then is this misuse? In a nutshell, Descartess answer is that he has two
mental faculties, the intellect and the will, and that error does not come
from either one of them separately, but from the joint action of the two.
The intellect is the capacity to have clear and distinct perceptions. Des-
cartes has two important doctrines about it. First, the intellect cannot
arm or deny anything; or, putting it more carefully, by using only the
intellect, a person or mind cannot arm or deny anything. The reason is
that the intellect, which Descartes also calls the faculty of understanding
(CSM II , SPW , AT VII ) and my understanding [which] comes
from God (CSM II , SPW , AT VII ), is only the capacity to under-
stand a proposition or an idea; it does not enable one to arm or to deny
any proposition, or to arm or deny that any of its ideas correspond to
reality. It immediately follows from this that the intellect alone cannot be
the source of error, since error consists precisely in arming a false prop-
osition or in denying a true proposition (or, Descartes says, in pursuing
evil or sin rather than good). Thus, Descartes can safely hold that the in-
tellect is in no way defective or faulty. Second, on the other hand, the
(human) intellect is nite or limited; it cannot grasp everything or under-
stand every question; there are some matters that it cannot perceive
clearly and distinctly. Descartes insists that the nitude of the human
intellect, per se, in no way diminishes Gods perfection. As he says,
I have no cause for complaint on the grounds that the power of

understanding or the natural light which God gave me is no
greater than it is; for it is in the nature of a nite intellect to
lack understanding of many things, and it is in the nature of
a created intellect to be nite. Indeed, I have reason to give
thanks to him who has never owed me anything for the great
bounty that he has shown me, rather than thinking myself
deprived or robbed of any gifts he did not bestow. (CSM II ,
SPW , AT VII )

Some commentators hold that for Descartes, even arming a true proposition or
denying a false one counts as an error, provided that the armation or denial is not based
on clear and distinct perceptions. See for example Noa Naaman-Zauderer, Descartes
Deontological Turn, pp. , , , , .
Meditation IV
The will, by contrast, is the capacity to arm or to deny (or more gen-
erally, to do X or not to do X, where X includes not only arming or de-
nying, but also pursuing or avoiding); it is thus necessarily involved in
assenting to any proposition or in making any judgment. Descartess
most important doctrine about the will is that, unlike the intellect, its
scope is unlimited. In the Fourth Meditation, he says that it is only the
will, or freedom of choice, which I experience within me to be so great
that the idea of any greater faculty is beyond my grasp; so much so that it
is above all in virtue of the will that I understand myself to bear in some
way the image and likeness of God (CSM II , SPW , AT VII ), and
in the Principles of Philosophy he even says that [the will] can in a certain
sense be called innite (CSM I , SPW , AT VIIIA ). What this
doctrine means is that there is no proposition grasped by the intellect
that the will can neither arm nor deny, and no course of action that it
cannot choose. However, since clear and distinct perceptions are assent-
compelling, it does not mean that for any proposition, we can always
either arm or deny it: when we clearly and distinctly perceive a propo-
sition, we can only arm it. Nor does it mean that we can actually carry
out any course of action that we choose, since unlike God we lack the
power to do many things.
It follows that one can choose to arm or to deny even propositions
that surpass one s intellect or understandingthat is, that one does not
clearly and distinctly perceiveand this is precisely how error arises. As
Descartes puts it,
When I look more closely into myself and inquire into the nature
of my errors ..., I notice that they depend on two concurrent
causes, namely on the faculty of knowledge which is in me, and
on the faculty of choice or freedom of the will; that is, they
depend on both the intellect and the will simultaneously ... [For]
the scope of the will is wider than that of the intellect; but instead
of restricting it within the same limits, I extend its use to mat-
ters which I do not understand. Since the will is indierent in
such cases, it easily turns aside from what is true and good, and
this is the source of my error and sin. (CSM II , SPW
, AT VII )

It is not clear whether Descartes means to include, among actions that we can choose,
ones that no human being can actually perform, such as jumping ten meters high.
Descartes
The corollary of this analysis of error, as the wills arming or denying prop-
ositions that the intellect does not clearly and distinctly perceive, is that if
one restricts ones judgments of truth or falsity (and, Descartes says, of
goodness and badness) to what one clearly and distinctly perceives, then one
cannot go wrongone is then infallible. Descartes fully accepts this result:
If, whenever I have to make a judgment, I restrain my will so that

it extends to what the intellect clearly and distinctly reveals, and
no further, then it is quite impossible for me to go wrong. (CSM
II , SPW , AT VII )
3. Some Possible Objections

Although the secondary literature on Descartess Fourth Meditation is
much smaller than that on any of the other ve, his theory of error raises
fascinating issues. We shall conne our discussion to two possible objec-
tions to the theory.
3.1 Assenting and Deciding to Believe

The most common objection to Descartess theory of error is that by at-
tributing armation and denial to the will, it wrongly implies that people
can simply decide what to believe, even if there is no basis at all for their
decision. E. M. Curley puts the objection this way:
Consider some proposition for which you have literally no evi-

dence at all one way or the other. It may not be as easy as you
might suppose to think of an example that strictly satises that
condition, but I am in that situation with respect to the proposi-
tion it rained three hours ago on Jupiter. Now, paying careful
attention to what is happening in your mind, believe it. Or, if you
prefer, disbelieve it. Or do both, in turn. Did anything happen?
Unless your experience is very unlike mine, I suspect not. Indeed.
I fear that if my salvation depended on my either believing or
disbelieving this particular proposition, I should be damned.

E. M. Curley, Descartes, Spinoza and the Ethics of Belief, p. .
Meditation IV
This objection seems to be directed against a view that Descartes never

arms and that he is not logically committed to. Nowhere does Descartes
say that we can simply decide what to believethat (a) for any proposition
p that I understand, I can just decide to believe or to disbelieve p. Rather,
he holds the dierent view that (b) for any proposition p that I understand
(with the exception of the negations of propositions that I currently per-
ceive clearly and distinctly) there is nothing to stop me from believing or
assenting to p. This is perfectly compatible with holding, as indeed there is
good reason to think Descartes did hold, that humans do not adopt a belief
unless there is some reason that moves them to do sothat the will does
not arm a proposition p in a perfectly arbitrary way, without any reason
whatsoever. That reason might be a poor or frivolous one; for example, I
might believe that it rained on Jupiter because my mother told me so, or
because the earth and Jupiter are aligned in a certain way, or because I
wishfully think that there are lush forests on Jupiter. But that is not to say
that I believe it arbitrarily, and whether humans ever do believe things
that way seems to be an empirical question.
Basically, Descartes is only making the common-sense point that we
are capable of arming and denying things before making a clear and
reasoned judgment. He is also saying that if we arm or deny something
before making a careful and reasoned judgment, then we are misusing
our freedom and that this misuse is the source of error. More grandly,
since he holds that there is no proposition we understand that we can
neither arm not deny, and calls the capacity to arm or deny a propo-
sition the will, he says that the scope of the will is innite. But there
is nothing in his texts to indicate that by this he means that we can just
decide to believe any proposition. At most, Descartes might be criticized
for not saying anything to forestall such a misunderstanding of his posi-
tion, especially given his talk of innity.
This is not to say that there are no issues about the extent to which we
can control our beliefs. On the contrary, the topic of doxastic volunta-
rism is a controversial one, on which there is a substantial body of liter-
ature that addresses questions like, Can we believe something in the face
of overwhelming evidence against it? Is wishful thinking, or indoctrina-
tion, or superstition, potentially sucient to neutralize the weight of
any such evidence in our minds, as Descartes would presumably say? But
it is not necessary to pursue these questions in order to exonerate Des-
cartes from the common criticism that on his view we can just arbitrarily
decide to believe anything at all.
Descartes
3.2 Irresistibility and Freedom

As John Cottingham observes, the central doctrine of Descartes Fourth
Meditation is what may be called the doctrine of the irresistibility of the
natural light: when the intellect is confronted with a clear and distinct
perception, the will is immediately and spontaneously impelled to give
its assent to the truth of the relevant proposition. Indeed, we have seen
that Descartes holds that clear and distinct perceptions are irresistible or
assent-compelling, that is, that during the time that one clearly and dis-
tinctly perceives some proposition p, one cannot help believing p, one
cannot doubt that p. In terms of his theory about the role of the intellect
and the will in making judgments, his position is that as long as the intel-
lect clearly and distinctly perceives p, the will cannot do otherwise than
assent to p. This doctrine is one that he insists upon in many places, for
example:
My nature is such that so long as I perceive something very

clearly and distinctly I cannot but believe that it is true. (CSM II
, AT VII , SPW )
I have already demonstrated that everything of which I am clearly
aware is true. And even if I had not demonstrated this, the nature
of my mind is such that I cannot but assent to these things, at least
so long as I clearly perceive them. (CSM II , SPW , AT VII )
So long as I attend to the proof [that the three angles of a triangle
are equal to two right angles], I cannot but believe this to be true.
So long as we attend to a truth which we perceive very clearly, we
cannot doubt it. (CSM II , AT VII )
On the other hand, Descartes is equally insistent that when we assent

to a clear and distinct perception, we do so freely; indeed he thinks that
we are then at our freestthat assenting to a clear and distinct percep-
tion is the best kind of freedom that we can have. When one brings to
mind the historical contextDescartess time was one when following
the dictates of reason or the lead of the evidence rather than accepting
the dogmas sanctioned by authority was neither the favored method of

John Cottingham, A Descartes Dictionary, pp. .
Meditation IV
xing belief nor conducive to ones personal safetyDescartess position

may well strike us as a salutary celebration of intellectual freedom. But
from a strictly philosophical point of view, his position gives rise to what
looks like a powerful objection: how can it possibly be true that we are
free when we assent to a clear and distinct perception, if such a percep-
tion compels our assentif we cannot do otherwise than to assent?
To see how Descartes could deal with this problem, we need to investi-
gate more closely what he thinks freedom is. The text that scholars fre-
quently cite as most important in this regard is this passage from
Meditation IV:
The will simply consists in our ability to do or not do something

(that is, to arm or deny, to pursue or to avoid); or rather, it
consists simply in the fact that when the intellect puts some-
thing forward for armation or denial or for pursuit or avoid-
ance, our inclinations are such that we do not feel we are
determined by any external force. (CSM II , AT VII , SPW
)
At least since the appearance of an important article by Anthony Kenny,

scholars have seen this passage as problematic. Its rst part (up to the
semicolon) looks like a straightforward armation of what C. P. Ragland,
slightly adapting a term coined by Harry Frankfurt, calls the principle
of alternative possibilitiesthe principle that X is a free act only if
the agent has the ability to either do X or not do X, or the ability to do
otherwise. This principle states a condition for freedom that is highly
intuitive and that many philosophers hold to be the sine qua non of free-
dom. Another place where Descartes seems to endorse the principle, and
indeed to apply it directly to acts of assenting to true propositions, is his
Principles of Philosophy, Part One, section :
We do not praise automatons for accurately producing all the

movements they were designed to perform, because the produc-
tion of these movements occurs necessarily. It is the designer

Anthony Kenny, Descartes on the Will, pp. .

C. P. Ragland, Alternative Possibilities in Descartess Fourth Meditation,
and Descartes on the Principle of Alternative Possibilities, . I say slightly adapt-
ing because Frankfurt treats the principle as a (proposed) necessary condition for moral
responsibility, whereas Ragland treats it as a necessary condition for freedom.
Descartes
who is praised for constructing such carefully-made devices; for

in constructing them he acted not out of necessity but freely. By
the same principle, when we embrace the truth, our doing so vol-
untarily is much more to our credit than would be the case if we
could not do otherwise. (CSM I , SPW , AT VIIIA
[emphasis added])
In the second part of the above passage from Meditation IV, however,
Descartes gives what looks like a quite dierent denition of freedom.
First, he brings in his view that any free act (of arming or denying,
seeking or avoiding) must follow upon some perception provided by the
intellect. Second, he seems to be saying that what makes the act free is
that we do not feel compelled to perform it by some external agent or
force or by any factor other than ourselves.
This second formulation gives rise to two questions: () Does Descartes
really mean that in order to act freely, we need only feel that our act is
spontaneous or undetermined by any external force, or does he mean
that our act must actually be undetermined or spontaneous? () How
does the second formulation (about not being determined by any exter-
nal force) relate to the rst one (about having the two-way power of
choice, or being able to do otherwise)? Does it, as some have suggested,
retract it, or does it merely explicate it in some way? With respect to ques-
tion (), we shall here simply follow the Israeli scholar Noa Zaaman-Zau-
derer, who says, in her recent ne book on Descartes: Descartes does not
mean that our freedom demands only that we feel ourselves undeter-
mined by external coercion, but also that we be so undetermined.
With respect to (), matters are more complicated. The second formu-
lation spells out something that must not be the case in order for the rst
formulation to apply: we must not be determined in our choice by any
external force. But the second formulation is certainly not equivalent to
the rst, for one could be undetermined to do X by any external force and
yet unable to do anything other than X. This would the case if one were
determined (compelled) to do X by some factor internal to oneself, for
example by an addiction, or a compulsion, or some mental illness. Indeed,
there are philosophers who believe that no human actions are free

Noa Naaman-Zauderer, Descartes Deontological Turn, p. n. Naaman-Zauderer
cites J. K. Campbell, C. P. Ragland, and Vere Chappell as commentators who rightly em-
phasize this point.
Meditation IV
because they are all caused by internal factors that are links in a causal
chain of events that ultimately extends to events that occurred before
one was born and that were therefore out of ones control. This hard
determinist position, as it is called, rests on the twin assumptions that
(a) all events including human actions are caused by antecedent events,
and (b) an events being caused by antecedent events is incompatible
with its being a free action. We shall not delve into the complex issue of
whether (a) and (b) are both true, since that would require a full-scale
treatment of the traditional philosophical problem of freedom and deter-
minism that we cannot undertake here. But it is clear that Descartes was
not a hard determinist, so that he would have rejected (a) or (b) or both
(a) and (b); for he holds that the will is free by denition, which entails
that acts of will, or volitions, are free. As Vere Chappell says,
It is [Descartess] view that every volition is free, and that it is so,

furthermore, of necessity. For it is the essence of will, as he puts it,
to act freely; willing is free by nature... . It is not that men have the
power of willing, some of whose exercises are free and some not.
Rather they just have the power of free-willing, or willing-freely.
It follows that an action that results from a volition or an act of willone

that is voluntaryis free; an action can fail to be free only by not result-
ing from a volition, by not being voluntary. But Descartes certainly holds
that there are voluntary actions, so he is logically committed to holding
that there are free actions.
Descartes also holds, however, that freedom is not an all-or-nothing
matter; rather, according to him there are degrees or grades of freedom.
These degrees are based on the extent to which a choice or action is based
on good reasons. The lowest grade of freedom, according to this criterion,
attaches to actions that result from choices that are, from a rational
point of view, arbitraryones that are made even though there are no
good reasons for preferring any of the possible alternatives, or when the
reasons for choosing among them are evenly balanced; Descartes calls
this grade of freedom freedom (liberty) of indierence. An action that

Vere Chappell, Descartess Compatibilism, p. . Chappell cites CSM II /AT VII
, AT XI and CSM I/AT , among others, as supporting passages.

He thus gives this term a dierent meaning than did the Scholastics, who used it sim-
ply to mean the two-way power to perform or not to perform an action.
Descartes
was motivated only by impulse, passion, or emotion would presumably

exhibit this lowest grade of freedom. Intermediate grades of freedom
would attach to choices that are based partly on good reasons; as Des-
cartes puts it in a letter of May , to Mesland: I did not say that a
person was indierent only if he lacked knowledge, but rather, that he is
more indierent the fewer reasons he knows which impel him to choose
one side rather than another (CSMK , AT IV ). The highest grade
of freedom attaches to actions (armations and denials, pursuits and
avoidances) that result from a clear and distinct perception of what is
true or what is good; Descartes calls this grade of freedom freedom of
spontaneity.
This reference to freedom of spontaneity brings us back to the prob-
lem of how assenting to a clearly and distinctly perceived proposition can
be a free act if we cannot but assent to itif such a proposition is assent-
compelling. It is clear that Descartes holds that in such cases we are free,
indeed maximally free, for he writes:
During these past few days I have been asking whether anything
in the world exists, and I have realized that from the very fact of
my raising this question it follows quite evidently that I exist.
I could not but judge that something which I understood so clearly
was true; but this was not because I was compelled so to judge by
any external force, but because a great light in the intellect was
followed by a great inclination in the will, and thus the sponta-
neity and freedom of my belief was all the greater in proportion to my
lack of indierence. (CSM II , AT VII , SPW ; emphasis
added)
Here it seems that Descartes is saying that when I clearly and distinctly
perceive the cogito (or, by implication, any clear and distinct proposition),
(a) I cannot do otherwise than assent to it, but (b) this assent is perfectly
and maximally free. Our problem is: how can this possibly be true? Isnt
it just incoherent?
4. The Coherence of Cartesian Freedom

Descartes could give at least two dierent responses. To understand the
rst response, it may be helpful to rst consider rst what he thinks
about choosing good, as opposed to evil or bad, things and actions. His
Meditation IV
position is that when one clearly and distinctly perceives X to be better

than Y, one has no choice but to select X. He rejects the possibility of
what philosophers call weakness of will (akrasia), which can be dened
as choosing something that one knows to be bad. As he puts it:
Since our will works so as to pursue or avoid only what our intel-
lect represents as good or bad, we need only judge well in order to
act well, and to judge as well as we can to do our best. (CSM I ,
SPW , AT VI )
If I always saw clearly what is true and good, I should never have to
deliberate about the right judgment or choice; in that case, although
I should be wholly free, it would be impossible for me ever to be in
a state of indierence. (CSM II , SPW , AT VII )
It seems to me certain that a great light in the intellect is fol-
lowed by a great inclination of the will; so that if we see very
clearly that a thing is good for us, it is very dicultand, on my
view, impossible, as long as one continues in the same thought
to stop the course of our desire. (CSMK , AT IV )
If we saw it [the sinfulness of some contemplated action] clearly,
it would be impossible for us to sin, as long as we saw it in that
fashion; that is why they say that whoever sins does so in igno-
rance. (CSMK , AT IV )
Notice how, in the second passage, Descartes treats a clearly known truth
in the same way as a clearly known gooda point to which we shall
return shortly.
Descartess view that when X is clearly known to be better than Y or,
more simply put, when X is good, one cannot but choose X, has a noble

This passage is from Descartes Discourse on the Method, originally written in French. I
have altered Cottinghams translation a little. Descartess highly idiomatic text says: dautant
que notre volont ne se portant suivre ni fuir aucune chose, que selon que notre entende-
ment la lui reprsente bonne ou mauvaise, il sut de bien juger, pour bien faire, et de juger le
mieux quon puisse, pour faire aussi tout son mieux. Cottingham translates ne se portant
suivre ni fuir as tends to pursue or avoid. That is too weak, insofar as tends suggests
or implies a mere tendency rather than an invariable way of operating, which is why I have
used the locution works in such a way as to pursue or avoid. A completely literal translation
of Descartess words would be, (Inasmuch as) our will carrying (carries) itself so as neither
to pursue nor to avoid anything except what our intellect represents to it as good or bad, it
suces to judge well in order to act well, and to judge as well as one can to also do ones best.
Descartes
lineage in the thought of Plato and Aristotle. The thought that if one
really knows that X is goodthat is, that X is the best of the available
alternativesone cannot fail to pursue X, is a powerful one. Descartes
would support it by holding that when one pursues the bad rather than
the good, this can only be because one is prevented from clearly and
distinctly perceiving the good by inattention, emotion, or passion, or
the likebecause, as the saying goes, one is blinded by these. But on
the other hand, the thought that a person may fail to pursue something
that she knows to be bestthat is, the thought that there is such a
thing as weakness of the willseems to be borne out by a common ex-
perience, namely, that of seeing people who we think know better
make bad choices. We shall not delve into the controversial question of
whether or not there is such a thing as weakness of the will. Rather, the
important point for our purposes is that even if knowing that X is good
inevitably leads to choosing Xthat is, even if there is no such thing as
weakness of willthis does not mean that the choice of X was unfree.
The mere knowledge that X is good obviously has no bearing on free-
dom or lack thereof. But adding that this knowledge inevitably leads us
to choose X (whether this addition is correct or not) does not seem to
render the choice of X unfree. To take a parallel case, most epistemolo-
gists hold that S knows that p entails S believes that p; it does not
follow that when someone knows that p, her belief that p or assent to p
is unfree, forced, or coerced.
This point bears on the issue of whether the assent-compellingness of
clear and distinct perceptions is compatible with free assent to them. For
as Margaret Wilson notes, there is in Descartess thought an assimila-
tion of truth and goodness. This assimilation comes out in Descartess
expressions, reasons of truth and goodness, what was true and good
(CSM II , SPW , AT VII ), what is true and good (CSM II , SPW
, AT VII ), wholly true and good (CSM II , SPW , AT VII ).
Thus, just as he thinks that the fact that a clear and distinct perception of
the good impels the will to choose the good does not entail that the choice
is unfree, so he thinks that the fact that a clear and distinct perception of
a true proposition is assent-compelling does not entail that the assent is
unfree. Even apart from the assimilation of truth to goodness, this latter
point seems right. Suppose that you are faced with a choice between
assenting to + = , + =, and + = , and that at time t

Margaret Wilson, Descartes, p. .
Meditation IV
you clearly and distinctly perceive + = . Then do you really have any
alternative at time t to assenting to that propositioncan you really do
otherwise? It seems not. Yet it also seems wrong to say that your assent
is unfree; on the contrary it seems to be, as Descartes would say, com-
pletely spontaneousa kind of eortless acquiescence with the obvious.
This arithmetical example gives rise, however, to a possible objection
that we should address before turning to the second response that Des-
cartes could give to the incoherence charge. Descartes considers ones
grasp or understanding of + = to be an act of the intellect, and ones
assent to + = to be an act of the will. But, as a number of commenta-
tors have said, anyone who understands what a simple mathematical
proposition like + = means thereby also sees that it is true, so there
does not seem to be any role for a distinct act of the will here. Bernard
Williams puts the point this way:
Clearly and distinctly to understand the proposition about the

angles of a triangle [that they must equal two right angles], for
example, is to see that it is necessarily true. But once this step
has been taken ... the theory of assent is in diculty. For if in
this sense I clearly and distinctly understand a proposition
that is to say, I can see that it is truethere is nothing else I have
to do in order to believe it: I already believe it. The will has
nothing to do which the understanding has not already done.
To answer this objection on Descartess behalf, we need to start with a

point about judgment. A judgment has two distinguishable compo-
nents: a propositional content and a propositional attitude toward that
content. For example, in the judgment, the cat is hiding under the table,
the propositional content is the cats hiding under the table, and the prop-
ositional attitude is that of arming this content. Now for Descartes,
the intellect is nothing but the capacity to grasp propositional content
(and also mere concepts like cat), and the will is simply the capacity to
arm or deny propositional content (and also to pursue or avoid per-
ceived goods and bads). Since we do in fact grasp propositional con-
tents as well as arm or deny them and pursue or avoid things, we do

Bernard Williams, Descartes: The Project of Pure Enquiry, p. .

The analysis of the role of intellect and will in judgment that follows is due to David
M. Rosenthal, Will and the Theory of Judgment, pp. .
Descartes
have those capacities, and this is arguably all that Descartes means by
saying that we have the faculties of intellect and will. Now in the case of
a clearly and distinctly perceived proposition, it is not the case that one
of those capacitiesthe capacity to arm or denyceases to function,
since we do arm the proposition. Nor is it the case, of course, that the
other capacity, of understanding the proposition, ceases to function,
since we cannot arm or deny something that we do not understand.
Rather, the activation of the capacity to understand the content is inevi-
tably accompanied by the activation of the capacity to arm it. Consider
an analogy. Suppose that Henri is a uent French speaker with normal
and currently unimpeded hearing and intelligence. Then if Henri dis-
tinctly hears the spoken words le chat se cache sous la table, he also
immediately and invariably understands what is saidthat the cat is
hiding under the table. But it does not follow that his capacity to hear
and his capacity to understand are identical, or that only one of them is
operating, or that one of them is superuous, or anything of the sort. It
will not do to respond that there are no possible worlds in which one
understands + = but withholds or denies it, for there are also no pos-
sible worlds in which a uent French speaker with currently unimpeded,
normal hearing and intelligence distinctly hears le chat se cache sous la
table but fails to understand what is being said.
Williamss objection seems to turn on a misleading personication of
the intellect and the will that Descartess language (and even common
language) encourages. Let us explain. The act of assenting to a clear and
distinct perception is, from a phenomenological (experiential) point of
view, a single act: we do not have an experience of understanding + =
, and another, distinct experience of assenting to it. Further, this single
act calls for only only one agent; indeed if there were two agents
involvedan understander and an assenterthen there would no
assent to the proposition that + = . Now obviously understanding
must be involved, since we cannot arm something that we do not
understand. So it may seem, just as Williams thinks, that any act of the
will would be superuous. But this line of thinking wrongly treats the
understanding and the will each as agents or actors. If they were, then
only one such agent would be needed and, since clearly the understanding
is needed, the will would be otiose. The truth of the matter, however, is
that the agent is a person or a mind, and that the will and the under-
standing are not agents but capacities of that person. There is nothing
suspect in saying that when a person assents to a proposition, both of
Meditation IV
these capacities are activated, and that in the special case of a clearly and
distinctly perceived proposition, the activation of the one necessitates
the activation of the other.
We can now turn to the second response that Descartes could give to
the charge that he cannot coherently hold that assenting to a clearly and
distinctly perceived proposition can be a free act if we cannot do other-
wise. This response turns on the interpretation of the principle of alter-
native possibilities. It is commonly assumed that this principle means
that for an act X performed at a time t to be free, it must be possible for
the agent to refrain from performing X at that very time, t. This entails
that for the act of assenting to a clearly and distinctly perceived proposi-
tion p a time t to be free, it must be possible for the assenter to deny or
withhold p at that very time, t. To be sure, some commentators have
rightly emphasized that it is only at the moment of illuminationonly at
the moment that one clearly and distinctly perceives pthat one cannot
but assent to p, and they also seem to think that this somehow mitigates
the incoherence of holding both that such assent is free and that it vio-
lates the principle of alternative possibilities. But, so far as this writer is
aware, no commentator has gone so far as to suggest that, at least in
Descartess eyes, the principle of alternative possibilities is still satised
in such a case. Instead, the fact that at the moment of illumination one
cannot withhold or deny p is seen as showing that Descartes is not fully
committed to that principle, since the principle is supposed to mean that
assent to a clearly and distinctly perceived proposition can be free only if
one can deny or withhold the proposition at the very moment of illumi-
nation. But it is not necessary to place such a strict interpretation on the
principle, nor does it seem fair to Descartes. For Descartes agrees, in fact
he asserts, that a proposition that could be clearly and distinctly per-
ceived can be denied because of a lack of concentration, inattention, con-
fusion, or any other factor that prevents one from perceiving it clearly
and distinctly. On any such occasion, one can do otherwise than to assent
to the proposition. Descartes claims only that at the moment that one
clearly and distinctly perceives a proposition, one is unable to refrain
from accepting it as true. He also says in several places that it is dicult
to keep ones focus on a clearly and distinctly perceived truth, and that
the moment that ones focus shifts or that one is distracted, one is no
longer impelled to assent to the proposition, and can even doubt its truth

See, for example, Cottingham, A Descartes Dictionary, pp. .
Descartes
(CSM II , AT VII ; CSM II , AT VII ; CSMK , AT IV ).

It seems perverse, then, to insist that for assent to a proposition to con-
form to the principle of alternative possibilities, one must be able to
reject or withhold that proposition even when one perceives it clearly
and distinctly. Furthermore, as Descartes emphasizes, arriving at a clear
and distinct perception often requires careful and sustained eort and
thought, in which one freely focuses ones attention on reasons and evi-
dence, and further exercises ones freedom in deliberating upon them, as
exemplied in the Meditations themselves. Is it reasonable to say, as the
usual interpretation of the principle of alternative possibilities would
make us say, that the fruit of such eortsnamely, a clear and distinct
perception in which a great light in the intellect [is] followed by a great
inclination in the willis an unfree or coerced assent to a truth?
Consider an analogy. Suppose that a skilled diver is poised at the end
of a diving board. After much training and practice, after taking position
and carefully rehearsing the dive in her mind, she dives. Of course, from
that point on she has no alternative but to fall into the water. Does this
mean that her dive, or anything that she did before, during, or after her
dive, was unfree or coerced? It seems not; for she had the alternative of
refraining from diving, which is what matters to her freedom. Now sup-
pose that, say, a mathematician works long and hard to arrive at the
proof of a theorem, discovers the proof, and then keeps the whole proof
in mind and perceives that the theorem follows from the premises so
clearly and distinctly that he cannot but assent to the theorem. Does this
mean that this assent is unfree? It seems not. At most, someone might
complain that calling the assent an act is misleading, since it suggests
that there is some act of assenting to the theorem that occurs later in
time than grasping the proof. But it is clear that the mathematician does
assent to or mentally arm the theorem, and one may be hard pressed to
improve on the Cartesian language that calls this assent an act of the will.
We propose, then, the following interpretation of Descartess posi-
tion: when Descartes says that our assent to a clear and distinct percep-
tion is free, he intends to include the thought that we can do otherwise
than assent to it: he fully accepts the principle of alternative possibilities.
But for him that principle entails only that we can withhold or deny the
relevant proposition except at the very time when are perceiving it clearly
and distinctly. At all other times, such as when we are working toward
attaining a clear and distinct perception of it, or about to perceive it
clearly and distinctly, or no longer perceiving it clearly and distinctly, or
Meditation IV
incapable of perceiving it clearly and distinctly, and so forth, we can cer-

tainly withhold or deny it, and for Descartes that suces to satisfy the
principle of alternative possibilities. One might call this the longitudi-
nal, as opposed to the cross-sectional, interpretation of the principle
of alternative possibilities.
5. Descartess Troubling Letter to Mesland

There is a letter in Descartess correspondence that may seem to clash
with our interpretation. This is his much-discussed letter of February ,
to the Jesuit priest, Denis Mesland, where he says:
Indierence ... seems to me strictly to mean that state of the

will when it is not impelled one way rather than another by any
perception of truth or goodness. This is the sense in which I took
it when I said that the lowest degree of freedom is that by which
we determine ourselves to things to which we are indierent. But
perhaps others mean by indierence a positive faculty of deter-
mining oneself to one or other of two contraries, that is to say, to
pursue or avoid, to arm or deny. I do not deny that the will has
this positive faculty. Indeed, I think that it has it not only with
respect to those actions to which it is not pushed by any evident
reasons on one side rather than on the other, but also with
respect to all other actions; so that when a very evident reason
moves us in one direction, although morally speaking we can
hardly move in the contrary direction, absolutely speaking we
can. For it is always open to us to hold back from pursuing a
clearly known good, or from admitting a clearly perceived truth,
provided we consider it a good thing to demonstrate the freedom
of our will by so doing. (CSMK , AT IV )
At least four things seem to be happening in this passage. First, Des-

cartes reminds Mesland that on Descartess own notion of indierence,
it applies just to exercises of the lowest degree of freedom, where no

As Anthony Kenny points out, there is some doubt as to the date and the addressee
of this letter (Descartes on the Will, pp. ). But so far as I know, no one doubts that
the author of the letter is Descartes.
Descartes
compelling reasons drive us to one choice rather than another. Second,

he recognizes that other people mean something very dierent by indif-
ference, namely, simply the two-way power to chose between opposites.
Thirdand this is where the trouble beginshe asserts that absolutely
speaking, though not morally speaking, we have this two-way power
even when a very evident reason (presumably, a clear and distinct per-
ception) moves us toward one of two opposite choices. Fourthand this
conrms the troublehe asserts that we can always refuse to go along
with a clear and distinct perception, in order to demonstrate our free-
dom of will. The last two points seem atly incompatible with saying that
we cannot refrain from assenting to clear and distinct perceptions at
times when have them.
Let us take up these two points in reverse order. Of course, Descartess
claim that we can always hold back from assenting to a clearly and dis-
tinctly perceived truth can be taken to mean that we can do so even at the
very moment, or during the very span of time, that we perceive it clearly
and distinctly. It would be silly to deny that Descartess words lend them-
selves to this interpretation. But the question is whether they must, or
should, be read that way; for there is certainly an alternative reading,
namely, that Descartes means only that we can always turn our careful
attention away from a truth that we perceive clearly and distinctly, and
that then we are no longer impelled to assent to it.
Now, there are at least three excellent reasons for adopting this alter-
native reading. First, it no longer saddles Descartes with contradicting
his clearly stated position in the Meditations, Principles of Philosophy, Ob-
jections and Replies, Correspondence and elsewhere that clear and distinct
perceptions are assent-compelling. After all, Descartes was if anything a
highly subtle thinker, and in answering his critics he was also quite ready
on occasion to appeal to straightforward common-sense distinctions,
such as the distinction between denying or withholding p at the very
same time that one clearly and distinctly perceives p, and denying or
withholding p at times when ones attention is not focused on p. For ex-
ample, he appeals to such a common-sense distinction, albeit a dierent
one in a dierent context, when he says to Gassendi: when we say that
an idea is innate in us, we do not mean that it is always there before us.
This would mean that no idea was innate. We simply mean that we have
within ourselves the faculty of summoning up the idea (CSM II ,
SPW , AT VII ). Likewise, then, Descartes may be interpreted as
saying to Mesland that when we say that we are free to deny a clearly and
Meditation IV
distinctly perceived proposition, we do not mean that we can deny it

even when we are clearly and distinctly perceiving it, for this would mean
that no proposition was assent-compelling. Rather, we simply mean that
we have the power to turn our careful attention away from it, and that
then we can deny it. That is both coherent and commonsensical. Why
then should one insist on attributing to Descartes a contradiction, when
there is an equally natural and perfectly straightforward way of reading
his words that does not run into contradiction?
Second, Descartess idea that we can demonstrate the freedom of our
will by holding back from pursuing a known good or assenting to a
clearly distinctly perceived truth (in our proposed sense of turning our
careful attention away from these) makes good sense, for dissenting
from a proposition that had only some evidence in its favor, or eschewing
pursuit of something that had only limited value, would not be as clear a
manifestation of our freedom of will as holding back from pursuing a
good or assenting to a proposition that we remembered having recently
very clearly and distinctly perceived. Indeed, by Descartess lights this
would appear to be the strongest possible demonstration of our freedom!
By contrast, the notion that we can best demonstrate our freedom of will
by withholding or denying a proposition at the very same time as we are
perceiving it clearly and distinctly seems absurdly over-demanding.
Third, as Naaman-Zauderer acutely says,
Descartes could not have abandoned his view that clear and dis-
tinct perceptions compel assent without undertaking a funda-
mental revision of his most basic epistemological doctrines. One
is his conception of error ... Had Descartes believed that we can
dissent from a clearly and distinctly perceived truth, he could not
have argued that it is quite impossible for us to go wrong so
long as we restrict the operations of the will to the clear and dis-
tinct perceptions of the intellect.
This is of course correct, since Descartes would then have to admit, con-
tra his claim that we cannot go wrong so long as we restrict our will to
clear and distinct perceptions, that one possible way for us to go wrong
is by denying a true proposition even though we clearly and distinctly
perceive it. If we can deny a clearly and distinctly perceived proposition,

Naaman-Zauderer, Descartes Deontological Turn, p. .
Descartes
then we can deny a true clearly and distinctly perceived proposition,

and so we can go wrong even with regard to something that we clearly
and distinctly perceive. Our only complaint against Namaan-Zauderer,
then, is not that she insists on the assent-compellingness of clear and
distinct perceptions: this insistence is surely correct. Rather, it is that,
like most commentators, she thinks that this assent-compellingness is
incompatible with the principle of alternative possibilities, so that she
is then in the position of having to maintain that assenting to a clear
and distinct proposition can be free even though it violates that prin-
ciple. By contrast, on our interpretation of the principle (or of how Des-
cartes understands the principle), Descartes can maintain both the
assent-compellingess of clear and distinct perception and the principle
of alternative possibilities.
Let us turn to Descartess remark that when a very evident reason
moves us in one direction, although morally speaking we can hardly
move in the contrary direction, absolutely speaking we can. Here Des-
cartes seems to be saying that in one sense (morally speaking) we
cannot deny a clear and distinct perception but in another sense (abso-
lutely speaking) we can. He does not explain what he means by mor-
ally speaking and absolutely speaking. As Zaaman-Zauderer points
out, his distinction cannot be understood in terms of the distinction he
makes elsewhere between moral and absolute or metaphysical cer-
tainty. For moral certainty pertains to cases where there are more rea-
sons on one side than on the other, whereas in a clear and distinct
perception, all the reasons are on one side only. So, cases of moral cer-
tainty are not analogous to cases of being morally speaking unable to
deny a clear and distinct perception, because in cases of moral certainty
we are able to go against the preponderance of reasons and elect the
weaker alternative, whereas in cases of being morally speaking unable
to deny a clear and distinct perception, the reasons are conclusive and
we are unable to go against them; thus it remains unclear what Des-
cartes means when he says that absolutely speaking, we can deny a
clear and distinct perception.
Naman-Zauderer considers but rejects two interpretations of what
Descartes may have meant, by Lilli Alanen and Alan Nelson. According
to Alanen, when Descartes states that moving in the contrary direc-
tion would be morally impossible for us, he means only that it would be

Naaman-Zauderer, Descartes Deontological Turn, p. n.
Meditation IV
nonrational or morally unjustied. Against this Namaan-Zauderer

objects, reasonably enough, that
[For Descartes], rationality hinges on the (negatively formulated)

duty to avoid assenting to matters we apprehend confusedly or
obscurely. Unsurprisingly, therefore, Descartes does not explic-
itly prescribe consistent assent to our clear and distinct percep-
tions while attending to them. Such a prescription is not required
because of the brute fact that psychologically or phenomenolog-
ically we cannot possibly do otherwise. So long as our attention
is focused on clearly and distinctly perceived reasons, we can be
wholly assured that we shall always assent to them.
In other words, Descartes never gives voice to the idea that we morally
ought to assent to clear and distinct perceptions, and this fact should not
surprise us since he so explicitly says that current clear and distinct per-
ceptions are irresistible, so that we have no choice but to assent to them.
According to Nelson, on the other hand, when Descartes writes that ab-
solutely speaking we can hold back from admitting a clearly perceived
truth, he is simply restating that the will can be diverted from a clear and
distinct perception. (Presumably, then, when Descartes says that mor-
ally speaking we cannot hold back from admitting a clearly perceived
truth, he simply means that we cannot do so while perceiving it clearly
and distinctly.) Against this, Naaman-Zauderer objects that
Nelsons interpretation does not t well with Descartes sugges-

tion that it is with regard to the same situation of choice that,
morally speaking, we cannot act or judge against the weight of
the evidence, whereas absolutely speaking, we can. This point
was made by Ragland: the two senses passage maintains that,
with respect to one and the same act of will, and at the same
time, we can be both morally unable to hold back and absolutely
able to hold back.



