Descartes, Malebranche, and the

Crisis of Perception
Descartes,
Malebranche, and the
Crisis of Perception

Walter Ott

1
 OUP CORRECTED PROOF – FINAL, 10/3/2017, SPi

3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Walter Ott 2017
The moral rights of the author have been asserted
First Edition published in 2017
Impression: 1
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 licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries 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
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2016962083
ISBN 978–0–19–879171–3
Printed in Great Britain by
Clays Ltd, St Ives plc
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents

Preface vii
Acknowledgments xi
Bibliographical Conventions xiii

Plan of the Book 1

1. The Crisis of Perception 6
1.1 Sensible Qualities 11
1.2 The Picture of Innocence 19

2. The Early Descartes 30

2.1 The Overlap Thesis 32
2.2 Corporeal Ideas and the Physiology of Perception 35
2.3 From Brains to Minds 42
2.3.1 Sensations and the analogy with language 44
2.3.2 Turning toward the brain 47
2.4 An Unnatural Geometry 57

3. The Meditations 63
3.1 Sensations, Ideas, and Intentional Resemblance 67
3.2 Material Falsity 79
3.3 Brain Images after L’Homme 84
3.4 The Special Modes of Thinking 86
3.5 Jobs for Sensations 93
3.6 Conclusion 97

4. The Dioptrique 100

4.1 Dethroning the Brain Image 101
4.2 The Proper Sensibles 105
4.3 Geometrical Qualities 107

5. Later Descartes 112

5.1 How to Paint Extension 113
5.2 Rapporter 122
5.3 Overview of Descartes 127
5.4 Malebranche’s Critique: the Selection Argument 131
vi CONTENTS

6. The Cartesians 137

6.1 Louis de la Forge 139
6.2 Robert Desgabets 144
6.3 Pierre-Sylvain Régis 148
6.4 Antoine Arnauld 150

7. Malebranche on Sensation 157

7.1 An Adverbial Theory? 159
7.2 Cartesian Pressures 166
7.3 The Role of Sensations 167

8. Early Malebranche 176

8.1 Robust Natural Judgments 176
8.2 Problems 179

9. Middle Malebranche 182

9.1 Five Things We Confuse in Sensory Experience 183
9.2 Seeing-as 187
9.3 Compound Sensations as Perceptions of Relations 190
9.4 Conclusion 192

10. Later Malebranche 194

10.1 Intelligible Extension 194
10.2 Efficacious Ideas 199
10.3 The Disappearing Of 202

11. Conclusion 218

Appendix: The Development of the Theory

of Natural Judgment 225

References 227
Index 237
Preface

I am keenly aware that scholars have trod, if not this particular path,
nearly every step of it on their way elsewhere. So I have tried to
confine myself to those topics about which I have something original
to say or, failing that, unavoidable steps on the road to my goal.
I have tried to write a book one might enjoy reading. As a result,
I have not written a review of the relevant literature for its own sake.
In the usual case, only those commentators whose work serves as a
useful foil for my own views appear in the text. But I have learned
from all the commentators I cite and many more besides. Without the
groundbreaking work of Douglas Lindberg and Gary Hatfield on the
history of the philosophy of perception I am sure I would never have
taken an interest in these matters. Reading John Sutton’s stimulating
and wide-ranging Philosophy and Memory Traces provided further
inspiration. And the work of Celia Wolf-Devine and Robert Pasnau,
to name just two more, provided guidance on important issues.
It is probably typical for philosophers to be most indebted to those
with whom they disagree. That is certainly the case here, and to an
unusual degree. If I should turn out to be right on some of these
points, it is only because others formulated clear and attractive theses
that forced me to think for myself. I have learned much from the
work of Alison Simmons, Nicholas Jolley, Steven Nadler, Tad
Schmaltz, Larry Nolan, and all the others with whose views I at
times take issue.
Let me say a quick word about my methods and assumptions.
Recent commentators, especially Peter Machamer and J.E.J. McGuire,
question the assumption that Descartes’s positions are largely stable
in his mature work.1 I share their skepticism. I spent an enormous
amount of time trying to shoehorn all of Descartes’s mature positions

1
See Machamer and McGuire (2009).
viii PREFACE

into a single view. Only when I gave up that ambition and let the texts
speak for themselves, on their own terms, did I begin to make any
progress.
No doubt there is a core set of doctrines that Descartes hews to
after 1637 or so. I’ll identify some of them below, and some of those
go right back to the even earlier Regulae. But in the cases I’ll cover, the
differences that emerge are so significant that one cannot mix views
from, say, Le Monde with those of the Principles. Conflating these
works means one has to cherry-pick some texts and ignore or ‘read
away’ others. As we shall see, some of the differences are superficial
and owe more to Descartes’s varied purposes—say, attaining enough
knowledge to aid ‘the designing of telescopes,’2 or giving an account
of the passions—than to any shift in doctrine. Nevertheless, there are,
I believe, substantial points of development.
Similar points apply to Malebranche’s work. Thomas Lennon and
Paul Olscamp’s translation of The Search after Truth is authoritative
partly because it works from the text that Malebranche himself, at the
end of his life, endorsed. But like many of us, Malebranche was keen
to elide the errors of his youth. The first edition of the Search differs
wildly from later ones on the issues that concern us, and so merits
separate treatment.3
My strategy, then, is to come at each text without the assumption
that it must be consistent with all the others. Consistency across
texts is a worthy goal and a welcome result when you find it. But if
someone asserts that p, I’ll take him to mean that p, unless evidence
within that same text shows otherwise.
I should also say something about why what follows is supposed to
be worthy of anyone’s attention. People are interested in the history

2
‘In the Optics my purpose was to show that one could make sufficient progress in
philosophy to enable one to achieve knowledge of the arts which are beneficial for life;
for the designing of telescopes, which I explained there, is one of the most difficult
projects ever attempted’ (Preface to the French edition of the Principles, AT IXB 15/
CSM I 187).
3
See esp. Geneviève Rodis-Lewis (1963) for an account of the differences among
the editions.
 PREFACE ix

of philosophy for all kinds of reasons. Some reasons are purely

antiquarian; others are not. One might want to see how a historical
figure deals with a problem in order to find a solution to it. Some-
times, one notices that a certain problem does not arise and wonders
what assumptions led to the problem in the first place. Are those
assumptions written in the nature of things, or are they optional
claims that can be jettisoned?
All of these motivations and more are perfectly legitimate, in my
view. But in this particular book, my primary interest is in finding out
what Descartes and Malebranche believe and why they believe it. At
least prima facie, their stories about perception are very strange:
Descartes at one stage thinks that in perception we ‘turn toward’
our own brains, and the images sketched there; Malebranche, that we
paint the idea of extension with the mind’s own colors. What lines of
argument would lead one to these conclusions? What assumptions
guide the debate? These are the chief questions I hope to answer.
Finally, I am wary of coming off as more dogmatic than I am. Prior
drafts were littered with expressions like ‘It seems to me that . . . ’ and
‘One might be forgiven for wondering whether . . . ’ Such verbiage is
hardly likely to propel the reader through dense prose. So I have
removed most of these qualifications. In their place, allow me to
preface the book with one great big ‘As far as I can tell . . . ’
Acknowledgments

Among all the debts I owe, the greatest is to Antonia LoLordo, who
suffered through a seminar I taught on these issues and even then was
willing to comment in detail on a draft. Her insightful comments
saved me from innumerable idiocies and shaped every page for the
better.
Peter Momtchiloff displayed super-human patience as the book
slowly took shape. I am especially indebted to the anonymous referees
for the press, one of whom provided a truly extraordinary set of
comments.
Ben Jantzen of Virginia Tech was an invaluable interlocutor on
these issues. Laura Keating was kind enough to offer helpful com-
ments on one of my earliest drafts of a paper on these issues in
Descartes.
Gallica (the website of the Bibliothèque Nationale de France),
Google Books, and the Bibliothèque électronique suisse have made
available digital versions of many rare and untranslated texts, obvi-
ating the need to travel to see the originals. Thanks a lot.
I am very grateful to Leslie Oakey for the cover art. Matthias Butler,
Clement Raj, and the design and production team at Oxford University
Press did a wonderful job seeing the book into print. Thanks to
Andrew Woodard and Michael Janes, whose acute copy-editing and
proofreading found all the mistakes I couldn’t.
My delightful neighbors, Gene and Jeanne Zechmeister, went to
the extraordinary length of reading and commenting on the final
draft, catching many errors and infelicities in the process.
I thank the editor and publishers of Philosophy and Phenomeno-
logical Research for permission to reprint portions of my ‘Male-
branche and the Riddle of Sensation’ 88, 3 (2014): 689–712.
xii ACKNOWLEDGMENTS

Thanks to Marty Fagin of Marty’s Wine Cellar in Ruckersville,

Virginia, who stocked the high-quality beer necessary for an endeavor
like this. Mention this book and get 5% off any order over $50.
I am very grateful to the University of Virginia and to my depart-
ment chair, Tal Brewer, who generously secured a (career-first) sab-
batical in the spring of 2015.
Finally, I owe a tremendous debt to Jorge Secada, who guided my
first sustained reading of Descartes twenty years ago.
Charlottesville, Virginia
January 1, 2017
 OUP CORRECTED PROOF – FINAL, 10/3/2017, SPi

Bibliographical Conventions

St. Thomas Aquinas

SCG Summa Contra Gentiles
ST Summa Theologicae
CDA Commentary on Aristotle’s De Anima
Francisco Suárez
MD Metaphysical Disputations
René Descartes
AT Oeuvres de Descartes. 11 vols. Ed. C. Adam
and P. Tannery.
CSM/CSMK The Philosophical Writings of Descartes. 3 vols. Vols. 1
and 2 ed. John Cottingham, Robert Stoothoff, and
Dugald Murdoch; vol. 3, ed. Cottingham, Stoothoff,
Murdoch, and Anthony Kenny.
G The World and Other Writings. Trans. Stephen
Gaukroger.
Antoine Arnauld
OA Oeuvres de Messire Antoine Arnauld. Ed. G. Du Pac de
Bellegarde and J. Hautefauge. 43 vols.
Nicolas Malebranche
OC Oeuvres Complètes de Malebranche, dir. A. Robinet
SAT The Search after Truth. Book.chapter.section: page
number in
LO SAT, trans. T. Lennon and P. Olscamp.
E Elucidations of SAT, in LO.
D Dialogues on Metaphysics and on Religion, dialogue and
page number in
JS D, trans. N. Jolley and D. Scott.
Robert Desgabets
RD Oeuvres philosophiques inédites
OUP CORRECTED PROOF – FINAL, 10/3/2017, SPi

xiv BIBLIOGRAPHICAL CONVENTIONS

Chronology and abbreviations

Descartes
Regulae (c.1628) Rules for the Direction of the Mind (unpublished)
L’Homme
(c.1629–32) Treatise on Man (unpublished)
La Lumiere
(c.1629–32) Treatise on Light1 (unpublished)
Dioptrique (1637) Optics
Discours (1637) Discourse on Method
Meditations (1641) Meditations on First Philosophy
Replies (1642) Replies to objections to the Meditations
Principles (1644) Principles of Philosophy
Notae (1647) Comments on a Certain Broadsheet
Passions (1649) Passions of the Soul
Arnauld and Nicole
Logique (1662) Logic, or the Art of Thinking
La Forge
Traitté (1666) Treatise on the Mind of Man
Malebranche
SAT (1674) The Search after Truth First edition (1674) Books 1–3
SAT Second
edition (1675) Books 4–6
Foucher
Critique (1675) Critique of the Search for the Truth in which is
examined at the same time a part of the Principles
of M. Descartes
Desgabets
CdC (1675a) Critique of the Critique of the Search after Truth
Supplément
(1675b) Supplement to the Philosophy of M. Descartes
Malebranche
SAT
Third Edition Elucidations of The Search after Truth

1
I follow Stephen Gaukroger’s practice of referring to these two treatises collect-
ively as Le Monde. See G vii.
 OUP CORRECTED PROOF – FINAL, 10/3/2017, SPi

BIBLIOGRAPHICAL CONVENTIONS xv

SAT The Search after Truth complete, including

Fourth edition Elucidations (subsequent editions include both the
(1678) Search and Elucidations)
Foucher
ND (1679) New Dissertation on the Search after Truth
Arnauld
Examen (1680) Examination of a Work entitled: Treatise of the Essence
of Body and the Union of Mind and Body, against the
Philosophy of M. Descartes (unpublished)
VFI (1683) On True and False Ideas
Malebranche
Réponse (1684) Response by the Author of the Search after Truth to
M. Arnauld’s Book, On True and False Ideas
SAT Fifth edition (1688)2
Dialogues (1688) Dialogues on Metaphysics and on Religion
Régis
Système (1690) General System according to the Principles of Descartes
Malebranche
SAT Sixth edition (1700)
SAT Seventh edition (1712)

Note on translations
Where no English version is cited, translations from French are my own.

2
This fifth edition claims to be the fourth, as does its 1678 predecessor. The
numbering of future editions is affected; the seventh edition purports to be the
sixth, the sixth the fifth. I am grateful to Tad Schmaltz for helping me sort out
the tangled history of the editions of the Search after Truth.
[B]y the most ancient, the most strange, and the most deeply
rooted of errors, we attribute to the external world what are
called sensible and corporeal qualities, such as light, heat, taste,
odor . . .
Robert Desgabets, Supplément de la philosophie de M. Descartes

It is obvious that one sees bodies only by color and that one can
only distinguish them as different by the difference of their colors.
No proof is necessary for this claim, save for a little reflection on
the effects of colors in painting.
Nicolas Malebranche, Réponse à Regis
 Plan of the Book

This is a book about how two French philosophers of the seventeenth

century, René Descartes and Nicolas Malebranche, along with some
of their countrymen, deal with a crisis in the theory of perception.
That century witnesses the demise of two central elements in philo-
sophical thinking about sensory perception, elements that go back
at least to Aristotle. First is the view that the proper objects of each
sense—qualities such as color and sound—are real members of the
mind-independent world. The common sensibles such as size, shape,
and motion are perceived by perceiving the proper sensibles. When
color and the rest are ejected from the mind-independent world,
philosophers find themselves compelled to offer a totally new account
of how it is that the mind comes to the common ones. Is it by
inferring size, shape, and motion on the basis of, say, color or felt
pressure? Or is it a purely automatic operation, accomplished by
divine decree? Our experience of the proper sensibles becomes prob-
lematic as well. How is it that a sensible quality like color gets
‘localized,’ that is, experienced as being on the surfaces of the body
that causes it?
The crisis also owes its origin to a second development. Since
Galen, philosophers such as Ibn al-Hacen and Roger Bacon contrib-
uted to a unified program known as the ‘Baconian synthesis.’ This
view posits species that pass into the eye and through the ventricles of
the brain, ultimately uniting with species from the other senses in the
organ of the common sense. Johannes Kepler in effect demolishes
the Baconian synthesis. A new story has to be substituted; and it must
be one that respects the new austerity of the world beyond the mind.
 PLAN OF THE BOOK

Descartes’s ambitious early work aims to develop a fresh account

of the physiology of perception. On this view, there is no material
species passing into the pineal gland, which Descartes takes to be the
organ of the common sense. Instead, motion reaches into the brain
and sketches an image on the pineal gland. Armed with this notion of
a brain image, Descartes believes he can explain the ‘perception’ of
non-human animals and even human beings when their minds are
not attending to their environments.
It would be a shame to have developed such a sophisticated
account of mindless ‘perception’ only to drop it when the activity of
a human mind experiencing its environment is at issue. The early
Descartes believes he can use his new physiological account to explain
how we cotton on to the common sensibles in a world devoid of
proper ones. On this view, or so I shall argue, it is by being aware
of the image sketched on the pineal gland that we become aware of
objects in the world. The purely corporeal brain image functions as a
representation because it resembles an object in the subject’s envir-
onment. When the mind is aware of the brain, it thinks through this
image out to the object. For this reason, the early Descartes, and even
the Descartes of the second Replies, does not hesitate to call the brain
image an ‘idea.’
Rather than struggle to read away this part of Descartes’s view,
I show how it makes sense in its context. For it is a mechanistic
counterpart of Aquinas’s position, according to which one thinks
about singular things by convertendo se ad phantasmata, ‘turning
toward the phantasms,’ which are purely corporeal images. Descar-
tes’s version of this conversio doctrine is not a bold and counter-
intuitive innovation but rather an effort to hew as closely to tradition
as his ontology will allow.
Even in his earliest stages, however, Descartes thinks human vision,
when accompanied by attention, is supplemented by a variety of
mental calculations, which he calls ‘natural geometry.’ (As we’ll see,
in offering two means of judging distance, one purely corporeal and
one involving reasoning of some kind, Descartes is following the well-
worn Baconian path.) Natural geometry is not necessary for perception,
 PLAN OF THE BOOK 

whether merely mechanical or mental. Non-human animals and

inattentive humans navigate their environments successfully without
it. Thus Descartes’s early view allows him to give a univocal explan-
ation of what I call the ‘overlap region,’ that set of behaviors exhibited
by both animals and inattentive humans. What is more, the same
physiological elements that underwrite ‘perception’ in the overlap
region are exploited by the theory of mental perception.
As time goes on, Descartes becomes aware of the manifest short-
comings of the brain image. That image is subject to numerous
distortions and explains far less about the phenomenology of percep-
tion than Descartes needs it to. As a result, his middle and later work
eschews the brain image. While we still need to be aware of motions
in the brain or in the eye, the brain image drops out of the account as
a mere side effect of those motions. In its place, Descartes relies more
heavily on the natural geometry announced in the early work. The
prospect of giving a univocal account of the overlap region dims. If
the human perception of distance and position requires the exercise
of natural geometry and is not merely supplemented by it, it is hard
to see how Descartes proposes to account for the apparently quite
sophisticated ‘perceptions’ of non-human animals. What is more,
Descartes now needs to provide new answers to our questions above:
if it is not by being aware of a brain image that one becomes aware
of the size, shape, and position of bodies outside the mind, how,
exactly, is that accomplished? And how does the mind come to localize
the sensible qualities it experiences?
On Descartes’s final view, the mind’s ideas and sensations are
simply triggered when it becomes aware of motions in the brain.
They then have to be projected (‘rapporter,’ ‘referred’) on to bodies.
Although Descartes never drops the view that the mind is aware of
the brain when it undergoes sensory experience, the explanatory
power of the brain is at its nadir in this stage. That makes it all the
more difficult to see just on what basis the mind is going to refer its
sensations and ideas to objects in the world.
It is at this array of views that Nicolas Malebranche levels his most
intriguing argument against Descartes, what I call the ‘selection
 PLAN OF THE BOOK

argument.’ In short, Malebranche thinks that nothing about the

physiology of the brain or eye can direct the mind to summon this
or that idea or sensation. And nothing can guide the mind in pro-
jecting its sensations on to bodies. It seems obvious to Malebranche
that Descartes’s picture is backwards: the mind has to use its sensa-
tions of color to individuate bodies in the first place. It will take some
work to tease out just which of Descartes’s views Malebranche has in
mind at various stages of the selection argument and judge its effect-
iveness. I shall argue that the selection argument poses a genuine
challenge for each iteration of Descartes’s views.
Although there are some orthodoxies among the Cartesians—chief
being the distinction between mind and body—their theories of
perception are a heterogeneous bunch. Louis de la Forge and Robert
Desgabets build on Descartes’s final and earliest works respectively,
while Pierre-Sylvain Régis is most influenced by Descartes’s middle
period. On Régis’s view, we sense color and sound and the rest but
only imagine size, shape, and motion: we have to discern the invisible
outlines of objects by clothing them in the mind’s qualities. For
Desgabets, the mind has to inspect a brain image. The case of Antoine
Arnauld is especially intriguing in this light. I argue that his depar-
tures from Descartes are less important than they seem, and that he,
too, feels the sting of Malebranche’s selection argument.
But if Malebranche has a keen eye for weaknesses in the Cartesian
position, his own early view is vulnerable to much the same objection.
By peeling away the layers of later additions and revisions, we’ll find
the first edition of The Search after Truth introducing the doctrine of
natural judgments. These judgments perform a variety of functions:
assessing distance and position, correcting for optical illusions and
distortions, and explaining the localization of sensible qualities.
Within three years, however, Malebranche changes his mind. The
second edition of the Search robs these natural judgments of their
force and indeed ownership: they are now done for us by God. Their
nature is altered as well: instead of being judgments in any interesting
sense, they are now simply compound sensations.
 PLAN OF THE BOOK 

Later developments travel even further from Descartes’s work.

Malebranche replaces the picture of the mind coming into contact
with a single idea of a body, which in turn represents that body.
Instead, Malebranche posits a single, entirely uniform idea of exten-
sion in God’s mind, which he calls ‘intelligible extension.’ In the end,
I argue that, as far as sensory perception is concerned, Malebranche
does not even believe that there are ideas, in any interesting sense. Put
differently, he aims to return the notion of an Idea to its proper
Platonic context. For intelligible extension does none of the work of
a Cartesian idea. As a result, I argue that, taken to its logical conclu-
sion, Malebranche’s position abjures the whole picture of the mind
using a representation to think about the material world.
Whether each step of this debate represents a historical accident
or a reasoned progression toward Malebranche’s final position is, of
course, among the questions I hope to answer. By working through
that sequence of positions and arguments, I hope to show, at a
minimum, just how formidable are the challenges created by the
twin demise of the Aristotelian and Baconian views, and just what
lengths our philosophers are willing to go to in order to solve them.
 1
The Crisis of Perception

The goal of this chapter is to lay out Descartes’s predicament. As

someone who wants to contribute both to natural philosophy and
metaphysics on the question of perception, he finds himself deprived
of the traditional sources. To put the point more positively: he finds
the field wide open, given the twin rejections of sensible qualities such
as color and the empirical theory embodied in the Baconian synthesis.
Before beginning, it would be wise to say something about what
I take Descartes and his followers to be trying to explain. To my mind,
the problem of sensory perception includes a wide array of questions,
which we can group under the following four headings:

- Epistemic. To what degree does undergoing a sensory experience

confer justification on the associated beliefs?
- Causal. What is it that brings about the rich experience any
sentient being is capable of? What role is played by the extra-
mental objects? What role, if any, is played by automatic sub-
conscious mental activity, or by introspectively available mental
inference?
- Phenomenological. What is the most accurate description of the
deliverances of experience, from the first person point of view?
- Representational. What, if anything, about the subject’s experi-
ence makes it the case that the subject is experiencing this or that
physical object?
I propose to set aside the epistemic question altogether, as pursuing
it alongside the others would be a distraction. The other three groups
 THE CRISIS OF PERCEPTION 

are so tightly intertwined that I cannot imagine separating them

completely. For instance, any answer to the causal question will
presumably have to be faithful to the phenomenology: if our theory
says that we see objects by seeing pictures etched on our pineal gland
(as Descartes at one time holds), then that causal story seems to
violate the deliverances of introspection. By the same token, a mental
representation of a particular body might well figure in one’s causal
account. Nevertheless, the causal question will be at the forefront of
the coming chapters. What is the means by which cognitive contact
with the ordinary objects of sense experience is achieved?
Before we can consider that question, something must be said
about the notion of representation that is in play. What follows in
this section necessarily engages contemporary philosophy, for some
of the commentators against whom I’ll be arguing attribute distinct-
ively twentieth-century positions to these figures. Descartes, in par-
ticular, has been read as holding a ‘teleosemantic’ theory of sensory
representation. So we can’t engage in these debates without fixing the
meanings of these terms of art.
Nearly all of the figures we’ll look at hold that states that exhibit
underived about-ness do so because they are mental representations.
At least since Aristotle’s On Interpretation, this claim has been a
ground-level assumption. It says only that something ‘in’ the mind
(or brain) allows us to think about the world. The real work of about-
ness is done at the level of the mental, and language, whether spoken
or written, is at best the shadow of what passes in the mind.
What do mental states represent? One intuitive—if only because
familiar—contemporary notion is something like this: a mental state
represents the world as being thus-and-so. A state represents its
satisfaction conditions. But, at least for the moderns, this notion of
representation takes up the story too late. According to the view we’ll
consider, to represent the world as being thus-and-so requires prop-
ositional content, which is a function of sub-propositional elements
that are themselves representations. So a representation, in the sense
we need, can’t be identified with a set of satisfaction conditions.
Rather, we need representations of something more basic, like properties
 THE CRISIS OF PERCEPTION

or objects. These representations can then be combined to form a

thought about how the world is: that the moss on the boulder is wet,
for example.
With this as background, we can make a first pass at defining
representation, in the moderns’ sense:

a is a representation of b for S just in case it is in virtue of a that S thinks of b.1

Even as a first pass, this is too weak, since ‘in virtue of ’ is notoriously
slippery. If it is purely a causal claim, then it seems false: a might
trigger a thought of b without in any way representing b, as when a
song makes you think of an event in your past.2 What we want for our
first pass, I think, is a notion of ‘in virtue of ’ that is fully explanatory
in the following sense: the explanation for S’s thinking of b has to
explain why S thinks of b and not c. In our song example, for instance,
the song triggers a thought of b but does not explain why the thought
is a thought of b rather than c or d. By contrast, a representation is
something that allows the mind thinking of it to think through it to
something else.
The schema above is designed to be neutral between two ways in
which it might be developed: ‘internalism’ and ‘externalism.’ I’m
using these terms in my own, purely stipulative, way. The distinction
I have in mind is not to be identified with that between ‘narrow’ and
‘wide’ content, which exercises many contemporary philosophers.
We’ll need these two -isms to understand the competing interpret-
ations of our figures.

1
One might worry that this definition forces the moderns into an objectionable
homuncularism. To see that it doesn’t, note that S in our definition can be satisfied
either by a physical subject, a Cartesian mind, or a system or sub-system within either
of them. Ruth Garrett Millikan helpfully distinguishes between the producer and the
consumer of a representation. See esp. her (1987) and (1989). These can, but need not,
be within a single organism. For example, a beaver that splashes its tail to warn of
danger counts as a representation-producer; the consumers are the other beavers who
flee as a result.
2
Nor do we want ‘in virtue of ’ to be construed only as realization: although one
might claim that state a just is S’s thinking of b, it would be unwise to build that into
the very notion of representation. I owe these points to Brie Gertler.
 THE CRISIS OF PERCEPTION 

The contrast we want, I think, lies between ‘current time slice

theories’ and ‘historical or genetic’ theories.3 The former hold that
what a representation is of turns solely on how things are right now.
‘Internalism’ will include any such time-slice view. A historical the-
ory, by contrast, claims that what a state represents is at least in part
a function of its causal or genetic background.4 ‘Externalisms’ are
historical, since they deny that all the facts about a given time-slice
suffice to determine the representandum of a representational state.
Instead, one has to look to the causal or selectional history of the state.
Consider the crudest possible analysis of representation: literal,
pictorial resemblance. This will be our paradigm case of internalism.
What makes the image of the Eiffel Tower represent the Eiffel Tower
turns on just two things: the image and the Eiffel Tower. Where the
image came from and what role that state was selected to play are
beside the point. Hence imagism is a current time-slice view.5
Our ‘externalists’ deny all this. On their view, the history of a state
makes a crucial contribution. The most popular such views point to
the causal origin of the state, or the role nature selected it to play in
the cognitive economy of the creature that has it. We might lump the
causal and teleosemantic variants of externalism together as ‘tracking’
views, since both take representation, at least in perceptual contexts,
to be a matter of tracking properties in the subject’s environment.
According to these tracking views, one might know all there is to
know about the intrinsic nature of a state and of the external world
(as it currently is) and yet, absent a knowledge of the causal or selec-
tional history of the state, still not know what it represents.

3
This is roughly how Alvin Goldman draws the line between epistemic internal-
ism and externalism. See Goldman (1979, 141).
4
I should note that ‘what’s true right now,’ in my use and Goldman’s, cannot
include facts about the past; if it did, the distinction would collapse.
5
Note that an internalist in my sense is perfectly free to hold that the extension of a
representation—what it picks out in the world—is determined partly by what is in that
world. Consider the inevitable Twin Earth scenario, where water is really XyZ and not
H2O. Your picture of watery stuff picks out H2O partly because H2O is what, in your
world, resembles your picture.
 THE CRISIS OF PERCEPTION

Externalisms—historical rather than current time-slice theories—

are perfectly free to treat sensations as blank effects, devoid of any
internal powers of representation. All that counts for representation
in this externalist sense is whether a state can track the features of
the environment the theory says it tracks. A tracking view can make
anything at all into a representation, provided it plays the right role.
Now, the representation must have some intrinsic nature or other, or
else the system wouldn’t be able to re-identify and exploit it. But
beyond that, I can see no real constraint on what sorts of intrinsic
natures will work. Although fairly obvious, this point is easy to miss
in the cloud of confusion that billows up around all talk of represen-
tation. When a representational theory is attributed to a philosopher,
it is vital to figure out from the start what kind of representation is on
offer. Again, I am not claiming that the moderns themselves so much
as entertain tracking theories; that is still in question. Nor am
I claiming that every modern philosopher thinks there is a substantive
account of representation to be had.
I take Descartes and Malebranche to be trying in part to explain the
simplest perceptual transactions, cases in which a subject experiences
the world around her. There are three features of sensory experience
in particular that need explaining:
1) Individuation. We experience objects as different from one
another.
2) Positioning. We experience bodies as occupying distinct regions
of space and hence as standing in relations of distance to each
other and to us.
3) Localization. We experience color and other sensible qualities
as belonging to objects outside of us.6 The red of the apple

6
The classic discussion of localization, and space perception in general, from Kant
onward is Gary Hatfield’s (1990). For a treatment of localization in Thomas Reid, see
Lorne Falkenstein (2000). As I shall understand it here, the notion refers to the
localization of what come to be called secondary qualities: whatever process it is by
which we come to experience, say, color as out there, in a determinate expanse of
space. Falkenstein (2000) clearly has this sense in mind, but builds in other tasks, such
as ‘localization on a two dimensional field of view, depth perception, perception of
 THE CRISIS OF PERCEPTION 

appears to be on the surface of the apple and not on the basket

the apple rests on.

Although these explananda become most clear in the work of his

successors, we shall also be asking how Descartes deals with them.
The demise of the Aristotelian and Baconian views makes these three
processes especially problematic, and so it is chiefly with these three
issues that we shall be occupied. I should make clear that I am not
assuming that individual subjects in fact perform these processes
themselves: that is very much up for argument. What is not up for
argument, I think, is that our experience, from the inside, really does
exhibit all three features. With all of this throat-clearing behind us, we
can proceed to the main business of the chapter: setting out the
relevant features of the intellectual situation in which Descartes
finds himself.

1.1 Sensible Qualities

Philosophers of the modern period have the at once charming
and deeply insulting habit of treating their forebears as intellectual
infants.7 In the Principles, Descartes presents the transition from
error to truth as a journey from childhood naïveté to sophistication.
‘The chief cause of error,’ he tells us, ‘arises from the preconceived
opinions of childhood.’8 These preconceived opinions turn out to be
just the views commonly held by the scholastics. Descartes’s opponents

objective magnitude, [and] coordination of visible with tangible spaces’ (2000, 306,
n.4). I shall be using ‘localization’ in its narrowest sense, and refer to these other issues
as such when they arise. I should also note one other point of departure. Falkenstein
calls localization ‘a complex achievement,’ which suggests that it is something that the
subject does. Not all of our philosophers will agree, and so I do not want to build that
feature into the very notion of localization.
7
This variety of scorn is heaped on their contemporaries as well. Antoine Arnauld,
for example, explicitly charges Malebranche with laboring under the misapprehen-
sions of childish beliefs; see Arnauld (1990, 58 f.).
8
Principles I.71, AT VIIIA 35–6/CSM I 218. Descartes’s point is not that we believe
things based on authority when we’re young; instead, he’s talking about the beliefs the
child naturally forms by being ‘so closely tied to the body.’
 THE CRISIS OF PERCEPTION

are, in effect, overgrown children—adults who never overcame their

native prejudices. Chief among the opinions of these innocents is the
attribution to objects of ‘tastes, smells, and so on.’9 Objects have, or at
least can have, ‘sizes, shapes, motions, and the like.’ But taste or smell
are not even possible modes of bodies.
From our jaded perspective, it’s surprisingly hard to articulate the
innocent view. Whether any philosopher managed to retain his
innocence long enough to write about it is open to question.10 Either
way, the philosophers of the seventeenth century typically credit
themselves with taking that innocence away. On their view, as Robert
Desgabets puts it in 1675, ‘color, sound, odor, and all that one calls
sensible qualities’ are at most the effects of objects acting on our sense
organs.11 The features that were real to the innocent are unmasked as
mere artifacts of our sensory interaction with the world. Whatever
properties bodies possess, the ‘sensible’ qualities on Desgabets’s list
are not among them. While this doctrine admits of considerable
variation in the seventeenth century, it was, in broad outlines, already
well entrenched by Desgabets’s time, having been articulated by
Galileo, Hobbes, Descartes, and others. Indeed, Desgabets, with all
the scorn of the newly-minted sophisticate, claims that a moment’s
reflection will suffice to show ‘how ridiculous one would make one-
self, if one imagined that there is anything in a flame beyond the
movement of the tiny parts that make it up.’12
The banishment of colors, sounds, and the rest from the mind-
independent world might be the distinctive break that divides the
seventeenth century from its medieval and ancient past. Desgabets
claims that ‘although some ancient philosophers, and even Saint
Augustine, were not entirely ignorant of the doctrine . . . one can

9
Principles I.71, AT VIIIA 36/CSM I 219.
10
Jonathan Lear (1988, 111 f.), for example, argues that even Aristotle offers a
dispositionalist account of color. I find Stephen Everson’s (1997, 112) arguments to
the contrary persuasive.
11
CdC 30.
12
CdC 32. La Forge (1666, 110) says much the same thing: if people could
persuade themselves that sensible qualities resemble something in the object, itself
no mean feat, ‘would they not become ridiculous?’
 THE CRISIS OF PERCEPTION 

nevertheless regard it as a discovery of our century, since it does not

seem that, before our time, anyone drew the consequences that
naturally flow from it.’13 From Desgabets’s point of view, the ejection
of sensible qualities from the world outside the mind is the fulcrum
on which everything else turns. Without draining such qualities from
the world, natural philosophy and metaphysics are doomed to wan-
der in darkness, and man is ‘an enigma to man himself.’14 More than
fifty years later, David Hume claims that ‘[t]he fundamental principle’
of modern philosophy ‘is the opinion concerning colours, sounds,
tastes, smells, heat and cold; which it asserts to be nothing but im-
pressions in the mind, derived from the operation of external objects,
and without any resemblance to the qualities of the objects.’15
The distinction Hume and Desgabets have in mind unquestionably
makes for a cleaner ontology where the external world is concerned,
since physical objects can now be regarded as instantiating only the
properties studied by geometry, plus motion.16 But if the distinction
simplifies one’s ontology, it spawns a variety of fascinating compli-
cations for the philosophy of perception. If sensible qualities are not
part of the world outside the mind, just what are they, and what role,
if any, do they play in our cognitive economy?
At this point, we find ourselves in a terminological minefield. There
is no uncontroversial way to refer to the two classes of properties
Descartes isolates. Locke, of course, will come to call them ‘secondary’
qualities, but just what he takes the difference between primary and
secondary qualities to be is controversial. Nor is Descartes himself
much help: rather than giving a name to the class of properties
the sophisticated mind jettisons from the world, Descartes typically

13
CdC 35.
14
Desgabets (Supplément, RD 165); cp. Malebranche, SAT I.x/LO 49.
15
Hume (2000), Book I, Part IV, section 4, paragraph 3.
16
Some philosophers, such as Robert Boyle and John Locke, entertain the possi-
bility of unobservable non-geometrical qualities in bodies. (See my 2009, 182 f. for
discussion.) Still, these unobservable qualities will not, trivially, appear on the list of
sensible qualities.
 THE CRISIS OF PERCEPTION

lists a few and then adds ‘and so on,’ a practice Galileo initiated in
the Assayer.17
That in fact is the best way to proceed. Instead of trying to give a
definition that will isolate all and only the qualities we want, we
should list them and then, for our own convenience, impose a
name. Just what these qualities consist in—what their real definition
is—will be a matter of debate, and so trying to give such a definition at
the outset would beg the question.
Descartes’s list in the Principles, which overlaps with Galileo’s,
includes:

- tastes
- smells
- sounds
- heat
- cold
- light
- colors18

Later in the Principles, Descartes addresses the senses one by one,

allowing us to supplement his original list with haptic qualities:

- heaviness
- heat
- humidity19

I propose to follow Desgabets, along with Antoine Arnauld, Pierre-

Sylvain Régis, and others, and call all such qualities ‘sensible qualities.’20
Whatever else they turn out to be, they are qualities that we confront in
sensory experience. We still lack a name for the other qualities we

17
For example, Galileo writes, ‘these tastes, odors, colors, etc., so far as their
objective existence is concerned, are nothing but mere names for something which
resides exclusively in our sensitive body’ (1623/2009, 22). In the Principles (I.71),
Descartes speaks of ‘the sensations of tastes, smells, sounds, heat, cold, light, colours
and so on’ (AT VIIIA 35/CSM I 219).
18
Descartes provides this list in the Principles I.71 (AT VIIIA 35–6/CSM I 219).
19
Principles IV.191 (AT VIIIA 318/CSM I 282). Cp. Louis de la Forge (1666, 108).
20
See Arnauld and Nicole (1996, 49) and Régis (1690, 77 f.).
 THE CRISIS OF PERCEPTION 

experience. Matters are no easier here. To call them ‘primary’ is to court

confusion by importing the associated doctrines of Boyle and
Locke and others. To call them ‘real’ is to suggest that, in line with
seventeenth-century usage, they are capable of existing apart from the
substance in which they inhere, an issue orthogonal to ours. I propose,
then, that we refer to them as ‘geometrical’ qualities, since, apart from
motion, the most prominent qualities on most philosophers’ catalogs
are precisely the properties singled out by geometry. But again, I do not
wish to make this part of their definition: what makes them geomet-
rical, in my sense, is merely their appearance on the lists of Galileo and
Descartes.
We should add to the list of geometrical qualities such features as
position and distance. These are clearly relational, whereas size and
shape are (at least arguably, and on the surface) monadic.21 Still, they
seem to deserve a place in our list of mind-independent features.
So we have merely sensible qualities on one hand and geometrical
qualities on the other. Innocents suppose bodies have both; sophisti-
cates, that they have only the latter. Now we need a way of talking
about our experiences of these qualities. With the proviso that I am
merely stipulating, I propose to use ‘sensory experience’ to refer to
whatever is going on in the subject (construed to include both mind
and body) in ordinary, veridical cases of perception. By ‘sensations,’
I will continue to mean our experiences of sensible qualities.22 No
term suggests itself for sensory experiences of geometrical qualities, so
I shall have to allow the context to make my meaning clear.

21
With Newton, philosophers come to recognize that a thing’s shape is partly
a function of the gravitational fields being exerted on it. It is then in some sense
relational rather than intrinsic. And even in our period, some philosophers, such as
Malebranche, argue that shape is a relation of points along an object’s surface. On
Malebranche’s view, shape will still count as a monadic property of the body that has
it, in the sense that, while relational, it is not a relation between this body and some
other.
22
Talking of sensations and ideas might suggest that Descartes holds what Alison
Simmons (2003) calls a ‘bifurcation’ account of experience, one that neatly slices
experiences of sensible qualities off from those of geometrical ones. Whether he holds
such an account or not is a matter of controversy, and I do not mean my stipulations
to decide the matter. We’ll return to this issue in chapter three.
 THE CRISIS OF PERCEPTION

Before proceeding, I have to emphasize the distinction between

sensations and sensible qualities. Sensations are sensings of sensible
qualities. Maybe further argument will conflate these, so that there is
no real sense in which a sensation is of anything at all. To talk about a
sensation of red, or F#, is just to refer to a certain kind of experience.
But this has to be the result of argument: we should not begin by
taking the conflation for granted.
Now we can ask just what thesis Galileo and Descartes mean to
assert when they reject sensible qualities. Galileo’s case is the clearest.
Galileo argues that sensible qualities ‘are nothing but mere names for
something which resides exclusively in our sensitive body, so that if
the perceiving creatures were removed, all of these qualities would be
abolished and annihilated from existence.’23 Sensible qualities are
mind dependent in a straightforward sense: no perceivers, no sensible
qualities, though there might well be geometrical qualities left over.
Galileo is not suggesting that sensible qualities are real features of the
world that emerge from the contact of subject and object; instead, he
is an eliminativist.
At times, Descartes seems to agree. In the Principles, he claims that
sensations of tastes, smells, and the rest ‘do not represent anything
located outside our thought,’ although the French edition goes on to
add that they do ‘vary according to the different movements which
pass from all parts of our body to the part of the brain to which our
mind is most closely joined and united.’24 Note that Descartes’s claim
is not merely that, as it happens, sensible qualities do not exist in the
world. It is the much stronger claim that such qualities are not
even features objects might have had. By contrast, we clearly perceive
‘size, shape and number to be actually or at least possibly present in
objects in a way exactly corresponding to our sensory perception of
understanding.’25
Later in the Principles, though, Descartes seems to identify sensible
qualities with dispositions of objects to cause sensations in us.

23 24
1623/2009, 22. Principles I.71 (AT VIIIA 35–6/CSM I 219).
25
Principles I.70 (AT VIIIA 34/CSM I 218).
 THE CRISIS OF PERCEPTION 

‘[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 various disposi-
tions in those objects which make them able to set up various kinds of
motions in our nerves.’26 The context makes it clear, however, that
Descartes means by ‘dispositiones’ what we would call arrangements.27
The French version replaces ‘dispositiones’ with ‘figures, situations,
sizes, and motions of their parts.’28 The name ‘red,’ then, is supposed
to refer to whatever arrangement of particles in a body that causes the
relevant sensation in us. This is not a dispositionalist account, for red
is getting identified with the categorical base of the disposition, not the
disposition itself.
One plausibly dispositional account is presented by Kenelm Digby,
in his 1658 Two Treatises. Consider Digby’s analysis of color: it is
‘nothing else, but the disposition of a bodies superficies [surface], as it
is more or less apt to reflect light.’29 Digby’s use of ‘apt’ suggests that
he means ‘disposition’ in our contemporary sense.
So we have three, seemingly very different, views: one that denies
that sensible qualities exist outside the mind; one that identifies such
qualities with the categorical bases of a body’s disposition to cause
sensations in us; and one that identifies the qualities with those
dispositions. The distinction seems important: after all, two positions
say that sensible qualities exist and the other one says they don’t.
But as Robert Boyle points out, the ‘famous Controversie’ among
these positions is merely verbal.30 After all, no eliminativist needs to
claim that our experiences of sensible qualities lack causes. Nothing
prevents her from agreeing with the other two views that there are

26
Principles IV.198 (AT VIIIA 322/CSM I 285).
27
Even seventeenth-century English allows for this use of ‘disposition’ to mean
‘arrangement.’ Locke, for example, writes that ‘we may in reason consider these four
degrees: the first and highest, is the discovering and finding out of Proofs; the second,
the regular and methodical Disposition of them, and laying them in a clear and fit
Order, to make their Connexion and Force be plainly and easily perceived’ (Essay
IV.xvii.3: 669).
28 29
AT IXB 317. 1658, 315.
30
See Boyle (1664, 74–5), quoted in Stuart (2013, 116, n.14).
 THE CRISIS OF PERCEPTION

bodies whose micro-level properties are responsible for our having

the experiences we do. Whether one chooses to dignify these capaci-
ties, or the properties that underwrite them, with the name ‘red’ or
‘humid’ is a practical question.31
In light of this, I shall always mean by ‘sensible quality’ what
Galileo, Descartes, and all the rest mean when they offer their lists.
I will have to talk about the dispositional or categorical base for our
experiences of secondary qualities as such, without the benefit of a
new term. This is the only way I can see to avoid confusion.
Before proceeding, we need to dip into contemporary philosophy
once again. For it is tempting to suppose that the moderns’ sensible
qualities can be identified with what we would call ‘qualia’ or ‘phe-
nomenal character.’ Succumbing to this temptation would, I think, be
a serious mistake.
The core of the debate over ‘what it’s like’ is what Joseph Levine
calls the ‘explanatory gap.’32 The gap stretches between neural states
on one hand and the what-it’s-like of experience on the other. How
can states of the hunk of meat between our ears explain mental states,
with their own peculiar phenomenal character?
This question differs in at least two ways from any of those we’ll be
dealing with. The first difference is obvious in retrospect but easy to
miss. Anger and pain are not sensible qualities, since they do not even
seem to be in or on bodies, and yet they are paradigmatic examples of
what-it’s-like-ness. Even our perceptions of qualities like size and
shape come with their own phenomenology. Thomas Nagel asked
what it’s like to be a bat, not what it’s like to be a bat tasting coconut.
So the first difference is one of scope: the problem of phenomenal

31
Many would disagree. Consider Gary Hatfield’s distinction between ‘Powers:
Secondary qualities in objects are powers for producing non-resembling sensory ideas
in perceivers’ and ‘Bare Causation: The physically basic properties of bodies serve as
bare causes for sensations of color, etc., and such colors exist only in the mind’ (2011,
310). While there are some contexts (e.g., the debate over how to account for
metamers) in which the distinction does some work, it is of considerably less
ontological significance than appears.
32
See Levine (1983, 354).
 THE CRISIS OF PERCEPTION 

character concerns more kinds of mental states than just our experi-
ences of sensible qualities.33
The second difference is more fundamental. The explanatory gap
lies between brain states and the distinctive phenomenology of
experience. Saul Kripke invites us to consider the firing of neurons
on one hand and pain on the other, and then wonder how they can be
the same thing. We can generate the same sense of wonder in the face
of more sophisticated views than the identity theory Kripke targeted:
Ned Block’s China brain does the same thing for functionalism. Each
ontology of mind compatible with physicalism generates its own
version of the explanatory gap.
But these worries are alien to our philosophers, simply because
none of them is trying to square physicalism with phenomenology.
Our philosophers are worried instead about a different gap altogether:
that between the qualities then recently evicted from the mind-
independent world on one side and the qualities bodies are still
permitted on the other. It’s not a question of what it’s like to see
red; it’s a question of how the red we see is related, if at all, to objects
out there in the world.34

1.2 The Picture of Innocence

If my purposes required an exhaustive summary of all relevant
thought about perception prior to the seventeenth century, I would

33
One might suppose that phenomenal character was never a problem until the
moderns ‘kicked the sensible qualities upstairs.’ But that would be a mistake: whatever
one does with the sensible qualities, phenomenal character will still be a problem, for
the simple reason that states having such a character are not exhausted by states
having to do with sensible qualities.
34
Sydney Shoemaker (2002) helpfully distinguishes between Levine’s ‘subjective’
explanatory gap and its ‘objective’ counterpart, which is of ‘greater antiquity.’ As
Shoemaker puts it, the objective gap asks ‘how colors, given their perceived nature,
can be, or be realized in, physical properties of things, given what we know about these
physical properties’ (2002, 2). Now, for most of the moderns, Shoemaker’s objective
gap is a question with a clear answer: sensible qualities are not, nor are they realized in,
physical properties. Still, Shoemaker’s objective gap helpfully steers us away from the
obsession with the relationship between phenomenal states and the brain.
 THE CRISIS OF PERCEPTION

give up now. Happily, there are only a few key points we need to
grasp, in order to understand what our figures are up to. Moreover,
how the early moderns understand their scholastic competitors is, for
our purposes, as important as what those competitors themselves
really think. I begin with a brief ‘big picture’ discussion of sense
perception in Aristotle and Aquinas before moving on to theories of
brain function, ultimately focusing on vision and the Baconian syn-
thesis that is falling apart just as Descartes begins his career.
For the Aristotelian, sensory experience is common to humans and
animals. Animals are entirely corporeal beings nevertheless capable
of sentience. Since we and animals share a sensitive soul, we share a
baseline of capacities and experiences, all of which are the result of
entirely corporeal entities and their faculties. In ordinary sensory
experience, sensible qualities are impressed on one of the external
senses. The sensible species that result are ultimately gathered by the
common sense, which unifies the impressions of the various senses.
The common sense achieves a representation of the object that caused
the initial sense impressions; what is more, it represents because it is
a similitudo of that thing.35 Nevertheless, this ‘sensible species is not
what is perceived, but rather that by which the sense perceives.’36
We and animals also share a capacity for memory and imagination.
This requires us to store sensible species, at which point they are
called phantasms. Phantasms are ‘images of individuals, and exist in
corporeal organs.’37 While they help explain the recognitional and
predictive capacities of human and non-human animals alike, phant-
asms play two further roles in distinctively human cognition.
Lacking an intellect, animals are confined to thinking of particu-
lars. Only beings equipped with an intellect are able to produce the
intelligible, rather than sensible, species that allow them to think of
universals. We are able to do this by abstracting the intelligible species

35
In what follows, I draw primarily on Aristotle’s de anima (especially II.6, II.13,
and III.1), Aquinas’s commentary on the de anima (CDA), as well as his ST 1 q.84 and
q.85. All quotations from Aristotle are from Aristotle (1984).
36
ST I q.85 art 2.
37
ST I q.85 art 1. reply to obj.3; see CDA III.13, 392.
 THE CRISIS OF PERCEPTION 

from the phantasm.38 It is only because we have a corporeal image

that represents a single individual that we are able to extract its latent
intelligible content. Where abstraction is concerned, phantasms are
indeed the object, rather than the means: ‘intellective cognition does
not occur without a body—but in such a way that the body serves as
the object, not the instrument.’39
There is, however, a problem lurking. The intellect’s proper object
is always the universal. If this is so, how is it able to think of
individuals, its improper objects? And here we see the second role
for phantasms. Aquinas tells us repeatedly that the only way to
intellectually cognize an individual is to ‘turn toward the phant-
asms.’40 The intellect cognizes individuals only by turning toward
something corporeal, namely, ‘the likeness of an individual thing.’41
This doctrine of ‘conversio’ will become important below, where we
will look at it in greater detail than is necessary now.
So much for what is happening in the subject. What must the world
outside the mind be like, if the Aristotelian view is to function? A
hallmark of innocence is its belief that bodies have sensible qualities.
Nevertheless, a standard Aristotelian view does draw a distinction
whose extension is nearly identical to that of the sensible/geometrical
distinction: that between proper and common sensibles. On Aristotle’s
view, such qualities as ‘motion, standing still, shape, size, number, and
unity’42 are revealed by multiple senses, while others are proper to one
sense only. Color is a proper sensible, but no less real for that. Note that
Aristotle’s list of common sensibles overlaps considerably with the
early moderns’ lists of geometrical (or, for Locke, ‘primary’) qualities.43

38
ST I q.85 art 1.
39
Sententia libri De Anima I, lecture 2, number 19, trans. Hain (2015, 54).
40
See, e.g., ST I q.84 art 7.
41
ST I q.84 art 7 reply to obj.3.
42
This is Aristotle’s list from de anima III.1, trans. Pasnau, in Aquinas (1999, 291).
See also de sensu 442b5-6 and de memoria 450a10, which adds time.
43
Compare Locke’s list of ‘simple Ideas of divers Senses’ (Essay II.v.1: 127)—
number, extension, figure, rest, and motion—with his primary qualities, which
include number, extension, figure, motion, and solidity. The one exception is solidity,
a primary quality detectable only by one sense, namely, touch.
 THE CRISIS OF PERCEPTION

The moderns on the whole jettison the proper sensibles. In so doing,

of course, they incur a new burden: explaining how the common
sensibles are grasped in sensory perception. The Aristotelians can
give this a straightforward answer: it is by and through seeing color,
for example, that one sees size, shape, and the rest. What is more, these
latter qualities are just the same qualities you experience if you touch
an object with your eyes closed. Here it is important to be clear: one is
not seeing color and shape, as if they were two different things. Instead,
one is seeing a thing’s shape by seeing its color. As Aquinas puts it, ‘the
only time sight apprehends size or shape is insofar as it apprehends
something colored.’44
Nevertheless, there are a number of threats to the innocence of this
picture. First, note that Aristotle and Aquinas speak of a sense as
judging the common sensibles on the basis of the proper ones. On
their view, a sense cannot err with respect to its per se object but can
with respect to its per accidens object, the common sensibles. So when
it comes time to judge ‘what that colored thing is or where it is,’ it can
make mistakes.45 In apprehending size and shape, sight is, it seems,
making a further judgment beyond that of color-presence, but it is
doing so on the basis of the color sensation. We need not be bothered
by Aristotle’s attributing judgment to the senses, for he might be
speaking loosely of judgments made in sense perception.46 Neverthe-
less, it is starting to look as if the common sensibles are really common
judgeables, not directly perceived by any one of the senses. We can call
this the ‘indirect’ reading.
The indirect reading seems confirmed when Aristotle argues, at the
start of de anima iii, against the view that the common sense has its
own external sense organ, over and above the other such organs.
Aristotle writes,

44 45
Aquinas (1999, 205). Aquinas (1999, 204).
46
Ronald Polansky (2007, 376, n.23) argues that Aristotle is indeed speaking
loosely, since he has not yet distinguished sense perception from thought, as he will
in III.iii.
 THE CRISIS OF PERCEPTION 

[T]here cannot be a special sense-organ for the common sensibles either, i.e.
the objects which we perceive incidentally through this or that special sense,
e.g. movement, rest, figure, magnitude, number, unity; for all these we
perceive by movement, e.g. magnitude by movement, and therefore also
figure (for figure is a species of magnitude), what is at rest by the absence
of movement: number is perceived by the negation of continuity, and by the
special sensibles; for each sense perceives one class of sensible objects.47

The problem is that the common sensibles seem to be perceived

incidentally. This suggests that we can be said to perceive shape thro-
ugh color only in the same way in which we perceive Cleon’s son by
perceiving a white thing that happens to be Cleon’s son. But then we
are no longer, strictly speaking, perceiving shape at all. We would be
perceiving something that in fact has shape, but not as shaped. As
Anna Marmodoro puts the point: the link between color and shape, or
between white and Cleon’s son, would be ‘ontological, not perceptual;
external, not part of the perceptual content.’48 We merely see a white
patch that turns out to coincide with Cleon’s son. On such a view, the
common sensibles are not really sensible at all.
But Aristotle explicitly tells us, in the same passage, that ‘in the case
of the common sensibles there is already in us a common sensibility
which enables us to perceive them non-incidentally; there is therefore
no special sense required for their perception.’49 Unless Aristotle has
contradicted himself, he cannot really mean what he seems to mean at
452a13.
Faced with this apparent contradiction, commentators have resorted
to a wide variety of maneuvers, including suggesting that Aristotle
equivocates on ‘incidental.’ Joseph Owens offers a more plausible
reading. On his view, Aristotle is presenting a view he does not agree
with. If there were a special organ for the common sensibles, then we