Naaman-Zauderer, Descartes Deontological Turn, p. ; quoting directly from Alan
Nelson, Descartes Ontology of Thought, p. .

Naaman-Zauderer, Descartes Deontological Turn, p. . The quote from Ragland is
from his Descartes on the Principle of Alternative Possibilities, p. .
Descartes
Having rejected both Alanens and Nelsons interpretations of

Descartess morally speaking versus absolutely speaking distinction,
Naaman-Zauderer then goes on to propose her own interpretation:
In my understanding, moral necessity or impossibility, in this

context, denotes a kind of practical or psychological impossi-
bility, whereas absolute possibility signies what is possible from
an absolute, metaphysical point of view, irrespective of our inner
experience. [Here Zaaman-Zauderer adds this footnote: Des-
cartes uses the phrase absolutely speaking (absolute loquendo) or
equivalents on other occasions as well, usually to signify the con-
sideration of things from Gods absolute point of view, as it were,
independently of any actual limitations of our minds. See, e.g.,
letter to Clerselier, April , (AT V ; CSMK );
Second Replies (AT VII ; CSM II )]. In saying that choosing
otherwise when a very evident reason moves us in one direction
is morally impossible for us, therefore, Descartes may mean that
for us to do so is psychologically or practically impossible. From
an absolute, purely metaphysical perspective, however, it is
always possible for our will to act independently of our intellect,
even in the face of clear and distinct ideas. The reason is that,
metaphysically speaking, our intellect and will remain distinct
from one another even when the intellects perceptions are clear
and distinct. Although we experience our intellect and our will as
unied at the very moment of illuminationan experience that
makes us morally (psychologically) incapable of acting against
the intellects overwhelming reasonsmetaphysically or abso-
lutely speaking we are capable of doing so.
This characterization of what is possible absolutely speaking is less

than convincing, for when Descartes says to Mesland that it is always
open to us to hold back from pursuing a clearly known good, or from ad-
mitting a clearly perceived truth, provided we consider it a good thing to
demonstrate the freedom of our will by so doing, he is surely talking
about a real experiential possibility, not merely about an abstract one
that can never be realized in experience, and that has to do only with the
metaphysical distinction between the will and the intellect. How could an

Meditation IV
act demonstrate to us our (real) freedom of the will, if that act occurred
irrespective of our inner experience and went beyond any actual limi-
tation of our minds? By contrast, our interpretation, according to which
Descartes is claiming that we can demonstrate our freedom to ourselves
by withholding even a truth that we remember having recently clearly
and distinctly perceived, gives a perfectly literal, concrete content to Des-
cartess claim. Naaman-Zauderer would presumably object, as she does
against Nelson, that our interpretation does not do justice to the claim
that it is the very same act of will that we are supposed to be morally un-
able to perform but absolutely able to perform. It seems, however, that
this objection is not really coherent. For given that an actual, single act of
the human will cannot intelligibly be said to belong both to the realm
of that human beings experience and to some abstract metaphysical
realm that exists irrespective of our inner experience and indepen-
dently of any actual limitations of our minds, it seems that the scenario
she tries to envisage could only be one in which there is a unitary act (of
holding back from assenting to a clear and distinct perception) that one
is both able and unable to performand that is atly contradictory.
Clearly, Descartes does mean to say at least that what is possible ab-
solutely speaking is not always possible morally speakingthat the
domain of what is possible absolutely speaking is broader than (though it
includes) the domain of what is possible morally speaking. Our interpre-
tation also ts well with this idea. For the class of propositions that can
be clearly and distinctly perceived is broader than, and also includes,
the class of propositions that are clearly and distinctly perceived by one
nite mind at a particular time. So when Descartes says that absolutely
speaking we can we can hold back from admitting a clearly perceived
truth, provided we consider it a good thing to demonstrate the freedom
of our will by so doing, he can plausibly be taken to mean that although
we can hold back our assent from any proposition that falls within the
metaphysically absolute totality of propositions that we could ever
clearly and distinctly perceive (even from ones that we recently clearly
and distinctly perceived), and may do so especially in order to demon-
strate our freedom of will in the strongest way possible for us, nonethe-
less we cannot at any given time do this for the members of that class
that also fall within the frequently altering sub-class of it that comprises

Here I put to one side Immanuel Kants theory that the same action can belong both
to the empirical world and to a noumenal world outside of space and time.
Descartes
only the propositions that we are perceiving clearly and distinctly at that
time.
There remains one question. In his Replies to the Second Set of Objec-
tions, as we saw in the previous chapter, Descartes says that some clear
and distinct propositions are so simple that one cannot even think of
them without clearly and distinctly perceiving them, so that, given the
assent-compellingness of clear and distinct perceptions, one can never
doubt these propositions at all:
Now some of these perceptions are so transparently clear and at

the same time so simple that we cannot ever think of them with-
out believing them to be true. The fact that I exist so long as I am
thinking, or that what is done cannot be undone, are examples of
truths in respect of which we manifestly possess this kind of cer-
tainty. For we cannot doubt them unless we think of them; but
we cannot think of them without at the same time believing that
they are true ... Hence we cannot doubt them without at the
same time believing that they are true; that is, we can never
doubt them. (CSM II , M , SPW , AT VII )
How is this to be squared with our proposal that when Descartes says to
Mesland that we can always hold back from assenting to a clear and dis-
tinctly perceived proposition, he means that we can do so provided we
turn our careful attention away from it? It seems that with regard to
these propositions, the only way that we can hold back our assent from
them is by simply not thinking about them. We suggest that this conse-
quence should be accepted, even though it may initially seems odd to say
that we are holding back our assent from a proposition simply by not
thinking about it. It might seem that to hold back our assent from a prop-
osition p, we must be thinking about p, since to doubt that p, we must be
thinking about p. But notice that holding back our assent from a clearly
and distinctly perceived proposition is not the same as doubting it. One
can hold back ones assent from it by ignoring it, or by resolving not to
think about any proposition that one can clearly and distinctly perceive,
perhaps by lling ones mind for ve minutes with thoughts about only
economics! This seems quite sucient, strictly speaking, for it to be true
that one is currently holding back from assenting to any clear and distinct
proposition, including the cogito and the principle of noncontradiction,
and that is all that Descartess position requires in order to be coherent.
Meditation IV
6. Error and Evil

People generally regard human error as something bad or evil. We try
to avoid error, and we highly value its opposite, knowledge. We erect
systems of education, and a large part of the process of education con-
sists in learning methods for avoiding error. In his Fourth Meditation,
Descartes seeks not only to diagnose the cause of error and to prescribe
the method for avoiding it, but also to show that the existence of this
particular brand of evil is compatible with the existence of a perfect God.
In that sense, the Fourth Meditation is Descartess main contribution to
the philosophical literature on the problem of evil. To put this contribu-
tion into a larger context, we shall begin with a general account of the
problem of evil; subsequently, we shall explain how Descartess views t
into this account.
6.1 The Problem of Evil

The classical problem of evil arises from the apparent contradiction
between the existence of evil and the belief in God. The problem is
that the existence of evil appears to be incompatible with the exalted
attributesomnipotence, omniscience, and supreme goodnessof
the God of Judeo-Christianity. Does evil exist because God is unable to
prevent it? Then he is not omnipotent. Does it exist because he does
not want to prevent it? Then he is not supremely good. Does it exist
because he is unaware of it? Then he is not omniscient. Thus, it seems
that while God could have any two of the attributes that are tradition-
ally taken as necessary to his perfection, he cannot have all of them.
The problem can be formulated in the following logically valid argu-
ment from evil:
() If God is omnipotent, then he can prevent evil.

() If God is perfectly good, then he wants to prevent evil.
() If God is omniscient, then he knows of any evil that exists.
() Evil exists.
() If evil exists, then either God is unable to prevent it, or he
does not want to prevent it, or he is unaware of it.
_____________________________________________________
? () God is either not omnipotent, or not perfectly good, or not
omniscient.
Descartes
Traditionally, premises () and () are omitted because they are regarded

as obvious and uncontroversial, and the last disjunct (or clause) of the
conclusion is left understood as well. Since the argument is valid, it is not
possible for all its premises to be true and its conclusion to be false. So a
theist (a believer in God) must, on pain of contradiction, either accept
the conclusion or reject at least one of the premises.
Could a theist simply accept the conclusion, that is, simply admit that
the fact of evil proves that God is limited in power or in goodness or in
knowledge? From a purely logical point of view, that is one option. Indeed,
there are religions in which God is conceived as limited or nite in power
and/or goodness. There are also religions, like Zoroastrianism, according
to which there exist both a good deity and an evil one, who are each limited
by the others power, and who are in perpetual strife. At one point in
David Humes Dialogues Concerning Natural Religion, the character Clean-
thes opts for a nite God, to which the character Philo, who is generally
taken to represent Humes own position, responds that even from a God
whose power, goodness and wisdom were limited, an unbiased judge
would expect a much better world than the actual world. Be that as it
may, the key point to recognize is that to accept the arguments conclu-
sion would be to give up the belief in God as he is conceived in Judeo-
Christianity. St. Anselm dened God in the twelfth century as the being,
than which none greater can be conceived, and Descartes denes him as
a supremely perfect being. These denitions are not arbitrary: they answer
to the concept of the God worshipped by Christians and Jewsa perfect
God whose greatness is unsurpassable, and who therefore cannot be even
a little bit limited in power, knowledge, or goodness. So even though the
argument from evil is not explicitly directed against the existence of God,
but purports only to prove that God cannot have all the divine attributes,
its conclusion is tantamount to showing that God (at least the God of
Judeo-Christianity) does not exist. Indeed the argument could be turned
into an attempted disproof of his existence, just by adding the premise
() If God exists, then he is omnipotent, perfectly good, and

omniscient.
Then it follows from () and () that
() God does not exist.

David Hume, Dialogues Concerning Natural Religion, pp. .
Meditation IV
Since theists (henceforth meaning by that term, believers in the

Judeo-Christian God) cannot accept the arguments conclusion, (), they
must deny at least one of its premises. Let us consider rst the possibility
of denying (), the premise that evil exists. What could such a denial
mean? It might mean that (a) the things we call evil, such as pain, suf-
fering, cruelty and so forth, do not exist, or (b) that these things exist,
but are not really evil. Now (a) can be immediately dismissed, since it
simply ies in the face of facts. Besides, we would then have to admit that
at least the illusion of pain, suering, and cruelty exist. But isnt that
already an evil? Perhaps the existence of the illusion of pain, suering,
and cruelty isnt as bad as real pain, suering and cruelty, but surely the
absence of both real and illusory pain, suering and cruelty would be still
better. As for option (b), it fares no better. For if the things we consider
to be evil, such as pain, suering and cruelty, do not count as evil, then
what would? If the answer is that nothing would even count as evil, then
the term evil signies nothingit is meaningless. But then so is its
opposite, good, in which case it is meaningless to say that God is good.
The point here is not that, as people sometimes say, there cannot be good
if there is no evil; there is no reason to accept that claim. Rather, the
point is that if nothing would even count or qualify as being evil, then the
contrast between good and evil vanishes at the conceptual level, so that
it becomes meaningless to call anything either good or evil.
Can theists then reject premise ()that if God is omnipotent, then
he can prevent evil? It may seem that they cannot, because by denition
an omnipotent being can do anything. In other words, it may seem that
() can be deduced from two analytic truths, as follows:
(i) If God is omnipotent, then he can do anything.

(ii) If God can do anything, then he can prevent evil.
__________________________________________
? () If God is omnipotent, then he can prevent evil.
However, matters are not quite so simple, for (i) would commit us to
saying that God can perform acts that involve a logical contradiction, for
example, construct a four-sided triangle, or cause the mercury in a ther-
mometer to be both one inch and two inches way from the bottom at the
same time. Now as we have seen, Descartes did hold that an omnipotent
could make contradictories true, so he would not be able to answer the
argument from evil by objecting to (i). But, quite aside from the fact that
Descartess doctrine of the creation of the eternal truths would threaten
Descartes
his doctrine that the human will closely resembles Gods will, it seems
quite incoherent. Aquinass view that nothing that implies a contra-
diction falls under the scope of Gods omnipotence seems to be the
coherent view of omnipotence. It is worth pointing out, therefore,
that () cannot really be avoided by objecting to (i), for (i) can easily
be qualied so as to accommodate Aquinass view of omnipotence, as
follows:
(ia) If God is omnipotent, then he can do anything that it is logi-

cally possible for him to do.
Presumably it would have been logically possible for God to prevent evil,
so the argument for () can be amended by also qualifying (ii) to read this
way:
(iia) If God can do anything that it is logically possible for him to

do, then he can prevent evil.
From (ia) and (iia), () follows as before.

So far, we have argued that a theist is committed to accepting premises
() and (). Premises () and () appear to be necessarily true, which is
presumably why they are usually left understood; so the argument is
commonly given an abbreviated formulation, which we shall adopt in the
rest of our discussion:
() If God is omnipotent, then he can prevent evil.

() If God is perfectly good, then he wants to prevent evil.
() Evil exists.
___________________________________________________
? () Either God is not omnipotent or he is not perfectly good.
It appears, then, that the only way to avoid the arguments conclusion is
to reject premise (). In other words, the theist must show that just
because God is perfectly good, it does not follow that he wants to prevent
all evil. This is indeed the main way that theists have tried to cope with
the problem of evil: they have tried to show that the existence of evil is
not incompatible with Gods perfect goodness. This endeavor is called
theodicy, and Descartess Fourth Meditation is, in part, a contribution to
theodicy.
Meditation IV
The rst step in developing a theodicy is to draw a distinction between

two kinds of evil: moral evil and nonmoral (or natural) evil. Moral evil
is evil that results from human beings (and, some would add, angels)
wickedness or wrongdoing. Nonmoral evil is evil that does not result
from human wickedness, but from causes like disease, plagues, earth-
quakes, tsunamis, droughts, and so forth. The problem for the theodicist
then divides into two questions: () Why does a perfectly good God who
is omnipotent allow the existence of moral evil, and () Why does such a
God allow the existence of nonmoral evil?
Theodicys traditional response to the rst question, commonly called
the free-will defense, starts from the point that God, in his goodness, cre-
ated a host of creatures possessing free will. Being omnipotent, he could of
course have instead created only automata who always behaved as he wished
them to. But this would not have been as good on his part as creating people
with free will because, as the philosopher of religion John Hick puts it, only
persons [i.e., beings endowed with free will] could, in any meaningful sense,
become children of God, capable of entering into a personal relationship
with their creator by a free and uncompelled response to his love. How-
ever, the theodicists response continues, the freedom of creatures has a
price: it entails that there may well be moral evil. For free creatures cannot
be guaranteed to always act rightly; rather, just because they are free, they
may at times perform hurtful, wicked, morally wrong actions.
A natural objection to this line of thought is that the freedom of crea-
tures implies the possibility of moral evil, but not its actuality. Why did
God, if he is perfectly good as well as omnipotent, not create people who
always freely choose to do what is morally right? In response, the theodi-
cist should concede that freedom implies only the possibility of moral
evil, not its actuality. Just because a creature is free to perform morally
wrong actions, it does not follow that it will perform any such actions;
conceivably, it might refrain from performing any such actions through-
out its entire life. The picture changes, however, when the anti-theodicist
adds that a perfectly good God would have created people who always
freely choose to do what is right. For this amounts to saying that God
would have caused or determined his creatures to freely choose only right
actions, which is contradictory. If God causes or determines us to choose
certain actions, then those actions cannot be completely free. Rather, we
are like puppets, or people acting under post-hypnotic suggestion.

John Hick, Philosophy of Religion, p. .
Descartes
The theodicist concludes that it is no reection on Gods perfect good-

ness that moral evil exists. For the only way that God could have pre-
vented it from existing would be by depriving his creatures of free will (or
at least of a large measure of free will, since they could never choose mor-
ally wrong actions, but only choose among morally right ones), which
would have undercut the possibility of moral goodness on their part and
of their entering into a meaningful relationship with him. To put is dif-
ferently, it is true that our freedom implies only the possibility of moral
evil, not its actuality. But our freedom also implies that God is not re-
sponsible for any moral evil that actually occurs, for to hold him respon-
sible for it would be to say that he should have prevented it. But if he had
prevented it, he would also have deprived his creatures of their freedom,
or signicantly reduced it.
The free-will defense, even if successful, provides only a partial solu-
tion to the problem of evil, for it leaves untouched the problem presented
by nonmoral evil; that is, by physical suering and mental anguish that
do not result from human wickedness or wrongdoing. At best, then, the
free-will defense forces a religious skeptic to reformulate the argument
from evil only in terms of nonmoral evil, as follows:
() If God is omnipotent, then he can prevent nonmoral evil.

() If God is perfectly good, then he wants to prevent nonmoral
evil.
() Nonmoral evil exists.
_____________________________________________________
There are two dierent traditional responses to this argument. The rst,
which stems from St. Augustine ( a.d.) is directed against pre-
mise (); the second, which stems from Irenaeus ( a.d.), is
directed against premise (). We shall consider them in turn.
The Augustinian theodicy holds, basically, that there is no such thing
as nonmoral evil; the only kind of evil that exists is moral evil, which is a
kind of turning away from God. Augustines denial of nonmoral evil rests

The discussion that follows in this section is strongly inuenced by the writings of
John Hick, cited in notes and . For a penetrating critique of Hicks Irenean theodicy,
see Edward H. Madden and Peter H. Hare, Evil and the Concept of God, especially pp. ,
, .
Meditation IV
on two pillars: his privative view of evil, and his aesthetic view of evil. The
privative view is that nonmoral evil is not something real or positive;
instead it is merely an absence or lack of perfection. According to Augus-
tine, everything that exists is good, but some things are better or more
nearly perfect than others. Most obviously, God, the creator, is perfect,
but created things are less than perfect; if they were perfect, they would
not be any dierent from God. Thus it is in the very nature of created
things to be imperfect, but this does not mean that anything that God
created is positively evilall things are good. This view, which comes
from Plotinus and ultimately from Plato, equates being with goodness.
It implies that just as there are degrees of goodness, there are degrees
of being, and that just as nonbeing is unreal, likewise nonmoral evil is
unreal.
The aesthetic view of nonmoral evil, which complements the privative
view and also derives from Plotinus, is that the things that we call evil
only appear to us as evil because we do not understand them in relation
to the whole creation. If we were to see them in relation to the whole,
then we would appreciate the fact that every part of creation, no matter
how it appears from our limited human perspective, contributes to the
goodness of the whole. As an analogy, consider a single measure of a Bach
fugue, heard in isolation from the rest of the composition. It can be dis-
sonant and hence ugly, but it may be an indispensable element for en-
hancing or maximizing the beauty of the whole piece. Likewise, a single
patch of color on a painting might utterly lack aesthetic appeal, but sig-
nicantly enhance the beauty of the whole painting.
We shall say more about the Augustinian theodicy when we return to
Descartes, but rst let us turn our attention to the Irenean theodicy, espe-
cially as defended by its modern advocate, John Hick. Irenean theodicy
does not deny the existence of nonmoral evil. Rather, it opposes premise
() of the above argumentthe premise that if God is perfectly good, then
he wants to prevent nonmoral evilon the ground that the nonmoral evil
that exists is required for the purpose of bringing about a greater good
than would be possible without it. Specically, there are certain valuable
human virtues that could not be developed or manifested in a world lack-
ing all nonmoral evil, such as courage, perseverance, sympathy, charity,
protectiveness, and so forth. In a world devoid of any adversity or obsta-
cles, there could never be any opportunity to exercise any of these virtues.

Wallace I. Matson, The Existence of God, p. .
Descartes
This line of thought can be supplemented by attacking what seems to

be an assumption behind premise (). This is the assumption that a per-
fectly good Gods purpose in creating the world must have been to pro-
duce a pleasure paradise for humans to inhabit, akin to building a cage
for some favorite pet animal. In other words, his aim, since he is per-
fectly good, must have been to provide humans with as safe, pleasurable,
and nondemanding an environment as possible. Irenean theodicys
response to this assumption is that it misrepresents Gods purpose. His
true purpose was to create a world of soul-making, in which creatures
would develop and exercise the nobler virtues, and for this purpose it is
essential that there be hardships, obstacles, and adversity to overcome.
The nobler virtues could not get a foothold if nonmoral evils did not
exist.
How might Irenean theodicy, as so far described, be criticized? Well, a
critic can point out that although it is true that sometimes nonmoral evil
breeds virtues like courage, patience, compassion, steadfastness, and so
on, sometimes it does just the opposite. Some people crumble under the
weight of adversity; some become permanently embittered, or even go
insane. Thus, a critic could reformulate the argument from evil this way:
() If God is omnipotent, then he can prevent nonmoral evil that

does not help to produce a greater good.
evil that does not help to produce a greater good.
() Nonmoral evil that does not help to produce a greater good
exists.
_____________________________________________________
In his book Evil and the God of Love, John Hick attempts to refute
even this carefully qualied version of the argument. He does so by ap-
pealing to one of the most characteristic beliefs of Christianity, namely
the belief in life after death. Hick argues that although the evils that
we encounter during life on earth do not all contribute to a greater
good, they may well do so if we take into consideration the afterlife. He
holds that the belief in an afterlife in which the evils suered in this
world ultimately lead to a greater good should be an intrinsic part of

John Hick, Evil and the God of Love, pp. .
Meditation IV
the Irenean theodicy that he favors. Notice that for the purpose of an-
swering the argument from evil, it is not necessary to arm that there
will be an afterlife. All that is necessary, strictly speaking, is to hold that
there may be an afterlife or, more accurately put, that it is not unrea-
sonable or irrational to arm this. For this maneuver forces the argu-
ment from evils proponent to qualify the argument once again, so that
it goes as follows:
() If God is omnipotent, then he can prevent nonmoral evil that

does not help to produce a greater good in either this life or the
afterlife.
evil that does not help to produce a greater good in either this life
or the afterlife.
() Nonmoral evil that does not help to produce a greater good in
either this life or the afterlife exists.
_____________________________________________________
However, at this point the theodicist could claim that the argument form
evil has become unreliable, because nobody has adequate evidence for
premise (); nobody can arm with justiable condence that there is
not an afterlife during which whatever nonmoral evil we endured in our
earthly life will lead to a greater good. Thus, the theodicists best hope
may lie in forcing the religious skeptic to quality premise () of the argu-
ment from evil to the point where premise () becomes so dubious that
the argument becomes unreliable.
To be sure, there are possible responses that a skeptic can still make.
One is that some nonmoral evils suered in this life are so crushing that
the mere memory of them would ruin or at least damage the quality of
the after-life. Another is to point to the fact of animal suering: non-
human animals, too, are sentient, conscious beings, and multitudes of
them suer great pain and distress. Why would a perfectly good God
allow this? Are we to believe that animals, too, have an afterlife in which
their past suering leads to a greater good? Yet another response would
be to argue that the belief in an afterlife is unreasonable. After all, we
know that when certain parts of the brain are damaged, corresponding
mental functions and forms of consciousness cease to occur. Is it not
highly probable, then, that when brain activity stops altogether, so does
Descartes
all mental functioning and consciousness? We leave further reection on

these questions to the reader.
6.2 Cartesian Theodicy

Against the background the general problem of evil, let us turn to Descartess
attempt to show that the existence of error is compatible with the existence
of an omnipotent, perfectly good, omniscient God. First, it is obvious that
Descartes treats human error in the same way as traditional theodicy treats
moral evil, for he applies the free-will defense to it. Human error results
from a certain misuse of our free will, namely that of making judgments re-
garding matters that we do not perceive clearly and distinctly, and it can be
avoided by restricting our judgments to matters that we do perceive clearly
and distinctly. Moreover, Descartes treats erroneous judgments in tandem
with morally bad choices, and his language implies that he regards mistakes
about what is true or false as morally blameworthy, as when he says that
even if through sheer luck I make a true judgment about something that I do
not perceive clearly and distinctly, I shall still be at fault since it is clear by
the natural light that the perception of the intellect should always precede
the determination of the will (CSM II , SPW , AT VII ).
Descartes also appeals to the privative view of evil: not without prece-
dent, he extends the privative view beyond nonmoral evil to (what he
regards as a species of) moral evil. Thus near the beginning of the Fourth
Meditation, he says
I realize that I am, as it were, something intermediate between

God and nothingness, or between supreme being and non-being:
my nature is such that in so far as I was created by the supreme
being, there is nothing in me to enable me to go wrong or lead me
astray; but in so far as I am not myself the supreme being and am
lacking in countless respects, it is no wonder that I make mis-
takes. I understand, then, that error as such is not something
real which depends on God, but merely a defect. Hence my going
wrong does not require me to have a faculty specially bestowed
on me by God; it simply happens as a result of the fact that the
faculty of true judgment which I have from God is in my case not
innite. (CSM II , SPW , AT VII )
Here Descartes is preparing the way for his view that neither of the fac-
ulties used in making judgments, the intellect and the will, have error
Meditation IV
built in to them, so that they are in no way defective. But the reason he
gives for this, that error is not something real, expresses the privative
view of evil.
In the next paragraph, however, Descartes nds that this appeal is
insucient:
But this is not entirely satisfactory. For error is not a pure nega-
tion, but rather a privation of lack of some knowledge which
somehow should be in me. (CSM II , SPW , AT VII )
Here Descartes makes a distinction between a mere negation or lack,

and a privation. A mere lack is the absence of a desirable characteristic
that a thing does not have because of its very nature; a privation is the
absence of a desirable characteristic that thing could have had while
keeping the same nature, and in that sense ought to have. For ex-
ample, the absence of eyes in certain deep-sea sh that live in total
darkness is a lack but not a privation, but the absence of functioning
eyes in a human being would be a privation. We say that a sightless
deep-water sh lacks sight but not that it is deprived of sight, but we say
of a human being who lacks functioning eyes that she is deprived of
sight and suers from blindness. Now, Descartes asks, why isnt the
occurrence of human errors a privation rather than a mere lack? His
answer is that, apart from the fact that we cannot fathom Gods every
purpose,
Whenever we are inquiring whether the works God are perfect,

we ought to look at the whole universe, not just one created thing
on its own. For what would perhaps rightly appear very imperfect
if it existed on its own is quite perfect when its function as part of
the universe is considered. (CSM II , SPW , AT VII )
This is essentially an appeal to the aesthetic conception of evil: parts of a

whole that may have little or no value in isolation may well contribute
signicantly to the goodness of the whole. Thus, in his attempt to recon-
cile the existence of human error with Gods perfection, Descartes ap-
peals not only to the free-will defense, but also to both strands of the
Augustinian theodicy, the privative view of evil and the aesthetic view of
evil.
Even if these appeals can solve the theodicean problem caused by the
actual occurrence of human errors, there remains a question: why did a
Descartes
perfect God create humans who are even prone to error? Why did he not,
instead, give us a nature such that we never go wrong? As Descartes
puts it:
There is ... no doubt that God could have given me a nature such
that I was never mistaken [since he is omnipotent]; again, there
is no doubt that he always wills what is best [since he is perfectly
good]. Is it then better that I should make mistakes than that I
should not do so? (CSM II , SPW , AT VII )
God could easily have brought it about that without loosing my

freedom, and despite the limitations in my knowledge, I should
nonetheless never make a mistake. He could, for example, have
endowed my intellect with a clear and distinct perception of
everything about which I was ever likely to deliberate... . Had
God made me this way ... I would have been more perfect than I
am now. (CSM II , SPW , AT VII )
As these passages show, Descartes readily admits that God could have
made humans incapable of error simply by giving them an understanding
with as great a scope as their wills. So, why did he not do so?
Desacartess response to this question is a straightforward appeal to
the aesthetic conception of evil:
I cannot ... deny that there may in some way be more perfection
in the universe as a whole because some of its parts are not im-
mune from error, while others are immune, than there would be if
all the parts were exactly alike. (CSM II . SPW , AT VII )
In other words, in the big picturefrom the point of view of the whole
universeit may be better if some beings are prone to error than if all
beings were infallible. In his Reply to Gassendis objections, Descartes
explains this point with a striking analogy:
A . . . comparison [could be made] . . . between someone who

wanted the whole of the human body to be covered with eyes so
as to look more beautiful (there being no part of the body more
beautiful than the eye), and someone who thinks that there
ought not to have been any creatures in the world who were
liable to error (i.e., not wholly perfect). (CSM II )
Meditation IV
It seems, then, that Descartess main way of responding to the question

of why a perfect God did not make humans immune to error is to appeal
to the aesthetic conception of evil.
Is there any place in Descartess thinking about error for the Irenean
theodicy? It seems not. The idea that we could perfect our souls by
learning from our errors, or by gradually increasing our knowledge and
wisdom through overcoming obstacles to the acquisition of knowledge,
seems foreign to Descartess thinking, despite his own life-long, prodi-
gious eorts to increase his knowledge. Thus, in the Third Meditation he
considers the possibility that he might be a greater being than he seems
to be, and even so great as to be able to cause the idea of God, because his
knowledge increases gradually and there is no reason why it could not
grow indenitely. His response is that the very fact that his knowledge
increases gradually shows how imperfect he is: this gradual increase in
knowledge is itself the surest sign of imperfection (CSM II , SPW ,
AT VII ). This seems incompatible with the idea that a soul which has
had to develop its intellectual virtues the hard way is more nearly per-
fect than one to whom everything comes easy.
6.3 Some Critical Reflections

In this section, we shall oer some brief critical reections on the the-
odicy that underlies Descartess attempt to reconcile the fact that humans
are prone to error with Gods perfect goodness. Since Descartes does not
appeal to the Irenean theodicy, we will focus only on the Augustinan
themes that we identied in Cartesian theodicythe privative and the
aesthetic conceptions of evil.
Few philosophers nowadays would look with favor on the equation of
being and goodness that lies at the heart of the privative conception of
evil. Contemporary philosophers commonly make a distinction, which
they regard as fundamental, between facts and values. This fact/value
distinction means that for anything that exists, it is an open question
whether it is good, bad, or indierenta question that can be reasonably
answered only on the basis of some theory of value or some ethical or
aesthetic principles. Thus, the distinction contradicts the idea that any-
thing is good, or valuable, just because it exists, or has being. But as for-
eign as the equation of being and goodness may be to much contemporary
philosophy, it is at the heart of Judeo-Christianity. Christians and Jews
believe that the world was created by and is looked over by a perfectly
Descartes
good God. From their point of view, such a world must be good, and to
discover that it was not good would be devastating and terrifying. It would
threaten the very meaning of their lives, since they could no longer see
humans as children of a good deity who cares for them and will deliver us
from evil and from eternal oblivion. Thus, the existence of evil cannot be
seen by Christians and Jews as a trivial blemish or a small imperfection;
instead, it confronts them with a formidable problem. No wonder, then,
that one classic response to the problem is deny the reality of evil, and to
see the belief that evil is real as based on some sort of misunderstanding
of Gods purposes.
All the same, however, the existence of horrible and prolonged phys-
ical pain, debilitating illness, extreme mental suering and anguish, and
of the natural forces and catastrophes that cause these, is obvious. Epi-
demics, droughts, earthquakes, tsunamis, tornadoes, agonizing and fatal
cancers, horrible work-related injuries, degenerative physical and mental
illnesses are just a few examples. The existence of these things presents
religious believers with a wrenching dilemma: either admit that they are
genuinely evil and utterly inexplicable by us as being works of God, or
deny that they are really evil. If one opts for the rst horn of the dilemma,
then ones religious belief seems to ignore the dictates of reason. If one
opts for the second horn, then one seems wholly callous: how can one
really believe, for example, that the painful death of a child from inoper-
able throat cancer is not evil? Further, even the notion that there is moral
evil then seems imperiled, since if pain and suering are not really evil in
themselves, it is hard to see how intentionally causing them should be
evil. Are we to say that they are evil when intentionally caused by humans,
but not when intentionally caused by God?
The aesthetic conception of evil is perhaps best seen as an attempt to
shore up the privative conception of evil, by denying the ultimate reality
of evil and thereby softening the second horn of the above dilemma. To
answer the question of how an innocent childs suering and death can
be seen as only a privation of good and not really an evil, the aesthetic
conception of evil would suggest that the childs suering and death be
seen as merely an absence of good in one part of the universe that is com-
pensated for by the resulting goodness of the whole.
This suggestion, however, overlooks an important point. The aes-
thetic conception of evil turns on an analogy between parts of an aes-
thetically pleasing whole like a work of art and conscious beings who are
parts of Gods whole creation. But the analogy is awed, for a part of an
Meditation IV
aesthetically pleasing whole, like a single measure of a Bach fugue or a

portion of a painting, is not a sentient being capable of suering or joy.
So, in judging the goodness of the fugue or the painting as a whole, the
only point of view that is relevant is that of the listener or the viewer.
But in judging the goodness of the whole creation, we must also consider
the points of view of each of its living, conscious parts.
In light of the fragility of the aesthetic conception of evil, it may now
seem that Descartess attempt to explain away the fallibility of humans
rests on very weak ground. But perhaps we can offer an alternative
response that Descartes does not give but that he could have given. He
could have held that God gave us an understanding as wide as our wills,
so that we are indeed capable of attaining clear and distinct perceptions
on all matters whatsoever. Then we would still have the freedom to err.
But this freedom would not be based on the mismatch between our intel-
lect and our willthe will would not extend beyond the intellect. Rather,
our freedom to err would stem only from our freedom not to engage in
the arduous work that would still be necessary for nite creatures to at-
tain clear and distinct perceptions regarding all matters. But availing our-
selves of this freedom would then not be a nonmoral evil, but rather a
moral evil manifesting a form of laziness or a lack of perseverance. Thus,
the capacity to make errors, no less than the actual commission of errors,
would reect only a moral aw in humans, and Descartes would not need
the aesthetic (nor even the privative) conception of nonmoral evil to ac-
count for it. Of course adopting this alternative would not help to solve
the problem posed by nonmoral evil, and it would require Descartes to
drop the view that there are some matters that the human intellect, no
matter how assiduously we use it, simply cannot grasp. We leave it to the
reader to consider whether this would have been a viable option for him.

Matson, The Existence of God, pp. .
| 5 |
Meditation V
The Ontological Argument for the Existence of God
1. Descartess Ontological Argument

The Ontological Argument was originally put forward by St. Anselm
(), who was Archbishop of Canterbury under William II and
one of the greatest Christian theologians. Anselm began by dening God
as a being than which nothing greater can be conceiveda denition
that beautifully expresses the Judeo-Christian concept of God as an abso-
lutely unsurpassable being. He then asked whether it is possible that this
being exists only in our minds, as a mere object of thought. He answered
that this is impossible, for then this very being than which nothing greater
can be conceived would be a being than which a greater can be conceived;
for if it had also existed in reality, then it would have been greater. Anselm
concluded that God exists both as an object of thought and in reality.
Although Descartes does not mention Anselm, and it is unclear whether
he was familiar with the Proslogium (the little book in which Anselm advanced
his argument), Descartess argument for Gods existence in Meditation V is
essentially a modernized version of Anselms argument. Instead of dening
God as a being than which nothing greater can be conceived, Descartes
denes God, more simply, as a supremely perfect being. He then argues
that it is impossible for such a being not to exist, because then it would not
be supremely perfect! Before we look at this arresting piece of reasoning
more closely, notice how well it suits Descartess purposes. As already men-
tioned, Descartess arguments for Gods existence must not employ any pre-
mises that refer to the material world, or indeed to anything existing outside
his own thought, because all such things are still doubtful. The Ontological
Argument satises this requirement perfectly, because it turns on the mere
denition or concept of God as an absolutely unsurpassable being.