47
425a13–15. See Joseph Owens (1982) for a thorough commentary on the issues
of translation involved.
48
Marmodoro (2014a, 177). Marmodoro reads Johansen (2012, 180) as appealing
to accidental perception to explain our grasp of the common sensibles. I am not sure if
this is quite what Johansen means. Nevertheless, Marmodoro’s take on accidental
perception seems to me to be entirely correct.
49
425a27–8.
 THE CRISIS OF PERCEPTION

would perceive the common sensibles only incidentally through the

other senses and not directly, as we in fact do.50 As Aristotle goes on to
say: if there were a special sense for a common sensible like movement,
‘our perception of it would be exactly parallel to our present perception
of what is sweet by vision.’51
Whether this is the correct reading of Aristotle’s text or not, it will
serve us handily, since it is consistent with Aquinas’s take on the
passage:
[W]hen [Aristotle] says, ‘Further, there cannot be a special sense-organ . . . ,’
because someone could say that there is some other sense cognitive of the
common sensibles, he eliminates this by the following argument. Whatever is
known by one sense, as its proper sensible, is not known by the other senses,
except accidentally; but, the common sensibles are not sensed accidentally by
any of the senses, but per se by many; therefore, the common sensibles are
not the proper objects of any one sense.52

The common sensibles are, then, perceived per se and not inciden-
tally. We do not infer shape from color; rather, both shape and color
are perceived per se. The difference is that color is perceived only by
vision; that is the sense in which color is its ‘proper’ object.
There remains the problem of judgment. As Aquinas says, we
apprehend shape through vision only by apprehending color. Color
is a real feature of the body that makes it visible. And since the color
ends where the body does, there is no need to infer to the shape of a
body from our experience of its color. How, then, can a sense be more
trustworthy with regard to its proper sensible than it is with regard to

50
Paul Symington (2011, 248) directs us to Aquinas’s commentary on Metaphysics
V, Lecture 15, 980, 982, 983, in Aquinas (1961). There, Aquinas argues that, since
whiteness is indivisible, it can only be individuated by belonging to the object in which
it inheres. So although whiteness exists in its own right, it does so only by inhering in
quantity. In this sense, whiteness is accidentally quantitative. None of that conflicts
with my interpretation, since it does not make our perception of that quantity through
sight accidental.
51
425a23.
52
Commentary on III.1 (425a13 f.) in Aquinas 1946, section 575. Pasnau’s trans-
lation (in Aquinas 1999, 296) of the same passage is slightly more awkward, though
the same in substance as Kocourek’s earlier version.
 THE CRISIS OF PERCEPTION 

the common ones? Doesn’t the whole notion of the senses judging
shape suggest that shape is not immediately perceived, but only
inferred?
Here it is helpful to consider a thought experiment suggested
by Aristotle but expanded on by Aquinas. Suppose we had only one
sense modality, say, sight. In such a case, we would not be able to tell
the difference between size or shape and color. As Aquinas writes,

[S]ince [sight] is concerned only with color, and since color and size follow
from one another (for sense is altered by size and color concurrently), we
could not distinguish between color and size; instead, they would seem to be
the same. But as things are, since size is sensed by a different sense than sight,
whereas color is not, this in itself makes it clear that color and size are
different. And it is the same for other common sense objects.53

From the perspective of a being that has only sight, there is no

difference between color and size. It is only through our experiences
from other senses that we are able to distinguish them: touch, for
instance, apprehends size but not color.54
This thought experiment lets us see just what judgment, in the
Aristotelian sense, contributes. One sense can reveal faults in an-
other’s judgments, as touch can correct optical illusions. Although
one is seeing shape by seeing color, and sensing shape by sensing, say,
warmth or pressure, it is in principle possible for the shape delivered
up by sight to be different from that delivered by touch, as when one
sees and touches a stick partially immersed in water. None of this
would be possible if we had only one sense. So the point Aristotle and
Aquinas are making is epistemic, rather than phenomenological.
When sight goes wrong about a common sensible, that is not because
it is making an inference on the basis of an unshaped color to this or
that shape. Rather, the color experience already includes an experience

53
Aquinas (1999, 298).
54
Thomas Johansen (2012, 178 f.) and Pavel Gregoric (2007, 69–82) use this
thought experiment to defend their ‘deflationary’ reading, according to which the
common sense is not another ability or capacity alongside that of the other senses. For
criticism, see Marmodoro (2014a).
 THE CRISIS OF PERCEPTION

of shape. In non-veridical cases, however, other senses can be used

to correct the (seen) shape. In short, the Aristotelian view remains
innocent.55
Even if what I’ve said so far captures the broad outlines of the
Aristotelian doctrine, we have to fill in some of the details. To do so,
we need to go well beyond Aristotle and even Aquinas. The doctrine
of common sensibles has to be underwritten by some physical pro-
cess. On the view we’re considering, sensory experience is the act of
the corporeal organs. And even if the common sense is not another
external sense like sight, it nevertheless requires its own corporeal
organ. I propose to focus on vision, though I hope the implications
for the other senses whose deliverances meet in the organ of common
sense are clear.
One influential—arguably dominant—theory of vision is the so-
called ‘Baconian synthesis,’ a combination of views with its roots
in the second century A.D. works of Galen, running through the
eleventh-century work of Ibn al-Haythem (also known as ‘Alhacen’),
and reaching fruition in the thirteenth-century works of Roger Bacon,
Witelo, and John Pecham.56
We can begin on the side of the object and work inward. Color is
what makes an object visible. The color exists formally in the object.
As it radiates out from that object, it exists in the transparent medium only
‘intentionally’ or ‘spiritually.’ So far we have not departed much from
Aristotle or Aquinas.57 As Malebranche sums up the scholastic view,

55
I of course realize there are many related problems one might raise about the
common sensibles; hence the prodigious efforts of Gregoric (2007), Marmodoro
(2011) and (2014b), and Johansen (2012). My goal here is to deal only with those
problems directly relevant to ours.
56
It is important not to exaggerate the degree of uniformity exhibited by writers
in this tradition. Roger Bacon (1928, vol. 2) dissents from Alhacen on a number of
points, including whether the eye itself emits species and whether the visible species
exist immaterially or materially in the intervening medium (though Bacon himself
continually downplays these differences). See Wolf-Devine (1993, 30) for Descartes’s
reading of Kepler, Witelo, Alhazen, and Pecham (but probably not Roger Bacon
himself).
57
See Pasnau 1997 on intentional or immaterial existence.
 THE CRISIS OF PERCEPTION 

The most commonly held opinion is that of the Peripatetics, who hold that
external objects transmit species that resemble them, and that these species
are carried to the common sense by the external senses. They call these
species impressed, because objects impress them on the external senses.58

Once these species reach the eye, they exist in it as a physical

representation. The image next has to pass through the optic nerve
to the ventricles, small cavities in the brain. An image projected on
the back of the eye would be too wide to pass through the narrow
channel of the optic nerve. So rather than make the image on the back
of the eye the first stopping point for these species, Alhacen and the
tradition that follows him single out a much narrower area in the
glacial humor.
Once in the brain, the image has to be collated with the input from
the other senses. The medievals typically locate the organ of the
common sense in the ventricles of the brain: the images or species
transmitted from the various senses are gathered there, where they
exist in the animal spirits. At this point, we need to consider brain
anatomy, as understood by the tradition. One useful way to approach
the view is to think of it as attributing to the cardiovascular system the
functions we would ordinarily attribute to the nervous system.59 The
corporeal medium of thought, in which the species transmitted from
objects ultimately exist, is blood. To be more precise, the medium is a
special kind of rarefied, almost spiritualized blood.
From Galen’s day until the work of Andreas Vesalius in the mid-
sixteenth century, the majority of writers on these topics noted that
the brain exhibits a curious structure at its base known as the rete
mirabile, or ‘marvelous net.’60 This network of fine blood vessels is
responsible, on their view, for refining the blood pumped up to the

58
SAT III.ii.2: 220. In the SAT, Malebranche assumes that the species are corpor-
eal; many of his arguments against the view depend on this assumption. But the
typical scholastic view holds instead that, at least in the space between the object and
the human eye, the species have an intentional or spiritual, rather than physical,
existence. Roger Bacon is a notable dissenter; see his Opus Majus (1928, vol.2, 459 f.).
59
See Wolf-Devine (1993, 61).
60
See Clarke and Dewhurst (1972).
 THE CRISIS OF PERCEPTION

head by the heart. In their terms, it transforms the vital spirits, in

which life consists, into the animal (from anima, or soul) spirits, in
which thought and representation occur.
As the image makes its way deeper into the brain, it reaches the
animal spirits and becomes increasingly rich in content. As A. Mark
Smith presents it, the physical representation at the surface of the eye
becomes the visual representation in the eye, which in turn becomes
perceptual and finally conceptual in the ventricles of the brain. This
process is a series of inferences, or quasi-inferences; its precise status,
and the degree of intellectual or conscious involvement in it, seems to
me unclear.
While Galen was right about the rete mirabile and its place at the
bottom of the brain, he was right only with regard to the animals
whose brains he examined: pigs and oxen. No such structure exists in
humans. And although the anatomist Andreas Vesalius knew this as
early as 1543, belief in the rete mirabile was curiously persistent. Even
Thomas Willis, a founding member of the Royal Society and an
acquaintance of Locke’s, in 1643 allowed that the rete mirabile exists
in some humans, though not all. When it does, it tends to make them
rather sluggish and slow-witted, as befits a structure found in pigs
and oxen.61
Johannes Kepler’s 1604 Ad Vitellionem Paralipomena presents a
challenge to the reigning orthodoxy.62 Kepler rejects the idea that the
image could be projected in the glacial humors in the center of the
eye. For him, what counts is the image projected at the back of the
eye. Kepler, of course, discovers that the image there is inverted both
left-right and top-down.63 If this is the means by which vision

61
Clarke and Dewhurst (1972, 78).
62
Both Douglas C. Lindberg (1976) and Gary Hatfield and William Epstein (1979)
emphasize the continuities between Kepler and the earlier orthodoxy. While their
points are well taken, the differences are more important than the similarities in the
present context.
63
See Kepler (2000, 181): ‘Vision thus occurs through a picture of the visible object
at the white of the retina and the concave wall; and those things that are on the right
outside, are depicted at the left side of the wall, the left at the right, the top at the
bottom, the bottom at the top.’
 THE CRISIS OF PERCEPTION 

happens, the Baconian synthesis is wrong; there is no way an image of

that size could be squeezed into the optic nerve and passed along to
the ventricles of the brain.
What happens after the image is formed on the retina is something
Kepler only briefly speculates about.64 ‘I leave to the natural philo-
sophers to argue’ about this. ‘For the arsenal of the optical writers
does not extend beyond [the] opaque wall [formed by the retina.]’65
In short, as Smith argues, Alhacen, Bacon, and the others are inter-
ested in explaining vision; Kepler is interested in optics.66
Kepler’s work spells the end of the Baconian synthesis but hardly
supplants it, for the simple reason that Kepler tells us very little about
what happens in vision after the formation of the retinal image.
Taken together, the demise of both the Baconian theory of perception
and the broader Aristotelian view of the proper and common sensi-
bles make for a crisis in thinking about perception that will shape
philosophy in the seventeenth century and beyond. And it is at
precisely this crossroads that we find Descartes, as he begins writing
Le Monde.

64
See, e.g., Kepler (2000, 180–1).
65
Kepler (2000, 180), quoted in Smith (2004, 193).
66
Smith might be slightly over-stating matters. As Hatfield and Epstein show,
Kepler in one work at least goes so far as to say that the immaterial image is somehow
transmitted to the organ of the common sense. Nevertheless, Kepler does not offer a
definitive and clear account of how this happens, or of how visual experience is
generated as a result. See Hatfield and Epstein (1979, 373 f.).
 2
The Early Descartes

Descartes’s first response to the crisis of perception is a bold attempt

to use the purely mechanical apparatus of the eye and brain to explain
the positioning and individuation of bodies in perception.1 There is
no substantive story about the localization or pairing of sensible with
geometrical qualities, but that in part is due to the early Descartes’s
ambition to explain perception in mechanical terms. After all, non-
human animals are mere machines, bereft of any sensations at all, and
yet they get along just fine.
I begin with a brief discussion of Descartes’s position on animals.
This allows us to fix our sights on exactly what abilities and states
Descartes thinks can be explained purely mechanically. The early
Descartes holds what I call the ‘overlap thesis’: any behaviors exhib-
ited by animals and by humans not attending to their environment
must receive the same mechanical explanation. We begin, then, by
investigating purely mechanical perception. Many will bristle at this
conjunction of words: shouldn’t perception be reserved for thinking
things? Nevertheless, as we’ll soon see, Descartes does speak of
‘corporeal memory,’ inviting us to apply that adjective to the other
states and processes that animals can also instantiate. In any case, or

1
This chapter takes up Descartes’s early works, which include the Regulae, Le
Monde, and Discours. Although he never published it, Descartes appears to have
been working on Le Monde in the early 1630s, around the same time as his Dioptrique.
(Desmond Clarke (2003, 55) tells us that ‘dioptrics’ is used, in the seventeenth century,
to refer to studies of light refraction, as opposed to reflection (‘catoptrics’). Descartes’s
ambitions go well beyond both.) I shall argue in chapter four that Le Monde and the
Dioptrique differ in significant ways and hence merit separate treatment.
 THE EARLY DESCARTES 

so I’ll argue, whether or not one wishes to add inverted commas to

such phrases is largely a matter of taste.
Corporeal perception, I argue, is grounded in the representational
abilities of an image etched in the brain. Descartes exploits the best
surviving remnants of the Baconian synthesis and unites them with
Kepler’s work to produce an ingenious, if false, empirical account
of the origin of the brain image. This image can then be stored in
corporeal memory and can interact with the animal spirits in such a
way as to produce goal-directed behavior.
The brain image has a role to play in attentive perception as well.
Descartes insists that the mind ‘turns toward’ the brain or body when
it perceives or imagines. That claim has been a source of vexation to
his readers almost from the beginning. I shall argue that Descartes
means just what he says. In the early work, his claim is well motivated.
Descartes, after all, holds the overlap thesis at this stage. It makes
sense for Descartes to treat attentive human perception as building on
the physiological basis already in place, rather than simply replacing
it altogether. It would be inelegant in the extreme if the same physio-
logical process that subserved animal and inattentive human percep-
tion turned out to be totally irrelevant to attentive human perception.
What is equally important, the view that the mind inspects the brain
when it senses or imagines has scholastic antecedents. In fact, I’ll
argue that far from being a novelty, it represents Descartes’s attempt
to fall in step with the Aristotelian view to the greatest degree
permitted by his ontological innovations.
In the end, we shall find the early Descartes positing three, usually
concurrent, processes to account for attentive sensory perception.
The mind becomes aware of the brain image, whose representative
powers allow the mind to become aware of the geometrical qualities
of the bodies around it. Second, the mind decodes the signals of the
brain and ‘represents to itself ’ the appropriate sensation. Finally, the
mind is able to deploy a series of semi-conscious inferences Descartes
calls ‘natural geometry’ to make its judgments of geometrical qualities
more accurate. After setting out the overlap thesis, I explore each of
these processes in turn.
 THE EARLY DESCARTES

2.1 The Overlap Thesis

It may surprise readers—as it did me—to learn that Descartes began
his posthumous career, not as the arch dualist responsible for creating
the problem of mind-body interaction, but as a crypto-materialist
who hid his real doctrines to escape ecclesiastical persecution.2 While
hardly accurate, this counter-caricature is a useful corrective. For
the early Descartes in particular assigns a vast array of functions to
the body alone that later dualists would come to attribute solely to the
mind.3 To get at this, we need to know something of Descartes’s
position on animals.
Descartes’s project with regard to animals must be seen in light
of his anti-Aristotelian program. Just as we can explain combustion
without any appeal to scholastic forms, so we can explain animal
behavior without appealing to sensitive, animal, rational, or indeed
Cartesian, souls. If we then apply the principle of parsimony, we can
say that animals are merely machines in that they lack souls in any of
these senses.4 Throughout his career, Descartes vacillates between the
claim that animals do not experience, see, or perceive, at all, and the
claim that they do all these things, just not in the same sense that we
do.5 To my mind, this is a distinction without difference. The core of

2
See, e.g., the opening of Stephen Gaukroger’s (1995) biography of Descartes,
which recounts the legend of its protagonist’s mechanical doll, Francine.
3
The early Descartes thus inverts the popular caricature of him in the non-
philosophical world. The influential psychologist Timothy D. Wilson, for example,
takes Descartes to be claiming that the mind is always transparent to itself (which is
fair enough) but then claims that if there were a Cartesian mind, its body would be
unable to get out of bed in the morning. For it would have no faculty of propriocep-
tion, which takes place at a level below consciousness (Wilson 2002, 18–19). As we’ll
see, Descartes thinks the brain is capable of performing all of the functions Wilson
attributes to his ‘adaptive unconscious.’ Since Wilson evidently thinks the adaptive
unconscious is a function of the brain, he and the early Descartes in fact agree.
4
As Lex Newman (2001) has persuasively argued.
5
In his correspondence, Descartes typically says that animals do not have ‘feelings
just like ours’ (to Mersenne, July 30 1640, AT III 121), or that they do not ‘think as
we do’ (to Marquess of Newcastle, 23 November 1646, AT I 576), or that they do not
‘have sensations like ours’ (to More, February 5 1649, AT V 276–7). (Stephen
Gaukroger collected and translated these passages (2002, 202)). On the other side,
he says in the sixth Replies that ‘brutes possess no thought whatsoever’ (AT VII 426/
 THE EARLY DESCARTES 

the bête machine doctrine is the claim that only mechanical principles
are needed to explain animal behavior. Whether one sees this as
debunking the pretensions of those who attribute thought to animals,
as Descartes seems to, or instead as bolstering the explanatory
powers of matter and motion, is a question of emphasis. Even in the
absence of a soul, animals can do many of the same things ensouled
bodies do.6
In his early work, Descartes goes beyond the bête machine doctrine
and defends the overlap thesis. First, Descartes isolates a region of
behavior that is shared by humans and animals, namely, all activities
that we perform without attending to them that animals can also
perform. Descartes says in the Discours that machines can perform
‘precisely those [functions] which may occur in us without our thinking
of them, and hence without any contribution from our soul (that is,
from that part of us . . . whose nature is simply to think). These functions
are just the ones in which animals without reason may be said to
resemble us.’7 Such functions include not just pedestrian ones like
respiration, reflexes, and digestion but, as Descartes tells us in the
Fourth Replies, ‘walking, singing, and the like, when these occur without
the mind attending to them.’8
The overlap thesis simply says that any behavior in the overlap
region must get the same mechanical explanation. In a 1637 letter to
Plempius, Descartes writes, ‘I explain [in L’Homme] quite explicitly
that my view is that animals do not see as we do when we are aware
that we see, but only as we do when our mind is elsewhere . . . In such
a case we too move just like automatons.’9 It is important to see that
the overlap thesis is independent of the bête machine doctrine: to say

CSM II 287). It is hard not to conclude that this difference is one of emphasis rather
than substance. In corresponding with people he takes seriously, Descartes presents
his view in a moderate form; when replying to the authors of the sixth Objections, he
takes a more polemical approach.
6
For the debate on sensory states in animals, see esp. John Cottingham (1978),
Stephen Gaukroger (2002, esp. 200 f.), Lex Newman (2001), Gordon Baker and
Katherine Morris (2002), Cecilia Wee (2005), and Janice Thomas (2006).
7 8
AT VI 46/CSM I 134. AT VII 230/CSM II 161.
9
AT I 413/CSMK III 61–2.
 THE EARLY DESCARTES

that animals are machines, instantiating no mental properties but

only those of res extensa, is not to say that behavior in the overlap
region must get the same explanation.
For us humans, it is only by virtue of being associated with such
a sophisticated machine that we can walk even while our thoughts
drift towards dinner. Otherwise, we would stumble about like Holly-
wood zombies each time our attention wandered. Indeed, in this early
stage, Descartes attributes quite a number of important functions to
the machine, which it can carry out even without being ensouled. In
L’Homme, he imagines God creating a purely mechanical human body
and denying it a soul. Consider the list of functions and states attribut-
able to such unensouled human bodies, which Descartes adduces in
L’Homme and largely repeats in the Discours:

. . . the reception by the external sense organs of light, sound, smells, tastes,
heat, and other such qualities, the imprinting of the ideas of these qualities
in the organ of the common sense and the imagination, the retention or
stamping of these ideas in the memory, the internal movements of the
appetites and passions, and finally the external movements of all the limbs
(movements which are so appropriate not only to the actions of objects
presented to the senses, but also to the passions and the impressions found in
the memory, that they imitate perfectly the movements of the real man).10

Although a mere machine, the counterfeit man is capable of combin-

ing perceptions, memories, and desires in such a way as to mimic the
actions of a real one, at least where behaviors in the overlap region are
concerned. Like other animals, the counterfeit man will flee danger,
seek food, and respond warily to known threats. The claim that a
mere machine can do all these things is itself startling, one Arnauld
finds difficult to believe.11 That the very same kinds of things happen

10
AT XI 202/CSM I 108. For the Discours, see AT XI 55–6/CSM I 139.
11
Arnauld writes: ‘[I]t seems incredible that it can come about, without the
assistance of a soul, that the light reflected from the body of a wolf onto the eyes of
a sheep should move the minute fibers of the optic nerves, and that on reaching the
brain this motion should spread the animal spirits throughout the nerves in the
manner necessary to precipitate the sheep’s flight’ (AT VII 205/CSM II 144). Roger
Bacon uses the sheep and wolf example to show that sensitive souls also contain an
 THE EARLY DESCARTES 

in human bodies when their souls are inattentive is a natural, but not
necessary, corollary.
Note that Descartes is not committed to giving precisely the same
explanation for attentive human behavior and a corresponding
inattentive behavior. Although we and a sheep might see a wolf at
the same time, and share many of the same physiological reflexes and
reactions, the attentive human has a soul that can act as fountain
keeper, directing the flow of animal spirits from the brain to the
limbs; this is what makes ensouled bodies capable of voluntary action.
Nevertheless, I shall argue that full-blown attentive human percep-
tion involves (though is not exhausted by) the same physiological
processes that allow for animal ‘perception.’

2.2 Corporeal Ideas and the Physiology

of Perception
Now that we know what kinds of things can be explained mechanic-
ally, let’s see how that explanation is supposed to work. For now,
I propose to ignore human minds and whatever roles they might play
and focus exclusively on the physiology of perception. And here the
intellectual background is especially important. Recall that Kepler’s
work on vision has come into conflict with the Baconian synthesis
just as the doctrine of the common sensibles, founded as it is on
Aristotelian innocence, is crumbling. Only when seen against the
backdrop of this crisis can Descartes’s early work come into focus.
I begin by sketching the general form of mechanical perception
before asking how it allows Descartes to incorporate Kepler into a
broadly Baconian picture of vision, even if the world beyond the mind
lacks sensible qualities. This background will allow me to argue that
corporeal perception requires corporeal representation, which turns
out to be literal resemblance.

‘estimative’ faculty, which, among other things, allows animals to ‘have some
perception in things advantageous and in things harmful’ (1928, vol.2, 425).
 THE EARLY DESCARTES

We should start with an overview of the full process of perception,

derived from L’Homme:

1. External objects act on the sense organs and produce qualities

like light, sound, odor, taste, and heat.
It would be truly extraordinary if by ‘light’ and ‘sound’ Descartes
meant something extra- or super-mechanical. That he does not is
already signaled in the Regulae, where Descartes insists on the literal
truth of Aristotle’s famous wax and seal analogy. There, he claims that
external objects make impressions of themselves on the sense
organs.12 What he means by ‘odor’ existing in the nose or brain
cannot be the sensible quality, but rather its corporeal analog.
2. Ideas of these qualities are impressed on the ‘common sense,’ which
is the seat of memory and imagination. That is, the motions of the
animal spirits trace tiny figures on the brain. (Only those figures that are
traced on the surface of the pineal gland should be counted as ‘ideas.’)13
By claiming that it is the motions of the animal spirits that create
the picture in the pineal gland, Descartes can avoid positing a numer-
ically identical thing that has to migrate from the sense organ to the
ventricles of the brain. When Descartes claims in the Regulae that
ideas come to the common sense ‘pure and unmixed,’ he is at pains to
point out that he does ‘not conceive anything real’ passing from the
ear, say, to the pineal gland, any more than a real entity passes from
one end of a quill pen to the other.14
3. The impression of these ideas on the memory results in their
being retained, in just the same way that needles might leave traces in
cloth ‘which would make [the original holes] very easy to open again.’
‘This shows how the recollection of one thing can be excited by that of

12
AT X 412/CSM I 40.
13
‘[It is only] those [figures] that are traced by the spirits in the surface of gland
H [the pineal gland], which is the seat of the imagination and of the common sense,
which should be taken for ideas, in other words, for the forms or images that the
rational soul immediately considers, when, being united to this machine, it imagines
or senses some object’ (AT XI 176–7).
14
AT X 414/CSM I 41.
 THE EARLY DESCARTES 

another, which had been imprinted in the memory at the same time.
For example, if I see two eyes with a nose, I immediately imagine a
forehead and a mouth, and all the other parts of a face, because I am
unaccustomed to seeing the one without the other.’15
The details of this fanciful account need not detain us. What is
important is the possibility of purely mechanical recognition, in the
same way that a vending machine can ‘recognize’ a coin.
4. The appetites and the passions act in concert with these brain
states to produce motion appropriate to the environment. Note that the
machine doesn’t merely follow through or initiate a set of movements;
it does so because of the causal connections among its memories,
desires, and perceptions.
Although nothing as detailed as step 3 appears in the Regulae,
Descartes is quite clear even there about step 4. He claims that he
can explain ‘how all the movements of the other animals come about,
even though we refuse to allow that they have any knowledge [‘cog-
nitio’] of things, but merely grant them a corporeal imagination.’16
Putting Descartes’s account in its intellectual context can help
make what is from our point of view utterly mysterious seem natural.
Descartes is willing to follow Kepler, his ‘1er maistre en optique,’17 as
far as Kepler is willing to go. As we have seen, Kepler explicitly
abjures speculation about what happens after the image is formed
on the retina. So after that point, Descartes’s only guides are the
Baconian synthesis and his own speculation. Descartes effects his
own synthesis, bringing together Kepler’s work and the best of the
Baconian tradition, while purging that tradition of its reliance on

15
AT XI 174–9/G 151. For more on Descartes’s mechanical account of pattern
recognition, see John Morris (1969).
16
AT X 415. Every other occurrence of ‘cognitio’ in this Rule is rendered as
‘knowledge’ by CSM. But for some reason, here they translate it as ‘awareness’
(CSM I 42). Now, Descartes might well at this stage be denying animals awareness
too; but, given that knowledge is the official and announced topic of this part of Rule
XII, it would seem odd for him to equivocate on it in the way CSM does.
17
See the letter to Mersenne of March 31 1638 (AT II 86), quoted in Wolf-Devine
(1993, 30). Wolf-Devine lumps Kepler in with the writers of the Baconian synthesis,
but as we’ve seen, to do so is to obscure their very different goals.
 THE EARLY DESCARTES

corporeal species or images that travel, intact, from the external

senses to the common sense.
That Descartes is still indebted to the Baconian tradition is obvious
from his reliance on the general framework of animal spirits and
brain ventricles. Indeed, at least in L’Homme, Descartes believes in the
rete mirabile, which refines the vital spirits into their more spiritual,
‘animal’ form.18 Chief among the barriers to uniting these two
systems—Kepler’s and the Baconian synthesis—are the difficulty of
getting the image through the optic nerve into the brain and the defects
in that image that might make it unusable by the common sense.
The first problem is solved by the ‘pure and unmixed’ way in which
the image is transferred. Only the motion is preserved, and there is
no need for a corporeal image to travel through the optic nerve. The
second is more pressing, for if the image at the back of the eye were
reproduced, just as it is, on the common sense, it would be inad-
equate, or so Descartes thinks, as a representation of the creature’s
visible environment, since, as Kepler notes, it is inverted left-right and
up-down.19
Descartes’s ventricular theory suggests to him an ingenious maneuver.
The pineal gland is a bit like a projection screen, sitting in a brain
cavity and having images sketched on it by the animal spirits. In this
respect, it is structurally similar to the eye itself: in both cases, there is
a surface on which an image is being traced. If the image on the retina
is flipped and inverted, because of the origin and arrangement of the
rays that produce it, why could that image not also be flipped and
inverted as it ‘passes’ through to the pineal gland? In fact, Descartes
goes on to explain precisely how that happens.20 In effect, the pineal

18
The arteries that carry the spirits from the heart, ‘after having divided into
countless small branches and having composed the little tissues that are stretched
out like tapestries at the bottom of the cavities of the brain’ (AT XI 129/G 105–6). Even
the account of the Passions requires that the animal spirits, generated by the heart, be
filtered out by blood vessels in the brain (see Passions I.10, AT XI 334–5/CSM I 331–2).
19
As Descartes explicitly notes in the Dioptrique, the image’s ‘parts are reversed,
that is to say, they are in a position totally contrary to those of objects’ (AT VI 123).
20
See AT XI 174 f./CSM I 105 f. Cp. figure 45 in the Dioptrique.
 THE EARLY DESCARTES 

gland becomes another retina, on which the properly oriented image

can be projected.21
Descartes’s achievement is not a trivial one. He exploits the resources
of an otherwise moribund theory to correct for the chief problems posed
by Kepler’s results. He now has a right-side up, properly oriented little
picture in the center of the brain. And that picture includes not just the
deliverances of vision but of the other senses as well.
All of this background puts us in a position to ask: how does the
corporeal image represent the object in the world? And by now it
should be obvious: resemblance. The corporeal idea represents the
extra-bodily object by virtue of resembling it in respect of size and
shape. Although this claim is controversial, it should seem natural
enough: after all, why would Descartes go to such great lengths to
preserve the exact resemblance of the brain image, if that resemblance
is to do no work?
We can confirm our analysis of corporeal representation as literal,
pictorial resemblance by considering Descartes’s account of corporeal
memory.22 Throughout his career, Descartes will insist that there is
such a thing and that it functions by means of resemblance.23 In the
Conversation with Burman, Descartes (reportedly) says, ‘when, for
example, on hearing that the word K-I-N-G signifies supreme power,
I commit this to my memory and then subsequently recall the
meaning by means of my memory, it must be the intellectual memory
that makes this possible. For there is certainly no relationship (‘affi-
nitas’) between the four letters (K-I-N-G) and their meaning, which
would enable me to derive the meaning from the letters.’24 Notice

21
Descartes’s illustrations in L’Homme indicate as much. Admittedly, the prov-
enance of those illustrations is not always clear. But the text confirms that the
illustration is an accurate reflection of Descartes’s meaning. Clarke (in his preface to
his translation of la Forge 1666 (1997), xiii) tells us that Clerselier engaged la Forge to
improve on Descartes’s own inferior sketches.
22
As John Morris (1969) argues, Descartes changes his mind over the course of his
career about the need for an intellectual memory. What is constant, however, is his
claim that there is at least a corporeal, purely physical memory.
23
For the earliest statement of it, see Regulae XII, AT X 416/CSM I 43.
24
AT V 150/CSMK III 336–7.
 THE EARLY DESCARTES

how the argument works. It is because there is no affinitas between

‘K-I-N-G’ and the supreme power that the intellectual memory must
be invoked. The missing connection would be supplied by resem-
blance. The implication is clear: where there is a resemblance, the
corporeal memory is sufficient.
Why, though, does the corporeal ‘idea’ merit its name? Why not
just call it a brain image? In the context of the Meditations, I shall
argue that a Cartesian idea, by its nature, allows the subject having it
to think through it out to an object in the world (if such there be).
That is, a Cartesian idea is a representation in the internalist sense
stipulated above. It makes sense, then, that at this stage brain images
count as ideas. For they can be exploited by both humans and animals
alike in picturing their environments.
The status of brain images as ideas provides another line of argu-
ment for the conclusion that corporeal representation is resemblance.
We know that it is supposed to play the role of an internalist rep-
resentation. What else besides pictorial resemblance could allow it to
play this role? No other candidates have emerged. And to Descartes’s
inheritors, such as Robert Desgabets, it is simply obvious that the
brain image is supposed to resemble the object in the world. Desga-
bets has nothing but contempt for Simon Foucher, who seemingly
cannot grasp this simple point: there is ‘an image formed in the brain,
in which one perceives some real and “entitative” resemblance with
the object, as they say in the Schools.’25
And yet, if I am right, we face another question: how can the
resemblance relation obtain, if Descartes, at this stage, thinks that
there are corporeal ideas of light, sound, color, and the rest? In some
way or other, these have to be understood as features of the image
traced on the pineal gland. But of course Descartes does not want to say
that the sensible qualities can exist in, or be perceived by, a soulless
substance.
Given that the image has to be ‘etched’ on the pineal gland by the
animal spirits, we are in essence dealing with an inverted bas-relief

25
1675, 94.
 THE EARLY DESCARTES 

representation.26 It is important to keep this in mind, since it removes

an obvious difficulty, namely, that of representing shape in a purely
physical way. As Malebranche was to emphasize, our visual mental
pictures of objects have to include color in order to represent the
boundaries of bodies. An uncolored two-dimensional image, it seems,
is no image at all. By contrast, a bas-relief representation can mark
those boundaries by differences in depth. That is precisely because,
unlike a mental image, it is itself in space.
Working with these meager materials, Descartes suggests, in Reg-
ula XII, a way of getting a kind of ersatz color into the brain image.

[W]hatever you may suppose colour to be, you will not deny that it is
extended and consequently has shape. So what troublesome consequences
could there be if . . . we simply make an abstraction, setting aside every feature
of colour apart from its possessing the character of shape, and conceive of the
differences between white, blue, red, etc., as being like the difference between
the following figures or similar ones? [Here Descartes gives three figures. The
first is five vertical lines; the second is a box composed of sixteen squares;
the third is the same box, only with each square bisected on the diagonal.]
The same can be said about everything perceivable by the senses, since it is
certain that the infinite multiplicity of figures is sufficient for the expression
[exprimendis]27 of all the differences in perceptible things.’28

The role of color is to mark off the boundaries of objects, revealing

their shapes. What Descartes sees here is that, for these mechanical
purposes, one does not need what we experience as color: all we need is
some feature that can be found in our etching that does the same job.
But it is not merely single objects that are corporeally represented.
Not only does he speak of corporeal images being traced on the pineal
gland and stored in the memory, Descartes also claims to be able to

26
In bas-relief, the surfaces of objects are higher than the background. In Descar-
tes’s case, the boundaries of objects are etched in the surface of the gland.
27
As Wolf-Devine (1993, 22) notes, exprimendis might mean either ‘expression’ or
‘explanation.’
28
AT X 413/CSM I 40–1. Keep in mind that Descartes here is speaking, not of
color as it is experienced by minds, but as it figures in the external senses and the
pineal gland.
 THE EARLY DESCARTES

explain how these ideas ‘cause movement of all the bodily parts.’29 To
play that role, corporeal ideas need some way of representing dis-
tance. What makes Descartes’s treatment of the corporeal represen-
tation of distance awkward is that he treats it as an idea, even though
it is clearly not another image among many. Descartes tells us that

[I]n order to understand what the idea of the distance of objects consists in,
you [should] assume that as the [pineal] gland’s position changes, the closer
the points on its surface are to the centre of the brain, the more distant are
the places corresponding to them, and that the further the points are from
it the closer the corresponding places are. Here, for example, we assume that
if b [a point on the top of the pineal gland] were pulled further back, it would
correspond to a place more distant than B [the center of the distal object],
and if it were made to lean further forward it would correspond to a place
that was closer.30

What has to happen, evidently, for the machine to ‘judge’ distance is

for it to keep track, in memory, of the relative positions of the points
on the gland’s surface. The idea of distance, then, is not another image
alongside all the others; instead, it is a feature that arises because the
ideas change. If we imagine our machine sitting utterly stationary in
front of an unchanging scene, it would have no way of becoming
aware of distance, although it could still figure out that one object
began and ended to the left or right of some other object.
To sum up: when we walk through a forest thinking of higher
things, our brains and eyes are performing in just the same way as
those of the squirrels and bears around us. A tiny little picture,
collating the input of all the senses, is constantly being traced and
revised on our pineal glands. What happens, though, when a nearby
twig snaps, and the mind begins to attend to its environment?

2.3 From Brains to Minds

I shall argue that the Descartes of L’Homme holds that the attentive
mind is immediately aware of what happens in the brain, so long as

29 30
AT XI 174/G 146. AT XI 183/G 155.
 THE EARLY DESCARTES 

the two are united. This reading is controversial, since it seems to

violate the deliverances of introspection.31 But the textual evidence
stubbornly resists reading away. What is more, the view fits nicely
with Descartes’s ambitious early project of explaining the overlap
region by positing a single mechanical process. If the brain can
represent the world mechanically, then the obvious way for the
mind to be aware of the world is for it to be aware of the brain.
Now, it is true that attentive human perception is overlaid with the
activity of natural geometry, as we shall see. Nevertheless, the early
view takes the brain image as the fundamental explanans in both
attentive and inattentive perception.
As time goes on, however, Descartes seems to lose confidence in his
account of mechanical perception, or at least in its details. Of course
he continues to assert that animals are mere automata. Still, after
L’Homme, one never again finds a detailed attempt to work out
precisely how mechanical perception operates. In the end, as I’ll
argue in the coming chapters, Descartes offers an equivocal account
of perception: although both humans and animals can be said to see
the apple or tiger in front of them, they are doing it in totally different
ways. The overlap thesis is quietly and tacitly withdrawn. Corres-
ponding to this shift is a diminished role for the brain image. While it
is still there, still resembles the object in the world to some degree, and
still has an important causal role to play, that role becomes purely
causal. That final position is still quite a distance off. I now hope to
make clear just how great that distance is.

31
See, e.g., Alison Simmons (2003) and Lili Alanen (2003). Simmons claims that
Descartes cannot mean what he says, since the role of the brain is purely causal, not
cognitive (2003, 561 f.). For her part, Alanen says that ‘the cerebral images are not
actually perceived at all.’ She takes talk of the mind ‘directing itself toward some part
of the brain’ as merely metaphorical and insists that ‘a literal reading of the metaphor
makes no sense’ (2003, 147). On the other side, Tom Vinci (2008, 477 f.), Peter
Machamer and J.E.J. McGuire (2009, 185), and Andrew Chignell (2009) argue, in
different ways, that Descartes does mean to say that the mind is aware of the brain in
perception and imagination.
 THE EARLY DESCARTES

2.3.1 Sensations and the analogy with language

Throughout his career, Descartes uses the metaphor of language to
capture the relation between some mental and brain states. Just as a
word can make us think of something it in no way resembles, so a
brain state can prompt us to think of something to which it bears no
resemblance. Encouraged by these remarks, many commentators have
supposed that there is no resemblance between any mental state or
object and the brain state with which it is associated, or between either
of those and the object in the world.32 Moreover, they have equally
supposed that the connection between brain state and sensation is
automatic, or ‘instituted by nature,’ requiring no effort on the part of
the mind. I shall argue that in fact Descartes’s claims are considerably
stranger than these, at least at this stage of his career. In this section,
I focus on sensible qualities, while the next section turns to the geomet-
rical ones.
One way to trace the path from brains to minds is to ask just what
happens when God installs a soul in the machine envisioned by
L’Homme. God ‘will place its principal seat in the brain and will
make its nature such that the soul will have different sensations
(‘sentimens’) depending on the different ways in which the nerves
open the entrances to the pores in the internal surface of the brain.’33
Note what Descartes does not say. He does not speak of ideas, or of
mental states generally, but only of sensations. When we turn to the
other part of Le Monde, La Lumiere, we find that his linguistic
analogy preserves this: what the brain states are linked with is not
any old mental state or object but sentimens in particular. In denying
resemblance through the linguistic analogy, Descartes is clearly con-
cerned to reject the scholastic view that the sensible qualities that we

32
For example, Desmond Clarke’s otherwise very useful (2003) seems to me to run
together Descartes’s denial of resemblance between the sensation of color and anything
in the corporeal world on one hand, and a denial of resemblance between brain images
and the objects in the world that help to cause them on the other. (See Clarke 2003, 55.)
On my view, the latter resemblance must hold in order for Descartes’s account of
corporeal memory (not to mention seemingly goal-directed animal behavior) to work.
33
AT XI 143/G 119.
 THE EARLY DESCARTES 

experience either just are or resemble the sensible qualities that exist
in the world. For Descartes insists that it is the nature of the mind,
rather than the body, that explains the sensations it undergoes.
Descartes begins La Lumiere by trying to convince the reader that
sensations need not resemble their causes:

Words, as you well know, bear no resemblance to the things they signify, and
yet they make us think of these things, frequently even without our paying
attention to the sound of the words or to their syllables. . . . Now if words,
which signify nothing except by human convention, suffice to make us think
of things to which they bear no resemblance, then why could nature not
also have established some sign which would make us have the sensation of
light, even if the sign contained nothing in itself which is similar to this
sensation?34

Descartes next considers an objection: someone might say that words

and tears do not really cause us to think of what they signify; instead,
‘it is our mind which, recollecting what the words and the counten-
ance signify, represents their meaning to us at the same time.’35 An
obvious reply would be that this is where the analogy breaks down: all
Descartes needs, after all, is to deny the resemblance relation between
brain states and sensations.
This is exactly the reply he does not give. Instead, Descartes
concedes that ‘it is our mind which represents to us the idea of light
each time our eye is affected by the action which signifies it.’36 In later
works, he will reply in a very different way. But we are dealing solely
with the views of the early Descartes, and we must do our best to
understand them, in all their oddity.
Two features of Descartes’s reply should bring us up short. Notice
that the mind represents to us (‘nous represente’) the sensation of light.37
Again, Descartes does not want to appeal to purely bodily states of

34 35
Le Monde AT XI 4/CSM I 81. AT XI 4/CSM I 81.
36
AT XI 4/CSM I 81.
37
In La Lumiere, Descartes seems to count sentimens among ideas. In L’Homme,
by contrast, he reserves idée for corporeal images sketched on the pineal gland, and
sentimens for distinctively mental states.
 THE EARLY DESCARTES

affairs to explain the nature of sensations. Since their connection

with brain states is as arbitrary as language, their nature cannot be
accounted for by the brain state alone. Descartes continues to claim
this throughout his career. In the Notae (1647), he says that the ideas
of pain, color, sound, and the like must be innate if ‘on the occasion of
certain corporeal motions, our mind is to be capable of representing
[‘exhibere’] them to itself.’38 Mere motion might account for a sensa-
tion’s occurrence, but its nature must be conjured from the mind’s
own resources. However odd it may sound, Descartes is committed to
the claim that the mind represents a sensation to itself.
The claim sounds considerably less odd if we distinguish this sense
of representation from the more common sense in which we’ve been
using the term. To say that the mind represents a sensation to itself is,
as ‘exhibere’ indicates, just to say that the mind presents itself with
that sensation, rather than passively receiving it from outside. Here
again the language analogy can be useful. When one hears ‘elephant,’
those three syllables do not carry with them the thought of an
elephant. That’s up to the mind experiencing those sounds to do. So
just as the mind has to present itself with the meaning of a word, so it
presents itself with a sensation.
But there is a further—and much more important—oddity in the
offing. When the mind presents itself with a sensation, it does so on
the basis of its awareness of a brain state. To say the least, this is an
unusual state of affairs. The brain state is the signifier, and the
sensation or idea is the thing signified. In the case of words and facial
expressions, the sign is perceived, and the mind must infer to what is
not (a thought or an emotion). On its face, the case of sensation
reverses this: the brain state (the sign) is not perceived, while the
sensation (the signified) is. This has led some commentators to
suggest that Descartes is introducing a novel notion of signification,
or perhaps picking up a rather obscure scholastic one.39 No such

38
AT VIIIB 359/CSM I 304. In rendering ‘exhibere’ as ‘representing,’ the trans-
lators of CSM seem to have elected to keep the Latin of Descartes’s Notae in line with
the French of La Lumiere, which, as we’ve just seen, uses ‘representer.’
39
For an interesting exchange on this issue, see Peter Slezak (2000) and John
Yolton (2000).
 THE EARLY DESCARTES 

drastic interpretive measures are called for. Descartes can be read

literally: he thinks the sign is perceived. So he really does think that
the mind is immediately aware of the brain.

2.3.2 Turning toward the brain

In attentive perception, the mind must inspect or become aware
of the image sketched on the pineal gland. Doing so is the mind’s
primary means of becoming aware of the geometrical qualities pos-
sessed by the bodies in its immediate environment. We have already
seen the evidence for the claim that the brain image pictorially
resembles the object in the world and indeed is capable of represent-
ing the relative positions and distances of objects. But as we’ve also
seen, reading even the early Descartes as claiming that the mind is
aware of the brain is highly controversial.
Let’s begin with the textual evidence. In Regula XII, when Descartes
moves from the corporeal to the mental, he tells us that the spiritual
power can be said to see, touch, and so on, when it applies ‘itself with
the imagination to the common sense’ (applicet se cum imaginatione
ad sensum communem).40 The meaning of ‘applicet’ is not obvious.
But the notion that the soul turns to the body reappears throughout
Descartes’s writings. In the sixth Meditation, for example, Descartes
says when the mind imagines, it ‘looks at something in the body
which conforms to an idea understood by the mind or perceived
by the senses.’41 The fifth Replies clarifies: that something toward
which the mind turns is a corporeal species, a speciem corpoream.42
And as Descartes reportedly tells Burman six years later, ‘[w]hen
external objects act on my senses, they print on them an idea, or
rather a figure, of themselves; and when the mind attends to these
images imprinted on the gland in this way, it is said to have sensory
perception.’43
I have warned already of assuming agreement across texts. So I am
not appealing to these later texts to support my reading of L’Homme.

40
AT X 416/CSM I 42. Note that Descartes here seems to be conceiving of the
imagination as a faculty of the immaterial soul, not as a corporeal function.
41 42 43
AT VII 73/CSM II 51. AT VII 389. AT V 162/CSMK III 344.
 THE EARLY DESCARTES

Happily, some of the most direct evidence comes from L’Homme itself.
Since my claim is controversial, permit me a slightly longer quotation:

Note that, by these figures, I don’t just mean the things that represent in
some way the position of lines and surfaces of objects, but also all those that
(according to what I said above) can give occasion to the soul to sense (sentir)
motion, size, distance, colors, sounds, smells, and other such qualities; and
even those that can make it sense tickling, pain, hunger, thirst, joy, sadness,
and other such passions . . . [It is only] those [figures] that are traced by the
spirits in the surface of gland H [the pineal gland], which is the seat of the
imagination and of the common sense, which should be taken for ideas, in
other words, for the forms or images that the rational soul immediately
considers [‘les formes ou images que l’ame raisonnable considerera imme-
diatement’], when, being united to this machine, it imagines or senses some
object.44

Note that Descartes explicitly says that corporeal figures ‘represent

in some way’ (representent en quelque sorte) lines and surfaces of
objects. In the second part of the passage, he claims that only some of
these figures merit the name ‘idea.’ Unsurprisingly, they are the
figures etched in the pineal gland, the seat of the common sense.
What happens when our fictional humanoid machine has a soul ins-
talled in it? How will this soul be able to imagine or sense? By immediate-
ment considering the image on the pineal gland. This cannot be a mere
slip on Descartes’s part, for it is the point on which he would want to
be most clear. Only here is he explaining how thinking beings like us
make use of the brain image. If it were not for this immediate consider-
ation, the image would play no role at all in making us sense or imagine.
When one reads the works of Descartes’s immediate followers, one
finds near unanimity on these points. For however much these
philosophers depart from their master on other issues, almost to a
man they endorse the claim that imagination and sensory experience
requires the mind to turn toward an image in the brain, which in turn
really and literally resembles the object itself.