Descartes
Descartes introduces the argument some way into Meditation V, as

follows:
But if the mere fact that I can produce from my thought the idea
of something entails that everything which I clearly and dis-
tinctly perceive to belong to that thing really does belong to it, is
not this a possible basis for another argument to prove the exis-
tence of God? Certainly, the idea of God, or a supremely perfect
being, is one which I nd within me just as surely as the idea of
any shape or number. And my understanding that it belongs to
his nature that he always exists is no less clear and distinct than
is the case when I prove of any shape or number that some prop-
erty belongs to its nature. Hence, even if it turned out that not
everything on which I have meditated in these past days is true,
I ought still to regard the existence of God as having at least the
same level of certainty as I have hitherto attributed to the truths
of mathematics. (CSM II , SPW , AT VII )
This passages opening sentence reveals that the Meditation V argu-

ment for Gods existence, unlike the Meditation III arguments, does not
purport to provide a vindication of the clarity-and-distinctness criterion
of truth. Instead, the Meditation V argument explicitly uses that criterion.
This dierence in the roles of the arguments is also revealed by Descartess
saying that the Meditation V argument shows that Gods existence has at
least the same level of certainty as I have hitherto attributed to the truths
of mathematics. This statement indicates that Descartes does not see the
Meditation V argument as somehow attaining a higher level of assurance
than mathematics, which would then enable it to guarantee the truths of
mathematics. Instead, he sees the argument as resting, just like mathe-
matics, on the principle that clear and distinct perceptions are true.
The fact that Descartes uses his criterion of truth in Meditation V is
not surprising; for by the time he reaches that Meditation, he takes him-
self to have already established the criterion. Thus, as Copleston has
argued, the problem of the circle does not arise for the Meditation V argu-
ment, as it did for the Meditation III arguments. It is true, as Copleston
also points out, that in the Principles of Philosophy Descartes presents the
Ontological Argument before the causal ones and derives his criterion of

Frederick Copleston, A History of Philosophy, pp. .
Meditation V
truth afterward. So in that work all the theological arguments are evi-
dently seen as contributing to the guarantee of clear and distinct percep-
tion. Perhaps, then, the dierence in the way Descartes treats the
arguments in the Meditations does not reect a deep-seated feature of his
thought. But the dierence is there and should be noted.
Let us now turn directly to Descartess Ontological Argument itself.
The passage that we have quoted from Meditation V does not make the
arguments structure very clear. But Descartes explained his reasoning
more fully in his Replies to both the rst and second sets of Objections:
My argument however was as follows: That which we clearly and

distinctly understand to belong to the true an immutable nature,
or essence, or form of something, can be truly asserted of that
thing. But once we have made a suciently careful investigation
of what God is, we clearly and distinctly understand that exis-
tence belongs to his true and immutable nature. Hence we can
now truly assert of God that he does exist. Here at least the con-
clusion does follow from the premises. But, what is more, the
major premise cannot be denied, because it has already been con-
ceded that whatever we clearly and distinctly understand is true.
(CSM II , M , SPW , AT VII )
But my major premise was this: That which we clearly under-
stand to belong to the nature of something can be truly armed
of that thing. Thus, if being an animal belongs to the nature of
man, it can be armed that man is an animal; and if having three
angles equal to two right angles belongs to the nature of a tri-
angle, it can be armed that a triangle has three angles equal to
two right angles; and if existence belongs to the nature of God, it
can be armed that God exists, and so on. Now the minor pre-
mise of my argument was: yet it belongs to the nature of God
that he exists. And from these two premises the evident conclu-
sion to be drawn is the one which I drew: Therefore it can truly be
armed of God that he exists. (CSM II , AT VII )
Drawing on these passages, we may formulate Descartess argument as

follows:
() Whatever I clearly and distinctly perceive to belong to the

nature or essence of a thing can be truly armed of that thing.
Descartes
() I clearly and distinctly perceive that existence belongs to the

nature or essence of a supremely perfect being.
____________________________________________________
? Existence can be truly armed of a supremely perfect being,
i.e., a supremely perfect being exists.
For brevitys sake, we have left understood the phrase true and immu-
table that Descartes puts in front of nature or essence in both pre-
mises. The point of this phrase will be seen in the next section.
Premise is based on two ideas: Descartess clarity-and-distinctness
criterion of truth and the connection (emphasized in the last-quoted
passage) between a propertys belonging to the essence of a thing and
belonging to (i.e., being such that it can be truly asserted of or truly
armed of) the thing itself. Thus, we can construct the following subar-
gument for premise :
Whatever I clearly and distinctly perceive to belong to the nature

or essence of a thing does belong to its nature or essence.
Whatever belongs to the nature or essence of a thing can be truly
armed of that thing.
____________________________________________________
? Whatever I clearly and distinctly perceive to belong to the
The rst premise of this subargument is a direct application of Descartess

criterion of truth. The second premise is an analytic truth. For essence
and nature here mean basically the same as denition. But if a prop-
erty P belongs to the very denition of a thing, then, of course, P can be
truly armed of (i.e., must actually belong to) that thing itself.
Premise of the main argument involves the key idea in the Ontolog-
ical Argumentthe connection between supreme perfection and existence.
Descartes invokes this connection in replying to the rst of three pos-
sible objections that he considers immediately after introducing the ar-
gument. The rst of these possible objections may be paraphrased:
I have always made a distinction between the essence and the existence
of a thing, between the question What is X? and the question Does X
exist? So why cant I make this distinction in the case of God, and thereby
conceive of God as not existing? Descartess reply is that on considering
Meditation V
the matter more carefully, it becomes obvious that the existence of God
cannot be separated from his essence anymore than having its angles
equal degrees can be separated from the essence of a triangle or
having a valley can be separated from the essence of a mountain. (Des-
cartes claried the second example in a letter of January , to
Gibieuf explaining that by a mountain he had meant merely an uphill
slope, and by a valley a downhill slope [CSMK , AT III ].) Descartes
is here saying that there is an obvious, logical connection between being
God (having Gods essence or nature) and existing, just as there is an
obvious, logical connection between being a triangle and having three
angles that equal degrees or having an uphill slope and having a
downhill slope.
Why is there such a connection between the essence and the existence
of God? Descartess answer comes in his statements that
it is just as much of a contradiction to think of God (that is, a

supremely perfect being) lacking existence (that is, lacking a per-
fection), as it is to think of a mountain without a valley. (CSM II
, SPW , AT VII )
and
I am not free to think of God without existence (that is, a su-

premely perfect being without a supreme perfection). (CSM II
, SPW , AT VII )
In eect, Descartes is here giving a subargument for the that-clause of

premise () of the main argument, which is the part of the premise in
which the crucial connection between supreme perfection and existence
is asserted. This subargument is:
All perfections belong to the essence of a supremely perfect

being.
Existence is a perfection.
____________________________________________________
? Existence belongs to the essence of a supremely perfect being.
This little argument expresses the key idea in Descartess Ontological

Argumentthat the nature or essence of a supremely perfect being must
include existence, because such a being must have all perfections and existence is
Descartes
itself a perfection. In other words, since existence is a perfection, anything that

failed to exist would not have all perfections and so could not be a supremely
perfect being. This is why there is an obvious, logically necessary connection
between being God (i.e., being a supremely perfect being) and existing.
In order to make his argument still clearerto, as it were, teach it to
usDescartes considers two further possible, hypothetical objections to
it. The rst may be paraphrased:
Although it is true that I cannot think of God without existence any-

more than of a mountain without a valley, it would be invalid to
argue: I think of a mountain with a valley; therefore, there exists a
mountain with a valley. Likewise, why isnt it invalid to argue: I
think of God as existing; therefore, God exists? After all, my thought
imposes no necessity on anything: just because, for example, I at-
tribute wings to a horse in my imagination, it does not follow that a
winged horse exists; likewise, just because I attribute existence to
God in my thinking, it does not follow that Go really exists.
In Descartess reply to this objection, he grants that it would be invalid

to argue:
I cannot think of a mountain without a valley.

______________________________________
? A mountain and/or a valley exists.
However, he claims, it is valid to argue:
I cannot think of a mountain without a valley.

______________________________________
? A mountain is inseparable from a valley.
(Of course, this assumes the general principle, which the objection does
not call into question, that if I cannot think of [i.e., clearly and distinctly
perceive] a thing X without a property P, then P is inseparable from X.)
Therefore, Descartes continues, it is also valid to argue:
(i) I cannot think of God without existence.

(ii) Existence is inseparable from God. (from step i)
(iii) God really exists. (from step ii)
Meditation V
Furthermore, he adds, it is the necessity of Gods existence that deter-

mines my thought, not vice-versa; for while I can think of a horse either
with or without wings, I cannot think of God (i.e., of a supremely perfect
being) without existence (i.e., without this particular perfection).
The last possible objection Descartes considers is a little more complex.
It goes as follows. If one supposes or assumes that
() A supremely perfect being has all perfections
then, since
() Existence is a perfection,
it follows that:
() A supremely perfect being exists.
However, the objection continues, the original supposition or assump-

tion, (), is no more necessary than the assumption that
() A circle can inscribe all quadrilateral gures
which, since
() A rhombus is a quadrilateral gure,
would lead to the false conclusion that
() A circle can inscribe a rhombus.
Since a rhombus is an oblique-angled parallelogram, it cannot be in-

scribed in a circle (i.e., placed within a circle with all four corners of the
rhombus touching the circles circumference). This is why () is false. The
point of the objection is that just as the argument from () and () to ()
must be unsound, so the argument from () and () to () is unsound.
Descartes replies, as one would expect, by rejecting the analogy
between the two argumentsspecically the analogy between () and
(). We may paraphrase his reply:
Its true that I do not have to think about a supremely perfect

being, anymore than about a circle. In that sense, and in that
Descartes
sense alone, () and () are alike, are both not necessary. But in
all other important respects, they are dierent. For whenever I do
think of a supremely perfect being, I perceive clearly and distinctly
that () is necessarily true (just as when I think of a triangle, I
perceive clearly and distinctly that it must have three angles). But
when I think of a circle, I do not, and indeed cannot, clearly and
distinctly perceive that () is necessarily true. On the contrary,
() is false, so no wonder that the falsehood () can be derived
from it.
Although the objection just considered may have looked weak even
before Descartess rebuttal, it is important. For in setting out this objec-
tion, Descartes has in eect given us a very simple version of the Onto-
logical Argument itself. This simple version consists of just the rst three
numbered statements from the objection:
() A supremely perfect being has all perfections.

() Existence is a perfection.
_________________________________________
? () A supremely perfect being exists.
This concise Ontological Argument, like the subargument for the that
-clause of premise () of the longer argument (see page ), explicitly
uses the premise Existence is a perfection. This idea, which will be
examined in the next section, is a crucial common element of Descartess
and Anselms ontological arguments. For both Anselm and Descartes,
existence contributes to a things greatness (Anselm) or perfection
(Descartes). So, having dened God, in accordance with Judaism and
Christianity, as a supremely perfect being (Descartes), or as one than
which nothing greater can be thought (Anselm), one is logically com-
pelled to say that such a being exists. This is really the heart of the Onto-
logical Argument.
2. Critique of the Ontological Argument

Most philosophers today agree in rejecting the Ontological Argument,
but there is less agreement as to exactly what is wrong with it. In this
section, we shall weigh three dierent objections to the argument.
Meditation V
2.1 Gaunilos Objection

The rst objection to the Ontological Argument was made by a monk
named Gaunilo against Anselms original argument. Gaunilo objected
that if Anselms argument were sound, then the same reasoning could be
used to prove the existence of things that do not exist, such as an island
than which none greater can be conceived. As applied to Descartess ar-
gument, Gaunilos objection might go as follows. Suppose that for pre-
mise () of Descartess initial argument, we substitute the premise that:

nature or essence of a most perfect island (most perfect lion,
most perfect cigar, most perfect mustache, etc.).
Then from () and premise (that whatever I perceive clearly and dis-
tinctly to belong to the nature or essence of a thing can be truly armed
of that thing) we can deduce that there really exists a most perfect island,
a most perfect lion, a most perfect cigar, a most perfect mustache
indeed a most perfect thing of any and every type! Surely, this absurd
consequence shows that something is wrong with Descartess argument.
To meet this objection, Descartes would probably have appealed to
the theory about natures or essences that he sketches at the beginning
of Meditation V, before stating the Ontological Argument itself. There
Descartes says that he has ideas of certain things that whether or not
they actually exist and whether or not he even thinks of them, have
natures or essences of their own, which he has not invented and cannot
change. Descartes calls such natures or essences true and immutable
natures. As an example, Descartes cites a triangle. He points out that
this geometrical gure has a nature or essence which he did not invent
and which does not depend in any way on his mind, since various prop-
erties can be proved of it, for example, that its longest side is opposite its
widest angle, that its three angles equal two right angles, and so forth.
For Descartes, this example is merely one illustration of a very general
viewthat geometrical gures and other mathematical objects, such as
numbers, have true and immutable natures or essences, which account
for the certainty of mathematics. Furthermore, at least one nonmathe-
matical object, namely, the supremely perfect being, has a true and im-
mutable nature or essenceone that diers from all other essences in
that it alone includes existence. By contrast, Descartes believes, a most
perfect island, lion or cigar has no true and immutable nature or essence.
Descartes
Such things are merely ctitious creatures invented by ourselves, like

the centaur (a mythical beast having the head, arms, and torso of a man
but the body and legs of a horse). Consequently, while premise of the
Ontological Argument is true, () is false (inasmuch as nature or es-
sence in both of these statements is short for true and immutable
nature or essence). Therefore, the argument cannot be used to prove
the existence of most perfect islands or lions or cigars.
This reply to Gaunilos objection, however, is not without diculty; for
it places an onus on Descartes to give a criterion for distinguishing
between things that do have a true and immutable nature or essence
and things that do not. Why does a triangle have a true and immutable
nature, whereas (say) a centaur does not? Descartes would doubtless say
that while the centaur is invented by us and dependent on our thought,
the triangle is not; and at rst this may sound good. But what does it re-
ally mean? It does not mean that while centaurs dont really exist in
nature, triangles do; for Descartes explicitly says that even if no triangle
exists or has ever existed outside his thought, this geometrical gure still
has a true and immutable nature. So there seems to be no relevant dier-
ence here; for just as the concept of a centaur exists only thanks to a
denition constructed by humans that species the properties of a cen-
taur, so the concept of a triangle exists only thanks to a denition con-
structed by humans that species the properties of a triangle. And just
as, having once dened a triangle, we can demonstrate various properties
of that gure (e.g., that its three angles equal degrees), so, having
once dened a centaur, we can demonstrate various properties of that
creature (e.g., having six limbstwo human arms and four equine legs).
How, then, does a thing that has a true and immutable nature dier
from one that does not? Unless this question can be satisfactorily
answered, Descartess theory of true and immutable natures cannot save
his Ontological Argument from Gaunilos objection.
Gaunilos objection, however, has a serious limitation: at best it shows
that something is wrong with the argument. But it does not show what is
wrong with itthat is, which premise is false or which step is fallacious.
By contrast, the next objection to be considered attempts to pinpoint the
error in the argument.

Ernest Sosa has pointed out that Descartes both arms and denies that a composite
gure, such as a triangle inscribed in a square, has a true and immutable nature (quoted
by Anthony Kenny, Descartes, p. ; cf. CSM II , M , SPW , AT VII ).
Meditation V
2.2 Kants Objection

The most famous objection to the Ontological argument was made by
Immanuel Kant in his Critique of Pure Reason (). Kant attacks a pre-
mise that we have seen to be common to both Anselms and Descartess
arguments, namely, the premise that existence contributes to a things
greatness, or that existence is a perfection. Kant attacks this premise in
a particularly deep way, by identifying and criticizing an assumption
behind it. This assumption is that existence is a property. The premise
that existence is a perfection rests on this assumption, because a
perfection is a particular type of property, namely, one that makes
a thing better than it would be without that property. To see this, con-
sider, for example, the perfections of God: he is said to be omnipotent
(all-powerful), omniscient (all-knowing), and omnibenevolent (all-good).
Clearly, each of these three perfections would be a characteristic, prop-
erty or attribute of God. This illustrates the general principle that perfec-
tions are properties. Now the Ontological Argument holds that another
of Gods perfections is existence. By the same principle, then, it too must
be a property. So we may take as established the statement
() If existence is a perfection, then existence is a property.
Now Kants strategy is to attack ()s consequent. If his attack succeeds in

showing that ()s consequent is false (i.e., that existence is not a prop-
erty) then, by modus tollens, it also shows that ()s antecedent is false,
that is, that existence is not a perfection. This would refute Descartess
Ontological Argument, because, as we have seen, the subargument for
the crucial part of the longer versions second premise, as well as the con-
cise version itself, both use the statement that existence is a perfection
as a premise.
But how does Kant propose to show that existence is not a property?
Here is part of what he actually says:
Being is obviously not a real predicate; that is, it is not a concept

of something which could be added to the concept of a thing. It is
merely the positing of a thing, or of certain determinations, as
existing in themselves ... The proposition, God is omnipotent,
contains two concepts, each of which has its objectGod and
omnipotence ... If, now, we take the subject (God) with all its
predicates (among which is omnipotence), and say God is, or
Descartes
There is a God, we attach no new predicate to the concept of

God, but only posit the subject itself with all its predicates, and
indeed posit it as being an object that stands in relation to my
concept. The content of both must be one and the same; nothing
can have been added to the concept, which expresses merely
what is possible, by my thinking its object (through the expres-
sion it is) as given absolutely. Otherwise stated, the real con-
tains no more than the merely possible. A hundred real thalers
do not contain the least coin more than a hundred possible tha-
lers . . . By whatever and by however many predicates we may
think a thingeven if we completely determine itwe do not
make the least addition to the thing when we further declare
that this thing is.
Kants thesis in this famous passage is his opening statement that

Being is obviously not a real predicate. Putting this into contemporary
terminology, Kant may be interpreted as saying
() Exists is not a descriptive predicate.
This claim, to which we shall return in a moment, has a direct bearing on

whether existence is a property; for the statement would be generally
accepted,
() If existence is a property, then exists is a descriptive

predicate.
But, from () and (), it follows by modus tollens that existence is not a
propertyin which case, as we have seen, it is not a perfection. Thus, if
() and () are both correct, then the Ontological Argument is refuted.
Statement () is certainly acceptable; for a property (e.g., redness) is a
characteristic, quality or feature that a thing may have or lack. So, a word
that designates a property (e.g., red) functions to describe things as
having or lacking that property; it is a descriptive word, or, in more tech-
nical vocabulary, a descriptive predicate; for example, if red(ness) is a
property, then the word red is a descriptive predicate. The general prin-
ciple involved herethat if X is a property, then a word that designates

Immanuel Kant, Critique of Pure Reason, pp. .
Meditation V
X is a descriptive predicateis an uncontroversial, analytic truth. State-

ment () simply applies this principle to existence.
Since () is unproblematic, the whole weight of Kants objection rests
on ()the contention that exists is not a descriptive predicate. Ac-
cordingly, we must consider exactly what Kant means by (). He does not
mean that the term exist cant take the predicate place in a sentence.
This would be obviously false, because in a sentence like Tame tigers
exist, the term exist is the grammatical predicate. (This is a case where
the verb and predicate are telescoped into one word; the untelescoped
subject-verb-predicate construction would read, Tame tigers are exis-
tent things, or the like.) But what Kant says is that exist is not a de-
scriptive predicate (not a real predicate). Why? Well, compare Tame
tigers exist with Tame tigers growl. Here growl is again the grammat-
ical predicate. But it also functions to describe tame tigers, to say some-
thing about what they are like: they growl. On the other hand, the term
exist in tame tigers exist, according to Kant, does not function to
describe tame tigers. If I tell you that tame tigers exist, I have not told
you anything about what tame tigers are like. This is why Kant says that
exists is not a descriptive predicate. A descriptive predicate is one that
tells what a thing is like. Since exists doesnt do this, it is merely a gram-
matical predicate, not a descriptive one. Thus Kant says that being is
not a real predicate.
What is the true function of the term exist in a sentence like Tame
tigers exist? According to Kant, its true function is to say that the concept
of a tame tiger applies to something. So when I say that tame tigers exist, I
am not saying that tame tigers have a certain property, namely, existence.
I am saying something about the concept of a tame tiger. I am saying that
this concept applies to something, or has instances, or is exemplied.
Today, some philosophers who agree with Kants basic point would put it
in terms of language, rather than concepts. They would say that the func-
tion of exist in Tame tigers exist is to say that the term tame tiger
applies to something.
The key to grasping Kants point, whether it is put in terms of concepts
or of language, is to see the contrast between (a) describing a thing and
(b) saying that a concept or a term applies to something. A further ex-
ample may help to make this contrast more vivid. Suppose I say God is
omnipotent. Then I am describing God; I am telling you something about
what he is like; I am ascribing a certain property or characteristic to him.
But suppose I say God exists. Then I am not describing God; I am not
Descartes
telling you anything about what he is like; I am not ascribing any prop-
erty or characteristic to him. Instead, I am saying that the term God,
unlike the term unicorn, applies to something; that the concept of God,
unlike the concept of a unicorn, is exemplied or instantiated. Thus,
there is a deep dierence between statements of the form X is of such-
and-such a kind and X exists. The former describe X, assign a property
or properties to X; the latter do not. Instead, they covertly mention the
concept of an X or the term X; they are equivalent to the concept X has
instances or the term X applies to something.
Kants view about existenceand with it his objection to the Ontolog-
ical Argumentis quite widely accepted by contemporary philosophers.
There are probably many philosophers who regard the proposition that
existence is not a property as being about as well-established as any phil-
osophical thesis can be. Kants view is even reected in the symbolic no-
tation of modern logic. A descriptive statement, like the Taj Mahal is
white, would be symbolized as Fa (read as a is F or a has the property
F), where a is a constant denoting the Taj Mahal and F is a predicate
designating the property, whiteness. But an armative existential state-
ment (i.e., a statement asserting that something exists), like the Taj
Mahal exists, would be symbolized as x (x = a) (read as there exists
something x such that x is identical with the Taj Mahal), where is
called the existential quantier and x is a variable. The point, which can
be appreciated without mastering the technicalities involved, is simply
that existence is not represented as a predicate designating a property
but by means of the quantier together with the variable, x (read as
There exists something x such that ...). Indeed, x (x = a) contains no
predicate expression at all.
Although many philosophers accept Kants criticism of the Ontolog-
ical Argument, there are also philosophers who dispute it. These philoso-
phers would point out that properties dier widely from each other
(compare, e.g., the property of whiteness and the property of omnipo-
tence). So why couldnt existence be a property, even if a rather special
one? Existence could be a property that such things as the Taj Mahal,
Australia and electrons have and such things as Santa Claus, Shangri-La,
and gremlins do not have. And exists could be a descriptive predicate
used to designate this property. The fact that existence is not usually
treated as a predicate in the symbolic notation of logic proves nothing;
for this notation is only meant to facilitate the evaluation of arguments
as valid or invalid and does not reveal any metaphysical truths. Besides,
Meditation V
even in this notation it is not impossible, but only inconvenient, to treat

existence as a property. The Taj Mahal exists, for example, can be sym-
bolized as x[(Ex) & (x = a)] (read as There is something x such that x
exists and x is identical with the Taj Mahal), where existence is repre-
sented by the predicate E instead of the quantier and variable, x.
Therefore, some philosophers would say, Kants position is only one pos-
sible view of the matter. To show that it is the true view, argument is
required.
As one would expect, a number of arguments have been oered in sup-
port of Kants view. But no such argument is generally regarded as con-
clusive. Kants own reasoning in the passage we quoted above, for
example, is unlikely to sway a philosopher who sees nothing wrong with
regarding existence as a property. Its core is contained in the following
segment:
If ... we say God is, or There is a God, we attach no new pred-

icate to the concept of God, but only posit the subject itself with
all its predicates, and indeed posit it as being an object that
stands in relation to my concept. The content of both must be one
and the same; nothing can have been added to the concept, which
expresses merely what is possible, by my thinking its object
(through the expression it is) as given absolutely. Otherwise
stated, the real contains no more than the merely possible.
A hundred real thalers [large silver coins formerly used through-
out Europe] do not contain the least coin more than a hundred
possible thalers ... By whatever and by however many predicates
we may think a thingeven if we completely determine itwe
do not make the least addition to the thing when we further
declare that this thing is.
Athough this passage may seem rhetorically quite powerful, its argument
is dicult to make out. Its gist seems to be that if existence were a prop-
erty, there would be a mismatch between our concept of God and the
object of that concept; for then when we say or think that God exists, our
concept of God would fail to include a property that we had attributed to
God, namely existence. The reason why our concept of God would fail to
include that property is that our concept of an object can never include
existence, even if we say or think that the object really exists. Why?
Because the concept of a real X cannot include any property that the
Descartes
concept of a merely possible X does not include, and the concept of a

merely possible X obviously does not include the property of existence.
To put it more formally, perhaps we can fairly reconstruct Kants argu-
ment this way:
() If existence is a property, then when we say that God exists

our concept of God matches its object only if our concept of God
includes (the property of) existence.
() If when we say that God exists our concept of God matches its
object only if our concept of God includes (the property of) exis-
tence, then (likewise) when we say that one hundred real thalers
(or any other real object; the hundred real thalers is only an ex-
ample) exists our concept of one hundred real thalers matches its
object only if our concept of one hundred real thalers includes
(the property of) existence.
() Our concept of one hundred real thalers does not include any
property that our concept of one hundred (merely) possible tha-
lers does not include.
() If our concept of one hundred real thalers does not include
any property that our concept of one hundred (merely) possible
thalers does not include, then it is not the case that our concept
of one hundred real thalers matches its object only if our concept
of one hundred real thalers includes (the property of) existence.
____________________________________________________
? () Existence is not a property.
This is a logically valid argument. It may even be sound, and we venture

to say that like many other Kantian arguments, it goes deep. The trouble
with it, however, is that anyone who holds that existence is a property

For readers who know basic logic, here is a proof of its validity.
() p (q r)
() (q r) (s t)
() u
() u ~(s t)
___________________/?~p
() ~(s t) (), (), Modus Ponens
() ~(q r) (), (), Modus Tollens
() ~p (), (), Modus Tollens
Meditation V
would say that premise () just begs the question; for she would say that
our concept of one hundred real thalers does include a property that our
concept of one hundred (merely) possible thalers does not include,
namely, the property of existence! Kant appears to think that he can
refute that claim by pointing out that a hundred real thalers do not con-
tain the least coin more than a hundred possible thalers, but this is irrel-
evant; for a person who holds that a hundred real thalers has the property
of existence and that a hundred merely possible thalers lacks that prop-
erty certainly does not mean that a hundred real thalers has the property
of containing more coins than a hundred merely possible thalers. Rather,
the person means that a hundred real thalers has a property that a hun-
dred merely possible thalers lacks, namely, the property of existence.
Likewise, to generalize the point, someone who holds that a real object
with properties P, P, P ... Pn (where none of these is existence) also
has a property that a merely possible object with properties P, P, P ...
Pn lacks does not mean that the real object has some property Px,
other than existence, that the possible object lacks, but rather that the
real object has a specic property that the merely possible one lacks:
existence.
In the following section, we shall oer a defense of Kants objection.
We shall not claim, however, that this defense is conclusive, but only that
it makes a reasonable and illustrative case for Kants position.
2.3 Further Consideration of Kants Objection

The following defense of Kants objection is somewhat indirect. We shall
begin by presenting a problem that at rst seems quite unrelated to
Kants objection. Then we shall argue that the most plausible solution to
this problem commits one to Kants view about existence.
Consider the statement
() Carnivorous cows do not exist.
This kind of statement, which denies the existence of something, is called

a negative existential statement, or just a negative existential. Nega-
tive existentials raise a simple yet perplexing problem that has puzzled

In presenting the problem and possible solutions, we shall draw on Richard Cart-
wright, Negative Existentials, .
Descartes
philosophers at least since the time of Plato. This problem is, How can
such a statement be true? To be true, it must be meaningful. But for the
statement to be meaningful, it seems that its subject-term must pick out
something, namely, carnivorous cows. But if the subject-term does pick
out carnivorous cows, then carnivorous cows exist after all, so the state-
ment that they dont exist must be false. For the same reason, it looks as
though all negative existentials must be falsewhich is surely absurd.
There are two classic solutions to this problem: Inationism and Dea-
tionism. The purpose of both solutions is to show how negative existen-
tials can be both meaningful and true.
Inationism, which is defended in the early writings of Bertrand Rus-
sell and in the works of the Austrian philosopher Alexius Meinong
(), tries to allow for true negative existentials by drawing a
distinction between existence and being or subsistence. The basic idea is
that anything that can be thought about or talked about must have being,
must subsist, even if it does not exist. As applied to our example, the idea
is that although carnivorous cows do not exist, they do subsist or have
being. So statement really says,
(a) Carnivorous cows subsist (have being), but dont exist.
In (a), the subject-term does pick out something, namely, subsisting

carnivorous cows. So the statement is meaningful. Moreover, it also hap-
pens to be true, since these subsisting carnivorous cows only subsist and
do not exist. The upshot is that by interpreting () to mean (a), () can
be seen to be both meaningful and true.
Inationism, however, is not a popular theory among contemporary
philosophers, for it has at least two very troublesome consequences.
First, the realm of being that Inationism postulates is very myste-
rious. Suppose one asks, Where are the carnivorous cows and other en-
tities that only subsist but dont exist? Well, the answer has to be that
they are nowhere; for if they were somewhere (i.e., if they were situated
in space) then they would exist. But if they are nowhere, what can it mean
to say that they subsist or have being? The Inationist view seems
utterly mysterious.
The second diculty is that when Inationism is applied to some neg-
ative existentials, it is downright paradoxical. Consider, for example,
Round squares do not exist. According to Inationism, this statement
really says that round squares subsist (have being), but dont exist. But
Meditation V
how can a round square, which is a contradiction in terms, subsist or

have being? It would seem that such an entityone that both has and
does not have anglescannot subsist or have being anymore than it can
exist. It cannot have any kind of being at all.
Deationism, which is defended in the mature writings of Bertrand
Russell and favored by most contemporary philosophers, attempts to
allow for true negative existentials without running into the paradoxes
of Inationism. The fundamental claim of Deationism is that negative
existentials are not really about their subjects at all. A sentence like Car-
nivorous cows do not exist may look as if it were about carnivorous cows,
but this is just a misleading appearance created by its grammatical form.
Since there are no carnivorous cows for the sentence to be about, it
cannot really be about carnivorous cows.
Deationists support this basic claim by oering paraphrases of nega-
tive existential statements that are meant to reveal their true meaning.
Since these paraphrases are supposed to reveal or to analyze the state-
ments true meaning, the paraphrases are called analyses of the state-
ments. Deationist analyses of negative existential statements dier
somewhat from one Deationist philosopher to another. Here are some
typical Deationist analyses of Carnivorous cows do not exist:
(b) The concept carnivorous cow does not apply to anything; i.e., this
concept is not exemplied: it has no instances (Kant, Frege).
(c) The term carnivorous cow does not apply to anything.
(d) Nothing has the dening characteristics of a carnivorous cow;
nothing combines the properties of being mammalian, bovine, and
meat-eating. (C. D. Broad)
(e) Nothing has the properties of bring both carnivorous and a cow
there are no xs such that x is carnivorous and x is a cow is true.
(Bertrand Russell)

Suppose, e.g., that a dragon is dened as a reptile which ies and breathes re.
Then the statement that dragons do not exist is equivalent to the statement that nothing
combines the three properties of being a reptile, of ying, and of breathing re (C. D.
Broad, Arguments for the Existence of God, www.ditexct.com/broad/aeg.html.).

This is an extrapolation from what Russell says about the whole realm of non-
entities in his famous essay On Denoting, p. , as reprinted in his Logic and Knowl-
edge: Essays , pp. .
Descartes
Contemporary philosophers regard these analyzes as being at least ro-

ughly equivalent to each other. One reason for this is that they would all be
symbolized in the same way in modern logic: ~(x)(Vx & Cx) (read as: It is
not the case that there exists an x such that x is carnivorous and x is a cow).
For our purposes, the important point is that these Deationist analyses do
not postulate any objects that subsist but dont exist. In light of the di-
culties such objects generate, therefore, Deationism appears to be a better
solution to the problem of negative existentials than Inationism.
How does this issue bear on Kants objection to the Ontological Argu-
ment? The answer is that Deationism commits one to Kants view about
existence. The reason for this is twofold: () if we analyze negative exis-
tentials in the Deationist manner, then we must analyze armative
existentials in a parallel or isomorphic manner; but () to analyze ar-
mative existentials in a manner parallel to the Deationist analysis of
negative existentials is to adopt Kants view that exists is not a descrip-
tive predicate. Let us explain each of these two points in turn.
To understand point , consider a pair of armative and negative exis-
tential statements pertaining to the same thing, for example, Unicorns
exist and Unicorns do not exist. Now these statements contradict each
other; they are what logic calls contradictories (of each other). This
means that they must have opposite truth-values: they cannot both be
true, nor can they both be false. If one is true, the other is false; and vice-
versa. Now the reason for () is that if we do not analyze the two state-
ments in a parallel fashion, then they wont contradict each other. To see
this, suppose that we analyze the negative existential Unicorns dont
exist as The concept unicorn is not exemplied and that we analyze the
corresponding armative existential, Unicorns exist, as a predicative
statement, that is, as Unicorns have the property of existence. Then it
can be shown that the two statements do not contradict each other.
To show this, we need to use three statements:
. Unicorns have the property of existence.

. The concept unicorn is not exemplied.
. Unicorns subsist or have being.
The following argument proves that () and ()which must be contra-

dictories if they are correct analyses of Unicorns exist and Unicorns
dont exist, respectivelyare not contradictories. Since contradictories
must have opposite truth-values, we know
Meditation V
(A) If () and () are contradictories, then not-() entails ().
Now, our discussion of Inationism has shown that some philosophers

maintain that not-() entails (). Without going that far, we can at least
agree with inationists that:
(B) Not-() does not entail not-().
Further, we may arm
(C) () entails not-().
This is because if there were such things as subsisting unicorns, or uni-

corns that lack existence but have being, then they would exemplify the
concept unicorn. Now, the statement is necessarily true,
(D) If not-() does not entail not-() and () does entail not-(),
then not-() does not entail ().
This is because logical entailment is a transitive relationship: if P entails

Q and Q entails R, then P entails R. So, if not-() did entail (), then it
would have to entail everything that () entails, including not-(). There-
fore, if not-() does not entail not-(), then it cannot entail (), either.
But now, there follows from (B), (C), and (D)
(E) Not-() does not entail ().
Finally, there follows from (A) and (E)
(F) () and () are not contradictories.
This proves that if we analyze unicorns exist as () and unicorns do

not exist as (), then these statements will not contradict each other
which is absurd.
How then should we analyze Unicorns exist, so as to make it contra-
dict The concept unicorn is not exemplied? The answer is obvious. We
must analyze it as: The concept unicorn is exemplied. But (and this is
the second point that we set out to explain) to analyze Unicorns exist
in this way is to adopt Kants view about existence; for it is to reject the
idea that the sentence describes what unicorns are like, in favor of the
Descartes
idea that the sentence says that the concept of a unicorn is exemplied,
or that the term unicorn applies to something. Thus, it is to reject the
view that exists is a descriptive predicate. And this, as we have seen, in
turn implies that existence is not a property and hence not a perfection,
so that the Ontological Argument is unsound. The upshot is that if Dea-
tionism is correct, then the Ontological Argument is unsound.
Although the foregoing defense of Kants objection seems plausible, it
is not wholly unproblematic. For Deationism is not without its di-
culties. The main diculty arises from negative existentials that deal
with mythological and ctional creatures. For example, consider the sen-
tence, Dragons do not exist. According to Deationism, this sentence is
not really about dragons at all, since there are no dragons for it to be
about. Rather, the sentence means The concept of a dragon is not exem-
plied or the term dragon does not apply to anything. Now while such
an analysis seems quite plausible when applied to carnivorous cows do
not exist, it sounds somewhat paradoxical when applied to a sentence
dealing with dragons. For dragons, unlike carnivorous cows, are mytho-
logical creatures; they have a place in mythic lore and literature and a
certain status, so to speak. The same diculty arises when Deation-
ism is applied to sentences about ctional creatures. For example, to say
that Hamlet did not exist is not really about Hamlet sounds paradox-
ical. Now it may well be that this diculty is not fatal to Deationism but
only calls for certain renements in the theory, designed to deal with
ctional discourse. Indeed, philosophers of language and of art pursue
research on this very topic. But pending a satisfactory analysis of myth-
ological and ctional discourse within the general framework of Dea-
tionism, it must be admitted that the foregoing defense of Kants
objection cannot be regarded as conclusive.
2.4 Cateruss Objection

Kants objection to the Ontological Argument, as we have seen, raises
complex and far-ranging issues. Yet, some philosophers feel that the On-
tological Argument commits a rather simple mistakeone that should
not require such elaborate argumentation, and even theorizing, to ex-
pose; for the argument tries to prove Gods existence from a mere deni-
tion of the word God. But, according to these philosophers, it is
impossible to deduce the existence of anything from a mere denition;
it is impossible to build bridges from the conceptual to the real. This
Meditation V
objection to the Ontological Argument is actually an old one. Aquinas

raised it in the Summa Theologica (Part I, Question , art. ); and it was
very concisely and clearly stated by the author of the rst set of Objec-
tions to Descartess Meditations, a priest named Caterus. Caterus wrote:
Even if it is granted that a supremely perfect being carries the

implication of existence in virtue of its very title, it still does not
follow that the existence in question is anything actual in the real
world; all that follows is that the concept of existence is insepa-
rably linked to the concept of a supreme being. (CSM II , M
, SPW , AT VII )
In this section, we shall argue that Cateruss objection (as we shall call
this criticism) is a decisive objection to Descartess version of the Onto-
logical Argument.
In order to explain Cateruss objection, we shall use a contemporary
distinction that was not known to Caterus but is implicit in what he
wrote, namely, the distinction between the material mode of speech and
the formal mode of speech. This distinction, as well as its bearing on the
Ontological Argument, is introduced in a brilliant passage by the English
philosopher, Antony Flew (). The Ontological Argument, Flew
writes, provides
a memorably persuasive illustration of the need to have and to

master a notation which can make absolutely and systemati-
cally clear the fundamental dierence between, on the one
hand, verbal and conceptual questions and, on the other, mat-
ters of fact and existence. For the premise of the Ontological
Argument is the denition of a word, whereas the proposed
conclusion would be the supreme fact of the universe. The
prime need is to distinguish, and the corresponding notational
need is for devices to mark the distinction, between: on the one
hand, discussion of the concept or concepts ofsayGod (talk
about the meaning or meanings of the word God, and about the
implications of its employment); and, on the other hand, dis-
cussion of the objects, if any, of these concepts (talk about the
things to which these words do or would refer) ... Once given
the prime distinction between concept and object, and a cor-
responding notation, these can be and have been developed
Descartes
to illuminate . . . logical relations. The extended notational

distinction is that between the Material mode of Speech
(MMS) and the Formal Mode of Speech (FMS). These labels
were introduced by Rudolf Carnap (b. ), a charter mem-
ber of the Vienna Circle of old original logical positivists... .
The man who says that Threehood necessarily involves Odd-
ness, that the idea of a triangle contains the idea of the equality
of its three angles to two right angles, or that existence is part
ofor isthe essence of God is thereby employing the Mate-
rial Mode of Speech. The alternative Formal Mode is, as might
be feared much more long-winded: to say There are three of
them and to deny There are an odd number of them would be
to contradict yourself; to say This is a triangle and to deny
This has three angles equal to two right angles would be to
contradict yourself; and which exists is part ofor isthe
denition of the word God ... No one would be so foolish as to
suggest that the Formal Mode ought completely and generally
to replace the Material Mode of Speech. Yet it is extremely
important to be able, and on occasion willing, to transpose pas-
sages from the one into the other. [An] excellent exercise would
be ... to transpose into the Formal Mode the passage from ...
Descartes quoted already [in which Descartes advances the On-
tological Argument].
In a moment, we will do the exercise that Flew proposes. First, how-

ever, let us make sure that we understand the distinction between the
material and formal modes of speech. The material mode is the more
common mode of discourse, in which people talk about objects in the
world, like rocks and trees. The formal mode, on the other hand, is a more
specialized mode, in which we talk about concepts and/or words. Here you
may object that concepts or words are in the world, just as rocks and
trees are. This is true. But there is still a distinction between language
and concepts on the one hand, and nonlinguistic or nonconceptual re-
ality on the other. Formal mode discourse has the former as its subject
matter; material mode discourse has the latter as its subject-matter. For
example, suppose that I say or write:

Antony Flew, An Introduction to Western Philosophy, pp. .
Meditation V
(a) Horses are large, solid-hoofed, herbivorous quadrupeds.
This statement is in the material mode, because it is about certain objects

in the world (i.e., certain nonlinguistic and nonconceptual items):
horses. But suppose that instead of (a), I say,
(b) Horse means large, solid-hoofed, herbivorous quadruped.
Then my statement is in the formal mode, because it is about certain

words or linguistic expressions: horse and large, solid-hoofed, herbiv-
orous quadruped. Another formal-mode statement would be:
(c) The concept horse may be applied when and only when the
concept large, solid-hoofed, herbivorous quadruped may be applied.
The reason (c) is in the formal mode, of course, is that it is about certain
concepts: the concepts horse and large, solid-hoofed, herbivorous quadru-
ped. It would still be in the formal mode if it were expanded to read
(c) The concept horse, or term horse, may be applied when and
only when the concept large, solid-hoofed, herbivorous quadruped,
or phrase large, solid-hoofed, herbivorous quadruped, may be
applied.
Now as Flew points out, it would be very foolish to suggest that the for-
mal mode is always preferable to the material mode. Indeed, if we had to
transpose all our material mode statements into formal mode ones, then
we simply could not say most of the things that we want to say; for we gen-
erally mean to talk about nonlinguistic reality, rather than about concepts
or language itself. The two types of discourse are not equivalent, except
perhaps in certain special instances, such as denitions, which are arguably
about words rather than things. (This is the reason for talking of trans-
posing, rather than translating, from one mode to the other. The notion
of transposition is borrowed from music. Just as transposing a piece of
music from one key into another may change its character, so transposing
a statement from one mode into the other may alter its meaning.)
Nevertheless, there are certain contexts or situations where it is
important to be able to transpose a material mode statement into a for-
mal mode one. For example, suppose I say,
Descartes
(d) A unicorn is a one-horned animal resembling a horse.
A person who did not know that a unicorn is a mythical beast might be
misled by my statement into thinking that there really are unicorns or
that I believe that there are unicorns; or a person who knew that there
are no unicorns might begin to wonder how my statement can be mean-
ingful and even true, since there are no unicorns for it to be about. In
such cases, it would he helpful to transpose (d) out of the material mode
into the formal mode, as follows:
(e) Unicorn means one-horned animal resembling a horse.
The transposition makes it clear that (d) is only a denition, and as such,
does not imply that there are any objects actually answering to the de-
nition. Like other denitions, (d) does not carry what philosophers call
existential import: it does not imply the existence of anything. All that
(d) implies is that we have a concept of a unicorn and a corresponding
term. Transposing (d) into (e) makes this point obvious.
We are now ready to apply what we have learned about the material
and formal modes to Descartess Ontological Argument. Let us begin
with the very simple version of the argument that, as we saw in section ,
he oers in connection with the last of his three possible objections to
the argument:
() A supremely perfect being has all perfections.