44
AT XI 176–7; emphasis mine.
 THE EARLY DESCARTES 

In a Paris bookstall in 1664, Nicolas Malebranche picks up a copy

of the Traité de L’Homme and finds what he read there so exciting he
nearly has a heart attack.45 One of its claims evidently sticks with him,
and he repeats it, in his own fashion, in his Search after Truth:

[T]he soul immediately resides in that part of the brain to which all sense
organs lead. . . . When I say that it resides there, I mean only that it is aware of
all the changes taking place there in relation to the objects that cause them, or
customarily cause them, and that it perceives what happens outside this part
only through the agency of the fibers ending there . . . 46

The claim can also be found in slightly more orthodox Cartesians

like Robert Desgabets.47 But not all of Descartes’s first readers were
happy to attribute this picture to him. In fact, the drive to obscure
Descartes’s early view is exhibited by perhaps its earliest published
commentary, Louis de la Forge’s Traitté de L’Esprit de L’Homme.48
His take on the text is worth considering, because he engages in some
truly energetic ‘reading away.’
First, he points out what he takes to be a common, indeed, nearly
universal, error: thinking that the mind looks at corporeal images
when it senses or imagines.49 The mistake is a big one, but common

45
See Lennon and Olscamp’s introduction to the Search in their translation
(Malebranche 1997a, viii), for the details of the story.
46
SAT I.x/LO 50. Nor should anyone dismiss this claim as an isolated one,
confined to the early pages of the Search; for much later, in SAT III.ii.3: 224, discussed
below, Malebranche describes how imagining a geometrical figure requires us to
sketch an image of it on the brain (though we do so according to the rule provided
by a non-sensory idea). Given the rest of Malebranche’s view, this is puzzling, as his
earliest readers notice. John Locke, for example, remarks that Malebranche ‘allows the
soul power to trace images on the brain, and perceive them. This, to me, is matter of
new perplexity in his hypothesis; for if the soul be so united to the brain as to trace
images on it, and perceive them, I do not see how this consists with what he says a
little before in the first chapter, viz. “that certainly material things cannot be united to
our souls after a manner necessary to its perceiving them” ’ (‘Examination’ section 15).
Locke’s reference is to SAT III.i/LO 217.
47
See Desgabets (1675, 94).
48
First published in 1666, two years after the publication of Descartes’s Traité.
49
The target la Forge singles out for abuse is the ‘celebrated doctor’ Fracastor,
a.k.a. Girolamo Fracastoro, now remembered only for having given syphilis its name,
in a three-part epic poem about the disease.
 THE EARLY DESCARTES

enough: ‘there is hardly anyone who does not imagine that his soul
is like a little Angel lodged in his brain, where it contemplates the
images that come to it by the senses.’50 Naturally, la Forge is well
aware of the textual evidence in favor of attributing precisely that
mistake to the Descartes of L’Homme. And yet:

These expressions, no matter how strong, should not be construed as if he

meant that the images are received in the proper substance of the mind, or
that the ideas which he contemplates immediately and these images are the
same thing. Instead, he applies ‘idea’ to the corporeal images that are traced
on the gland and says that the soul contemplates them immediately, because
it is to these images alone that the ideas of the soul (which represent objects
to us) are immediately united. And because of the tight bond that there is
between mind and body, one can in some sense say that they [corporeal
images] modify the mind itself, because together mind and body make up
one thing.51

It is not clear just what la Forge’s maneuver amounts to. Earlier, he

has told us, largely on the basis of evidence from the Dioptrique rather
than L’Homme, that the ‘corporeal species’ are not images at all but
changes in the brain that ultimately occasion or ‘oblige’ the soul to
think of this or that.52 Here he reverts to speaking in terms of images
traced on the gland, which I find hard to construe as changes. Trying
to explain how corporeal images are united to the idea the soul
summons on their occasion, he resorts to fudging: since they go
together, and since one is bodily and the other mental, the mind-
body union allows us ‘in some sense’ to identify them. Not only is this
a very creative reading, it seems to make a hash of la Forge’s own
interpretation, according to which it is the brain motions, not brain
images, that count as the ‘corporeal species.’ Even that interpretation
is suspect, since la Forge is reading the Dioptrique account into
L’Homme, where it doesn’t belong. Below, I argue that the two
works diverge on important points. But even if la Forge were right,
and the brain image were not an image at all but a series of changes, it

50 51
La Forge (1666, 100). La Forge (1666, 114–15).
52
La Forge (1666, 111–12).
 THE EARLY DESCARTES 

would be hard to see how in any sense one could identify those
changes with an idea in an immaterial substance.
Perhaps the best reason to take Descartes at his word is the
scholastic context. It is worth noting that, as far as I know, Descartes’s
scholastic critics such as Caterus (author of the first set of Objections)
never take issue with him on this point. In fact, in the context of the
Aristotelian picture, it is a rather unremarkable claim. Turning to that
context can help illuminate just what is novel about Descartes’s view.
On the scholastic view, perception is the act of a corporeal organ,
while intellection is not. As a result, since like knows like, the proper
objects of perception are singulars, while the intellect’s proper objects
are universals. This creates a puzzle. For there is seemingly no way for
the intellect to grasp singulars. The solution is for the intellect to ‘turn
to’ the phantasms. Aquinas tells us that a phantasm is a corporeal
image that ‘is the likeness of an individual thing.’53 As a corporeal
thing itself, phantasms can only represent individuals, not universals;
what is more, they do so by resemblance. Nor are they purely visual,
since a phantasm is ‘a passion of the common sense.’54 The intellect
needs to ‘turn to’ the phantasms both in order to cognize particulars
and in order to imagine.
Descartes’s Latin closely tracks that of Aquinas on this point. In the
Fifth Replies, for example, Descartes says that the mind imagines ‘by
directing itself towards a corporeal species,’ convertendo se ad speciem
corpoream;55 in the Summa, Aquinas says that the intellect grasps
particular things ‘by directing itself towards phantasms,’ convertendo
se ad phantasmata.56 In the sixth Meditation, Descartes says that
the mind se convertat ad corpus;57 in the Summa, Aquinas says that
the agent intellect understands particulars ex conversione . . . supra

53
ST I q.84 art. 7, reply to objection 3.
54
This is Aquinas’s quotation of Aristotle in the former’s Sentencia Libri De
Memoria et Reminiscia, quoted in Pasnau (2002, 280).
55
AT VII 389.
56
ST I q.86 art.1. This phrase appears no fewer than thirteen times in the so-called
Treatise on Man. I am not, of course, claiming that Descartes and Aquinas share a
theory of perception.
57
AT VII 73.
 THE EARLY DESCARTES

phantasmata.58 So the notion that the mind might have to ‘turn

towards’ something corporeal would not have surprised Descartes’s
scholastic readers.
What does surprise them, of course, is his dualism. Given this
innovation, Descartes simply does not encounter the same puzzle
that led Aquinas to his doctrine of phantasmata. Or rather, he encoun-
ters his own exaggerated version of it: once one makes the mind a
substance, it becomes all the harder to see how the mind could be in
cognitive contact with the individual objects that surround the body it
is tied to. In appealing to a brain image, Descartes is making novel use
of entirely traditional tools. Instead of the agent intellect doing the
turning, we have the immaterial mind. The difference between the two
is a product of Descartes’s dualism and his rejection of hylomorphism.
Given those two massive changes, Descartes’s talk of the mind se
convertat ad corpus is as traditional as it can be.59
There is a further complication that proves illuminating. In both
Aristotle and Aquinas, the role of phantasmata is not entirely clear.
Some commentators on Aquinas, such as Anthony Kenny, claim
that phantasms are involved in every act of sensing.60 Part of the
reasoning here is that a phantasm is defined by Aristotle as ‘a passion
of the common sense.’61 Since phantasms seem necessary for the
transmission of sense perception to the common sense, it is hard to
see how perception could be accomplished without them.62 Other
texts suggest that phantasms are involved in memory, imagination,
and the abstraction of intelligible species, but not in perception.63

58
ST I q.85 art. 1 reply to objection 3. The Latin is from the Leonine edition,
available online.
59
It is intriguing to note that Averroes appeals to the intellect’s connection with
the phantasms to explain the nature of the intellect’s union with the body (see
Aquinas, ST 1 q.76 and Pasnau 2002, 76). Perhaps the key to the Cartesian mind-
body union is the activity of conversio. Might Descartes not have this possibility in the
back of his mind?
60 61
Kenny (1993, 37–8). On Memory 450a12.
62
For the claim that the primary role of phantasms in Aristotle’s theory of
perception is precisely this transmission to the sensus communis, see Everson
(1997), chapters 4 and 5, cited by Dorothea Frede in her very helpful (2001).
63
See Pasnau (2002, 280 f.).
 THE EARLY DESCARTES 

After all, Aquinas mentions the need to turn toward the body when it
comes time for the intellect to think of singulars and abstract intel-
ligible species; it is not obvious that sense perception necessarily
involves either of these intellectual acts. The corresponding claim in
Aristotle, then, is that aisthêsis and phantasia are distinct and have
different roles to play.
The fact that Descartes focuses on imagination at the start of the
sixth Meditation may well reflect the former interpretation, which
does not deploy phantasms to explain sense perception. It is uncon-
troversial that the Aristotelian tradition at least invokes phantasms
and the conversio of the intellect to the body where imagination is
concerned. As we have seen, Descartes wants to go further. On his
view, the very same procedure is necessary to explain sensory per-
ception. The difference between sense perception and imagination,
for him, lies in the origin of the brain image, not in the need to inspect
it. Assuming his readers are already on board with his picture of the
imagination, he chooses to begin there. Whether Aristotle and Aqui-
nas hold that phantasms have a role to play in sense perception itself
is beside the point. At a minimum, Descartes is adopting their story
about imagination and leveraging it into a story about sense percep-
tion. And even if Descartes’s corporeal images are playing a role that
Aquinas splits between sensible species (perception) and phantasms
(imagination and memory), these Thomistic entities differ only in
etiology, not in kind.
I conclude, then, that at least in L’Homme, Descartes means what
he says: the mind is immediately aware of the brain. Below, we shall
explore further evidence for what I shall argue is a constant in
Descartes’s thought. It would be best, though, if there were some
way to take the sting out of attributing this position to Descartes, even
in his earliest work. For the view is at the furthest limits of the
plausible.
First, one might object that the resemblance of the brain image
to the extra-bodily object makes no sense in a dualist framework.
Why should the mind need to look at a brain image? If it can look at
anything at all without an idea as an intermediary, what stops it from
 THE EARLY DESCARTES

gazing on objects directly? And from what perspective would it ‘see’

this image? If the orientation of the image on the retina is a problem,
why not just install the soul upside down and backwards, to get the
orientation to come out right?64
Against this barrage of objections, Descartes might appeal to the
unique status of the pineal gland as the organ of the common sense.
That the soul should inspect the brain image is a result of God’s
simplicity in designing the world. God has already had to create
brains for other animals that allow for inattentive (because unen-
souled) perception; why should he not use the same materials for
attentive perception? And now that he’s gone to the trouble of
ensuring that the image comes out with the proper orientation in
the brain, the simplest course to allow for attentive perception is to
make the soul experience the pineal gland.
Descartes should also insist that vision is the wrong metaphor. Since
the brain image is etched in the gland, there is no reason he couldn’t
appeal to the mind’s ability of proprioception. The better analogy then
would be with touch, as one might be able to figure out, even blind-
folded, what image had recently been tattooed on one’s flesh.
Even though these objections can be parried, the fact remains that
Descartes’s view seems phenomenologically implausible. A philoso-
pher who insists she is immediately aware of her own brain, or the
backs of her eyes, is, in any era, liable to be suspected of lying, or
reporting a very strange medical condition.
A tempting thought presents itself here. Aquinas famously claims
that the intelligible species is not that which is understood but that by
which understanding happens. In just the same way, the sensible
species is not that which is sensed but that by which sensing hap-
pens.65 Although Aquinas does not say so explicitly in this part of the
Summa, it is natural to suppose that this doctrine extends to the
phantasms. When the intellect turns toward the corporeal images that
allow it to think, however derivatively, of particulars, it is natural to
suppose that those images are, like the intelligible species itself, only

64 65
I owe these objections to Ben Jantzen. ST I q.85 art 2, discussed above.
 THE EARLY DESCARTES 

that by which thought of an individual happens, not themselves the

objects of that thought. Might not Descartes make a parallel move
with his corporeal images, and relegate them to part of the means by
which sensing, imagining, and corporeal memory happen, rather
than that which is sensed?
There are two issues with this move. The first concerns how we
read Aquinas on this point. As Robert Pasnau has made clear, it is far
from obvious just what Aquinas means to say here.66 The view
Aquinas attacks is not one that takes the intelligible species to inter-
pose itself between the mind and the world. Instead, Aquinas clearly
takes the opposition to hold that only the species is perceived. This is
clear in his arguments. For example, he claims that the things we
understand are the objects of science, and so ‘if what we understand is
merely the intelligible species in the soul, it would follow that every
science would be concerned, not with things outside the soul, but only
with the intelligible species within the soul.’67 What Aquinas rejects,
then, is only the claim that understanding and perception end at the
(intelligible or sensible) species. That leaves open, of course, that we
think through those species out to the objects in the world.
A second issue is more telling, I think. This move would in effect
rob the corporeal images of their explanatory power. If they never
figured as the objects of perception, then they would do nothing to
explain why perception has the content it does. In other words, their
role in the causal explanation of experience requires them to appear
as at least the immediate objects of perception.
Another way to make Descartes’s view more plausible is to follow
Andrew Chignell in distinguishing between intensional and exten-
sional awareness. We can say that the mind is extensionally aware of
brain states, without saying that they are transparently available to, or
present in, consciousness.68
For my part, I find it a bit difficult to grasp this picture. Here’s one
way to construe the difference. One might see the President on the

66 67
See Pasnau (1997, 195 f.). ST I q.85 art 2, my emphasis.
68
See Chignell (2009, 11).
 THE EARLY DESCARTES

street but not see him as President. One is ‘intensionally’ aware of a

human being, or a patch of color, or whatever you like, but ‘exten-
sionally’ aware of the President, since that human being is, in fact, the
President. The trouble comes when we try to carry the analogy
through. Our Presidential example relies on the identity between
the President and the human being who is represented in thought.
But nothing like this is the case when it comes to the brain: the brain
image is not the thing in the world under some other description.
Perhaps this is not quite what Chignell has in mind. Let’s consider
a different analogy, one between brain images and mirrors.69 When
looking at a mirror, one can equally well be said to see the mirror and
the objects reflected in it. And of course there are cases, as in the use
of periscopes in submarines, where one is unable to see the objects
directly but has to ‘turn toward’ the mirror to do so. We can imagine
other cases where the subject is not aware she is looking at a mirror;
presumably that’s the analogue of the benighted state of la Forge and
all those who reject this picture of perception.
This analogy takes us much of the way toward making Descartes’s
(and the Aristotelians’) conversio doctrine intelligible. But it cannot
take us all the way, for there is one important disanalogy: the object
and the image in the mirror will both at least seem to have sensible
qualities; by contrast, the brain image is experienced through pro-
prioception and does not seem to have such qualities. While this is
pure speculation, I think Descartes might tell a story roughly parallel
to (though hardly identical with) the one he will go on to tell in the
Principles.70 In childhood, one is indeed aware only of the brain
states. But one notices that they are accompanied by pleasure and
pain, and all the other sensations. Soon one begins to associate
features of the brain image with these sensations. Before long one is
projecting one’s experiences out on to the world. Such projection has
become so ingrained that the infant’s experience is inaccessible, if not
inconceivable, to the adult.

69
I owe this analogy to an anonymous referee.
70
See Principles I.71, discussed in the next chapter.
 THE EARLY DESCARTES 

Even if I fail to make Descartes’s view plausible to a contemporary

reader, the view is intelligible in its context. Still, I think it is less silly
than it seems. It is worth remembering Theodor Erisman and Ivo
Kohler’s inverting goggles experiments from 1950. Given enough
time to form associations, the nature of the initial stimulus seems
not to matter much.71 With increased experience, subjects report no
difference between their original perceptions and those had with the
goggles. My point is only that systematic and profound perceptual
adaptation suggests that it would be unwise to dismiss Descartes’s
early view as too implausible even to be his early view. To my mind,
its real troubles lie elsewhere, as we shall see soon enough.

2.4 An Unnatural Geometry

Although L’Homme does not have much to say beyond the linguistic
analogy about how the soul summons the appropriate sensations of
color or other proper sensibles or projects them on to bodies,72 it does
offer an intriguing story about the common sensibles: ‘position,
shape, distance, size, and other similar qualities which are not related
to one sense.’73 So far, we have been focused on material from relatively
late in the work, where Descartes explains how the image is traced on
the brain. Once installed there, the soul will contemplate that image.
Given that the image itself is able to represent all of the common
sensibles, one would think Descartes would leave it at that.
Earlier in the book, however, he has already given a very different
account of how a soul comes to judge the common sensibles. As he
will in the Dioptrique, Descartes adduces his analogy of the blind man
using two sticks to determine the location of a single object before
him. The blind man knows the positions of his two hands and the
angles at which he holds the sticks. When they strike an object, he can

71
Film of these experiments is available on YouTube.
72
Pace what Descartes says: he claims that he has ‘already dealt with’ ‘the details of
what makes it possible for the soul to conceive of all the differences in colour’ (AT XI
158/G 131), but as Gaukroger points out, he has not.
73
AT XI 159/G 131.
 THE EARLY DESCARTES

tell, ‘as if by a natural geometry,’ where the object is.74 And in

precisely the same way, the soul can use the position of its eyes to
tell where a distal object is.
There are other means of judging distance as well, such as a change
in the humors of the eye or the arrangement of colors. We’ll explore
them in greater detail when we come to the Dioptrique. For now, we
should deal with the obvious problems this natural geometry raises
for the early Descartes’s view.
Not surprisingly, commentators have found the notion of uncon-
scious judgments in tension with Descartes’s official view of the mind
as unfailingly aware of its own activities.75 But the inclusion of
unconscious judgments is not (or not just) a mistake on Descartes’s
part; it is a direct result of his standing at the end of the Baconian
tradition. Despite his hostility to many of its planks, Descartes pre-
serves one of its key features: the notion of successive inferences, on
the basis of more fundamental data. In those accounts, it is often
extremely hard to tell what is metaphor and what isn’t: as the image
travels from the eye to the animal spirits in the ventricles and becomes
richer in intentional content, is the subject performing inferences or
judgments, or are these processes only inference-like, instantiated as
they are in purely corporeal stuff?76 While Descartes rejects much of
this picture, he retains the notion of a kind of inference or judgment
happening below the level of conscious experience.

74
AT XI 160/G 133.
75
This seems to follow from Descartes’s definition of thought: ‘By the term
“thought,” I understand everything which we are aware of as happening within us,
in so far as we have awareness of it’ (Principles I.9, AT VIIIA 7/CSM I 195). Margaret
Wilson (1978), among others, has questioned Descartes’s commitment to this ‘trans-
parency of the mental.’
76
On this issue in Alhacen and his followers, see esp. Smith (2004, 188 f.) and
Abdelhamid I. Sabra (1978). Kepler is surely poking fun at this unclarity in the
Baconian synthesis when he writes in judicial terms: ‘How this image or picture [at the
back of the eye] is joined together with the visual spirits that reside in the retina and in
the nerve, and whether it is arraigned within by the spirits into the caverns of the
cerebrum to the tribunal of the soul or of the visual faculty; whether the visual faculty,
like a magistrate given by the soul, descending from the headquarters of the cerebrum
outside to the visual nerve itself and the retina, as to lower courts, might go forth to meet
this image—this, I say, I leave to the natural philosophers to argue about’ (2000, 180).
 THE EARLY DESCARTES 

Writers in the Baconian tradition do not draw a sharp and clear

line between the physiological and the mental.77 As a result, there is
sometimes no informative answer to questions about the ontological or
psychological status of the transitions from one stage to another. Is it a
mental inference, accessible to introspection, or an automatic function
happening beneath the layer of consciousness, or a purely physiological
process? If we have to excuse earlier writers for not answering such
questions, it is correspondingly harder to excuse Descartes, who draws
the sharpest possible distinction between mental and physical.
For our purposes, a second problem is more salient. Descartes is
here describing how the soul judges distance, and this seems to be an
entirely different process from that by which the machine ‘judges’
distance. His ‘natural geometry’ is thus unnatural in the sense that it
is not deployed by any non-human animals, or indeed by human
bodies when their minds are ‘elsewhere.’
But even here there is an illuminating connection with the Baconian
tradition. In fact, the tension in Descartes’s work between purely
corporeal and rational means of judging distance is a direct descendant
of a long-standing and unresolved conflict within that tradition.
Roughly four centuries earlier, Roger Bacon offers a similarly two-
layered view, in his Opus Majus. Bacon posits the common sense and
a pair of further faculties, the ‘estimative’ and ‘cognitive’ faculties, to
account for perception of the common sensibles, including distance.78
This is how sense perceives distance; but we also perceive distance
through reasoning. In particular, we can use the positions of inter-
vening objects to work out an object’s distance from us; this is why
it’s difficult to see just how high the clouds are above us.79 As Bacon
recognizes, this suggests that the rational soul is responsible for this
second means of judging distance. But to Bacon, the conclusion is
intolerable: ‘in no way can the fact be hidden that animals perceive
the distance of objects, and motion, and rest.’80 Bacon’s resolution of

77
As Hatfield and Epstein (1979) point out.
78 79
Bacon 1928, vol.2, 423 f. Bacon 1928, vol.2, 523.
80
Bacon 1928, vol.2, 544.
 THE EARLY DESCARTES

this tension is rather unsatisfying. He chooses to deny reasoning per

se to animals, since that is the work of the rational soul; instead, we
should speak of ‘a certain gathering together of several facts into one
in consequence of a natural purpose and instinct, the several facts
resembling premises and the resultant one a conclusion . . . ’81 What
animals do is similar to reasoning, and ‘for this reason authors on
Perspective call it argument and syllogism.’82 Note that this solution
to the tension, to the degree that it is one, is not available to Descartes.
For it still requires attributing an awareness of facts to the animal, and
an ability to think, if not ‘reason,’ about them.
Faced with this predicament, Stephen Gaukroger has taken the
bold step of asserting that natural geometry is not the exclusive
property of minds.83 If animals need to judge distance, and natural
geometry is how it’s done, then animals must be using natural
geometry. It is certainly true that the Cartesian soul does not need
to perform these geometrical calculations consciously; if it did, pre-
sumably many more people would be bumping into things. That
seems to fall short, however, of allowing that mere machines can
perform these calculations as well.
If we do not wish to tough it out and attribute natural geometry
to mere machines, we are still left with some interpretive options. The
best is to say that automata and attentive ensouled machines are both
exploiting the purely mechanical means of judging distance; natural
geometry is an overlay that only minds can place on top of the
awareness purchased by images drawn on the pineal gland. Human
distance perception is richer than that of mere machines.
It looks as though there is a price to be paid for this maneuver. If
Descartes were right, we should expect to see animals performing more
poorly than humans where distance vision, and detection of the com-
mon sensibles in general, is concerned. Just the opposite seems to be the
case. Descartes can reply, however, that natural geometry allows us to
perform better than we would otherwise. The superior acuity of other

81
Bacon 1928, vol.2, 544.
82 83
Bacon 1928, vol.2, 545. See Gaukroger 2002, 200 f.
 THE EARLY DESCARTES 

animals is a testament to their clever design. Presumably God could

have obviated the need for natural geometry if he wanted to, by endow-
ing us with superior eyes and brains. Why he did not do so is just one
more thing lost ‘in the inscrutable abyss of his wisdom.’84
None of that is to say that Descartes’s early position is without its
problems, of course. Descartes abandons much of the early view and
we will try to uncover his own reasons for doing so in the coming
chapters. Before pressing on, we should stop to gather up the threads.
We have three very different processes happening in attentive
human perception. First, we have the experience of sensible qualities.
This is to be explained by analogy with language. To use Descartes’s
example: when one senses color, one of the tiny tubes that empty into
the pineal gland is ‘opened differently as the action that causes it
differs.’85 This difference in how the tube is opened gives the soul
occasion to sense a color, just as hearing ‘dog’ can give the mind
occasion to think of dogs.
Alongside that process, of course, is the image that is being traced
on the pineal gland. The surface of the pineal gland is ‘the seat of the
imagination and the common sense.’86 In perception and imagin-
ation, the soul immediately considers the image on the pineal gland.
As we have seen, Descartes believes there is a corporeal counterpart to
color as we experience it. This pseudo-color helps capture the rela-
tions among the bodies perceived, such as distance and position.
Then we find yet a third process, natural geometry. In L’Homme,
this process is chiefly designed to make estimations of distance and
relies on such data as the behavior of the fluid in the eye and the
opinions the soul already has about relative position.
The challenge is to see how all three processes are supposed to fit
together in attentive perception. How is it that the mind comes to see
colors spread out on objects? Descartes might have explained how the
mind summons its sensations of color, but nothing tells us how those

84
From the replies to Gassendi, AT VII 375/CSM II 258; see Meditation Four, AT
VII 55/CSM II 39.
85 86
AT XI 176/G 149. AT XI 176/G 149.
 THE EARLY DESCARTES

sensations are being related to mind-independent objects. Presum-

ably that process has to take a loop through the brain image: the mind
must be uniting its own sensations with the corporeal idea in order to
produce technicolor experience. And I have no idea how the mind
could be doing that.
Descartes is not bothered by this problem of ‘localizing’ sensation
because his chief interest in this part of L’Homme is setting out the
purely mechanical means by which activities in the overlap area are
carried out. On this early view, we do not need a sensation of color or
any other sensible qualities in order to judge the common sensibles.
The real work is being done by the brain image, while natural geo-
metry is brought in to aid in our judgments of distance. Only in
natural geometry is there even a mention of the use of color to judge
the geometrical relations among bodies. Nothing in the text suggests
that sensible qualities are relevant to anything but this one aspect of
natural geometry. Sensible qualities like color and sound are, for the
most part, epiphenomenal. They are simply irrelevant to positioning
and individuating bodies in experience.
In that respect, Descartes’s early theory of perception is indeed a
radical departure from those of his predecessors. The Aristotelian
tradition is seemingly in unanimous agreement on one point: that the
only way to become aware of the common sensibles is through the
proper ones. Only in receiving light and color can a person see shape,
size, motion, and the rest. The early Descartes rejects that picture
wholesale. As far as the early Descartes is concerned, the only thing a
person who did not experience light and color would be missing out
on is one extra means of judging distance. In both attentive and
inattentive perception, such a person would be just as able as some
of the unensouled machines at navigating the world. So although
Descartes begins with the materials at hand—the remnants of the
Baconian synthesis, the doctrine of turning toward the phantasms—
when he constructs his new theory of perception, his final product
could hardly be more distant from those sources.
 3
The Meditations

Descartes’s early philosophy of perception is not an aberration. Its

core is carried right through to the last of the Meditations. If so, our
inquiry has to diverge from the strict chronology of publication.
Otherwise, we would move straight to the Dioptrique, which appears
in 1637. My thesis is that the Descartes of the Meditations is substan-
tially closer to that of L’Homme than the Dioptrique, even though the
Dioptrique comes between these two texts.1 Later on, I’ll offer some
hypotheses about why this is so. If I am right, moving directly to
the Dioptrique would only breed confusion.
The ‘I’ of the Meditations is not Descartes himself but a fictional
character, one who begins her meditations from a thoroughly Aris-
totelian standpoint.2 The meditator’s thoughts follow a gradual pro-
gression, beginning in the confusion of scholastic innocence and
ending in the clear light of Cartesian sophistication. If so, we cannot

1
Throughout his career, Descartes will direct readers to the Dioptrique for his
views, even in the course of saying something blatantly inconsistent with that text. But
note that in those referrals, Descartes typically points to a very specific claim he has
defended in the Dioptrique, rather than to that text as a whole. (For example, in the
sixth Replies (AT VII 437/CSM II 295), Descartes appeals to the Dioptrique’s rejection
of intentional species, a rejection that is consistent throughout his career.) There is no
compelling reason to shoehorn whatever later texts one comes across into the
framework provided by the Dioptrique. At any rate, I shall resist doing so, in order
to see what view emerges if we take Descartes at his word.
2
Several others (e.g., Jorge Secada 2000) have made these two points. For textual
evidence, consider that Descartes tells us in the Synopsis of the Meditations that ‘no
sane person has ever doubted’ ‘that there really is a world, and that human beings have
bodies’ (AT VII 16/CSM II 11). The meditator will of course go on to doubt precisely
these things. Assuming Descartes considers himself sane (something even the insane
have been known to do), the meditator cannot be Descartes himself.
 THE MEDITATIONS

assume that everything the meditator says in the early days of her
progress represents Descartes’s considered view. As a result, it makes
sense to take the sixth and final Meditation as announcing Descartes’s
position, using material from earlier Meditations with some caution
and only when it is consistent with the last of the Meditations.
This developmental reading of the Meditations is, of course, noth-
ing new. What is new, I think, is the prospect it introduces of taking
some material, even from the third Meditation, as merely provisional,
to be revised in light of later arguments. If the meditator makes
progress as she frees herself from the bonds of the senses, there will,
trivially, be claims that she rejects or would reject on further reflec-
tion. Many commentators assume that once we begin the second
Meditation, more or less everything that follows has the same canon-
ical status. I hope to show that we can make headway on vexing issues
such as material falsity if we abandon that assumption.
At the very beginning of our inquiry, we saw that the Francophone
early moderns typically cast themselves as intellectual adults, in
opposition to their Aristotelian and scholastic predecessors, whom
they see as never having overcome the prejudices of childhood. This
theme is sounded throughout many of Descartes’s other works.3 But
in the Meditations, it is present in the very first sentence: the medi-
tator is ‘struck by the large number of falsehoods’ she had ‘accepted
as true in childhood.’4 In the opening two meditations, she pauses to
review her prior beliefs: she believed that she was a human being, a
rational animal; that all knowledge comes from the senses; that the
qualities she senses either are or resemble mind-independent prop-
erties of objects.5 All three claims are, of course, articles of the
Aristotelian faith. Early in the sixth Meditation, she recounts her
initial beliefs yet again.6 On her prior view, she had sensations of

3 4
See Principles I.70 and 71. AT VII 17/CSM II 12.
5
See AT VII 26/CSM 17 for the meditator’s prior beliefs about her nature and AT
VII 18/CSM II 12 for her previous views on knowledge.
6
‘To begin with, I will go back over all the things which I previously took to be
perceived by the senses, and reckoned to be true; and I will go over my reasons for
thinking this’ (AT VII 74/CSM I 51).
 THE MEDITATIONS 

‘light, colours, smells, tastes and sounds, the variety of which enabled
[her] to distinguish the sky, the earth, the seas, and all other bodies,
from one another.’7 Note that the meditator’s original view casts
sensible qualities as the means by which individuation happens.
Such a process would have been unproblematic, precisely because
prior to meditating, our protagonist thought that the objects of her
sensings either resembled or just were mind-independent features of
those objects.
But just as the meditator will give up on that resemblance claim,
so too will she give up on the prospect of using her sensations to
individuate bodies. Instead, I’ll argue that the work of directing the
mind toward bodies in the world is accomplished by the brain image.
In the Meditations, Descartes of course is not interested in setting out
a full-blown theory of vision, let alone perception in general. What he
does say comports with L’Homme and is silent on the whole question
of natural geometry. And since the only role for sensible qualities in
L’Homme was in the context of those sub-conscious inferences, the
Meditations makes no demands on sensible qualities in perceptual
experience.
I shall argue that, on the view the meditator ultimately achieves,
sensations are not representations in any interesting sense. What,
then, is happening to us, when we survey the world around us?
Objects act on our sense organs, which then, by means of motion
alone, transmit an image to the pineal gland. In sensory experience,
our attention turns toward the brain. The image sketched in the brain,
while not a mental state, nevertheless deserves to be called an ‘idea,’
since it is responsible for directing our attention through it out into
the world.
My reading of the Meditations is controversial in several ways.
First, as with L’Homme, it attributes to Descartes a limited, if very
odd, form of direct realism: in sensory and imaginative experience,
the mind is immediately aware of an extra-mental thing, namely, the
brain. Second, my reading does not require the subject to deploy

7
AT VII 75/CSM I 52.
 THE MEDITATIONS

the innate idea of extension in sensory experience. Finally, sensations

are not representations, on my reading. Nothing in the text requires
or suggests that they play any role in representation, and several
Cartesian positions at least prima facie seem to block them from
doing so. But if so, don’t they become so much ‘cognitively vacuous’
‘mental decoration’ (in Alison Simmons’s apt phrases)?8 Even if they
do not represent bodies in the internalist’s sense, why shouldn’t they
be representations in the sense specified by externalism, that is,
mental items that allow us to track repeatable properties or states
of affairs? I shall argue that such a view overplays the teleological
language of the sixth Meditation. Sensations have two important jobs
to do, but neither of them requires them to be representations. We
shall address these and other objections in due course, but they are
worth keeping in mind as we go.
So much for our destination. To get there, I begin with a discussion
of the nature of ideas, taking the third Meditation as a guide. There, it
becomes clear that ideas represent their objects by resembling them,
in a sense we’ll have to determine. In that Meditation, the protagonist
simply cannot tell whether her idea of cold, for example, is an idea of
a sensation or of a mind-independent quality of objects. In line with
the developmental reading, I argue in section two that Descartes’s
discussion of material falsity, particularly in the fourth Replies, shows
that the idea of a sensible quality is an idea of a sensation. By the
end of her Meditations, she has learned that the sensation itself is
not a representation of anything but merely a blank effect. How, then,
is perceptual experience to be explained? Section three explores
the Meditations’ version of ‘turning toward the body.’ Brain images
are needed to do the jobs the meditator originally assigned to sensible
qualities. Perhaps ironically, a paradigmatically mental state like the
sensation of yellow is not, in the strict sense, an idea; and yet a paltry
etching on the pineal gland, in some circumstances, is. Section four
sets out Descartes’s sixth Meditation account of sensory experience,

8
1999, 349.
 THE MEDITATIONS 

while the fifth ties up the loose ends by exploring the roles Descartes
thinks sensations must play.

3.1 Sensations, Ideas, and Intentional

Resemblance
Does the Descartes of the Meditations think sensations are represen-
tations, or not? This must be an extraordinarily difficult question,
since it has been answered so often, and in so many different ways.
On some readings, mental experiences of sensible qualities are blank
effects, to be understood on analogy with pain.9 On other readings,
Descartes offers an externalist account, whereby sensory states rep-
resent their causes, or what God intended to be their causes.10 Still
other interpretations cast Descartes as an internalist who holds that
sensations must involve the mind’s innate idea of extension.11 On
that reading, to speak of ‘sensations’ is to refer to an aspect of an
idea that also includes geometrical features.12 Such ‘fusion’ accounts
deny that sensations and ideas of geometrical qualities are distinct
ingredients in sensory experience.
What accounts for this seeming free-for-all? When it comes to
externalist, tracking readings of sensations, I think there are a number
of tightly nested confusions, which take some work to root out. I’ll
address these in the last section of this chapter. (Until then, I shall
always mean ‘representation’ in the internalist sense.) The debate
over the other issues is less tractable. Each side can point to a favored
text and seemingly cinch the case. This suggests that we be more
cautious in selecting our source texts. If we take the developmental

9
Pain might itself be representational for Descartes, so the analogy is not uncon-
troversial. The best defense of the non-representational reading of Cartesian sensation
in general is Laura Keating (1999). Ann Wilbur Mackenzie (1989, 182) also argues for
this conclusion, on the grounds that the purported objects of Cartesian sensations are
impossible, and Descartes thinks that ideas can only be of possibilia.
10
For the causal reading, see Margaret Wilson (1990); for the teleosemantic
interpretation, see Alison Simmons (1999).
11
See De Rosa (2010).
12
Simmons (2003) and De Rosa (2010) both argue for the fusion account.
 THE MEDITATIONS

reading seriously, we should expect to find inconsistencies across

Meditations, as the protagonist wends her way from scholastic dark-
ness to the clear light of the sixth day.
Although the second Meditation reveals that the essence of any
body such as a piece of wax is a determinable, rather than any
determinate, property, it says nothing about the status of sensible
qualities. The third Meditation finds our protagonist unsure whether
a sensible quality can exist outside the mind. But by the time she
reaches the last Meditation, she sees that the only mind-independent
properties of bodies are geometrical ones. As a result, she sees that her
sensations cannot be representations; they not only do not but cannot
represent anything. To make this case, I shall argue that Cartesian
representation is resemblance.13 It is the failure of sensible qualities to
be among even the possible properties of physical things that accounts
for the failure of sensations to resemble, and hence to represent,
anything.
Let us begin with the third Meditation, that half-way house on the
road to enlightenment. There, the protagonist sets about cataloguing
her thoughts. Her first distinction is between a strict and loose
sense of ‘idea.’ Some thoughts are ‘as it were images of things’—
these are ideas in the strict sense. Other thoughts must include a
‘likeness’ but also have ‘something more’: these are volitions, affects,
and judgments.14 In what follows, I shall always take ‘idea’ in the
strict sense. The loose sense includes what we would call propos-
itional attitudes: judging that something is the case, or fearing that
something might be. Strict-sense ideas are the sub-propositional
elements that are ingredients of loose-sense ideas.
Since the meditator has already argued that thoughts are not
limited to sensations, the talk of ‘as it were’ images must emphasize

13
In a very helpful paper, David Scott (2010) also reaches this conclusion.
14
AT VII 37/CSM II 25–6. Nicholas Jolley (1990, 23) offers an intriguing alternative
reading of the distinction. On his view, ideas in the loose sense include judgments, affects,
and non-representational states; in the strict sense, they are only representations. I find
this reading strained. Ideas in the loose sense are representations plus something else,
such as an affect or judgment, not representations or non-representations.
 THE MEDITATIONS 

the ‘as it were’ and play down the ‘image.’ What all ideas in the strict
sense have in common, whether they are images or not, is their role as
representations, as being of something. This point is worth empha-
sizing. The meditator claims that ‘ideas can only be, as it were, of
things,’15 and I see no evidence that Descartes departs from this
position in the Meditations. Nothing guarantees that the object of
an idea in fact exists. This is why Descartes uses ‘as it were’ a second
time: not only are they merely as it were images, since not all are
images, they are only as it were of things, because only when those
things exist is there anything for them to be of. We might say, then,
that all ideas have in common the feature of representing some being
that is at least possible, if not actual. A hallucinatory experience, no
matter how inspired, still presents the basic features of bodies in a
logically possible array.16
Rather more darkly, the meditator claims that ‘the objective mode
of being belongs to ideas by their very nature.’17 We’ll soon deal with
the ‘objective reality’ of ideas, a phrase Descartes lifts from the scho-
lastics.18 For now, the crucial point is that, whatever one makes of
the objective mode of being, all ideas have it, and they have it not
contingently but as a result of their very nature. The immediate object
an idea presents is something internal to the mind, or, as we shall see,
to the mind-body union.19 Whether anything outside the body resem-
bles that idea is a further question.

15
AT VII 44/CSM II 30.
16
See the discussion of the ‘simpler and more universal’ things, such as corporeal
nature in general, that are ‘as it were the true colours from which we form all the images
of things, whether true or false, that occur in our thought’ (AT VII 20/CSM II 14).
17
AT VII 42/CSM II 29.
18
A very useful treatment of the scholastic origins of Descartes’s jargon is Calvin
Normore’s (1986). For a recent and intriguing treatment of objective reality in
Descartes, see Lionel Shapiro (2012).
19
Since Wilson (1990), it is common to distinguish between an idea’s presenta-
tional and referential content. The presentational content of an idea is the way an
object is presented to the mind; an idea’s referential content is the object the idea is
‘about.’ (For discussion, see De Rosa 2010, 32.) I find this distinction unhelpful in the
present context. Talking of the mode or way an object is presented invites anachron-
ism, since those terms have already been claimed by Gottlob Frege (or his translators)
for a very different use. I don’t see any evidence for a notion of Fregean Sinn in the
 THE MEDITATIONS

If I am right so far, all ideas are as it were of something. But what

are ideas themselves? In the Second Replies, Descartes gives an offi-
cial, if maddeningly opaque, definition of the term: ‘By the word
“idea” I understand the form of any thought through the immediate
perception of which I am aware of that very same thought.’20 Suppose
we take ‘thought’ in the most general possible sense. Then we might
think of the form of the thought as whatever features make it the kind
of thought it is (judgment, volition, etc.) and make it of whatever it is
of. The definition in the Second Replies, then, is a definition of ‘idea’
in the loose sense of Meditation Three. In the narrow sense, individ-
ual ideas give a propositional thought the content that it has. If I am
thinking that the Eiffel Tower is hideous, what makes my thought of
the Eiffel Tower and not the Washington Monument is the idea that
figures in the subject-place.
In virtue of what can ideas play this role? What is it about them
that allows the mind to use them to ‘zero-in’ on some object, whether
actual or possible? Descartes isolates two aspects of ideas: their formal
aspect, that is, their status as modes of minds, and their objective
aspect, that is, their status as representations.21 The formal and
objective distinction is borrowed from scholastics such as Francisco
Suárez; but, as Secada aptly puts it, ‘[a]s in other instances, [Descartes]
took the robe but not the flesh.’22

Meditations. Now, someone might say that the way an object is presented need not
evoke any Fregean doctrines, but simply points to the introspectible content or nature
of the idea. If so, I suppose this way of talking is unobjectionable.
20
AT VII 160/CSM II 113. What follows in this paragraph builds on Andrew
Pessin’s (2009, 6), though in ways I doubt he would approve of.
21
I shall not spend much time on the three distinctions Descartes draws among (1)
formal/objective reality; (2) ideas in the formal/material sense; and (3) ideas in the
objective/material sense. I agree with Jolley (1990, 14–15) that distinctions (2) and (3)
amount to the same thing. For our purposes, the relevant distinction is between an
idea qua a mode of the mind and an idea as a representation.
22
Secada (2000, 80). See Francisco Suárez’s MD II, 1, 1, quoted in Secada (2000,
80) for the formal/objective distinction. Secada argues that for Suárez, the ‘objective
concept,’ unlike the ‘formal concept,’ is not really what we would call a ‘concept’ at all:
instead, it is the extra-mental thing that is the object of the formal concept. As Secada
reads him, Descartes retains the claim that the idea qua objective concept is the object
 THE MEDITATIONS 

Rather than pursue the intricacies of scholastic terminology, our

investigation into how ideas represent is best pursued by beginning
with the common ground between Descartes and the scholastics. As
Suárez puts it, ‘almost all philosophers and theologians . . . place cer-
tain similitudes of the objects in the cognizing powers in order to
unite through them the objects and the powers.’23 So far, Descartes
agrees: all judgments, volitions, and affects include some likeness of
the thing. It is important to note that the thing an idea resembles
need not exist; or, to remove the air of paradox, we can say that the
thing it would resemble need not exist. After all, Descartes’s examples
of propositional attitudes here include things like willing, fearing,
affirming and denying. This is quite a grab bag: one can be said
to fear x, but only to affirm or deny that x is F. Still, it would be
absurd to assume that no one fears what does not exist, or affirms a
property of an inexistent.
Where Descartes parts ways with Suárez, Caterus, and their ilk is
on the status of this similitudo. For it is not merely an aspect of an act,
or part of the machinery whereby an act is accomplished. It is itself
the immediate object of thought.
Consider how Descartes replies to Caterus. Caterus argues that
to speak of objective being in the mind is simply to speak of ‘the
determination of an act of the intellect by means of an object, and
this is merely an extraneous label which adds nothing to the thing
itself.’24 For Caterus, the objective being is the object itself, which
determines the mind in a particular way. Talking of that objective
being as being ‘in the mind’ (or intellect) is just a slightly misleading
way of indicating that the object in the world is being thought about.
This is not a real property of the object itself but an extrinsic, purely
relational one. Descartes agrees but then charges Caterus with having

of a mental act. But for Descartes, ideas are instead internal to the mind or the mind-
body union.
23
Commentaries on Aristotle’s de anima, d. V. q. 1, art. 3, in Suárez (1978–81),
quoted in Secada (2000, 84).
24
AT VII 102/CSM II 74.
 THE MEDITATIONS

misunderstood him. Descartes wasn’t talking about the object in the

world at all. ‘I was speaking of the idea, which is never outside the
intellect, and in this sense “objective being” simply means being in
the intellect in the way in which objects are normally there.’25
This notoriously unhelpful last phrase has given rise to a number of
different interpretations. Some take Descartes to mean that objects
exist formally in the world and immaterially or objectively in the
intellect.26 On this view, a single thing, or perhaps its essence, has two
manners of existence. Such a view is hardly unheard of in the seven-
teenth century; it is precisely the position John Sergeant defends,
extending Aquinas’s view.27 And it certainly has a foundation in the
text. At times, Descartes says that the thing in the world exists, not in
the mind per se, but in the idea: the objective reality of an idea is ‘the
being of the thing which is represented by an idea, insofar as this
exists in the idea.’28
There is a less mysterious alternative. If we put Descartes’s first
Replies together with the third Meditation, we can say that objects are
normally in the intellect by means of similitudo or imago, likeness or
image. That is, they are there only in the sense that the mind contains
a likeness of them. In fact, it’s not clear that there is much daylight
between these two interpretations. As Locke argues against Sergeant,
if one thing has two manners of existence, then it seems to be one

25
AT VII 102/CSM II 74.
26
For example, Andrew Pessin claims that ‘Fred’s thinking of the sun, then, is not
entirely external to the sun, but in some way involves the very same sun which exists
“out there” ’ (2009, 8); ‘To say the sun itself exists in the intellect would, then, be to say
that a certain essence or nature, of a body with certain mechanical properties
including size and location, exists objectively, the very same essence which, realized
in nature, is the formal sun’ (2009, 9).
27
See Sergeant 1984 (first published 1697), 37–8. For Aquinas, see esp. his
commentary on the de anima, where he claims that an essence ‘can have two different
modes of being: material being insofar as it is in natural matter; and immaterial being
insofar as it is in the intellect’ (Commentary on the de anima, II.12, trans. Claude
Panaccio 2001, 186–7). Sergeant departs from Aquinas on this point, since he thinks it
is not merely the thing’s essence but the thing itself that has the two manners of
existence.
28
AT VII 161/CSM II 113–14, discussed in Vinci (2008, 465 f.).
 THE MEDITATIONS 

only qualitatively, not numerically.29 And at that point, the only

relation between the two is similarity, not strict identity. Indeed,
Aquinas himself, while insisting on the identity of the species in the
mind and the world, at the same time speaks of things’ being in the
intellect by way of similitudo.30
All of this suggests that the likeness in the mind plays its repre-
sentational role precisely in virtue of being a likeness. Any reader who
has followed me this far will be tempted to close the book in disgust.
Resemblance is a relation that can obtain only between a picture and
an original; how could a Cartesian idea, especially an idea of the
intellect like that of extension, resemble anything?31
But here it is worth keeping in mind the variety of notions of
‘resemblance’ on offer in scholastic and early modern thinking. Aqui-
nas, for example, explicitly claims that the kind of likeness involved in
representation is not pictorial likeness. In fact, it is sui generis, neither
identifiable with visual resemblance nor reducible to any other, famil-
iar kind of resemblance.32 Since Descartes explicitly tells us that all
ideas are only ‘as it were’ images, he might well think that ideas ‘as it
were’ resemble their objects. Now, Descartes never says just what he

29
Sergeant claims that his ‘notions’ just are the things themselves, existing imma-
terially in the intellect. In his marginal note, Locke writes, ‘And soe our good Author
has at last proved his Notions are likenesses of things’ (in Sergeant 1984, 38).
30
For more on this aspect of Aquinas’s view, see Panaccio’s excellent (2001, 196 f.).
31
There is some irony here. In twentieth-century philosophy of mind, resemblance
is often ruled out as a candidate for the means of representation because it is too
promiscuous, since everything resembles everything else in some respect or other. In
the Descartes literature, it is ruled out for being too restrictive.
32
See Pasnau (1997, 105–13) and Brower and Brower-Toland (2008). Pasnau
claims that representation is similarity for Aquinas, while Brower and Brower-Toland
argue that Aquinas takes representation to be a primitive. I am unsure how deep this
disagreement is, since Pasnau emphasizes that, for Aquinas, the relevant sense of
resemblance is itself sui generis. As Brower and Brower-Toland point out, ‘the most
Aquinas ever says about intentional similarity or likeness is that “it can be understood
as representation” (QDV 2.5 ad 7; see also 2.3 ad 9, 4.4 ad 2, ST 1.85.8 ad 3, and SCG
2.46)’ (2008, 221). Pasnau (1997, 112) recognizes this. The difference is that Pasnau
takes Aquinas’s silence to indicate a promissory note, a reflection of a gap in his
knowledge of the workings of the mind that later empirical investigation might fill.
For Brower and Brower-Toland, there is no such gap, since Aquinas never intends to
provide or even make room for a substantive theory of representation.
 THE MEDITATIONS

takes the resemblance relation to be. Following Robert Desgabets and

Louis de la Forge, I will call it ‘intentional’ resemblance, which they
distinguish from ‘real’ or ‘entitative’ resemblance. Although the medi-
tator often calls alleged resemblance relations into question, these
attacks are, as David Scott argues, confined to the claim that sensible
qualities resemble anything in objects.33 Ideas, by contrast, do (inten-
tionally) resemble their objects (when those objects in fact exist
outside the mind).
It is hard to see what else we could make of the meditator’s repeated
insistence that ideas are tanquam rerum imagines.34 In the very passage
that draws the strict/loose distinction, she claims that an idea in the
loose sense ‘includes something more than the likeness of the thing (‘res
similitudinem’).35 And that of course implies that such an idea includes
at least the likeness of the thing. Moreover, she seemingly takes this to be
convertible with the claim that ‘there can be no ideas that are not as it
were of things’ [‘tanquam rerum esse possunt’].36 In the French trans-
lation, this becomes the claim that ‘ideas, being like images, must in each
case appear to us to represent something.’37 And we have already seen
that intentional likeness is a traditional way of spelling out how things in
the world can be ‘in’ the mind.38

33
As Scott puts it: ‘In effect, I am claiming that Descartes’ departure from the
Scholastics on the question of resemblance is only partial, confined as it is to sensation.
If it is true that at the intellectual level the Scholastics affirm resemblance, then
Descartes is in at least formal agreement with his predecessors on this score’
(2010, 503).
34
AT VII 37/CSM II 25. I am grateful to an anonymous referee, who pointed out
that Descartes’s insistence that we cannot fully comprehend God’s nature might make
trouble for the claim that our idea of God resembles him in any way, even merely
intentionally. In my view, this is a problem for any reading of Descartes; put
differently, it is a problem for Descartes.
35 36
AT VII 37/CSM II 26. AT VII 44/CSM II 30.
37
As noted in CSM II 30 n.2.
38
And as Scott notes, Meditation Three is framed as an inquiry into which of the
subject’s ideas resemble things outside the mind and which don’t. Consider the two
ideas of the sun the meditator finds in herself (AT VI 39/CSM II 27). Scott writes,
‘Descartes reckons that two quite disparate ideas of “the sun itself,” one drawn from
reason and the other from sensation, cannot both resemble the same object; and the
phrasing shows that for him the issue is clearly which one of those ideas resembles. His
 THE MEDITATIONS 

It is, naturally, very difficult to say much about intentional resem-

blance, since we have been told so little. We know from Aquinas that
it is not to be identified with agreement in nature, that is, the sharing
of forms.39 What counts for our purposes is that, however distant
Descartes’s intentional resemblance is from ordinary pictorial resem-
blance, it is still an internal relation. If it obtains, it does so in virtue of
the intrinsic properties of the relata. No one should think resem-
blance’s being an internal relation entails that it is itself an intrinsic
property of an idea. That would have the absurd consequence of
guaranteeing the existence of any object of thought.
Reading Cartesian representation as resemblance is defensible not
just in terms of the texts and the intellectual context in which Descartes
writes; it is in fact required by the logic of Descartes’s own position.
Simon Foucher’s 1675 Critique of Descartes and Malebranche makes
this point very clearly.40 Foucher poses a dilemma:

If [ideas] can represent without being like, not only all the ideas we have,
whether in our senses, imagination, or some other way whatever it be, have
as much right to represent, the ones as much as the others, but all our ideas,
whatever they are, would be able to represent one and the same object, a
position one cannot support.
If it is necessary that our ideas be like to represent, either one must
conclude from this that we cannot have science, or else that the notions of
the soul and of matter on which the author relies are entirely contrary to
what is true of them.41

Suppose one abandons the resemblance requirement. One would

have no grounds for denying that any of our mental contents are

premise, in other words, is that “both cannot” resemble, not that “neither can” (which
it would have to be were he out to impugn resemblance)’ (2010, 499).
39
In a very useful paper, Claude Panaccio (2001, 197) argues that it would be going
too far to say that this means there is no natural or ordinary resemblance involved.
There must be some core of features in common between the two kinds of
resemblance.
40
Foucher is attacking both Descartes and Malebranche, as the full title makes
clear: ‘Critique of the Search for the Truth in which is examined at the same time a part
of the Principles of Mr. Descartes.’
41
See Foucher 1995, 32–3.
 THE MEDITATIONS

representations. There would be no reason to deny a sensation of cold

the status of a representation. Worse, there would be no reason for
saying that any two ideas, no matter how different, did not in fact
represent one and the same object.42 If representation swings free of
resemblance, then nothing about the ideas themselves can tell you
what that idea is an idea of. Foucher is assuming that one can discern
what an idea represents on the basis of its monadic properties only if
representation is resemblance. One might question this. But certainly
the resemblance relation is, in this period, the only account of repre-
sentation going that would allow one to know that idea x represents
object y (if y exists), just on the basis of the intrinsic, introspectively
available features of x.
Foucher’s second horn is hardly more appealing. For if we have
ideas of sensible qualities that in fact do resemble something in the
world outside of the mind, then the whole distinction between sens-
ible and geometrical qualities goes by the board. All qualities would
then be equally real, and the correct notion of matter would be
entirely different from the austere notion of extension.
But there is a third alternative Foucher misses: although ideas must
resemble in order to represent, sensations, as Descartes explicitly tells
us in the Principles, do not represent anything outside the mind.43 As
a result, their failure to resemble anything in the world is entirely
unproblematic; they neither resemble nor represent.
Descartes emerges from Foucher’s attack unscathed, at least on this
score. The important point for us is, once again, the limitation it puts
on intentional resemblance. While it is not property sharing, it must
still resemble ordinary resemblance in one respect: being aware of one
of the relata allows you to know something about the other.
Before leaving this section, I should note a further objection. I said
earlier that the meditator argues against the resemblance of sensible
qualities to anything else; the only properties of bodies are modes of

42
1995, 32.
43
Descartes claims that sensations ‘do not represent anything located outside our
thought’ (Principles I.71, AT VIIIA 35–6/CSM I 219, discussed above).
 THE MEDITATIONS 

extension. But there is at least one text, albeit from the third Medi-
tation, where she argues instead against the resemblance of a seem-
ingly non-sensory idea to an object. The meditator purports to have
two ‘ideas’ of the sun, one gotten from the senses, and one from the
intellect.44 Only the intellectual idea resembles the sun. But if repre-
sentation is resemblance, how can both ideas be of the sun?
It isn’t only resemblance readings that face this problem. Raffaella
De Rosa develops a descriptive account of ideas, according to which
an idea represents its object by virtue of describing it. It then becomes
obscure how an idea can be of something it misdescribes. On her
account, sensory ideas ‘present bodies confusedly, as possessing some
properties that they do not actually present.’45 If this works as a
solution for her view, it also works for the resemblance account. For
one could then say that resemblance is a matter of degree. The
astronomical idea of the sun has more of the features the real sun
has than the idea given through sensation. Unfortunately for both of
us, the meditator explicitly rejects this option. To her, it is obvious
that ‘both these ideas cannot resemble the sun.’46
Happily, there is another way out of this maze. A natural reading
of the Meditations passage takes the two ‘ideas’ of the sun to be ideas
only in the loose sense. One idea ‘shows the sun to be several times
larger than the earth,’ while the other ‘makes the sun appear very
small.’47 If these ideas have truth values, that is, they present the sun
as having some property or other, they cannot be ideas in the strict
sense. What we have are competing beliefs, one acquired ‘as it were’
from the senses.48 That telling phrase, from the third Meditation,

44
For an intriguing, if challenging, treatment of the two ideas of the sun, see
Norman Wells (2003).
45 46
De Rosa (2010, 25). AT VII 39/CSM II 27.
47
AT VII 39/CSM II 27.
48
D.T.J. Bailey (2006) argues against the claim that Descartes is simply being
careless in not distinguishing between representations and complexes of such repre-
sentations that admit of a truth value. I do think that Descartes’s claim in the
Principles (I.48) that ‘[a]ll the objects of our perception may be regarded either as
things or affections of things, or as eternal truths’ (AT VIIIA 22/CSM I 208) seems
designed to elide the difference between propositions, objects, and properties. Still,
 THE MEDITATIONS

signals that the senses themselves do not produce these ‘ideas’ at all.
In the sixth Replies, as we’ll see, Descartes makes clear that to speak of
the senses as producing beliefs is misleading; strictly speaking, we are
talking about ‘the judgments we formed without any reflection in our
early childhood.’49 Both judgments can be about the sun, by virtue of
both including an idea of the sun, even if only one judgment is true.
We can find further evidence for taking ‘idea’ in the loose sense if
we ask: why is it so obvious to Descartes that intentional resemblance
where loose-sense ideas are concerned doesn’t admit of degrees? After
all, in the Dioptrique, Descartes is happy to speak of the varying
degrees to which the image in the brain and the eye resembles the
object in the world. What makes intentional resemblance, at least at
the level of loose-sense ideas, so different?
In the fifth Objections, Pierre Gassendi presses Descartes on
degrees of resemblance. Why not say that ‘both these ideas of the
sun resemble the sun and are true, or conform to the sun, though one
does so more than the other’?50 Descartes’s reply is prima facie odd:

Your taking the two ideas as one on the grounds that they are referred to
(‘referuntur’) only one sun is like saying that a true statement does not differ
from a false one because it is asserted of the same subject.51

But Gassendi never said that the two ideas are one.52 Descartes must
be assuming that if two ideas resemble or are referred to the same
object, then they are not two ideas, but one. This is why it was so
obvious to the meditator that only one of these ideas could resemble
the sun. If both did, then they would be the same.

I do not think the Descartes of the Meditations can be accused of carelessness.

Descartes has already distinguished between ideas in the strict and loose senses and
so feels free to go on using the term in both senses, relying on readers to disambiguate
where necessary.
49 50
AT VII 438/CSM II 295. AT VII 283/CSM II 197.
51
AT VII 363/CSM II 251.
52
In fact, in the next sentence of his reply, Descartes says that Gassendi claims that
‘the idea we arrive at by astronomical reasoning is not in fact an idea.’ But if it’s not an
idea at all, Gassendi can hardly have meant that the astronomical idea of the sun is the
same idea as the sensory one.
 THE MEDITATIONS 

None of this reasoning makes the slightest sense if we have to read

‘idea’ in the strict sense. If ideas are ‘as it were images,’ then they seem
to admit of differing degrees of resemblance, just like images them-
selves. Of course there will be many differences between images and
ideas, but none of them, as far as I can tell, rules out degrees of
resemblance. But suppose Descartes is taking ideas in the loose sense,
as a complex of strict-sense ideas and a propositional thought.
A loose-sense idea of the sun, then, asserts that its subject has a set
of properties. What is intentional resemblance at the level of loose-
sense ideas? It’s simply truth or falsity. One ‘idea’ of the sun has it as
very small; one has it as very large. At least one of these ideas will be
false and hence fail to resemble its object.
By way of concluding this section, let us take a step back and see
how what we’ve learned illuminates the question of sensory repre-
sentation. If I am right so far, sensations cannot be representations,
for there is nothing for them to resemble, whether intentionally or
in any other way. My argument so far has turned on a controversial
analysis of Cartesian ideas and representation, together with the
uncontroversial claim that the Cartesian world beyond the mind
has no place for sensible qualities. But we can get to the same result
by examining another doctrine from the third Meditation, one that
will help clarify the relations among sensations, ideas, and qualities:
the doctrine of material falsity.

3.2 Material Falsity

I have been arguing that the third Meditation is committed to the
claim that all ideas are representations. At this stage, the meditator
does not yet know whether the sensible qualities she experiences have
any existence outside of her. This introduces the possibility that her
ideas of such qualities are materially false. This doctrine has been the
subject of much scrutiny. I think that, armed with the developmental
reading and a clear distinction between a sensation and an idea of it,
we can clear up the mystery.
The meditator says,
 THE MEDITATIONS

[A]lthough, as I have noted before, falsity in the strict sense, or formal falsity,
can occur only in judgments, there is another kind of falsity, material falsity,
which occurs in ideas, when they represent non-things as things. For example,
the ideas which I have of heat and cold contain so little clarity and distinctness
that they do not enable me to tell whether cold is merely the absence of heat or
vice versa, or whether both of them are real qualities, or neither is. And since
there can be no ideas which are not as it were of things, if it is true that cold is
nothing but the absence of heat, the idea which represents it to me as
something real and positive deserves to be called false; and the same goes for
other ideas of this kind.53

The first thing to note about the passage above is that its point is
epistemic. The meditator notices that she has ideas of cold and heat. If
so, then these ideas must by definition be representations of some-
thing. And since, whatever intentional similitude is, it is a function of
the introspectible nature of an idea, the meditator ought to be able to
tell what they represent (or at least, what they would represent if there
were an external world). The problem is that these ideas are too
confused to enable her to discern what each idea is an idea of. Is the
idea of cold the idea of a genuine quality existing outside the mind, or
merely the idea of one of the mind’s own states, namely, a sensation?
The meditator is simply not equipped at this stage of her development
to answer that question.
Note, too, that the passage is conditional: if cold is merely the
absence of heat, then the idea of cold presents nothing as if it were
something. The meditator is not asserting the antecedent. In fact, she
is not saying that the ideas of cold and heat are materially false; each
would be so only if it represented nothing as something. And that is
just what the confusion of these ideas prevents the meditator from
knowing. In the context of this Meditation, that is all the protagonist
needs to establish, since she is trying to show that the ideas of heat
and cold do not have enough objective reality to demand a cause
distinct from herself.

53
AT VII 44/CSM II 30, my emphasis. The doctrine of material falsity is not a
permanent feature of Descartes’s view. As Margaret Wilson and others have noted,
this doctrine is confined to the Meditations, the fourth Replies, and the Conversation
with Burman.
 THE MEDITATIONS 

The reader will forgive—and possibly thank—me if I do not rehearse

all the arguments and positions staked out on material falsity.54 For our
purposes, the chief problem is this: how can an idea of cold be of cold, if
it represents nothing? Arnauld presses the question:
[I]f cold is an absence, it cannot exist objectively in the intellect by means
of an idea whose objective existence is a positive entity. Therefore, if cold is
merely an absence, there cannot ever be a positive idea of it, and hence there
cannot be an idea which is materially false.55

The problem, then, is that if cold should turn out to be the absence of
heat, it would be impossible to have an idea of cold. And yet we plainly
do have an idea of cold.
The true solution cannot be given until we reach the end of the
Meditations. By then, we have learned that matter is nothing but exten-
sion. Any mode that is not a mode of extension cannot belong to body.
Once this is in place, we can see how to answer Arnauld properly.
For on this austere ontology, the only sensible qualities are to be
identified with sensations. Cold is not whatever is represented by the
sensation. If it were, it would make sense to think of it as something
outside the mind, and in particular, to wonder whether it is a positive
thing or merely an absence. But we cannot think of it in that way at
all. To call a sensation ‘of cold’ is simply to mark it as being the
sensation it is. It does not require either us or Descartes to posit some
mind-independent feature of the world as its representandum. That is
not to say, however, that there cannot be an idea of cold. It is simply
the idea of the sensation itself.
This point comes out in the reply to Arnauld. Arnauld’s worry is
that, without something positive to feature as the content of the idea,

54
Among the relevant sources here, in addition to those already mentioned, are
Richard Field (1993), Catherine Wilson (2003), and Cecilia Wee (2006). For a
thorough treatment of the literature, along with a novel interpretation, see Lionel
Shapiro (2012). My account is indebted to the work of Alan Nelson (1996), who also
distinguishes a sensation from the idea of that sensation (though we differ on other
points).
55
AT VII 206/CSM II 145. As Nelson (1996) shows, Arnauld himself ends up
endorsing a Cartesian account of material falsity.
 THE MEDITATIONS

there will be no idea at all. Descartes replies that the idea of cold really
does have ‘something positive’ in it, ‘namely, the actual sensation
involved.’56 Descartes goes on to speak of ‘the idea of the sensation
of cold (‘ideam sensûs frigoris’).’57 Cold is not an idea but what is
represented by an idea; the mistake comes in considering that the
thing represented is a quality or thing that can exist outside the
mind. It is merely a sensation, and there is no barrier to having
ideas of those.58
In other words, the idea of cold must have the objective mode of
being: unlike a sensation, an idea must truly be of something. And it is
of something: it is of the sensation at issue. True, Descartes tells us
that ideas of sensations ‘do not represent anything real.’59 But what is
‘real’ here is simply what exists in the world outside the mind.
Wondering whether the idea of cold represents a nothing as a some-
thing is just wondering whether the sensible quality we confront in
experience exists outside the mind. And a clear and distinct percep-
tion of the nature of body, such as the fifth Meditation supplies,
shows that sensible qualities are not possible modes of bodies.
Not yet having grasped any of this in the third Meditation, the
meditator simply does not know whether an idea of cold reaches out
to represent something outside the mind. That is why she cannot
simply appeal to the nature of extension as excluding cold, heat, and
all the rest. But after undergoing the following meditations, she comes
to see that, just as the fourth Replies has it, the idea of cold is the idea
of a sensation, not of a genuine mind-independent quality.
The idea of cold is materially false in the sense that it is obscure and
confused and so provides material for judging falsely. As the medita-
tor says, ‘I am unable to judge whether or not what it represents to me

56 57
AT VII 234/CSM II 164. AT VII 235/CSM II 164.
58
For further confirmation, see The Conversation with Burman. Descartes (report-
edly) says that I may err if I ‘consider the idea of colour, and say that it is a thing or a
quality; or rather I may say that the colour itself, which is represented by this idea (qui
per eam ideam repræsentetur), is something of the kind’ (AT V 152/CSMK IIII 337).
A color, then, is not an idea but what is represented by an idea, namely, a sensation.
59
AT VII 234/CSM II 163.
 THE MEDITATIONS 

is something positive which exists outside of my sensation.’60 Notice

that it is the idea, not the sensation, that is obscure and confused. It is
materially false only in the sense that it provides material for, and a
temptation to, false judgment. It is not materially false in the sense
that it is an idea of nothing. It’s the sensation, not the idea, that fails to
be of.
Here it is helpful to contrast ideas of geometrical qualities. In these
cases, Descartes tells us in the Principles, ‘we clearly perceive [them]
to be actually or at least possibly present in objects in a way exactly
corresponding to our sensory perception or understanding.’61 An
idea of a shape, for example, carries with it the possibility of being a
mode of a body (and of nothing else). It wears its instantiable
character on its face. Not so with ideas of sensible qualities. As the
third Meditation shows, you cannot inspect the idea of a sensation
and figure out that the sensation it represents cannot belong to a
body. For that, you need a grasp of the nature of body. The idea of a
sensation, then, is confused and obscure because, unlike ideas of
geometrical qualities, whether and how anything can instantiate it is
not obvious from the idea alone.
Later philosophers will in fact point to our experience of sensible
qualities as evidence of the real distinction between mind and body.
For Desgabets, for example, ‘the doctrine of the nature of sensible
qualities is the fundamental proof of the distinction between soul and
body.’62 Descartes himself never argues in that way, for the simple
reason that one learns of the diminished status of sensible qualities
only through the investigation of geometrical ones.
To sum up: we have to distinguish between an idea of a sensation
and the sensation itself. That distinction is clearly present in the
fourth Replies. The key to applying it is to see that the meditator is
not in a position when the doctrine of material falsity is first intro-
duced to make an informed judgment about the status of sensations.

60
AT VII 234/CSM II 164.
61
Principles I.70, AT VIIIB 34–5/CSM I 218, my emphasis.
62
Supplément 8.1, in Desgabets (1985, 197).
 THE MEDITATIONS

Once she is, it becomes obvious to her that nothing like a sensation
can exist outside the mind. As a result, there can be no ‘intelligible
resemblance’ between a sensation and something in the world.63
Given what we’ve learned of Descartes’s theory of intentional resem-
blance, we can render that last claim in a cleaner way: sensations are
not representations.
Our next task is to figure out just how Descartes believes sensory
experience happens. In the final meditation, we’ve seen our protag-
onist confess that, prior to meditating, she ‘had the sensations of light,
colours, smells, tastes, and sounds, the variety of which enabled [her]
to distinguish the sky, the earth, the seas, and all other bodies one
from another.’64 She used to believe that only the sensible qualities of
things allowed her to individuate bodies in experiences. She was, in
short, a card-carrying Aristotelian who believed she needed sensible
qualities to mark the boundaries of bodies. Although a natural enough
assumption, it is one that turns out to be false. And, oddly enough, just
as sensations are deprived of their status as ideas, images in the brain
achieve it. Or so I shall now argue.

3.3 Brain Images after L’Homme

I have argued that in his early stages, Descartes takes there to be a literal
resemblance between the brain image sketched on the pineal gland by
the animal spirits and the object in the world it represents. This image
has two jobs to do. First, as the key player in the overlap region, it has to
explain how purely mechanical perception and cognition work. Second,
it has to explain how a human mind, once installed in the machine, can
be aware of its environment. It is important to note that the mind’s
awareness of the brain is not mediated by (mental) ideas. Brain states
already ‘representent’ the objects they resemble.65 It is through the brain
state that the mind thinks out to the world, and this is enough to earn
them the title of ‘ideas.’ While this is clearly true in L’Homme, it is less

63
Principles I.70, AT VIIIB 34/CSM I 218.
64 65
AT VII 72/CSM II 52, emphasis added. AT XI 176–7, quoted above.
 THE MEDITATIONS 

clear later on. In the series of definitions offered in the Second Replies,
Descartes writes,

[I]t is not only the images depicted in the imagination which I call ‘ideas.’
Indeed, in so far as these images are in the corporeal imagination, that is, are
depicted in some part of the brain, I do not call them ‘ideas’ at all; I call them
‘ideas’ only in so far as they give form to the mind itself, when it is directed
towards that part of the brain.66

Until the final clause, this seems clear enough: from now on, Des-
cartes tells us, he won’t call brain images ‘ideas.’ The last clause takes
this back. The antecedent of ‘them’ in the text is the brain images; it
is those images that are getting called ‘ideas.’67 They merit this title in
virtue of giving ‘form to the mind,’ when the mind ‘focuses’ on them.
If we take ‘form’ in this text to be connected with representation, this
makes good sense: brain images determine or provide content to
mental states in the sense that they are the immediate object of the
mind’s attention and allow it to think of the object in the world.68 (In
the next section, I argue for this claim in more detail.)
I think we can see the forces driving Descartes both to take brain
states as ideas and to deny them this status.69 On one hand, the brain
state is corporeal and so on the official view should not count as an
idea. On the other, Descartes needs it to do the work usually assigned
to ideas, namely, serving as a representation in the internalist sense.70

66
AT VII 161/CSM II 113. Descartes offers a different view in the Letter to
Mersenne of July 1641: ‘in so far as these images are in the corporeal imagination,
I do not use that term [‘idea’] for them at all’ (AT III 392–3/CSMK 185).
67
Not everyone agrees. For example, Alanen (2003, 147) writes, ‘Descartes reserves
the term “idea” in his mature work for what ‘informs’ the mind when it is ‘directed to’
or affected by certain traces in the brain (as contrasted with the images depicted by
those traces in the brain).’
68
Thus ‘form’ in this context is not to be identified with the formal reality of ideas.
69
In his conversation with Burman, Descartes also hesitates to call a brain state an
‘idea.’ ‘When external objects act on my senses, they print on them an idea, or rather a
figure, of themselves; and when the mind attends to these images imprinted on the
gland in this way, it is said to have sensory perception’ (AT V 162/CSMK III 344).
70
There is another reason to take brain states as ideas. Stated baldly, the internalist
view is vulnerable to a regress argument: in making some state an intermediary between
mind and world, one invites the demand for a further intermediary between the original
one and the mind, and so on. To stop the regress, one needs something that directs the
mind to its ultimate object but can be immediately grasped, and not only by means of
 THE MEDITATIONS

The apparent contradiction is resolved in the second Replies.

There, Descartes treats brain states as ideas, but only when they
function for a mind as a representation of bodies. This last point is
worth emphasizing. For it allows Descartes to say that, strictly speak-
ing, non-human animals do not have ideas. The second Replies claims
that it is only insofar as these brain states are used by a mind that they
count as ideas. For in such a context they ‘give form [‘informant’] to
the mind itself, when it is directed towards that part of the brain.’71 To
give ‘form’ to the mind is to determine it to think of the object it does.
On its own, the brain state is not an idea because it is not a
representation. Why not? The problem cannot be merely that it
does not resemble the object in the world. As I have argued, corporeal
representation, when it happens, is for Descartes always a matter of
(entitative) resemblance. What bars brain states in the absence of
attention from counting as ideas is that mere resemblance is not
enough to secure representation. It is only when ‘lit up’ by the mind
that a brain state becomes a representation, precisely because it then
becomes the means by which a mind thinks of the object in the world.

3.4 The Special Modes of Thinking

With all of this in place, we can examine in more detail the sixth
Meditation’s account of sensory perception. A good place to start
is the imagination; both are ‘special modes of thinking’ that depend
on the mind-body union, and the meditator gives us a substantial
account of the imagination’s workings.
The meditator speaks of the imagination as the power to see
something as if it were present before her.72 The easiest, but hardly
the only, way to explain how that happens it so say that the same kind
of thing that happens in veridical perception happens in imagination.
Whether that is so or not, it is clearly what Descartes holds.73 So we

some further representation. This is just what brain states are supposed to do: the mind
can inspect them directly, without having to use a (mental) idea.
71 72
AT VII 161/CSM II 113. AT VII 72/CSM II 50.
73
See the ‘Conversation with Burman,’ where Descartes (reportedly) says, ‘The
difference between sense-perception and imagination is thus really just this, that in
 THE MEDITATIONS 

can learn a good deal about ordinary sensory perception from looking
at the imagination.
In the sixth Meditation, just before reviewing her former (and now
recognizably false) beliefs, the meditator considers the difference
between pure intellection and imagination. If the imagination belongs
not to the meditator’s essence but to the mind-body union, then there
is some reason to believe that physical stuff in general exists after all.74
What interests us here is just how Descartes sees the ‘special modes
of thinking’ as somehow spanning the mind-body gap. Imagination is
after all something a mind experiences, and yet it requires the presence
of a body. Imagining is a power that is in the subject but not part of the
subject’s essence.75 It is not possible for a disembodied soul; under-
standing is. In understanding, the meditator conjectures that the mind
‘turns towards itself and inspects one of the ideas which are within it.’
When intellectually grasping the chiliagon, the mind uses only its own
resources, and is not dependent on its contingent association with a
body. By contrast, as we have seen, ‘when [the soul] imagines, it turns
towards the body and looks at [intueatur, considere]76 something in the
body which conforms to an idea understood by the mind or perceived
by the senses.’77

sense-perception the images are imprinted by external objects which are actually
present, whilst in imagination the images are imprinted by the mind without any
external objects’ (AT V 162/CSMK III 345). Granted, the Conversation is not on its
own decisive evidence. But I believe it can be persuasive evidence, at least when
nothing in the rest of Descartes’s view conflicts with it.
74
As the meditator notes, ‘this is only a probability; and despite a careful and
comprehensive investigation, I do not yet see how the distinct idea of corporeal nature
which I find in my imagination can provide any basis for a necessary inference that
some body exists’ (AT VII 74/CSM II 51).
75
See AT VII 73/CSM II 51.
76
AT VII 73/AT XIB 58 (the French translation of the Meditations). ‘Intueatur’
(inf., ‘intueor’) is the same verb Descartes uses on the previous page, when the
meditator claims that to imagine a triangle is to ‘see the three lines with my mind’s
eye as if they were present before me’ (AT VII 72/CSM II 50). Though it can simply
mean ‘look at’ or ‘see,’ intueor can also carry the sense of looking closely at or
inspecting something.
77
AT VII 73/CSM II 51. CSM follows the original Latin fairly closely here; the
French translation speaks not of ‘an idea understood by the mind’ but of ‘an idea it has
formed on its own.’
 THE MEDITATIONS

Nor is this role for the brain image confined to imagination. In the
fifth Replies, Descartes takes on Gassendi’s challenge to explain mind-
body interaction in the context of sensory perception:

Here you ask how I think that I, an unextended subject, could receive the
semblance or idea of a body that is extended. I answer that the mind does not
receive any corporeal semblance [speciem corpoream]; the pure understand-
ing both of corporeal and incorporeal things occurs without any corporeal
semblance. In the case of the imagination, however, which can have only
corporeal things as its object, we do indeed require a semblance which is a
real body: the mind applies itself to this semblance but does not receive it.78

Here it is important to tease out what exactly in Gassendi’s objection

is being challenged. Part of what is at issue here is the theory of
corporeal species migrating from the eyes to the thinking subject,
which of course is a key plank in the Baconian view. Notice just where
Descartes parts ways with this theory: it is not at the stage at which a
corporeal semblance occurs in the body, but at the point where it is
supposed to be transmitted to the mind. There is no such corporeal
transmission; instead, the mind can simply ‘apply itself ’ to the cor-
poreal image. In sensory experience, the mind is hardly choosing to
do so. Nevertheless, whenever we have a thought whose object can
only be corporeal (as is the case with imagination and perception), the
mind must be aware of the speciem corpoream, which can only be an
image in the brain.
To imagine the triangle is to make it appear vividly before the
mind, almost as if it were present in sensory experience. But how?
That kind of sensory ‘punch’ has to come from the mind’s association
with the brain. Since this instance of imagination is voluntary (not all
are, of course), we have to picture the mind creating in the brain an
image that ‘conforms’ to the idea. What is this conformity? Here the
original texts are no help: both the Latin and French are cognates of
‘conform.’79

78
AT VII 387/CSM II 265.
79
In fact, they are orthographically identical in Latin and French: ‘conforme.’
 THE MEDITATIONS 

Malebranche presents an account of imagination that can be of

use here.80 When we imagine something, say, a geometrical figure,
‘we perceive it by tracing an image of it for ourselves in the brain.’ The
intellectual idea serves as a model or rule (règle) for producing
the image, and the image is accurate only insofar as it ‘conforms’ to
the idea.81 The image conforms to the idea because the idea is the
model on which the image is constructed. By contrast, in sense
perception the mind is simply aware of the speciem corpoream with-
out having played any role in its generation.
The reading I am defending is controversial, for it has no place for
the innate idea of extension in sensory experience. By contrast, many
authors have argued that even sense perception, for Descartes,
requires the deployment of the innate idea of extension.82 On my
view, the idea of extension is necessary for grasping the essence of
material things, which itself is a step on the road to proving that those
things even exist. But it is not necessary for ordinary sense perception.
In short, the idea of extension is shouldered aside by the speciem
corpoream, the corporeal idea sketched on the pineal gland.
Someone might well object that my view contradicts the second
Meditation. There, understanding, sense perception, and imagining
are equally said to be forms of thinking, and all seemingly belong to
the meditator’s essence. Even if none of what the meditator imagines
is real, ‘the power of imagination is something which really exists and
is part of my thinking.’83 How, then, can imagination and perception
depend on the existence of body? Or is the famous second Meditation
list of kinds of thinking simply wrong, and corrected in this much
later Meditation?
The developmental reading argues for the latter option. Of course,
sensation and imagination cannot be understood ‘without an intellectual

80
This passage comes in the course of his argument by elimination for the doctrine
that we see all things in God. The competitor Malebranche wishes to challenge is the
view that we produce ideas on our own, by means of the imagination.
81
SAT III.ii.3/LO 224.
82
See esp. De Rosa (2010, 128 f.) and Simmons (2003).
83
AT VII 29/CSM II 19.
 THE MEDITATIONS

substance to inhere in,’ since ‘an intellectual act is included in their

very definition.’84 In other words, being a mind is a necessary condi-
tion for taking on the modes of thought that are sensing and imagin-
ing.85 Nevertheless, as the French edition explains, these modes of
thought ‘are quite special and distinct from me.’86 They are distinct
from the meditator, since the meditator will lose them when she dies.
But they nevertheless involve an act of the mind.
A second objection comes from the role of skepticism. After all, the
first Meditation considers the possibility that there is no physical stuff
at all. But if I am right, there is no possible world in which the
meditator undergoes precisely the same experiences she would have
were she embodied and yet is not installed in a body.
The best response here is to point out that external world
skepticism need only be an epistemic, not alethic, possibility. All the
Meditations requires is that the meditator not be in a position to
rule it out. Nothing suggests that Descartes wishes to leave open the
metaphysical possibility the meditator considers at the start of her
progress. This is hardly a subtle point. God’s necessary existence and
nature by themselves rule out the massive deception such a state of
affairs would imply. So we already know that the skeptical scenario
only seemed possible, without ever genuinely being so.
A related objection is that my account threatens the proof of the
real distinction between mind and body. For if I’m right, then there is
no way for the meditator to conceive of herself, with all her current
sensory states, as existing in a world in which body does not exist. The

84
AT VII 78/CSM II 54. CSM’s translation is rather loose here. Descartes doesn’t
speak of the ‘very definition’ but of a thing’s ‘formal concept.’ The French translation
expands on this a bit: ‘Car dans la notion que nous auons de ces facultez, ou (pour me
seruir des terms de l’ecole) dans leur concept formel, elles enferment quelque sorte
d’intellection’ (AT IXB 62). (‘For in the notion that we have of these faculties, or (to
use the terms of the School) in the formal concept, they include some kind of
intellection.’)
85
By ‘intellectual act’ I take Descartes to be referring to an act of the mind. (The
French translation confirms this, since it speaks indifferently of intellectual acts and
acts of thinking.)
86
AT IX-A 62: these faculties are ‘toutes particulieres, & distinctes de moy.’
 THE MEDITATIONS 

reply here is straightforward: the proof of the real distinction requires

us to conceive ourselves as thinking, non-extended things. It does not
require us to conceive of ourselves as thinking things with precisely
these thoughts and no others. The clear and distinct understanding of
the mind is the thought of a thing that thinks, not a thing that thinks
this or that in particular.
In fact, the sixth Meditation passages we have been wrestling with
immediately follow the proof of the real distinction. Descartes makes
just the points one would expect, given his second Meditation dis-
cussion of thought: it is easy to conceive oneself without the active
faculties of sensory perception and imagination, since one will, after
the death of the body, in fact exist without them. Conversely, the
meditator cannot conceive of these faculties without some mind or
other.87 Nothing about Descartes’s view requires it to be possible, or
even all-things-considered-conceivable, that the meditator have pre-
cisely the same sensory experiences she does even when body does
not exist.
Perhaps the most intriguing objection comes from the piece of wax
example. One might argue that that text shows that the intellect is
involved even in sensory perception. The claim would be that the
intellect has to summon and deploy its idea of extension even to
perceive the determinate shape and size of the wax. Descartes does
say that ‘the perception I have of it [the wax] is a case not of vision
or touch or imagination—nor has it ever been, despite previous
appearances—but of purely mental scrutiny.’88 Unfortunately, this
text is ambiguous: it does not specify what it was about the wax
that was being perceived. As the Meditation continues, however, it

87
AT VII 78/CSM II 54.
88
AT VII 31/CSM II 21. Tad Schmaltz (2000, 68) notes this passage, but stops a bit
short, I think, of attributing to Descartes the claim that the intellect is involved in the
prior, sensory perception of the wax; Schmaltz claims that Descartes here ‘indicate[s]
that purely intellectual perception of the wax is present not only when we clearly and
distinctly understand the nature of the wax but also when we have an “imperfect and
confused” sensory perception of the wax itself.’
 THE MEDITATIONS

becomes apparent that the intellect is needed to discern the essence

of the wax, and not its fully determinate features.
For further confirmation, we can point to the denigration of the
epistemic role of the senses in the wax passage. It is telling, I think,
that the meditator asks, ‘[w]as there anything in it [her prior sensory
perception of the wax] which an animal could not possess?’89 That
suggests that the intellect is simply not necessary to the perception of
determinate features. As in Descartes’s earliest work, the brain image
alone is sufficient to explain the experience of geometrical qualities.
The meditator contrasts the sensory awareness she had prior to
meditating, when she believed she knew the wax through her senses,
with her present, intellectual awareness of the essence of the wax.
Naturally, the meditator claims her knowledge of the wax is now
more ‘perfect’ than it was before. For our purposes, the key point is
that the senses are limited to fully determinate properties: the wax has
precisely this size, shape, and even taste at one moment, and another
fully determinate complement of qualities at another. What the
intellect allows the meditator to do is to ‘take its clothes off ’ and
consider it ‘naked.’90 When stripped of its determinate properties, the
wax is revealed as a determinable, that is, something that will always
have some determinate set of properties or other, none of which is
essential to it.91
Descartes’s emphasis is clearly on the inadequacy of mere sense
perception to deliver up knowledge of the wax’s essence. But the contrast
he draws in this second Meditation presupposes that the senses by
themselves are enough to deliver experience of determinate qualities.
To extend Descartes’s metaphor, the senses alone show us at least the
clothes, though not the body underneath.

89 90
AT VII 32/CSM II 22. AT VII 32/CSM II 22.
91
Note that Descartes is not concerned in the wax passage to distinguish between
sensible and geometrical properties; that move has to await Meditations five and six.
The present point is that the essence of the wax is a determinable and as such always
inaccessible to mere sensory perception.
 THE MEDITATIONS 

3.5 Jobs for Sensations

I have been arguing that the brain image plays the role in sensory
perception that a mode of an immaterial mind plays in pure thought—
that is, it serves as a representation, directing the mind. Sensation and
imagination both involve the mind’s turning to the brain image. That
image drives out the need for a mental idea of extension, at least
where necessarily embodied experience is concerned. What is more, it
also drives out the need for sensations. If we are asking how subjects
attain ‘experiential contact’ with objects in the world, and register
their positions, shapes, distances, and all the rest, the answer will
simply be: the mind inspects the brain image.92 And as in Descartes’s
earlier work, the same motions that engrave the corporeal image give
‘the mind its signal [signum] for having a certain sensation.’93
But if sensations are not representations, what good are they? This
worry animates much of the opposition to my view. For if sensations
are mere side effects of the mind’s union with the body and especially
the brain, it is unclear what work they are doing. Even if I have
successfully argued that sensations are not representations in the
internalist’s sense, that is, that they are not really ideas at all, that

92
To my mind, this fact makes trouble for the otherwise attractive ‘fusion’ view of
Simmons (2003) and De Rosa (2010). On this view, Descartes refuses to ‘bifurcate’
experiences of sensible and geometrical qualities; these experiences are instead fused
into a single idea. As evidence, Simmons (2003, 557) cites L’Homme, where Descartes
tells us that the pineal gland’s images ‘give occasion to the soul to sense (sentir)
motion, size, distance, colors, sounds, smells, and other such qualities’ (AT XI 176).
While this does at first sight suggest that he is bundling all these qualities together in a
single fused experience, this cannot really be the case. But given what we know so far,
we are forced to say that the brain image works in two very different ways: by
prompting the mind to present sensations to itself, and by exhibiting to the mind
the object represented by the brain image. In fact, this last point undermines both the
bifurcated and fusion accounts. For both presuppose that in sensory experience we
have (mental) ideas of geometrical qualities. To posit a mental idea of size, shape, and
position on top of the corporeal idea is to multiply entities beyond necessity. As I have
argued, ‘idea’ in these contexts means corporeal idea, that is, an image drawn on the
pineal gland. And in calling these images ‘ideas,’ he is pointing to their ability to serve
as representations for the mind in their own right.
93
AT VII 88/CSM II 60.
 THE MEDITATIONS

claim is consistent with their being representations in the externalist’s

sense. Sensations might still track repeatable features of the environ-
ment. While the key elements of my view are, I think, consistent with
these tracking readings, I do not believe such readings will withstand
scrutiny.94 I first sketch my positive account of the roles of sensations
before turning to tracking theories in more detail.
The sixth Meditation suggests two roles for sensations. First, each
brain motion is paired with precisely one sensation,95 namely, the one
that best conduces to the preservation of the mind-body union.96 It’s a
bit hard to believe that each of the indefinitely (if not infinitely) many
sensations of which we are capable can play such a role: tastes like
sourness and acidity surely can, but what about the scent of lavender
or the sound of F#? In any case, this role is an epistemic one: although
our reflex reactions to some unpleasant sensations are automatic,
nothing prevents us from using sensible qualities that co-vary with
pleasurable or painful states of affairs to seek them out or avoid them.
A second role is even more clearly epistemic: sensations provide a
justification for thinking that ‘bodies possess differences correspond-
ing’ to the ‘great variety of colours, sounds, smells, and tastes.’97 But

94
Elsewhere, I have argued that Descartes is an occasionalist regarding body-body
interaction; see my (2009). If that is right, then tracking views will have a hard time
getting off the ground. But throughout this enquiry I have bracketed the whole
question of Descartes’s alleged occasionalism, so it would be inappropriate to argue
against tracking views on those grounds. And of course my own use of causal terms
regarding any body-body interaction should be understood as neutral between genu-
ine causation and mere occasioning.
95
‘[A]ny given movement occurring in the part of the brain that immediately
affects the mind produces just one corresponding sensation’ (AT VII 87/CSM II 60).
96
This is what Descartes means in the Principles II.3 (AT VIIIA 41/CSM I 224) by
saying that the senses only show us ‘what is beneficial or harmful to man’s composite
nature.’ Principles IV.198 (AT VIIIA/CSM I 284) is prima facie inconsistent with this
text, since there Descartes claims that the senses show us ‘nothing in external objects
beyond their shapes, sizes, and motions.’ The French translation of IV.198 clears this
up: there, the text reads, ‘[q]u’il n’y a rien dans les corps qui puisse exciter en nous
quelque sentiment, excepté le mouvement, la figure ou situation, & la grandeur de leurs
parties’ (AT IXB 316) (‘there is nothing in bodies that can excite a sensation in us
other than the motion, figure, position, and size of their parts’).
97
AT VII 81/CSM II 56.
 THE MEDITATIONS 

this role as an indication or sign of such states of affairs does not turn
the sensation into a representation in any interesting sense, anymore
than the width of the metal legs of the chair you might be sitting
in ‘represents’ the ambient temperature. This is an epistemic, not
intentional, role.
These considerations are not decisive in themselves. We need to
examine causal and teleological views in more detail. A causal view
runs into problems fairly quickly. If a mental state represents what
causes it, then, oddly enough, our sensations turn out to represent
brain states rather than things in the world. My argument here is not
an appeal to the traditional ‘depth’ problem: if a given mental state
stands at the end of a long causal chain, what entitles us to say that
that state represents just one node on the chain, and not others? The
problem instead is that the brain states must count as the cause, since
only they co-vary with sensations. What happens outside the brain
state varies considerably.98
A teleosemantic story can solve this problem. For what God intended
to bring about a sensation is not, or not just, a state of the brain, but of
an object in the world. Alison Simmons has cleverly defended this
position. On her view, sensations track ‘ecological’ properties, that is,
properties of objects that are relevant to the preservation of the mind-
body union.99 Descartes certainly does claim that the chief purpose of
sensory experience is to give the mind notice of what harms or benefits

98
The meditator tells us that ‘a given motion in the brain must always produce the
same sensation in the mind,’ regardless of that brain motion’s ultimate origin in the
world outside the brain (AT VII 88/CSM II 61).
99
For further criticism, see Pessin (2009) and De Rosa (2010, chapter four). De
Rosa makes heavy weather of the fact that Simmons’s ecological properties are not
intrinsic properties of res extensa. Simmons (2008), which De Rosa does not address,
clarifies her view, pointing out that such properties are relational: ‘[a]lthough they do
not have a place in Cartesian physics, properties like health, damage, and nutrition do
have a place in Cartesian anthropology, that is in the context of the mind-body union’
(2008, 106). Obviously no part of extension is intrinsically poisonous or nutritious; it
is only when considered in relation to the mind-body union that it becomes so.
I would go further and say that, in relation to non-human animals, it also makes
sense to speak of ecological properties.
 THE MEDITATIONS

the composite substance.100 But notice that my interpretation can

accommodate Descartes’s claims about the usefulness of sensations.
Against this maneuver, Simmons argues that, for sensations to play
the role of markers, they must be representations. Sensations, Sim-
mons points out, ‘provide the basis for a number of true judgments
about the corporeal world.’101 The senses tell me that I have a body,
and that there are other bodies around me.102 How could they do so,
she asks, unless they were representations?
This is the crucial question. The non-representationalist reading
can make a plausible reply, I believe. Many things, states, and acts
provide a basis for judgment in this sense without being representa-
tions. Even if mental state S is not a representation, I can still form a
true judgment about its source. Pain is a good example. Whether
Descartes holds that pain is a representational state or not is an open
question; assume for the moment that it is not and that it is a mere
blank effect. I can nevertheless judge truly or falsely about its causal
origin. It, too, would be the basis for a true judgment—‘something is
hurting me right now’—without being a representation. It is these
causal judgments that are needed to preserve the mind-body union.
Simmons, then, is quite right to say that mental sensations can be
used to infer ecological features. The mistake is to infer from this that
mental sensations represent these features. In fact, I suspect the
tracking reading has things backwards. Using x to track y presup-
poses that I already have some independent way of representing y. If
so, it would be misleading to say that x represents y; at most, x
provides grounds for an inference to y.
And there is textual evidence as well. Descartes insists in the sixth
Meditation that our sensations are caused by objects outside of us. If
Descartes held a tracking theory of representation, he would then be
entitled—indeed, required—to say that sensations in fact represent

100
‘[T]he proper purpose of the sensory perceptions given me by nature is simply
to inform the mind of what is beneficial or harmful for the composite of which the
mind is a part . . . ’ (AT VII 83/CSM II 57).
101 102
1999, 352. AT VII 80-1/CSM II 55–6.
 THE MEDITATIONS 

geometrical qualities in bodies. That he does not say so means that he

must be operating with an internalist theory of representation, such
as intentional resemblance.

3.6 Conclusion
If there has been any surprise in this chapter, it is the degree to which
the Descartes of the Meditations retains the views he first sets out in
L’Homme. True, Descartes says nothing about the theory of natural
geometry from L’Homme. That is surely due to his different purposes
in these texts. Nothing in the Meditations rules out natural geometry
as a supplementary means of judging distance and position. Nor is
there anything inconsistent with the overlap thesis. And there is
positive evidence for the continued importance of brain imagery. As
in the earlier work, the basic mechanisms of sensory perception are,
in fact, mechanisms.
That is not to say that the mind plays no role when attentive
perception is at issue. It is striking, however, how much the Descartes
of the Meditations concedes to his Aristotelian opponents. Although the
idea of extension might be required for knowing the existence and
essence of bodies, it is not required for ordinary perception, whether
attentive or not. And although Descartes dismantles the hylomorphist
ontology and with it much of the scholastics’ philosophy of perception,
he still claims that the mind ‘turns toward the body’ in imagination and
perception. That is more than a faint echo of Aquinas’s claim that the
mind turns toward the phantasms, which are purely corporeal. It is,
I have argued, the closest Descartes can come to endorsing that Aristo-
telian position without violating his ontology.
The Meditations, then, is in a sense an earlier work than its chron-
ology allows. As in the earlier work, there is no reason to think that
experiences of sensible qualities function as representations.103 In

103
I am hardly the first reader of Descartes to reach this conclusion; at most,
I have added arguments to the already impressive arsenal developed by Ann Wilbur
Mackenzie and Laura Keating, in the works noted above.
 THE MEDITATIONS

L’Homme, we saw that Descartes occasionally appeals to changes in

colors to explain how the soul judges distance, a point we’ll return to
below. But as far as I can tell, nothing about that process requires the
colors to represent distance. Using colors to judge distance in this way
in fact presupposes that one has some other way to represent the
bodies at issue and their relative positions, just as using the width of
a metal chair leg to judge temperature presupposes that one has
already succeeded in thinking about temperature in the first place.
The representational question matters for us not just in itself but
because it has ramifications for Descartes’s answers to the causal
questions of perception. As the sixth Meditation tells us, one comes
to imagine—and by extension, perceive—by becoming aware of the
brain image. There is no role for distinctively mental ideas in a
transaction where something corporeal functions as an idea. Nor, it
seems to me, is there any distinctive role for sensations.
This last point is bound up with another way in which Descartes
differs from his predecessors and successors. An influential eight-
eenth- and nineteenth-century view has it that visual experience
immediately presents us with a flat, two-dimensional mental image.
On the basis of this ‘sensory core,’ the mind makes a series of succ-
essive judgments that imbue it with three-dimensionality. In James
J. Gibson’s terms, one begins with a visual field and ends with a visual
world.104 If Hatfield and Epstein are correct, even some of Descartes’s
predecessors and near-contemporaries hold this view.105
So we have another way to see just how radical Descartes’s view is:
he rejects the whole notion of a sensory core. There is no suggestion
that the mind begins with a two-dimensional mental image and fills
it out by making judgments. First, the image the mind immediately
considers is not a mental image at all; it is a purely corporeal image

104
See Gibson (1950), cited in Hatfield and Epstein (1979, 363).
105
See Hatfield and Epstein (1979). Hatfield and Epstein discuss the Descartes of
the sixth Replies, where it certainly does seem as if he has a notion of a sensory core.
But they take this feature of Descartes’s view to be constant throughout his corpus,
whereas I do not.
 THE MEDITATIONS 

in the brain. Second, that image is, as we have seen, already three-
dimensional. And even where natural geometry does bring in
judgments of distance, they are not initially responsible for the
distance represented in one’s visual field. That feature is given in
the brain image; natural geometry simply allows the soul to judge
distance more accurately. Unlike the vast majority of other thinkers
on the causal mechanisms of visual perception, the early Descartes
has no truck with a sensory core.
 4
The Dioptrique

Even though its composition antedates the Meditations, the Dioptri-

que in fact departs from that text and Descartes’s early view in
general. (Why this should be so is a question I turn to at the end of
this chapter.) One hallmark of the early Descartes’s view is the
richness and exactitude of the brain image. It is by being aware of
this image that the mind comes to be aware of its environment.
Minds, of course, can apply natural geometry and achieve more
accurate representations of their world. But the brain image itself is
pretty good; indeed, it must be, if unensouled machines and inatten-
tive humans get by on it alone.
The key to explaining the shift in Descartes’s views, I think, is to see
that he loses faith in the ability of the brain image on its own to serve
as a representation, that is, as a (corporeal) idea. As a result, he has to
attribute more and more to the mind rather than the body. The
inverse relationship is simple and inescapable: the more sophisticated
and powerful the mechanical explanation is, the less need there is for
sophisticated and powerful mental processing. Conversely, the less
work the brain image does, the more impressive the human or divine
judgments that substitute for it must be.1
Demoting the brain image, and the mechanical account of percep-
tion generally, is not without its costs. For one thing, the early view
has an elegance that is lost. The prospects of giving a single mechanical

1
My thoughts here are inspired by a point Richard Sorabji (1993) makes in a
different context. Sorabji in effect shows that, as one denies animals thought and
reason, one is compelled to enrich their perceptions to make up for the gap.
 THE DIOPTRIQUE 

explanation for animal and inattentive human behavior are now

vanishingly small. If distance perception in attentive humans, for
example, requires a mind, then whatever is going on with inattentive
humans and animals must be very different indeed. The new account
of perception, in short, undermines the overlap thesis. What is worse,
while Descartes can retain his bête machine doctrine, his story about
human perception makes one suspect that in principle there is no
substantive account of animal and inattentive ‘perception’ to be had.
This is all the more disappointing since, as the early Descartes points
out, humans and some animals share many of the same behaviors and
physiological structures. And trivially, attentive and inattentive
humans are physiologically identical.
Being unable to rely on the brain image as he had in other works,
Descartes makes two important moves. First, he relies on what had
been the merely supplemental theory of natural geometry. Second, he
introduces a purely causal account, whereby some brain motions are
ordained by nature not just to trigger sensations but to bring about
ideas of geometrical qualities as well. These two moves will bring with
them a host of problems that will bedevil not just Descartes himself
but his successors. We should begin by seeing just how the Dioptrique
pushes the brain image to the side.

4.1 Dethroning the Brain Image

To a reader of L’Homme and the Meditations, the start of Discourse
Five of the Dioptrique comes as something of a shock. There, Descartes
announces that ‘the soul does not need to contemplate any images
resembling the things which it perceives.’ The claim is ambiguous: is
Descartes saying that the mind doesn’t need to contemplate an image
that resembles the object, or that it doesn’t need to contemplate any
image whatsoever, or both? To see what Descartes is rejecting, we need
some more context.
Descartes clearly has two targets in his sights. The first is the
traditional view that corporeal species are transmitted from objects
to the brain, where they can be contemplated by the soul. This is not
 THE DIOPTRIQUE

surprising: we’ve already seen the earliest Descartes (in the Regulae)
arguing that nothing numerically identical passes through the optic
nerve to the brain. Descartes here adds that the image in the brain
need not resemble its object in all respects, a point which is close to
trivially true, since if it did, it would just be its object.2 His second
target is the claim that we perceive objects by perceiving images in the
eye.3 The images in the eye often include ‘only ovals when they make
us see circles,’ a feature that is also seen in engravings.4
But all of this leaves us with a question: just how defective does
Descartes think the images in the eye (and brain) really are? The
text seems equivocal and has produced a variety of readings. Prob-
ably most commentators read the Dioptrique as denying that either
brain image or retinal image resembles the object that caused it.
Richard Watson argues for just the opposite claim: he thinks that
Descartes never retreats from the view that the brain image resembles
the object.5
Consider that, according to Discourse Four, only a little resem-
blance is required in order for something to serve as a representation
of its object, as we can see in engravings. Why, then, does Discourse
Six take the fact that retinal images only contain ovals when they
represent circles as evidence that we don’t judge objects by means of
those images? If we take the lesson about engraving to heart, surely we
ought to draw just the opposite conclusion: namely, that the retinal
image resembles its object enough for us to be able to see it. What is
more, the intervening Discourse (Five) follows up the rejection of the
mind contemplating resembling images with this: ‘And yet, for all
that, the objects we look at really do imprint quite perfect images of
themselves on the back of our eyes.’6

2
As Descartes himself argues; see AT VI 113/CSM I 165.
3
See AT VI 140–141/CSM I 172.
4
AT VI 140–141/CSMI I 172; for the claim about engravings, see AT VI 113/
CSM I 165.
5
See Watson (1995).
6
AT VI 114/CSM I 166. Descartes of course goes on to list some of the retinal
image’s defects. ‘The primary and principal’ such fault is that ‘no matter what shape
the parts of the eye may have, it is impossible that these shapes should make the rays
 THE DIOPTRIQUE 

In fact, I think we can extract two rather different arguments

against Descartes’s earlier view. The Dioptrique argues that retinal
images cannot be the means by which we judge size and shape. The
argument appeals to size constancy: very close objects imprint a larger
image on the retina than do those that are far away, and yet we do not
see the close objects as larger in size. A parallel point may contribute
to Descartes’s abandonment of his early view, where the brain image,
rather than the retinal image, is at stake. The early Descartes has no
mechanical means of accounting for size or color constancy.7 An
approaching object takes up more and more of our visual field, and
is often subject to changes in color and shading. Nevertheless, we do
not experience it as getting bigger or as shifting in color. Even when
undergoing inattentive perception, we are not usually vulnerable to
illusions of size, as can be seen through our behavior. And of course
animals do not behave as if the sizes of things depended on their
distance from them.8
It is easy to see why size constancy would be a particular problem
for the early Descartes. The inverted bas-relief brain image makes a
certain amount of sense, if we are considering a static situation in
which the machine has to ‘recognize’ objects by their shapes. We
even saw how Descartes thinks distance can be included in the brain
image. What happens, though, when we put the scene in motion? Ceteris
paribus, the closer the object, the larger the image. And Descartes
suggests no way of mechanically correcting the brain image in order to

that emanate from diverse points assemble together converge in as equally many
diverse points’ (AT VI 121). The other defects ‘consist in the fact,’ which Kepler had
demonstrated, that ‘the parts [of the image] are inverted, that is to say, in a position
totally contrary to that of objects’ (AT V 123). While both are serious problems for the
retinal image, both would be corrected for in the brain image, if the processes outlined
in L’Homme were allowed to operate. So neither can help explain why he abandons
the early view.
7
See Wolf-Devine (1993, 64 f.).
8
One wonders just how such a deficit would manifest itself in behavior. It would
certainly seem that a mistaken honey bee reporting on the richness of a food source
through its waggle dance would be prone to exaggeration and fall out of favor with its
fellows.
 THE DIOPTRIQUE

achieve size constancy. So the Dioptrique view does not appeal to the
brain image at all; instead, when humans attentively perceive their
environment, they have to use the ‘knowledge or opinion’ they have
of the object’s distance, not just the retinal image, to judge the other
geometrical qualities.9
If this is Descartes’s reasoning, it is not made explicit. The only
explicit argument I can find against taking vision to proceed through
the brain image comes at the start of Discourse Six. Descartes con-
cedes that there is some resemblance between the brain image and the
object in the world. But he warns that ‘we must not think that it is by
means of this resemblance that the picture causes our sensory per-
ception of these objects—as if there were yet other eyes within our
brain with which we could perceive [the brain image].’10
This is a curious kind of argument. Descartes does not deny that
there is a brain image; it is a side effect of the very same brain motions
that, as we shall see, prompt sensations. Nor does the image fail to
resemble, at least to the degree that some engravings and other art
works resemble their objects. The objection here seems to be that the
brain image story is locked in a vicious regress. We begin by positing a
little image for the soul to look at. Through this image, it sees objects.
But why think the brain image is any easier for a disembodied soul to
experience than the objects that caused that image in the first place?
Wouldn’t we need yet another set of eyes to perceive the brain image?
But note that this argument would apply equally well to all of the
bodily states and properties we will find Descartes appealing to in
the text that follows. Descartes says, for example, that we judge
distance partly by the shape of the eye. Why couldn’t this same
argument be deployed against Descartes’s new position? Why not
say that that view is silly, since we would need another pair of eyes to
see the shape of our own eyes? In short, the shape of the eye seems no
more or less difficult for a soul to experience than a brain image.
Moreover, the pineal gland at least has the advantage of being the
‘seat of the common sense’11 and the ‘location’ of the soul.

9 10 11
AT VI 140/CSM I 172. AT VI 130/CSM I 167. AT VI 129.
 THE DIOPTRIQUE 

The argument seems to assume that the only way for the soul to
experience something in the brain is through vision. Without this
assumption, the regress never gets going. But the earlier Descartes
would simply reject that assumption; the soul, as I have argued,
perceives the image engraved on the pineal gland through a kind of
internal touch, not unlike proprioception. You do not need another
pair of eyeballs in your brain to experience the image traced on the
pineal gland. Nor, I think, can the Descartes of the Dioptrique con-
sistently believe that you would, for he is about to help himself to
a number of processes that require the mind to be aware of the
brain and eye.
To return to the puzzle with which we began: what exactly does
Descartes mean to reject? It now seems clear that he rejects the notion
of the mind turning toward the brain, where it contemplates an
image. But his explicit arguments show at most that you do not
need to perceive an exactly resembling image in the brain in order
to experience a material object. They do not, it seems to me, refute his
earlier view. If we need to reconstruct an argument on his behalf, it is
that the brain image alone seems insufficient to account for size and
color constancy.
In any case, Descartes deprives himself of one means of being
aware of external objects, namely, by being aware of the brain
image. Let us begin our investigation of his replacement by looking
at the new treatment of the proper sensibles before advancing to the
common ones.

4.2 The Proper Sensibles

Rather than appeal to a picture in the brain, Descartes tells us we
must hold that ‘it is the movements composing this picture which,
acting directly on our soul in so far as it is united to the body, are
ordained by nature to make it have such sensations.’12 This move is
a genuine change from the semiotic account of Le Monde, which

12
AT VI 130/CSM I 167.
 THE DIOPTRIQUE

required the mind to decode the brain’s signs, which come to it in

the form of motion.
In the case of vision, there are only six principal qualities: ‘light,
colour, position, distance, size, and shape.’13 Light and color are the
only qualities ‘belonging properly to the sense of sight.’14 Awareness
of these proper sensibles is explained purely by the institution of
nature, which dictates that brain motion x cause sensation y. Here,
there is no need whatsoever for the mind to be aware of motions in
the brain. The process is purely automatic.
The shift to automation is most clearly exhibited in Descartes’s use
of the linguistic analogy. Recall the objection Descartes considers in
La Lumiere. There, the objector wonders just how a mind can under-
stand the ‘words’ of the brain. The early Descartes stuck to his guns.
Such a quick, perhaps even barely conscious, inference is no different
in kind from our natural ability to ‘read’ facial expressions as signs of
joy or sorrow.
In the Principles, Descartes gives the same linguistic analogy and
faces precisely the same objection. His reply, however, could not be
more different:

It may be objected that speech or writing does not immediately excite in the
mind any emotions, or images of things apart from the words themselves; it
merely occasions various acts of understanding which afterwards result in the
soul’s constructing within itself the images of various things. But what then
will be said of the sensations of pain and pleasure? A sword strikes our body
and cuts it; but the ensuing pain is completely different from the local motion
of the sword or of the body that is cut—as different as colour or sound or smell
or taste. We clearly see, then, that the sensation of pain is excited in us merely
by the local motion of some parts of our body in contact with another body; so
we may conclude that the nature of our mind is such that it can be subject to
all the other sensations merely as a result of other local motions.15

There is no inference to be made: when x happens in the brain, it

causes sensation y in the mind. Descartes has replaced the picture of

13 14
AT VI 130/CSM I 167. AT VI 130/CSM I 167.
15
AT VIIIA 321/CSM I 284.
 THE DIOPTRIQUE 

the mind decoding the brain’s signals with a brute causal picture. As
Descartes puts it in the Dioptrique itself: the same motions that
engrave the image on the pineal gland ‘are ordained by nature to
make it [the soul] have . . . sensations.’16
If we stay at the level of the proper sensibles in the Dioptrique, this
causal model simply replaces the early semiotic one. One might
expect Descartes to stick to the same brute causal story when it
comes to the improper objects of vision. As we shall see, matters are
not so simple.

4.3 Geometrical Qualities

The Descartes of the Dioptrique presents us with two ways to explain
visual experience of geometrical qualities: a purely causal one, insti-
tuted by nature, and a more intellectual one. The former has no use
for awareness of the brain or eye; the latter, it turns out, demands it.
The automated and inferential processes are not mutually exclusive.
Each kind of process can happen at once. And they can be even more
intimately intermingled, as when we use color and light sensations,
triggered by brain motions, to judge relative distances.17
Descartes thinks that the awareness of distance and position are
logically prior to that of size and shape. So it makes sense to begin with
these. Descartes offers at least four different explanations for distance
perception. Let me briefly lay them out, noting whether they cause (by
the ordination of nature) an awareness of a common sensible, or form
the basis for judgment about it (by natural geometry).

1) the changing shape of the eye and corresponding brain change

(ordained by nature)
2) the relation of the eyes (natural geometry)

16
AT VI 130/CSM I 167.
17
‘And when we look at two ships out at sea, one smaller than the other but
proportionately nearer so that they appear equal in size, we can use the difference in
their shapes and colours, and in the light they send to us, to judge which is the more
distant’ (AT VI 140/CSM I 172).
 THE DIOPTRIQUE

3) the changing position of a single eye (natural geometry)

4) the distinctness or indistinctness of the shape plus strength of
the light (natural geometry)18

Only one of these, 4), appeals to our awareness of a sensible quality.