() Existence is a perfection.
________________________________________
? () A supremely perfect being exists.
To refute this argument, we shall make a series of four remarks.

. Premise () is a denition; for although it is not explicitly couched in
the form of a denition, it turns on the point that a supremely perfect
being must, by denition, possess every perfection. Now since () is a
denition, it does not carry existential import; it does not imply or pre-
suppose the existence of anything. Specically, (), or the use of the noun
phrase a supremely perfect being within (), does not imply or presup-
pose the existence of a supremely perfect being. This is fortunate because
if () did imply or presuppose the existence of a supremely perfect being,
then the argument would of course beg the question; for then its rst
Meditation V
premise would covertly assert the very proposition that the argument is
intended to establish, so that it would be impossible to know that pre-
mise to be true without already knowing the conclusion to be true.
. Since () is a denition, it can easily be transposed out of the mate-
rial mode into the formal mode:
(F) Supremely perfect being means being that has all

perfections.
As already indicated, some philosophers would even say that transposing

() into (F) does not alter ()s meaning, because () is a denition and
denitions are, strictly speaking, about words rather than things. But the
criticism we are developing does not depend on whether this point is
correct, so we need not insist on it.
. A proponent of the argument must agree to let us substitute (F) for
() in the argument, on pain of begging the question; for why should he
refuse to allow this substitution? The only possible reason would be that
he wishes to treat () as more than just a denitionthat he is interpret-
ing (), or the noun phrase a supremely perfect being in (), as carrying
existential import, that is, as implying the existence of a supremely per-
fect being. But then the argument begs the question, as we have seen.
. However, when (F) is substituted for (), the argument becomes
invalid. For () does not logically follow from (F) and (). The only con-
clusion that may be derivable from (F) and () is
(F) Supremely perfect being means being that (among other

things) exists.
But this conclusion does not say that a supremely perfect being exists. It
merely says that only a being whose denition says that (among other things)
it exists would satisfy the denition of a supremely perfect being; or, as it
might very misleadingly be put, that only a being that exists by deni-
tion would satisfy the denition of a supremely perfect being. But this
does not show that any being does satisfy this denition; it does not
prove that there is a supremely perfect being. Look at it this way: To be
supremely perfect, a being would have to be omnipotent, omniscient, and
omnibenevolent. Now how could the fact that only a being whose deni-
tion says that it exists would satisfy the denition of a supremely perfect
being possibly prove that there really is an omnipotent, omniscient, and
Descartes
omnibenevolent being? That fact is simply irrelevant to such a conclusion.

To put the point still dierently: (F) is, so to speak, purely negative in-
formation: it just means that no being whose very denition does not
say (imply) that it exists qualies for the title, supremely perfect being.
But it doesnt follow from this that any being does qualify for this title:
perhaps the title simply has no holder.
At this point, you should reread the passage from Caterus quoted at
the beginning of this section. Can you see how the criticism just oered
merely spells out Cateruss point more fully? All we have done is to use
the contemporary distinction between the material and formal modes to
explicate Cateruss insight; for Caterus can be seen as pointing out that
the formal mode statement A supremely perfect being carries the impli-
cation of its existence by its very title (= statement F) does not entail
the material mode statement that The existence in question is anything
actual in the real world (= statement ) but only another formal-mode
statement to the eect that the concept of existence is inseparable from
the concept of a supreme being. Notice also that Cateruss objection is
completely independent of Kants. Cateruss refutation works even if one
grants the premise that existence is a perfection; for granting this pre-
mise merely compels one to dene the concept of a supremely perfect
being in such a way that only a being whose denition says that it exists
can satisfy the denition. But, as Caterus shows, this does not mean that
anything does satisfy the denition.
In his replies to the First Set of Objections, Descartes attempts to
answer Cateruss objection by making a distinction between possible exis-
tence and necessary existence. He writes:
It must be noted that possible existence is contained in the con-

cept or idea of anything that we clearly and distinctly under-
stand; but in no case is necessary existence so contained, except
in the case of the idea of God. Those who carefully attend to this
dierence between the idea of God and every other idea will un-
doubtedly perceive that even though our understanding of other
things always involves understanding them as if they were exist-
ing things, it does not follow that they do exist, but merely that
they are capable of existing. For our understanding does not
show us that it is necessary for actual existence to be conjoined
with their other properties. But, from the fact that we under-
stand that actual existence is necessarily and always contained
Meditation V
with the other attributes of God, it certainly does follow that

God exists. (CSM II , M , SPW , AT VII )
In replying to Caterus, Descartes also oers an explanation of his notion

of necessary existence. Unlike contemporary philosophers, who think of
necessary existence as existence in all possible worlds, Descartes thinks
of it as existence by virtue of a beings own power. Thus he writes:
When we attend to the immense power of this being, we shall be

unable to think of its existence as [even] possible without also
recognizing that it can exist by its own power; and we shall infer
from this that this being does really exist and has existed from
eternity, since it is quite evident by the natural light that what
can exist by its own power always exists. So we shall come to
understand that necessary existence is contained in the idea of a
supremely powerful being, not by any ction of the intellect, but
because it belongs to the true and immutable nature of such a
being that it exists. (CSM II , M , SPW , AT VII )
Descartes also alludes to necessary existence in two footnotes that he

added in the French version of Meditation V. Since the French version of
the Meditations appeared in , six years after the Latin version
appeared along with (all but the seventh set of) the Objections and Replies,
it seems likely that Descartes added these two footnotes with his reply to
Caterus in mind. In them, he expands the segment of the Latin version
that Cottingham translates as apart from God, there is nothing of which
I am capable of thinking such that existence belongs to its essence into
apart from God, there is nothing of which I am capable of thinking such
that existence necessarily belongs to its essence, and he expands the
translated Latin versions what is more self-evident than the fact that
the supreme being exists, or that God, to whose essence alone existence
belongs, exists? into what is more self-evident than to think that there
is a God, that is to say, a sovereign and perfect being, in the idea of whom
alone necessary or eternal existence is comprised, and who consequently
exists? (CSM II , SPW , AT VII ).

I have slightly amended Cottinghams translation of the second-quoted passage from
the French version.
Descartes
Does Descartess appeal to the notion of necessary existence help to

answer Cateruss objection? It seems not; for, rst, the curious principle
that what can exist by its own power always exists, contrary to Descartess
claim that it is quite evident by the natural light, is anything but obvious:
compare it to if I am thinking, then I exist or to + = . Indeed, this
principle seems to mean, or at least to imply, that just because we have the
concept of something that can exist by its own power, therefore this con-
cept must be instantiated, which is surely false. After all, just because we
can form the concept of something that exists without any external cause,
it does not follow that that concept is instantiated; but if we now add to the
content of our concept of this thing that it exists because of an internal
cause (or by its own power, of because it is the cause of itself), then even
assumingwhat might be questionedthat this addition makes sense, it
still does not follow that the concept is instantiated. Second, Caterus could
easily amend his objection, quoted above, to read this way:
Even if it is granted that a supremely perfect being carries the

implication of necessary existence in virtue of its very title, it
still does not follow that the existence in question is anything
actual in the real world; all that follows is that the concept of
necessary existence is inseparably linked to the concept of a
supreme being. (CSM II , M , SPW , AT VII , with
necessary added)
In short, just because one builds necessary existence into the concept of
a supremely perfect being, it does not follow that anything answers to
that concept. Thus, suppose that we import necessary existence into the
formal-mode version of Descartess shortest version of the Ontological
Argument, so that it reads this way:
(F) Supremely perfect being means being that has all

perfections.
(N) Necessary existence is a perfection.
____________________________________________________
? (FN) Supremely perfect being means being that (among
other things) necessarily exists.
This argument still does not show that a supremely perfect being
necessarily exists, but only that a being whose denition says that (amo-
ng other things) it necessarily exists would satisfy the denition of a
Meditation V
supremely perfect being. On the other hand, if we insist on keeping the

argument in the material mode, then as before it begs the question.
Cateruss objection also applies to the version of the Ontological Argu-
ment that Descartes gives in answering his second hypothetical objec-
tion. That version went as follows:
(i) I cannot think of God without existence.

(ii) Existence is inseparable from God. (from step i)
(iii) God really exists. (from step ii)
The objection is then that (ii) does not really follow from (i). All that re-
ally follows from (i) is:
(iiF) The concept of existence is inseparable from the concept of God,
where this means that one cant think of God without thinking of exis-
tence (not that one cant think of existence without thinking of God). But
even if (iiF) is true, this does not show that there is anything answering to
the concept of God, or that this concept is exemplied. Again, the basic
point is that just because we include existence in a concept, it does not
follow that that concept is exemplied or instantiatedi.e., that there is
anything answering to the concept. Nor would it help to substitute
necessary existence for existence in (i), (ii), and (iiF) and necessarily
exists for exists in (iii).
Before concluding our examination of the Ontological Argument, we
should ask whether Cateruss refutation also works against Descartess
initial, longer formulation of the argument. That formulation was:
() Whatever I clearly and distinctly perceive to belong to the

nature or essence of a supremely perfect being.
____________________________________________________
?() Existence can be truly armed of a supremely perfect
being, i.e., a supremely perfect being exists.
To see that Cateruss objection does indeed bear on this argument, we

need only transpose it out of the material mode into the formal mode.
One way to do this is as follows:
Descartes
(F) Whatever I clearly and distinctly perceive to be part of the

denition of X can be truly armed of all Xs, if there are any.
(F) I clearly and distinctly perceive that exists is part of the
denition of supremely perfect being.
____________________________________________________
?(F) Exists can be truly armed of all supremely perfect be-
ings, if there are any.
This argument obviously does not prove that there are any supremely per-
fect beings or (more to the point) that there is even one supremely per-
fect being. Rather, all it proves is that if there are any supremely perfect
beings, then they exist or (more to the point) that if there is a supremely
perfect being, then it exists. In other words, the argument does not prove
that God exists, but only (the tautology that) if God exists, then he ex-
ists. Notice that this point continues to hold if the phrase, if there are
any is deleted from the argument: this phrase is used only for the sake
of emphasis. Notice also that again, nothing would be gained by sub-
stituting exists necessarily (or exists by virtue of its own power or
exists in all possible worlds) for exists in the argument. At best, the
argument might then show that if God existed in the actual world, then
he would exist in all possible worlds.
It might be objected that this way of dealing with Descartess argu-
ment is too quick. For why should Descartes not simply refuse to allow
the substitution of the formal mode argument for his material mode
one? In response, we may make two points.
. The formal mode version is an improved formulation of what was
confusedly and tendentiously expressed in Descartess formulation. For,
to quote Flew once again,
[A]n inferior notation may ... encourage and express actually er-
roneous ideas. This is the reason for writing transpose rather
than translate; for the FMS analogue may sometimes be a sub-
stantial improvement on, and hence not equivalent to, the MMS
original. Thus . . . most of those, from Aristotle onwards, who
have spoken of the essences of things would have been reluctant
to allow that all they were saying was expressed in some FMS

This last observation is due to Krasimira Filcheva.
Meditation V
statement about the denitions of the words; though they might

have had even greater diculty in explaining precisely what more
they had in mind, and in justifying their beliefs about it.
. We need not alter Descartess formulation as drastically as we have

done in order to make Cateruss objection. Even the following formula-
tion, in which only premise () is altered (and quite minimally so), will
serve:
(Fa) Whatever I clearly and distinctly perceive to belong to the

essence of an X can be truly armed of all Xs, if there are any.
essence of a supremely perfect being.
____________________________________________________
? (Fa) Existence can be truly armed of all supremely perfect
beings, if there are any.
Again, this argument obviously does not prove either that there are any
supremely perfect beings or (more to the point) that there is even one
supremely perfect being. Again, it yields only the tautologies that if there
are supremely perfect beings, then they exist, or that that if there is a
supremely perfect being, then it exists. Again, this point holds even if the
phrase, if there are any, is deleted from the argument. Finally, again,
importing the notion of necessary existence into the argument would
not help.
But what if Descartes refused to allow the substitution of (F) or (Fa)
for ()? Then our reply would be that substituting (F) or (Fa) for ()
makes it clear that not all things whose denitions (or essences) we
clearly and distinctly perceive to include certain properties must really
exist. By contrast, refusing to substitute (F) or (Fa) for () amounts to
insisting that all things whose denitions (or essences) we clearly and
distinctly perceive to include certain properties must really exist. But
this would be obviously false. Furthermore, Descartes himself did not
believe it; for in presenting his theory about true and immutable
natures near the beginning of Meditation V, he explicitly says that many
things whose true and immutable natures or essences he clearly and

Antony Flew, An Introduction to Western Philosophy, p. .
Descartes
distinctly perceives to include various properties may not exist any-

where outside me. For example, he says, a geometrical gure like a tri-
angle has a true and immutable nature, whose constitutive properties he
clearly recognize[s], even if perhaps no such gure exists, or has ever
existed, anywhere outside my thought (CSM II , SPW , AT VII ).
3. Some Implications for Descartess System

We may conclude that Descartess Meditation V argument for the exis-
tence of God is no more successful than his Meditation III arguments for
Gods existence. This negative verdict on Descartess philosophical the-
ology has major implications for the rest of his system. As we have said,
his overall strategy, as shown in Figure ., is to extract his criterion of
truth from the cogito, to provide a vindication of this criterion by means
of the Meditation III causal proofs of Gods existence, and then to use the
criterion in his Meditation V ontological argument for Gods existence, in
his Meditation VI attempt to establish that mind is a dierent substance
than any matter which may exist, and nally in his attempt to prove, also
in Meditation VI, that matter really does exist.
The rst implication of the failure of Descartess arguments for the
existence of God, then, is simply that his attempt to establish Gods exis-
tence must be judged unsuccessful. In view of the human signicance of
the question of Gods existence, and given that the subtitle of the Medi-
tations includes the words, in which are demonstrated the existence of
God, this, of course, represents a major failure in Descartess overall
scheme.
The second implication is that Descartess attempt to vindicate his cri-
terion of truth by appealing to the existence of a perfect God must be
also be regarded as a failure. If the solution to the problem of the Cartesian
Figure .
Meditation V
circle oered in the chapter is satisfactory, then this failure does not
stem from any circularity in Descartess procedure but simply from the
fact that his specic arguments for Gods existence do not work. As we
have noted, the Meditation V argument seems not to be intended to con-
tribute to the vindication of Descartess criterion of truth anyway. Nev-
ertheless, if it were a sound argument, then perhaps it could play such a
role provided the appeal it makes to clear and distinct perception were
understood in the same way as for the Meditation III arguments. How-
ever, in light of the failure of both the Meditation III and the Meditation V
arguments, it seems that we must conclude that Descartess attempt to
provide a divine vindication for his criterion of truth is ultimately unsuc-
cessful. Therefore, if we hold that Descartes cannot legitimately use this
criterion unless it can be vindicated, then we must also conclude that he
cannot advance beyond the cogito. In particular, he cannot establish the
two remaining major theses of his Meditations, the distinction between
mind and (any) body (which may exist) and the existence of the material
world, because his arguments for both of them rely on his criterion
of truth.
Descartess arguments for these two theses, however, are worth con-
sidering regardless of whether the criterion of truth that they rely upon
can be vindicated. So in the next chapter, we shall adopt the following
policy. We shall assume that the clarity-and-distinctness criterion of
truth can stand on its own merits, or solely on the basis of the extraction
argument (from the cogito) that was presented in chapter , section . In
the course of examining Descartess arguments for the distinction
between mind and body and for the existence of matter, however, we
shall inquire whether the failure of his philosophical theology has any
implications for those arguments, beyond the fact that the criterion of
truth that they employ must stand on its own merits. Our suggestion
will be that the failure does not seriously compromise Descartess argu-
ment for the distinction between mind and body, but that it profoundly
aects his argument for the existence of matter.
| 6 |
Meditation VI
Dualism and the Material World
1. The Scope of Meditation VI

Meditation VI is a longish text in which Descartes completes his basic
agenda and addresses a number of subsidiary themes. It can be divided
into the following segments or episodes (the paragraph numbers are
those of the Cottingham translation): () a discussion of the imagination
and of how this faculty makes it probable that material things exist
(paragraphs ); () a review of what Descartes initially believed about
the senses and of the reasons he subsequently found for doubting them
(paragraphs ); () the proof of the Real Distinction between mind
and body (i.e., of Cartesian Dualism) (paragraph ); () the proof that
the material world exists (paragraph ); () an account of certain par-
ticular beliefs about material things that, in light of this proof, can safely
be accepted (paragraphs ); () an account of certain plausible but
nevertheless erroneous beliefs about material things (paragraph ); ()
a physiological explanation of why we sometimes pursue ends that are
bad for us and of why this fact does not reect adversely on Gods per-
fection (paragraphs ); and () a resolution of the dream problem
(paragraph ). In addition, episode contains a subsidiary argument
for mind-body dualism (paragraph ). We shall not try to discuss all of
these topics. Rather, we shall focus on Cartesian Dualism and on Des-
cartess theory of the physical world. We shall examine Descartess main
proof of dualism, his proof that material things exist, and some of his
basic views about their nature; discuss the most famous problem that
arises from Descartess dualism; and, nally, oer an overall assessment
of Cartesian Dualism.

Descartes
2. Descartess Proof of the Real Distinction

between Mind and Body
Descartes presents his main argument for dualism in the following para-
graph, which we have divided into three segments in order to facilitate
discussion:
[A] First, I know that everything which I clearly and distinctly

understand is capable of being created by God so as to corre-
spond exactly with my understanding of it. Hence the fact that I
can clearly and distinctly understand one thing apart from an-
other is enough to make me certain that the two things are dis-
tinct, since they are capable of being separated, at least by God.
The question of what kind of power is required to bring about
such a separation does not aect the judgement that the two
things are distinct.
[B] Thus, simply by knowing that I exist and seeing at the same
time that absolutely nothing else belongs to my nature or es-
sence except that I am a thinking thing, I can infer correctly that
my essence consists solely in the fact that I am a thinking thing.
[C] It is true that I may have (or, to anticipate, that I certainly
have) a body that is very closely joined to me. But nevertheless,
on the one hand I have a clear and distinct idea of myself, insofar
as I am simply a thinking, non-extended thing; and on the other
hand I have a distinct idea of body, insofar as this is simply an
extended, non-thinking thing. And accordingly, it is certain that
I am really distinct from my body, and can exist without it. (CSM
II , SPW , AT VII )
Let us begin by examining segment B, since it raises special diculties

that need to be noticed before we can properly analyze Descartess argu-
ment. On the face of it, segment B seems to be arguing
(i) I know that nothing belongs to my nature or essence except

that I am a thinking thing.
_____________________________________________________
? (ii) My essence consists solely in the fact that I am a thinking
thing.
Meditation VI
Now in light of what philosophers call the truth-condition for knowl-

edge, this is obviously a valid argument; for that condition says that,
simply by virtue of the denition of knowledge, S knows that p entails
that p is true. But still the argument is quite unsatisfactory; for nowhere
in his previous Meditations has Descartes established the truth of its
premise. Rather, all he has established (in Meditation II) is that the only
property he knows for certain belongs to his nature or essence is
thinking. This suggests that the argument in segment B really ought to
go as follows:
(ia) I do not know that anything other than thinking belongs to

my nature or essence.
____________________________________________________
?(ii) My essence consists solely in the fact that I am a thinking
thing.
Indeed, in the French translation of the Meditations, which Descartes

himself approved, this is essentially how the argument does go. The
French version of segment [B] says,
Thus, just because I know with certainty that I exist, and that
meanwhile I do not notice that anything else necessarily belongs
to my nature or essence except that I am a thinking thing, I
rightly conclude that my essence consists solely in the fact that I
am a thinking thing.
As E. M. Curley has pointed out, the Latin version (on which the transla-
tion by John Cottingham from which we are quoting is based) is ambig-
uous and could also have been translated in this way. We suggest that this
would actually have been a better way for Cottingham to translate seg-
ment B. For it cannot be overemphasized that premise (ia), not premise
(i), is all that Descartes is entitled to assert at this point in his Meditations:
He showed, in Meditation II, that the only property which he knew for
certain to belong to his essence was thinking; he did not show (but is now,
in Meditation VI, trying to show) that the only property which in fact
belongs to his essence is thinking. Therefore, the argument in segment B

Edwin M. Curley, Descartes Against the Skeptics, p. .
Descartes
should be interpreted, despite Cottinghams translation, as going from

(ia) to (ii), not as going from (i) to (ii).
This observation, however, brings us face to face with the basic di-
culty in the segment B argument: it is simply not valid. One cannot val-
idly go from the premise that P is the only property that is certainly
known to belong to ones essence, to the conclusion that P is the only
property that really belongs to ones essence. Such an argument from
ignorance is not signicantly dierent from the obviously invalid argu-
ment from doubt that, as we saw, Descartes seems to use in his Discourse
on the Method and Search After Truth: I cannot doubt that I exist; I can
doubt that (my) body exists; therefore I am not a body.
As we saw, however, in the more careful Second Meditation, Descartes
admitted that this is an invalid argument; for directly after stating it, he
went on to admit that, for all he knew, he might be a body (see page ):
And yet may it not perhaps be the case that these very things
which I am supposing to be nothing, because they are unknown
to me, are in reality identical with the I of which I am aware? I
do not know, and for the moment I shall not argue the point,
since I can make judgements only about things which are known
to me. (CSM II , SPW , AT VII )
Yet, as the word thus at the beginning of segment B indicates, Des-

cartes now thinks that the material in segment A somehow legitimizes
the reasoning in [B]. The material in [C], as we shall argue in a moment,
also contributes to his legitimization of [B]. (More obviously, segment C
spells out the consequence that follows when the general principles in
[A] are combined with the thesis Descartes will soon prove [i.e., that he
has a body] and with the clear and distinct conceptions of mind and body
attained in his previous Meditations, namely, that his body is really dis-
tinct from his mind.) In any case, the paragraph as a whole is partly
intended to show that an argument that was deemed unsatisfactory at an
earlier stage, when supplemented by further principles, can be safely ac-
cepted. This reading is conrmed by what Descartes says in the Medita-
tions Preface to the Reader. There he notes that it had been objected,
against the Discourse on the Method that
From the fact that the human mind, when directed toward itself,
does not perceive itself to be anything other than a thinking
Meditation VI
thing, it does not follow that its nature or essence consists only
in its being a thinking thing, where the word only excludes
everything else that could be said to belong to the nature of the
soul. (CSM II , AT VII )
After complaining that this objection is based on a misinterpretation of

his intentions in the Discourse, Descartes adds, signicantly,
I shall, however, show below how it follows from the fact that I
am aware of nothing else belonging to my essence, that nothing
else does in fact belong to it. (CSM II , AT VII )
This certainly raises a question of interpretation; for if an argument is

invalid, then no magic can transform that very argument into a valid one.
At best, the addition of new premises can only yield a new argument that
includes the premises of the old one and is valid. But segments A and C
do not seem to stand even in that relation to the argument in [B]. So
what does Descartes mean by suggesting that the reasoning in [B] has
now been vindicated? To answer this question, we must rst examine
what Descartes says in segments A and C.
The rst sentence in [A] can be formulated as follows:
() If I can clearly and distinctly conceive X existing in a certain

way, then X can really exist in that way, at least by Gods power.
This opening premise, which Descartes also asserts at the very beginning
of Meditation VI (where he says, there is no doubt that God is capable of
creating everything that I am capable of perceiving in this manner [i.e.,
clearly and distinctly] [CSM II , SPW , AT VII ], follows directly
from his clarity-and-distinctness criterion of truth). In the next sentence,
Descartes derives a general principle from ()that if he can clearly and
distinctly conceive X existing apart from Y, then X is really a dierent
thing from Y. (In expressing this principle we use the variables X and Y
instead of Descartess expression one thing apart from another, because
that expression may misleadingly suggest that in the if-clause of the prin-
ciple, Descartes is already assuming that he is referring to two dierent
things; whereas, in fact, the if-clause sets forth the condition under which
we can know that we are referring to two dierent things rather than to
only one, or that the variables X and Y stand for two dierent things
Descartes
rather than for one and the same thingi.e., the condition under which
we can know that the then-clause is true.) The inference from () to this
principle, however, requires two intermediate steps. The rst one is ellip-
tically stated in the clause, since they are capable of being separated, at
least by God. Untelescoped, this is a premise saying,
() If I can clearly and distinctly conceive X existing apart from Y,

then X really can exist without Y, at least by Gods power.
This follows directly from () (because existing apart from Y counts as

existing in a certain way), and leads to Descartess principle. To obtain
that principle, however, the second intermediate step is required. This is
a premise that Descartes states, somewhat obliquely, in his next sen-
tence: The question of what kind of power is required to bring about
such a separation does not aect the judgment that the two things are
distinct. This can be put as follows:
() If X really can exist without Y, no matter what power it takes,

then X and Y are really two dierent things.
The idea behind () is that if X can really exist without or apart from Y, then
even if it takes so much as Gods power for this to happen, X and Y must be dif-
ferent things; for not even God could make a thing exist without or apart
from itself, since that is not even logically possible. (Here we are deliberately
ignoring the extraordinary view about omnipotence that, as we saw in chap-
ter , Descartes apparently held, according to which an omnipotent God
could do logically impossible thingsfor example, make a four-sided tri-
angle or make two contradictory statements both true. Descartes does not
mention this mind-boggling view in the Meditations, and it would certainly
ruin his argument for dualism. Indeed, he seems to put this view aside at
the start of Meditation VI when he says, I have never judged that something
could not be made by him [God] except on the grounds that there would be
a contradiction in my perceiving it distinctly (CSM II , SPW , AT
VII ). From () and (), we can now derive Descartess principle that:
() If I can clearly and distinctly conceive X existing apart from Y,

then X and Y are really two dierent things.
This brings us to the end of the material in segment [A].

Meditation VI
The references to clearness and distinctness found in ()() are absent

from the next segment of our passage, [B]. But they are picked up again
in the following segment, [C], where Descartes advances another pre-
mise and then draws two conclusions. The premise, worded so as to make
it connect clearly with () and (), is
() I can clearly and distinctly conceive myself, as a thinking and

nonextended thing, existing apart from (my) body, as an ex-
tended and nonthinking thing.
This premise rests squarely on Descartess long and careful discussion of

the self in Meditation II; for surely, if anything was established by that
discussion, it was that Descartes could form a clear and distinct concep-
tion of himself as simply a thing that thinks (arms, denies, wills,
imagines, seems to perceive, etc.)a conception which, in the passage
about the wax, he went on to contrast with the conception of an ex-
tended thing. But with the help of (), the two conclusions that Des-
cartes draws at the end of segment [C] can be derived. For there follows
from () and ()
() I am really a dierent thing from my body.
And there follows from () and ()
() I can really exist without my body, at least by Gods power.
These two conclusions assert the Real Distinction between mind and
body. In his Principles of Philosophy, Descartes explains that a real distinc-
tion is one between two or more substances; he contrasts such a distinc-
tion with a modal distinction, which is one between a substance and a
mode or between two modes of the same substance, and with a concep-
tual distinction or distinction of reason, which is one between a sub-
stance and an attribute without which it cannot be intelligibly conceived
(he gives a substance and its duration as an example) or two such attrib-
utes of the same substance (he gives extension and being divided into
parts as an example) (CSM I , SPW , AT VIIIA ).
Notice that despite the high prole of () in Descartess text, this prin-
ciple is not really needed in the argument; for it is not used to derive (),
and () can be derived without it from (), (), and (). Indeed, as we shall
Descartes
stress in section , lines (), (), and () are the arguments fundamental
premises (see p. .)
Before discussing the argument, let us try to answer the question of
interpretation raised above: How is this argument, which is extracted
only from segments A and C, and which is essentially an argument from
the independent conceivability of mind and body, supposed to legitimize
the reasoning in segment B, which is an epistemological argument in-
voking the certainty of thought and (implicitly) the doubtfulness of
body? Well, compare steps ()() with [B]. In ()(), Descartes goes
from the premise that he can clearly and distinctly conceive himself
existing as only a thinking thing to the conclusion that he is distinct from
his body and could exist without it. This certainly resembles the argument
in [B], where he goes from the premise that thinking is the only property
which he knows for certain belongs to his essence, to the conclusion that
he is essentially only a thinking thing. But we can go further: we can say
that the argument in [B] is merely a simplied versionwhich Descartes
does not claim to be valid as it standsof the argument extracted from
[A] and [C]. To see this, let us compare the two arguments. In the rst
place, their premises are intimately related; for Descartess claim in Medi-
tation II that thought was the only property that he knew for certain
belonged to his essence already had much more content than the words
Thinking is the only property that I know for certain belongs to my es-
sence reect. What more? Well, he had a clear conception of his thinking;
he was forming a clear conception of body; he could appreciate the con-
trast between those two conceptions; and he could clearly conceive,
through the doubt, the possibility that the object of the former concep-
tion might exist though the object of the latter did not. But this is virtu-
ally what () says. In the second place, the two arguments conclusions are
intimately related; for, as Descartes uses these notions, to say that X can
exist without Y entails that Y is not part of Xs essence: [I]f something
can exist without some attribute, then it seems to me that that attribute
is not included in its essence (fourth set of Replies, CSM II , SPW ,
AT VII ). Thus, when Descartes arms in () that he can exist without
his body, this entails that extension is not part of his essence, which
leaves only thinking as his essence. But this is just the conclusion of [B].
Of course, the argument in [B] directly corresponds only to steps
()() of the argument in [A] and [C]. The longer argument goes beyond
the short one, by presenting the general principles that are needed to le-
gitimize the inference from a clearly and distinctly conceived distinction
Meditation VI
between mind and body to a real distinction between them, namely,

()(). But this should not stop us from seeing that the argument in [B]
is a highly simplied version of what Descartes fully and adequately pre-
sents in ()().
This interpretation can be conrmed by three observations. First, the
conclusion-indicator word thus that introduces segment B functions in
a very unusual way. It does not function, in the normal way, to introduce
the conclusion of an argument. Rather, it introduces an entire argument,
namely, the segment B argument. Clearly then, Descartes does think that
the longer argument somehow legitimizes the shorter one. But how? In
the remark we quoted above from the Preface to the Reader Descartes
seems to concedeand in Meditation II he plainly does concedethat
the segment B argument is invalid. Presumably, then, Descartes is not
saying that the argument has now been shown to be valid after all. Yet in
the Preface to the Reader Descartes also promises to show below how
such an argument can nonetheless be legitimate. The place where he ful-
lls this promise is the very paragraph in Meditation VI that we are exam-
ining. But surely his way of fullling the promise cannot be just to assert
baldly that the argument earlier conceded to be invalid is really valid.
What then is the meaning of Descartess thus? The answer, we suggest,
is that this word indicates that the simplied, heuristic sketch in [B] is
now legitimate, because it serves as a stand-in for the more complex argu-
ment in [A] and [C]. In other words, thus indicates that segments A and
C provide a kind of second-order commentary on segment B, to the eect
that the reasoning in [B] is now legitimate not because it is itself valid
but rather because it can do duty for the more complex argument that
Descartes has just presented.
Second, as Jakko Hintikka emphasized in his famous article on the
Cogito, Cartesian doubt is no merely passive aair. Rather (as Descartess
litmus test of certainty shows), it amounts to an active attempt to
think the contrary of what we usually believe. Thus, when Descartes
says that he can be certain of his existence as a thinking thing while
doubting the existence of his body, part of what this means is that he can
clearly and distinctly conceive the possibility that his mind exists while
his body does not. It should not be surprising, therefore, that Descartes
initially presents an argument from the independent conceivability of

Jakko Hintikka, Cogito, Ergo Sum: Inference or Performance? p. ; also in
Alexander Sesonske and Noel Fleming, eds., Meta-Meditations: Studies in Descartes, p. .
Descartes
mind and body in the guise of an epistemological argument turning on

certainty about the mind and doubt about the body.
Finally, when Descartes responded to a request that he present his
main arguments in geometrical fashion, he himself gave, as his demon-
stration of the Real Distinction, essentially the argument in [A] and
[C], omitting the one in [B] altogether (CSM II , SPW , AT VII
). We may reasonably conclude, then, that the argument we have
extracted from segments A and B is Descartess nished, ocial argu-
ment for dualism.
It behooves us, therefore, to understand exactly what this argument
is, and is not, supposed to prove. It is supposed to prove, in conclusions
and , what Descartes calls the Real Distinction between Mind and
Body. But exactly what does this distinction amount to? Well, Descartes
is not claiming to have proved in () that his mind does, at any time, exist
without his body (i.e., without the body that he will shortly argue he
possesses). Rather, he is claiming to have proved that the mind could
exist without the body. As line indicates, Descartes holds that the mere
possibility that mind can exist apart from body is enough to establish
that they are dierent. As previously mentioned, his idea is that even if
it would take so much as Gods power for this possibility to be actual-
ized, the mere logical possibility is sucient for mind and body to be
dierent things; since a thing cannot possibly exist apart from itself.
And (as he notes at the end of segment A) whether it would require Gods
power for them to exist separately does not matter. If any power would
allow them to exist separately, then they must be dierent. Thus, as the
late, eminent Descartes scholar Margaret Wilson pointed out, it would
be a fundamental misunderstanding to complain that
Descartess argument can show at best that mind and body are
possibly or potentially distinct (would be distinct if God should
choose to separate them)not that they are distinct. [For] Des-
cartes holds that two things are really distinct if it is possible for
them to exist in separation. On his view actual distinctness does
not entail separateness.
Notice, then, that Descartess argument is not intended to establish that

the mind or soul actually exists after the body stops functioning, that is,

Margaret Wilson, Descartes, p. .
Meditation VI
that the immortality of the soul is a fact. Rather, it is only intended to

show that the mind could exist without the body, that is, that immortality
is possible. Notice also, however, that if the mind is just (some part of) the
body, then immortality is not even possibleunless, of course, bodily
resurrection is possible, a question we shall not discuss here. Thus, while
Descartess dualism is not sucient to establish immortality, it does seem
to be necessary (assuming that bodily resurrection is not possible).
Now that we have Descartess main argument for dualism before us, it
is time to evaluate it. Let us start by addressing a possible objection that
has probably occurred to you already. This is that since the premises of
Descartess argument refer to Gods power and since his arguments for
Gods existence are unsuccessful, his argument for dualism collapses.
Although this objection is a very natural one to make, it does not really
refute Descartess argument; for the statements in the argument that
refer to Gods power, namely, (), (), and (), need not depend for their
truth on whether God exercises that power or even on whether God ex-
ists. To see this, consider premise , If I can clearly and distinctly con-
ceive X existing in a certain way, then X can exist in that way, at least
by Gods power. This premise is similar in form to the statement If I
can see normally, then I can see molecules, at least with the help of a
microscope. But for this statement to be true, microscopes do not have
to exist: the statement was not false before microscopes came into exis-
tence, and it would remain true even if they were all destroyed. Likewise,
Descartess premise can be true whether or not God exists. The same goes
for step of his argument, since it follows from premise . Signicantly,
then, the conclusion drawn in step (I am really a dierent thing from
my body) does not depend for its truth on Gods existence either.
It might seem, however, that even if Descartess premises do not
depend on Gods existence for their truth, one must at least admit that
the conclusion drawn in step (I can exist without my body, at least by
Gods power) loses all interest if God does not exist. But since God may
exist even if it has not been proved that he does exist and since Descartess
argument is only meant to show the possibility of disembodied survival
anyway, even () is not deprived of all its signicance by the failure of
Descartess philosophical theology. It must be admitted, however, that
the subtlety of Descartess argument makes it quite dicult to assess its
exact signicance. We shall return to this matter in the nal section of
this chapter. The point we wish to emphasize for now is just that the
soundness of Descartess argument does not depend on Gods existence.
Descartes
To reinforce this point in one last way, it may be useful to look at

the matter differently. Suppose that the phrase at least by Gods
power were simply deleted from Descartess entire argument. Then
would not the argument be just as strong as before? If so, then this
suggests that there is no essential connection between Descartess
case for dualism and his philosophical theology. Rather, his argument
for dualism is, as previously noted, essentially an argument from the
independent conceivability of mind and matter.
Let us now turn to a second possible objection to Descartess argu-
ment. This is the objection made by Antoine Arnauld, in the fourth set of
Objections to the Meditations. Arnauld was generally very sympathetic
to the Cartesian philosophy, but he acutely criticized a number of Des-
cartess arguments. (As we have seen, it was he who most clearly raised
the problem of the circle.) In a clear and forceful way, he presented what
is probably still the most penetrating objection to the proof of the real
distinction.
Arnauld could not see that Descartess proof was any improvement
over what we have called his preliminary, abbreviated version of the ar-
gument. As Arnauld put it:
But so far as I can see, the only result that follows from this is
that I can obtain some knowledge of myself without knowledge
of the body. But it is not yet transparently clear to me that this
knowledge is complete and adequate, so as to enable me to be
certain that I am not mistaken in excluding body from my es-
sence. (CSM II , M , SPW , AT VII )
To clarify his point, Arnauld went on to give a parody of Descartess

argument:
Suppose someone knows for certain that the angle in a semi-

circle is a right angle, and hence that the triangle formed by this
angle and the diameter of the circle is right-angled. In spite of
this, he may doubt, or not yet have grasped for certain, that the
square on the hypotenuse is equal to the squares on the other
two sides; indeed he may even deny this if he is misled by some
fallacy. But now, if he uses the same argument as that proposed
by our illustrious author, he may appear to have conrmation of
his false belief, as follows: ... [H]e may say ... I know ... that
Meditation VI
everything which I clearly and distinctly understand is capable of

being created by God so as to correspond exactly with my under-
standing of it. And hence the fact that I can clearly and distinctly
understand one thing apart from another is enough to make me
certain that the two things are distinct, since they are capable of
being separated by God. Yet I clearly and distinctly understand
that this triangle is right-angled, without understanding that the
square on the hypotenuse is equal to the squares on the other
sides. It follows on this reasoning that God, at least, could create
a right-angled triangle with the square on its hypotenuse not
equal to the squares on the other sides.
I do not see any possible reply here, except that the person in
this example does not clearly and distinctly perceive that the tri-
angle is right-angled. But how is my perception of the nature of
my mind any clearer than his perception of the nature of the tri-
angle? He is as certain that the triangle in the semi-circle has one
right angle (which is the criterion of a right-angled triangle) as
I am certain that I exist because I am thinking.
Now although the man in the example clearly and distinctly
knows that the triangle is right-angled, he is wrong in thinking
that the aforesaid relationship between the squares on the sides
does not belong to the nature of the triangle. Similarly, although
I clearly and distinctly know my nature to be something that
thinks, may I, too, not perhaps be wrong in thinking that nothing
else belongs to my nature apart from the fact that I am a thinking
thing? Perhaps the fact that I am an extended thing may also
belong to my nature. (CSM II , M , SPW , AT
VII )
The logic of Arnaulds objection can be summarized as follows. Suppose

that premise () of Descartess proof is replaced with
() I can clearly and distinctly conceive a right triangle existing

apart from the square on its hypotenuse being equal to the sum
of the squares on its other two sides.
Then it follows from () and line () of Descartess proof that:
() A right triangle can really exist without the square on its hypot-
enuse being equal to the sum of the squares on its other two sides.
Descartes
Since () is absurdly false, and is validly deduced from () and (), either
() or () must be false. But (), as the example of the angle in the semi-
circle is supposed to show, is true. So () is false; therefore Descartess
proof is unsound. (Notice also that if () is false, then so is (), since ()
follows from ()and also that Descartess criterion of truth is false,
since () follows from it. Moreover, neither (), (), nor () can be estab-
lished; all that remains of the argument is () and ().)
In his fourth set of Replies, Descartes wrestles with Arnaulds objec-
tion for several pages. Without trying to cover all of his points, let us
focus only on the most instructive ones.
Descartess rst point is that the example of the right triangle is not
parallel to the case of mind and body. Arnaulds example, he says,
diers in many respects from the case under discussion. First of

all, though a triangle can perhaps be taken concretely as a sub-
stance having triangular shape, it is certain that the property of
having the square on the hypotenuse equal to the squares on the
other sides is not a substance. (CSM II , M , SPW , AT
VII )
Descartess point is simply that having the square on its hypotenuse

equal to the squares on its other sides is a property of a right triangle (we
might call it the Pythagorean property, after Pythagoras, who demon-
strated the Pythagorean Theorem that describes this property), not an-
other substance. Thus, the most that Arnaulds example can show is that
() is false when X designates a substance and Y designates a prop-
erty. It cannot show that () is false when, as in Descartess argument,
X and Y both designate substances.
Although the example of the right triangle does dier from the case of
mind and body in the way Descartes indicates, his point is a weak one: it
does not get to the heart of Arnaulds objection; for, as Anthony Kenny
incisively points out,
[I]t may be replied that being extended is a property, and

Arnaulds argument shows that being able to conceive mind
without this property does not show that mind in fact lacks it.