Taken together, they show Descartes availing himself both of purely
automatic processes and the mind’s ability to perform sub- or un-
conscious inferences.
Let us consider 1). In order to focus on an object as it moves toward
or away from us, our eyes must change their shape. When they do so,
a corresponding part of our brain changes as well, ‘in a manner that
is ordained by nature to make our soul perceive this distance.’ This
ordinarily happens ‘without our reflecting upon it.’19 This is the
causal story. Note that there is no attempt at real explanation here:
just the stipulation of a brute connection between two events, one in
the brain, one in the soul.20 The same causal process is also invoked to
explain our awareness of position. Changes in the minute parts of the
brain cause changes in our experience of our own body and of its
position relative to others in its visual field.21
Descartes, however, wants to retain the natural geometry of his
early view. In this process, the mind performs ‘a kind of reasoning
quite similar to that used by surveyors.’22 In the case of binocular
distance vision, Descartes appeals to the analogy of the blind man
with his sticks. He can use the resistance each stick meets to detect
bodies and calculate their distance from him and among each other.
Descartes does not waver from the claim that we are aware of our
bodies without the intervention of mental ideas. Just as the blind man
knows the relation of his two sticks, so we must know the relation of
our eyes to one another.
Once we have distance and position, we find we have already
achieved awareness of size and shape. The early account of shape
perception in L’Homme is very straightforward: an object traces an

18 19
AT VI 137–40/CSM I 170–2. AT VI 137/CSM I 170.
20 21
AT VI 137/CSM I 170. AT VI 134/CSM I 169.
22
AT VI 138/CSM I 170.
 THE DIOPTRIQUE 

image on the pineal gland with a shape that ‘corresponds exactly to its
[the object’s] own.’23 By contrast, in the Dioptrique, we are said to
judge (‘juger’) shape based on the knowledge or opinion we have of
the distance of an object, compared with the size of the images
imprinted on the back of the eye.24 Without a judgment about the
relative distances of the parts of an object, there is no way to tell
whether one is seeing the true shape, as one would if it were seen
straight on, or seeing a different figure as a result of the object’s being
tilted relative to the subject. Looking at a penny held at angle, I see an
ellipse; but if I judge that some parts of the penny are more distant
from me, I judge that the penny is in fact circular.
The perception of the proper sensibles has been given over to the
institution of nature. And what is more important, perception of the
common sensibles (distance, position, size, and shape) now requires a
mind. That mind is either undergoing the ideas ordained by nature,
or making judgments about the objects perceived, or both. If so, either
animals are not able to ‘perceive’ those common sensibles in any
sense, no matter how weak or metaphorical, or else they do it in a
completely different way, since they lack the advantage of a mind. But
again, given that animals do not stumble around blindly, they must be
accomplishing, with their mere meat-minds, the same kind of thing
we do with our exalted souls. The same is true of human ‘perception’
when we are not attending to our environments. It follows that there
must be a very different, completely corporeal story about vision that
does not require the activity of a mind. This duplication of methods is
the chief cost of abandoning Descartes’s earlier position.
Two problems with the present account merit discussion. Obvi-
ously, Descartes has done nothing in the Dioptrique to square the
unconscious inferences of natural geometry with the transparency of
the mental. As in his explication of natural geometry in L’Homme,
Descartes remains maddeningly opaque on the degree to which
inference, conscious or otherwise, is involved. In fact, as Celia
Wolf-Devine points out, the terms Descartes chooses for this process

23 24
AT XI 159; see Wolf-Devine (1993, 80 f.). AT VI 140/CSMI 172.
 THE DIOPTRIQUE

seem designed for obfuscation.25 At various times, Descartes uses

verbs like juger, apperçevoir, voir, connaître, and savoir, a fluidity
which seems designed to confuse the reader. Descartes retains, with-
out right, the Baconian’s deliberately elastic way of describing the
perceptual process.
Second, there is a problem internal to the Dioptrique. It is very hard
to see how Descartes proposes to explain the localization of qualities.
We have three different kinds of processes running, in any case of
attentive vision: the automatic and quasi-inferential processes that
generate experience of the common sensibles, and the automatic
process that produces our experience of sensible qualities. How is it
that the color experiences ‘ordained by nature’ to follow from this or
that jangling of the nerve origins in the brain match up perfectly with
the surfaces of bodies? Moreover, why do we see colors as being
outside of us at all? These are questions to which the Dioptrique
provides no answers.
These two problem areas—the status of unconscious inferences
and the localization of sensible qualities—might well be behind
some of the equally dramatic developments we are about to see in
Descartes’s philosophy of perception. There is one important point of
continuity with the earlier work: the Dioptrique does not have the
concept of a ‘sensory core.’ The ‘inputs’ for sensory awareness are
almost all physiological states. To that extent, and no further, the
Dioptrique is in agreement with Descartes’s earlier account.
We cannot leave the Dioptrique without confronting the chrono-
logical question: if Descartes has already abandoned the view of Le
Monde by 1637, why does he hark back to that view in the Medita-
tions of 1641? As he sits to compose the sixth Meditation, to say
nothing of the second Replies, he has already given up on the doctrine
of conversio. Why include it in the Meditations at all?26
Descartes doesn’t tell us, so the best we can do is speculate. First we
have to note the very different motivations behind these works. Just as

25
As Wolf-Devine (1993, 73) notes.
26
I am grateful to an anonymous referee for pressing this question.
 THE DIOPTRIQUE 

the Dioptrique cheerfully ignores the whole problem of squaring

natural geometry with the official definition of thought, so the Medi-
tations shuns any sustained discussion of the details of perception.
Motivations matter in another way as well. I have emphasized the
scholastic pedigree of Descartes’s conversio doctrine. Given that
one of Descartes’s primary goals in the Meditations is converting
the scholastics to his new philosophy, it is not surprising that he
would include his earlier conversio doctrine and pass over his more
recent thoughts in silence.
Descartes’s views are very much in flux in this middle period. For,
as I am about to show, the sixth Replies offers yet a third approach to
sensory perception, one that can be squared neither with the Diop-
trique nor with the text of the Meditations itself. Any attempt to make
these three views consistent would necessarily involve taking Pro-
crustean measures.
 5
Later Descartes

So far, we’ve seen two quite different pictures of sensory experience

in Descartes. In this chapter, we’ll consider two more. To review:
Descartes’s earliest work posits immediate awareness of a fairly
accurate brain image to account for attentive sensory experience of
geometrical qualities. On top of that story, we find natural geometry
operating seemingly just out of the reach of awareness. Natural
geometry supplements the deliverances of conversio but does not
supplant it. Our awareness of sensible qualities, by contrast, is a
matter of the mind’s decoding the motions in the brain and sum-
moning the appropriate sensation as needed.
A second view appears in the Dioptrique. In this picture, the mind
has to be aware of what’s happening in the eyes and brain, but no
special role is given to the corporeal image. The brain image is a mere
mechanical side effect, not an idea, not a representation. In its place
is a portmanteau account of our awareness of geometrical qualities:
a purely causal one, where the ordination of nature just makes us
perceive those qualities, and a quasi-inferential one, where the mind
is performing a series of judgments. The brute causal story is also
applied to the sensible qualities.
Neither picture has an indispensable role for these sensible qual-
ities; they seem largely epiphenomenal. At most, variations in color
provide an extra way of judging distance. This is a startling fact. As far
as I can tell, nearly everyone in the Baconian and Aristotelian tradi-
tions agrees with Aquinas: ‘the only time sight apprehends size or
shape is insofar as it apprehends something colored.’1 The common

1
Aquinas, CDA II.13, 205.
 LATER DESCARTES 

sensibles are supposed to be common to multiple senses; they reveal

themselves through the proper objects of each sense. By contrast,
Descartes’s geometrical qualities are revealed through the mind’s
experience of the brain and eyes to which it is attached. And neither
of the pictures we’ve looked at has room for a ‘sensory core,’ a two-
dimensional image that is enriched through judgment and inference.
As we investigate Descartes’s later work, it is worth keeping in
mind the inverse relationship between the adequacy of the brain
image and the mental work that needs to be done. If we, like the
early Descartes, take the brain image to be largely in good order (and
certainly adequate to the needs of inattentive humans), then the work
of the attentive mind can largely be confined to becoming aware of
that image. Sophisticated judgments and inferences can be brought to
bear, of course, and can produce a more accurate mental picture of
the environment. But once one deprives the brain image of its role in
perception, the mind (whether ours or God’s) will be called upon to
do more of the work. The Dioptrique tries to do it through natural
geometry and the ordination of nature. The sixth Replies, as we’ll see,
takes a different tack, demanding that the mind judge geometrical
qualities on the basis of sensations and bodily states.

5.1 How to Paint Extension

We can now turn to the sixth Replies and its famous three grades of
sensory response. I am doing so only after discussing the Dioptrique
for two reasons: unlike the body of the Meditations, this reply expli-
citly refers to the Dioptrique; and, more interestingly, it represents a
departure from both the Dioptrique and the Meditations.
I have argued that the Meditations treats our awareness of the
geometrical qualities of bodies as largely passive; it’s a matter of
the mind’s becoming aware of the brain. When that happens, the
mind’s sensations are also triggered. These sensations play epistemic
roles, which help preserve the mind-body union; but nothing about
these roles requires them to be representations, even if in some cases
they provide a further means of judging distance.
 LATER DESCARTES

The sixth Replies offers a radically different view, with problems of

its own. Some commentators put it at the heart of their interpretation,
reading all other texts in its light; others take it to be an aberration.2
I suggest we approach it on its own terms, making connections with
other texts where we can but allowing it to speak for itself.
In these Replies, Descartes is responding to an objection about
epistemology. The objectors claim that the senses are self-correcting
and don’t need any help from a Cartesian intellect. In the case of the
stick bent in water illusion, we are able to correct our earlier impres-
sion simply through further experience from another sense modality,
namely touch.3 What is at stake is just what degree of justification is
conferred by sensory processes alone. To sort this out, Descartes will
go on to explain what is happening at each grade of sensory experi-
ence. His reply, in short, is that there is no certainty to be had at the
first two grades, simply because they lack the propositional structure
necessary for either truth or falsity.4 What his objectors attribute to
the senses—our awareness through touch of the stick as remaining
straight as it is immersed in water, for example—in fact should be
referred to the intellect.

[The] movement in the brain, which is common to us and the brutes, is the
first grade . . . This leads to the second grade, which extends to the mere
perception of the colour and light reflected from the stick; it arises from
the fact that the mind is so intimately conjoined with the body that it is
affected by the movements which occur in it. Nothing more than this should
be referred to the sensory faculty, if we wish to distinguish it carefully from
the intellect. But suppose that, as a result of being affected by this sensation of
colour, I judge that a stick, located outside me, is coloured; and suppose that
on the basis of the extension of the color and its boundaries together with its
position in relation to the parts of the brain, I make a rational calculation
about the size, shape and distance of the stick: although such reasoning is
commonly assigned to the senses (which is why I have here referred it to the

2
Gaukroger (2002, 200 f.) is a good example of the latter tendency, Simmons
(2003) of the former, though she reads the sixth Replies in light of the Dioptrique (see
esp. 2003, 558 f.).
3 4
AT VII 418/CSM II 282. AT VII 438/CSM II 295–6.
 LATER DESCARTES 

third grade of sensory response), it is clear that it depends solely on the

intellect.5

What happens at each grade is one of the most disputed topics in all
of Descartes scholarship. We can isolate two areas of controversy.
First, there is a debate over just how, and in what sense, the third
grade is intellectual: some commentators take it to be a matter of a
priori reasoning, or of the deployment of the innate idea of extension,
or the development of justified beliefs.6 Second, there is the question
of where to put the mind’s awareness of geometrical properties: is it
confined to the third grade, or is it already present in the second?7
Before we can answer these questions, we should first ask just what
Descartes is taking over from the Dioptrique. He refers us to that text
on two points: first, that there are no species migrating into the brain,
but only motions that travel through the optic nerve; and second, that
reasoning alone allows us to work out any one of size, shape, and
distance on the basis of the other two. The first point is unexceptional,
and is consistent throughout Descartes’s work all the way back to the
Regulae. The second points us toward the doctrine of natural geom-
etry. But although this Reply and the Dioptrique both invoke judg-
ment, the raw data on which that judgment is exercised is importantly
different.
Let me reconstruct, as carefully as possible, just what happens at
each of the three grades.

1. Rays of light bounce off the stick and create motions in the optic
nerve that are then transmitted to the brain. (This is common to
us and animals.)
2. Given the mind-body union, we undergo ‘the mere perception
of colour and light.’ This is all ‘the sensory faculty’ is capable of.

5
AT VII 437–8/CSM II 295; emphasis mine.
6
In order: Nancy Maull (1978), Secada (2000, 132), Atherton (1990, 30). I am
indebted to Simmons’s (2003, 555) taxonomy.
7
Simmons (2003, 555) reads the literature as converging on the claim that it is only
at the third grade that ‘primary’ qualities appear, while she defends the opposing view
that such properties are already on the scene in the second grade.
 LATER DESCARTES

3.1 As a result of undergoing this sensation of color, we judge that

a stick, which is outside of us, is colored.
3.2 On the basis of the extension and boundaries of the color,
together with the position of the color relative to parts of the
brain,8 we calculate the size, shape, and distance of the stick.

Note that at stage 2, Descartes deliberately omits the common

sensibles. As in the Dioptrique, the proper objects of vision are merely
color and light, not size, shape, or distance. This is strong evidence
that, pace Simmons and others, the second grade does not present us
with determinate shapes. This is an important point that is easily
missed if we run together steps 2, 3.1, and 3.2. It’s true, as Simmons
notes, that color itself is said to have extension.9 But this does not
entail that the color as it is present in stage 2 is extended. It gets that
extension only after the process described in 3.1.
Somehow, we have to move from an awareness of light and
color that does not itself have or represent shape and the rest to an
awareness that does. Note that Descartes still does not accept the
notion of a sensory core, pace Hatfield and Epstein. The sensory core
is supposed to present the world in two dimensions, only needing to
be filled out with distance to generate three-dimensional experience.
The raw material for the Cartesian mind now is an unshaped expanse.
By contrast, the Dioptrique insists that the mind calculates dis-
tance, shape, and size on the basis of its awareness of parts of the
brain and the eyes. (Only in the case of the blurred image as a guide to
distance does Descartes suggest that our awareness of sensible qual-
ities has any role to play.) What accounts for this change?
Part of it must be due to the dialectical context. His interlocutors
have exaggerated what mere brains and passive sense perception can
do. Neither the first nor second grades presents the world as being
thus-and-so. Just after the quoted passage, with its three grades,

8
The French translation makes it clear that it’s the position of the color, not the
extension or the stick, that is judged relative to the brain parts; see AT IX-A 237.
9
Simmons (2003, 557).
 LATER DESCARTES 

Descartes reminds the objectors that when he claimed ‘the reliability

of the intellect is much greater than that of the senses,’ he did not
mean that the senses deliver up one picture of how the world is and
the intellect another. To talk about the reliability of the senses is just
to talk about the truth or falsity of ‘the judgments we formed without
any reflection in our early childhood.’10 The senses don’t, strictly
speaking, judge anything. Only at the third grade does the possibility
for truth or falsity appear.11 What his objectors have not understood
is that the senses themselves cannot be a source of certainty because
they do not deliver up states with truth values at all.
At stage 2, we have an unshaped, unsized, unpositioned array of
light and color. It’s very hard to know what to make of this. It
certainly doesn’t seem available to introspection, though of course
Descartes has an easy answer to that: we are so accustomed to making
stage 3 judgments that it is all but impossible to think them away. On
the basis of this sensory experience, we judge (3.1) that a stick, located
outside the body, is colored. This is the key step, for it takes us from
shapeless visual sensations to an experience of a colored object
outside of us. Just how it happens is obscure, and we’ll return to it
below. For now, we should note that it is judgment that gets us an
experience as of physical objects.12 The raw material of sensation does
not, of itself, represent anything at all outside of us. We have to take
that raw material and project it on to objects in order to see that
material as outside of us.
There is a complication we should pause over here. On Descartes’s
official view, judgment is only responsible for assent, dissent, or the
suspension of either. As an aspect of volition, judgment allows us to
commit ourselves or not to a proposition or its negation. The

10
AT VII 438/CSM II 295.
11
‘It is clear that we are not here [in the context of speaking about the reliability of
the senses and intellect] dealing with the first and second grades, because no falsity
can occur in them’ (AT VII 438/CSM II 295–6). CSM inserts ‘of response’ after ‘first
and second grades,’ but nothing in the Latin corresponds to this interpolation.
12
A point Secada emphasizes: ‘[t]he representative function of the mind relative to
things outside itself is the result of judgment’ (2000, 92).
 LATER DESCARTES

construction of a proposition out of individual ideas is, then, left to

the understanding.13 Strictly speaking, then, the understanding is
doing the projecting, and the will is judging the resulting proposition,
namely, that the stick is outside of us and bent. However the work of
proposition construction and affirmation is divided up, however,
Descartes’s point is clear: constructing the proposition and affirming
it are simply not the jobs of the sensory faculty. All the passive faculty
can do, in the case of vision, is receive the impressions of light and
color. Except in contexts where the distinction matters, I’ll speak of
the construction and affirmation of the relevant propositions under
the blanket heading of judgment.
Step 3.1 comes into sharper focus when we contrast it with the
scholastic picture. Unlike the scholastics, Descartes thinks the aware-
ness of, say, shape is a result of mental activity, not already contained
in our awareness of the proper sensibles. The proper sensibles are
mere sensations, which do not, of themselves, represent anything in
the world. The mind has to paint them on to the geometrical prop-
erties of bodies in order to make those properties visible.
So at step 3.1, we as it were paint a region of mind-independent
extension with colors. By being applied through judgment to the stick,
the color only now takes on a determinate size and shape. Rather than
simply finding that color and shape are co-terminous, as the scholas-
tics would have it, the mind has to make the two coincide. The
procedure is a bit like throwing a sheet over an invisible couch: by
using something foreign to the object, one is able to determine its
outlines. 3.1, then, gets us a color that coincides with a region of
extension. The next step is to judge size, shape, and position by
comparing the extension of the color to the parts of the brain.

13
See Notae AT VIIIB 363/CSM I 307 for a clear discussion of the difference
between judgment and understanding. Some Cartesians, especially the Port-
Royalians, use ‘judgment’ in a very different way, to mean the uniting of ideas into
a proposition (see Loqique II.2: 79). In terms of my (2002) and (2004, 44 f.), the Port-
Royalians’ ‘judgment’ is sub-propositional, while Descartes’s (official) use is super-
propositional. Even those who do not accept my reading of the Cartesians on
judgment can recognize the sub-propositional use of ‘judgment’ in Arnauld and
Nicole; see, e.g., Jill Vance Buroker (1994).
 LATER DESCARTES 

Step 3.2 has rightly puzzled many commentators. It is bad enough

to think of a sensation as having a determinate extension. Moreover, it
is hard to see how we could compare a feature of a sensation to parts of
the brain. Even if we are somehow aware of the parts of the brain, the
comparison is of apples and oranges. Finally, as Celia Wolf-Devine
argues, the present story seems viciously circular: it presupposes that
we are aware of the parts of the brain. But being aware of them is no
different in kind from being aware of the common sensible properties
of objects outside the body.14
Although I agree that the sixth Replies’ account is problematic,
matters are not quite so dire. First, Descartes can appeal to the mind-
body union to explain why we can have an unmediated awareness
of the brain and its parts and not of things outside the body. The
account, then, is not viciously circular. Second, it is not the sensation
itself that has extension. Only the stick has extension. One might try
to argue that the sensation has extension objectively, in that it repre-
sents the extended thing as colored.15 But the color sensation is not
intrinsically of the stick; it has to be applied to that region of exten-
sion. The color is extended only in that it is judged to be co-terminous
with the stick.
Finally, what sense does it make to compare an image of colored
extension with parts of the brain? Happily, Descartes is not telling us
to compare an image with part of a brain. For at this stage, it is no
longer the image but the stick itself that is the object of the compari-
son. Since 3.1 already got us an awareness of colored extension
outside of us, we can use that mind-independent (though not really
colored) extension as one term of the comparison. If so, Descartes can
run part of the account he offered much earlier, in L’Homme. There,
he claims that the closer some of the points on the surface of the
pineal gland are to the center of the brain, the more distant must be
the places on the seen object that correspond to them.16

14 15
See Wolf-Devine (1993, 89). See Simmons (2003, 569).
16
See AT VI 183/G 155.
 LATER DESCARTES

All of this leaves the question that has exercised commentators:

how can the sophisticated process set out in 3.2 be taking place
without our noticing it? On some views, what we have here is not
really a matter of judgment at all, but something sub-intellectual, sub-
judgmental.17 In fact, Louis de la Forge anticipates this contemporary
view, attributing such ‘judgments’ to the imagination, rather than the
intellect.18 On Simmons’s view, these ‘sensory judgments’ happening
at the third grade cannot be ordinary Cartesian judgments, because
they are not beliefs and do not have a truth value.19 Nor are they
voluntary and resistible, in the way that official Cartesian judgments
might seem to be.
Although this is a very attractive view, I doubt Descartes can
embrace it. Here it is important to keep the context of the three
grades in mind. Part of the point of going through the details of
perception is to show that there are in fact judgments arising at the
third grade. To demote these judgments to mere images is to obscure
this dialectic.
We can take at least some of the sting out of my interpretation by
realizing that their status as judgments does not by itself make these
third-grade operations optional. The mere fact that we’ve been mak-
ing such judgments from childhood makes them very difficult to
resist. But voluntary judgments are not necessarily resistible; the
fourth Meditation’s paradigmatically free judgments are such that
one cannot fail to make them once one sees the proposition at issue.
What is more important, our perceptual apparatus requires that we
continue making these judgments if we are to navigate through our
environments. There is still a problem of transparency, however: if
one thinks that the Cartesian mind is always and entirely ‘phosphor-
escent,’ in Gilbert Ryle’s phrase, then it is mysterious how these
operations could take place without our knowing it. Merely to reply
that they become automatic through habituation, in the way that our
understanding of language does, fails to solve the problem, for it

17
See Wilson (1993, 169) and Simmons (2003, 555).
18 19
I discuss la Forge’s views below, section 6.1. Simmons (2003, 567).
 LATER DESCARTES 

implies that at least at an early stage we were capable of ‘catching

ourselves in the act,’ as it were.
But to my mind, the biggest problems with the sixth Replies have
not yet been set out. At stage 3.1, we are supposed, ‘from’ or ‘on the
basis of ’ sensations, to judge that there is a colored stick outside of us.
Habituation can explain why we do this, but not how. The sensations
on their own are not sensations of anything; they need to be projected
in judgment. But what guidance is there for this projection? It’s not
enough for the mind to judge that something outside of me is causing
this sensation. The mind has to judge that this thing, which begins
here and ends there, is responsible for the sensation.
To see what makes the problem intractable, given Descartes’s set-
up, consider the following dilemma: do we have any awareness of the
region of extension that constitutes the stick independently of our
awareness of color, or not? If we do, then color is otiose: we would be
able simply to judge that there is a determinate region of extension
we’re calling ‘the stick’ without undergoing any sensations. Clearly,
that is not Descartes’s view. But if our only visual awareness of
extension comes through color, then there is nothing to guide us
when it comes to judge where to apply it. Being told to paint-by-
numbers is an intelligible command only if there are outlines within
which we are supposed to apply the colors.
Someone might reply that this isn’t quite fair. Suppose the visual
field is indeed a punctiform array of varied colors. One picks out the
points that resemble each other in respect of color, and then applies a
concept like ‘stick’ to bring them together. The first problem is that
Descartes doesn’t say any of that. But even if he did, I’m skeptical that
it would be of much use. After all, as Descartes himself is surely aware,
one of the things the mind has to accomplish with its judgments is
color constancy. The visual array will present a single object as
varying, sometimes dramatically, in its color. A deeper problem,
perhaps, is that the visual array cannot already be spatial, and have
spatial relations in it. That’s the sort of thing the mind needs to
produce at stage 3. In sum, the sixth Replies’ account of perception
is deeply problematic. That is no surprise if, as I suspect, he is
 LATER DESCARTES

exaggerating the role of the intellect as a rhetorical, nearly ad homi-

nem, device. In any case, later texts abandon the present view entirely
and try again.

5.2 Rapporter
I shall argue that the Principles and Passions stake out yet a fourth
position. These later texts depart from the views of the middle period.
In their place, we have two distinct kinds of events combining to
generate perceptual experience: the triggering of sensations and ideas
on one hand and the referring (rapporter) of those sensations and
ideas to objects in the environment on the other. Although rapporter
is anticipated in the sixth Replies, it takes on a new importance here,
as the chief mechanism by which we come to be aware of the world
around us.20
One way to approach this new view is to contrast it with that of Le
Monde. There, on experiencing something in the body, the mind
represents to itself the sensation of light.21 It was up to the mind to
decode, as it were, the brain signals and summon the appropriate
sensation. To say that the mind represents a sensation to itself seems
to mean just that it calls forth that sensation. By contrast, we are now
told that the brain motions represent sensations. As Descartes puts it
in the Passions, ‘only motions in the brain represent these objects
[namely, sounds, smells, tastes, heat, and the rest] to the soul.’22 Nor
is this an isolated slip.23 The motions themselves are doing the
representing or summoning here, not the mind. This fits well with
Descartes’s claim in the Passions that these motions cause the mind to

20
The doctrine of ‘rapporter’ also makes a (brief) appearance in the Notae (AT
VIIIB 358–9/CSM I 304). It is striking that the Passions’s recounting of the Dioptrique
account only discusses sensations and does not even mention ideas, or thoughts of
geometrical properties at all (Passions I.13, AT XI 338–9/CSM I 333). The Principles,
however, does make it clear that ideas (representations of geometrical qualities) are
also referred to objects (Principles I.71, AT VIII-A 35–6/CSM I 218–19).
21
Le Monde AT XI 4/CSM I 81, discussed above.
22
Passions I.13 AT XI 338/CSM I 333.
23
See Passions I.50 AT XI 369/CSM I 348, discussed below.
 LATER DESCARTES 

have certain sensations. There is no decoding necessary, in keeping

with the Principles’ revision of the linguistic analogy of La Lumiere.24
Not only is the present view at variance with that of Le Monde, as
one might well expect, it is also a serious departure from the Medi-
tations and Replies. Gone is the need for the mind to be aware of an
image in the brain. Nor is there any indication that the mind must be
aware of the position of the eyes or the parts of the brain.
Nevertheless, there seems to be one point of continuity: the mind
must still be aware of something in the body to which it is united,
even if that something has changed. In the Passions, Descartes pro-
duces what might be the two most variously translated sentences of
his entire corpus:

Those [perceptions] we refer to things that are outside of us, i.e., to the
objects of our senses, are caused (at least when our judgments in this regard
aren’t false) by those objects, which, exciting some movements in the organs
of the external senses, also excite, by the intervention of nerves, other
movements in the brain that make our souls sense them.
Thus when we see the light of a torch and hear the sound of a bell, this
sound and this light are two different actions, which, simply by exciting two
different motions in some of our nerves, and through them in the brain, give
to the soul two different sensations, which we refer to the subjects we
suppose to be their causes, in such a way that we think we see the torch
itself, and hear the bell, and not feel (sentir) only the movements coming
from them.25

24
See above, section 4.1.
25
Passions of the Soul Part I article 23, AT XI 346 (my translation). CSM I 337
takes considerable liberties. The translation given here is my own, and is designed to
be as literal as possible. In particular, I preserve the ambiguity of the first sentence.
Here are four candidates for the antecedent of ‘them,’ at the end of Sentence 1:
i) The perceptions/sensations
ii) The physical objects out there in the world, e.g., the bell or the torch
iii) The motions in the external senses
iv) The motion of nerves in the brain
CSM and Voss’s translation in Descartes (1989) both choose (ii), though only Voss
mentions the interpretive problem. Suppose CSM and Voss were right. Then Sentence
1 says that we sense the objects themselves. But Sentence 2 denies this. We take
ourselves to sense the objects but we don’t: we sense only the movements coming from
them.
 LATER DESCARTES

Although the standard translation of this text obscures the point,

Descartes’s implication in the final sentence is clear: we really do
sense only the motions coming from the torch and the bell. It is the
processing of referring that allows us to think we see the torch and
hear the bell, even though this is not really true.
The same account, I submit, is present in the Notae. Descartes
endorses article nineteen of the broadsheet he is critiquing. That
article reads in part: ‘Sense-perception consists entirely in the per-
ception of some corporeal motion.’26 As Descartes himself puts it,
‘nothing reaches our mind from external objects . . . except certain
corporeal motions.’27 The motions are the immediate objects of
experience, and they give the mind occasion to form ideas and
sensations, that is, perceptions of geometrical and sensible qualities.
We have then a two-step process. In being aware of the brain
motions, the mind is presented with the sensations and ideas caused
by those motions. It then has to project or refer these sensations and
ideas on to the objects it takes to be the ultimate source of those brain
motions.28 The first step is inelegant: if the mind is not decoding the
brain’s signals in accordance with the institution of nature, as the
earlier view had it, why think that the mind has to be aware of
anything happening in the brain at all?
For our purposes, it is the second step that is crucial. If we stopped
at the first step, we would not see objects outside us as being outside
us; we would simply undergo a panoply of sensations and ideas. In the
Passions, Descartes provides very few clues as to the nature of rap-
porter. It is best, then, to turn to the Principles for enlightenment.29
There, Descartes distinguishes two ways of making judgments. In the
first, one judges that the object has a property that either is or resembles

26 27
AT VIIIB 346/CSM I 296. AT VIIIB 359/CSM I 304.
28
Descartes’s focus in the Passions is on sensations. For the claim that the mind
refers both sensations and ideas, see esp. Principles I.71 AT VIIIB 35-6/CSM I 218–19.
29
Principles I.70, AT VIIIA 34/CSM I 218. In the next section, Descartes does
speak of the mind referring (‘referebat’) its thoughts to things outside itself. But he
does not explicitly identify this referring with either of the two kinds of judgment in
the previous section.
 LATER DESCARTES 

the one given in sensation. Such a judgment is either false or inco-

herent. In the second, one thinks only that there is something or other
in the object that produces the sensation.
Steven Voss builds on the second, purely causal judgment: ‘we
“refer” our perception to an object just in case we spontaneously
judge that the action causing our perception is within that object.’30
For her part, Deborah Brown glosses rapporter differently: ‘To refer a
perception A to B is to experience B as modified by A.’ Or, what she
takes to be equivalent, it is ‘for B to appear as modified by A.’31 One
advantage she claims for invoking ‘seeing-as’ is that it allows us to
preserve the difference between ‘perception and belief or judgment.’32
As I see matters, Brown and Voss each has half of the truth. Brown
is right in thinking that there is more to referring than simply judging
that object B caused sensation A. Where commentators like Voss
have gone wrong is not in thinking that judgments of causal origin
play a role in referring but in thinking that that is all there is to it.
A mere judgment of causal attribution does not explain our experi-
ence of the world around us. We don’t merely think that the objects in
our environment are causing our ideas and sensations. After all, we
sometimes (correctly) judge that a thorn has caused us pain, and yet
we do not experience the pain as being outside of us. So it is not
enough for us to think that the moss of the boulder is somehow
responsible for our experience of green. We have to go further and
think that the object is green. Only this second judgment explains the
richness of experience.33 Moreover, we can easily see how the two
judgments go together: if one is to attribute color to an expanse of

30
Descartes (1989, 30, n.23), quoted in Brown (2006, 98).
31 32
2006, 101. 2006, 101.
33
Someone might object that merely judging that a is F doesn’t entail that one
experiences a as F. For example, I might judge that the jell-o on the counter two yards
away is mushy without experiencing it as such (for that, I have to touch it). (I owe this
point to Ben Jantzen.) Descartes might reply that since we are actually experiencing
the sensible quality in question, to judge that the object has that quality is to
experience it as having it. (Note that in the jell-o case I am not experiencing the
mushiness.)
 LATER DESCARTES

extension, one must first judge that that expanse is responsible for
one’s color experience. If we somehow thought the sky was causing
our sensation of green, we would presumably see it as green.
We can accommodate Brown’s gloss on referring by seeing that
both kinds of judgment are taking place. We are not just judging that
the mossy boulder caused us to have a sensation of green; we are
in fact projecting that sensible quality on to the boulder. A helpful
synonym for ‘refer’ might be ‘attribute to.’ Like the Principles’ judg-
ments, it is ambiguous: one might attribute F to a body in the sense
that one judges that body to be the cause of F; or one might think that
the body is F. Referring, on my view, includes both judgments.
In sensory experience, we are systematically but harmlessly
deluded. Recall that Descartes says ‘we think we see the torch itself,
and hear the bell,’ even though we really only sense motions in the
brain. This belief is corrected by later reflection; but that does not stop
us from making the same judgments in sensory perception as we did
prior to our enlightenment. That many of our sensory beliefs are false
might be forced on us in any case by Descartes’s talk of judgment.
Strictly speaking, to judge, as we have seen, is to assent to a propos-
ition or deny it. It is up to the intellect to produce propositions that
can be affirmed or denied. If rapporter just is the judgments of the
Principles, then Descartes is telling us that the mind assents to the
proposition that this rock is green, for example.
The best defense for my reading of rapporter comes from Descar-
tes’s account of error in the Principles. There, Descartes sketches the
development of a child’s beliefs, and suggests that the ‘chief cause’ of
error is simply not growing up. At the start, the child undergoes both
sensations and ideas of geometrical qualities. It doesn’t ‘attribute’
either of these to bodies until its own body begins twisting ‘around
aimlessly in all directions in its random attempts to pursue the
beneficial and avoid the harmful.’ Only at this point does the mind
realize that the objects around it have an existence outside itself; as a
result, it attributes ‘to them not only sizes, shapes, and motions and
the like, which it perceived as things or modes of things, but also
tastes, smells and so on, the sensations of which were, it realized,
 LATER DESCARTES 

produced by the objects in question.’34 Notice that there are two

judgments being made here: the child thinks that the objects around
it are causing its sensations and it thinks that the objects are colored,
have taste, and so on. If it reaches adulthood, the child should realize
that there cannot be any ‘intelligible resemblance’ between its sensa-
tions and objects.35 Sensations, unlike ideas of geometrical qualities,
‘do not represent anything located outside our thought.’36
Referring, then, has to include both judgments. The purely causal
judgment is usually accurate; the judgment that paints objects with
the mind’s color is always false. But it is one that we find ourselves
compelled to make. The stick that looks bent in water still looks bent
even after you conclude that it isn’t. Similarly, even an enlightened
Cartesian sees objects as colored. And in Descartes’s system, the only
way to do that is to judge that they are colored.

5.3 Overview of Descartes

The Aristotelian view has easy answers to our three chief questions
about perception, namely, how it is that individuating, positioning,
and localizing are achieved. The proper sensibles are mind independ-
ently real. So it makes no more sense to ask why we see the rock as
gray (localizing) than why we see the rock having the shape and size
that it does. By being aware of these proper sensibles, we are at once
aware of the differences among objects (individuation) and their
relations to each other (positioning). Mistakes can be made, of course,
and one sense can correct another; but the reality of the proper
sensibles and their revelation of the common ones means that there
is no special mystery about how our perceptual apparatus achieves
contact with its environment. Over the centuries, the innocent view
gradually develops a sophisticated empirical account of the physical
processes that underlie perception.

34
Principles I.71, AT VIIIA 35-6/CSM I 219.
35
Principles I.70, AT VIIIA 34/CSM I 218.
36
Principles I.71, AT VIIIA 35/CSM I 219.
 LATER DESCARTES

None of those easy answers is available to Descartes. Following

Galileo, he splits the proper sensibles from the common ones. His
radical distinction between mind and body places sensible qualities
in the mind, as mere sensations. The gulf that has now opened up
between the two classes of quality threatens to remain forever open. It
is now a genuine question why we should see the rock as gray and not
the grass around it. Nor is it obvious how we are able to individuate
the rock from its surroundings and see it as occupying a distinct
position in space. What is more, Kepler’s work dooms the Baconian
synthesis, and so the empirical story that underwrites the innocent
view is unrecoverable.
Descartes begins with a twofold strategy. First, he appeals to the
transmission of motion, rather than of literal image, from the back of
the eye, through the optic nerve, and into the ventricles, where it can
ultimately paint a (properly oriented) picture on the pineal gland.
This ingenious maneuver allows him to preserve the basic structure of
the Baconian synthesis while avoiding Kepler’s critique. Second, he
takes a cue from the scholastics, who hold that, at least in abstraction
and in thinking of particulars, the intellect turns toward a corporeal
image. This doctrine of ‘conversio’ allows Descartes to exploit the
brain image in attentive human perception. Descartes simply takes
over the scholastic view, broadens it to include perception and
imagination, and casts his own immaterial substance in the role of
the gazing intellect. Even at this first stage, Descartes acknowledges
that some supplementary inferences (or quasi-inferences), the ‘nat-
ural geometry,’ can be made that allow us humans to get a more
accurate picture of the common sensibles than we otherwise would.
For all that, the chief mechanism of perception, for us as for non-
human animals, is the brain image.
As we’ve seen throughout these chapters, there is an inverse rela-
tionship between the accuracy of the brain image on one hand and
the need for the activity of minds (whether ours or God’s) on the
other. The less Descartes relies on the brain image, the more he is
forced to attribute the perceptual process to minds, whether finite or
infinite.
 LATER DESCARTES 

At the second stage, represented by the Dioptrique, Descartes

claims that minds both passively receive ideas of geometrical qualities
and make judgments about them through natural geometry. The
inputs for these judgments are, for the most part, an awareness of
parts of the brain and the shape of the eye. Such quasi-inferential
transitions are of course commonplace in the Baconian tradition.
That tradition could arguably afford to be vague about just how
much distinctively mental processing was happening, partly because
they do not draw the sharp line Descartes does between the physical
and the mental.37 But the deliberate vagueness of Descartes’s discus-
sions of these natural judgments is problematic, precisely because
he does draw such a sharp line. This stage also abandons the early
account of the proper sensibles. Now the mind does not read the signs
in the brain but merely suffers whatever sensations have been
ordained to be caused by the relevant brain motions.
The third stage is represented by the sixth Replies. There, Descartes
claims that the mind experiences a non-spatial field of color. It then
judges that there is an extended object in the subject’s sensory field.
Finally, the mind compares the now seemingly colored extension
outside the mind with its position relative to the parts of the brain.
This is the only stage at which the mind must make use of sensible
qualities to judge geometrical ones.
The fourth and final stage (Principles, Passions) retains the mind’s
awareness of the brain, although by now it is an awareness only of
motions in the brain and not the brain image, which has been
demoted to a mere side effect. There is now a two-step process: the
triggering of sensations and ideas by the perception of brain motions
and the rapporter of those sensations and ideas to physical objects.
Along the way, I have argued for some controversial theses. Des-
cartes at times thinks the mind turns toward the body when it
perceives or imagines. While this is surely, to a contemporary eye,
one of the strangest elements of his text, it is in fact Descartes’s

37
I owe this point to Hatfield and Epstein (1979).
 LATER DESCARTES

attempt to co-opt a standard Aristotelian view. We saw that the

doctrine of conversio has a prominent antecedent in the work of
St. Thomas Aquinas. In the scholastic context, there is nothing odd
about saying the intellect has to turn toward a corporeal likeness in
order to imagine or even perceive something.
And this leads us to a second contentious claim. As I read him, the
early Descartes believes that ideas can be corporeal. The image in the
brain is a representation and hence an idea: it is through that image
that the mind is able to perceive objects in the world. Moreover,
that image represents by resembling. This fits neatly with the early
Descartes’s ambitious overlap thesis, which offers to give a univocal
explanation for inattentive human and inhuman ‘perception.’
I have also argued that Cartesian representation generally is resem-
blance. Once again, that is nothing new or shocking in this intellec-
tual context: as la Forge and Suárez point out, hardly anyone in the
period disagrees. In the case of corporeal representation, the resem-
blance is literal; it is that of a picture to its original. The representa-
tional powers of mental ideas also turn on resemblance. Now,
whether mental representation requires what his followers call ‘enti-
tative’ or instead ‘intentional’ resemblance is unclear. But whatever
Descartes has in mind, representational resemblance, whether cor-
poreal or mental, must be an internal relation, since it must be such
that one can learn about the thing represented by looking at the thing
doing the representing.
For that very reason, among others, sensations are not representa-
tions. Nothing outside the soul can resemble a sensible quality.
Descartes is explicit on this point in the Principles. I have argued
that it follows from his views in the Meditations and in particular
from his doctrine of material falsity. Ideas of sensations represent
sensations. But sensations themselves represent nothing at all.
Each of these four stages contains the seeds of its own destruction.
Having severed the connection between proper and common sensi-
bles, there is nothing left to guide the mind as it conjures its sensa-
tions or ideas and refers them to physical objects. That is the point of
an incisive argument of Malebranche’s, to which we now turn.
 LATER DESCARTES 

5.4 Malebranche’s Critique:

the Selection Argument
Nicolas Malebranche’s best argument against Descartes’s theory of
perception comes, oddly enough, in the course of his attack on innate
ideas. That attack is in ill-repute, and for good reason: he gives a
demonstrably uncharitable reading of his target.
Malebranche’s first argument is an appeal to simplicity. Is it likely,
he asks, that ‘God created so many things along with the mind of
man’?38 Why fill each finite mind with its own store of ideas when
they can all be housed in God? For Malebranche, there is no reason to
assume that each mind has access only to its own modes. Once we
reject this assumption, it then becomes possible to think of minds
having access to a single set of ideas. Malebranche in effect divides the
labor of act and object between the finite human mind and an object
in the divine mind.
If this move is permitted, then Malebranche’s simplicity argument
against innate ideas seems to have a point: if the same work can be
accomplished with fewer entities, then surely that is the means God
would prefer. And while some of Descartes’s early work is a bit coy on
just which and how many ideas are innate, the 1647 Notae makes clear
that ‘there is nothing in our ideas which is not innate in the mind.’39
The simplicity argument assumes that for an idea to be innate is for
it to be in the mind. Descartes certainly encourages this assumption;
in the fifth Meditation, Descartes says he can ‘bring forth the idea of
God from the treasure house [thesauro] of my mind as it were.’40 Yet
Tad Schmaltz is surely right to put the emphasis on tanquam and not
thesauro.41 On Descartes’s real view, ideas are innate in the way
generosity or cancer is innate in a family. Thus ideas are ‘in’ us only
potentially, not actually. As Descartes tells Hobbes, ‘when we say that
an idea is innate in us, we do not mean that it is always before us . . . We
simply mean that we have within ourselves the faculty of summoning

38 39
SAT III.ii.4 LO 227. AT VIIB 358/CSM I 304.
40 41
AT VII 67/CSM II 46. Schmaltz (1996, 97).
 LATER DESCARTES

up the idea.’42 So Malebranche’s first argument misses its target.43

Does that mean that Descartes automatically wins the simplicity
contest?
It certainly improves his chances. Since Descartes does not stock
finite minds with infinite sets of ideas, his view is not nearly as
ontologically promiscuous as Malebranche supposes. It remains
true, however, that when the disposition is actualized, a numerically
distinct idea is produced in the finite mind, one which, from Male-
branche’s point of view, needlessly doubles the idea in the mind of
God. So even after we recognize the dispositional nature of ideas,
ontological simplicity still favors Malebranche’s view. If we’re just
counting ideas, Malebranche’s world is the smaller of the two.
So the issue of simplicity cannot be dodged by appealing to
dispositions. Nevertheless, simplicity arguments—especially those
narrowly focused on counting up entities—are not particularly per-
suasive. That one view posits fewer entities than another is a virtue
only if the questioned entities do no explanatory work. Malebranche
needs to argue for this claim. Happily, he does so, in what I’ll call the
‘selection argument.’44
In a dense passage, Malebranche both presents his argument and
considers and rejects possible maneuvers Descartes might make in
response. In fact, it almost seems as if Malebranche recapitulates the
development of Descartes’s view, taking on what we might call the
static model, according to which there is an image in the brain that
directs the mind, before moving on to the dynamic model, which
makes brain or eye motions play that role.

42
AT VII 189/CSM II 132. See also Notae AT VIIIB 357–8/CSM I 303–4, quoted
in Nadler (1992, 128–9).
43
This is the consensus in the literature. See, e.g., Nadler (1992, 128), Tad Schmaltz
(1996, 97), Andrew Pyle (2003, 54); as Pyle puts it, ‘it is hard to find anyone who
actually held’ the storehouse view. See also Jolley (1988, 79).
44
The selection argument has a close cousin in Wittgenstein’s work: see esp. the
Blue Book in Wittgenstein (1960, 3). As Jeffrey McDonough pointed out to me,
selection arguments tend to prove too much. In Malebranche’s case, they leave us
with the doctrine of the vision in God; in Wittgenstein’s, with a kind of evasive
behaviorism.
 LATER DESCARTES 

For ease of reference, I’ve divided the selection argument into three
sections:

[A] But even if the mind had a storehouse [un magasin] of all the ideas
necessary for it to perceive objects, yet it would be impossible to explain how
the soul could choose them to represent them to itself, how, for example, the
soul could make itself instantly perceive all the objects whose size, figure,
distance and motion it discovers when it opens its eyes in the countryside.
[B] Through this means it could not even perceive a single object such as the
sun when it is before the body’s eyes. For, since the image the sun imprints
in the brain does not resemble the idea we have of it (as we have proved
elsewhere), and [C] as the soul does not perceive the motion the sun
produces in the brain and in the fundus of the eyes, it is inconceivable that
it should be able to determine precisely which among the infinite number of
its ideas it would have to represent to itself in order to imagine or see the sun
and to see it as having a determinate size.45

In one scenario [A], you open your eyes in the countryside and see a
variety of objects. Note what the model under attack has to explain:
not only must you summon the ideas of each object, with the appro-
priate size and figure, you must also put it in relation to the ideas of
the other objects around you. In [A], Malebranche does not mention
sensations but presumably they too must be conjured and matched
with their objects.
Malebranche next makes the situation less complex with a new
scenario [B]. Are matters improved if we forget about the countryside
and think about a single object? The first matching problem is now
gone, only to be replaced by a second. For now the question is: given
just one object, which of the competing ideas that I could summon
should I now summon? Note that nothing about either of these
matching problems depends on the mind’s being a storehouse of
ideas. Whether the ideas are always, or only potentially, in the
mind, the problem remains: which idea do I call forth?

45
SAT III.ii.4/LO 227. The basic idea is there in the first edition (1674, 361), but
gets expanded with later editions. My translation departs from LO only in the last
clause: the French is, ‘telle grandeur determinée.’ For the purposes of this discussion,
I will not distinguish between ideas and sensations.
 LATER DESCARTES

In [B], Malebranche considers a single reply that would answer

both matching problems: it is the brain image that determines the
idea. At times, Descartes claims that you don’t need a (mental) idea to
have sensory experience, at least of geometrical qualities: the brain
image directs the mind to the object and hence plays the role of an
idea. But as we’ve seen, Descartes begins to realize just how inad-
equate the brain image would be. In [B], Malebranche makes just this
point: the image imprinted on the brain does not resemble the idea we
have of the sun. Nor does the brain image resemble the sun itself.
The inadequacy of brain images leads Descartes to the dynamic
view: it is ‘on the occasion of ’ certain motions in the brain that the
mind summons the correct idea. Once he gives up on the static image
model, however, there is no way for Descartes to answer the matching
problems. He cannot appeal to the literal resemblance between the
mind’s immediate object and the object in the world. So if brain
images no longer resemble objects in the world, then there is no
real story about how the mind picks the right idea. Here it is import-
ant to note that Malebranche never denies that we perceive the brain
image. His claim is only that that image is irrelevant to the matching
problems.
In [C], Malebranche addresses Descartes’s dynamic model. Male-
branche now flatly denies that we perceive the motion that the sun
produces in the brain or in the eyes. Given Malebranche’s claim
earlier in the Search that the soul ‘resides’ in the brain, in the sense
that it is aware of everything that happens in it, it is hard to see why he
should feel entitled to this blanket denial. I am not sure why he thinks
he can just rule out of court our being aware of brain motions,
especially as he is willing to allow that we are aware of brain images.46
However that may be, once the resemblance claim is given up,

46
SAT I.x/LO 51 suggests a teleological argument: it would do the mind no good to
be aware of these motions. That is why instead we are aware of sensations, which
differ from each other in kind, whereas motions differ only in degree. This argument
is orthogonal to our present concern, since the issue isn’t (only) about sensations. The
argument also begs the question: if being aware of motions allowed us to summon the
relevant idea(s), then there would be a good reason for us to be aware of them.
 LATER DESCARTES 

Malebranche can argue that the dynamic model provides no guidance

in the selection of ideas. Nothing about the motions in the brain can
tell you which idea to summon.
Descartes might reply: there’s only one innate idea that has to be
summoned in order to think of extended objects, namely, the idea of
extension. The innate idea of extension is the idea of a determinable;
it is not the idea of this or that extension. Fair enough; but the
essential problem remains. Somehow, I must be able to contract, as
it were, the idea of the determinable into that of a determinate.
Malebranche’s point, then, is that neither brain images nor motions
are of any help in this task. What the Cartesian mind is said to be
unable to do in [C] is to call up an idea that allows it to see the sun as
having a determinate size.
In sum, whatever its other flaws, the argument of [A]–[C] above is
not vitiated by its use of ‘storehouse’ (magasin). Nor does it present a
caricature of Descartes’s views on the relation between brain images
and motions and ideas. What is more, the selection argument gets to
the real issue over innateness. What most of Malebranche’s targets
have in common is hubris: they attribute to the mind powers that
should be reserved for God.47 And among these is selecting which
idea to think of in which experiential context.
If Malebranche is right, Descartes has no way to explain the mind’s
ability to make the transition from brain state to mental state. This
leaves open one escape route: why couldn’t Descartes give up on the
whole picture of the mind ‘representing to itself ’ the relevant ideas
and sensations, and just say that the connection is a brute causal one,
‘ordained by nature?’48 Even this move does not go far enough. For it
leaves the mind with the job of referring the ideas and sensations to
objects, and the same difficulty recurs: how does the mind know to
match its idea of a roughly circular shape to the boulder and not to,
say, the waterfall in the distance? How does it decide to match the
green sensation with the moss and not the tree bark? In localizing

47
This is one of the themes of Jolley (1990).
48
Daisie Radner envisions precisely this reply (1978, 49).
 LATER DESCARTES

sensations, it seems, the subject does have some decisions to make.

And Descartes has not provided a basis on which they can be made.
To parry Malebranche’s argument, the ‘automatic’ option must claim
that rapporter, too, is a purely causal process, ‘ordained by nature.’ So
the automatic view has to help itself to a vast array of incredibly
complex causal connections set up by God.
By itself, that does not seem like a devastating objection. But here it
is important to keep the dialectical situation straight. Descartes faces a
dilemma: does the mind need its own stock of innate ideas and
sensations so it can summon them in sensory experience, or not?
If it does, then the selection argument kicks in: innate ideas explain
nothing about sensory experience. If it doesn’t—if Descartes concedes
this point, and retreats to the automatic view—then he is vulnerable
to the simplicity argument: positing innate ideas for the mind to
exploit is an extravagance, if God can simply make these ideas
available to us. Descartes can resist one of the arguments only by
leaving himself open to the other.
Malebranche’s overall argument, then, succeeds. And although it
doesn’t quite make contact with every iteration of Descartes’s view
that we’ve explored, I noted further problems with each iteration as
we went along. Many of those problems were, in their own fashion,
simply a foreshadowing of the selection argument. Can any of the
Cartesians do better than their master? Before returning to Male-
branche’s own view, which will occupy the last chapters of the book,
I want to look at the variety of views Descartes’s other successors
produce.
 6
The Cartesians

Descartes’s kaleidoscopic array of positions and arguments leaves his

immediate successors, particularly those who take themselves to be
his faithful expositors, in a difficult spot. Already we’ve seen Louis de
la Forge’s energetic attempts to read away key features of L’Homme in
light of later texts. La Forge’s mélange of texts, and his own readings
of them, make for a view that does not fit easily into any one stage of
Descartes’s development.
If la Forge imports Descartes’s final works, such as the Passions,
Robert Desgabets remains, for all his innovations, firmly entrenched
in the earliest. Desgabets is perhaps best known now for his claims
that all substances are eternal (the ‘indefectibility thesis’) and that
everything one can think of exists, as well as for his alleged empiri-
cism.1 Desgabets has a profound influence on Pierre-Sylvain Régis,
although it does not, evidently, extend to the issues that will concern
us.2 Both philosophers take a quasi-Spinozistic view of matter, treat-
ing it as a single substance. In this they might have followed Descartes

1
For the indefectibility thesis, see Patricia Easton (2005). Monte Cook’s (2008) is
the best discussion of the sense in which Desgabets might be an empiricist. On
Cook’s view, Desgabets denies that our concepts come from experience; all experi-
ence does is cause us to have the thoughts we do. On the claim that anything one can
think about exists, see Cook (2002) and Timothy D. Miller (2008). That claim is
significantly less exciting than it seems, for Desgabets restricts it to substances. For
instance, when one thinks of a unicorn, one is thinking of material substance (which
does exist) modified in certain ways (in which it is not in fact modified). So strictly
speaking, the unicorn, qua material substance, does exist; it just doesn’t have the
modes one imagines it to have.
2
It’s hard to know just what works of Desgabets Régis has read; but see (Schmaltz
2002, 7 f.) for evidence of the connection between them.
 THE CARTESIANS

himself, but Régis’s treatment of finite minds as illusory aspects of

a single thinking thing is indeed a Spinozistic doctrine from which
Descartes, and possibly even Desgabets, would have recoiled in
horror.3 None of these claims, however interesting in itself, will detain
us long here. For all three philosophers—la Forge, Desgabets, and
Régis—agree that the world does not contain sensible qualities. While
none of them shows the empirical curiosity of the early Descartes,
each attempts to assemble a workable philosophy of perception out of
Cartesian materials.
In addition to these three lesser known Cartesians, we shall also
look at the work of Malebranche’s most vituperative critic, Antoine
Arnauld. Best known as a critic of Descartes, Arnauld devotes many
more pages to his controversy with Malebranche. But his own positive
view is worthy of attention. On its face, Arnauld’s view is problematic,
for he seemingly endorses a naïve realism about the structure of
perception while rejecting naïve realism about sensible qualities. Even
if he succeeds in excising the detested ‘ideas’ that interpose themselves
between subject and world, he nevertheless faces the same problems as
his countrymen.
Although Malebranche’s own major work appears right at the
heart of these developments, in 1674–5, we will look at his work
separately. If the order of chapters needs a defense, it is that Male-
branche continually revises the Search after Truth right up until 1712,
long after the other four figures are dead.
In what sense are the philosophers we’ll discuss in this chapter
‘Cartesians’? La Forge takes himself to be explaining, not replacing,
Descartes’s views; where he finds something implausible, he reads it
away by appealing to other texts. Arnauld is a more original thinker
but tends to exaggerate the degree to which he and Descartes agree.
The case is altogether different with Desgabets and Régis, given their
self-conscious departures from Cartesian teachings. And yet both
would argue that they merit the title ‘Cartesian.’ Desgabets tells us

3
On whether Desgabets anticipates Régis’s Spinozism about the mind, see esp.
Timothy D. Miller (2008), who argues persuasively that he does not.
 THE CARTESIANS 

that ‘one finds many things in [Descartes’s] writings which are

suitable for rectifying whatever is faulty in them.’4 Hence Desgabets
thinks he stands a good chance of becoming a better Cartesian than
Descartes himself.5

6.1 Louis de la Forge

La Forge’s commentary on L’Homme is notable for its creativity.
Like Desgabets, la Forge reads Cartesian representation as requiring
resemblance. But it is vital to see that la Forge is not, despite his
own protestations, simply adducing the Scholastics’ notion of non-
pictorial resemblance. What we called above intentional resemblance
is a sui generis irreducible notion, which puts a name to a lacuna:
somehow or other, there is representation happening in a way that
does not involve the sharing of qualities. La Forge’s own notion of
non-pictorial resemblance is quite different and will result in a thor-
oughly automated and purely causal account of sense experience.
Literal resemblance obtains in two ways, according to la Forge.
First, a can resemble b when a and b are of the same species, as when
one rhesus monkey resembles another. More to the point, such
resemblance can also obtain when a ‘has some relation [‘rapport’]
with [b], either in shape, color, or some other sensible accident, as a
picture does with the thing it represents or an echo with the voice of
whoever is singing.’6
By contrast, a can resemble b ‘with respect to its representational
being when, without having any real, positive or sensible resemblance
to the other, it is still able to make us think about it and to give us
the sensation we have of it.’7 To see why he defines representational
resemblance as he does, we have to recall that it comes up in the
context of the relation between the corporeal image in the brain and

4
Supplément, RD 5, 164.
5
As Schmaltz aptly puts it: ‘For Desgabets, Cartesianism is not a fixed position that
can simply be extracted from Descartes, but rather a work in progress that starts with
Descartes’s insights but that subjects his views to revision and correction’ (2002, 11).
6 7
1666, 103. 1666, 103.
 THE CARTESIANS

the object in the world. La Forge is anxious to avoid any literal

resemblance even between the physical image and its object, an anxiety
his fellow Cartesians simply do not share, or so I shall argue. For la
Forge, to say that a resembles b representationally is to say that a makes
us think of b and causes us to have the sensations we do. This is
something brain states can do even when no real resemblance obtains.
To deny any real resemblance between the corporeal image and its
object in the world is to reject Descartes’s early view, which relies on
that resemblance to explain how the soul that is aware of them can
use them to think out to objects in the world. We’ve already seen la
Forge deny that the soul is aware of anything in the brain, so it is no
surprise that he sees no need to make the brain image do any real
work. La Forge has no choice but to acknowledge that for the
Descartes of L’Homme the brain image is a representation; even he
cannot read away the words on the page. But he distorts Descartes’s
view by understanding the ‘representation’ at issue in a purely causal
sense. It does nothing except cause us to think of the object in the
world and trigger our sensations.
How, exactly, does the brain image do this? Nothing about the
image’s intrinsic properties, its shape or figure or position relative to
the eyes, can explain its power to make us think of an object in the
world. Indeed, as we’ve seen, la Forge doesn’t even think of the brain
image as an image, but as a set of changes (‘changements’) or motions
in the brain.
Unsurprisingly, la Forge appeals to God. God joined ‘all the
thoughts we have on the occasion of observing external objects’ to
bodily motions in the brain.8 When a person runs her hand along a
boulder, a series of changes happen in her body and ultimately her
brain to which God has annexed the thought of a boulder. In fact,
God’s annexing has to be considerably more precise than that: he has
annexed the thought of a boulder in just that position, having just
those dimensions that it has in her current environment. One need not

8
1666, 267.
 THE CARTESIANS 

be as attracted to divine simplicity as Malebranche to wonder whether

this is the most efficient means of bringing about human perception.
Even if we grant la Forge his fully automated account, he still faces
some difficulties. Notice that his definition of intentional resemblance
separates the thought of the object from the sensations we have in its
presence. Both are equally triggered by the subject’s presence in front
of the boulder; but that by itself does nothing to explain why one
should feel the boulder as mossy, and not the surrounding air, or why
one sees the boulder as green and not purple or gray. Sensations
might well have a role in alerting us, or livening up the mind to its
environment, but they are not representations and have no role to
play in explaining the directedness of perceptual experience.9 But
again, that does not get him off the hook: the fact is that from the
subject’s point of view, the boulder looks and feels mossy, and that
still needs an explanation. Put more sharply: we don’t just experience
mossy-look and mossy-feel when looking at and touching the boulder.
We experience mossiness-starting-and-ending-just-here-and-just-
there. A mere causal connection is too coarse-grained to explain
that pairing.
La Forge is not entirely without resources here, however. To see
them, we have to move from the role of the brain motions to the
relationship between the sensations caused by these motions and
the object in the world. Here as well, la Forge is keen to deny any
real or entitative resemblance. Instead, the senses, he tells us, ‘repre-
sent to us the action of the object, whether in the object itself and
outside our body, or at least in the extremity of one of our mem-
bers.’10 If la Forge here means to invoke his idiosyncratic version of
representational resemblance, he can gloss this as the claim that the
senses ‘make us think’ of the action of the object on us as if that action
were in the object itself. So even if a subject’s sensation of green
swings free of the thought of the mossy boulder, she can still experi-
ence the boulder as the cause of her sensation. To do so, she has to

9
See 1666, 276: ‘the senses . . . represent the good and evil of the body.’
10
1666, 216.
 THE CARTESIANS

commit (or rather, suffer) an error: thinking that the action of the
boulder that ultimately triggered her sensation is in fact in the
boulder itself.
That move raises a further problem: how is the subject able to think
of the boulder as the cause of her sensation, without using that
sensation itself to individuate the boulder from its surroundings?
Thinking of the boulder as the cause of her sensation presupposes
that the subject has already individuated the boulder from its
environment.
La Forge’s only solution is to automate the whole process of sense
perception so fully that the subject becomes entirely passive. She is
not even forming the judgments or thoughts of objects in her envir-
onment; she is undergoing them. This comes out most clearly in la
Forge’s treatment of the three grades of perception of the sixth
Replies. (In expounding and defending L’Homme, la Forge’s eclecti-
cism knows no bounds.)
The first grade is purely physiological, and here la Forge treats the
brain image as motion, as we have seen. At the second grade, we find a
confused perception that ‘always accompanies the motion of the nerve
fibers’ when it reaches the pineal gland. La Forge is not very clear about
whether the second grade includes mere perceptions or fully-formed
thoughts or judgments. He speaks indifferently of ‘confused percep-
tion’ and confused ‘sensory knowledge.’11 Descartes could hardly be
happy with this. For in the sixth Replies, he is, as we have seen, chiefly
concerned to pry apart what belongs to judgment and intellect from
what belongs to passive sensory reception.
What explains this indifference on la Forge’s part? If anything, it is
that he does not see any need for the activity of judgment in sensory
perception at any of the three grades. As a result, he necessarily makes
a hash of them. The second grade, in la Forge’s taxonomy, includes
much that Descartes puts in the third. In particular, at the second
grade, the senses are said to ‘represent the qualities which we perceive
by means of them as if they were in our sense organs or in the objects

11
1666, 271.
 THE CARTESIANS 

which impinge on them.’ Given la Forge’s treatment of represen-

tation, this amounts to the claim that sensory experience triggers in
us the thought that the qualities we perceive are outside of us. This
is a reflection of divine will, not an error deeply ingrained by
custom or experience. What is striking in all this is the degree to
which the senses are a source of error. It is not just sensible
qualities that are falsely painted on to objects; even the geometrical
qualities as presented (i.e., occasioned) by sense perception ‘are
hardly ever represented to us precisely in the way in which they
occur in the object.’12
La Forge thus presents us with the most anemic of the Cartesian
accounts we shall examine. It is no accident that la Forge does not
discuss the third grade of perception and simply refers the reader to
the Passions for its treatment of ‘rapporter.’13 For these Cartesian
texts cannot be made to fit with la Forge’s entirely passive account of
sensory experience.
La Forge’s account is anemic in a second way: there is no story here
about how the contents of sensory experience ‘hook on’ to the world.
We are told that a subject thinks of the boulder as causing her
sensations, and thinks of those sensations as somehow on the boulder
itself. What is the relationship between the world as her ideas and
sensations present it and the world as it is in itself? There is no room
in la Forge for real resemblance to play any role here. But neither,
I have argued, is there any room for an extrinsic relation such as
causation to give us a foothold. To claim that God causes us to have
certain thoughts on certain occasions is not to explain how those
thoughts get to be about what is in the subject’s environment.
It’s no surprise, then, to find la Forge’s successors abandoning
his view and appealing to other bits of machinery in Descartes’s
corpus, particularly the brain image. Our next philosopher relies on
la Forge’s explicit distinction between kinds of resemblance but
sticks doggedly to the claim that the brain image literally resembles
its object.

12 13
1666, 272. 1666, 278.
 THE CARTESIANS

6.2 Robert Desgabets

At first sight, Desgabets is an unlikely figure to make this move. First,
he denies that there is any such thing as sensation. And he is known
in the literature on these issues as a direct realist. We can begin by
exploring the precise sense in which he merits that title before setting
out his views on perception and evaluating them in light of Male-
branche’s selection argument.
From Desgabets’s point of view, Descartes’s main achievement is
the discovery that sensible qualities are illusory. This discovery ‘at last
opens to us a marvelous quarry from which to take the foundations of
a true philosophy.’14 In fact, Desgabets thinks he can deduce quite a
number of his central doctrines from the true nature of sensible
qualities, including the real distinction between mind and body.15
As Desgabets sees it, Descartes took a number of wrong turns after
this discovery. In particular, the whole strategy of examining one’s
ideas to sort the clear and distinct from the confused is itself confused,
since there are no ideas, only acts of thought and their objects. ‘To have
an idea of x’ is just a confused way of talking about knowing x.16
Desgabets deplores the habit, so common in his otherwise enlight-
ened times, of imagining that ‘ideas are like things and that these are
objects between thought and the object thought about.’ ‘The act by
which we know an object ends immediately and directly at the object
itself, and not at an idea that is merely a representative and objective
medium, as if it were a portrait in which and by means of which one
saw the King.’17

14
Supplément (RD 5, 165).
15
See Supplément RD (5, 197). Desgabets writes, ‘la doctrine de la nature des
qualités sensibles est la preuve fondamentale de la distinction de l’âme et du corps’ (‘the
doctrine of the nature of sensible qualities is the fundamental proof of the distinction
between mind and body’).
16
‘Connaître donc un objet et en avoir l’idée [sont] la même chose’ (Supplément
220) (‘knowing an object and having an idea [of it] are the same thing’).
17
‘[L]’acte par lequel nous connaissons un objet se termine immédiatement et
directement à l’objet même, et non pas à l’idée qui n’est aucunement un milieu
representatif et objectif, comme serait un portrait dans lequel et par le moyen duquel
on verrait le Roi’ (Supplément 220). See CdC 94.
 THE CARTESIANS 

All of that surely sounds like direct realism. It becomes clear in the
Critique de la Critique, however, that only in pure intellection does
thought reach its object directly. In other cases, as we’ll see, there is an
intermediary, although that intermediary is almost never a mental act
or object. Desgabets’s claim that there is no mental idea between
thought and object should not be allowed to obscure his claim that
imagination requires an intermediary. Just like the early Descartes,
Desgabets thinks that sensory experience is accomplished when the
soul perceives a brain image that entitatively resembles an object in
the world.
We now have to turn to imagination because Desgabets does not
think there is, strictly speaking, such a thing as sensation. There is
surely sensory perception of some kind or other, but on examination
it resolves itself into two: pure intellection and imagination. As
Desgabets puts it in the Critique:

. . . there are only two general ways of perceiving things, namely, knowing by
pure intellection or imagination, for sensation, by which we have originally
come by all our knowings [‘connaissances’], pertains sometimes to one,
sometimes the other. For very clear experience teaches us that the faculties
of knowing sometimes reach their objects immediately, and that in such a
case one has an idea without envisioning any intermediary or image in order
to perceive it, as happens when we think of the divine perfections, or of the
truths of faith and philosophy. By contrast, in imagination we apply our-
selves to an image formed in the brain, in which one perceives some real and
‘entitative’ resemblance with the object, as they say in the Schools.18

Let’s take each of the two processes involved in sensory experience—

intellection and imagination—in turn. Recall that, for Desgabets, the
ejection of sensible qualities from the world puts them entirely on
the side of the mind. So when one experiences a taste, or a color, one
‘knows himself intuitively.’19 The sensation is just a thought of oneself
existing in a certain state. The soul is at once the subject and object of
its own experience. As a result, Desgabets was mocked—as Male-
branche would be—for thinking that in sensing green the soul turns

18 19
CdC 93–4. CdC 97.
 THE CARTESIANS

green.20 But Desgabets’s point is that, on any account consistent with

the rejection of sensible qualities, those qualities must be folded into
the states, acts, or intra-mental objects of minds. In undergoing a
sensation, then, the mind is revealed to itself, as (minimally) in, and
capable of being in, the state in question. Although it sounds para-
doxical, when the mind experiences a sensation it attains pure and
intellectual knowledge of itself.21
Those who think that the objects of the senses are external objects
have not entirely freed themselves from the prejudice that sensible
qualities are outside of us, and in or on objects. What the senses
reveal, as their objects, can only be the soul itself and its own states.
Even though the objects of sensings are always the soul itself, they
nevertheless trigger in us thoughts of external objects. They could
hardly play their role of alerting us to the ecological properties of
bodies otherwise.
To see how that triggering happens, we have to move to the second
sense of ‘sensation.’ In this sense of the term, one is not experiencing
sensible qualities. Instead, one is imagining, that is, turning toward the
brain image. Whether a token imagining counts as an instance of
sensation is an extrinsic matter, one that turns on whether there is in
fact an object present and, perhaps, on the causal history of the brain
image. This, I suggest, is why Desgabets refuses to acknowledge sensa-
tion as a sui generis faculty of the mind, as nearly everyone else does.
Desgabets can claim Cartesian inspiration for this terminological div-
ision, since, as we have seen, Descartes himself thinks the distinction
between sensory perception and imagining is an extrinsic one.
To see how imagination works, we must keep in mind la Forge’s
distinction between his own brand of intentional resemblance, when

20
See Schmaltz (2002, 138). As I argue below, Malebranche is driven to his
rainbow-colored soul by entirely different pressures, including the metaphysics of
modes.
21
‘Second, it follows [from the doctrine of sensible qualities] that it is properly the
senses that give us this intellectual knowledge [of ourselves], because it is clear that
they provide the perceptions in question, and it is indubitable that these perceptions
are our ideas or thoughts, i.e., that they are our true knowings of ourselves insofar are
as we are in that state’ (CdC 98).
 THE CARTESIANS 

one thing simply makes us think of something else, and real or

‘entitative’ resemblance, which is internal resemblance in the ordinary
sense.22 Unlike la Forge, Desgabets insists that the brain image bears a
real and entitative resemblance to the object in the world. It is by
means of this image that the soul is able to think of the object.23 This
representational relation could not be accomplished in sensory experi-
ence without the brain image and its real resemblance to bodies.
In this respect, even though sensory experience includes sensible
qualities, it nevertheless is a source of knowledge:

One must know that the senses, whether interior or exterior, make nothing
known except whatever they make us think of, and of which they give us the
idea by the force of the union of mind and body; if they make us think of
matter, rest, figure, and all that can result from them, they make us know
these things, which are outside of us, and they do so so effectively that we
know them by a clear and simple knowledge.24

Note that Desgabets has no room for natural geometry: only when it
is a question of the ‘precise size’ of an object do we bother to do any
reasoning about it at all.25 Instead, as in L’Homme, it is the brain
image itself that directs the mind to the object in the world.
Even if Desgabets can appeal to the brain image to explain how the
subject individuates bodies and positions them relative to each other,
he stumbles when it comes time to pair sensations with their objects.
On his own view, sensations direct the mind inward, not outward;
they provide knowledge of the soul but not of any body. It is only the

22
For the distinction in Desgabets, see, e.g., CdC 117.
23
See CdC 94, quoted above, as well as CdC 121: ‘the corporeal species traced in the
brain has some true resemblance with its object, and that the soul turns itself toward
this image to form a spiritual idea so that the soul thinks of this object by means of this
image. By contrast, in pure intellection the soul receives simply an idea that represents
its object without perceiving the corporeal species that is tied to this idea and which is
lodged in the storehouse of memory.’
24
Supplément RD 5, 165–6; see CdC 171–2: ‘by having the sensation one knows
confusedly something outside of him that gives him this sensation, whatever it may
be, and this is always true. Beyond all that, the sensation presents to us clearly size,
roundness, distance, movement, which are corporeal and outside of us.’
25
CdC 132.
 THE CARTESIANS

brain image, not the sensation, that directs the mind outward. This
leaves open precisely how one is to pair the mossiness with the
boulder and nothing else.

6.3 Pierre-Sylvain Régis

One of the welcome features of Régis’s Système is precisely its sys-
tematic nature. Intended as a comprehensive textbook of Cartesian-
ism, the Système has a kind of clarity and order the works of other
Cartesians often lack. So it is no surprise that, when it comes time to
give an account of sensory perception, Régis gives an admirably clear
picture that, for that very reason, is largely novel.
Unlike Desgabets, Régis acknowledges that there is such a thing as
sensation, taken as an irreducible mental faculty. What sensation qua
faculty provides is simply sensations: experiences of sensible qualities.
Geometrical qualities are not themselves the object of sensation;
instead, one has to deploy imagination in order to discover them.
An ordinary sensory experience features both elements. When the
subject touches a boulder, the motion of the nerves, propagated to
the brain, causes her to have a sensation of mossiness and pressure.
She then refers (‘rapporte’) this sensation to the place where she
thinks the motions in her hand originate, namely, the boulder.26
Like the Descartes of the sixth Replies, Régis thinks we have to use
sensible qualities to trace the outlines of geometrical ones. The
extension of the boulder is, strictly speaking, invisible and insens-
ible; it can be revealed only when clothed in sensible qualities by the
imagination.

26
‘For example, when I look at a horse, the light which is reflected to my eyes,
excites a sensation of color that I refer (rapporte), according to the institution of
nature, to the place from which comes the impression of motion that causes it [the
sensation]. This is what makes it the case that I see in this place the extension of the
horse that was previously invisible to me. It follows from this that seeing in general is
nothing but a sensation of light or color that makes visible in a certain way an
extension that otherwise could not be seen’ (Système 162). Régis goes on to say that
the same account applies, mutatis mutandis, to the other senses. Hence I’ve chosen
tactile experience for my example.
 THE CARTESIANS 

Régis of course recognizes that we typically do not talk this way:

[The act of imagination] is really mixed with the sensation that precedes it,
which one doesn’t usually bother to distinguish from it. For example, when
I see a man who is near me, instead of saying that I see his color and imagine
his extension and figure, I just say that I see him, and in this way I attribute to
sensation alone what partially concerns imagination.27

Nevertheless, strictly speaking, one sees colors, and hears sounds, and
only imagines extension and other geometrical qualities like position
and distance.
The key step is, of course, the referring of the sensible qualities to
the objects in the world. Régis tells us that we refer these qualities
‘suivant l’institution de la nature,’28 according to the institution of
nature. There are only two alternatives here: either referring amounts
to a kind of judgment, by which the mind projects sensible qualities
on to bodies, or it is a purely automatic process.
Each has its own problems. If referring is a kind of judgment, then
Régis faces the familiar problems encapsulated in Malebranche’s
selection argument. The subject has to take her awareness of the
mossiness of the boulder and imagine it onto the boulder. Why
does she choose the boulder rather than the waterfall? Presumably
because she judges that the boulder is the cause of her sensation. But
then she must have some way of representing the boulder that does
not require her to be aware of its extension, since that awareness is
what the projection of the sensible qualities is designed to explain.
Unless Régis can account for this ur-awareness of the boulder’s
extension, his account is circular.
One possible maneuver would be to say that the subject needn’t be
aware of the extension of the boulder as such. All she needs is to be
able to judge that her mossy sensation originated in such-and-such a
position in the world outside her. She then uses this judgment of
causal origin to select a region of extension, and then in imagination
paints that extension with mossiness. But again, if she can have an

27 28
Système 163. Système 162.
 THE CARTESIANS

awareness of these positions without deploying the imagination, what

work is the imagined projection doing?
It seems more likely, then, that Régis does not think the subject is
herself doing any of this. Régis is describing a fully automated
process. We return, then, to the same dilemma that confronts Des-
cartes: if the process is automated, all of the detail is unnecessary.
Why not just say that God set things up in such a way that the subject
thinks there is a boulder of such-and-such a position, size, and shape,
when and (hallucinations aside) only when there is such a boulder?
Nicolas Malebranche’s selection argument in effect poses just this
dilemma. Whether he can improve on the work of the other Carte-
sians remains to be seen.

6.4 Antoine Arnauld

Before treating Malebranche in detail, we should come to terms with
the work of his sharp-tongued critic, Antoine Arnauld. We must
begin with the question of Arnauld’s alleged direct realism. No one
we have seen so far merits that title. Even Desgabets is a direct realist
only in the rather strange sense in which the early Descartes is one:
both hold that one is immediately aware of an image sketched in the
brain. The typical direct realist wants to say that we see ordinary
objects without any intermediary at all, whether idea or brain image.
What such a view would make of sensible qualities is an interesting
question.
Whether Arnauld will give us a chance to consider it is a vexed
issue. For it is far from clear that he holds any version of direct
realism.29 At first sight, his 1684 On True and False Ideas seems
promising. Suppose we choose, among all the possible definitions of
‘direct realism,’ the view that there are no ideas taken as mental
objects distinct from acts. On that score, Arnauld is very clear. Instead

29
See Nadler (1989) for the classic direct realist reading of Arnauld; for criticism,
see Kremer (1994) and Hoffman (2002).
 THE CARTESIANS 

of treating ideas as representative objects, Arnauld recommends

treating them as acts:

I have said that I take the perception and the idea to be the same thing.
Nevertheless, it must be remarked that this thing, although single, stands in
two relations: one to the soul which it modifies, the other to the thing
perceived, in so far as it exists objectively in the soul. The word perception
more directly indicates the first relation; the word idea, the latter.30

To talk of the thing existing in the soul objectively is, for Arnauld,
simply to say that the soul performs an act that is directed at that
thing. An immediate consequence of this view is that a perceptual act
cannot be, or fail to be, a representation in virtue of its connection to
some other object; instead, ‘all our perceptions are essentially repre-
sentative modalities.’31
But identifying ideas with mental acts is only one notion of direct
realism.32 What seems most at issue is whether and in what sense
the subject can be said to be in ‘direct’ contact with the things she
perceives. My suspicion is that whenever Arnauld seems to be arguing
against the view that ideas interpose themselves between subject and
world, he is really targeting the view that there is a distinction
between our acts of perception and their immediate objects. He
does not deny that there are such immediate objects; he just wants
to make them intrinsic to the acts in which they figure.
Malebranche is a natural target for Arnauld, not because of his
representationalism, but because he clearly and sharply distinguishes
between the act of perception and the idea toward which it is directed.
As we’ll see, only the act is in the finite mind; the idea must be in God.
There could hardly be a greater cleavage between act and idea than to
locate them in two such different substances.

30 31
Arnauld (1990, 66), trans. Nadler (1989, 109). Arnauld (1990, 66).
32
Hoffman makes this point: ‘it is not sufficient for direct realism that
intentionality is taken to be an intrinsic property of acts of awareness. A theory of
perception that held that we perceive external objects by perceiving the content
intrinsic to acts of perception would still be representationalist, because that
intrinsic content would be serving as an intermediate object’ (2002, 174).
 THE CARTESIANS

There are two main lines of argument for this reading of Arnauld
and his target. First, consider how Arnauld deals with illusion.
A contemporary direct realist has to offer a disjunctive account:
in veridical perception, we are in direct contact with objects, while
in illusory contexts something very different (though indistinguish-
able from the first-person perspective) is happening. That is not
Arnauld’s view.
Instead, Arnauld helps himself to the representationalist’s analysis
of illusion, by pointing to the identity between act and representation:

[I]f it is not necessary that there be something external similar to the

representation, then it is no longer necessary that there be outside of us
anything existing which is similar to the perception that I have of the sun. It
follows that there is no reason why I must have recourse to these represen-
tations distinct from perceptions, which enable me to see the sun even if
there were no sun.33

Arnauld objects, not to representations per se, but to representations

that are distinct from perceptions. If he folds his representations into
the acts of thinking, he is entitled to the very same response as the
representationalist: one can be thinking, and even thinking of some-
thing, whether the ultimate object represented exists or not.
Stronger evidence comes from Arnauld’s treatment of objective
being. When he says that ideas are representations, Arnauld insists he
is saying something ‘completely different from saying that pictures
represent their originals and are the images of them . . . ’ This seems
promising, but all Arnauld means is that ideas and perceptions
represent things in the sense that ‘the things we conceive are object-
ively in our mind.’34 And the relationship between objective being
and the thing represented is sui generis, not reducible to literal
resemblance. So once again where it looked as if Arnauld was about
to challenge the whole picture of the mind using a representation to
think about things in the world, we find him instead taking aim at the
distinction between ideas and acts.

33 34
Arnauld (1990, 87). Arnauld (1990, 66).
 THE CARTESIANS 

What little Arnauld does say about objective being should give the
direct realist pause. A direct realist should hope that, on Arnauld’s
account, for x to be ‘objectively’ in the mind is just for x to be the
thing one is thinking about. Unfortunately, as Arnauld tells us, he
‘does not just mean that [x] is the object which my thought is about,
but that it is in my mind intelligibly, as is customary for objects which
are in my mind.’35
The fact that Arnauld doesn’t help himself to a deflationary reading
of objective being does not, by itself, block his route to direct realism.
The objective being of x in the intellect need not be the immediate
object of thought; it might merely be part of the causal story that
explains how the intellect comes to have x as its object. It’s too bad,
then, that Arnauld goes on to block this maneuver, too:
[I]f I think of the sun, the objective reality of the sun, which is present to my
mind, is the immediate object of this perception; and the possible or existing
sun, which is outside my mind, is so to speak its mediate object. It is clear
from this that, without invoking representations distinct from perceptions, it
is true that, not only in the case of material things but generally in regard to
all things, it is our ideas that we see immediately and which are the immediate
object of our thought . . . 36

Once again, it’s not immediately perceived representations that are

the target, but only representations that are distinct from mental acts.
As Aloyse Raymond Ndiaye puts it, ‘in the perception of body,
everything happens as if the idea is a window through which
I perceive things. In a sense, I can say that I perceive nothing directly,
since I only perceive through this window.’37
So it is hard to believe that Arnauld is a direct realist in the relevant
sense, namely, someone who thinks that perception of objects is
direct and not mediated by representations. But even if he were, he
is clearly not a naïve realist: like his fellow Cartesians, he denies that

35
Arnauld (1990, 67). Arnauld’s last phrase self-consciously invokes Descartes’s
first Replies, examined above.
36 37
Arnauld (1990, 71–2). Ndiaye (1991, 70).
 THE CARTESIANS

sensible qualities belong to extension.38 Indeed, the question of the

distinction between ideas and acts aside, the rest of his philosophy
of perception is largely Cartesian.39 As a result, he faces the same
challenges.
In an unpublished treatise probably composed around 1680 (Exa-
men), Arnauld considers three ways in which our sensations might
come about. First: it might be that corporeal motions in the body
cause our sensations. The problem here is that bodies can only knock
other bodies about; they cannot, on pain of violating the causal
likeness principle, cause sensations.40 Second: the soul itself might
cause these sensations, on the occasion of movements in the brain.
This, of course, is precisely the view of the early Descartes. Arnauld
does not dismiss it, for the third alternative—deus ex machina—is
unpalatable. But he simply cannot see how the soul could give these
sensations to itself
so appropriately and with such marvelous speed; since the soul would not
know when it must give them to itself, not knowing which corporeal motions
in the organs of sense these perceptions must follow, and always follow by
the order established by nature. It is as if one claimed that because a man
knows how to speak French well, he can respond appropriately to everything
that is asked of him in Hebrew: for all the motions that are in the organs of
our senses, are as Hebrew to our soul, which knows them not.41

Arnauld’s objection is not that the soul is not aware of the brain.
Rather, it’s that the soul cannot decode the language of the brain.42
Arnauld is making a point that might have been inspired by

38
See, e.g., Logique I.9: 49 f.
39
Not that the difference (if it is one) would be trivial. As Malebranche puts it,
reacting to Arnauld’s boasts of Cartesian orthodoxy: ‘M. Arnauld says that “the modi-
fications of the soul are essentially representative”; M. Descartes, that ‘the thoughts of
the soul considered as being simply modifications of thought are all equal and represent
nothing.” M. Arnauld says that nothing is lacking to the soul to enable it to perceive
objects; M. Descartes claims that it is by the diversity of objective realities that the soul
can perceive diverse objects. What grand accord, Sir, between these two sentiments!’
(First Letter to Arnauld, OC 6: 216–17).
40 41
OA 38: 146. OA 38: 147.
42
Kremer’s otherwise excellent (2012) neglects this point. Instead, as Kremer reads
him, Arnauld argues that ‘the soul cannot form sensible perceptions in itself on the
 THE CARTESIANS 

Malebranche’s selection argument: how can the soul decide which

sensation to summon, on which occasion? Arnauld, then, is driven
to the third alternative, namely, that God causes the appropriate
sensations when the occasion demands.
The occasionalist move takes Arnauld only part of the way. What
will he do with these sensations, once they have been caused? Are
such perceptions of sensible qualities representations? If so, of what?
It is at this vulnerable point that Malebranche’s attacks find their
target. To see this, we should consider Malebranche’s reply to one of
Arnauld’s more churlish objections.
Arnauld mocks the doctrine of the vision in God for its clash with
common sense. On Malebranche’s view, according to Arnauld,
‘women who idolise their beauty see God in looking at themselves
in the mirror, because the face they see is not their own but an
intelligible face which resembles it and which is part of the infinite
intelligible extension contained in God.’43
In his 1684 Réponse, Malebranche uses Arnauld’s view of color
against him. Malebranche writes:

It pains me to reply to these juvenile objections, which are fit to ensnare

children and simpletons. When one sees a woman, it’s only the color of her
face that makes her visible; and if there were no color, would we see it? For,
according to M. Arnauld, color isn’t in the woman, but is instead a modifi-
cation of the soul. Thus, by his own reasoning, no man ever sees or loves a
woman. For one loves only what one sees; and one sees only color, or colored
extension, which is only a mode of the mind.44

Arnauld is just as far from common sense as Malebranche is. In fact,

Arnauld uses Malebranche’s own language of ‘spreading’ (‘repandre’)

occasion of particular motions in the bodily sensory apparatus because the soul is not
aware of those motions.’ My reading, I believe, agrees with Nadler’s (1989, 57–8).
43
Arnauld (1990, 148).
44
Réponse OC 6: 77. As Gaukroger (in Arnauld 1990, 227) notes, ‘there is not a
great deal to be learned from the long dispute between Arnauld and Malebranche.’
Only at isolated moments does one catch something worthwhile. On the whole
reading these texts is a dismal experience, like reading unedited comments on a
political blog.
 THE CARTESIANS

sensations over things. Arnauld writes, ‘[m]en are only partly mis-
taken when they regard colours as being spread over objects. For
while they are not really spread over them, nevertheless the intention
of the Author of Nature is that our soul attach colours to them and
apply them (y attache & les appliquent en quelque sorte) to bodies in
some way, in order to distinguish among them more easily.’45 The
most substantial story one can find in Arnauld is in the Logique, and
that story is a thinly veiled paraphrase of the corresponding account
in Descartes’s Principles.46
In short, Arnauld owes us an account of how the soul spreads the
objects of its own acts on to bodies. Whatever conclusion one reaches
about his debate with Malebranche, it seems to me beyond question
that Malebranche goes much further in giving us such an account.
And instead of challenging representationalism, Arnauld defends
an isotope of it. Truly challenging the Cartesian picture of the mind
thinking through ideas to individual objects they represent will,
ironically, be left for Malebranche.

45
OA 38: 314, trans. Gaukroger in Arnauld (1990, 175); see the Port-Royal Logic
(1996, 49).
46
Compare Logique (1662) I.9: 49 f. with Descartes’s Principles (1644) I.71 (AT
VIIIA 35/CSM I 218–9).
 7
Malebranche on Sensation

Malebranche’s philosophy of perception is best explored in light of

the selection argument he lodges against Descartes. For that argu-
ment is liable to turn against whomever wields it. The core problem is
this: having severed sensations of sensible qualities from ideas of
geometrical ones, it is surprisingly hard to knit them back together.
In an effort to satisfy the demands he himself lays out, Malebranche
develops a wonderfully intricate network of positions in metaphysics
and the philosophy of mind.
The next three chapters explore the main strategies Malebranche
uses to solve the problem he was the first to clearly articulate. After
exploring the ontology of perception, we turn to what I call the
‘robust’ theory of natural judgments, found only in the first (1674)
edition of the Search.1 That edition hews closely to Descartes’s early
work, requiring that the mind be aware of what takes place in the
brain. In this work, Malebranche claims that it is up to us to take the
sensations we have and apply them to bodies, rendering those bodies
visible and sensible. We do so by making judgments that are natural
in the sense of being habitual; we fail to notice them because we are so
accustomed to making them. Such judgments—that a given spherical
object is red, for instance—are always false.
Whether or not he sees that this theory is unworkable for the
reasons I’ll give, Malebranche performs an astonishing about-face.
In the 1678 edition, Malebranche introduces what I call ‘anemic

1
The most useful discussion of the relevant differences between the first edition
and later ones I have found is Rodis-Lewis (1963, 46 f.).
 MALEBRANCHE ON SENSATION

natural judgments,’ which are merely compound sensations. It will

take some work to see just what this phrase means. On its face, the
view is not promising, since these judgments are not really judgments
at all. I argue that any attempt to repair the view would run afoul of
Malebranche’s own metaphysics of sensation.
Further versions of Malebranche’s view include two developments.
In the early period, Malebranche claims that there is an idea in God
corresponding to every individual body in the world. But on the later
view there is only one such idea in God’s mind—what Malebranche
calls ‘intelligible extension’—with which we are in cognitive contact.
There is a second difference: the doctrine of efficacious ideas. In his
later work, Malebranche insists that, although God is the only true
cause, ideas nevertheless can be said to act on minds. There is thus
some sense in which the idea of extension causes my sensations.
Arnauld accuses Malebranche of changing his mind, and many
commentators follow suit. Ferdinand Alquié, for example, argues
that Malebranche has replaced the vision in God with the vision by
God.2 These two doctrines—intelligible extension and efficacious
ideas—make for a radical change in Malebranche’s position, despite
his protestations to the contrary.3
I shall argue that Malebranche begins with an ambitious program
that makes minds active in the construction of their own experiences
and ends with a view that deprives minds of this role. How can we
explain that transition? A natural thought is that it is the full flower of
his Augustinian world-view, which, in sharp contrast to that of
Descartes, deprives human minds even of the power to house—let
alone produce—their own ideas.4 Malebranche might well have come

2
See Alquié (1974, 209 f.) As Alquié notes, Malebranche goes right on talking as if
there were particular ideas of bodies in God’s mind, even after he introduces the
notion of intelligible extension.
3
Giving a precise date to these developments is difficult. The first edition of books
one through three of SAT is 1674, with the other three following in 1675, and the
Elucidations in 1678. Certainly by the time we reach Elucidation X, intelligible
extension is in place. The doctrine of efficacious ideas is rather harder to pin down;
see Nolan (2012, 27, n.26) for more.
4
This way of drawing the contrast between Descartes and Malebranche is one of
the most enlightening contributions of Jolley (1990).
 MALEBRANCHE ON SENSATION 

to regard his early view as not going far enough, since it still accords
minds the power to project sensations on to ideas and to make
individuating and positioning judgments.5 His motivations need
not be purely theological, however, since each of his earlier views
has internal problems that make it unstable. The final iteration of
his position takes a dramatic turn, one that comes into focus only in
his often churlish exchanges with other philosophers. Taken to its
logical conclusion, or so I’ll argue, Malebranche displaces the whole
Cartesian picture of the mind using a representative idea to perceive
the world of extension. For intelligible extension really does none of
the work a Cartesian idea is supposed to do.
A central plank in my reading is the claim that at no stage of his
development are Malebranche’s sensations representations in any of
the competing senses on offer: they resemble nothing, either really or
merely intentionally. And although of course they have a causal
history, that history is of no help in solving our puzzles, nor does it
turn them into representations. Since this claim follows, or so I argue,
from Malebranche’s ontology of perception, we must start there.

7.1 An Adverbial Theory?

We must begin with a statement of Malebranche’s general theory. In
its broadest outlines, the theory claims that we see bodies neither
directly nor by means of representations in our minds; instead, we see
them only through ideas in God’s mind. God’s ideas being purely
intellectual, they can have no sensory element. Sensations must then
be placed in the mind of the perceiver. Malebranche opens the Search
after Truth with a distinction:

We can . . . say that the soul’s perceptions of ideas are of two kinds. The first,
which are called pure perceptions, are, as it were, superficial to the soul: they
do not make an impression on it and do not sensibly modify it. The second,
which are called sensible, make a more or less vivid impression on it. Such
are pleasure and pain, light and colors, tastes, odors, and so on. For it will be

5
I owe this insight to an anonymous referee.
 MALEBRANCHE ON SENSATION

seen later on that sensations are nothing but modes of the mind [manieres
d’être de l’esprit], and it is for this reason that I call them modifications of
the mind.6

Note that Malebranche here is not making quite the distinction one
might expect. The present distinction is between two ways of per-
ceiving ideas: purely and sensibly. Pure perceptions here are the sort
of non-imagistic, intellectual thoughts one might have while doing
geometry. Sensible perceptions are still perceptions of ideas, though
those ideas are accompanied by sensations, which make—or, as we’ll
see, just are—vivid impressions on the mind. So the contrast is not
between two aspects of perception, but between two kinds of percep-
tion. Nevertheless, the distinction we are seeking is implicit here, for
sensible perception includes both ideas and sensations. This comes
out more clearly much later in the Search:

When we perceive something sensible, two things are found in our percep-
tion: sensation and pure idea. The sensation is a modification of our soul, and
it is God who causes it in us . . . As for the idea found in conjunction with the
sensation, it is in God, and we see it because it pleases God to reveal it to us.
God joins the sensation to the idea when objects are present so that we may
believe them to be present and that we may have all the feelings and passions
we should have in relation to them.7

At this stage, it is utterly obscure how this ‘joining’ happens. It

seems to mean nothing more than that God does two things at
once: causing a sensation and revealing an idea. But this won’t do.
It is not merely that the soul undergoes these two changes at the
same time, in the way one might suddenly feel hungry while work-
ing on a proof in geometry. The idea and sensation seem to be joined
in a much more robust sense. The work of understanding that
joining of idea and sensation belongs to the next three chapters. In
the remainder of this one, I want to get clearer on Malebranche’s
ontology of perception.

6 7
SAT I.i/LO 2. SAT III.ii.6/LO 234.
 MALEBRANCHE ON SENSATION 

While much about that ontology of perception is controversial,

most commentators converge on one point: Malebranche defends an
adverbial theory.8 On such a view, a sensation is not a mental object,
but merely a way in which an act of sensing happens. Cast in his
ontology, the claim is that the soul perceives the idea of extension red-
ly, or carrion-scented-ly, or D#-ly. Malebranche garners praise for
anticipating developments in twentieth-century philosophy of mind
by pioneering a new way to locate sensations in a mechanical world.9
I shall argue that this reading cannot be right. At the same time,
it is onto something important about Malebranche’s philosophy of
perception.
Contemporary adverbialists are united in thinking of a given color
sensation, call it C, as modifying an act. There is the act of sensing
the table, and when this act is done C-ly, we get the appearance of a
C-colored table. The crucial point here is that sensations are adverbial
modifications of acts, not of substances. No adverbialist would want
to say that sensations are adverbial modifications of the brain; this
would be a category mistake, since adverbs modify verbs, not nouns.
A possible dodge would be to try to make ‘existing’ the relevant verb,
so that Bobo’s sensing a red table involves Bobo’s brain existing redly.
But then why not just say that C is a way the table has of existing? To
say that a table exists redly is just a rather odd way of saying that the

8
See, e.g., Cottingham (2008), Jolley (1990, 1994, and 2000), Nadler (1992), and
Schmaltz (1996). Of these figures, Schmaltz is the most circumspect; he claims that
‘[t]here is considerable reason to attribute to Desgabets and Malebranche alike an
adverbial account of sensation according to which we have, for instance, a sensation of
green not by perceiving a green mental object but rather by perceiving greenly’ (1996,
257, n.138). Lawrence Nolan (2012) also argues against the adverbial reading, albeit
on different grounds.
9
See Jolley (2000). Whether the adverbial theory is in fact philosophically defens-
ible is a further issue. For my part, I find the adverbial view itself elusive. I just don’t
know what it could mean to say that I sense an object brown-ly. I am hardly alone in
this; see Laurence BonJour (2004) and (2007). And Frank Jackson’s ‘many-properties’
objection to adverbialism (1977, 64 f.) has not, as far as I can tell, really been answered,
though of course philosophers have tried (see, e.g., Tye 1984, though I should note
that Tye 2007 seems to agree that adverbialism is indefensible). Jackson’s ‘many
properties’ objection to adverbialism is a special case of the problem of localizing
sensible qualities.
 MALEBRANCHE ON SENSATION

table is red. So if we’re unhappy with saying that the table is red, it’s
hard to see why casting the brain, or whatever it is that thinks, as the
thing that is red should be preferable. Any adverbialism about sen-
sations worthy of the name has to identify those sensations with ways
that acts of perception are performed.
But this isn’t Malebranche’s view. Consider the famous (or infam-
ous) rainbow-colored soul passage:

You even make a fool of yourself before certain Cartesians if you say that the
soul actually becomes blue, red, or yellow, and that the soul is painted with
colors of the rainbow when looking at it. There are many people who have
doubts, and even more who do not believe, that when we smell carrion the
soul becomes formally rotten, and that the taste of sugar or of pepper or salt,
is something belonging to the soul.10

Nicholas Jolley uses this passage to introduce the adverbial reading.11

But it’s poor evidence for this reading precisely because it takes
sensations to be modifications of a substance, not an act. Consider
what a genuinely adverbial view would have to look like in the context
of Malebranche’s ontology and vision-in-God doctrine. An adverbi-
alist Malebranche would have to claim that when I perceive an idea in
God, that act of perception has two aspects: it is directed at the idea,
and it is performed C-ly, where ‘C’ specifies a color. That’s not what
Malebranche says. Instead, it is the mind—not the act—that exists in
a certain way. Contemporary adverbialists would reject this, even if
we replace ‘mind’ with ‘brain’ rather than ‘soul.’ In short, it is not just
‘certain Cartesians’ who will think you a fool if you say that sensing
blue just is the sensing thing’s turning blue.12
To this, defenders of the adverbial reading might shrug their
shoulders. Yes, Malebranche departs from the adverbial theory on
this score. Sensations are modes of the soul rather than aspects of

10
E XI/LO 634. As Schmaltz (1996, 257, n.136) shows, this view was in fact
attributed to Malebranche by his critics, particularly Anselme of Paris, in 1678.
Note that Malebranche does not disavow it here.
11
Jolley (2000, 37).
12
See Jolley (2000, 37) for an ingenious defense of Malebranche on this point.
 MALEBRANCHE ON SENSATION 

sensory acts. And yet it remains the case that sensation does not
exhibit an act-object structure. Like the adverbialist, Malebranche
denies that there is a color that is sensed.
Before replying, we must clear up a terminological issue, if only by
stipulation. When discussing Malebranche, it is useful to distinguish
between representation and intentionality. I follow Nadler in giving a
minimal content to ‘intentionality’: an object or act is intentional if
and only if it is directed at something outside itself. Representation
includes, but means more than, mere directedness.13 To see the
difference, note that no Malebranchean act is representational; there
is nothing intrinsic to the act of pure perception, for example, that
explains why it is a (pure) perception of an idea of a square.14
Representation is the exclusive province of ideas, which are (some-
times) the intentional objects of mental acts. If those intentional
objects are themselves representations, the mind can think through
them to the representanda.
I now want to go further and argue that Malebranche does indeed
treat sense experience on the act/object model. For Malebranche,
sensings are intentional, in our sense: they are directed at an object.
If sensing counts as thinking, then this is inevitable: to think is to
think of something.15 Sensing is thinking and so must have an
object.16 The only question is, what is its object? Is it an idea in
God’s mind, or something else?

13
For the distinction, see Nadler (1992, 81 f.); Simmons (2009, 106; 116) follows
him in this. Nadler takes intentionality to obtain between a perceiving and an idea,
while representation links the idea with its object. (Nadler, of course, takes the
‘perception’ relation in this context to be intellectual, not sensory, which enables
him to read Malebranche as a direct realist.)
14
Simmons (2009, 124) makes this point.
15
‘As M. Arnauld says, to think of nothing is not to think at all, i.e., there is no
thought that lacks its object’ (OC 9: 945).
16
‘[B]y the words thought, mode of thinking, or modification of the soul, I generally
understand all those things that cannot be in the soul without the soul being aware of
them through the inner sensation it has of itself—such as sensations, imaginings, pure
intellections, or simply conceptions, as well as its passions and natural inclinations’
(SAT III.ii.1/LO 218).
 MALEBRANCHE ON SENSATION

Something else. The object of a sensing is a mode of the mind. In a

December 1690 letter, Malebranche writes, ‘your soul is green, or has
a modification of green that you see when you are in the middle of a
meadow, your eyes open.’17 Such passages could easily be multi-
plied.18 Sensing is a kind of sentiment intérieur, by which the mind
is aware of its own modes; what it is aware of are sensations, or modes
of the mind. So the adverbial reading has two strikes against it:
sensations are not ways in which sensings are performed but modes
of the soul that senses. And sensory experience has an act-object
structure, with the mode of the soul serving as the object.
We cannot leave matters there, for the adverbial reading has one
very important thing right. As we have seen, to say that a state or act
is directed at an object and hence intentional is not to say that it is
representational. Even in pure perception, the act of the mind is
intentional, and yet it is not representational. The work of represen-
tation is left for the object of thought to perform, if it can.
On my reading, Malebranche’s sensings are intentional without
being representational. In this, they are perfectly symmetrical with
pure perceptions. The difference is that the object of a sensing is not
an idea endowed with representational powers but a tenebreuse
modification of the soul. Acts of pure perceiving are ‘representational’
in a derivative sense, in virtue of the representational powers of their
direct objects. But the objects of sensings have no such powers.
Sensations really are blank effects.19 Like the feeling of being tickled,
they are neither intentional nor representational.
Thus, while I disagree with the adverbial interpretation, its defend-
ers are right in one respect: Malebranchean sensations are not
the appropriate sort of thing to point to or be about anything at all.
This is so even though sense experience, as I have argued, has an
act-object structure. For we are now concerned with sensations—
things sensed—rather than acts of sensing.

17
OC 19: 564, trans. and quoted in Schmaltz (1996, 83).
18
See, e.g., SAT I.xiii/LO 61.
19
See Ayers (1991, vol.1, 22) for the origin of this phrase.
 MALEBRANCHE ON SENSATION 

This ‘blank effect’ reading has recently come under attack, in

helpful analyses by Lawrence Nolan and Alison Simmons. While
I agree with these commentators that Malebranche desperately
needs sensations to be representations, his ontology makes it impos-
sible for him to meet this need. It is in the very nature of an idea that
one can, by perceiving it, think ‘through’ it to what it represents;
by contrast, there is no means by which one might think ‘through’ a
mode to an idea, or anything else.20 Malebranche makes this explicit
in the Elucidations: sensations, as modes of the mind, ‘are in no
way different from us, and . . . as a result can never represent anything
different from ourselves.’21 In his exchange with Arnauld, he makes the
same point: ‘it is in the soul that one sees colors or senses pain. For it
would be a contradiction to say that the modification of a substance
can be located where the substance is not.’22 That is precisely why
Malebranche needs his theory of natural judgments and all the rest.
To get at this point, we need to consider the Cartesian background
of Malebranche’s theory. Doing so will help show why sensations

20
Simmons (2009, 122–3) argues that ‘sensations-2’ (what I have been calling
simply ‘sensations,’ i.e., the objects of sensings) are ‘Janus-faced,’ in that they can be
considered as modes of the mind and at the same time as acts directed toward the idea
of extension. Her argument seems to be the following: the mind, for Malebranche, has
only two faculties: understanding (the faculty of ‘receiving various ideas, that is, of
perceiving various things’ (SAT I.i1/LO 2) and the will. Simmons concludes that ‘there
is simply no room [in Malebranche’s framework] for non-intentional sensations that
are neither ways of perceiving nor ways of willing’ (2009, 123). But this is a bit hasty.
On the next page of SAT, Malebranche seems to retract his claim that the under-
standing only receives ideas. ‘[I]t is the understanding that perceives modifications of
the soul, or that senses them’ (SAT I.i.1/LO 3). This is exactly what my reading
predicts: the modification of the soul is the object sensed. Now, Simmons is quite right
that Malebranche cannot abide sensings (what she calls ‘sensations-1’) that lack an
object. But this does not show that sensings are sensings of ideas.
21
SAT E X/LO 621, my emphasis. In this passage, Malebranche warns that we
must not suppose ‘que les sensations ou modifications de nôtre ame puissent repre-
senter les objets à l’occasion desquels Dieu les excite en nous: mais c’est que nos
sensations qui ne sont point distinguées de nous, & qui par consequent ne peuvent
jamais representer rien de distingué de nous,’ although sensations ‘peuvent nous fair
juger qu’ils [bodies] existent’ (OC 3, 142, my emphasis).
22
OC 6, 211; cp. OC 6: 55.
 MALEBRANCHE ON SENSATION

cannot be representational. And it might even generate some sym-

pathy for Malebranche’s rainbow-colored, carrion-scented soul.

7.2 Cartesian Pressures

There are pressures within Descartes’s substance/mode ontology that
make Malebranche’s view hard to avoid. In fact, I believe Male-
branche’s view on this score is exactly what Descartes is committed
to, whether he is aware of it or not.
Let’s begin with the relation between a substance and its essence. If
one thinks of the essence as a property that inheres in the substance,
the substance itself is in danger of becoming a bare particular, char-
acterizable only as ‘that in which essence E inheres.’ Instead, Des-
cartes claims there is only a conceptual distinction between substance
and essence. As he puts it in the Principles: ‘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 itself and extended substance itself—that is, as
mind and body.’23
What then does it mean to call a determinate extension a mode of a
body? As Descartes glosses it, all this amounts to is that ‘one and the
same body, with its quantity unchanged, may be extended in many
different ways.’24 To call extension a mode of body is not to say that
that mode must inhere in some underlying substratum. Rather, the
point is that no determinate extension is essential to the body. The
squareness of this table is just extension existing in a particular way.
Modes, in short, are ways substances/essences have of existing.
The consequences for sensations are dramatic. Descartes is com-
mitted to treating modes of mind and body symmetrically. So just as
the shape of this table is its way of existing, so Bobo’s thought of the
table is a way his mind has of existing. Where these mental modes are
ideas, Descartes can distinguish their objective from their formal
reality. The mind does not become the table when it thinks of it;

23 24
AT VIIIA 30–1/CSM I 215; my emphasis. AT VIIIA 31/CSM I 215.
 MALEBRANCHE ON SENSATION 

instead, it is modified by a mode that has objective being and hence

represents the table. So the mode/content distinction, however one
wants to make this out, allows Descartes to avoid saying that the mind
is modified with the objects of its representational modes.
But if one denies that sensory modes are ideas, and hence that they
have objective reality, there is no way to stave off this counter-
intuitive consequence. It is only by pointing to the representational
nature of ideas that Descartes can resist saying that the mind is
modified by the objects of its ideas. The symmetry of Descartes’s
analyses of modes of thought and extension requires us to say that
a sensation is a way the mind exists. Without a formal/objective
distinction, the putative content collapses back into the mode.
So if sensations are not êtres representatifs, there is no distinction
between their status as modes and what they represent.25 What on a
representational account could be safely sequestered in the content of
an idea must now characterize the mind that has that sensation, just
as square characterizes a square body.26
It is by accepting Descartes’s metaphysics, then, that Malebranche
finds himself committed to his rainbow-colored soul. His peculiar
position is not an idiosyncrasy but a straightforward consequence of
the metaphysics of mind both he and Descartes are committed to.

7.3 The Role of Sensations

For all that, Malebranche cannot afford to make sensations the idle
by-product of sensory experience. Unlike the early Descartes, he has
no alternative mechanism to explain our experiential grip on the
objects in our environment.