Anthony Kenny, Descartes: A Study of His Philosophy, p. .
Meditation VI
In other words: since extension is a property, and Descartes is trying to

prove that a mind can exist without this property, he cannot aord to
allow that () is false even when Y designates only a property.
Later, however, Descartes gives a better reply to Arnauld. He says:
It is true that the triangle is intelligible even though we do not

think of the ratio which obtains between the square on the hy-
potenuse and the squares on the other sides; but it is not intelli-
gible that this ratio should be denied of the triangle. In the case of
the mind, by contrast, not only do we understand it to exist with-
out the body, but, what is more, all the attributes which belong to
the body can be denied of it. (CSM II , M , SPW ,
AT VII ; emphasis mine)
[T]here is no way in which the triangle can be distinctly under-
stood if the ratio which obtains between the square on the hy-
potenuse and the squares on the other sides is said not to hold.
(CSM II , M , SPW , AT VII ; emphasis mine)
To analyze this reply, we can make three points:
. The reply turns on Descartess interpretation of the phrase, clearly

and distinctly conceive (understand, perceive) X apart from Y. Des-
cartes explains in several places that this phrase does not mean just
clearly and distinctly conceive X without thinking of Y; for he real-
izes that from ones ability to do that, it does not follow that X can
really exist without Y. Thus, in rst set of Replies he grants that one
can clearly and distinctly conceive of motion without thinking of a

There is a rather similar passage in one of Descartess letters:
There is a great dierence between abstraction and exclusion. If I said simply that
the idea which I have of my soul does not represent it to me as being dependent
on the body and identied with it, this would be merely an abstraction, from
which I could form only a negative argument which would be unsound. But I
say that this idea represents it to me as a substance which can exist even though
everything belonging to the body be excluded from it; from which I form a
positive argument, and conclude that it can exist without the body. And this
exclusion can be clearly seen in the nature of the soul, from the fact that one
cannot think of a half of a thinking thing. (Letter of May , to Mesland,
CSMK , AT IV )
Descartes
body: one can form a clear and distinct idea of motion without
thinking of a moving body. But, he implies, it does not follow from
this that motion can exist on its own, without being the motion of
some body (CSM II , AT VII ). This is why, in the proof of the
Real Distinction, clearly and distinctly conceive X apart from Y
cannot mean clearly and distinctly conceive of X without thinking of
Y. Rather, it means clearly and distinctly conceive X existing while
conceiving that Y does not exist. (It is to suggest this reading that, in
formulating the steps of his proof, we put the word existing after
X.) As Descartes says in the sixth set of Replies:
I found that the distinction between ... mind and body ... is
much greater than the distinction between things which are such
that when we think of both of them we do not see how one can
exist apart from the other (even though we may be able to under-
stand one without thinking of the other). For example, we can
understand the immeasurable greatness of God even though we
do not attend to his justice; but if we attend to both, it is quite
self-contradictory to suppose that he is immeasurably great but
not just. (CSM II , AT VII )
. Descartes could have argued that the phrase, clearly and distinctly
conceive X apart from Y, cannot mean clearly and distinctly con-
ceive X existing while conceiving that Y does not exist in Arnaulds
(), since I can clearly and distinctly conceive a right triangle exist-
ing while conceiving that the property of having the square on its
hypotenuse equal to the squares on its other two sidesthe Pythag-
orean propertydoes not exist makes little if any sense. Rather, in
Arnaulds (), the phrase must have some meaning other than the
one it has in Descartess (). Therefore, Arnaulds example fails to
refute (), because () cannot be deduced from () and () if the
key phrase means one thing in () and something else in (). But
this would not have answered the point that Arnaulds example
does refute () when Y designates a property.
. To meet this point, Descartes allows the key phrase to have a
meaning that ts Arnaulds example; namely, clearly and distinctly
conceive X existing while conceiving that X is not Y (i.e., that X does
not have property Y). This diers from Descartess own interpreta-
tion of the phrase, since it does treat Y as a property that X might or
Meditation VI
might not have, rather than insisting that Y be a thing. But by inter-
preting the phrase in this way, Descartes is able to cut to the core of
Arnaulds objection; for he can now show that the objection fails to
prove that () is false even when Y designates a property. For Des-
cartess premise () now means
(a) I can clearly and distinctly conceive myself existing while

conceiving that I am not extended (i.e., that extension is not a
property of me).
Arnaulds (), on the other hand, now means
(a) I can clearly and distinctly conceive a right triangle existing

while conceiving that the square on its hypotenuse is not equal to
the sum of the squares on its other two sides.
However, Descartes points out, while (a) is true, (a) is false; for
although I can think of a right triangle, and presumably even conceive it
clearly and distinctly, without any thought of the ratio between the
square on its hypotenuse and the squares on its sides, I cannot clearly
and distinctly conceive a right triangle while conceiving that the square
on its hypotenuse is not equal to the squares on its sides. As he puts it,
there is no way in which the triangle can be distinctly understood if the
ratio which obtains between the squares on the hypotenuse and the
squares on the other sides is said not to hold. Therefore, Arnaulds ex-
ample fails to show that () is false even when Y designates a property.
Admittedly, this reply invites further questions. One would like to
know exactly why it is possible to clearly and distinctly conceive a right
triangle without thinking of the Pythagorean property but not possible
to clearly and distinctly conceive a right triangle while denying that
property of it. Descartess idea seems to be that when X is P follows
from ones conception of X, there is something worse about conceiving
that X is not P than about just not thinking of P while conceiving X. He
seems to be committed to the following criterion for clear and distinct
conception:
S clearly and distinctly conceives X only if, for every property P

such that X is P follows from Ss conception of X, it is not the
case that S can conceive that X is not P.
Descartes
Obviously, it would be desirable to have an account of clear and distinct

conception to support this criterion. But even in the absence of such an
account, the criterion seems quite plausible. There is something to be
said for the idea that while one can have a clear and distinct conception
of X without being aware of all the properties that follow from ones con-
ception of X, one fails to have a clear and distinct conception of X if one
conceives it not to have a property that follows from ones conception of
X. Furthermore, it will not do to object that for all Descartes knows, per-
haps I am extended does follow from his conception of himself; for
Descartess careful examination of his idea of himself in Meditation II
seems to show that this is not the case (whereas a careful examination of
his conception of a right triangle would show that the Pythagorean prop-
erty does follow from that conception).
We may conclude, therefore, that Arnaulds objection fails to refute
Descartess argument. Since that objection is probably the most acute
one ever raised against Descartess argument, we may accept the argu-
ment, at least provisionally. We shall have more to say about it in our
overall assessment of Cartesian Dualism in section .
3. Descartess Proof of the Material World

If one accepts Cartesian Dualism, then one faces a dicult question that
has come to be known as the mind-body problem: What is the relation-
ship between mind and matter, specically between the mind and body
of a single person? Before discussing this issue, however, we need to
consider Descartess views about the material world. How does he -
nally overcome his doubt about the very existence of such a world? And
what are material objects really like, according to him? These questions
are addressed, respectively, in the present and in the next section. Then,
in the last two sections, we shall be ready to consider the mind-body
problem.
In the paragraph that directly follows his proof of the Real Distinction,
Descartes nally overthrows his doubt concerning material things. He
does so, as we might expect, by advancing a proof of their existence. Des-
cartes begins the paragraph very cautiously, by noting that he has certain
faculties, namely, imagination and sensation, which are distinct from
him in the same way as the modes of a physical thing would be from that
thing, but which do not require the existence of any physical things. These
Meditation VI
faculties require only his existence as a thinking substance, for they are
merely modes or properties of a thinking substance, as is shown by the
fact that they cannot be conceived to exist without a thinking substance
to which they belong. Next Descartes notes that he also recognizes (i.e.,
has the clear and distinct conception of) certain other faculties or powers,
such as motion and change of shape, that would, if they really exist, require
the existence of an extended or material substance; for they clearly
require extensionare modes of extensionand hence can really exist
only if extended substance exists. The question is whether any extended
substance does exist.
Descartes nally turns to this question about one-third of the way
into the paragraph, where he says:
Now there is in me a passive faculty of sensory perception, that

is, a faculty for receiving and recognizing the ideas of sensible
objects; but I could not make use of it unless there was also an
active faculty, either in me or in something else, which produced
or brought about these ideas. (CSM II , SPW , AT VII )
Here Descartess terminology may get in the way of understanding the

point he is making. By ideas of sensible objects, he means sensory expe-
riences, which he here regards as a type or subclass of ideas, just as he did
in Meditation III, where he classied hearing a noise, seeing the sun,
and feeling the re as falling among his adventitious ideas (CSM II ,
SPW , AT VII ). Thus, as we shall see in a moment, he ascribes to
these ideas the same notion of objective reality as he ascribes to all
other ideas. By saying that he has a passive faculty of sensory percep-
tion, he means that he has the capacity to obtain these ideas of sensible
objects without any awareness of producing them himself. Thus, his basic
claim is just that he has ideas of sensible objects (= sensory experiences)
that he is not aware of actively producing or conjuring up himself. Now,
claims Descartes, these ideas must have some cause: his point after the
semicolon is not merely the pragmatic one that if there were no cause of
his sensory experiences, his faculty or capacity for receiving and rec-
ognizing the ideas would be useless to him, but, rather, that it would not
even be activated, so that he would never even have the ideas. Descartes,
then, is here relying on the principle that his ideas must have some
causea principle that (as we have seen) is also central to his Meditation
III proof of Gods existence.
Descartes
We may now summarize the rst step of Descartess argument:
() I have ideas of sensible objects that I do not seem to produce

myself; these ideas must be produced by some cause.
The question now becomes: What is this cause? Descartes will argue that
it must be material things. His argument proceeds by a process of elimi-
nation, that is, by ruling out all other possible causes of the ideas. We
shall present it in a somewhat informal manner, because its formal struc-
ture is not as important as its overall conception.
The rst possibility that Descartes seeks to rule out is that he himself
is the cause of the ideas. (This possibility is not ruled out by (), which
only says that he does not seem to produce the ideas himself.) We may
paraphrase what he says, at the start of the sentence immediately fol-
lowing the one quoted above, and incorporating the helpful emendation
about not presuppose[ing] any thought on my part that Descartes
substituted into the French version:
() The cause of my ideas of sensible things is not in me as a

purely thinking thing, because (a) it does not presuppose my
thought, and (b) the experiences it produces come independently
of my will.
To understand this, we need to know what Descartes means by some-

things presupposing thought. Something presupposes thought if it
could not exist unless thought existed. For example, doubting, believing,
and feeling all presuppose thought, because they could not exist unless
thought existed. This is because they are specic kinds of thought or, in
Descartess own terminology, modes of thought (just as squareness
and circularity presuppose extension, because they are specic forms, or
modes, that extension can take on and that cannot really exist unless
extension exists). What Descartes means by (a), then, is that whatever
the cause of his ideas of sensible objects may be, it is not a mode of
thought. For this cause is something that produces sensory experiences;
it is an active faculty. But to doubt, to believe, or to feel is manifestly
not to produce or cause some sensory experience; and in (a) Descartes is
making the general point that no mode of thought consists in causing
sensory experiences. In point (b), he anticipates a possible objection to
this generalization. The objection is that there is a mode of thought that
Meditation VI
could be the cause of his ideas of sensible objects, namely, the will. In
other words, perhaps I deliberately will (i.e., conjure up) these ideas.
Descartess reply to this objection is that it is simply false that I will or
conjure up the ideas, because they manifestly occur quite independently
of my will.
It is important to understand that step depends on the results of the
proof of the real distinction; for Descartes is assuming that he is a purely
thinking substance (or in any case that the part of him which thinks is
distinct from any body that may also exist) and giving reasons why the
cause of his ideas of sensible objects must be something other than this
thinking substance. In other words, () has the quite limited function of
showing that the cause of these ideas is not the thinking substance whose
existence was asserted in the cogito and whose distinctness from any-
thing physical was nally shown by the proof of the Real Distinction.
Beyond this negative claim, () tells us nothing about the identity of this
cause.
Descartess next step goes as follows:
() The cause of my ideas of sensible things must therefore be

some substance other than myself. Furthermore, this substance
must contain, either formally or eminently, all the reality that
the ideas it produces contain objectively.
Here Descartes uses once again the principle that the cause of an idea
must have at least as much formal reality as the idea contains objective
reality; for, as we saw in chapter , whenever X contains formally or emi-
nently all the reality that idea I contains objectively, X has at least as
much formal reality as I contains objective reality, and so it is possible for
X to cause I.
Next, Descartes gives a breakdown of all the possible causes of his
ideas of sensible things allowed by ():
() this substance is either
(a) body (i.e., matter), a substance that contains formally all the
reality that the ideas it produces contain objectively; or
(b) God Himself, in which case it is a substance that contains em-
inently all the reality that the ideas it produces contain objec-
tively; or
Descartes
(c) some created thing more noble than body, in which case,
again, it is a substance which contains eminently all the re-
ality that the ideas it produces contain objectively.
There seems to be a minor diculty here. Descartes seems to overlook

the possibility that the substance that causes his sensory experiences
could be some created thing which is dierent from body but exactly as
noble as (instead of more noble than) body. For example, suppose that
this substance were some nite, nonphysical substance other than him-
self. Then Descartes seems to assume that this substance would have to
possess more formal reality than the ideas it causes contain objective
realityto be nobler than the bodies these ideas portray. But the basis
of this assumption is not obvious: Why couldnt the substance in ques-
tion possess exactly the same degree of formal reality as the ideas it
causes contain objective reality? It would seem that in order to cover this
possibility, Descartes should have formulated option (c) this way:
(c) some created thing other than body, which contains emi-
nently all the reality that the ideas it produces contain objectively.
(Remember that according to the explanation of eminently contains of-

fered in chapter , a cause that contains exactly as much but a dierent kind
of formal reality than its eect contains that eects reality eminently.)
Despite this somewhat esoteric diculty, however, the, general pat-
tern of Descartess reasoning in step () is clear enough: his ideas of
sensible objects could be caused by either (a) physical objects, (b) God
himself, or (c) some other thing created by Godsome deputy of God,
so to speak.
In his next step, which is the crucial step in the argument, Descartes
eliminates (b) and (c). We may paraphrase what he says as follows:
() The cause of my ideas of sensible objects cannot be God or

any created substance other than bodies; for God has given me
no way to spot that this is so, but, instead, a very powerful incli-
nation to believe that these ideas come from bodies (material
objects). So God would be a deceiver if the ideas were produced in
any other way. But since God is a supremely perfect being, he
cannot be a deceiver.
() Therefore, bodies exist.
Meditation VI
If, as Descartes had supposed might be the case in Meditation I, all of his
sensory experiences were caused by God himself or, so to speak, by some
deputy of God, Descartes would have absolutely no way to detect this: he
would be subject to a permanent, undetectable hallucination. Moreover,
he would still have a virtually irresistible feelingwhat the twentieth-
century Spanish-American philosopher George Santayana called animal
faiththat his sensory experiences did come from bodies. (Try actually
to doubt that the sensory experiences you have at this very moment are
caused by a written page. Is this not very dicult to do? This illustrates
Descartess and Santayanas point.) Thus, he would be irremediably
deceived. But this means that God, who Descartes assumes has been
clearly proved by his previous Meditations to be the creator of Descartes
and whatever else exists, would be a deceiver; for a deceiving God would
be precisely one who allows any falsity in my opinions which cannot be
corrected by some other faculty supplied by God (CSM II , SPW ,
AT VII ). But Descartes also takes himself to have abundantly shown
that God is a perfect being, who therefore cannot be a deceiver, because,
as he said in the Third Meditation, It is manifest by the light of nature
that all fraud and deception depend on some defect (CSM II , SPW ,
AT VII ). It follows, then, that sensory experiences are caused by mate-
rial things: the doubt of the existence of material things generated in
Meditation I is, at last, overthrown.
Having so proved the existence of the material world, Descartes im-
mediately goes on to add an important qualication, which we may para-
phrase as follows:
() Material things may not be exactly as we perceive them to be

by our senses, since such perception is often obscure and con-
fused. But they must really contain all that is clearly and dis-
tinctly perceived in them by the mind, i.e., the geometrical
propertiesextension and its modes.
This remark sets the stage for Descartess views about the nature of ma-
terial things, which will be considered in the next section.
The conclusion drawn in () is the very general one that corporeal
things (bodies, physical things) exist. This conclusion does not tell us
what physical things there are, or even whether there is only one physical
thing in the universe or more than one. But Descartes now also feels en-
titled to draw certain more specic conclusions that, as he puts it, I am
Descartes
taught by nature. By this phrase, which must not be confused with the
light of nature (= reason, or the faculty of clear and distinct perception),
Descartes is indicating that there are many things which he is naturally
or spontaneously inclined to believe. Now that he has overcome his gen-
eralized doubt about the existence of matter, he believes he can safely
accept some of these things. It is important to note, before listing these,
that Descartes says he can accept them just because of the very fact that
God is not a deceiver (CSM II , SPW , AT VII ). In other words,
Descartes now takes himself to know for certain, solely on the grounds
that God is not a deceiver, several specic things about the material
world that he nds himself naturally impelled to believe. The things he
lists are these:
. I have a body.
. I am very closely joined to this body.
. There are other bodies (physical things).
. My perceptions of colors, sounds, tastes, temperatures, and hard-
nesses enable me correctly to infer that the physical things that cause
these perceptions have properties that vary as widely as, but may not
resemble, those perceptions.
. These bodies can aect me both benecially and harmfully.
Points , , and relate to Descartess views about the relationship

between mind and body. We shall say more about this important issue in
section . Point relates to Descartess views about the nature of the
material world, specically, to a view called the theory of primary and
secondary qualities. We shall discuss this topic in the next section. First,
however, we need to assess Descartess proof of the material world.
Given what Descartes has argued in his previous Meditations, his proof
of the material world has a certain cogency. It makes a logical use of the-
ses for which he has already argued, such as the existence of a perfect
God and the real distinction between mind and body. It is a natural devel-
opment and culmination of the overall argument of the Meditations.
This is not to say, however, that the proof is invulnerable, even within
the framework of Descartess own theism and dualism. One possible ob-
jection, for example, was made by the German Rationalist philosopher
Leibniz (). Leibniz questioned Descartess claim that a perfect
God would not allow us to be deceived about the causes of our sensory
experiences. Perhaps, Leibniz suggested, God has certain good reasons
Meditation VI
for allowing us to be so deceived, just as he has good reasons for allowing

certain other evils. Perhaps the deception even works for our benet.
Even if Descartes could have answered this sort of criticism, there
remains a much more fundamental objection. In the proof of the real
distinction, as we have seen, God serves only to guarantee the truth of
the clear and distinct perceptions contained in the proof. By contrast, the
proof of the material world does not rely on God only as a guarantor of
clear and distinct perceptions. Rather, it relies on God, in step , to rule
out the skeptical hypothesis that despite our powerful inclination to
believe that our sense experiences are caused by bodies, they may really
be caused by an evil deceiver, or in some other bizarre way. So if there
may be no God, then this skeptical hypothesis remains in place: no rea-
son has been given that removes it. But the thesis that there is a God
depends on the proofs of the existence of God in Meditations III and V,
and those proofs are unsuccessful. Therefore, Descartess proof of the
material world must be regarded as a failure. The implication is that the
radical doubt of Meditation I remains unanswered. For all Descartes
knows, there may be no material world, and his sensory experiences may
be produced by the deceiver or in some other extraordinarily bizarre way;
for all he can be absolutely certain of, the only thing that exists is the
purely thinking self of the cogito.
The signicance of the failure of Descartess proof of the material
world for philosophy after Descartes can hardly be understated. Most
major philosophers since Descartes have wrestled with the problem of
the external world that he uncovered but failed to solve. Some, like
Locke in the seventeenth century and Russell in the twentieth century,
have tried to show that material things are the causes of our sense expe-
riences by a quasi-scientic argument to the best explanation that
does not appeal to God. Some, like Berkeley (), John Stuart
Mill () and A. J. Ayer (), have argued that material
things are in some way composed out of, or constructed from, the sense
experiences themselves. Others, like David Hume (), have

Gottfried Leibniz, Critical Remarks Concerning the General Part of Descartes Prin-
ciples, p. .

For critical discussion of this approach, see Georges Dicker, Berkeleys Idealism, pp.
and pp. and Perceptual Knowledge, pp. .

For critical discussion of this approach, see Dicker, Berkeleys Idealism, pp.
and pp. and Perceptual Knowledge, pp. .
Descartes
held that Descartess problem is insoluble and that skepticism is the

only rational position. Immanuel Kant () argued that the
mere ability to order our own past experiences in time proves that those
experiences are caused by successive states of physical objects that we
perceive in space. Still others philosophers, like Dewey ()
and Wittgenstein (), have argued that Descartess doubt of
the material world is illegitimate and should not have arisen in the rst
place.
4. Descartes on the Nature of the Material World

Although Descartess proof of the existence of material things is a failure,
it does not follow that his views about their nature are false or uninter-
esting. It is to these views that we turn in the present section.
As we saw in our discussion of the wax example of Meditation II, Des-
cartes held that the only property included in a clear and distinct concep-
tion of body is extension, or three-dimensionality. For Descartes, bodies
are, so to speak, units of extension, or units having three-dimensional
shape and size. His conception of matter is thus a purely geometrical one.
As he puts it in The Principles of Philosophy:
The nature of body consists not in weight, hardness, colour, or

the like, but simply in extension... . [T]he nature of matter, or
body considered in general, consists not in its being something
which is hard or heavy or coloured, or which aects the senses
in any way, but simply in its being something which is ex-
tended in length, breadth and depth. (CSM I , SPW , AT
VIII A )
This conception of matter raises many questions, of which the most fun-
damental are probably these three:

For an analysis of Humes skeptical position on the belief in an external world, see
Georges Dicker, Three Questions about Treatise .., and Humes Metaphysics
and Epistemology, pp. .

For a defense of this approach, see Georges Dicker, Kants Refutation of Idealism,
, Kants Refutation of Idealism: a Reply to Chignell, , Kants Refuta-
tion of Idealism: Once More unto the Breach, , and Kants Theory of Knowledge,
pp. .
Meditation VI
. How do properties other than extension and its modes, such as color,
taste, smell, sound, and heat and cold, t in? Are they merely illu-
sions of our senses? Or do they in some way also belong to bodies, and
if so how?
. How does matter dier from space, given that space, like Cartesian
body, seems to be nothing but extension?
. How does one body dier from another body, given that bodies are
just units of extension? What demarcates one such unit from an-
other?
In the three subsections to follow, we shall consider Descartess answers

to each of these questions.
4.1 Primary and Secondary Qualities

In order to answer the rst question, Descartes resorted to a theory that
can be traced back to the Greek atomist Democritus ( b.c.), that
was accepted by nearly all seventeenth-century philosophers and scien-
tists, and that is, in one form or another, still defended by many philoso-
phers todaythe theory of primary and secondary qualities. The classic
exposition of this theory is given by John Locke () in his Essay
Concerning Human Understanding, written some years after Descartess
death. In the following discussion we shall draw substantially on Lockes
account, in order to put Descartess version of the theory into better per-
spective. But since Lockes account, despite its insightfulness, is ambig-
uous and sometimes even inconsistent, our exposition of it will be
somewhat modernized and idealized, and we will take note of its principal
inconsistency at the end of this subsection.
The theory makes a fundamental distinction between two kinds of
qualities or properties of material things: primary qualities and sec-
ondary qualities. Primary qualities are dened as those that a thing must
possess in order to be a physical object. According to Locke, these include
shape, size, solidity, mobility, and number. Imagine, says Locke, that a
grain of wheat is divided in half, that each half is again divided in half,
and so on as far as you like. No matter how long this process of division
continues, each part must still have some size, some shape, solidity, the
ability to be moved; and, of course, there is always some number of parts.

John Locke, An Essay Concerning Human Understanding, Book , chapter .
Descartes
This shows, as Locke puts it, that these qualities are inseparable from [a]
body, in what estate soever it be. In other words, the primary qualities
are those that are supposed to belong necessarily to any portion of
matter whatsoever: they serve to dene matter. As we might expect in
view of Descartess geometricized concept of matter, his list of primary
qualities is shorter than Lockes. For Descartes, the primary qualities
include only extension and its modes, plus motion. Notice, then, that for
Descartes solidity is not one of the primary qualities.
The list of secondary qualities includes at least: colors, tastes, smells,
sounds, heat and cold. Descartes adds solidity to the list. As we shall see,
however, there are strong reasons for agreeing with Locke that solidity is
a primary quality. So let us leave solidity out of our list of secondary qual-
ities; we shall try to justify this stipulation in the next subsection. The
classication of heat and cold as secondary qualities is also problematic
because, at least today, we think of these as falling along the scalar mag-
nitude that we call temperature, and arguably any body must have some
temperature or other, in which case temperature is a primary quality and
heat and cold are determinates of it. But Locke may have thought of heat
and cold rather as opposites that characterize some but not all bodies, in
which case they would qualify as secondary qualities. That said, we shall
not pursue this point further, since it does not aect the basis contours
of the theory.
The denition of secondary qualities, which is the most important fea-
ture of the entire theory, is that secondary qualities are only capacities,
powers, or dispositions of physical objects to cause experiences of color,
taste, smell, sound, and heat and cold in a perceiver, under normal condi-
tions of observation. For example, the color red is only the capacity of
certain objects (e.g., ripe tomatoes and re engines) to cause experiences
of red in a normal perceiver under normal light; and a sweet taste is only
the capacity of certain objects (e.g., sugar cubes and ripe strawberries) to
produce experiences of sweetness in a perceiver under normal conditions
for gustatory perception. To introduce a piece of contemporary termi-
nology, the denition of secondary qualities says that they are disposi-
tional properties. A dispositional property (or, for short, a disposition) is
a capacity to cause or to undergo some change. For example: explosiveness
is a dispositional property of gunpowder, fragility is a dispositional prop-
erty of glassware, solubility in water is a dispositional property of sugar,

Locke, Essay, Book , chapter , section .
Meditation VI
and corrosiveness is a dispositional property of some acids. The denition

of secondary qualities, then, says that they are dispositional properties
of a special kind: they are capacities to cause certain experiences in a
sentient organism. Of course, in order for such capacities to operate, the
right conditions must be present. For instance, in order for a red objects
capacity to cause experiences of red to operate, there must be an appro-
priately situated perceiver equipped with normally functioning eyes, a
sucient amount of light, and so forth. But the same thing is true of
other kinds of dispositional properties; for example, in order for the cor-
rosiveness of an acid to operate, there must be a piece of metal of a cer-
tain sort, oxygen, a temperature not so cold that the acid freezes or so
hot that the metal melts, and so forth.
As the terms primary and secondary may suggest, the theory also
holds that the secondary qualities depend upon the primary qualities. In
other words, it holds that whether or not an object has, say, the capacity
to cause experiences of red in normal light depends upon whether or not
it possesses certain primary qualities. Locke, who accepted the seven-
teenth-century atomism (corpuscularism) of scientists like Boyle and
Galileo, ties this point to their atomic theory of matter. He says that the
capacities in question depend upon the primary qualities of the objects
atomic parts. As he puts it:
[S]uch qualities which in truth are nothing in the objects them-

selves but powers to produce various sensations in us by their
primary qualities, i.e., by the bulk, gure, texture and motion of
their insensible parts, as colours, sounds, tastes, etc. These I call
secondary qualities.
This contains two statements followed by Lockes definition of sec-

ondary qualities: () that physical things are composed of insensible
parts (i.e., atoms or corpuscles), each of which have the primary qual-
ities: a certain size, shape, solidity and velocity and () that because
of the primary qualities of these atomic parts, physical things have
powers (the word Locke always uses for capacities or dispositions) to
cause perceivers to experience sensations of color, sound, taste, smell,
heat and cold. It is these specic powers or capacities that Locke calls
secondary qualities.

Locke, Essay, Book , chapter , section .
Descartes
What is the rationale for the theory of primary and secondary qual-
ities? There are at least two reasons that support the theory. The rst is
that it provides a way of tting the secondary qualities into a scientic
account of matter. Scientic descriptions of matter generally do not refer
to secondary qualities. For example, no atomic theory of matter, whether
ancient, seventeenth-century, or contemporary, ascribes such qualities
as color, smell, or taste to the atoms themselves. These qualities are
regarded by science as phenomena to be explained by the atomic structure
of matter and must therefore not be ascribed to the atoms themselves.
How, then, do these qualities relate to matter as it is described by physics?
The answer proposed by Locke is a plausible one: colors, tastes, smells,
and so on are really only capacities that objects have to aect perceivers
in certain ways, because of their particular atomic structure. For ex-
ample, suppose that a certain tabletop is brown. Then Locke would say
that its being brown consists in its having the capacity to cause experi-
ences of brown in us under normal conditions and that it has this
capacity because of the molecular structure of its surface.
Notice that the scientic motivation for the theory of primary and
secondary qualities need not necessarily be tied to an atomic theory of
matter. We can see this by looking at Descartess version of the theory.
Unlike Locke, Descartes was not an atomist; for he believed that exten-
sion can be innitely divided, whereas atoms would have to be indivisible
particles. Nonetheless, scientic concerns are no less central to his ver-
sion of the theory than to Lockes. For Descartes, the scientic descrip-
tion of matter is the clear and distinct one of matter as extension. Color,
taste, smell, sound, and heat and cold are no part of this conception. So
again, how do these qualities relate to matter as conceived by science?
Descartess answer, in his Principles of Philosophy, is that
[T]he properties in external objects to which we apply the terms

light, colour, smell, taste, sound, heat and cold-as well as the
other tactile qualities ... are, so far as we can see, simply var-
ious dispositions in those object which make them able to set up
various kinds of motions in our nerves ... which are required to
produce all the various sensations in our soul... . [L]ight, colour,
smell, taste, sound and tactile qualities ... are nothing else in the
objectsor at least we cannot apprehend them as being any-
thing elsebut certain dispositions depending on size, shape
and motion. (CSM I , SPW , AT VIII A )
Meditation VI
In other words, secondary qualities are merely dispositions of objects

which scientically considered are composed only of shapes and sizes in
motionto aect perceivers in certain ways.
The scientic aspect of the theory of primary and secondary qualities
explains some of the remarks Descartes makes after giving his proof of
the physical world. As we saw in the previous section, immediately after
the proof Descartes suggests that physical things are not exactly as we
perceive them by our senses. A bit later he makes this more specic,
saying that bodies have certain dierences that explain, but do not re-
semble, the colors, tastes, sounds, smells, temperatures and hardnesses
that we perceive. These points are implications of the theory of primary
and secondary qualities. Descartes means that the experiences of color,
taste, smell, and so on that bodies cause in us do not resemble the var-
ious shapes, sizes and motions in virtue of which bodies (have the dispo-
sitions to) cause such experiences in us. Essentially the same point was
more prominently and famously made by Locke, who put it by saying
that the ideas of secondary qualities do not resemble anything in mate-
rial objects. Locke also claimed that, by contrast, the ideas of primary
qualities do resemble those qualities themselves.
The second reason that supports the theory of primary and secondary
qualities comes to light if one reects on the following question: What is
a color, such as the redness of an apple, when no one is looking at it? At
such times, the apple is obviously not causing any experiences of red. To
put this point more strikingly, the apple does not then look red; for there
is no one to whom it looks red. What then is it for an apple to be red,
when no one is looking at it? This question is much like the better-known
puzzle, Does a tree that falls in a forest with no one there to hear it make
a sound? A plausible answer to this puzzle is that if by making a sound
one means having the capacity to cause an auditory experience in a per-
ceiver (by producing vibrations in the air, called sound waves, which can
stimulate the auditory sense receptors), then the falling tree does make a
sound; for it does have this capacity (while it is actually breaking and
crashing through the forest and into the ground.) But if by making a
sound one means actually causing an auditory experience (or sounding
in some way), then the tree makes no sound, because it cannot cause an
auditory experience unless a perceiver is present. In other words, the tree
does not then sound in some way, because it cannot sound in any way at

Compare Kenny, Descartes, p. .
Descartes
all unless it sounds some way to someone. Thus the tree does make a
sound; but this only means that it has the capacity or disposition to cause
an auditory experience in a perceiverthat if a perceiver were present
the tree would cause such an experience or would sound in some way to
that perceiver. Now, the theory of primary and secondary qualities oers
a parallel answer to our question about color (and to strictly analogous
questions that could be asked about taste, smell, and heat and cold). A
red object that no one is looking at is still red, because it still has the dis-
position or capacity to cause an experience of red in a normal perceiver
under normal conditions; it is still true of the object that if a perceiver
were to look at it under normal conditions, it would cause such an expe-
rience in him or her. But of course, a red object that no one is looking at
doesnt then cause any such experience. It doesnt then look or appear
red, since there is no one to whom it looks or appears red. Thus, if we ask,
what exactly is a things redness, insofar as it exists on the things surface
whether or not anyone is looking at it, the answer is that it is the power,
capacity, or disposition of the thing to cause an experience of red in a
perceiver (to look or appear red to a perceiver) under appropriate condi-
tions. The thing has this disposition whether or not someone is looking
at it; so it is red whether or not it is being perceived. And if we add that it
has this disposition because of the molecular structure of its surface,
then we have Lockes view that a things secondary qualities depend upon
the primary qualities of its atomic parts.
Although the dispositional account of secondary qualities just given
has much to recommend it, it needs supplementation. For as it stands it
implies that colors, sounds, tastes, and smells are only powers, capac-
ities or dispositions of a special sort, but that view is too narrow, for at
least two reasons. One reason, which might be called the liking-it ar-
gument, is this. Suppose I tell you that my favorite color is aquamarine,
or that I like the taste of chocolate. Surely I am not then saying that I
like the power, capacity or disposition of things to produce certain vi-
sual or gustatory sensations. The mere capacity or disposition is not
what I care about, like, or enjoy. Nor does what I say mean simply that I
like things that have these powers or dispositions. For aquamarine could
be my favorite color even though I disliked aquamarine things because
(say) they all came in ugly shapes, and I could like the taste of chocolate
even though I disliked things that had that taste because they all con-
tained an unpalatable stung. But if, when I say that aquamarine is my
favorite color or that I like the taste of chocolate, I am talking neither
Meditation VI
about certain dispositions nor about the things that have those disposi-
tions, then what am I talking about? The answer seems to be that I am
talking about a distinctive, qualitative aspect of aquamarine things and
of chocolatean aspect that I cannot define verbally but that I am
familiar with from experience, and that cannot be identied with or
reduced to a mere power or capacity.
The other reason why the dispositional account of secondary qualities
is too narrow comes out if one reects on the following passage from
Berkeleys Principles of Human Knowledge:
[C]an there be a nicer strain of abstraction than to distinguish

the existence of sensible objects from their being perceived, so as
to conceive them existing unperceived? Light and colours, heat
and cold, extension and gures, in a word, the things we see and
feel, what are they but so many sensations, notions, ideas or im-
pressions on the sense; and is it possible to separate, even in
thought, any of these from perception? For my part, I might as
easily divide a thing from itself.
One way to read this passage is as a challenge, which we may call Berke-
leys Challenge, to distinguish the existence of any sensible quality (what
Berkeley called its esse) from its being actually perceived (what he called
its percipi). Berkeleys challenge says, in eect: you cannot even con-
ceive a color existing apart from its being seen, a sound existing apart
from its being heard, an odor existing apart from its being smelled, etc.
Berkeley uses his challenge to support his own idealist view that all the
qualities we perceive are really only ideas or sensations in our minds.
Berkeleys challenge seems ineffective when applied to primary
qualitieswhy cant one conceive, for example, of an unperceived cube
simply as a shape that lls a certain volume, or of an unperceived solid
object as one that excludes other bodies from the space it occupies? But
the puzzle about the tree that falls in a forest with no one to hear it shows
that the challenge has at least some force with respect to sounds, and
analogous puzzle cases can be described for tastes, smells, and even
colors. To see this better, consider the following thought experiment.