25
Here it is intriguing to compare Berkeley’s remarks in A Treatise concerning the
Principles of Human Knowledge Part I §49. Elsewhere (2006), I argue that Berkeley’s
treatment of perception is designed precisely to avoid this Malebranchean consequence.
26
It’s tempting to object that a parallel argument could show that, for Male-
branche, the soul in thinking of extension must itself be extended. But notice that in
the case of the perception of ideas, it is only the act and not the object that exists ‘in’
the soul. By contrast, the objects of sensory acts just are modes of the soul itself.
 MALEBRANCHE ON SENSATION

Extension, whether as instantiated or as it exists in the mind of

God, is devoid of sensible qualities; nor can sensible qualities them-
selves be extended. As Aristes says in the Dialogues, ‘I deny, Theo-
dore, that color is extended. We see it as extended, but our eyes
deceive us.’27 Malebranche’s challenge is to explain this ‘seeing-as.’
Malebranche must use whatever resources he can muster—robust or
anemic natural judgments, intelligible extension, or efficacious
ideas—to explain how our sensory experiences achieve contact with
objects in a material world. The remainder of this section is devoted
to making this last demand precise: what is it that Malebranche thinks
sensations need to do?
All of the functions we’ll look at can be subsumed under one
heading: preserving the mind-body union. Nothing the senses do
conduces to knowledge of the essence of body. By ‘the senses,’
Malebranche means more than just sensation, taken as the mode of
a mind. The faculty of sense is the mind’s passive ability to receive
‘from God ideas mixed (‘confonduës’) with sensations, i.e., sensible
ideas.’28 So we have to be careful to pry apart three things Male-
branche might mean by ‘sensation’ or ‘the senses’: the mode of the
mind, the mode of the mind ‘mixed’ or ‘confused’ with the idea, and
the faculty of suffering such modes and experiencing ideas.29 I shall
always use ‘sensation’ in the strict sense, as the mode of the mind.
First, the senses are required to ‘wake the soul up.’ Pure intellec-
tions don’t make much of an impression on the soul. Without a
sensation, our experience of external objects would be incapable of
making the soul pay attention to those features of the environment
that are relevant to the body’s survival.30 Chief among these, of

27
D 5/JS 75.
28
SAT Conclusion of the first three books, LO 261/1674 416. Note that elsewhere
‘confondre’ and its conjugations are translated as ‘confuse,’ as at LO 52, which speaks
of the ‘four things we confuse’ in every sensation.
29
Note that this list is quite distinct from Malebranche’s own list of the ‘four
things’ we confuse; see SAT I.x/LO 52, discussed below.
30
SAT I.i/LO 2. See SAT I.xii/LO 59. Note that the last sentence of the first full
paragraph on LO 59 does not appear in the corresponding chapter (I.xi) of the 1674
addition.
 MALEBRANCHE ON SENSATION 

course, is the existence of other bodies besides one’s own. As Male-

branche puts it in the Dialogues, ‘the idea represents their [bodies’]
essence, and the sensation informs us of their existence.’31
If sensations are going to help preserve the mind-body union, they
somehow have to be associated with or joined to ideas. To see this,
consider just how the senses are conducive to survival. Summing up
the first book of the Search, Malebranche writes,
We have . . . seen (a) that our senses are accurate and precise for informing us
about the relations our body has with all the bodies surrounding, but that
they cannot tell us what these bodies are in themselves; (b) that their proper
use is only to preserve its life and health; and (c) that they are to be
thoroughly rejected when they attempt to dominate the mind.32

Here again we see ‘the senses’ incorporating both an idea and a

sensation. Indeed, Malebranche seems sometimes deliberately to
blur the two: earlier, he speaks of the ‘idea or sensation of size,’ a
locution he retains throughout all editions of the Search and one that
seems designed to confuse the reader.33
Our chief interest will be in (a). Accomplishing (a) is a necessary
condition for being of use in preserving the mind-body union. Male-
branche here tells us that senses are ‘tres-fideles & tres exacts’ in
informing us of the relative positions of our body and others around
it. Earlier, however, he says that we can only ‘sometimes judge through
sight the approximate relations bodies have to our own as well as among
themselves; but we must never believe that they have the size they seem
to us to have.’34 At a minimum, the senses have to be reasonably
accurate with regard to the relative positions of bodies; getting a cup
of coffee, much less dodging trains, would be impossible otherwise.
The best place to begin is in fact with the roles Malebranche needs
ideas to perform. First, ideas play a role in experience: they are the

31
D 5/JS 74.
32
SAT I.x/LO 85. LO helpfully adds the lettered divisions here.
33
SAT I.vi/LO 32.
34
SAT I.vi/LO 32. Both this text and the one just quoted (SAT I.x/LO 85) appear in
all editions of SAT.
 MALEBRANCHE ON SENSATION

mind’s ‘immediate object’ when it sees a physical object like the sun.
This role as intermediary is the central feature by which Malebranche
defines ideas.35 Immediate perception requires that its object be
present to the mind, and since the mind cannot stroll about the
heavens uniting itself with the sun, it needs some immaterial object
to function as an intermediary.36
Ideas must also play a quite different role, one that is, on its face,
hard to square with their being immediate objects of experience: they
have to function as concepts.37 The very same idea can be involved
both in perception and in pure intellection. The difference between
the two is a function of the vivacity of the sensations, if any, accom-
panying the idea, and of course, if our sense perception is veridical,
the presence of something in the world corresponding to the idea.
Malebranche cannot afford to posit two different ideas, one for the
intellect’s use and one for the senses’. If that doubling happened, there
would be no guarantee that the geometrical properties one deduces in
pure intellection will be instantiated in the world around us.38 The
validity of geometry requires that there be only one idea involved in

35
SAT III.ii.1/LO 217. In 1674, 344–5, he concludes the first paragraph by writing,
‘Ainsi par ce mot idée, je n’entends icy autre chose, que ce qui est l’objet immediat, ou le
plus proche de l’esprit quand il apperçoit quelque chose.’ Later, he adds the claim that
an idea is that which ‘affects and modifies’ the mind. As Nadler (1992, 67 f.) has
shown, Malebranche makes this revision in part in response to an objection from
Arnauld, who claims not to understand what it could mean for something to be
immediately present, or close, to the mind when it perceives. Malebranche thus offers
a causal gloss to his definition, in effect introducing the doctrine of efficacious ideas,
which we’ll examine below.
36
See esp. Nadler (1992, 67 f.) on the ‘strolling soul’ argument.
37
Nadler ingeniously exploits this feature to argue that in fact ideas play only this
role. Malebranche thus becomes a kind of direct realist: we immediately perceive
bodies, and ideas function only as the concepts by which we are able to think of them
(Nadler 1992, 51 f.) Others have critiqued Nadler’s reading at length, and I have no
wish to repeat their arguments here; see esp. Pyle (2003, 50 f.). Although I agree with
these criticisms, Nadler nevertheless brings to our attention a key feature of ideas: they
allow us to think of and indeed to know all of a thing’s actual and possible properties.
It is by consulting the idea of extension, for example, that we are able to see that
sensible qualities cannot belong to bodies (E XI LO 634).
38
I owe this point to Alquié (1974); see esp. his appendix on Kant and
Malebranche.
 MALEBRANCHE ON SENSATION 

perception and abstract thought. Even if one rejects that argument as

fancifully proto-Kantian, its conclusion is bolstered by the logic of
Malebranche’s argument for the existence of ideas. His second such
argument explicitly divorces those ideas from their role in experience.
To think of nothing, Malebranche argues, is not to think at all; and
what one thinks of one, regardless of what is the case with the mind-
independent world, always has some properties or others.39 Ideas,
then, are real beings, since nothing has no properties. Whether this
argument is any good or not is unimportant. What counts is that
Malebranche introduces it as a second support for his conclusion that
ideas exist, and shows no sign that he is positing a separate or
different kind of entity from the one secured by the initial argument
from veridical perception, where the idea of the sun is the immediate
object of experience when one sees the sun. As Malebranche claims in
the Dialogues, there are not two ideas of extension.40
It helps to make sense of this dual role of ideas if we compare
Malebranche with a scholastic like Aquinas. As Thomas Lennon has
argued, Malebranche continually links the intelligible with the gen-
eral. As Malebranche puts it, ‘the mind never sees clearly what is not
universal.’41 Similarly, a core feature of Thomism is the claim that the

39
SAT III.ii.1/LO 217–18. Schmaltz (1996, 110) argues that Malebranche allows
for two kinds of ideas, one intellectual, and hence in God, and one purely sensory, and
hence a modification of the mind. While this is an intriguing position, I do not think it
can be squared with Malebranche’s continual insistence that modes of the mind are
neither ideas nor representations. Nor does the textual evidence seem persuasive;
the chief passage Schmaltz points to (Dialogue 1, OC 12: 46/JS 17) indicates, on the
contrary, that there is just one idea, namely, intelligible extension, that modifies the
soul in the different ways we call sensation, imagination, and pure intellection. And as
Schmaltz is aware, Dialogue 2 (OC 12: 60/JS 29) explicitly conflates sensible and
intelligible ideas. Schmaltz argues that this passage is consistent with his reading, since
a sensory idea, in his terms, is a mode of a soul, and not a separate entity. What
Theodore denies, according to Schmaltz, is only that there are two external causes
of sensations; as a mode of the soul, Schmaltz’s sensory ideas are not external causes of
sensations. I have a hard time seeing how this can be what Malebranche means,
however. For the idea(s) in question are at least causes of sensory experiences; if an
idea were a mode of the mind, it would not be a cause of anything, it would just be a
sensory experience.
40
D 2, OC XII 60/JS 28.
41
SAT I.i/LO 5. I discuss this issue in further detail below (section 10.2).
 MALEBRANCHE ON SENSATION

intellect only knows universals; thinking about particulars requires

turning toward the phantasms.42 On Aquinas’s view, this produces
the rather odd result that what we experience in sensory contexts is
not something the intellect can fully understand. The best it can do is
abstract out the intelligible species and turn toward the phantasms.
But since matter is the principle of individuation, and matter is in
itself unknowable, the intellect never achieves what we might think of
as ‘full cognitive contact’ with the ordinary objects around it.
Malebranche retains the connection between generality and intel-
ligibility. But by replacing the scholastics’ ordinary middle-sized dry
goods with ideas as the immediate object of thought, he removes the
oddity of their view. What the mind directly confronts in perception
is of its own nature intrinsically intelligible.
This maneuver brings its own problems. Where Aquinas has to
explain how we go from the (at best partially) potentially intelligible
particular to the actually intelligible species, Malebranche has to do the
opposite: if we begin with an experience of an inherently intelligible,
general object—an idea of an extended object, or intelligible extension—
he then has to explain how we come to have thoughts of particulars.
Somehow or other, sensations will have to be brought in. In the
Réponse to Arnauld, Malebranche says, ‘the difference of ideas of
visible bodies comes only from the difference of colors.’43 He illus-
trates this with an example: ‘For when on white paper I see a black
body, this [sensation] determines me to see the black body as a
particular body, which, without this color difference, would appear
to me to be the same [as the paper].’44 Similarly, in his exchange with

42
For a good discussion of these points in Aquinas, see esp. Anthony Kenny (1993,
chapter 9).
43
Réponse OC 6: 61. This is a later work (1684) and hence is committed to the
denial of particular ideas of bodies. But the doctrine of intelligible extension need not
be in tension with our quotation: there are, Malebranche might claim, no real
differences among visible ideas, since there is just one of these, viz., intelligible
extension. But insofar as there seem to be such differences, its only explanation is
the essential difference among color sensations.
44
Réponse OC 6: 61. See also D 1/JS 17: ‘ . . . it is only by the variety of colors that
we judge the difference between the objects we see.’
 MALEBRANCHE ON SENSATION 

Régis, Malebranche declares that ‘it is obvious that one sees bodies
only by color and that one can only distinguish them as different by
the difference of their colors. No proof is necessary for this claim,
save for a little reflection on the effects of colors in painting.’45 As
Émile Bréhier aptly puts it, ‘color becomes a sort of “principium
individuationis” in homogenous extension and without it nothing
would be distinct.’46
Although Bréhier is thinking of Malebranche’s doctrine of intelli-
gible extension, according to which there is a single idea in God’s
mind of utterly uniform and infinite extension, we must note that
Malebranche needs the senses to individuate regardless of whether he
holds that doctrine or not. In the early period, Malebranche thinks
there is a single distinct idea for every object we encounter or think
about. Still, there is an important sense in which all ideas, in any
period of Malebranche’s thought, are intrinsically general. Nothing
about the idea of the pine tree to the left of the subject pins it to just
that tree. As Nadler puts it, ‘[e]ven the idea of a particular geometrical
figure (e.g., the right triangle) is really the idea of a kind of figure.’47
What makes the idea of the pine tree a representation of it equally
makes it the representation of any precisely similar geometrical object.
There is a further problem worth noting. In Malebranche’s
example, the only way to distinguish pen from paper is to pair the
sensation of black with the idea of the pen, and white with that of the
paper. Why don’t we experience color as it is, namely, a mode of our
own souls? Why do we attach color to the idea of the pen, and
ultimately the pen itself, at all? Note that this is a why, not a how,
question; the how will be the subject of the coming chapters, and
changes dramatically over time. On the why question, Malebranche is
remarkably consistent.
The broad answer to ‘why’ is teleological: if the purpose of the
senses is to help a mind navigate through its environment, then it
does us no good to experience black or green as a mode of the soul.

45 46
Réponse à Regis OC XVII–1 281. Bréhier (1938, 149).
47
Nadler (1992, 37).
 MALEBRANCHE ON SENSATION

This answer isn’t sufficient, since it does not discriminate between

those sensations we attribute to our own bodies from those we
attribute to bodies outside of us. Both play a role in preserving the
mind-body union; yet only one is experienced as a feature of the
world outside the body.
Malebranche can make two replies. First, he claims that whenever a
body acts on us through the means of imperceptible particles, we
assign our sensation to the object that we take to be the cause of that
motion.48 By contrast, when a body acts on us through visible motion,
as when a sword strikes someone’s leg, the subject assigns the pain
that results to the leg and not the sword. This is a slightly odd view.
Imagine that an object causes pain from a distance, through the
motion of imperceptible particles in the intervening space. In such a
case, we would assign the pain to the distant object, in just the same
way we assign it color. Objects causing pain through radiation would
appear to us to have ‘pain patches,’ just as they can have green or red
patches.49 And pain seems not to be the right sort of thing to cover an
object. (Conversely, if we were aware of the tiny particles that cause
our sensations of colors, ‘we should judge or perceive the colors at the
fundus of the eye just as we judge that heat is in our hand, if our
senses were given us to discover the truth.’)50
Malebranche’s second story appeals to the liveliness of a sensation.
A very strong sensation gets attributed to the soul alone; slightly
weaker ones, to the soul and the object; and weaker ones still to the
object alone. This account, too, is problematic, for a pain can be very
much weaker than a sensation of yellow, and still one will not be
tempted to think of the pain as ‘out there’ in the world.51 In fact, the
example Malebranche gives to illustrate his point replaces intensity

48
SAT I.xi/LO 55.
49
The ‘pain patches’ thought experiment comes from Wittgenstein (1953, 312).
50
SAT I.xii/LO 57.
51
Malebranche might appeal to the mind’s habits here: never having before
experienced such a situation, we would continue to think of the pain as in the body.
That does nothing to account for the counterfactual where all pains are weak relative
to color sensations. In such a situation, it still seems to me that we would not think of
the pain as outside of the body.
 MALEBRANCHE ON SENSATION 

with distance. When you see a candle from a distance, you judge that
the yellow color is only in the candle itself. When it gets very close to
the eye, you judge that the yellow is in both the candle and the eye.52
Here it is the position that the mind judges to obtain that makes all
the difference. Even if Malebranche’s two replies were persuasive,
neither can help us with the how, as opposed to the why. That will
be the subject of the coming chapters.
Since Malebranche’s ontology deprives sensations of representa-
tional power, he will need some more machinery to solve the prob-
lems he has raised. He insists time and again that sensible qualities
are necessary to individuate bodies in experience. He recognizes
that minds—whether ours or God’s—have to localize those sensible
qualities, that is, see them in a determinate region of space. Lastly,
minds have to become aware of the relative positions of bodies,
somehow using sensible qualities to position objects in space. Part
of what drives Malebranche through each iteration of his view is the
simple question: how are all of these processes possible, if sensations
are not representations? Somehow the soul spreads ‘itself onto the
objects it considers by clothing them with what it has stripped from
itself.’53 How?

52 53
SAT I.ii/LO 59. SAT I.xii/LO 58.
 8
Early Malebranche

I’ve divided Malebranche’s work into three stages. The first is repre-
sented only by the first edition of the first three books of the Search.
While much of this material is retained throughout later editions,
it is overlaid with a thick crust of accretions. For this chapter, I’ve
consulted the 1674 edition; an appendix lists the most significant
divergences between this edition and some others.
Most obviously, the doctrines of efficacious ideas and intelligible
extension do not make an appearance in this first edition. Just as
significant are the robust natural judgments Malebranche posits,
which are in effect Cartesian rapporter transplanted into Malebranche’s
scheme, supplemented with a kind of natural geometry from Descar-
tes’s earlier periods. This early account runs into some significant
problems. As we’ll see, Malebranche changes his mind in later versions.
Our job here is to develop the early account, peeling away the revisions
The Search undergoes on its journey through multiple editions.

8.1 Robust Natural Judgments

The structure of sensory experience, on any of Malebranche’s views,
involves both an idea in God’s mind and a sensation in ours. But as
Malebranche is keen to point out, there are other things involved:
In almost all our sensations, there are four different things, which we
confuse, because they are all together, and in an instant, which is the
foundation of all other errors of our senses.
-The first is the action of the object, i.e., in heat for example, the pulse and
the motion of the little parts of the wood against the fibers of the hand.
 EARLY MALEBRANCHE 

-The second is the passion of the organ of sense, i.e., the agitation of the
fibers of the hand caused by those of the small parts of the fire, which
agitation communicates itself up into the brain, because otherwise the soul
would not sense it.
-The third is the passion, the sensation, or the perception of the soul, i.e.,
what each of us senses in spite of himself when he is near the fire.
-The fourth is the judgment the soul makes, that what it senses is in the
hand, and in the fire. For this judgment is so prompt, and the soul is so used
to making it, that it [the soul] considers it [the judgment] to be nothing more
than a simple sensation.1

These judgments become so familiar that we take them for granted,

and cease to be able to recognize them as judgments at all. They are
not judgments we could refrain from, or make otherwise than we do.
That is hardly a barrier to their being the work of the will, however,
since Malebranche, as much as Descartes, denies that it is ‘necessary
that our actions be indifferent to be voluntary.’2
We can in fact discern two layers of natural judgments. The first of
these provides material to be corrected and adjusted by the second.
Here is Malebranche’s discussion of the issue, in 1674:
When we look at a cube, for example, it is clear that all its sides, which we see,
almost never make a projection or image of equal size in the fundus of our
eyes; for this image, which is painted on each of the sides on the retina, is very
similar to a cube painted in perspective. And thus the sensation which we
have must represent the sides as unequal, since they are unequal in the cube
in perspective. Nevertheless we see them all equal by the judgment which we
make, that the parts of the cube that are farthest away cannot form images as
large as those formed by the sides that are closer; this is what corrects the
errors of our senses.
And one must remark that these judgments of which I speak, are not
arbitrary judgments, that we can make or not as it pleases us; they are natural

1
SAT I.ix in 1674, 83–4; the passage is altered in later editions (SAT I.x/LO 52–3,
examined below). It is difficult not to believe that Malebranche is here inspired by
Descartes’s three grades from the sixth Replies. An anonymous referee pointed me to a
list nearly identical to Malebranche’s in the Port-Royal Logic (I.11: 59). I believe the
three grades to be the common source of both Malebranche’s and Arnauld and
Nicole’s catalogs.
2
SAT I.ii/LO 9.
 EARLY MALEBRANCHE

to us, and we cannot form others that are contrary to them. They serve to
correct our senses in a thousand different ways, and without them we would
be fooled almost all the time; but nevertheless they can provide occasion
for error . . . 3

First-layer natural judgments must operate in order to generate

something for second-layer judgments to correct in these thousand
ways. If we left matters with the first layer of judgment, we would end
up living in a deeply skewed phenomenological world. Cubes would
appear to have unequal sizes; table tops would look like trapezoids.
Malebranche captures the sense in which they can look this way. That
skewed picture is exactly what you get if you think away the corrective
judgments of the second level. There, the mind uses the tools of
Cartesian natural geometry to make these corrections.4
Note that, even on this early view, we have yet to find a two-
dimensional ‘sensory core’ that has to be enriched through inference.
The first layer judgments are already providing an awareness of three-
dimensional objects; it’s just that they stand in need of correction. It’s
natural to suppose that the first layer includes a version of Descartes’s
notion of rapporter. Natural judgment is responsible for our thinking

3
SAT 1674 I.vi.4, 54.
4
Malebranche lifts some ideas from the Dioptrique. Here is a comparison of the
two texts:
DESCARTES—Dioptrique AT VI 13740/CSM I 170–2.
D1 the changing shape of the eye and corresponding brain
change (ordained by nature)
D2 relation of the eyes (natural geometry)
D3 changing position of a single eye (natural geometry)
D4 distinctness or indistinctness of the shape plus strength of
the light
MALEBRANCHE—SAT I.viii (1674) or I.ix (later editions)/
LO 41 f.
M1 angle made by rays of our eyes (nat. geo.)
M2 (D1) changing shape of the eye
M3 size of the image in the fundus of the eye
M4 (related to D4) the force with which the object acts on our eyes
M5 (related to D4) distinctness and clarity of the image
M6 relation of judged object to all others in the visual field
 EARLY MALEBRANCHE 

that the sensation is both in and occasioned by the object. Note that
this layer has a propositional component. As Malebranche puts it in
this edition, ‘all those who see the stars in the heavens make a false
judgment, since they judge, that these same stars that they see are in
the heavens.’5 No judgment can be false unless it is a judgment that
such-and-such is the case.

8.2 Problems
In later editions, Malebranche either erases or otherwise takes back
his doctrine of robust natural judgments. It isn’t always clear why; as
we’ll see, the later editions provide some clues. Before moving on, we
should try to assess the early view on its own merits. Whatever
Malebranche’s own motivations turn out to have been, he was quite
right to abandon his early view.
First, consider the relation between judging and understanding.
The understanding’s job is always and only to perceive; the only
differences one can locate within understanding are in terms of its
objects. So far as the understanding is concerned, a simple perception,
a judgment, and an inference are all of them perceptions. In the first
case, its object is a single thing; in judgment, it is the relation between
two or more things; and in inference or reasoning, it is the relation
among the relations of things.6
Understanding here is doing much of the work other philosophers
reserve for the will. What is the will doing, when the understanding
perceives first-order relations? The will judges ‘by voluntarily remain-
ing with what the understanding represents to it.’7 It’s hard to escape
the conclusion that the will is wheeled in only to make us responsible

5
SAT I.xiii: 106. The corresponding text in later editions I.xiv/LO 68, which alters
this passage considerably and interjects a rumination on ‘free’ judgments.
6
SAT I.ii/LO 7. Note that this chapter does not appear in the 1674 edition. That,
however, is a cosmetic matter: the material of 1674’s I.i has been broken into two
chapters.
7
SAT I.ii/LO 8.
 EARLY MALEBRANCHE

for errors, when it does not rest content with what the understanding
proposes to it. Strictly speaking, Malebranche tells us, judgment and
inference take place when the will stops pressing the understanding to
examine its subject in new relations.8 Whether this ‘repose’ ought to
count as judgment is doubtful.
For our purposes, the important point is that judgment, whether
assigned to the understanding, the will, or both, takes as its object
relations. This model fits geometrical inference, conducted in pure
intellection, perfectly. For the true properties of extension, whether in
idea or in the world, are nothing but relations of distance.9
It wreaks havoc, however, where sensations are concerned. Con-
sider how Malebranche argues for the real distinction between mind
and body. In place of a Cartesian conceivability argument, we get, just
as we do with Desgabets, a list of the properties that can belong to
each. Modes of extension can only be successive or relatively per-
manent relations of distance among the parts that make up the body.
Now, no one, Malebranche says, can so much as really think that
pleasure, pain, heat, taste, and color, are really disguised relations
of distance.10
Sensible qualities are known ‘only by an interior sentiment, and
[are such that] one cannot discover the relations they bear to one
another,’ let alone to the idea of extension.11 Sensations differ among
each other essentially, and not merely by degrees, even though
the corresponding brain motions themselves are of the same kind.
Malebranche’s favorite example is tickling and pain: the brain
motions and the motions affecting the skin might differ only in
degree but their associated sensations differ in kind.12 Given this, it
is easy to see why Malebranche would think that sensations cannot be
‘calibrated’ to one another, or stand in relations that we can discern.

8
SAT I.ii/LO 8.
9
‘This idea [namely, the idea of extension] can represent only successive or
permanent relations of distance, i.e., instances of motion and figure, for one can
perceive in extension only what it contains’ (SAT I.x/LO 49).
10 11 12
SAT I.x/LO 49. OC 9: 956–7. SAT I.x/LO 51.
 EARLY MALEBRANCHE 

If we can’t even put sensations in relation to each other, how are we

supposed to put them in relation to the points that make up an idea of
an extended object? Someone might object that the subject need not
be able to see any relations among her sensations in order to see the
mossy boulder as green and the water as gray. All she has to do is get
the green to match up with the boundaries of the boulder, and gray
with the water. Fair enough. But how is she to do this, given the gulf
between sensations and extension? Sensations are, after all, modes of
the mind and not representations.
A second problem with Malebranche’s account is much more
straightforward: the early view is viciously circular. If it’s up to the
subject to judge that this rock is green, she must first have individu-
ated the rock from its surroundings. And yet the only way to do this,
Malebranche says again and again, is by means of color. But she
cannot use color sensations to individuate bodies if this process
presupposes that those bodies are already individuated. The same
point holds, mutatis mutandis, for the other sense modalities.
And if the subject can’t get off the ground where individuation is
concerned, all efforts at positioning are equally doomed. In short, the
early view is vulnerable to the same kinds of arguments Malebranche
deploys against Descartes in his selection argument.
 9
Middle Malebranche

Although Malebranche never wavers from his claim that we need

sensations in order to experience the world of extension, his explan-
ation for how they help changes dramatically. In later editions of the
Search, he rejects the early view of robust natural judgments. He
does so, not by removing that material, but by supplementing it. In
one case, he simply adds a note that cancels the reader’s impression
that the mind is in any way active in its natural judgments.1 Male-
branche’s corpus contains other relevant innovations, particularly
the doctrines of efficacious ideas and intelligible extension. These
are the subject of the tenth chapter. My purpose here is to see how
well Malebranche’s revised view can get along without his later
doctrines.2

1
See the note at SAT I.ix/LO 41.
2
Someone might argue that, since the Elucidations appear (in a separate
volume) in 1678, along with the first revision of Books I–III of SAT, there really
is no gap between Malebranche’s doctrine of anemic judgments and his introduc-
tion of intelligible extension in E. The first reason to treat the two (anemic
judgments and intelligible extension) separately is philosophical: since one does
not entail the other, it makes sense to ask whether they can help explain perception
independently of each other. The second reason is textual. Intelligible extension,
perhaps oddly enough, only really appears in E X; in the body of SAT, it never
makes more than a cameo appearance. Finally, E X, as I shall argue, in fact fits
poorly with the doctrine of anemic judgments, since even after introducing intel-
ligible extension, Malebranche continues to speak of the mind projecting or
spreading its sensations on to objects, a task only robust natural judgments can
accomplish.
 MIDDLE MALEBRANCHE 

9.1 Five Things We Confuse

in Sensory Experience
On the surface, Malebranche retains his list of four things we confuse
in sensation. The first three elements are the same. Even the last seems
untouched: it ‘is the judgment the soul makes when it perceives that
what it perceives is in the hand and in the fire.’ But Malebranche’s new
gloss on natural judgment is startling: ‘Now, this natural judgment is
only a composite sensation (‘sensation composée’), but the sensation or
natural judgment is almost always followed by another, free judgment
that the soul makes so habitually it is almost unable to avoid it.’3 So we
have five things, not four, happening in sensory experience:

i) the act of the object

ii) the change in the sense organs
iii) sensations (modes of the mind)
iv) composite sensations (what I’ll call ‘anemic natural judgments’)
v) free (though habitual and difficult to resist) judgments4

The robust natural judgments of the first edition might have been
taken by the soul to be sensations because of their familiarity; for all
that, it was clear that they were in fact judgments, with a truth value
(namely, false). Among these robust natural judgments, I’ve argued, is
a second layer that is capable of correcting for distortions and illu-
sions in the frst. Malebranche now changes course dramatically:
natural judgments really are (and are not mistaken for) sensations.
Genuine judgment comes in only at (v), where we make free judg-
ments that can be resisted.5 In principle, one can, it seems, be cured of

3
SAT I.x/LO 52. LO’s translation omits the word ‘composée.’
4
As an anonymous referee pointed out to me, the judgment at stage (v) is free and
can be resisted. So it seems odd to suppose that that judgment is really part of our
sensory experience. I suppose Malebranche might mean simply that most people
make the free judgment alongside sensory experience and fail to distinguish it from
the other four elements he isolates.
5
Consider SAT I.xiv/LO 68: since the stars we immediately see are not in the
heavens but in the mind of God, ‘everyone who sees the stars in the heavens makes a
false judgment and who then voluntarily judges that they are there performs two false
 MIDDLE MALEBRANCHE

the tendency to form the belief that sensible qualities are in bodies;
such a philosophical saint would still have the compound sensations
necessary to think of the objects in her environment.
The precise relationship between the new free and natural judg-
ments is hard to discern. At times, it looks as if they are affirmations
of the same propositional content: ‘for men not only judge through a
natural judgment that pain, for example, is in their hand, they also
judge it by a free judgment; not only do they feel it there, they believe
it to be there . . . ’6 So the difference is between feeling that something
is the case and believing that it is. The problem is that we can negate
the free judgment while still making the natural one. Indeed, this is
just what the philosopher who has thrown off the beliefs of childhood
should do. Such a paragon of wisdom would then be involved in
inconsistency at every moment of perception, helplessly affirming the
very same proposition she willfully denies.
However that may be, we should wonder in what sense anemic
natural judgments are really judgments at all. Here is Malebranche’s
fullest defense of his use of the term to describe mere sensations. (This
passage is inserted at the very end of the chapter on distance
perception.)

I feel I must warn again that judgments about the distance, size, and so on, of
objects are formed in the ways I have just explained, not by the soul, but by
God according to the laws of the union of the soul and the body. I have
therefore called these judgments natural in order to emphasize that they
occur in us independently of us, and even in spite of us. But as God fashions
them in and for us in such a way that we could form them ourselves if we
knew optics and geometry as God does, if we knew everything that occurs in

judgments, one of which is natural, the other free. One is a judgment of the senses or a
compound sensation, which is within us, occurs independently of us, and even in spite
of us, and according to which no judgment should be made. The other is a free
judgment of the will, which can be avoided, and which consequently we must not
make if we are to avoid error.’ The corresponding passage in the 1674 edition (I.xiii/
106) is content simply to say that those who see the stars in the heavens make a false
judgment. There is no talk of this error as involving a compound sensation and a free
judgment.
6
SAT I.iv/LO 69. Note that this passage does not appear in the first, 1674 edition.
 MIDDLE MALEBRANCHE 

our eyes and brain, and if our soul could act on its own and cause its own
sensations, I attribute to the soul the performance of judgments and infer-
ences as well as the subsequent production of its sensations, which can be the
effect only of an infinite power and intelligence.7

All the work of perceptual cognition has been outsourced to God.8 It

can be attributed to us only in the counterfactual sense that if we
knew as much as God does, we would make these judgments for
ourselves. (By the same reasoning, I can be said to make all the same
moves in chess that a grand master does, because if I knew as much as
Karpov, I would move my pieces in the way he does.) But again, since
these judgments are just compound sensations, what Malebranche
must mean is that we would combine our sensations in the way God
combines them for us.
In fact, it’s far from obvious just what a sensation composée is
supposed to be. The notion is best explored by examining its origins,
in Malebranche’s own critique of his earlier view. In the 1674 version
of I.vi (in LO, I.vii), Malebranche posited such judgments to correct
for the initial presentation of the cube as having unequal sides.9 This
second layer of robust judgment is responsible for our grasp of the
sides of the cube as in fact equal.
But in later editions, Malebranche presents his old account as if it
were merely something that ‘might be said.’ Arguing against his own
earlier self, he writes, ‘ . . . [A]s it is given to the senses only to sense and
never, properly speaking, to judge, it is clear that this natural judgment

7
SAT I.ix/LO 46–7.
8
See Theo C. Meyering (1989) for an intriguing discussion of Malebranche in light
of cognitive science. Meyering takes Malebranche to anticipate some contemporary
developments, with the exception, of course, of the outsourcing of cognitive ‘processing’
to God.
9
There is some irony in the fact that, when Locke argues against Malebranche’s
later position, he ends up doing so on behalf of the earlier one. Locke writes, ‘In the
next place where he says that when we look on a cube “we see all its sides equal.” This,
I think, is a mistake; and I have in another place [presumably Essay II.ix] shown how
the idea we have from a regular solid, is not the true idea of that solid, but such an one
as by custom (as the name of it does) serves to excite our judgment to form such an
one’ (‘An Examination of P. Malebranche’s Opinion of Seeing all Things in God,’ in
Locke 1823, vol.8, section 16).
 MIDDLE MALEBRANCHE

[which presents the sides of the cube as equal] is but a compound

sensation that consequently can sometimes be mistaken.’10 Thus even
the correction of the original sensory experience, which one might
have thought had the best chance of surviving as a genuine judgment,
is now reduced to a compound sensation.
On its face, it isn’t obvious why Malebranche thinks the sensation
can be true or false simply in virtue of being compound. There is no
doubt that he has in mind a scholastic commonplace, traceable back
to Aristotle’s On Interpretation:
[F]alsity and truth have to do with combination and separation. Thus names
and verbs by themselves—for instance ‘man’ or ‘white’ when nothing further
is added—are like the thoughts that are without combination or separation;
for so far they are neither true nor false.11

Combining or separating simple elements allows truth and falsity to

come on the scene. Malebranche might well be thinking that this
commonplace justifies his attributing truth or falsity to anemic nat-
ural judgments. The problem comes when we look at the next line
from Aristotle’s text: ‘[E]ven ‘goat-stag’ signifies something but not,
as yet, anything true or false—unless ‘is’ or ‘is not’ is added.’12 It is not
just any joining of simple elements that generates truth and falsity;
only joining or separating by ‘is’ or ‘is not’ can do that.
So far, I have been talking as if compound sensations were one or
more sensations somehow combined by the mind. But there is at least
one text that undermines that assumption. Malebranche explains that
he calls a natural judgment a ‘compound sensation’ only because

it depends on two or more impressions occurring in the eye at the same time.
When I look at a man walking toward me, for example, it is certain that, as he
approaches, the image or impression of his height traced in the fundus of my
eyes continually increases and is finally doubled as he moves from ten to five feet

10
SAT I.vii/LO 34.
11
On Interpretation 16a11–16. I am indebted to an anonymous referee for direct-
ing me to this text and suggesting it lies behind Malebranche’s confidence that anemic
natural judgments achieve truth values.
12
On Interpretation 16a17–18.
 MIDDLE MALEBRANCHE 

away. But because the impression of distance decreases in the same proportion
as the other increases, I see him as always having the same size. Thus the
sensation I have of the man always depends on two different impressions, not
counting the change in the eyes’ position and other matters . . . 13

Anemic natural judgments are compound sensations only in a histor-

ical sense: they are the result of multiple impressions occurring in the
eye. The sensation that we ultimately experience might be uniform. We
experience the hawk as uniform in color even as it moves from sunlight
to shade. But we must keep in mind that Malebranche’s ‘impressions’
are purely physical, taking place in the eyeball. So he is at most
describing the states of the extended world that occasion God’s causing
the relevant experiences in us. Whether compound sensations are
composed of two different mental states or simply occasioned by
more than one physical impression in a sense organ, they seem incap-
able of doing the work Malebranche needs them to do. We need a
substitute for robust natural judgments that admit of truth and falsity,
and neither reading of sensation composée does the trick.
The theory of anemic natural judgments is Malebranche’s attempt
to walk a very fine line. On one side lies the earlier view: a set of
unconscious or quasi-conscious inferences that allow us to produce
a full-fledged sensory experience of ordinary three-dimensional
objects. On the other side lies the abyss of blank, non-representational
sensations. Merely compounding two such sensations hardly seems
sufficient to generate the robust end-product of sensory experience.
Moreover, we have seen that Malebranche needs the natural judg-
ments to have propositional content. But mere sensations do not have
such content: one needs ideas for that. Let’s turn now to one possible
means of navigating between these two perils.

9.2 Seeing-as
It’s not hard to see why Malebranche might want to abandon
robust natural judgments. First, the early view is viciously circular: it

13
SAT I.vii/LO 34. This text does not appear in the 1674 edition.
 MIDDLE MALEBRANCHE

presupposes just the ability it is trying to explain, namely, our ability to

individuate objects in sensory experience. Second, the change is in
keeping with Malebranche’s Augustinian theology, which emphasizes
our dependence on God. On the new view, individuating objects, local-
izing sensations, positioning bodies relative to each other, and correcting
for distortions are not the work of feeble finite minds like ours.
Nevertheless, Malebranche still needs some account of how these
three tasks are accomplished, whoever is doing the work. Can anemic
natural judgments explain sensory perception? Above, we considered
Deborah Brown’s intriguing suggestion that Cartesian rapporter is
‘seeing-as.’ Lawrence Nolan uses this same notion to explicate
Malebranche’s natural judgments. And at first sight, this is a tremen-
dous improvement over Malebranche’s early view.
Nolan of course recognizes that a natural judgment is merely a
compound sensation and hence nothing like Descartes’s rapporter,
which corresponds to the first layer of robust natural judgments. But
neither, he claims, is it a mere mode of the mind, devoid of inten-
tionality. Instead, a natural judgment is a case of ‘seeing-as.’ In
perception, the soul is modified by sensations and in this way is led
to see the idea in God’s mind as red or smelly or what have you. To
the circularity objection, Nolan’s notion of ‘seeing-as’ can offer a
reply. An object is presented to the mind in sensation as already
individuated and colored. It is not as if I am first presented with an
unindividuated object and then must project color on to it in order to
make it particular. As Malebranche now says, ‘it is not up to us to
assign the sensation of whiteness to snow or to see it as white.’14
Thus the individuation problem disappears. Compound sensations
come fully-formed as of objects with colors and dimensions. When
we reflect and ‘freely’ judge that colors are outside of us, our error is
explained by two facts: that objects ‘act’ on us through the motion of
invisible particles to produce color sensations, and that color sensa-
tions are ‘weak’ as compared to sensations of pain.15 But neither of

14 15
SAT I.xi/LO 55. SAT I.xii/LO 57 f., discussed above, section 7.3.
 MIDDLE MALEBRANCHE 

these forces plays a role in our natural judgments. These are inflicted
on us by God.
Nolan’s interpretation strikes me as exactly right. But to my mind it
re-locates, rather than solves, the problem. For if natural judgments
are not a matter of Cartesian rapporter, where the mind takes an
active role in projecting sensations on to objects, we are left with no
real way to understand them. The notion of seeing-as seems designed
to navigate a middle course between a robust judgment and mere
co-occurrence of sensations. For my part, I think this middle course is
illusory: seeing-as must collapse into one or the other. To see this, we
need to consider the notion of ‘seeing-as’ in more detail.
Its use calls to mind duck-rabbits, where a Gestalt shift can present
one with either a duck or a rabbit. I take it that neither Brown nor
Nolan means to use the term exclusively in this way: no Gestalt shift
can make the table in front of me look anything other than white.
Now, at the second layer of anemic judgments, I suppose there might
be just this kind of Gestalt switch, where the sides of the cube can
appear either equal or unequal. But for the lower level judgments,
which involve individuation and positioning, this notion of ‘seeing-
as’ does not seem to apply.
What, then, is left? I suppose another use of ‘seeing-as’ involves
concepts: someone without the concept horse can fail to see a horse as
a horse. But anemic natural judgments are compound sensations, not
complexes of sensations and concepts or ideas.
A final option is to say that ‘seeing x as a cube’ is meant, not to
describe the deployment of a concept, but the contents of the experi-
ence. One sees that x has such-and-such a shape, whether that shape
is one for which one has a concept and a word or not. In that case,
however, I fail to see how ‘seeing-as’ does much more than re-describe
our original problem. ‘Seeing x as F’ in this sense is just a paraphrase
of the more perspicuous ‘taking it that x is F,’ in which case it is an
old-fashioned judgment, after all.
Even if we leave this point aside, a further question obtrudes itself.
Why is it that a sensation on its own is not a seeing-as, but when
compounded with others, it becomes one? A natural answer is that a
 MIDDLE MALEBRANCHE

sensation of white takes up a certain region of visual space, and

defines a region of extension, only when it is combined with the
sensations of other colors that limit its extension. But Malebranche
is not entitled to say this, since colors do not define regions of
extension except insofar as they are paired with an idea of extension.
And it is precisely this pairing we are trying to understand.
It is worth remembering that this story is supposed to apply to all
sensations, of any modality. Take two tastes, or one taste and one
smell, or any combination you like. It is hard to see how such
aggregates of sensations can allow us to ‘see-as’ (or ‘taste-as,’ or
‘smell-as’). Mere aggregation seems incapable of transforming sen-
sations into judgments, in however attenuated a sense you like. And
above, we saw that it is not even necessary that there be two sensations
for the sensation to be compounded, in Malebranche’s idiosyncratic
sense.
The core problem is this: where, in a sensation composée, does an
idée figure? If the whole point of natural judgments is to attach our
sensations to ideas, and through them to objects, treating natural
judgments as composite sensations is bound to be a serious mistake.

9.3 Compound Sensations as Perceptions

of Relations
It would be a shame to leave the middle Malebranche in this predica-
ment. Perhaps there are other resources we have not yet exploited. In
this section, I look at two possible moves before turning, in the next
chapter, to Malebranche’s final view, with its doctrines of efficacious
ideas and intelligible extension.
There might be more to the notion of ‘seeing-as’ than I have so
far allowed. Above, I distinguished several senses of the phrase and
argued that none quite fits the bill. But if we recall Malebranche’s
theory of the understanding, a new gloss on ‘seeing-as’ suggests itself.
Recall that ‘there is no difference on the part of the understanding
between a simple perception, a judgment, and an inference, other
than that the understanding by a simple perception perceives a simple
 MIDDLE MALEBRANCHE 

thing without any relation to anything else whatsoever, that in judg-

ments it perceives the relations between two or more things, and
that in inferences it perceives the relations among the relations of
things.’16 Suppose a compound sensation involves, not just two (or
more) sensations, but the mind’s being aware of the relations between
the two. True, this reading has to ignore the fact that a composite
sensation, for Malebranche, need not be a composite of two sensa-
tions. For as we have seen, such a sensation can be composite only in
the sense that it causally depends on more than one impression or
image in the eye. I propose that we waive that objection and see what
can be done with this new proposal.
On this reading, when the subject senses the mossiness of the
boulder with her hand and its green color with her eyes, she also
detects or thinks of the relation between them; in this case, she thinks
of them as belonging to a single object. Moreover, she can put this
collection of sensations into relation with other sensations, say, the
brown bark of the tree. Now, she can’t really be doing this; it must be a
feature of the sensations that God inflicts on her.
In what sense is this a judgment? Malebranche assigns the detec-
tion of relations to the understanding, not the will. But since we’ve
already abandoned any hope of assigning natural judgments to the
will in any case, perhaps this isn’t a high price to pay. What God does
is provide the sensations, complete with their relations; all the subject
does is see those relations.
The chief problem here is that sensations, as Malebranche insists
throughout his career, simply do not stand in relations to each other,
or at least, in any relations the mind can detect.17 Sensations are
modes of the mind and hence do not admit of the kinds of systematic
combination or re-combination that modes of body do.

16
SAT I.ii/LO 7. Although Nolan does not discuss this text, he does seem to think
that compound sensations involve relations.
17
See OC 9: 956–7, discussed above. Of course, as an anonymous referee points
out, sensations do stand in some knowable relations to other things: they are related to
souls as modes, for instance.
 MIDDLE MALEBRANCHE

9.4 Conclusion
Anemic natural judgments are not without their attractions. They can
sidestep the circularity objection, since the soul is passive in ‘making’
these judgments. What is more, they obviate the need for processing
to be taking place beneath the level of consciousness, something that
always fits poorly with the transparent Cartesian mind.
Unfortunately, anemic natural judgments are unable to play the
roles Malebranche assigns them. Consider the very first thing Male-
branche needs them to do: they have to ‘wake the soul up’ to its
environment. But as modes of the mind, they ought, as Desgabets
thought, to direct the mind to itself, if anywhere. They certainly get
the mind’s attention, since they come to it involuntarily. But what the
mind attends to is a mode of itself.
Natural judgments are also supposed to explain how we can think
of heat as in the fire, or mossiness as on the boulder. If a natural
judgment is just a sensation composée, it does not include or make
reference to an idea. There is, then, no way to connect the mossiness
or heat the subject experiences to the object outside the mind, which
has to be represented by an idea.
Nor is the individuation problem solved. Recall that Malebranche
happily bites the bullet and claims that when the soul perceives color
C, it exists C-ly. If he is willing to tough out the counter-intuitive
consequences of such a view and really means to say that the soul
exists black-ly when it senses black, he is left with no resources to
individuate ideas by means of color. To do this, he needs the soul to
experience at least two colors simultaneously, so that one color can
form the boundary of another. At the risk of seeming silly, we would
then have to ask, can a soul be black all over and white all over at the
same time? If instead it is somehow ideal ‘regions’ of the soul that
turn black, how are these to be mapped on to the idea?
The fundamental problem is how to knit together the blank modes
of a finite mind and the ideas in the mind of God. The geometrical
qualities exist in God; sensible qualities, the objects of sensings, in us,
as modes. The immediate object of an act of sensory perceiving, then,
 MIDDLE MALEBRANCHE 

is a hopelessly bifurcated creature: some of its qualities exist in God,

others in us. The best Malebranche can do, I think, is to say that God
causes sensations in us when he reveals an idea to us. But that
connection, I have argued, is simply too loose. Seeing a white horse
is nothing like thinking of an octagon and feeling hungry.
To sum up: anemic natural judgment does nothing to explain how
sensations function in perceptual experience. If Malebranche simply
stipulates that in natural judgments we are presented with already
individuated, sensible, and particular ideas, then he has named a
problem, not solved it. That is not necessarily a criticism of Male-
branche: perhaps, given his materials, that is the best he could do.
Nevertheless, Malebranche does introduce two innovations—the doc-
trines of intelligible extension and efficacious ideas—into his account
of sensory perception. It is to those innovations that we should
now turn.
 10
Later Malebranche

Within a few years of publishing the first three books of the Search,
Malebranche detects a serious problem with his theory of ideas:
seeing a body through a single idea would prevent us from experi-
encing changes in perspective while perceiving one and the same
object. In response, he introduces the doctrine of intelligible exten-
sion, according to which there is only one idea of extension in God.
This new claim exacerbates the problems we’ve been examining.
Whether as a result of his sensitivity to these problems or not,
Malebranche also introduces the doctrine of efficacious ideas.
Attributing causal power to ideas fits well with the doctrine of
anemic natural judgments. And it might help us understand how
mere composite sensations are supposed to allow for the individu-
ation and positioning of objects, along with the localization of their
sensible qualities.1

10.1 Intelligible Extension

The phrase ‘intelligible extension’ doesn’t appear in the body of
the Search until 1700, and then only as an example mentioned in

1
By my estimation, there is at least a ten-year gap between the introduction of the
doctrine of intelligible extension (1678) and the first clear appearance of efficacious
ideas (1695), though Schmaltz traces the doctrine to the Dialogues of 1688 (Schmaltz
2000, 78). An anonymous referee suggested that the doctrine seems to be present in
Dialogue II.11: 29, where Theodore admonishes Aristes to ‘[j]udge the reality of ideas
not by the sensation you have of them which confusedly indicates their action to you,
but by the intelligible light which reveals their nature to you.’
 LATER MALEBRANCHE 

passing.2 The doctrine is introduced in the tenth Elucidation,

appended to the edition of 1678.3 Some philosophers—such as
Lennon and Nolan—take Malebranche at his word when he claims
only to be clarifying and not altering his view.4 They argue that
Malebranche believes from the start that what is intelligible is inher-
ently general. There is, then, no room for ideas of particular bodies.
In my view, this argument embodies a subtle confusion. That the
intelligible is general does not entail that it is without shape or size.
Even the idea of a particular cube is inherently general already: it
represents any cube of the same dimensions. Extension need not be
entirely undifferentiated in order to be general. Hence the addition of
the tenth Elucidation is a genuine change in Malebranche’s position.
For now Malebranche rejects the claim that there is a one-one
correspondence between individual bodies and ideas in God’s mind;
instead, there is only one such idea, namely, intelligible extension. We
should begin by noting his motivation for introducing intelligible
extension:

[W]hen I said that we see different bodies through the knowledge we have of
God’s perfections that represent them, I did not exactly mean that there are
in God certain particular ideas that represent each body individually, and
that we see such an idea when we see the body; for we certainly could not see
this body as sometimes great, sometimes small, sometimes round, sometimes
square, if we saw it through a particular idea that would always be the same.5

Distance changes the apparent size of objects, and this apparent size
can only be corrected through further natural judgments. It remains
the case that one and the same material object seems now large, now
small, depending on the perceiver’s distance and perspective. But if
the immediate object of thought is only an idea, how can it undergo

2
See IV.11/LO 321. For more on the dating of the doctrine, see esp. Rodis-Lewis
(1963, 82).
3
Malebranche 1678, 547 ff. See also Robinet’s n.80 in OC III 366.
4
(2012, 27): ‘I agree with Thomas Lennon (“Malebranche’s Argument for Ideas,”
61–2) that Malebranche always subscribed to the “intelligibility principle,” which
asserts that only general things are intelligible.’
5
E X/LO 627-8; cp. Radner (1978, 86).
 LATER MALEBRANCHE

these changes and yet remain the same? Ideas are, after all, supposed
to be eternal and immutable. So Malebranche now rejects his
earlier claim (or, as he would have it, his earlier façon de parler).
Instead of a duplication of objects in an array of divine ideas, there
is now said to be only one such idea, which is both infinite and
entirely uniform.
Once again, we must keep in mind that the idea we access in
intellectual thought is numerically identical to the idea we encounter
in sensory experience. Some readers of Malebranche have been
tempted by the parallels with Kant’s theory of space: both are
necessarily singular and I suppose in some sense ‘contain’ a mani-
fold of possible figures (though later on I will raise some suspicions
about that).6 In this case, it might be more enlightening to look
backward. In Metaphysics Z, Aristotle introduces the notion of ‘hulé
noété,’ ‘intelligible matter.’ Intelligible matter is the matter of ‘the
objects of mathematics.’7 If such objects are to be real in any
sense, they must, like everything else, be a form/matter compound.
Unlike intelligible extension, Aristotle’s intelligible matter is
abstracted from ordinary objects; it is ‘present in sensible things’
though ‘not qua sensible.’8 Nevertheless, both serve as the subject
matter for geometry; and both are featureless in the sense of
being without internal division.9 It may well be that Malebranche
intends his view to be received as a deliberate variation on this
Aristotelian theme.
Having introduced intelligible extension, Malebranche now needs
sensations to do more than just pin down an idea of a determinate
object with fixed dimensions. It is not just that an idea of a cube is
inherently general; there is now no idea of a cube, only an utterly

6
See Alquié (1974, 236) and Radner (1994, 68). Alquié adds in an appendix, ‘[L]ike
Kantian space, intelligible extension plays, in Malebranche, a dual role. It has its own
laws and, as such, offers to the mind the essential structure of mathematics. At the
same time, it serves to make the content of sensation objective’ (1974, 506).
7 8
Metaphysics Z 1036a11. Metaphysics Z 1036a11–12.
9
For a helpful discussion of intelligible matter in Aristotle and its role in geometry,
see Gaukroger (1980).
 LATER MALEBRANCHE 

homogeneous idea of extension. Color must ‘light up’ a region of

intelligible extension to give it a cube-ish shape.10
It is not up to us to go in search of the idea: God ‘applies’ the
idea of extension to the soul ‘in a thousand different ways.’ Notice
that Malebranche does not speak of the idea itself as a cause, as he
will later; he still speaks of the mind ‘spreading’ its sensations on to
the idea. Since God applies the idea to us in so many different ways,
‘intelligible extension contains all the differences of bodies due to
the different sensations the soul spreads [‘répand’]11 on the ideas
affecting it upon the occasion of these same bodies.’12 We are also
said to ‘attach’ the sensation of color ‘to different parts of ’ intelli-
gible extension.
It’s vexing to find Malebranche still speaking of ‘ideas’ precisely
when he’s at pains to say there is only one idea of extension in God. In
the Dialogues, he explicitly treats parts of intelligible extension as
ideas.13 The trouble is that intelligible extension cannot really have
parts at all. These ‘regions’ (for they cannot really be regions) of
intelligible extension are individuated only from the point of view
of the finite mind that attaches sensations, especially color sensations,
to them.
For us the interesting question is, on what basis does the soul
spread or attach its sensations to one region, rather than another?

10
‘[W]hen you sense or see [the circle], a determinate part of this [intelligible]
extension sensibly touches your soul and modifies it by the sensation of some color.
For intelligible extension becomes visible and represents a certain body in particular
only by means of color, because it is only by the variety of colors that we judge the
difference between the objects we see. All the intelligible parts of intelligible extension
are of the same nature insofar as they are ideas, just as all the parts of local or material
extension have the same nature as a substance. But as the sensations of color are
essentially different, by means of them we judge the variety of bodies’ (D 1/JS 17; cp.
Third Letter to Arnauld, OC 9: 959).
11
Throughout Elucidation X, LO renders ‘répand’ and similar words as ‘project.’
I prefer the more literal ‘spread.’
12
E X/LO 628.
13
‘All the intelligible parts of intelligible extension are of the same nature insofar as
they are ideas, just as all the parts of local or material extension have the same nature
as a substance. But as the sensations of color are essentially different, by means of
them we judge the variety of bodies’ (D 1/JS 17).
 LATER MALEBRANCHE

Consider what the picture now looks like. The subject is in a forest,
seeing and touching a mossy boulder. She undergoes the modifica-
tions of the soul we call ‘green’ and ‘mossy.’ God impresses the idea of
extension on her. Now she must spread or attach her sensations to
this idea. But without any awareness of the states of her brain and
body, it is impossible to see what would guide her in this act of
spreading. Note that I am not assuming that she does this con-
sciously, or voluntarily, if that means she could choose to attach her
sensations otherwise than she does. The problem is generated merely
by the fact that Malebranche says that she is doing the attaching.
The reader will have noticed a further oddity. Elucidation X is
included in the 1678 edition of the Search. And that edition includes
the doctrine of anemic natural judgments. But on that doctrine, the
soul is entirely passive in perception. The same curious juxtaposition
happens in the 1688 Dialogues, where Malebranche defends the
doctrine of anemic natural judgments but continues to speak of the
soul as attaching colors to an idea.14 Malebranche should not be
telling us that it’s up to the subject to do the spreading and attaching;
she should simply be suffering compound sensations.
Since he retains this activity on the side of the subject, we have to
re-evaluate the argument that introduces intelligible extension. The
problem is supposed to be this: the one body–one idea view sacrifices
the numerical identity of the immediate object of experience over
time, since the immediate object changes with perspective. But sup-
pose we were still working with robust natural judgments. The very
same idea might figure in those judgments, which are in part respon-
sible for correcting the defects of perspective and achieving size
constancy. If Malebranche retains the doctrine of robust natural
judgments, the introduction of intelligible extension is unmotivated.
It is only after the mind has been deprived of its activity, with the
doctrine of anemic natural judgments, that one needs to posit

14
Malebranche comes close to paraphrasing Descartes’s Principles I.70: ‘by a kind
of natural judgment, which is not free, we always refer sensations to what most suits
the good of the body . . . as for color, we refer it to objects alone’ (D 12/JS 219).
 LATER MALEBRANCHE 

intelligible extension to account for the numerical identity of the

idea one perceives.
Malebranche will eventually go on to reject the ‘spreading’ view we
find in Elucidation X. He will combine the doctrines of intelligible
extension and anemic natural judgments with another innovation:
efficacious ideas.

10.2 Efficacious Ideas

Malebranche seems not to have felt any tension between his doctrine
of anemic judgments, present in the Search from 1678 on, and his
claim that the mind attaches or spreads sensations on to ideas. In his
1684 Réponse to Arnauld, he claims that ‘the mind must have the idea
of extension, so it can attach, so to speak, the sensation of color to it:
just as a painter needs a canvas, so he can apply colors to it.’15
When he begins to think of ideas themselves as causes, however, he
inverts the metaphor.16 In 1704, he writes,

In order to see different bodies, it suffices that the idea of extension or

intelligible extension affects or touches the soul with diverse colors. For
indeed, painters need only an extended canvas and diverse material colors
in order to represent all visible bodies, by distributing diverse material colors
on their canvas in accordance with their art.17

Instead of the idea being the canvas and the mind the painter, now the
soul is the canvas, to be painted on by intelligible extension. However
odd it sounds, this reversal of the metaphor makes for a much better
fit with the mature Malebranche’s doctrine of anemic judgments.
And indeed the doctrine of efficacious ideas might hold out some
promise for solving the problems I have raised. This is Nolan’s, as well

15
Réponse OC 6: 78. For a useful discussion of Malebranche’s use of the canvas
metaphor, and how it changes over time, see esp. Lawrence Nolan (2012).
16
I am indebted to Nolan’s excellent (2012, 29) work, which documents the shift in
metaphors.
17
OC 9: 1066; trans Nolan 2012, 29.
 LATER MALEBRANCHE

as Schmaltz’s, position.18 Note that, with efficacious ideas in place, all

talk of projection or spreading drops out. Instead, we can be said to
‘sense’ the idea of extension when it modifies our minds. Once we see
that natural judgments are just compound sensations, there is no
special problem in combining them with efficacious ideas. But have
we made any gains in understanding how the overall view is supposed
to work?
Combining the two new doctrines, we can say, with Nolan, that
intelligible extension causes us to think of it in different ways
‘depending on how it affects the soul.’19 We see a region of intelligible
extension as extended, under certain defined limits. Nolan acknow-
ledges that we could not do this without sensations. On his view, as on
Simmons’s, sensations are intentional in the sense that they are
directed at regions of intelligible extension.
But if anything, intelligible extension makes it more difficult to see
how an idea can act on the soul. To see this, let’s turn back to
Malebranche’s example of a black body (say, a pen) resting on a
piece of white paper. In virtue of what does (the idea of) the pen
appear black, and (the idea of) the paper, white? There is nothing
intrinsic to the idea(s) that can explain this, because as we’ve noted
there are no real regions in intelligible extension, and even if there
were, they would be entirely uniform. And how can a merely ideal
region of intelligible extension act on my soul? If it acts at all, it must
do so as undifferentiated, utterly uniform extension.