George Berkeley, A Treatise Concerning the Principles of Human Knowledge, p. .

For a defense of this way of reading the passage, see Dicker, Berkeleys Idealism,
pp. .
Descartes
Suppose that there existed only one red thing in the entire universe, that
a normal perceiver saw that thing in sunlight at a time t, but that no one
at all saw it at time t. Then there is a sense in which the quality, red(ness),
would exist at t but not at t; likewise there is a sense in which tastes,
sounds, and smells would vanish from the universe if (or at least during
the times when) no one perceived any objects having those qualities. We
suggest that one test that any adequate view about the nature of colors,
tastes, sounds and smells should meet is to make clear in what sense this
is true. The purely dispositional account of them that we have so far given
does not meet that test, since objects continue have the dispositions to
cause the relevant experiences even when they are not being perceived.
In order to meet the test, we should recognize that colors, tastes,
smells, and sounds have a certain complexity; they have two aspects. On
the one hand, they are capacities, powers, or dispositions of things to
appear in certain ways to sentient beings under certain conditionsthis
we may call their dispositional aspect. (Note that this aspect corresponds
to Lockes denition of a secondary quality.) On the other hand, they are
the manifestations of those dispositionsthis we may call their manifest
aspect. For example, the color red is the disposition of some objects, such
as re engines, trac stoplights, and ripe tomatoes, to look red to nor-
mal perceivers in standard light. But red(ness) is also the distinctive qual-
itative event or episode, involving consciousness, that occurs when the
conditions are right for this disposition to manifest itself, that is, when
there is a perceiver with normal vision looking at a red object in normal
light. Accordingly, there are two true answers to the question: What is
the color red? One answer is that red is the disposition of certain objects
to look red to normal perceivers in normal light. The other answer is that
red is the event, occurrence or episode that constitutes the manifestation
of this disposition. We have chosen to speak of two aspects simply for
lack of a better term. This term should not obscure the fact that the two
aspects are related in a specic, unmysterious manner: the manifest
aspect is the manifestation of the dispositional aspect, just as the process
of corrosion is the manifestation of corrosiveness.
The distinction between the dispositional and the manifest aspect
answers both the liking-it argument and Berkeleys Challenge. For what I
like or enjoy if I like (say) red is the manifest aspect of red, not its dispo-
sitional aspect. Furthermore, what ceases to exist when no one perceives
red is again its manifest aspect, not its dispositional aspect. But since the
manifest aspect of red is in one sense identical with the quality red, it
Meditation VI
provides a sense in which, just as Berkeley held, red(ness) exists only by

being perceived. On the other hand, the distinction does not concede to
Berkeley his idealist view that red, as well as any other sensible quality, is
nothing but an idea; it preserves the objectivity of sensible qualities. For
their dispositional aspect continues to exist in material things whether
those things are perceived or not. Furthermore, even the manifest aspect
of red is not purely subjective, since it exists only when an objects ca-
pacity to look red is manifested or exercisedthat is, only when an object
like a ripe tomato or a re engine actually looks red to someone. Thus, an
objects looking red can be categorized as a relational property of the
object, that is, as a property that it can have only by standing in some
specic relation to another object, akin to the property of being a sibling,
being North of, or being taller than. In this case, the specic relation
involved is that of looking red to someone; in other cases it might be that
of smelling sweet to someone, or tasting bitter to someone, and so on.
Locke was sensitive to the complexity of secondary qualities. But for
reasons that would take us too far aeld to pursue here, he does not make
the distinction between their two aspects. Instead, he denes secondary
qualities simply as powers or dispositions to cause sensations or ideas of
color, taste, smell, sound, and heat or cold in us, and he sticks consistently
to that denition. But his answer to the question What is a color, a taste,
a smell, a sound, or an instance of hear or cold? is inconsistent. He oscil-
lates between equating colors, tastes, sounds, smells, heat and cold with
only the dispositions to cause in us sensations or ideas of colors, tastes,
sounds, smells, heat and cold (i.e., only with what his denition calls sec-
ondary qualities, or what we have called the dispositional aspect of a sec-
ondary quality), and equating them with only those ideas themselves (i.e.,
only with what we have called the manifest aspect of a secondary quality,
except that he construes the manifest aspect as an idea caused in a certain
way rather than as a relational property of the object). Sometimes Des-
cartes too is read as identifying colors, tastes, sounds, smells, and heat/
cold with mere ideas, as did Galileo and other scientists of the day, though
it is debatable whether this is a completely fair reading of him.

See the next note.

There is a fuller account of primary and secondary qualities, of Lockes treatment of
them, and of the distinction between the dispositional aspect and the manifest aspect of sec-
ondary qualities, in Dicker, Berkeleys Idealism, chapter . See also Georges Dicker, Primary
and Secondary Qualities, , reprinted in Walter E. Creery, George Berkeley, pp. .
Descartes
4.2 Matter, Space, and Solidity

We now turn to the second question raised by Descartess geometri-
cized concept of matter, namely, the question of how matter diers
from space. This question is bound to cause trouble for Descartes; for
how does body, if it is a substance whose whole nature is extension or
three-dimensionality, dier from space, whose nature seems to be ex-
actly the same? By dening matter as mere extension, has not Descartes
abolished the dierence between matter and space?
Descartes attempts to explain the dierence between the two in his
Principles of Philosophy, where he writes:
There is no real distinction between space ... and the corporeal

substance contained in it; the only dierence lies in the way in
which we are accustomed to conceive of them. For in reality the
extension in length, breadth and depth which constitutes a space
is exactly the same as that which constitutes a body. The dier-
ence arises as follows: in the case of a body, we regard the exten-
sion as something particular, and thus think of it as changing
whenever there is a new body; but in the case of a space, we at-
tribute to the extension only a generic unity, so that when a new
body comes to occupy the space, the extension of the space is
reckoned not to change but to remain one and the same, so long
as it retains the same size and shape and keeps the same position
relative to certain external bodies which we use to determine the
space in question ... [I]f a stone is removed from the space or
place where it is, we think that its extension has also been
removed from that place, since we regard the extension as some-
thing particular and inseparable from the stone. But at the same
time we think that the extension of the place where the stone
used to be remains, and is the same as before. (CSM I ,
SPW , AT VIII A )
Descartes is here making two points: () the dierence between a body

and the space or place it occupies lies only in the way we conceive of the
two; () the body is that which we conceive as removable from the space,
while the space is conceived as that which must remain when the body is
removed from it. The rst point already seems wrong, because it denies
that there is a real dierence (one independent of how we conceive
things) between matter and space. But let us focus on the second point.
Meditation VI
The key idea in it is that of removing something, say a stone, from the
space or place that it occupies. If that idea fails to make sense, then Des-
cartess attempt to preserve even a merely conceptual distinction
between space and body collapses. But the trouble is precisely that this
idea, considered in light of Descartess identication of both body and
space with mere extension, does not make sense. For suppose we ask:
When we conceive that the stone is removed from the place it occupied,
exactly what, according to Descartes, do we conceive as being removed
from the place? Descartes cannot answer that we conceive that some-
thing colored, or smelly, or noisy, or tasty, or hot or cold, is being removed
from the place; for these secondary qualities are only capacities or dispo-
sitions grounded in the extension of bodies, and so we can understand
their removal only if we can make sense of the removal of the exten-
sion itselfwhich is precisely our diculty. Nor can Descartes say that
we conceive that something solid is being removed from the place,
because according to him, solidity also is merely a capacity, grounded in
extension, to cause tactile sensations in a perceiver. The upshot is that
we seem to be removing only a volume from the placeor a volume from
a volume. But this seems quite unintelligible.
It would seem, then, that some property or properties other than ex-
tension must also be included in any adequate conception of matter. One
property that naturally suggests itself is solidity, and this is indeed what
Locke proposed:
[S]olidity ..., of all others, seems to be the idea most intimately

connected with and essential to body, so as nowhere else to be
found or imagined, but only in matter... . By this idea of solidity
is the extension of body distinguished from the extension of
space: the extension of body being nothing but the cohesion or
continuity of solid, separable, moveable parts; and the extension
of space, the continuity of unsolid, inseparable, and immoveable
parts.
But, you may ask, what exactly is solidity? Well, one step toward clari-
fying this notion is to distinguish it from hardness. A hard body is one
whose parts are not easily moved relative to each other; a soft body, like
a pillow or a quantity of water, is one whose parts are easily displaced

Locke, Essay, book , chapter , sections and .
Descartes
relative to each other. Solidity, by contrast, consists in lling up space or

in excluding all other bodies from the space actually occupied. It is not
that a solid mass cannot be moved from the place it occupies, but rather
that no other mass can occupy that same place so long as the rst one
still occupies it. So, even a soft body is solid. As Locke put it:
All the bodies in the world, pressing a drop of water on all sides,
will never be able to overcome the resistance which it will make,
as soft as it is, to their approaching one another, till it be moved
out of their way: whereby our idea of solidity is distinguished
both from pure space, which is capable neither of resistance nor
motion, and from the ordinary idea of hardness.
Solidity, then, is virtually the same as impenetrability; whereas hardness

is immovability relative to the immediately surrounding matter. To quote
once more from Lockes insightful discussion:
Solidity is hereby also dierenced from hardness, in that solidity

consists in repletion, and so an utter exclusion of other bodies
out of the space it possesses: but hardness, in a rm cohesion of
the parts of matter making up masses of a sensible bulk, so that
the whole body does not easily change its gure.
Insofar as Locke is oering a criterion for distinguishing body or

matter from space, his view seems plausible. But we must not rashly con-
clude that he has provided an adequate denition of body or of matter,
for to say that something is solid seems to mean just that it excludes other
bodies from the space it occupies. Thus if Locke is read as oering a de-
nition, then his denition seems to dene body in terms of solidity and
solidity in terms of body, which is circular. Possibly this circularity could
be avoided by dening solidity in a slightly dierent way, as the property
of excluding all but one extended thing (including a portion of empty
space) from a place. Then a body could be dened as an extended and solid
thing. Extension would still have to be included in the denition since, as

Locke, Essay, section .

Locke, Essay, section . Descartes sometimes says that the concept of body involves
its excluding other bodies from the space it occupies (e.g., CSM II , SPW , AT VII ).
The question, however, is whether he is entitled to say this, given his purely geometrical
account of the nature of matter.
Meditation VI
Descartes and Locke both thought, a thing could not be a body without
being extended. But the denition would not be circular since solidity
would not be dened merely as the ability to exclude a body from a place,
and a body could be thought of as a an extended thing that excludes all
but one extended item (namely, itself) from the place that it occupies.
4.3 Bodies as Substances versus Bodies as Modes of Substance

Descartess inability to dierentiate matter from space leads naturally to
our nal question about his theory of the material world: how does one
body dier from another, given Descartess view of matter as extension?
Descartes seems to have given two incompatible answers to this ques-
tion. One answer, which he gives in his Principles of Philosophy, is the tra-
ditional one that one body diers from another because they are two
dierent substances. The other answer, which he implicitly gives in the
Synopsis of the Meditations, is the surprising one, which we already
described in chapter , that two or more bodies are just dierent prop-
erties of a single substance. We shall examine these opposed passages and
argue that only the latter answer is compatible with Descartess doctrine
that matter is merely extension.
In his Principles, Descartes denes a real distinction as one between
two or more substances. (This is why he calls his proof that mind and
body are two dierent substances a proof of the Real Distinction between
them.) He then states his principle that we can perceive that two sub-
stances are really distinct simply from the fact that we can clearly and
distinctly understand one apart from the other, from which he says it
follows that each and every part of [extended or corporeal substance],
as delimited by us in our thought, is really distinct from the other parts
of the same substance (CSM I , SPW , AT VIII A ). Descartess
language is not careful here. In particular, the phrase really distinct
from the other parts of the same substance is poorly chosen, since he
has just dened a real distinction as one between dierent substances, not
between parts of the same substance. Nevertheless, it seems evident
that Descartes here means that any portion of extension that we can
clearly and distinctly conceive apart from other portions is really distinct
fromand hence a dierent substance fromother substances of the
same kind (from the other parts of extension taken as a generic to-
tality). This interpretation is conrmed in the next sentence, which
reads, in part: Each of us ... is really distinct from every other thinking
Descartes
substance and from every corporeal substance; for the phrase every
corporeal substance implies that there is a plurality of corporeal sub-
stances, each of which can only be a dierent body. This reading is fur-
ther conrmed in the next principle, where Descartes says:
A dierent case ... is the distinction by which the mode of one

substance is dierent from another substance... . An example of
this is the way in which the motion of one body is distinct from
another body... . It seems ... appropriate to call this kind of dis-
tinction a real distinction, since the modes in question cannot be
clearly understood apart from the really distinct substances of
which they are modes. (CSM I , SPW , AT VIII A )
Here Descartes gives, as an example of a case where the mode of one

substance is distinct from another substance (i.e., other than the one to
which that mode belongs), the case where the motion of one body is dis-
tinct from another body (i.e., other than the one to which that motion
belongs). This implies that dierent bodies are dierent substances. The
passage also implies that dierent bodies are really distinct substances,
which entails (both because substances is plural and because a real
distinction is by denition one between two or more substances) that
dierent bodies are dierent substances.
So far, we have merely made the textual point that Descartes says that
dierent bodies are dierent substances. The question, however, is whether
he is entitled to say this, given his view that bodies are nothing but exten-
sion. The answer, we suggest, is that he is not. To see why, note rst of all
that saying that dierent bodies are dierent extended substances obvi-
ously assumes that there can be more than one extended substancethat
there can be a plurality of extended substances. We may safely assert, then,
() If there cannot be more than one purely extended substance

(i.e., substance consisting of nothing but extension), then the
dierence between two or more bodies cannot be that they are
dierent purely extended substances.
But the diculty is precisely that Descartes is not entitled to say that
there can be more than one purely extended substance; for although ex-
tension can be divided into regions, it does not have separable parts. It
makes no sense to talk of separating one part of pure extension from
Meditation VI
another anymore than of separating one part of space from another:

Where would you put it? We may also assert, then,
() Extension does not have separable parts.
But this implies that there could not be a plurality of purely extended
substances, for what would make them dierent? We cannot say that just
being dierent regions or portions of extension would make them dif-
ferent, because even a single extended substance must have dierent
spatial portionsotherwise it wouldnt be extended. We cannot appeal
to properties other than extension, like color or hardness. The only thing
that could make two purely extended substances dierent would be that
they could, at least in principle, be separated or disjoined from each
other. But this is precisely what we cannot say about portions of pure
extension. It seems clear, then, that we can say,
() If extension does not have separable parts, then there cannot

be more than one purely extended substance.
There follows from () and (), however,
() There cannot be more than one purely extended substance.
Further, there follows from () and (),
() The dierence between two or more bodies cannot be that

they are dierent purely extended substances.
What, then, is the difference between two or more bodies? The only
answer possible for Descartes seems to be that the bodies are dierent
modes or accidental properties of a single, all-encompassing extended
thing or substance (res extensa). This single extended substance, which
might be called matter-space or space-matter, constitutes the entire
physical universe. There may be many thinking substances or mindsas
many as there are beings who could say or think Cogito, ergo sumbut
there can be only one extended substance.
This one-substance view of the physical world is, interestingly
enough, implicit in the passage from the Synopsis that we mentioned. In
the context of explaining what would have to be done in order to prove
that the soul is immortal, Descartes writes:
Descartes
[T]he premises which lead to the conclusion that the soul is

immortal depend on an account of the whole of physics. This is
required for two reasons. First, we need to know that absolutely all
substances ... are by their nature incorruptible ... [S]econdly, we
need to recognize that body, taken in the general sense, is a sub-
stance, so that it too never perishes. But the human body, in so far
as it diers from other bodies, is simply made up of a certain con-
guration of limbs and other accidents of this sort; whereas the
human mind is not made up of any accidents in this way, but is a
pure substance. For even if all the accidents of the mind change, so
that it has dierent objects of the understanding and dierent
desires and sensations, it does not on that account become a dif-
ferent mind; whereas the human body loses its identity merely as
a result of a change in the shape of some of its parts. And it follows
from this that while the body can very easily perish, the mind is
immortal by its very nature. (CSM II , SPW , AT VII )
We have met (part of) this passage before; it is the place where Descartes
uses the Argument from Change to show that the mind is a substance.
And we shall meet the passage again in Section , when we oer an over-
all assessment of Cartesian Dualism. But what interests us now is what
Descartes here says about body. He draws a major contrast between
body, taken in the general sense and the human body. The former
refers to the totality of extension. Descartes here regards this totality as
a single, unitary, all-encompassing extended substance, or res extensa. He
says that a complete physics could show that this substance, as well as
the soul, is naturally incorruptible and therefore never perishes. By
contrast, he says that a human body is made up of accidents, so that it
can easily perish when some of those accidents are changed. In other
words, Descartes here regards a human body as merely an aggregate of
accidental properties or modes, which perishes when some of these
modes are altered. Now there is every reason to suppose that Descartes
would treat other sorts of bodies in the same way as human bodies. But
in that case, Descartes is here implying that all particular bodies (e.g.,
chairs, rocks, planets, etc., as well as human bodies) are (clusters of) acci-
dental properties, or modes. And if we ask what they are modes of, then the
answer is that they are modes of the one, all-encompassing, incorrupt-
ible extended substance. The view that emerges is that there is only one
extended substance and that particular physical objects are modes of it.
Meditation VI
On this one-substance view of matter, the famous illustration of the

wax in Meditation II would have to be interpreted dierently than it was
in chapter , section . Appearances to the contrary notwithstanding, the
illustration would have to be seen as not concerned with what is required
for a piece of wax to continue existing as the same wax or even for a chunk
of matter to continue existing as the same chunk of matter. Rather, the
piece of wax would have to be seen, as suggested in chapter , section
., as a miniature model that Descartes uses to represent the whole
physical universe. The point of the illustration would then be that as long
as the physical universe retains the determinable property of extension,
it continues to exist as the same physical universe. The underlying sub-
stance that remains the same throughout a process of change would be
identical with the determinable property of extension. More strictly
speaking (since the term property is often used to stand for the univer-
sal, which is then distinguished from its particular instances, exemplica-
tions, or occurrences), the underlying substance would be identical with
an instance of the determinable property of extensionthe sole or
unique instance, given the one-substance view. Since a determinate prop-
erty cannot occur (cannot be exemplied or instantiated) unless its cor-
responding determinable property occurs, the principle that there are no
free-oating properties (i.e., that if there is a property, then there must
be a substance to which it belongs) would still hold, although it would
now mean that if a determinate property occurs, then the appropriate
determinable property must occur. Conversely, since a determinable
property cannot occur unless one of its determinates occurs, the prin-
ciple that a substance cannot exist without having some properties would
also hold, although it would now mean that the determinable properties
of extension and thought cannot occur unless there exist determinate
shapes and sizes and determinate thoughts.
This way of interpreting Descartes receives some textual support from
his Principles of Philosophy, part (especially principles and ). In
those places, Descartes seems to equate a substance with its dening at-
tribute. Thus, in principle he says that each substance has one princi-
pal property which constitutes its nature or essence (CSM I , SPW
, AT VIIIA ). And in principle he says,
Thought and extension can be regarded as constituting the

natures of intelligent substance and corporeal substance; they
must then be considered as nothing else but thinking substance
Descartes
itself and extended substance itselfthat is, as mind and body.

(CSM I , SPW , AT VIIIA )
Here Descartes seems to be using the terms extended substance,

thinking substance, mind, and body to stand partly for types, rather
than for individual substances, minds, or bodies; and to be saying
that these types are identical with their dening attributes. The passage
also strongly suggests, however, that an individual mind is identical with
an instance of the attribute of thought, and that an individual body is
identical with an instance of the attribute of extension.
As noted earlier, on the view that a substance is identical with (an
instance of) its dening attribute, Descartess version of the substance
theory would not be vulnerable to the standard empiricist objection to
substance that we discussed, namely that substance is in principle unper-
ceivable. It might be argued that this objection would still apply, because
a determinable property like extension cannot be perceived by itself,
that is, without perceiving some determinate of it like a square shape or
a circular shape. But even if it is true that a determinable cannot be per-
ceived without perceiving one of it determinates, this does not mean that
the determinable property is unperceivable. It means, rather, that to per-
ceive a determinate property is also to perceive the corresponding deter-
minablefor example, to perceive red is also to perceive color, to perceive
squareness is also to perceive shape.
On the other hand, if (a) physical substance is identical with (an
instance of) the determinable property of extension, then the substance
theory can no longer even purport to give an account of the continuing
existence of an ordinary physical object like a piece of wax or a stone. For,
as we argued in chapter , section ., retaining some shape or other, or
some size or other, or some color or other, would not be enough for an
ordinary physical object, for example your car, to continue existing even
though all of its determinate properties had changed. But since on the
view in question there is really only one immense physical object, the
physical world itself, of which pieces of wax, stones, human bodies, cars,
and so forth are only modes or accidental properties, the need to provide
an account of such things identity through time is less pressing. Of
course, a philosopher who wished to give an account of the identity
through time of ordinary physical objects (despite their modal status)

See chapter , section ..
Meditation VI
could still do so, in terms of some feature other than substance (e.g., spa-
tio-temporal continuity under a sortal, as also mentioned in chapter ).
But Descartes himself seems, in the end, not to have been very interested
in this question.
It is noteworthy that Spinoza (), a close student of Des-
cartess philosophy who felt that Descartes had failed to follow out con-
sistently the logical implications of his own thought, proposed just such
a one-substance view of the physical world as we have described. Spi-
noza, however, went even further. In his major work, entitled Ethics
Demonstrated in Geometrical Order, he argued that the one extended sub-
stance is also the one and only thinking substance, which he called God
or Nature. The resulting view is that the entire universe consists of only
one substance that is both thinking and extended.
To summarize: Descartes apparently held two incompatible views
about bodies. One view, which is the traditional view derived from Aris-
totle and the Scholastics, is that each body is a distinct substance. The
other view, which foreshadows Spinozas theory, is that bodies are modes
of a single substance. Only the latter view seems consistent with Des-
cartess doctrine that matter is merely extension. Finally, it should be
noted that the former view becomes once again defensible if, as we sug-
gested in the previous subsection, solidity is included in the conception
of matter; for unlike a mere portion of space, one solid body can be sepa-
rated from another, even if the two happen to be in contact.
5. Dualism and the Problem of Interaction

Descartess dualism of mind and matter implies a certain conception of
human beings or persons. A person is a composite entity, made up of two
distinct components: a mind, or soul, and a body. Of course, Descartes
gives a certain priority, so to speak, to the mind; for he holds that he
could exist without his body, from which he concludes (with the help of
the principle that nothing without which a thing can exist is included in
its essence) that his body is not part of his essence. By contrast, he would
certainly deny that he could exist without his mind, and he insists that
his mind is his essence. And in the Synopsis of the Meditations, as we saw
in the last section, he goes so far as to say that while the mind is a sub-
stance, the body is composed only of accidents. But despite Descartess
emphasis on the mind, he recognized that it would be fantastic to deny
Descartes
that human beings are, at least during their earthly lives, embodied and
that their embodiment is a salient fact of their existence. So, even while
maintaining his sharp dualism, he tried to do justice to the close and
intimate relation that each of us bears to his or her own body. This comes
out, for example, in the fact that Descartes lists I have a body as the
rst and most obvious particular fact that he can accept about the mate-
rial world and I am very closely joined to this body as the second. It also
comes out in the language that he goes on to use in order to describe this
close union:
Nature also teaches me, by these sensations of pain, hunger, thirst

and so on, that I am not merely present in my body as a pilot in his
ship, but that I am very closely joined and, as it were, intermin-
gled with it, so that I and the body form a unit. If this were not so,
I, who am nothing but a thinking thing, would not feel pain when
the body was hurt, but would perceive the damage purely by the
intellect, just as a sailor perceives by sight if anything in his ship is
broken. Similarly, when the body needed food or drink, I should
have an explicit understanding of the fact, instead of having con-
fused sensations of hunger and thirst. For these sensations of
hunger, thirst, pain and so on are nothing but confused modes of
thinking which arise from the union and, as it were, intermingling
of the mind with the body. (CSM II , SPW , AT VII )
Here and in some other places, Descartes goes for far as to imply that
a purely thinking thing could not have sensations of pain, hunger, or
thirst; rather, such sensations belong to a hybrid category that emerges
from the union of mind and body. John Cottingham has emphasized this
aspect of Descartess thought, calling it Cartesian Trialism. The basic
idea, which Cottingham explains lucidly and in detail, but which he also
concedes does not t well with Descartess radical dualism, is that
although there are only two kinds of substancemental and physical
some of the attributes of substance, especially sensation, fall into a third
category that is neither purely mental nor purely physical but somehow
emerges from a substantial union of mind and body.

John Cottingham, Descartes, pp. .

For a stimulating discussion of this kind of view, see Alison Simmons, Re-Humanizing
Descartes, .
Meditation VI
Descartess desire to do justice to the fact of embodiment raises one of

the most perplexing problems of his entire philosophy: What exactly is
this relationship of substantial union between the mind and the body of
a person? To appreciate why this question is such a dicult one for Des-
cartes, remember that according to him the mind and the body are radi-
cally dierent and even opposite: mind is a thinking and nonextended
substance, body is an extended and nonthinking substance. How then
can there be a close union or intermingling of the two? What can this
union possibly consist in? John Cottingham puts the diculty this way:
The notion that two dierent substances can unite to form a

single thing is not, in itself, obscure or problematic. We are famil-
iar nowadays, for example, with the idea that hydrogen and ox-
ygen can unite to form water; furthermore, this substantial
union generates emergent propertieswater has properties
such as that of being drinkable which were not present in its con-
stituent elementsand this (though it is not of course Des-
cartess own example) might be thought to give some grip on the
Cartesian notion that events like sensations emerge or arise
when mind and body are united, even though they are not part of
the essence of either res cogitans or res extensa. Nevertheless,
Descartes himself clearly felt that his notion of the substantial
union of mind and body presented problems. For mind and
body, as dened throughout [Descartess] writings, are not just
dierent, but utterly incompatible substances: in terms of their
characteristics, they mutually exclude one another, since mind is
dened as nonextended and indivisible, whereas matter is by its
nature extended and divisible. And it is not easy to see how in-
compatible items can be, in any intelligible sense, united. As
Descartes rather ruefully put it:
it does not seem to me that the human mind is capable of
conceiving at the same time the distinction and the union
between body and soul, because for this it is necessary to
conceive them as a single thing and at the same time to con-
ceive them as two things, and this is absurd.

John Cottingham, Descartes: Metaphysics and the Philosophy of Mind, p. .
The quotation from Descartes is from his letter of June , to Princess Elizabeth
(CSMK , AT III .
Descartes
Moreover, Descartess account of the mind-body union is ambiguous

and at times even inconsistent. The chief ambiguity concerns whether
the mind is joined to the whole body or only to a certain part of it. There
are texts supporting each interpretation. In some places, such as the pas-
sage just cited, Descartes seems to hold that the mind is intermingled
with the entire body; in other places he says that it is joined to the body
only by virtue of being united with a particular part of the brain (com-
pare, e.g., paragraphs and of the Sixth Meditation). Although
strongly inuenced by Descartess overall position, later dualists like
C. D. Broad () and C. J. Ducasse () have invariably
assumed that only the latter view is plausible; indeed the former view is
seldom, if ever, mentioned in recent presentations of dualism. No doubt
this is partly because it seems quite impossible to take Descartess talk of
the intermingling of mind and body literally; for an nonextended mind
cannot be intermingled with an extended body, as, for example, pow-
dered milk can be intermingled with water: the mind would have to be
composed of small extended parts that could be dispersed throughout the
body. So it seems that Descartess talk of intermingling can at best be
taken as a way of describing how we seem to ourselves to be joined to our
bodies, rather than as a literal, true account of the connection. Indeed, in
the passage from Meditation VI just quoted, Descartes himself hedges his
statements, saying that the mind is as it were intermingled with the
body. It must be admitted, nevertheless, that sometimes Descartes does
straightforwardly assert that the mind is joined to the whole body, as in
the following passage from The Passions of the Soul, a late work in which
he tried to describe the union of mind and body in some detail.
[T]he soul is really joined to the whole body, and ... we cannot
properly say that it exists in any one part of the body to the
exclusion of the others. (CSM I , SPW , AT XI )
But whether the mind is joined to the whole body or only to some part
of the brain, a much more fundamental question remains: how is the
mind joined to (part of) the body? What exactly is the special relation-
ship between the mind and the (part of the) body to which it is joined?
On the standard interpretation of Descartess position, he had a denite
answer to this question. This is that mind and body are causally related.
More specically, Descartes is usually interpreted as having held that
there is a two-way causal relation between mind and body: () the mind
Meditation VI
causally aects the body, and () the body causally aects the mind.
Mind-to-body causation occurs especially in voluntary action; for ex-
ample, your willing or deciding to raise your arm is an act of mind that
causes a physical occurrence: your arm goes up. Body-to-mind causation
occurs especially in sense perception; for example, a clap of thunder is a
physical occurrence that causes physical changes in your ears, nerves and
brain, which in turn cause a conscious, auditory experience in your mind
(which, incidentally, must on the dualistic view obviously not be con-
fused with your brain). This theory of the mind-body relationship is
called dualistic interactionism (where dualistic refers to the twon-
ess or duality, of mind and matter, and interactionism, to the causal
interaction between the two).
Dualistic interactionism has an undeniable appeal, for at least two rea-
sons. First, it seems to harmonize very well with our experience: it cer-
tainly seems to us that our decisions and volitions frequently cause our
bodies to behave in various waysthat our minds do, to an important
extent, control our bodies. And it seems just as obvious that what hap-
pens to our bodies causes a multitude of dierent conscious experiences
in our minds. Mind-body interaction thus seems to be a continuing and
pervasive feature of our ordinary experience. Second, dualistic interac-
tionism provides a plausible way to understand the close relationship
that each of us has to his or her own body and to no other body. What
makes a certain body my body is that I have a direct control over it that
no one else has and that what happens to it has a direct eect on me that
it has on no one else. Only I can cause that body to move by a mere voli-
tion; only I will feel pain if that body is injured. The body which I call
mine, then, is the body over whose movements I have direct control and
whose vicissitudes have a direct eect on me.
Despite its initial appeal, however, dualistic interactionism also faces
deep diculties. The most striking and famous one surfaces as soon as
we ask the following question: How can the mind causally aect the body,
and vice-versa? Remember, again, that according to Descartes the mind
is a thinking substance that has no spatial dimensions, while body is a
thoughtless, three-dimensional substance. How do these two things
interact? We obviously cannot say that the one makes contact with or
pushes the other, for this would require that they both have spatial di-
mensions and spatial surfaces. We cannot say that the mind applies a
physical force to the body, since force is mass times acceleration and the
mind has no mass. We cannot say that the body imparts thoughts to the
Descartes
mind, since the body has no thoughts. When we actually try to conceive
the interaction between an extended nonthinking thing and a nonex-
tended thinking thing, it seems quite inconceivableas inconceivable as
driving a nail with an immaterial hammer, or denting a spirit with rock.
In fairness, we should note that Descartes tried to give an account of
mind-body interaction. He did so in the texts where he says that the
mind is joined to a particular part of the brain. There he species the part
of the brain in question, and describes the immediate consequences of its
interaction with the mind. In The Passions of the Soul, he says:
[T]he part of the body in which the soul directly exercises its
functions is not the heart at all, or the whole of the brain. It is
rather the innermost part of the brain, which is a certain very
small gland situated in the middle of the brains substance and
suspended above the passage through which the spirits in the
brains anterior cavities communicate with those in its posterior
cavities. The slightest movements on the part of this gland may
alter very greatly the course of these spirits, and conversely any
change, however slight, taking place in the course of the spirits
may do much to change the movements of the gland. (CSM I ,
SPW , AT XI )
The gland to which Descartes is here referring is called the pineal gland
(also called the conarion). Descartes believed that this tiny gland must be
the point where mind and body interact, because it is the only part of the
brain that does not have a double (CSMK , ). He theorized that
it is surrounded by a rened material substance called animal spirits,
which interact via tube-like nerves with the muscles that control various
parts of our bodies. The pineal gland itself, he thought, interacts directly
with the mind: a given event in the mindsay, the willing to raise ones
armmoves the gland, causing it to drive the animal spirits through the
nerves to the muscles, which then contract, thus raising the arm. Con-
versely, a given bodily eventsay, stimulation of the retinadrives the
animal spirits through the nerves to the pineal gland, whose oscillation
then aects the mind in such a way that it has a certain visual experience.
As Descartes put it:
[T]he small gland which is the principal seat of the soul is sus-
pended within the cavities containing these spirits, so that it can
Meditation VI
be moved by them in as many dierent ways as there are percep-

tible dierences in the objects. But it can also be moved in var-
ious dierent ways by the soul, whose nature is such that it
receives as many dierent impressionsthat is, it has as many
dierent perceptions as there occur dierent movements in this
gland. And conversely, the mechanism of our body is so con-
structed that simply by this glands being moved in any way by
the soul or by any other cause, it drives the surrounding spirits
towards the pores of the brain, which direct them through the
nerves to the muscles; and in this way the gland makes the spirits
move the limbs. (CSM I , SPW , AT XI )
Although the scientic details of this pineal gland theory are obso-
lete, the theory is nonetheless very instructive; for it provides a vivid,
concrete illustration of two points. First, the dualistic interactionist who
holds that the mind interacts with the brain must say that some such
account as Descartess is literally true, even if the details of Descartess
own account are wrong. In other words, the interactionist is committed
to the view that certain specic brain events cause what we may call
mental events and that certain specic mental events cause certain
specic brain events. Second, such causal interaction of brain events and
mental events is extremely hard to comprehend. In other words, even if
we assume that the specic events involved have been correctly identi-
ed, the mystery of how the mental events cause physical ones, and vice-
versa, remains. John Cottingham puts the problem this way:
What strikes the reader here is not so much the wealth of obso-
lete physiological detail (modern readers will readily be able to
substitute electrochemical events in the cerebral cortex for Des-
cartess movements of the pineal gland and animal spirits) as
the way in which that physiological detail is expected to mesh
with events in the nonphysical realm of the soul. Descartes has
managed to supply a host of mechanisms whereby movements,
once initiated in the pineal gland, can be transferred to other
parts of the brain and body; but he does not seem to have tackled
the central issue of how an incorporeal soul can initiate such
movements in the rst place. And the same problem will apply
when the causal ow is in the other direction. Descartes devotes
a lot of attention to the physiological mechanisms whereby bodily
Descartes
stimuli of various kinds cause changes in the nervous system and

brain which dispose the soul to feel emotions like anger or fear.
But he does not explain how mere brain events, however complex
their physiological genesis, could have the power to arouse or
excite events in the mental realm.
Another recent writer, Fred Feldman, gives an illustration that makes the
mystery seem even worse when it is put in more up-to-date terms than
Descartess pineal gland theory:
[S]uppose someone dips his toe into a swimming pool to test

the water. The cold water quickly cools the skin on his toe, and
changes the temperature of the nerve endings that are scat-
tered there. Then some sort of electrical charge ows up the
nerve, jumping across various gaps between one nerve and the
next. Perhaps there are stages in which the electrical event causes
some chemical change, which in turn causes a suitable electrical
event in the next nerve. This purely physical chain of events
eventually reaches some part of the brain. Here is where the
trouble begins. How does it make the last step, the one that gets
it from the physical apparatus of the nervous system, and into
the mind?. . . . How does the body nally inuence the mind?
How does all the electrical and chemical activity in the nervous
system nally bring about that distinctive feeling of cold that
reveals that the water is too chilly for swimming? Many philoso-
phers would say that this alleged causal connection is simply
inconceivable.
The causal connection in the other direction is no easier to
understand. Suppose you have been thinking about a certain
friend. You decide to call her on the telephone. Precisely how
does this decision, apparently a mental event, give rise to the
rst physical event in the causal chain that ultimately leads to
the movements of your ngers? Your mind cannot rub against
the nearest nerves in your brain, nor can it give o heat, or light,
or chemicals, or an electric charge. Only a physical object could
do such things. Thus, causal interaction is equally perplexing,
whether it is mind-body interaction, or body-mind interaction.

Meditation VI
And Cartesian dualism cannot be true unless such interaction

happens all the time... . [T]his ... constitutes the most impor-
tant objection to Cartesian dualism. . . . [H]istorically, . . . no
other argument crops up more frequently in the anti-dualistic
literature. It is a formidable problem.
Indeed, mind-body interaction has, from Descartess own day to ours,

often been seen as the weakest point of his philosophy. Princess Elizabeth of
Bohemia, with whom Descartes had an important philosophical correspon-
dence, wrote to him that she found it easier to conceive that the mind itself
is extended, than to conceive how it interacts with extension. And Gassendi,
in his Objections to the Meditations, powerfully attacked the idea of interac-
tion (see, e.g., CSM II, , AT VII ). Still other contemporaries
of Descartes proposed various ways of amending his system so as to avoid
interaction altogether. For example, the Flemish thinker Arnold Geulincx
() and the French philosopher Nicolas Malebranche ()
proposed a theory called Occasionalism. According to this theory, there is no
interaction between mind and body. Instead, on each separate occasion
where mind-body interaction seems to take place, it is really God who brings
about the eect. So, for instance, when you decide to raise your arm, God
himself steps in and raises it for you; when you accidentally touch a hot
stove, God produces pain in you. The great German rationalist Leibniz pro-
posed a somewhat dierent theory, called Preestablished Harmony. Leibniz
agreed with the Occasionalists that there is no interaction between mental
and physical events. However, he thought it unbetting to Gods perfection
that he should need to constantly intervene so as to coordinate these events.
Accordingly, he theorized that God, when he created the world, also insti-
tuted a universal correspondence or harmony between mental and physical
events, somewhat like a skillful clockmaker who builds two clocks in such a
way that they will always keep exactly the same time. Today, philosophers
who work on the mind-body problem regard both Occasionalism and the
Preestablished Harmony as theories that are too fantastic to warrant se-
rious consideration. We mention them here because they vividly show how
radical Descartess problem of interaction seemed to be to his own contem-
poraries. The problem seemed so intractable that even theories that made
God the intermediary between the mind and body of a person were seen by
leading thinkers of the day as genuine improvements over interactionism.

Fred Feldman, A Cartesian Introduction to Philosophy, pp. .
Descartes
Indeed, some scholars have suggested that Descartes himself was ulti-
mately drawn to occasionalism. Any attempt to defend an occasiona-
list reading of Descartes, however, would have to contend with passages
like the following, where he seems unequivocally to arm mind-body
interaction:
That the mind, which is incorporeal, can set the body in mo-
tion is something which is shown to us not by any reasoning
or comparison with other matters, but by the surest and plain-
est everyday experience. It is one of those self-evident things
which we only make obscure when we try to explain it in terms
of other things. (Letter to Arnauld, July , ; CSMK )
It is not necessary for the mind to be a body, although it has the
power of moving the body. (Fifth Replies, CSM II , AT II )
[Burman] But how can this be, and how can the soul be aected
by the body and vice-versa, when their natures are completely
dierent? [Descartes] This is very dicult to explain; but here
our experience is sucient, since it is so clear on this point that
it just cannot be gainsaid. (Conversation with Burman, CSMK ,
AT V )
I would say: . . . the brain alone can directly act on the mind.
(Letter to Regius, May , CSMK , AT III )
Is the problem of interaction a fatal diculty for dualistic interaction-

ism? Although the diculty is clearly a major one, there is reason to
think that it is not insuperable. For some recent dualists (e.g., Broad and
Ducasse) have suggested ingenious ways of dealing with it. Broad argued
that the problem arises from the assumption that cause and eect must
be similar, and that this assumption is false even in the purely physical
realm: for example, a magnetic force can attract iron lings, but there is
no resemblance between the magnetic eld and the motions of little
pieces of iron. To this, it may be countered that a magnetic force and the
motions of iron lings are at least both physical phenomena that can be
investigated by physics, whereas this is not the case for mind and body.
But Ducasses reply to the diculty is especially incisive. He writes:

C. D. Broad, Mind and its Place in Nature, p. . Broads own example is that of a draft
causing a cold.
Meditation VI
[T]he objection that we cannot understand how a psychical [i.e.

mental] event could cause a physical one (or vice versa) has no
basis other than blindness to the fact that the how of causa-
tion is capable at all of being either mysterious or understood
only in cases of remote causation, never in cases of proximate
causation. For the question as to the how of causation of a
given event by a given other event never has any other sense
than through what intermediary causal steps does one cause the
other.
We can illustrate Ducasses point with an example. Suppose someone is

riding a bicycle, and that it is asked, How does pushing on the pedals cause
the bicycle to move? The answer is that pushing on the pedals causes them
to move, which causes the front sprocket to turn, which causes the chain
to rotate, which causes the rear wheel to turn, which causes the bicycle to
roll forward. In giving this answer, we have specied several causal steps
that are the intermediaries between the pushing of the pedals and the
movement of the bicycle. In specifying these intermediary causal steps,
we have explained how pushing the pedals causes the bicycle to move, or
how the cause produces the eect. But now suppose somebody were to
ask: But how does pushing on the pedals cause them to move? The
answer is that it just does. We cannot explain how it does, because there
are no intermediary causal steps between the cause and the eect that we
could cite in order to answer this how question. We have reached a case
of proximate causation, one where there are no intervening steps between
the cause and the eect; and such cases cannot be explained. Instead,
they have to be simply accepted as brute facts: Thats just how the world
works.
How, exactly, does Ducasses point bear on the problem of interaction?
The answer is that dualistic interactionists have always taken mind-body
interactions to be cases of proximate causation. Therefore, the question
of how a given mental event causes a brain eventor how a given brain
event causes a mental eventis just as illegitimate as the question of
how pushing on the pedals causes them to move. To make this more con-
crete, suppose (as C. D. Broad suggested in his book Mind and Its Place in
Nature) that the immediate eect of a mental event on the brain is to
lower the electrical resistance of certain synapses and to raise that of

C. J. Ducasse, In Defense of Dualism, p. .
Descartes
others, so that the directions of the paths taken by nervous impulses

through those synapses are aected. And suppose it is asked, How does
the desire to raise ones arm cause it to go up? The answer, in outline, is
that this mental event alters the distribution of resistances in certain
synapses, which aects the direction of the paths taken by certain
nervous impulses, which causes certain muscles to contract, which causes
ones arm to go up. But now, if someone asks how desiring to raise ones
arm causes the electrical resistance in some synapses to go down and in
others to go up, then the only answer that can be given is that this is a
case of proximate causation. As such, it cannot be explained. It must be
simply accepted as a brute fact.
At this point, you may want to raise the following objection. Surely,
when a bicycle rider causes the pedals to move by pushing on them, there
is an explanation of their motion, in terms of laws of nature involving
force, momentum, transfer of energy, or the like. The interactionist
cannot sensibly deny this. But then, how can he say that such cases of
proximate causation are just inexplicable brute facts?
A possible reply is that the interactionist need not at all deny that
laws of nature, involving various scientic concepts, govern cases of
proximate causation. But this does not mean that the proximate causal
connections are any the less brute facts. For what are these laws of
nature? According to one important and plausible view, they are basi-
cally regularities between events, to the eect that whenever an event of
a given kind occurs, then an event of another kind occurs. For instance,
whenever a given force is applied to a body (under certain speciable
conditions), it moves. The key point to grasp here is that the law merely
states that the case in question is an instance of what always happens in
similar cases. It does not explain how applying a force to a body (e.g.,
pushing on a bicycle pedal) makes it move, except in the sense of citing
a regularity in nature, to the eect that whenever such a force is applied,
the body moves. This regularity itself, and particular cases of the regu-
larity, remain as much brute facts as ever. The most we can do by way of
explaining the regularity is to derive it from a still more general regu-
larity. But then the latter is itself a brute fact: it is just a very general
principle describing the way nature happens to behave. All that the
interactionist is claiming is that, among the regularities in nature, there
are some to the eect that whenever a certain kind of mental event oc-
curs, a certain kind of brain event occurs; and some to the eect that
Meditation VI
whenever a certain kind of brain event occurs, a certain kind of mental

event occurs. Such regularities are no more nor less mysterious than
those linking purely physical events.
It must be acknowledged that the regularity conception of a law of
nature just sketched, which derives from David Humes famous and
inuential analysis of causality, is not uncontroversial. It is, however,
one of the leading philosophical theories about laws of nature. To the
extent that it is plausible, it supports Ducasses solution to the problem
of interaction (though we should note that Ducasse himself did not ac-
cept Humes theory and so would probably not have approved of our in-
voking it to support his solution to the problem).
6. An Assessment of Cartesian Dualism

In this chapters rst section, we presented Descartess proof of dualism
and Arnaulds objection to it and argued that Descartes gives a plausible
reply to that objection. In the last section, we presented the problem of
interaction and argued that this problem does not constitute a decisive
objection to dualistic interactionism. Does this mean that we should ac-
cept Cartesian Dualism (i.e., Descartess entire theory of mind and
matter as two dierent but causally interacting substances)? No: that
would be a hasty conclusion for at least three reasons. First, although
we have discussed the most famous objection to dualistic interaction-
ism, there are other objections to it that we have not considered. Sec-
ond, there are other dualistic theories that we also have not considered,
such as epiphenomenalism, a theory maintaining that physical events
cause mental events but that mental events never cause physical events.
Third, many (perhaps most) contemporary philosophers reject all forms
of dualism in favor of dualisms main rival, which is materialism. Materi-
alism is the view that whatever exists is material. According to materi-
alism, there are no such things as mental substances, mental events,
or mental states. If minds exist at all, then they can only be brains,
or brains together with nervous systems. If thoughts, desires, feel-
ings and sensations exist (something which is actually denied by the
most radical materialist philosophers, called eliminative materialists),
then they are merely neurological events, states or processes occurring
within the brain and central nervous system. Today, there are many
Descartes
dierent versions of materialism, with dierent names such as elimi-

native materialism, logical behaviorism, reductive materialism,
and some versions of functionalism.
We shall not examine materialist theories of mind, nor shall we exam-
ine in detail other objections to dualistic interactionism or other dualistic
theories like epiphenomenalism. It is not that these topics are unimpor-
tant or uninteresting. On the contrary, the area of philosophy to which
they belong, called Philosophy of Mind, is a very active and exciting eld.
Our reason for not going more deeply into the philosophy of mind is sim-
ply that doing so would take us too far from our study of Descartes and
would indeed require another book.
Nevertheless, we shall conclude this book by oering an overall assess-
ment of Cartesian Dualism. In the course of our discussion, we shall
touch on a few of the issues raised in contemporary philosophy of mind.
But let us start with a note of caution: no nal judgment as to the truth
of Cartesian Dualism can reasonably be made apart from the detailed
examination of other objections and rival theories that we have decided
to forego. So, our assessment will not take the form of pronouncing upon
the truth or falsity of the theory. Rather, we shall defend the more modest
thesis that even if Descartess proof of dualism is sound, this does not
have the weighty implications that are often associated with Cartesian
Dualismimplications prized by dualists and shunned by materialists.
To defend this thesis, we shall use a four-point strategy. First, we shall
present a pruned, or streamlined, version of Descartess proof of du-
alism that preserves the basic intuitions or insights on which Descartess
own argument turns. Second, we shall argue that even if this stream-
lined argument is sound, it does not have the signicant implications
commonly associated with dualism. Third, we shall argue that appear-
ances to the contrary notwithstanding, these implications do not follow
from Descartess own argument, eitherthat his argument does not
have weightier implications than the streamlined one. Finally, we shall
show that Descartes was aware of thisthat he was not fooled by his
arguments impressive look.
Descartess proof of dualism turns on three basic ideas: the conceiv-
ability of the separate existence of the mental and the physical, the
principle that what is conceivable is logically possible, and the principle
that the logical possibility of separate existence implies distinctness or
nonidentity. Thus, our streamlined version of his argument starts from
the premise
Meditation VI
() If X is any conscious state (i.e., any thought, desire, mental

image, sensory experience, or the like) and Y is any physical state
or process (e.g., a brain event), then it is conceivable that X exists
and Y does not exist.
This premise resembles, and rests on the same considerations as, Des-
cartess richer premise (stated as line of his proof of dualism in section
of this chapter) that he can clearly and distinctly conceive himself, as a
thinking and nonextended thing, existing apart from his body, as an ex-
tended and nonthinking thing. For brevitys sake, we have used the term
conceivable, instead of clearly and distinctly conceivable, in stating
premise of the streamlined argument. But of course, the qualication,
clearly and distinctly, should be understood in () and throughout the
rest of the streamlined argument. The second premise of that argument is
() If it is conceivable that X exists and Y does not, then it is log-

ically possible that X exists and Y does not.
This premise is very close to Descartess premise (stated as line of his

proof of dualism) that if he can clearly and distinctly conceive X existing
apart from Y, then X really can exist without Y, at least by Gods power.
The third premise is
() If it is logically possible that X exists and Y does not, then X is

not identical with Y.
This is close to Descartess premise (stated as line of his proof of du-

alism) that if X can really exist without Y, no matter by what power, then
X and Y are really two dierent things. The basic point, as before, is that
nothing could possibly exist without or apart from itself. Of course, nei-
ther () nor () of the streamlined argument makes any reference to
Gods power. For Descartess philosophical theology, as previously sug-
gested, plays no essential role in his proof of dualism (unlike his proof of
the physical world). Rather, the key ideas in the proof of dualism are the
three listed at the start of this paragraph, each of which is expressed in
just one of the streamlined arguments three premises. The conclusion
that follows from those premises is
() If X is any conscious state and Y is any physical state or

process, then X is not identical with Y.
Descartes
This conclusion is weaker than Descartess own conclusion that he, as a

thinking thing, is not identical with his body. For one thing, () asserts
only a dualism of states or processes, as opposed to a dualism of substances.
For another, () contains no reference to I, or to the self. Below, when we
come to our point that Descartess own argument does not have signi-
cantly weightier implications than does our streamlined argument, we
shall argue that these dierences do not matter very much. For the mo-
ment, however, we need only emphasize that () does, of course, assert a
dualism of the mental and the physical; for it implies (assuming that any
conscious states and physical states exist at all) what all materialists deny,
namely, that in addition to material things and processes, there are purely
mental existents. Since () asserts a dualism of states or properties rather
than of substances, it is the minimum that a dualist must hold; it is a
minimal form of dualismbut () is still dualism, since it means that
mental states are irreducibly dierent from anything physical. It also di-
rectly contradicts one of the most widely-held recent versions of materi-
alism, the identity theory, also called reductive materialism. This theory
doesnt deny that there are mental states, but it just baldly identies
these with brain states or processes. Thus the streamlined argument
brings Cartesian considerations to bear directly against a major form of
contemporary materialism.
Our streamlined version of Descartess proof of the Real Distinction
can doubtless be attacked. For example, some philosophers today
would question (), for () rests on the general principle that whatever
is (clearly and distinctly) conceivable is logically possible. But these
philosophers think that there are counter-examples to this principle.
One favorite example is that of Goldbachs Conjecture. Christian Gold-
bach () was a Russian mathematician who conjectured that
every even number greater than is the sum of two prime numbers. So
far, every such number which has been tested has been found to be
the sum of two prime numbers; but, of course, this does not mean that
some very large even number that no one has yet considered isnt the
sum of two primes. Furthermore, no mathematician has been able
either to prove or to disprove Goldbachs Conjecture. On the strength
of this example, some philosophers argue that the general principle
behind () is false. The argument is that since Goldbachs Conjecture is
a statement of mathematics, it is logically necessary if true at all,
and logically impossible if false. But we can conceive of Goldbachs
Conjecture being either true or false (i.e., we can conceive of both
Meditation VI
alternatives), say, by conceiving that some supercomputer tells us that

it is true or that it is false. Therefore, we can conceive of something
that is not even logically possible: the truth of Godlbachs conjecture
(in case the conjecture is actually false), or the falsity of the conjecture
(in case it is actually true). So, the principle that whatever is conceiv-
able is logically possible is false. It seems, however, that this objection
commits a rather simple mistake. It is one thing to conceive of a com-
puter telling us that Goldbachs conjecture is true (or false) and quite
another to conceive that the conjecture itself is true or false. I can con-
ceive of a computer telling me, + = ; it does not follow that I can
conceive + = . As Descartes would put it, such a conception would
hardly be clear and distinct. It seems, therefore, that the principle
behind ()a principle that many philosophers rely uponcannot be
so easily dismissed.
We need not insist, however, that the streamlined argument can with-
stand this or other objections that might be raised against it; for although
the streamlined argument seems quite powerful, our point is not that it
is sound. Rather, our point (the second one in the four-point strategy
outlined above) is that even if the streamlined argument is sound, the
implications are not very signicant. The basic reason for this is that
even if the logical possibility that conscious states might exist apart from
physical states shows that conscious states are not identical with physical
states, it does not show that it is causally possible for conscious states to exist
apart from physical states. Thus, for all that the argument can show, con-
sciousness may exist only as an eect of certain sorts of brain processes;
that is, it may never exist apart from the physical processes that cause it.
This may be so even if (as interactionism maintains but epiphenome-
nalism denies) mental events causally aect brain processes or events;
for it may be that those mental events are in turn caused by prior brain
events and never occur unless caused by these brain events. A contempo-
rary philosopher, the late James Cornman, drawing on Broads sugges-
tions about mind-body interaction, suggested one way in which this
might happen. Cornman suggests that the causal relationships between
mental and brain events could be as shown in Figure ..
If this is the way mental and brain events are interrelated, then there
are two notable consequences. First, as Cornman emphasizes, a scientist
could give an adequate neurological explanation of human thought and
behavior without ever mentioning the mental events; for although M
causally aects S, M is in turn caused by B. So a neurophysiologist who
Descartes
Figure 6.1 Figure 6.2
Source: Slightly modied from James W. Cornman, Keith Lehrer, and George Pappas,
Philosophical Problems and Arguments: An Introduction, d ed. (Indianapolis and Cambridge:
Hackett Publishing Company, ), p. . Chapter of this book contains an excellent
discussion of the various objections and alternatives to dualistic interactionism.
knew that B was occurring could predict and explain S and N without
ever mentioning Meven, indeed, if the neurophysiologist were a mate-
rialist who denied Ms existence and believed that matters stand as in
Figure .. Second, since on this view M would not occur unless B (or
some other brain event) produced it, consciousness is totally dependent
on a functioning brain for its existence.
Once these points are recognized, some of the objections that recent
materialists raise against dualism seem strangely empty and rhetorical.
For example, one leading contemporary materialists chief objection to du-
alism is that it postulates an irreducibly psychical something, a ghost
stu ... or ripples in an underlying ghost stu, dierent from anything
recognized by the natural sciences, thereby violating the principle of
theoretical simplicity. (The principle of theoretical simplicity, also known
as Ockhams Razor, is a very general methodological principle saying that
a theory should not multiply entities unnecessarily, that is, should not
postulate more dierent kinds of entities than are needed to explain the
facts that the theory is designed to explain. A very simple illustration
would be that since tornadoes can be explained in terms of certain atmo-
spheric conditions, a meteorological theory should not postulate evil
spirits to explain their occurrence. A historical example of the principles

J. J. C. Smart, Sensations and Brain Processes, p. , p. . in C. V. Borst, ed., The
Mind-Brain Identity Theory (New York: St. Martins Press, ), pp. .
Meditation VI
application would be the abandonment of the phlogiston theory after the

discovery of the role of oxygen in combustion.) But if dualism concedes
that mental existents may be completely causally dependent on physical
ones, then it is hard to see much force in this complaint. To be sure, if the
argument for dualism is sound, then conscious states cannot be reduced to
brain states: in that sense they are indeed irreducibly mental or psychical.
But phrases like irreducibly psychical something, ghost stu, and the
like also suggest a kind of independence of mental existents from physical
ones which is wholly belied by such a dualism. And while the notion of
theoretical simplicity is a complex one, its applicability here is question-
able; for it seems reasonable to hold that its sphere of application is limited
to entities that are held to play some essential explanatory role (especially
in scientic explanation), though there is no independent evidence for
their existence. But then the principle of simplicity cannot be legitimately
invoked to rule out a mental event like M, since M plays no essential ex-
planatory role and there is independent evidence for its existence (i.e., the
streamlined argument together with the evidence of introspection).
The other side of the coin, of course, is that if the dualist concedes that
the mental may be completely causally dependent on the physical, then it is
hard to see the great signicance of dualism; for such a dualist agrees with
contemporary materialists who think that a complete neurophysiology
could causally explain human thought and behavior and that there is good
reason to believe that consciousness ends when brain activity stops. Such a
dualist insists only on the logical possibility of consciousness existing apart
from brain activity and its consequent nonidentity with brain activity. This
is to abandon some of the ideas commonly associated with dualism, for
example, that the mental component of human beings stands in the way of
any adequate scientic explanation of human thought and behavior and
oers some positive evidence for immortality. Thus, the objections of
materialists are blunted only at the cost of foregoing some of the most
cherished ideas commonly associated with dualism. The disagreement
between dualists and materialists seems to become at least partly a verbal
one, over labels like spiritual and material, and associated ideological
beliefs that do not logically follow even if the case for dualism is sound.
Appearances to the contrary notwithstanding, much the same assess-
ment holds for Descartess own proof of dualism. This may at rst not be
apparent, because Descartess proof of the Real Distinction between
mind and body may seem to yield a much stronger form of dualism than
does the streamlined argument. But this is a misleading impression.
Descartes
First of all, remember that Descartess argument does not try to show
that the mind does, at any time, exist without the body. Rather, as we
saw in Section , it only tries to show that the mind could exist without
the body. As we saw, Descartes holds that the mere logical possibility that
mind can exist apart from body is enough to show that they are dierent
things, or that there is a real distinction between them.
But even in light of a proper grasp of the Real Distinction, it may still
seem that Descartess argument has weightier implications than our
streamlined version of it. There are two reasons for this. The rst is that
Descartess own argument is intended to establish a dualism of sub-
stances, rather than merely a dualism of states. The second is that Des-
cartes calls one of these substances I, or self, and says that it constitutes
his essence, thereby giving it a special priority over the other.
Neither of these points, however, is as signicant as it may seem. Let
us consider the second point rst. Suppose we grant that Descartess
argument shows that it is logically possible that I exist without my body
and that it is logically possible that I exist as only a mind, so that my
minds existence is logically sucient for my existence. Suppose we also
agree that it is not logically possible that I exist without my mind, that
is, that my minds existence is logically necessary for my existence.
Finally, suppose we agree that I should denote only that which is both
logically necessary and sucient for me to exist. It then follows that I
denotes my mind, not my body. It also follows, by the principle that a
things essence comprises whatever is necessary and sucient for its
existence, that my essence is my mind, not my body. To put it more for-
mally, suppose that we grant the entire argument:
. My minds existence is logically sucient for my existence.

. My minds existence is logically necessary for my existence.
. I denotes only what is both logically necessary and logically su-
cient for my existence.
. I denotes my mind and not my body (from , , and ).
. A things essence contains only what is both logically necessary and
logically sucient for its existence.
. My essence is my mind, not my body (from , , and ).
Now even if we accept this reasoning completely, it only establishes a

point of logical, linguistic, or conceptual priority. It has no tendency
whatsoever to show that what I denotes, or that what is comprised in
my essence, is causally independent of my body.
Meditation VI
Consider next the fact that Descartess own argument purports to

establish a dualism of substances, rather than merely a dualism of states.
The term substance is associated, especially within philosophy, with a
certain permanence, durability, even indestructibility. This can make it
look as if Descartess argument would, if it is sound, establish a dualism
that supports traditional beliefs about the causal independence of the
soul from the body and its prospects for immortality. Now Empiricist
criticisms of the concept of substance have cast a shadow over Descartess
reliance on this concept. These criticisms also go a long way toward
explaining why, today, dualism is usually formulated in terms of mental
versus physical states or events, rather than mental versus physical sub-
stance. But the important point for us is that even if dualism is formu-
lated in terms of mental and physical substance, our basic assessment of
Descartess proof of dualism still holds good; for the argument that Des-
cartes gives for the substance theory, namely the argument from change,
certainly does not show that a mental substance would be causally inde-
pendent of the body, or that it would be indestructible, or that it must
outlast the body, or anything of the sort. It follows that nothing in the
concept of substance, at least insofar as its employment is supposed to
be justied by this argument, entails that a mental substance and its
states must be causally independent of a material substance and its
states. Further, there do not seem to be any other arguments concerning
substance that would support such a conclusion. Certainly, the proof of
the Real Distinction contributes nothing itself that might strengthen the
concept of substance; it only uses the concept of substance that was
already operative in Meditation II. So far as Descartess argument for
dualism goes, then, mind might be totally dependent on matter.
Indeed, this would be the case if mental states or events were com-
pletely causally dependent on physical ones. To see this, recall once again
that just as there cannot be a property without a substance to which it
belongs, so there cannot be a substance without any properties. It fol-
lows that a mental substance cannot exist unless it has properties. Now
the properties of a mental substance must be its various thoughtsits
mental states. Thus, a mental substance cannot exist unless it has mental
states. So, if those mental states are totally causally dependent on phys-
ical states or events (e.g., on brain states or brain events), then it follows
that mental substance itself depends for its existence on the occurrence
of these physical states or events. Furthermore, since those physical
states are themselves properties that cannot exist apart from a material
substance, it also follows that mental substance depends for its existence
Descartes
on material substance. Putting this a bit more formally, we can give the
following argument. The substance theory, combined with the view that
a mental substance must have the appropriate sort properties (namely,
mental properties), implies that
() If there are no instantiated mental properties (= mental

states), then there are no mental substances.
But the causal dependence of mental states on physical states, which is

not ruled out by Descartess proof of the Real Distinction, means that
() If there are no instantiated physical properties, then there are

no instantiated mental properties.
From () and (), it follows that
() If there are no instantiated physical properties, then there are

no mental substances.
Now the substance theory, combined with the view that physical prop-
erties can belong only to the appropriate kind of substance (namely, ma-
terial substance) implies that
() If there are no material substances, then there are no instan-

tiated physical properties.
Finally, there follows from () and () that
() If there are no material substances, then there are no mental

substances.
The point can also be shown schematically, as in Figure . below.

Nothing in Descartess case for dualism rules out such dependence of
res cogitans on res extensa. At best Descartess arguments give a certain
epistemological priority to res cogitansshow that its existence can be
known before that of res extensa. But this does not mean that res cogitans
has any metaphysical prioritythat it can actually exist independently of
res extensa.
Did Descartes himself recognize the limitations of his case for du-
alism? It might seem that he did not; for in principle of the Principles
Meditation VI
Figure .
of Philosophy he writes, 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 (CSM I , SPW , AT VIII A ). At rst sight, this def-
inition seems to contradict atly our claims about the possible causal
dependence of mental substance on material substance. However, in the
next sentence Descartes says:
And there is only one substance which can be understood to

depend on no other thing whatsoever, namely God. In the case of
all other substances, we perceive that they can exist only with the
help of Gods concurrence. Hence the term substance does not
apply univocally [= with one and the same meaning] ... to God
and to other things; that is, there is no distinctly intelligible
meaning of the term which is common to God and his creatures.
<In the case of created things, some are of such a nature that
they cannot exist without other things, while some need only the
ordinary concurrence of God in order to exist. We make this dis-
tinction by calling the latter substances and the former qual-
ities or attributes of those substances.>

The last sentence is in angle brackets because Descartes added it in French to the
original Latin text of the Principles.
Descartes
Here, Descartes admits that the denition of substance he has just given
applies only to God. (Later, Spinoza was to argue that Descartes should
have stuck to this denition and so embraced Spinozas own view that
God is the only substance that exists or can be conceived to exist.) Cre-
ated substances cannot be dened as things that need nothing else in
order to exist, because they require Gods constant concurrence in order
to exist. They qualify as substances only in the weaker sense that they
need nothing else except Gods concurrence in order to exist. As Descartes
says in principle , as for corporeal substance and mind (or created
thinking substance), these can be understood to fall under this common
concept: things that need only the concurrence of God in order to exist
(CSM I , SPW , AT VIIIA ). This need for Gods concurrence
stems from Descartess doctrine, introduced in Meditation III, that cre-
ated substances, both mental and physical, need to be continuously cre-
ated by God in order to remain in existence (CSM II , SPW , AT VII
). Although we need not examine this doctrine (which we briey
noted in chapter ) for its own sake, we should take it into account in
considering whether Descartes thought that his case for dualism rules
out the causal dependence of one created substance on another. In al-
luding to the doctrine in the Principles of Philosophy, Descartes does not
explicitly dierentiate between saying that (a) created substances exist
only if God sustains them, and (b) created substances exist if only God
sustains them (in other words, if and only if God sustains them). Thus,
Principle says that other [i.e., created] substances ... can exist only
with the help of Gods concurrence, while Principle says that created
substances are things that need only the concurrence of God in order to
exist (statements (a) and (b), respectively). The former is consistent
with one created substances causally depending on another (as well as
on God), whereas the latter is not.
Now it seems that Descartes must have held (b) as well as (a). For
surely, an omnipotent God could, if he wished, sustain a substance in
existence without the help of any other thing. However, it does not follow
that Descartes has shownor that he believes he has shownthat the
mind is a substance in the sense stipulated by (b). For all that the proof
of the real distinction has shown, it might be that a human mind, like a
quality or attribute is one of those created things, [that has] ... such
a nature that they cannot exist without other things. The world might
work in such a wayGod might have designed it in such a way
that human minds, even though an omnipotent being could of course
Meditation VI
(miraculously) sustain them in existence without the help of anything

physical, naturally depend on physical things and, therefore, do not
count as substances even in the weaker sense that applies to created
things. That Descartes has not shown the contrary of this should be ob-
vious from what has already been said about his rationale for using the
concept of substance. More interestingly, there is excellent evidence that
Descartes realized he had not shown it. In the Synopsis of the Medita-
tions, he issues a warning to people [who] may perhaps expect argu-
ments for the immortality of the soul in the Meditations (CSM II , SPW
, AT VII ). He explains that since his aim in the Meditations was to
assert only what he could rigorously prove, he was obliged to follow
strictly his method of never asserting a proposition before setting out all
the premises on which it depended. He then declares that
the rst and most important prerequisite for knowledge of the

immortality of the soul is to form a concept of the soul which is
as clear as possible and is also quite distinct from every concept
of body; and that is just what has been done in this section [in
Meditation II]. (CSM II , SPW , AT VII )
Next, he outlines the further steps that will nally lead up to his asser-
tion of the Real Distinction in Meditation VI. Then he adds this crucial
comment (most of which we have already quoted in discussing other
points):
But I have not pursued this topic [i.e., immortality] further in

this book, rst because these arguments are enough to show that
the decay of the body does not imply the destruction of the mind,
and are hence enough to give mortals the hope of an after-life,
and secondly because the premises which lead to the conclusion
that the soul is immortal depend on an account of the whole of
physics. This is required for two reasons. First, we need to know
that absolutely all substances, or things which must be created
by God in order to exist, are by their nature incorruptible and
cannot ever cease to exist unless they are reduced to nothingness
by Gods denying his concurrence to them. Secondly, we need to
recognize that body, taken in the general sense, is a substance, so
that it too never perishes. But the human body, in so far as
it diers from other bodies, is simply made up of a certain
Descartes
conguration of limbs and other accidents of this sort; whereas

the human mind is not made up of any accidents in this way, but
is a pure substance. For even if all the accidents of the mind
change, so that it has dierent objects of the understanding and
dierent desires and sensations, it does not on that account
become a dierent mind; whereas the human body loses its iden-
tity merely as a result of a change in the shape of some of its
parts. And it follows from this that while the body can very easily
perish, the mind is immortal by its very nature. (CSM II , SPW
, AT VII )
There are two important points to notice about what Descartes says
here. First, he represents the thesis that the absolutely all substances,
or things which must be created by God in order to exist are naturally
incorruptible (i.e., that they need only Gods concurrence in order to
exist) as something to be established. Notice that in calling all the things
which must be created by God in order to exist substances but granting
that their incorruptibility needs to be proved, he is using the term sub-
stances purely denotatively to refer to mind and body; he is not as-
suming that these substances even need to meet the denition of
substance in his weaker sense. Thus, he is prepared to admit that for all
his metaphysical arguments in the Meditations have shown, neither the
soul nor body is incorruptible, or is a substance even in the weaker sense.
Although a careless reader might think that Descartes retracts this ad-
mission at the very end of the passage, where he seems to arm the
immortality of the soul, in fact he makes no such retraction; for part of
the this from which the souls immortality follows is the thesis, yet to
be established, that all substances . . . are by their nature incorrupt-
iblethat is, that all the things that he has been calling substances
really are substances in the weaker sense. In other words, Descartes here
arms the souls immortality only conditionally. His point is that if the
thesis that the things we call substances are incorruptible were estab-
lished, then, given that the soul is one of those things, its immortality
would be established, too. The second point is even more signicant.
Descartes says that proving the immortality of the soul depends upon
the completion of his whole physics. Now although it may be hard for us
to see how a physics could show that a purely mental thing is incorrupt-
ible, the vital point is the implication that any knowledge we could have
of the souls immortality must rest on a completed natural science, which
Meditation VI
might reveal as-yet-unknown causal connections and dependencies bet-

ween mind and body. Descartes realized that the logical possibility of the
minds existing without the body, and its consequent nonidentity with
the body (which is all he had argued for in the Meditations) is consistent
with the minds causal dependence on the body. He knewand openly
acknowledgedhow far his case for dualism was from establishing the
religious beliefs that he himself undoubtedly held and cherished.
BIBLIOGRAPHY
The secondary literature on Descartes is immense. The most complete bibliographies are:
Chappell, Vere and Doney, Willis eds. Twenty-Five Years of Descartes Scholarship, 19601984:
A Bibliography. New York: Garland, 1987.
Sebba, Gregor. Bibliographica Cartesiana: a Critical Guide to the Descartes Literature
18001960. The Hague: Martinus Nijho, 1964.
The former has an informative introduction and cross-indexes works by topic; the lat-
ter contains abstracts of many of the works. Some shorter but useful bibliographies
can be found in:
Caton, Hiram. P. The Origin of Subjectivity: An Essay on Descartes. New Haven: Yale Uni-
versity Press, 1973.
Cottingham, John, ed., The Cambridge Companion to Descartes. Cambridge: Cambridge
University Press, 1992.
Doney, Willis. Some Recent Work on Descartes: a Bibliography. In Michael Hooker,
ed., Descartes: Critical and Interpretive Essays. Baltimore: Johns Hopkins Univer-
sity Press, 1978.
Gaukroger, Stephen. The Blackwell Guide to Descartes Meditations. Malden: Blackwell,
2006.
Grene, Marjorie. Descartes. Minneapolis: University of Minnesota Press, 1985.
Naaman-Zauderer. Descartes Deontological Turn. Cambridge: Cambridge University
Press, 2010.
The list of works given below includes only the sources cited or mentioned in this book.
Among the items listed, the books by John Cottingham, E. M. Curley, and Margaret
Wilson contain useful bibliographies. So do Willis Doneys rst edited collection and
Alan Gewirths The Cartesian Circle Reconsidered.
Aquinas, Thomas. Summa Theologica. In Anton C. Pegis, ed., The Basic Writings of
Saint Thomas Aquinas. New York: Random House, 1945.
Aristotle. De Anima (On the Soul). In Richard M. McKeon, ed., The Basic Works of
Aristotle. New York: Random House, 1941.

Bibliography
Berkeley, George. A Treatise Concerning the Principles of Human Knowledge. Oxford:

Oxford University Press, 1998.
Borst, Clive V., ed., The Mind-Brain Identity Theory. New York: St. Martins Press, 1970.
Broad, Charlie D. Mind and Its Place in Nature. London: Routledge & Kegan Paul, 1962.
Burnyeat, Myles, ed., The Skeptical Tradition. Berkeley: University of California Press,
1983.
Butler R.J., ed. Cartesian Studies. Oxford: Basil Blackwell, 1972.
Cartwright, Richard. Negative Existentials. Journal of Philosophy 57 (20/21) (1960):
62939.
Chappell, Vere. Descartess Compatibilism. In John Cottingham, ed., Reason, Will and
Sensation. Oxford: Oxford University Press, 1994.
Copleston, Frederick. A History of Philosophy, Book 2, vol. 4. New York: Doubleday,
1985.
Cornman, James W., Keith Lehrer, and George Pappas. Philosophical Problems and Ar-
guments, 3d ed. Indianapolis and Cambridge: Hackett Publishing Company, 1987.
Cottingham, John. A Descartes Dictionnary. Oxford: Blackwell, 1993.
Cottingham, John. Descartes. Oxford: Basil Blackwell, 1986.
Cottingham, John. Descartes: Metaphysics and the Philosophy of Mind. In G.H.R.
Parkinson, ed., Routledge History of Philosophy, Vol, IV. London: Routledge, 1993.
Cottingham, John, ed., Reason, Will and Sensation. Oxford: Oxford University Press,
1994.
Curley, Edwin M. Descartes Against the Skeptics. Cambridge, MA: Harvard University
Press, 1978.
Curley, Edwin M. Descartes, Spinoza and the Ethics of Belief. In Eugene Freeman, ed.,
Spinoza: Essays in Interpretation. La Salle: Open Court Publishing Company, 1975.
Danto, Arthur, What Philosophy Is. New York: Harper & Row, 1968.
De George, Richard T., ed. Classical and Contemporary Metaphysics. New York: Holt,
Rinehart and Winston, 1962.
Dennett, Daniel C. Descartess Argument from Design. Journal of Philosophy 105:7
(July 2008): 33345.
Descartes, Ren, Descartes Conversation with Burman. Trans. John Cottingham.
Oxford: Oxford University Press, 1976.
Descartes, Ren. Descartes: Philosophical Writings. Trans. and ed. Elizabeth Anscombe
and Peter Geach. New York: Macmillan, 1971.
Descartes, Ren. Meditations on First Philosophy with Selections from the Objections and
Replies. Trans. John Cottingham. Cambridge: Cambridge University Press, 1986.
Descartes, Ren. Oeuvres de Descartes, 11 vols., Rev. ed. Charles Adam and Paul Tan-
nery. Paris: Librairie Philosophique J. Vrin and Le Centre National de la Recherche
Scientique, 19641976.
Descartes, Ren. The Philosophical Writings of Descartes. Trans. John Cottingham, Rob-
ert Stootho, and Dugald Murdoch, Volumes 1 and 2. Cambridge: Cambridge Uni-
versity Press, 19841985.
Descartes, Ren. The Philosophical Writings of Descartes. Trans. John Cottingham, Rob-
ert Stootho, Dugald Murdoch, and Anthony Kenny. Volume 3: The Correspon-
dence. Cambridge: Cambridge University Press, 1991.
Descartes, Ren. Selected Philosophical Writings. Trans. John Cottingham, Robert
Stootho, and Dugald Murdoch. Cambridge: Cambridge University Press, 1988.
Dicker, Georges. Berkeleys Idealism: A Critical Examination. New York: Oxford Univer-
sity Press, 2011.
Dicker, Georges. Humes Epistemology and Metaphysics: An Introduction. London: Rout-
ledge, 1998.
Bibliography
Dicker, Georges. Is There A Problem About Perception and Knowledge? American Phil-
osophical Quarterly 15:3 (July 1978): 16576.
Dicker, Georges. Kants Refutation of Idealism. Nos 47:1 (March 2008): 80108.
Dicker, Georges. Kants Refutation of Idealism: a Reply to Chignell. Philosophical
Quarterly 61:242 (January 2011): 17583.
Dicker, Georges. Kants Refutation of Idealism: Once More unto the Breach. Kantian
Review 17 (2) (July 2012): 19195.
Dicker, Georges. Kants Theory of Knowledge: An Analytical Introduction. New York:
Dicker, Georges. Perceptual Knowledge: An Analytical and Historical Study. Boston:
Reidel, 1980.
Dicker, Georges. Primary and Secondary Qualities: A Proposed Modication of the
Lockean Account. The Southern Journal of Philosophy, 15 (Winter 1977): 45771.
Reprinted in Walter E. Creery, ed., George Berkeley: Critical Assesments, Volume II,
3 vols. London: Routledge, 1991.
Dicker, Georges. Three Questions about Treatise 1.4.2. Hume Studies 33: 1 (April
2007): 11553.
Doney, Willis, ed. Descartes: A Collection of Critical Essays. Garden City, NY: Doubleday/
Anchor Books, 1967.
Doney, Willis. Descartess Conception of Perfect Knowledge. Journal of the History
of Philosophy 8 (4) (1970): 387403. Reprinted in his (ed.) Eternal Truths and the
Cartesian Circle: A Collection of Studies. New York: Garland, 1987.
Doney, Willis. The Cartesian Circle. Journal of the History of Ideas, 16:3 (June 1955):
32438. Reprinted in his Eternal Truths and the Cartesian Circle: A Collection of
Studies. New York: Garland, 1987.
Doney, Willis, ed. Eternal Truths and the Cartesian Circle: A Collection of Studies. New
York: Garland, 1987.
Ducasse, C.J. In Defense of Dualism. In Sydney Hook, ed., Dimensions of Mind. New
York: Macmillan, 1961.
Feldman, Fred. A Cartesian Introduction to Philosophy. New York: McGraw-Hill, 1986.
Flew, Antony, An Introduction to Western Philosophy: Ideas and Arguments from Plato to
Sartre. Rev. ed. London: Thames and Hudson, 1989.
Frankfurt, Harry G. Demons, Dreamers, and Madmen: The Defense of Reason in Des-
cartess Meditations. New York: Bobbs-Merrill, 1970.
Frankfurt, Harry G. Descartes Validation of Reason. American Philosophical Quarterly
2:2 (April 1965): 14956. Reprinted in Willis Doney, ed., Descartes: A Collection of
Critical Essays. Garden City, NY: Doubleday/Anchor Books, 1967.
Frankfurt, Harry G. Memory and the Cartesian Circle. Philosophical Review 71 (4)
(1962): 50411. Reprinted in Willis Doney, ed., Eternal Truths and the Cartesian
Circle: A Collection of Studies. New York: Garland, 1987.
Gewirth, Alan. The Cartesian Circle. Philosophical Review 50 (4) (1941): 36895. Re-
printed in Willis Doney, ed., Eternal Truths and the Cartesian Circle: A Collection of
Studies. New York: Garland, 1987.
Gewirth, Alan. The Cartesian Circle Reconsidered. Journal of Philosophy 67:19 (1970):
66885. (1987). Reprinted in Willis Doney, ed., Eternal Truths and the Cartesian
Circle: A Collection of Studies. New York: Garland, 1987.
Gewirth, Alan. Descartes: Two Disputed Questions. Journal of Philosophy 68:9 (1971),
28896. Reprinted in Willis Doney, ed., Eternal Truths and the Cartesian Circle: A
Collection of Studies. New York: Garland, 1987.
Freeman, Eugene, ed. Spinoza: Essays in Interpretation. La Salle: Open Court Publishing
Company, 1975.
Bibliography
Grice, Paul H. The Causal Theory of Perception. Reprinted in Robert J. Swartz, ed., Perceiv-
ing, Sensing, and Knowing. Berkeley: University of California Press, 1976, pp. 43872.
Hateld, Gary. Descartes and the Meditations. London: Routledge, 2003.
Hick, John. Evil and the God of Love. London: Fontana, 1968.
Hick, John. Philosophy of Religion, Second Edition. Englewood Clis. NJ: Prentice-Hall,
1973.
Hintikka, Jaako. Cogito, Ergo Sum: Inference or Performance. In Willis Doney,
ed., Descartes: A Collection of Critical Essays. Garden City, NY: Doubleday/An-
chor Books, 1967. Also in Alexander Sesonske and Noel Fleming, eds., Meta-
Meditations: Studies in Descartes. Belmont, CA: Wadsworth Publishing Company,
1965. Originally published in Philosophical Review 71 (1) (1962): 332.
Hook, Sydney, ed. Dimensions of Mind. New York: Macmillan, 1961.
Hume, David. A Treatise of Human Nature. 2d. ed. Ed. Lewis A. Selby-Bigge and P. H.
Nidditch. Oxford: Oxford University Press, 1975.
Hume, David. An Enquiry Concerning Human Understanding. 3d ed. Ed. L.A. Selby- Bigge
and P.H. Nidditch. Oxford: Oxford University Press, 1985.
Hume, David. Dialogues Concerning Natural Religion. Indianapolis: Hackett Publishing
Company, 1980.
Kant, Immanuel. Critique of Pure Reason. Trans. Norman Kemp Smith. London: Mac-
millan, 1963.
Kenny, Anthony. Descartes: A Study of his Philosophy. New York: Random House, 1968;
reprint ed. New York: Garland, 1987.
Kenny, Anthony. Descartes on the Will. In R. J. Butler, ed., Cartesian Studies. Oxford:
Basil Blackwell, 1972.
Kenny, Anthony. The Cartesian Circle and the Eternal Truths. Journal of Philosophy,
67 (19) (1970): 685700.
Leibniz, Gottfried W. Critical Remarks Concerning the General Part of Descartes
Principles. In his Monadology and Other Philosophical Essays, trans. and ed. Paul
Schrecker and Anne Martin Schrecker. New York: Macmillan, 1965.
Locke, John. An Essay Concerning Human Understanding. Abridged and ed. Andrew S.
Pringle-Pattison. Oxford: Oxford University Press, 1924.
Locke, John. An Essay Concerning Human Understanding. Ed. Peter H. Nidditch. Oxford:
Lovejoy, Arthur O. The Meanings of Emergence and Its Modes. In Richard T. De
George, ed., Classical and Contemporary Metaphysics. New York: Holt, Rinehart and
Winston, 1962. Originally published in Proceedings of the Sixth International Con-
gress of Philosophy (1926), 2033.
Mackie, John. The Miracle of Theism. Oxford: Oxford University Press, 1982.
Madden, Edward H and Hare, Peter H. Evil and the Concept of God. Springeld, IL:
Charles C. Thomas, 1968.
Mandelbaum, Maurice. Philosophy, Science, and Sense Perception: Historical and Critical
Studies. Baltimore: Johns Hopkins University Press, 1964.
Matson, Wallace I. The Existence of God. Ithaca: Cornell University Press, 1965.
Montaigne, Michel Eyquem de. Montaignes Essays and Selected Writings. Trans. and ed.
Donald M. Frame. New York: St Martins Press, 1963.
Naaman-Zauderer, Noa. Descartes Deontological Turn. Cambridge: Cambridge Univer-
sity Press, 2010.
Nelson, Alan. Descartes Ontology of Thought. Topoi 16 (2) (1997) 16378.
Parkinson, G.H.R., ed. Routledge History of Philosophy, Vol, IV. London: Routledge,
1993.
Bibliography
Price, Henry H. Thinking and Experience. Cambridge, MA: Harvard University Press,
1962.
Prichard, H. A. Descartess Meditations. In Willis Doney. ed., Descartes: A Collection of
Critical Essays. Garden City, NY: Doubleday/Anchor Books, 1967.
Ragland C. P. Alternative Possibilities in Descartess Fourth Meditation. British Jour-
nal for the History of Philosophy 14:3 (August 2006): 379400.
Ragland C. P. Descartes on the Principle of Alternative Possibilities. Journal of the
History of Philosophy 44:3 (July 2006): 37794.
Rorty, Amlie O., ed. Essays on Descartes Meditations. Berkeley: University of California
Press, 1986.
Rosenthal, David M. Will and the Theory of Judgment. In Amlie O. Rorty, ed., Essays
on Descartes Meditations. Berkeley: University of California Press, 1986.
Russell, Bertrand. A History of Western Philosophy. New York: Simon and Schuster,
1945.
Russell, Bertrand. Logic and Knowledge: Essays 19011950. Ed. Robert C. Marsh. New
York: Capricorn Books, 1971.
Schacht, Richard, Classical Modern Philosophers: Descartes to Kant. Boston: Routledge &
Kegan Paul, 1984.
Sesonske, Alexander and Noel Fleming, eds., Meta-Meditations: Studies in Descartes.
Belmont, CA: Wadsworth Publishing Company, 1965.
Simmons, Alison. Re-Humanizing Descartes. Philosophic Exchange 41 (20102011):
5271.
Smart, J.J.C. Sensations and Brain Processes. Reprinted in Clive V. Borst, ed., The
Mind-Brain Identity Theory. New York: St. Martins Press, 1970. Originally pub-
lished in Philosophical Review 68 (2) (1959): 14156.
Spinoza, Benedictus. Ethics Demonstrated in Geometrical Order. In Edwin. M. Curley,
ed. and trans., The Collected Works of Spinoza, Volume I. Princeton, NJ: Princeton
University Press, 1985.
Swartz, Robert J., ed. Perceiving, Sensing, and Knowing. Berkeley: University of Califor-
nia Press, 1965.
Thomas, Stephen. Practical Reasoning in Natural Language. 3d. ed. Englewood Clis, NJ:
Prentice-Hall, 1986.
Van Cleve, James. Foundationalism, Epistemic Principles, and the Cartesian Circle.
Philosophical Review, 88:1 (January 1979): 5591. Reprinted in Willis Doney, ed.,
Eternal Truths and the Cartesian Circle: A Collection of Studies. New York: Garland,
1987.
Van Cleve, James. On a Little-Noticed Fallacy in Descartes. Manuscript.
Williams, Bernard. Descartes: The Project of Pure Enquiry. Harmondsworth: Penguin
Books, 1978.
Williams, Bernard. Descartess Use of Skepticism. In Myles Burnyeat, ed., The Skepti-
cal Tradition. Berkeley: University of California Press, 1983.
Wilson, Margaret. Descartes. Boston: Routledge & Kegan Paul, 1978.
Wol, Robert Paul. Kants Theory of Mental Activity. Cambridge, MA: Harvard Univer-
sity Press, 1963.
Wittgenstein, Ludwig. Philosophical Investigations, 3d. ed. Trans. Elizabeth Anscombe.
Oxford: Blackwell, 1958.
INDEX
a posteriori knowledge, 2324 as distinct from the mind, 86, 90, 258,
a priori knowledge, 2324, 3334, 3738, 260, 263269, 271272, 278,
134 280, 298, 301, 303, 310, 313,
accidents. See properties 316, 320, 325326
Alanen, Lilli, 200202 doubt regarding, 26, 40, 44, 75, 82, 88
analytic statements, 3335, 3738 existence of, 258, 260, 264266,
animal faith, 279 271272, 280, 302
animal spirits, 306307 idea of, 258
Anscombe, Elizabeth, 109 interaction with mind and, 274,
Anselm, St., 206, 221, 228229, 231 301311, 315, 327
Aquinas, Thomas, 83, 101, 168169, self and, 80, 8889
208, 243 soul and, 304, 306307, 310, 321, 325
Archimedes, 39 Boyle, Sir Robert, 285
argument from change, 5054, 5864, Brahe, Tycho, 6
298, 321 brain
argument from doubt, 8890, 260 brain events and, 307308, 311313,
argument from ignorance, 260 317318, 321
Aristotle, 56, 50, 65, 67, 8183, 87, mind and, 304307, 310, 313,
137, 192, 252, 301 317318
Arnauld, Antoine, 144145, 150, 168, neuroscience and, 38, 307, 311312,
268274 317319
astronomy, 56 Broad, C.D., 239, 304, 310311, 317
atomism, 283, 285 bundle theory, 5051
Augustine, St., 210211, 215, 217 Burman, Frans, 24, 310
Ayer, Alfred Jules, 281
Carnap, Rudolph, 244
basic principles, 1415 carnivorous cows example, 237240, 242
Berkeley, Bishop George, 50, 66, 70, 139, Cartesian circle
281, 289291 Arnauld on, 144145, 150, 268
Blum, Roland P., 172n49 clear and distinct perceptions and, viii,
body, the. See also matter; physical world 144145
clear and distinct perceptions and, 86, creation of eternal truths and,
260 166170, 173
Cogito and, 75 general rule defense and, 153164

Index
Cartesian circle (continued) certainty. See also clear and distinct

memory defense and, viii perceptions
radical doubt of reason and, 164166, basic principles and, 14
169, 171173 conditions of observation and, 17
restriction defense and, 147154 Descartess search for, 4, 811, 39
validation of reason and, 170176 knowledge and, 4
Cartesian dualism. See also real logic and, 2223
distinction between mind and mathematics and, 9, 2123
body negation and, 1213
arguments for, 254 of ones own existence, 40, 259
assessment of, 313327 of ones own thoughts, 4344
criticisms of, 308309 senses and, 1517, 2324, 31
dened, 86 Chappell, Vere, 189
immortality and, 87, 267 circle and rhombus example, 227228
implications of, 68 clarity-and-distinctness criterion of
mind and body in, 86, 90, 258, 260, truth, 114, 126
263269, 271272, 278, 280, clear and distinct perceptions
298, 301303, 310311, 313, assent-compelling nature of, 128,
316, 320, 325326 143144, 152156, 158,
physics and, 87 160163, 169170, 186187,
Platonic thought and, 83 190, 192196, 198204
problem of interaction and, xi, 309313 beliefs from, 159160
Cartesian Trialism, 302 the body and, 86, 260
Cartwright, Richard L., 237n5 Cartesian circle and, viii, 144145
Caterus, xiii, 243, 248251, 253 Cogito and, 9394, 9697, 263
causal adequacy principle, 117118, 133, compared to moral certainty, 200
135, 143, 146 as criterion for truth and, 97100,
causal conception of perception (CCP), 127129, 143151, 153158,
3437 161164, 166167, 172,
causal principle, 118119, 122, 139 183184, 186, 222224,
cause 251253, 255, 261263, 269, 275
eect and, 115119, 121122, 124, denition of, 9495, 97
128136, 140143, 310 degrees of reality and, 139
ecient and total, 115, 117, 167 doubt and, 99100, 152153,
events and, 131132, 189, 310 155156, 158165, 173, 176
nal, 87 drug analogy and, 162163
formal and eminent containment and, rst item of knowledge and, 9293
119122, 278 God as guarantor of, 128129, 144,
of ideas, 104, 110114, 117120, 147148, 150154, 160161,
122125, 129, 142144 163, 166, 255, 281
intermediate, 311 Gods existence and, 97100, 127,
performationist assumption and, 129, 150151, 153, 165, 171,
130131, 135 175, 222224, 252253
proximate, 311312 intellect and, 129, 155, 163, 176, 182,
reality and, 115126, 128136, 184, 188, 202, 216
140143 mathematics and, 9596, 99, 155,
remote, 311 157, 159, 222, 269
of sensory perceptions, 3436, 38, principle of alternative possibilities
275282, 286 and, 195197, 200
centaur example, 230 will and, 186187, 190, 192204, 219
Index
cogitationes (thoughts), 44, 54, 86, 92 degrees of reality and, 125126

Cogito sequel to, 126129
additional premise question, 7579 subargument for rst premise of core
the body and, 75 of, 114124, 128129
clear and distinct perceptions and, Cottingham, John
9394, 9697, 263 scholarship of, 41, 111112, 117, 120,
components of, 93, 9596 132133, 135, 143144,
criticisms of, 6971, 7476 160161, 186, 302303, 307308
Descartess belief in the indubitability translations of Descartes by, viii, 41,
of, 157159, 163, 190, 204, 254, 99n3, 120, 171, 191n10, 249,
258, 281 259260
I premise within, 4748, 5758, criterion-not-needed strategy, 147,
6974, 77 153164, 175176. See also
Descartess various formulations of, general rule defense
41 Curley, Edwin M., 99, 108n9, 114n12,
dualism and, 75 142, 184, 259
extraction argument and, 94
performative interpretation of, 4849 Darwin, Charles, 6
question-begging and, vii, 4749, 57, deceiver argument. See also God,
7172, 7475, 77 deception and
Russells treatment of, 47 a priori premises of, 34, 3738
six-step reconstruction of, 5458 boundary-setting aspect of, 83
substance theory and, vii, xi, xiii, 49, causal conception of perception (CCP)
5355, 5758, 6870, 79 and, 3437
unreconstructed version of, 7179, Descartess response to, 2425, 4041,
92, 95 44, 126127
thinking as proof of existence and, evil demon and, 2627, 3638, 40,
4143, 71, 7375, 7779, 9197, 4445, 8081
99, 113 thinking and, 8283
conceivability, logical possibility and, deception
264266, 268, 314, 317 God and, 2425, 99100, 126129,
conditions of observation, 17, 19, 2931, 143147, 149152, 156,
284285, 287290 161163, 165, 170176,
consciousness, 133, 213214, 290, 181182, 279281
317319 senses and, 78, 1516, 1920, 22,
continuous creation, 138, 169, 324 2732, 38, 46, 103, 128, 163
contradictories, 167, 207, 240241, 262 deationism, 238240, 242
contraposition, 13 degrees of reality
Copernicus, Nicholas, 6 existence of God and, 108, 125126
Copleston, Frederick, 95, 222223 nite substances and, 107108,
core argument for Gods existence. See 137138, 177
under cosmological argument for ideas and, 107114, 129, 142143,
Gods existence (Descartes) 176179
Cornman, James, 317318 innite substances and, 107, 137138,
cosmological argument for Gods 177178
existence (Aquinas), 101 modes and, 137138
cosmological argument for Gods substance and, 138139
existence (Descartes) Democritus, 283
core of, 110114, 122, 124, 128129 Dennett, Daniel, 180
criticisms of, 129144 descriptive predicates, 232234, 240, 242
Index
determinable properties, 5152, 6264, denition and, 252253

299300 extension as, 299301
determinate properties, 5154, 59, mathematics and, 229
6164, 299300 ontological argument for Gods
Dewey, John, 3, 282 existence and, 229230
dispositions, 284291, 293 thinking as, 258261, 264, 269,
double negation, 1314, 37 299300, 320
doubt eternal truths, creation of, 166170,
grounds or reasons for, 3132, 173, 207208
171175 events
of logic, 2223 brain, 307308, 311313, 317318,
of mathematics, 2224 321
method of, 3, 26, 265266 cause-and-eect relationships and,
of ones own bodys existence, 26, 40, 131132, 189, 310
44, 75, 82, 88 mental, 307308, 311313, 317319,
of ones own existence, 40, 75, 8081, 88 321
of the physical world, 22, 2526, physical, 309, 311
3940, 44, 53, 75, 82, 8788, 98, evil
101, 134, 169, 274 aesthetic conception of, 215219
of the reliability of human reason, the afterlife and, 212213
164166, 169, 171173 fact/value distinction and, 217
of the senses, 3133, 98, 152, 166, 257 fullness of creation and, 217218
doxastic attitudes, 1112 God and, 181, 205213
doxastic voluntarism, 185 moral evil, 209210, 214
dragon example, 242 nonmoral evil, 209214
dream argument, 17, 1922, 24, 27, private view of, 214215, 217219
3032, 257 problem of, 181, 205212, 218
dreams, 1922, 3032, 42, 45, 104 theodicy and, 208217
dualism. See Cartesian dualism will and, 209210
dualistic interactionism, 305, 310, evil demon. See under deceiver argument
313314, 318 existence
Ducasse, Curt J., 304, 310311, 313 the body and, 258, 260, 264266,
271272, 280, 302
Elizabeth, Princess of Bohemia, 309 compared to subsistence, 238239,
eminent reality, 120. See also formal and 241
eminent containment doubting ones own, 40, 75, 8081, 88
emotions, 190, 192, 308 of God, 25, 118, 126128, 134,
epiphenomenalism, 313314, 317 143147, 149153, 161163,
epistemology, 24, 38, 79, 264, 266, 322 166167, 173176, 179180,
error 221, 225228, 231, 233236,
deciding to believe and, 184185 242244, 246255, 267268,
evil and, 205 280
fullness of creation and, 215216 Kants view of, 231236, 240
intellect and, 182, 219 necessary existence, 248250,
natural impulses and, 104 252253
will and, 25, 104, 129, 181185, 199, of the physical world, xi, 22, 2526,
205, 214217, 219 3940, 44, 53, 75, 82, 8788,
essence 9798, 101, 134, 169, 254255,
clear and distinct perceptions of, 257, 274, 277282, 287
223224, 253254 possible existence, 248
Index
as a property, 231237, 242 Descartess denition of, 221, 228

real predicate question and, 231233 eternal truths and, 166169
thinking as proof of, 4143, 4749, evil and, 181, 205213
71, 7375, 7779, 8283, 88, existence of, 25, 118, 126128, 134,
9197, 99, 106, 113, 258259, 143147, 149153, 161163,
263264 166167, 173176, 179180,
existential import, 246247 221, 225228, 231, 233236,
extension 242244, 246255, 267268, 280
characteristics of, 297 goodness of, 2425, 175, 179180,
dreams and, 20 205214, 216, 218, 231, 248, 272
as quality of physical existence, 40, as guarantor of clear and distinct
52, 6364, 85, 102n6, 121, perceptions, 128129, 144,
263, 270271, 273, 275276, 147148, 150154, 160161,
282284, 286, 292301, 303 163, 166, 255, 281
wax example and, 8586 idea of, 101, 105108, 110, 113114,
extraction argument, 94 116, 124125, 129, 143,
176177, 180, 221, 226, 248, 251
Feldman, Fred, 45, 308309 innite nature of, 118, 125, 177179,
nal causes, 87 249
Flew, Antony, 243245, 252253 Judeo-Christianity and, 205206,
formal and eminent containment, 111, 217218, 221, 228
117, 119121, 278 omnipotence of, 2122, 25, 99100,
formal mode of speech, 122, 243248, 125, 144, 152, 154, 156, 165,
251253 168170, 172175, 179180,
formal reality, 109114, 117126, 129, 205209, 212214, 216, 221,
143, 176177, 179180, 277278 231, 233, 247, 249, 262, 272,
Frankfurt, Harry G., xiii, 4n1, 24, 324325
147n33, 163164, 170172, 187 omniscience of, 205, 214, 231, 247
free will. See will perfection of, 98, 113114, 116,
Frege, Gottlob, 239 126127, 129, 143, 174175,
180, 211, 214, 221, 224228,
Galilei, Galileo, 67, 87, 285, 291 231, 246253, 257, 279281
Gassendi, Pierre, 42, 73, 116, 177180, gold statute example, 6062
198, 216, 309 Goldbachs conjecture, 316317
Gaunilo, 229230 Grice, Paul, 3536
Geach, Peter, 109
general conservation principles, 133135 hallucination, 22, 2526, 38, 45, 279
general rule defense, 153164 Hateld, Gary, 102n6, 147148
geometry. See mathematics; triangle heirloom view of causation, 132, 135
examples Hick, John, 209, 211213
Geulincx, Arnold, 309 hierarchy
Gewirth, Alan, xiii, 170172 dependence and, 138139
God. See also core argument for Gods Descartess three levels of, 109110,
existence (Descartes); ontological 137139
argument for Gods existence medieval philosophy and, 56
Anselms denition of, 206, 221, 228 Hintikka, Jaakko, 48, 76n34, 265
deception and, 2425, 99100, Hobbes, Thomas, 57, 66, 137
126129, 143147, 149152, Hume, David, 50, 66, 70, 101, 118119,
156, 161163, 165, 170176, 131132, 140, 142, 164165,
181182, 279281 206, 281282, 313
Index
Hyperaspistes, 7071, 116 innite, human conception of the, 177179

hypothetical syllogism, 13 inationism, 238241
insanity, viii, 1718, 4647
I intellect
Descartess use of in the Meditations, clear and distinct perceptions and,
14 129, 155, 163, 176, 182, 184,
special semantics of, 7172 188, 202, 216
reference to the mind and, 320 error and, 182, 219
I think, therefore I am. See Cogito will and, 183, 186, 190191, 193, 202,
ideas 214, 219
causes of, 104, 110114, 117120, Irenaeus, 210213, 217
122125, 129, 142144 Iseminger, Gary, vii, 72n12
degrees of reality and, 107114, 129,
142143, 176179 Kant, Immanuel, 23, 101, 134, 203n23,
formal reality and, 110114, 122124, 231237, 239242, 248, 282
129, 176177, 179180, 277278 Kenny, Anthony, xii, 5758, 7071, 137,
of God, 101, 105108, 110, 113114, 153154, 156158, 161, 170n46,
116, 124125, 129, 143, 176177, 187, 197n15, 270
180, 221, 226, 248, 251 Kepler, Johannes, 6, 87
judgments and, 102103 knowledge. See also clear and distinct
material falsity and, 102n6 perceptions
material objects and, 104106, certainty and, 4
112113, 120, 123, 125 Descartess notions of, 8
modes of thought and, 105106, natural light and, 73, 104, 115, 127,
112113, 125 138, 146, 182, 214, 249250
natural light and, 104 a posteriori, 2324
nature of, 101103, 105 a priori, 2324, 3334, 3738, 134
objective reality and, 108114, 118,
122124, 129, 176177, 180, Lehrer, Keith, 11
277278 Leibniz, Gottfried von, 50, 168, 280281,
perfection and, 107 309
pictures and, 101, 111 Locke, John, 6162, 101, 281, 283288,
representational content of, 123 290291, 293295
self and, 124125 Lovejoy, Arthur O., 130, 135
truth/falsity and, 102
types of, 103 Mack, Chelsie, 139n25
identity through time, 6163, 300301 Mackie, John L., 133135
imagination Malebranche, Nicolas, 309
physical world and, 8485, 257 Mandelbaum, Maurice, 28
self and, 91 manifest aspect. See under secondary
senses and, 8485 qualities
immortality, 87, 267, 297298, 319, 321, masked man fallacy, 89
325326 material mode of speech, 243248,
indierence, Descartess conception of, 251252
197198 material world. See physical world
indubitability materialism, 313314, 316, 318319
of certain propositions, 910, 45, 57, mathematics
80, 93, 142, 162163, 170 certainty and, 9, 2123
compared to truth, 163164 clear and distinct perceptions and,
innite regress, 17, 66, 122 9596, 99, 155, 157, 159, 222, 269
Index
Descartess knowledge of, 8, 10 knowledge and, 73, 104, 115, 127,

doubt of, 2224 138, 146, 164, 182, 214, 249250
eternal truths and, 166168 reason and, 58
necessary truths and, 96, 99 negative existentials, 237240, 242
physics and, 21 Nelson, Alan, 200203
a priori knowledge and, 2324 Neo-Platonists, 107
matter. See also physical world nerves, 286, 306308
atomic theory of, 286 Newton, Isaac, 67
characteristics of, 86 nominalism, 6668
compared to space, 283, 292295 noncontradiction, principle of, 99, 156,
mind as distinct from, 53, 68, 8788, 158159, 163, 204
9798, 254255, 268, 321 nothingness, 5556, 116117, 119,
solidity and, 294 135136, 139141
medieval philosophy, 56, 50, 5556, 69,
116117, 130. See also Anselm St.; objective reality. See under ideas
Aquinas,Thomas; Augustine, St. occasionalism, 309310
Meinong, Alexius, 238 Ockhams Razor, 318
memory, 29, 147, 149 omnipotence
memory defense, viii, 147 God and, 2122, 25, 99100, 125, 144,
mental events, 307308, 311313, 152, 154, 156, 165, 168170,
317319, 321 172175, 179180, 205209,
Mesland, Denis, 197199, 202, 204 212214, 216, 221, 231, 233,
Mill, John Stuart, 281 247, 249, 262, 272, 324325
mind logical possibility and, 168169,
brain and, 304307, 310, 313, 207208, 262
317318 ontological argument for Gods existence
as distinct from matter, 53, 68, 8788, Cateruss objection to, 248251, 253
9798, 254255, 268, 321 criticisms of, 228254
as distinct from the body, 86, 90, 258, Descartess version of, viii, xii, 25, 101,
260, 263269, 271272, 278, 221228, 254255
280, 298, 301, 303, 310, 313, formal mode of speech and, 122,
316, 320, 325326 243248, 251253
interaction with body and, 274, Gaunilos objections to, 229230
301311, 315, 327 Kants objection to, viii, xiii, 231237,
mind-body problem, 274, 309. See also 240242
problem of interaction material mode of speech and, 243248,
Moderate Realism, 65, 6768 251252
Modus Ponens, 23, 29, 37, 74, 76, 236 necessary existence and, 248251
Modus Tollens, 23, 37, 231232, 236 real predicate question, 231233
Montaigne, Michel, 78, 22 ontology, 50, 6668
Moores Paradox, 157
mountain and valley example, 225226 perception. See clear and distinct
perceptions; senses
natural light perfect island example, 229230
clear and distinct perceptions and, 146 philosophy of mind, 314
contrasted with nature, 104, 279280 physical world
doctrine of irresistibility regarding, Descartess view of the nature of the,
186 282301
existence and, 42 dierent conceptions of substance
ideas and, 104 and, 295301
Index
physical world (continued) essential, 54

dispositions and, 284291, 293 nothingness and, 5556
doubting the existence of, 22, 2526, problems of universals and, 6568
3940, 44, 53, 75, 82, 8788, 98, spatial location and, 66
101, 134, 169, 274 thoughts as, 56, 64, 6869, 79
God and, 277279 Ptolemy, 5
imagination and, 8485, 257 Pyrrho, 7, 22
matter and, 277279
one-substance view of, 64, 297299, qualities. See primary qualities;
301 properties; secondary qualities
primary qualities and, 283285, question-begging arguments, 4749, 57,
287289 7172, 7477, 92
proving the existence of, xi, 9798, Quine, W.V., 33n11
254255, 257, 274, 277282,
287, 315 Ragland, C. P., 187, 201
secondary qualities and, 102n6, real distinction between mind and body
283291, 293 Arnaulds exchange with Descartes on,
senses and, 276283, 286287 xii, 268274
solidity and, 283285, 292295, 301 Descartess proof of, 257258,
physics 260269, 271274, 277, 281,
Aristotle and, 56, 87 314316, 319322, 325
immortality and, 326 real predicates, 231233
mathematics and, 21 reason. See also clear and distinct
Newton and, 67 perception; natural light
teleology and, 67, 87 doubt of, 164166, 169, 171173
pineal gland, 306308 validation of, 170176
Plato, 6567, 83, 107, 116, 192, 211, 238 representational theory of perception,
Platonic Realism, 6568 103104
Plotinus, 211 representative reality. See objective
possible worlds, 96, 168, 194, 249, 252 reality
precontainment principle, 115116, res cogitans, 8687, 303, 322
130136, 140142 res extensa, 8687, 297298, 303, 322
preestablished harmony theory, 309 restriction defense, 147154
Prichard, H. A., 165 resurrection, 267
primary qualities, 280, 283285, Russell, Bertrand, 4748, 50, 57, 67,
287289 238239, 281
principle of alternative possibilities, 187,
195197, 200 Santayana, George, 279
Pringle-Pattison, Andrew S., 105106 Schacht, Richard, 6970
private language, 38n14 science, the rise of, 57, 10, 83, 87
problem of interaction, 309313 scientic realists, 60
properties secondary qualities, 102n6, 280,
accidental, 54, 106, 297298, 300, 309 283291, 293
attributes and, 55 dispositional aspect of, 290291
common, 6566 manifest aspect of, 290291
compared to substance, 68 self, the
determinable, 5152, 6264, 299300 body and, 80, 8889
determinate, 5154, 59, 6164, doubting of the, 8081
299300 ideas and, 124125
dispositional, 284285, 288 imagination and, 91
Index
senses and, 91 extension and, 6364

soul and, 8081, 83, 88 God and, 323324, 326
thinking and, 80, 86, 8893, 96, 106, mental versus physical substances and,
121 321324
self-refuting argument, xii, 2733, objections to Descartess version of,
3738 300, 321
senses substance-property principle, 55,
a posteriori knowledge and, 23 138139
causal elements of, 3436, 38, supremely perfect being. See God
275282, 286 syllogisms, 13, 74, 7778
certainty and, 1517, 2324, 31 synthetic statements, 33
deception and, 78, 1516, 1920, 22,
2732, 38, 46, 103, 128, 163 Taj Mahal example, 234235
doubt of, 3133, 98, 152, 166, 257 tautologies, 252253
dreams and, 1921, 3031 teleology, 57, 87
imagination and, 8485 thalers example, 232, 235237
physical world and, 276283, 286287 theodicy. See also evil
skeptics and, 78 denition of, 208209, 208213
thinking and, 276277 Augustine and, 210211, 215, 217
simplicity, principle of theoretical, Descartes and, 214217
318319 Irenaeus and, 210213, 217
skepticism thinking
arguments for, 78, 10, 22, 29 Descartess conception of, 84
Descartess response to, 8, 1516 indubitable quality of, 4445, 47,
solidity, 283285, 292295, 301 5758, 80, 93
solipsism, 124125 as proof of existence, 4143, 4749,
Sosa, Ernest, 230n2 71, 7375, 7779, 8283, 88,
soul 9197, 99, 106, 113, 258259,
Aristotles notion of, 8183 263264
body and, 304, 306307, 310, 321, the self and, 80, 86, 8893, 96, 106, 121
325 triangle examples
Cartesian dualism and, 87 Arnaulds use of, 268270
development of, 212, 217 Descartess use of, 148149, 155, 178,
existence of, 266267, 271n5 186, 223, 225, 228230, 254,
nite nature of, 179 269272
immortality and, 297298, 325326 Pythagorean property and, 270274
Judeo-Christian notion of, 83
the self and, 8081, 83, 88 unicorn example, 240242, 246
world-soul notion and, 7071 universals, the problem of, 6568
space, 292295
spatio-temporal continuity, 6164, 301 Van Cleve, James, xii, 100, 136, 145n31,
Spinoza, Benedictus de, 50, 301, 324 146n32, 153154, 156, 170n45,
subsistence. See under existence 174n45
substance theory vindication-not-needed-strategy,
argument from change and, 5053, 147153, 159, 175176
59, 321
corollary of, 6468 wax
criticisms of, 5861, 300 as a model of the entire physical world,
denitions of, 4950, 323324 63, 8486, 263, 282, 299
degrees of reality and, 138139 substance theory and, 5153, 59
Index
will morally speaking versus absolutely

clear and distinct perceptions and, speaking, 197198, 200203
186187, 190, 192204, principle of alternative possibilities
219 and, 187
degrees of freedom and, 189190, pursuing the good and, 191193,
197198 198199
Descartes conception of, 187190 Williams, Bernard, xii, 73n13, 79n16,
error and, 25, 104, 129, 181185, 199, 112n11, 118, 137, 141, 153154,
205, 214217, 219 159161, 179180, 193194
evil and, 209210 Wilson, Margaret, 170n46, 192, 266
hard determinism and, 189 winged horse example, 226227
indierence and, 197198 withholding belief, 1012, 14, 26, 39
intellect and, 183, 186, 190191, 193, Wittgenstein, Ludwig , 38n14, 151, 282
202, 214, 219
internal compulsion and, 188189 Zaaman-Zauderer, Noa, 188, 199203
judgment and, 193 Zoroastrianism, 206

Georges Dicker-Descartes - An Analytical and Historical Introduction-Oxford University Press (2013)

Uploaded by

Copyright:

Available Formats

Georges Dicker-Descartes - An Analytical and Historical Introduction-Oxford University Press (2013)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Georges Dicker-Descartes - An Analytical and Historical Introduction-Oxford University Press (2013)

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

Who is the author of the book?

Who is the author of the book?

Descartes

This page intentionally left blank

Oxford New York

Oxford is a registered trade mark of Oxford University Press

Georges Dicker 2013

You must not circulate this work in any other form

Library of Congress Cataloging-in-Publication Data

Filcheva, Richard Glauser, Gary Iseminger, Marie Jayasekera, and Peter

Descartess Meditations speaks to the philosophical novice as well as the

Locke). I have tried to do these things in a language and style accessible

All references to the works of Descartes are given in parentheses within

Ren Descartes, Meditations on First Philosophy with Selections

volumes, edited by C. Adam and P. Tannery (Revised edition, Paris:

Preface to the Second Edition vii

1. Meditation I and the Method of Doubt 3

2. Meditation II: The Cogito and the Self 39

3. Meditation III: The Criterion of Truth and the Existence of God 91

4. Meditation IV: Error, Freedom, and Evil 181

5. Meditation V: The Ontological Argument for the Existence

6 : Meditation VI: Dualism and the Material World 257

contemporary commentator on Descartes points out, it is proper for the

purpose. According to the medieval conception, everything in the uni-

We no longer know what things are in truth; for nothing comes to

shall be t to judge these dierences?... . If he is old, he cannot

Those long chains composed of very simple and easy reasonings,

The purpose of the doubt that Descartes adopts in Meditation I can

2. The Cartesian Doubt

. I want to discover what, if anything, is absolutely certain.

Descartes here declares that he will be just as wary of what is even

against his policy; for often the negation of an uncertain statement is

() If not-p is certain, then p has no credibility.

The only other premise is:

() If p has no credibility, then p is not uncertain.

() If not-p is certain, then p is not uncertain.

() If p is uncertain, then not-p is not certain.

The step from () to () is valid, because it has the following form:

Another way to see why the step from () to () is valid is to apply to it

. To do this, I need not examine all my beliefs individually, but

In order to understand what comes next, it is important to realize that

is a book on your desk or that someone is speaking to you. What are

. Until now, everything Ive accepted as most obvious and cer-

It is one thing, however, to say that something is accepted as certain,

Sometimes towers which had looked round from a distance

The occasional deceptiveness of the senses, then, provides a reason not

completely those who have deceived us even once (CSM II , SPW ,

. But the senses have sometimes deceived me; Id better not

With this step, Descartes introduces a theme that occupies him

Yet although the senses occasionally deceive us with respect to

Descartess point here is one that naturally occurs to anyone thought-

. Perceptions occurring in poor conditions of observation (e.g.,

The position reached in step would seem to be simply intelligent

The fact of dreaming, then, provides a reason to doubt even perceptions

. No, for even perceptions occurring in ideal conditions of ob-

In step , Descartes has reached a skeptical position with regard to the

[W]hether I am awake or asleep, two and three added together