18
Schmaltz recognizes that it is unclear ‘what it could mean to say that an idea
‘becomes sensible’ by means of sensation,’ or that the soul ‘attaches’ colors to an idea.
Schmaltz argues that it is only when Malebranche has efficacious ideas in place that he
is in a position to cash out these metaphors (Schmaltz (2000, 76 f.)). Nolan (2012, 31 f.)
puts similar weight on efficacious ideas. As Susan Peppers-Bates (2005) has argued,
however, that doctrine is in tension with Malebranche’s official view that God alone is
a true cause. Calling ideas efficacious can at best be a misleading way of speaking.
Although I find Peppers-Bates’s case convincing, I shall table the issue here, in order
to give the fairest possible hearing to those views that depend on imbuing ideas with
causal power.
19
Nolan (2012, 41). In a similar vein, Jolley (1994, 216) claims that acting on the
mind is both necessary and sufficient for a thing to be the mind’s immediate object.
 LATER MALEBRANCHE 

But the problem goes deeper. Whether we have intelligible exten-

sion in place or instead are working with an infinite variety of distinct
ideas, imbuing them with causal powers is of no help with our
problem. Merely being caused by an idea in God’s mind is insufficient
to imbue sensation with directedness. In pure perception, the mind’s
act of thinking of a square is directed at an idea. But in sensation, this
is not the case; the soul sees only itself. What causal relations obtain is
beside the point, or so I shall now argue.
To make his case for the intentional powers of sensations, Nolan
appeals to passages like the following:

When the idea of extension affects or modifies the soul with a pure percep-
tion, then the soul conceives simply this extension. But when the idea of
extension touches the soul more vividly, and affects it with a sensible
perception, then the soul sees or feels extension. The soul feels it or perceives
it still more vividly when the perception with which intelligible extension
modifies it is a pain. For color, pain, and all the other sensations are only
sensible perceptions, produced in intelligences by intelligible ideas.20

While this might establish that intelligible extension causes sensa-

tions, it does nothing to show that sensations are directed at intelli-
gible extension. Malebranche’s placing color, pain, smell, and sound
all on one level makes it hard to see how sensations could have
intelligible extension as their intentional object. What sensations
contribute, Malebranche is telling us, is vividness. They wake the
soul up. The distinction Malebranche draws above is between pure
perception and ‘sensible’ perception. But again, a sensible perception
is not a perception through sensation to an idea. It is a pure percep-
tion that is accompanied by sensation.
My point here is simple: that state S is caused by x does not
establish that S is directed toward or is about x.21 Let us leave aside
Malebranche’s metaphysics for a moment to make this general point.

20
Christian Conversations, OC 4: 75–6, trans. Nolan (2012, 42).
21
Nolan writes, ‘intelligible extension can cause us to cognize it in different ways
depending on how it affects the soul. So intelligible extension is the immediate object
of all our perceptions, sensory or otherwise’ (2012, 41). As far as I can tell, the
conclusion doesn’t follow, at least absent some further premises.
 LATER MALEBRANCHE

A state of sadness might be caused by hearing a certain song, but it

would be odd to say that that state is therefore directed to or about the
song. The mere fact that intelligible extension causes us to have
sensations does nothing to supply those sensations with representa-
tion or intentionality. To suppose otherwise is to strawman the
defenders of the blank effect reading: they would have to deny that
sensations are caused. But this is absurd; they deny only that sensa-
tions are themselves intentional or representational.
Thus, for all their differences, Schmaltz, Jolley, Simmons, and
Nolan all place too much weight on the causal connection between
idea and sensation. With Malebranche’s later view, we trade one
mystery—the soul’s projecting color on to an idea so as to see it as
F—for another—the mechanism by which intelligible extension
causes us to see it as F.
Malebranche stakes out a stark position. Sensations are not repre-
sentations, nor do they exhibit intentionality; they are nothing but
modifications of a mind. We have seen that this position cannot be
rehabilitated by lumping it in with contemporary adverbial theories.
Nor can Malebranche’s ingenious strategies—natural judgments,
intelligible extension, efficacious ideas, or any combination of
these—help him give sensations object-directedness. What is sensed
is always and only a mode of the mind and not extension, intelligible
or otherwise.

10.3 The Disappearing Of

It is hard not to conclude that the doctrine of intelligible extension,
with or without efficacious ideas, makes the challenges of positioning,
localization, and individuation even more daunting than ever. Taken
to its logical conclusion, however, this doctrine brings about a total
reorientation of the philosophy of perception. Indeed, I shall now
argue that the doctrine, in combination with the implications of other
of Malebranche’s views, makes for a radically anti-Cartesian ontology
of perception. On this view, or so I’ll suggest, some of the presup-
positions of the problems we have been considering simply fall away.
 LATER MALEBRANCHE 

In fact, the whole picture of the finite mind using a representation to

think of a single physical object vanishes. That does not by itself solve
the problems we have been considering: the phenomenology still
needs to be accounted for. But there is no longer any real pressure,
I shall argue, to make that phenomenology walk in lock-step with the
ontology.
For the most part, Malebranche seems content to allow his readers
to think of intelligible extension simply as the idea of extension that
exists in God’s mind. Like any Cartesian idea, it would have to have
two aspects: formal and objective. Its objective reality is what makes it
a representation: the idea can direct the mind, through its own nature,
to an object out there in the world.
But Malebranche’s final view is much more radical. To see it, we
must go beyond the Search and the Dialogues. For it is only in the
vitriolic exchanges with Arnauld and Régis that the true extent of
Malebranche’s philosophical heresy is revealed. In the end, Male-
branche rejects the distinction between formal and objective being
and with it, the very notion of a Cartesian idea. In other words, there
is no room in intelligible extension for a distinction between the idea
and what it is an idea of. And that means that there is no sense in
which intelligible extension is functioning as a representation. As
we’ll see, when pressed on the nature of intelligible extension, Male-
branche replaces representation with participation. Intelligible exten-
sion does none of the work required of Cartesian ideas.
Looked at from another point of view, Malebranche is not so much
rejecting ideas as restoring them to the place they had before Des-
cartes came on the scene. As is well known, Descartes chooses the
term ‘idea’ because it ‘was the standard philosophical term used to
refer to the forms of perception belonging to the divine mind.’22
Intelligible extension for Malebranche is not a Cartesian idea but a
divine archetype. In that sense, he is not rejecting the notion of an
idea but stealing it back.

22
AT VII 181/CSM II 127. For a useful discussion of this point, see esp. Jolley
1990, 12).
 LATER MALEBRANCHE

It is rare to find Malebranche explicitly addressing the formal/

objective distinction. These terms do not appear anywhere in all of
the Search or the Elucidations, and in the Dialogues, objective reality is
mentioned exactly once.23 In fact, it is only when pressed by other
philosophers that Malebranche explicitly attacks his opponents’ dis-
tinction between what is true of an idea in essendo (of the idea tout
court) and in repraesentando (of what it represents). To understand
Malebranche’s full assault on that distinction, we need to look at some
of his arguments against taking ideas to be modes.24 For it is as a
response to this line of argument that the distinction enters the debate.
In the Elucidations, Malebranche argues that a finite mind cannot
represent extension without itself taking on extension as a mode.25
Malebranche claims that ‘as the soul is a particular being, a limited
being, it cannot have extension in it without becoming material,
without being composed of two substances.’ This is a puzzling argu-
ment: why think that a soul cannot ‘contain’ infinity by way of an
idea without itself being infinite? The idea of the infinite need not be
itself infinite.
On one reading, Malebranche is claiming that representation
requires literal resemblance. So if a mode of a mind represented
extension, it would have to be itself extended.26 As we’ve seen,
however, this argument would be a non-starter: most Cartesians

23
D VIII/JS 128. Aristes speaks of the object of his thought of infinity as ‘the
objective reality in which his mind gets lost.’ Although Malebranche’s spokesman
Theodore agrees with what Aristes says, he does not use that phrase when making the
same point himself.
24
There is of course a sense in which Malebranche’s indirect argument for the
vision in God in SAT III.ii.1–6 counts as an argument against taking ideas to be modes
(at least of finite minds). But it is worth noting that, although some of the views
Malebranche criticizes are committed to that claim, Malebranche’s focus is elsewhere.
Thus, for example, in criticizing an empiricist account of perception, Malebranche
focuses on the notion that objects emit material species that penetrate the eyeball and
ultimately lodge in the brain. For a good discussion and evaluation of the indirect
argument for the vision in God, see Andrew Pyle (2003, 50 f.).
25
E X/LO 624.
26
Richard Watson made this suggestion to Nadler; see Nadler (1989, 37, n.34). See
also Jonathan Hill (2011).
 LATER MALEBRANCHE 

would distinguish between entitative resemblance, which requires

property sharing, and intentional resemblance, which does not.
Daisie Radner and Steven Nadler suggest a very different reading of
the argument.27 According to this argument, an infinite or general
idea must ‘contain or represent an infinite number of things (or many
things) at once. But something that is particular and finite cannot do
this job, since it can only take on a certain determinate, limited, and
particular form or content.’28 Although this might well be what
Malebranche has in mind, it still strikes me as assuming what is to
be proved. Why should the finitude of the mind and its modes
prevent the idea-qua-mode from being an idea of the infinite?
The same circularity infects a very similar argument from the
Réponse to Régis. Malebranche writes that
[t]he idea of a circle in general represents only what it contains [‘renferme’].
Now, this idea includes nothing general, since it is only a particular mode of
the soul, according to Mr. Régis. Thus the idea of a circle in general
represents nothing general. The contradiction is obvious . . . 29

This ‘container’ argument begs the question by assuming that finite

modes do not admit of the objective/formal distinction. If an idea can
contain (renferme, contient) something as its content, that is, object-
ively, then there is no barrier to a finite mode containing the infinite
in that sense. It is only if ‘contain’ means ‘possess as a property’ that
the argument goes through. In other words, it is only if the content of
an idea qua mode has been squashed into its features that the idea of
the infinite must itself be infinite.
These arguments from the Elucidations and the exchange with
Régis suffer from the same flaw: they assume that the formal/objective
distinction does not hold. Malebranche was made aware of this
objection. In a letter to Arnauld, Malebranche reiterates his claim that

27
See Radner (1978, 51 and 116) and Nadler (1992, 43).
28
Nadler 1992, 43.
29
OC 17–1: 302–3. For a similar argument, see D II/JS 28. Alquié (1974, 156)
discusses this line of argument but does not note its flaws.
 LATER MALEBRANCHE

the modes of the soul are finite, and that they thus cannot represent the
infinite, since nothing can represent or make seen what it lacks; that nothing
cannot be seen; that the soul cannot perceive what is not represented to it.
The most ordinary response is this: I distinguish: the modes of the soul are
finite in themselves, in essendo, I admit, but in repraesentando, I deny. I have
noted that I am not content with so cheap a defeat. For it is obvious that it is
to suppose exactly what is in question and not to respond to the proof we
have just seen. It is to admit in barbarous Latin that I speak the truth, but that
one does not want to yield to it. Nevertheless Mr. Arnauld found this
response fitting: and I am hardly astonished, since it appears to him that
the modes of the soul are representations, and that the whole is greater than
its parts. ‘You are not content,’ he says to me, ‘with this distinction; too bad
for you.’30

So far, Malebranche’s reply is not very convincing. Accusing his

opponent of begging the question would seem slightly more persua-
sive if the distinction between what’s true of a representation and
what it represents were not so entrenched. Malebranche goes on to
invoke the principle that ‘nothingness is not visible,’ and hence that it
is impossible to see the infinite in the finite. Thus no finite mode can
represent the infinite. Unfortunately, that move seems to repeat the
error Arnauld accuses him of, since it is effective only if one already
denies the in essendo/in repraesentando distinction.31
I conclude, then, that the debate between Malebranche and the
other Cartesians on the formal/objective reality distinction ends in a
stalemate. For our purposes, the key point is simply that Malebranche

30
Third Letter to Arnauld (1699), OC IX: 952–3.
31
In his Third Letter to Arnauld (1699), Malebranche writes, ‘[t]hat which is finite
in itself in essendo cannot be infinite in repraesentando. Thus, even if we had a clear
knowledge of our soul and its modes (which we do not, since we know our soul only
by an interior sentiment); since these modes are finite, we could never discover among
them the infinite, because nothing is not visible, and because one cannot perceive in
the soul what is not there. Similarly, if I perceive in the circle an infinity of equal
diameters, that is, that these diameters are equal in repraesentando, I must conclude
that they are really equal in essendo. For in fact a circle includes the reality of an
infinity of diameters. Thus in order that a reality be present to the mind, and affect the
mind, and for the mind to perceive or receive it, it necessarily must really exist, since
nothing cannot be perceived, and to see nothing is not to see (OC 9: 954). See
Malebranche’s similar remarks to Régis (OC XVII-1 302–3).
 LATER MALEBRANCHE 

does reject it. His rejection is not limited to modes of finite beings.
Even if God were to have modes, they could not exhibit the curious
duality the Cartesians require of their ideas.
In rejecting the distinction between what is true of an idea in
essendo and in repraesentando, Malebranche has gutted the very
notion of a Cartesian idea.32 For it is essential to that notion, whether
it is developed in a direct or indirect realist direction, that it retain
some distinction between the idea as a mode or act of a mind and the
representational content of that idea. For Malebranche, by contrast,
to represent something to the soul is simply for that thing to be
present to the soul.33
It may still seem, however, that Malebranche is a straightforward
indirect realist. On that view, a typical act of perception includes three
terms and two relations. The first relation holds between the mind
and the idea, and the second between the idea and the object in the
world (if there is one). Now, Malebranche is not denying or rejecting
the mind-idea relation. Minds exhibit intentionality in the sense
that they perceive ideas in God’s mind. If Malebranche were simply
relocating Cartesian ideas, one would expect representation to hold, not
between mind and idea, but between idea and object. Representation
is typically understood among the Cartesians, whatever their other
differences, as a function of the objective reality of ideas. No such
objective reality can appear at the mind-idea juncture; it is the idea, if

32
Nadler (1992, 41) makes a perceptive remark in this context. Malebranche seems
to be claiming that ‘if we are to see either a thing or a property or a feature of reality by
representation, what is apprehended by the mind (i.e., the representans) must either
actually be that thing or formally possess that property or feature. But then what we are
talking about is not representation at all, but some kind of direct presentation of the
thing or property itself.’ Nadler of course goes on to claim that this is not Malebranche’s
position. But Nadler is right in thinking that if all that is left of representation is
presentation, then we are no longer dealing with representation at all.
33
For example, Malebranche speaks of something ‘representing to the soul the
infinite, that is, making it formally perceive it’ (OC IX 947); he says that ‘the modes of
the soul are finite, and that they cannot represent the infinite, since one cannot
represent or make seen (ou faire voire) what one lacks’ (OC IX 952–3).
 LATER MALEBRANCHE

anything, that exhibits objective reality, and is directed to its ultimate

object in the world by that reality.34
To see how Malebranche proposes to supplant their theory, we need
to turn to the nature of the idea in God’s mind, that is, to intelligible
extension. What exactly is the link between idea and world, if not
representation in virtue of objective being?
To answer this, we need to know more about the ontology of
intelligible extension. It cannot be a Cartesian idea, since those ideas
are modes, and Malebranche’s God has no modes. What, exactly, is it,
then? When pressed on its nature, Malebranche typically genuflects to
Aquinas and declares that it is God’s ‘essence, insofar as it is particip-
able or imperfectly imitable.’35 Jasper Reid argues convincingly that
Malebranche’s Trinitarian theology is tied to the way in which intelli-
gible extension can be God’s essence.36 As Reid puts it, intelligible
extension is the Word, which ‘is eternally begotten as a second person
within [the divine substance] as God reflected on His various perfec-
tions and saw how they could be imperfectly imitated by creatures.’37
Intelligible extension just is God, considered as that aspect of God the
created world of extension is fit to imitate. The gulf between intelligible
extension and a Cartesian idea could hardly be wider.

34
In a letter of 1684, Malebranche writes, ‘I beg you to consider whether the
distinctions between formal being, objective being, in essendo, in repraesentando,
awaken clear ideas. Since for my part, I find that, with regard to creatures and their
ideas, these terms can have a good sense, but in regard to God or the infinite, they can
only dispose one to error, since they are the inventions of the human mind which
ignores the intimate presence and continual operation of this universal Reason which
illuminates minds by a manifestation of its substance, as Augustine says, and in which
are the fundamental ideas of all created and possible beings, and generally all truths’
(OC XVIII 281). Malebranche, of course, denies that creatures have ideas. And so the
‘good sense’ that these terms can be given vanishes.
35
E X/LO 625; for the same point in the exchange with Arnauld, see the Third
Letter, OC IX 910, 952, and 955.
36
See Reid (2003).
37
Reid (2003, 596). The Word has intelligible parts; it cannot, of course, have
literal parts, for that way lies Spinozism, something of which Malebranche was often
accused. See, for example his exchange with Dortous de Marain (trans. Marjorie
Grene, included in Foucher 1995). For a helpful discussion of the correspondence
and the accusation of Spinozism, see esp. Fred Ablondi (1998).
 LATER MALEBRANCHE 

This ontological difference brings in its train a series of further

differences. First, on the Cartesian view, it is the objective reality of
the idea that directs the mind towards its ultimate object. The object-
ive being of the idea of a square directs the mind toward the square
object in the material world. Now, there are many different ways of
understanding what is going on there. But on any of them, the
representational content of the idea is responsible for directing the
mind to its ultimate object.
Nothing like this is true of intelligible extension. It is not intelligible
extension itself but the sensations the mind projects on to it that allow
the mind to think of a square object. Even when Malebranche is most
anxious to elide the differences between the view embodied in the first
edition of the Search and this seemingly novel doctrine of intelligible
extension, this fact is apparent: ‘I do say that we see all things in God
through the efficacy of His substance, and particularly sensible things,
through God’s applying intelligible extension to our mind in a thou-
sand different ways, and that thus intelligible extension contains all
the perfections, or rather, all the differences of bodies due to the
different sensations that the soul projects [‘répand’] on the ideas
affecting it upon the occasion of these same bodies.’38
What is the relation between intelligible extension and extended
bodies? We have already seen that it is not one of picture-original,
since nothing about intelligible extension limits it to representing any
determinate world of extension.39 If it is indifferently ‘of ’ any physical
object at all, whether possible or actual, there is no sense to be made of
the claim that it is picture-like. Even intentional resemblance cannot
link intelligible extension and extension.
What Malebranche does say strongly suggests that he is replacing
representation with participation.40 He tells Arnauld, for example,

38
E X/LO 628.
39
In the Méditations Chrétiennes et Métaphysiques of 1707, Malebranche writes
that intelligible extension ‘is the idea of an infinity of possible worlds’ (OC X 99).
40
For a slightly different view, see Susan Peppers-Bates’s (2005, 97 f.) Peppers-
Bates concludes that the representational relation between intelligible extension and
extension itself is left unexplained and sui generis.
 LATER MALEBRANCHE

that no modes, whether of God or of finite minds, represent matter; it

is ‘only the idea of matter which is of itself representative, that is to
say, that it is only the divine nature, insofar as it eminently and
divinely contains all that there is of reality and perfection in matter,
which is representative of matter.’41 This passage invites a reductive
reading: to say that ‘God’s idea of matter represents matter’ is to say
that ‘the divine nature contains, in a higher and more noble way, all
the perfection and reality of matter.’
These departures from Cartesian orthodoxy come together in a
passage from 1704:

In expounding my doctrine, I never said: that one sees bodies by represen-

tative beings [êtres representatifs] that are in God; but always that one sees
them by what is in God that represents them, or in the substance of God
insofar as it is participable by bodies or representative of bodies, or by their
ideas or archetypes that can only be found in God, and which alone illumin-
ate, affect, or modify minds: the Author nevertheless . . . attacks me as if
I believed that there is in God formally such a being, representative of each
created thing, which representative being, in affecting my soul, makes me
see the created being: an opinion that he well knows I have refuted in my
first Book.42

The first sentence is startling: how can Malebranche pretend he has

never said that one sees bodies by seeing representations of them in
God? Such claims are to be found throughout his writings, including
the tenth Elucidation, where the doctrine of intelligible extension is
introduced.

41
OC IX 952.
42
OC IX 1068. Malebranche continues: ‘You have only to open your eyes to his
pretended ‘demonstrations’ to see that he gives my views on ideas this false or at least
equivocal exposition. Here, then, is my position, which one can easily find in my
Books if one takes the time to examine them. The general idea of created extension,
which includes the ideas of all particular bodies, that is, from which the particular
ideas of all bodies can be drawn, just as one can form or show all particular bodies
from created extension: the idea, I say, of local extension, or intelligible extension is
nothing but the substance of God, not taken or conceived absolutely, but perceived
insofar as it is relative to created extension, or insofar as it is imperfectly imitable or
participable, to speak as saint Thomas does’ (OC IX 1068–9).
 LATER MALEBRANCHE 

Nevertheless, Malebranche is usually careful to mention in these

texts that what we see in God is the substance of God himself. His
readers might be forgiven, however, for taking literally his talk of
intelligible extension as an ‘idea’ and taking his perplexing talk of the
participable substance of God as mere metaphor or pious ornament.
Just the opposite is the case.
Someone might argue that the relation between intelligible exten-
sion and the world must be one of representation. Malebranche, after
all, continually speaks as if we see created extension by seeing intelli-
gible extension, and as if created extension were the indirect object of
experience and as such, trivially, among the objects of experience.
How then can he be said to have severed the connection?
Of course there is some relation between the two kinds of exten-
sion. In some sense, they must share an essence; creation participates
in the nature of the divine substance. In my view, this is the only
relation to be found between the two. When Malebranche talks of
God’s creating something that ‘answers’ or ‘corresponds to’ (réponde)
intelligible extension, he can only mean that God creates something
that participates in intelligible extension.43 One can learn all of the
possible properties of extension by contemplating God. But that does
not mean that one can thereby learn its actual and hence fully
determinate properties.
Although this might seem a disappointing conclusion, it is the
inevitable consequence of Malebranche’s account of sense perception.
Remember that sensations do not characterize intelligible extension.
They are modes of a mind that either projects these sensations on to it

43
Malebranche claims (in the response to Régis) that ‘[w]ith regard to my ideas,
I believe that they only represent directly, that I see directly and immediately only
what they include; for to see nothing is not to see. But if God has created something
that answers [réponde] to my idea as to its archetype, I can say that my idea represents
this being, and that in seeing one directly I see the other indirectly. To know the
properties of this being, I consult the idea, and not my modalities, since it is the idea
and not my modality that is the archetype according to which God has formed it [this
being]. But I conclude nothing about the actual existence of this being, because God
does not necessarily create what these ideas represent, that is, beings who answer to
these ideas; their creation is arbitrary’ (OC XVII–1 303).
 LATER MALEBRANCHE

or suffers those sensations when intelligible extension acts on it. In

neither case is there anything intrinsic to intelligible extension that
provides it with any determinate shape. So the most that intelligible
and created extension could ever have shared is an essence.
Some will still find this response unsatisfying. Doesn’t God, at least,
need to have a fully determinate idea of the world he is about to
create? Not acting blindly in creating the world, God must use an
aspect of himself as its model, and a model that is equally a model of
every possible world is no model at all.
In fact, Malebranche continually inflates the sense in which intel-
ligible extension is the archetype of created extension. To see this,
consider Malebranche’s reply to Arnauld’s charge that his view
invites Pyrrhonian skepticism. Malebranche cleverly turns the accus-
ation on its head: it is Arnauld who opens the door to Pyrrhonism by
making ideas modifications of finite minds. By contrast, the fact that
God has used intelligible extension as ‘the model of extended creation
according to which all bodies are formed’ guarantees that this idea
conforms to its object.44
That guarantee holds only where the essence of extension is con-
cerned. One can certainly learn the properties and possible arrange-
ments of extension through God. But nothing in God pins intelligible
extension to this or that determinate arrangement. In other words,
nothing about intelligible extension tells the subject which of the
infinite possible arrangements of matter is actual. Given that, Male-
branche’s doctrine is cold comfort in the face of Pyrrhonism. Never-
theless, if Malebranche is right, we at least have access to the nature of
created beings. And God at least is guided by a model of the kind of
thing he is about to create.
Malebranche can make a stronger reply to the charge of Pyrrhon-
ism. The objection has to assume that there are mind-independent
differences among regions of created extension that need to be rep-
resented in the model God has of it prior to the creation. But that
assumption is not obviously correct.

44
OC IX 925–6.
 LATER MALEBRANCHE 

There is no void to be detected between objects. The only proper-

ties bodies admit of are relations of distance.45 What makes a square a
square is simply the relations that hold among the points that
constitute its surface. But in virtue of what are those points them-
selves selected as the boundaries of an object? Without a void, that
selection seems arbitrary. We know that visual experience relies on
colors to do this job; but it is far from obvious that there is any mind-
independent fact of the matter to which such experience can be held
accountable.
My suggestion is that Malebranche conceives of both intelligible
extension and extension as equally uniform and undifferentiated.
Both motion and the individuation of bodies are entirely phenom-
enal, a question solely of the sensations caused in minds by intelligible
extension. If extension were individuated into distinct bodies, and
intelligible extension were not, then seeing intelligible extension in
himself would provide God with no substantive model of the universe
he is about to create. Since the conditional seems correct, we have
reason to doubt its antecedent. On the model I am suggesting, the
only differences among the possible worlds God might have created
are subject-dependent. They amount, in the end, only to the different
sensations the minds living ‘in’ them would experience.46
Once Malebranche denies any distinctions among the ‘parts’ of
intelligible extension, he seems well on the road to this final position.

45
See SAT I.x.1/LO 49.
46
Thomas Lennon (1993) and (2007) argues, naturally on different grounds, that
the multiplicity of bodies is illusory for Descartes, since Descartes uses motion to
individuate bodies. But, as Lennon puts it, ‘matter, or body, and the space it occupies
are really identical; but body really moves only if it changes that space, which cannot
happen if it is identical to that space’ (2007, 29). I am unsure whether Lennon’s
conclusion is something Descartes really means to endorse or whether it is an unseen
consequence of his views on motion and individuation. Emily Thomas argues against
Lennon but at the cost of claiming that ‘the parts of space can move’ (2015, 761).
Another important contribution to this debate is Alice Sowaal’s (2004). I take it as
uncontroversial (based on the texts we are discussing) that Malebranche is much more
explicit about these matters. For an exhaustive bibliography and suggestive treatment
of the question in Descartes, see Reid (2014) (although Reid himself sets to one side
the whole question of metaphysical distinction among bodies; see his 2014, 39).
 LATER MALEBRANCHE

Consider Jasper Reid’s ingenious attempt to make sense of intelligible

extension as a representation. As Reid sees things, Malebranche takes
representation to be isomorphism. Reid asks us to consider a created
right triangle. The lengths of its side are first-order relations of
distance between each of the vertices; the Pythagorean relation will
be a second order relation that holds among the first order relations.
‘Although the three first-order relations between the intelligible
vertices will be different,’ Reid writes, ‘for they will not be distances,
they will still stand in just the same numerical proportion amongst
themselves.’47 The problem is that the intelligible vertices are not
there in intelligible extension. Intelligible extension has such parts
only in the sense that we, at times, perceive it that way.48 Suppose
I claim that a painting represents a forest scene by virtue of having a
particular array of colors on its surface. If it turns out that the
painting is really a blank canvas but prompts in us the thought of a
forest because of some optical illusion, it would be perverse to insist
that it is nevertheless of a forest scene.49
To sum up: when Malebranche argues that ideas are not modes of
created minds, much more is at stake than whose ideas they are. They
are not modes of finite minds like ours because they are not modes at
all. Nor, trivially, are they modes of God, since God has no modes to
begin with. In place of the idea of extension, Malebranche substitutes
God’s essence, taken under the concept of one of his persons. One
perceives intelligible extension by undergoing sensations when intel-
ligible extension acts on the mind. Whether there is anything else out
there that ‘answers’ to what one experiences is not discernible by
reason. What is far more important is that such ‘answering’ or
‘corresponding’ amounts only to the fact that there is something
outside of God that partakes of his nature.

47
Reid (2003, 602).
48
As Reid is well aware; see his (2003, 594).
49
Other Cartesians might tell a story about representation that would make this
result natural rather than perverse (consider, for example, the purely causal notion of
intentional resemblance explored in chapter six above). But there is no sign that
Malebranche would agree with them.
 LATER MALEBRANCHE 

It is easy to underestimate Malebranche’s dialectical position. To

reject the of-ness of even divine ‘ideas’ is to take a fairly natural path
from the nature of Cartesian modes as manieres d’existence. Nor is
it simply obvious that the alternative path followed by the other
Cartesians really makes sense. How well-placed are Malebranche’s
critics to complain about his view? Not one of them can say anything
substantive about representation and the curious objective being of
mental modes. All one gets from la Forge, Desgabets, and Arnauld is
the claim that representation is in some sense similar to, but at the
same time entirely different from, resemblance. Only familiarity
makes the Cartesian view seem substantive and plausible. It is hard
not to agree with Simon Foucher: ‘What does it mean to say that our
ideas are intentionally or representationally similar, if not that they
are similar insofar as they must be to represent? And that is exactly
what is in question. One must know in what consists the similitude of
ideas in respect of which they represent, and when one says that this
similitude is representational or intentional, one only repeats the
question . . . in more barbarous terms.’50
Nothing in this section solves the core problems Malebranche’s
philosophy of perception faces. The puzzling connection between
sensations and intelligible extension remains puzzling. But it does
cast those problems in a new light. The explananda is no longer our
ability to experientially ‘cotton on’ to mind-independent distinctions,
distances, and locations in the world of extension. Instead, what needs
explaining is how the physical world ever came to seem to have those
features in the first place.
Let me close this section by looking forward beyond the proper
scope of this book to Malebranche’s closest descendant in England,
George Berkeley. For I suspect that Malebranche’s flattening of ideas—
the disappearing of—finds its full expression in Berkeley’s immaterial-
ism.51 At least some of Berkeley’s arguments for immaterialism silently

50
ND, 34–5.
51
I am very much indebted to Martha Bolton’s (1987) and (2008). She convin-
cingly argues that Berkeley erases the idea/content distinction (in the way I’ll specify
 LATER MALEBRANCHE

assume that the ideas we experience do not admit of the Janus-faced

nature of Cartesian ideas. To take just one example: in his so-called
‘master argument,’ Berkeley challenges the reader to conceive books
and trees existing unconceived. ‘Nothing seems easier. But what is all
this, I beseech you, more than framing in your mind certain ideas which
you call books and trees, and at the same time omitting to frame the idea
of anyone that may perceive them? But do you not yourself perceive or
think of them all the while?’52
The argument seems almost comically wide of the mark. True, the
ideas of the books and trees must be conceived or perceived; but why
think that that feature is automatically preserved in the objects these
ideas represent? An idea of a tree is perceived when thought about;
yet nothing stops it from being of a tree that is unperceived.53
Enormous effort has been expended to rehabilitate the master
argument.54 And yet it is hard to escape the suspicion that Berkeley
elides the difference between an idea and what the idea is an idea of,
just as Malebranche does. Once the ‘of ’ of content disappears, there is
no barrier to prevent the properties of the idea qua datable mental
item from becoming those of its content.
Whatever one makes of the master argument, it is clear that
Berkeley’s own positive view of ideas has no room for intrinsic of-
ness. If ‘houses, mountains, rivers,’ and everything else are ideas, then
ideas lack content, for there is no sense in which a river is a river of

in the text) and that Berkeley assumes this erasure in most of his arguments in the
Principles of Human Knowledge.
52
Principles I.23 in Berkeley (1948–1957, vol. 2). In what follows, I am hewing
very closely to Bolton’s own (1987) treatment of the master argument. Bolton’s
reading has of course been challenged in the literature; see especially Robert
Muehlmann (1992, 53–5).
53
George Pitcher (1977, 153) raises this objection forcefully in his Robinson
Crusoe analogy. Berkeley’s view would have the absurd consequence that there cannot
be a public performance of a play about Robinson Crusoe: the presence of the
audience would prevent its protagonist from being alone on the island.
54
For an ingenious treatment of the master argument, see André Gallois (1974).
Kenneth Winkler’s (1989, 183–7) remains one of the best discussions of the argument.
As Winkler reads it, it is a highly compressed summary of what has gone before, not a
stand-alone argument.
 LATER MALEBRANCHE 

anything.55 I do not mean merely that such ideas have no extra-

mental material object, which is trivially true for an immaterialist.
The point instead is that these ideas are not of anything in the sense
that they lack representational content.56
The list of Berkeley’s departures from Malebranche is enormous.
But two of Berkeley’s early anonymous critics, who call him a ‘Mal-
branchiste de bonne foi,’ and name Malebranche ‘his master,’ are
perceptive.57 In erasing the distinction between formal and objective
being, and abandoning extension, both intelligible and otherwise,
Berkeley can make a good claim to be following Malebranche’s
insights—or, if you prefer, mistakes—to their logical conclusion. If
so, the French debate over perception as we have traced it presents us
with a miniature of the development of modern philosophy as a
whole, from Cartesianism to Berkeleyan idealism.

55
PHK I 4. As Bolton puts the same point: Berkeley’s classifying ordinary objects
as ideas ‘clearly implies that at least some ideas lack intentional objects: we see and feel
a fire, but a fire is not “of ” something’ (2008, 80). Bolton goes on to note that ideas, for
Berkeley, can be ‘of ’ other ideas by virtue of resembling them; see especially Notebook
entries 823 and 843. It is only the ideas we encounter in sensory experience that lack
of-ness. Moreover, as an anonymous referee pointed out, it is important to emphasize
that such ideas lack intrinsic content; by standing in relations with other ideas, they
become part of the divine language (see Bolton 1987, 68–9).
56
In at least one place in the Principles (I.49), Berkeley clearly seems to have
Malebranche’s Elucidations in mind. Berkeley considers the objection that, if exten-
sion exists in the mind, the mind will be extended. Like Malebranche, he argues that
extension is not ‘in’ the mind as a mode, nor is the idea of extension a mode of a mind.
Berkeley goes further than Malebranche, of course, and rejects the whole substance/
mode ontology, replacing that relation with a primitive relation of perception.
57
See Charles McCracken (1983, 205–6) for the early reception of Berkeley’s work.
McCracken provides a detailed account of just where the two thinkers diverge.
 11
Conclusion

Our philosophers see themselves as overturning a naïve view of the

world. For all their sophisticated doctrine and jargon, their Aristotel-
ian predecessors never outgrew the prejudices of childhood. First
among these is the assumption that the world is largely as it appears
to us. Again and again, throughout the seventeenth century and well
into the next, the moderns congratulate themselves on ousting sens-
ible qualities from the world beyond the mind.
That world suddenly becomes more tractable: we can in principle
account for any causal transactions we see in terms of geometrical
qualities alone, and hence such qualities are the only ones we need
postulate.1 But now it’s the world inside the mind that is mysterious.
If bodies have no sensible qualities, how is it that we come to see them
otherwise? More important, how (if at all) do we use these sensible
qualities to perceive our environment? We have focused on three
ways in which sensations at least seem to matter to sensory percep-
tion: the individuation of bodies, their positioning in space relative to
each other, and the localization of sensible qualities.
The Aristotelians can provide straightforward answers to these
questions. We see the world as having color because it does. Our
experience of color allows us to perceive shape, distance, position, and
the rest, because colors are equally properties of bodies. The brown

1
I am ignoring natural phenomena that remained intractable in this period, such
as magnetism. And the period of optimism about the explanatory powers of geomet-
rical qualities is relatively short-lived, ending, arguably, with Newton. For more, see
my (2009).
 CONCLUSION 

bark of the tree tells us where that tree is because its brownness is one
of its intrinsic features. In sensing a color we are also sensing shape,
position, and distance.
In the seventeenth century, these easy answers are closed off by the
ejection of sensible qualities. Yoked to the now moribund Aristotelian
position is the Baconian synthesis, a theory about how sensation
generally, and vision in particular, works. It is not too much to say
that the whole edifice that had grown up around the broadly Aristo-
telian tradition was crumbling. Our authors try their best to replace it.
Of course no theory of anything can start completely from scratch.
Descartes, Malebranche, and the others often help themselves to the
notions and strategies of their predecessors.
Descartes begins with a striking and radical position. Ideas and
hence representations are not confined to the realm of the mind. The
image in the brain—arrived at by a process inspired by the Baconian
synthesis, married to Kepler—functions as a representation. Once
installed in its machine, the mind ‘turns toward’ that brain image in
order to perceive. We saw that this doctrine of conversio is not an
innovation but an adaptation of the scholastics’ notion of the intellect
turning toward corporeal images. For the early Descartes, then,
corporeal ideas represent their objects by resembling them. These
corporeal ideas allow the unensouled machines we call animals (and
indeed inattentive humans) to act on their ‘desires’ and to ‘recognize’
objects in their environments.
In the case of brain images, the resemblance in question must be
literal. It is this ordinary sense of resemblance that makes the brain
image an idea of its object in the world. While this is a relatively
short-lived feature of Descartes’s position, his general attachment to
representation as resemblance is not. Indeed, as Foucher in effect
shows, Descartes is committed throughout his career to thinking of
representation as resemblance. To many readers of Descartes, this
reading will sound disastrous. But once we see Descartes’s intellec-
tual context, it becomes clear that, like the conversio doctrine, this
claim is thoroughly traditional. While some of his antecedents and
followers are committed to a sui generis notion of representational
 CONCLUSION

resemblance, it is not entirely clear that that is Descartes’s notion.

Still, it is perfectly open to Descartes to treat corporeal representation
as literal resemblance and the representational nature of ideas as
merely intentional resemblance.
Descartes’s views are anything but static. Although he never gives
up on his claim that the mind is aware of the brain, he comes to see
the brain image as a mere side effect of brain motions. Its defects
make it unsuitable to play the role Descartes once assigned it. As I’ve
argued, there seems to be an inverse relationship between the sophis-
tication and accuracy of the brain and eye on one hand and the need
for mental activity on the other. We saw Descartes appealing to a
variety of mental processes, including natural geometry, to explain
sensory perception. Such a move has troubling consequences for
Descartes’s bête machine doctrine, since such sophisticated inferences
are not available to animals and inattentive humans, who seem to get
by just fine.
His final position makes the summoning of ideas of geometrical
qualities and sensations a purely causal affair, the result of the order
of nature. There is still work for the mind to do: it has to attribute or
refer these sensations and ideas to bodies. As Malebranche will in
effect point out, this process is mysterious. Malebranche believes that
the only way for minds to individuate the objects of vision is by
means of color. Hence one cannot begin with an idea of a body
characterized solely by geometrical properties and then project a
sensation on to it.
Throughout, we saw little evidence that Descartes’s sensations are
representations, and much evidence that they are not. Descartes
makes the point explicit in the Principles. But it is there in the
Meditations’ doctrine of material falsity, especially as clarified in the
Replies. The developmental reading of the Meditations allowed us to
cut the Gordian knot the doctrine of material falsity has become.
Only in the sixth Replies does Descartes suggest that the mind needs
to use sensible qualities to individuate bodies.
Each of Descartes’s immediate successors chooses a different ver-
sion of his ancestor’s view on which to build. For Louis de la Forge,
 CONCLUSION 

the purely automatic view of Descartes’s last works is the most

defensible. For Robert Desgabets, it is the earliest work, with its
doctrine of corporeal ideas and conversio, that provides the clue to
the truth. Only Pierre-Sylvain Régis selects a view that, like the sixth
Replies, relies on sensible qualities to generate an awareness of the
geometrical qualities of bodies. Finally, Antoine Arnauld endorses
the view of the Principles and hence faces Malebranche’s selection
argument, just like his confederates.
Much of the development of Malebranche’s own view can be seen
as an attempt to meet the challenge of the selection argument.
Throughout his career, Malebranche believes that sensations have
to be the raw material on which the mind (whether ours or God’s)
judges geometrical properties, including distance and position. Only
by seeing color can we be said to ‘see’ the position and dimension of
physical objects.
Pulling back layers of revision, we were able to find Malebranche’s
original doctrine of natural judgments. These judgments are per-
formed by the finite mind and are responsible for generating full-
fledged experience out of the co-occurrence of an idea and a sensa-
tion. His view ultimately falls prey to the same kind of objection he
lodges against Descartes. With the second edition of the Search,
Malebranche rejects these natural judgments and demotes them to
the level of mere compound sensations.
At times we found Malebranche recapitulating some stages of
Descartes’s own career. He moves from attributing sophisticated
mental processes to finite minds, as Descartes did in the Dioptrique,
to withholding them. Yet their final positions could hardly be more
different. I argued that, when pressed on the nature of intelligible
extension, Malebranche abandons the whole picture of minds per-
ceiving an être representatif in God’s mind. And with it goes the
whole apparatus of formal and objective being. Finite minds still
experience ideal regions of intelligible extension painted with their
own sensations. But there is now no sense in which those regions
represent finite bodies. That turns out to be an acceptable result, I’ve
argued, because there really are no finite bodies to speak of in any
 CONCLUSION

case. Intelligible extension as an aspect of God’s essence acts on finite

minds to produce their experiences. Whether God happens to have
created extension itself is a further matter, one that can only be
known by revelation. But in any case, extension participates in intel-
ligible extension. Malebranche has replaced representation with
participation.
In its own way, the history of the philosophy of perception in
seventeenth-century France provides a model of that history in the
entire modern period. It is not merely that the issues that come to
dominate the French thinkers—the localization of sensible qualities,
the individuation and positioning of objects in experience—also
feature prominently in the modern period, especially in the work of
George Berkeley and Immanuel Kant.2 More than this, there is a slide
from a stark, anti-Aristotelian ontology of the external world into
some version of idealism. It is hard to believe this is an accident.
Perhaps it turns out to be so hard to tie sensible and geometrical
qualities together that philosophers like Berkeley and Hume are
driven to identify them again. Finding the way back to an innocent
Aristotelian view well blocked, they are faced with the fact that
sensible qualities cannot exist outside the mind. And if geometrical
and sensible qualities cannot be separated, then the world of extended
things can exist only in the mind.
Making that case would demand a monograph on its own. So let me
close with another, equally speculative, glance to the future that
followed the modern period. It is hard to deny that the Cartesians
respond to the crisis of perception in ways that to us are bound to look
somewhat absurd if not positively desperate. I have done my best to
place these views in their historical context, where their motivations
can become intelligible. Nevertheless, hardly any of these responses, as
they stand, is likely to be appealing today. But before we look back on
the seventeenth-century Cartesians with smug self-satisfaction, we
should first be certain that we have ourselves arrived at a thoroughly

2
On Kant, see Falkenstein (1995); Falkenstein (2000) takes up the issue in Thomas
Reid.
 CONCLUSION 

defensible response to the crisis of perception. For those of us who still

accept the distinction between primary and secondary qualities, the
problems of localization, individuation, and positioning remain,
though often obscured by the miasma of confusion surrounding
phenomenal character. Although the Cartesians’ solutions are some-
times fanciful, their problems, unfortunately, are not.
Appendix: The Development
of the Theory of Natural Judgment

Comparison of the 1674 edition of the first three books of the Search with the
1678 and 1712 editions (the last of which is used by LO).
1. LO I.ii 7 f. is not to be found in 1674. Beginning with the 1678 edition
(in fact the second edition of the first three books), the numbering of
chapters in Book I is affected. The 1674 chapter x, for example,
corresponds to LO chapter ix. This is more a matter of organization
than anything; the content of 1678 I.i–ii does not vary much from that
of 1674 I.i.
2. I.vi.4 in 1674: 56 is largely replaced in 1678 I.vii: 30 and LO I.vii.4: 34.
3. The footnote at LO I.ix: 41 is in neither 1674 nor 1678.
4. The last two paragraphs of LO I.ix: 46–7 are in neither 1674 nor 1678;
compare I.viii in 1674: 76.
5. I.ix in 1674: 83–5 is revised in I.x 1678: 45 and LO I.x.6: 52–3 (‘the four
things we confuse’).
6. LO I.xi: 55 adds a new final paragraph, not to be found in I.x 1674: 87
or the edition of 1678.
7. I.xiv in 1678: 57 f. and LO I.xiv: 67 f. depart from I.xiii in 1674: 106 f. in
ways too numerous to list here.
8. LO II.i.5: 101–6 does not exist in either the 1674 or 1678 texts.
References

Ablondi, Fred. 1998. ‘Le Spinoziste Malgré Lui?: Malebranche, De Mairan,

and Intelligible Extension.’ History of Philosophy Quarterly 15, 2: 191–203.
Alanen, Lili. 2003. Descartes’s Concept of Mind. Cambridge, MA: Harvard
University Press.
Alquié, Ferdinand. 1974. Le Cartésianisme de Malebranche. Paris: J. Vrin.
Aquinas, Thomas. 1945. Basic Writings, ed. Anton Pegis. 2 vols. New York:
Random House.
Aquinas, Thomas. 1946. The Commentary of St. Thomas Aquinas on Aris-
totle’s Treatise on the Soul, trans. R.A. Kocourek. New York: College of
St. Thomas Press.
Aquinas, Thomas. 1961. Commentary on Aristotle’s Metaphysics, trans. John
P. Rowan. Notre Dame: Dumb Ox Books.
Aquinas, Thomas. 1999. A Commentary on Aristotle’s De anima, trans.
Robert Pasnau. New Haven: Yale University Press.
Aristotle. 1984. The Complete Works of Aristotle, ed. J. Barnes. 2 vols.
Princeton: Princeton University Press.
Arnauld, Antoine. 1775–1783. Oeuvres de Messire Antoine Arnauld, eds.
G. Du Pac de Bellegarde and J. Hautefauge. 43 vols. Paris-Lausanne:
Sigismond D’Arnay et Cie.
Arnauld, Antoine. 1990. On True and False Ideas, trans. S. Gaukroger.
Manchester, UK: Manchester University Press.
Arnauld, Antoine and Pierre Nicole. 1996. Logic or the Art of the Thinking,
trans. Jill V. Buroker. Cambridge: Cambridge University Press.
Atherton, Margaret. 1990. Berkeley’s Revolution in Vision. Ithaca, NY:
Cornell University Press.
Ayers, Michael. 1991. Locke. 2 vols. London: Routledge.
Bacon, Roger. 1928 (1267). The Opus Majus of Roger Bacon, trans.
R.B. Burke. 2 vols. Philadelphia: University of Pennsylvania Press.
Bailey, D.T.J. 2006. ‘Descartes on the Logical Properties of Ideas.’ British
Journal for the History of Philosophy 14, 3: 401–11.
Baker, Gordon and Katherine Morris. 2002. Descartes’ Dualism. London:
Routledge.
Bolton, Martha. 1987. ‘Berkeley’s Objection to Abstract Ideas and Uncon-
ceived Objects.’ In Essays on the Philosophy of George Berkeley, ed. E. Sosa.
New York: Springer.
 REFERENCES

Bolton, Martha. 2008. ‘Berkeley and Material Representation: Why not a

Lockean Theory of Ideas?’ In New Interpretations of Berkeley’s Thought,
ed. S. Daniel. New York: Humanities Books.
BonJour, Laurence. 2004. ‘In Search of Direct Realism.’ Philosophy and
Phenomenological Research 69, 2: 349–67.
BonJour, Laurence. 2007. ‘Epistemological Problems of Perception.’ In The
Stanford Encyclopedia of Philosophy (Summer 2007 edition), ed. E.N. Zalta.
Available at: https://plato.stanford.edu/archives/sum2007/entries/perception-
episprob (last accessed: December 18, 2016).
Boyle, Robert. 1664. Experiments and Considerations Touching Colours.
London: Henry Harringman.
Bréhier, Émile. 1938. ‘Les “jugements naturels” chez Malebranche.’ Revue
Philosophique de la France et de l’Étranger 125, 3/4: 142–50.
Brower, Jeffrey E. and Susan Brower-Toland. 2008. ‘Aquinas on Mental
Representation: Concepts and Intentionality.’ The Philosophical Review
117, 2: 193–243.
Brown, Deborah. 2006. Descartes and the Passionate Mind. Cambridge:
Cambridge University Press.
Buroker, Jill Vance. 1994. ‘Arnauld on Judging and the Will.’ In Interpreting
Arnauld, ed. E. Kremer. Toronto: University of Toronto Press.
Chignell, Andrew. 2009. ‘Descartes on Sensation.’ Philosopher’s Imprint
9, 5: 1–22.
Clarke, Desmond. 2003. Descartes’s Theory of Mind. Oxford: Oxford Uni-
versity Press.
Clarke, Edwin and Kenneth Dewhurst. 1972. An Illustrated History of Brain
Function. Oxford: Landford Publications.
Cook, Monte. 2002. ‘Robert Desgabets’s Representation Principle.’ Journal of
the History of Philosophy 40, 2: 189–200.
Cook, Monte. 2008. ‘Desgabets as a Cartesian Empiricist.’ Journal of the
History of Philosophy 46, 4: 501–15.
Cottingham, John. 1978. ‘A Brute to the Brutes?’ Philosophy 53 (206): 551–9.
Cottingham, John. 1992. ‘Cartesian Dualism: Theology, Metaphysics, and
Science.’ In The Cambridge Companion to Descartes, ed. Cottingham.
Cambridge: Cambridge University Press.
Cottingham, John. 2008. Cartesian Reflections. Oxford: Oxford University Press.
De Rosa, Raffaella. 2010. Descartes and the Puzzle of Sensory Representation.
Oxford: Oxford University Press.
Descartes, René. 1984. The Philosophical Writings of Descartes, ed. John
Cottingham, Robert Stoothoff, and Dugald Murdoch. 3 vols. New York:
Cambridge University Press.
 REFERENCES 

Descartes, René. 1989. The Passions of the Soul, trans. Stephen Voss. Indian-
apolis: Hackett.
Descartes, René. 1904. Oeuvres de Descartes, ed. C. Adam and P. Tannery.
11 vols. Paris: J. Vrin.
Descartes, René. 1998. The World and Other Writings, ed. and trans.
S. Gaukroger. Cambridge: Cambridge University Press.
Desgabets, Robert. 1675. Critique de la Critique de la Recherche de la Vérité.
Paris: Jean Dupuis.
Desgabets, Robert. 1983–1985. Oeuvres Philosophiques Inédites, ed. J. Beaude.
7 vols. Amsterdam: Quadratures.
Digby, Sir Kenelm. 1658. Two Treatises, in the One of which, the Nature of
Bodies, in the other, the Nature of Mans Soul is Looked Into. Printed for
J. Williams. London.
Easton, Patricia. 2005. ‘Desgabets’s Indefectibility Thesis—a Step Too Far?’
In Receptions of Descartes, ed. Tad M. Schmaltz. London: Routledge.
Everson, Stephen. 1997. Aristotle on Perception. Oxford: Clarendon Press.
Falkenstein, Lorne. 1995. Kant’s Intuitionism. Toronto: University of
Toronto Press.
Falkenstein, Lorne. 2000. ‘Reid’s Account of Localization.’ Philosophy and
Phenomenological Research 61, 2: 305–28.
Field, Richard. 1993. ‘Descartes on the Material Falsity of Ideas.’ The Philo-
sophical Review 102, 3: 309–33.
Foucher, Simon. 1675. Critique de la Recherche de la vérité, où l’on examine
en méme temps une partie des principes de Mr Descartes. Paris: Martin
Coustelier.
Foucher, Simon. 1679. Nouvelle dissertation sur la Recherche de la vérité,
contenant la Réponse à la Critique de la Critique de la Recherche de la
Verité. Paris: Robert de la Caille.
Foucher, Simon. 1675/1995. Critique of the Search for the Truth in which is
examined at the same time a part of the Principles of Mr. Descartes, trans.
Richard Watson. In Malebranche’s First and Last Critics, eds. R. Watson
and M. Grene. Carbondale: Southern Illinois University Press. 1995.
Frede, Dorothea. 2001. ‘Aquinas on phantasia.’ In Perler ed., 2001.
Galilei, Galileo. 1623/2009. The Assayer, trans. A.C. Danto. In Modern
Philosophy, eds. R. Ariew and E. Watkins. Indianapolis: Hackett.
Gallois, André. 1974. ‘Berkeley’s Master Argument.’ The Philosophical
Review 83, 1: 55–69.
Gaukroger, Stephen. 1980. ‘Aristotle on Intelligible Matter.’ Phronesis 25,
2: 187–97.
Gaukroger, Stephen. 1995. Descartes. Oxford: Oxford University Press.
 REFERENCES

Gaukroger, Stephen. 2002. Descartes’ System of Natural Philosophy. Cam-

bridge: Cambridge University Press.
Gaukroger, Stephen, Shuster, John, and John Sutton, eds. 2000. Descartes’
Natural Philosophy, London: Routledge.
Gibson, James. 1950. The Perception of the Visual World. Boston: Houghton
Mifflin.
Goldman, Alvin. 1979. ‘What is Justified Belief?’ In Justification and Know-
ledge, ed. George Pappas. Boston: D. Reidel.
Gregoric, Pavel. 2007. Aristotle on the Common Sense. Oxford: Oxford
University Press.
Hain, Raymond. 2015. ‘Aquinas and Aristotelian Hylomorphism.’ In Aristotle
in Aquinas’s Theology, eds. G. Emery and M. Levering. Oxford: Oxford
University Press.
Hatfield, Gary C. 1990. The Natural and the Normative: Theories of
Spatial Perception from Kant to Helmholtz. Cambridge, MA: The MIT
Press.
Hatfield, Gary C. 2011. ‘Kant and Helmholtz on Primary and Secondary
Qualities.’ In Nolan ed., 2011.
Hatfield, Gary C. and William Epstein. 1979. ‘The Sensory Core and the
Medieval Foundations of Early Modern Perceptual Theory.’ Isis 70,
3: 363–84.
Hill, Jonathan. 2011. ‘Berkeley’s Missing Argument.’ British Journal for the
History of Philosophy 19, 1: 47–77.
Hoffman, Paul. 2002. ‘Direct Realism, Intentionality, and the Objective Being
of Ideas.’ Pacific Philosophical Quarterly 83, 2: 163–79.
Hume, David. 2000. A Treatise of Human Nature, eds. D.F. and M. Norton.
Oxford: Clarendon.
Jackson, Frank. 1977. Perception: A Representative Theory. Cambridge:
Cambridge University Press.
Johansen, Thomas K. 2012. The Powers of Aristotle’s Soul. Oxford: Oxford
University Press.
Jolley, Nicholas. 1988. ‘Leibniz and Malebranche on Innate Ideas.’ The
Philosophical Review 97, 1: 71–91.
Jolley, Nicholas. 1990. The Light of the Soul. Oxford: Oxford University Press.
Jolley, Nicholas. 1994. ‘Intellect and Illumination in Malebranche.’ Journal of
the History of Philosophy 32: 209–23.
Jolley, Nicholas. 2000. ‘Malebranche on the Soul.’ In S. Nadler ed. 2000.
Keating, Laura. 1999. ‘Mechanism and the Representational Nature
of Sensation in Descartes.’ Canadian Journal of Philosophy 29, 3: 411–30.
Kenny, Anthony. 1993. Aquinas on Mind. London: Routledge.
 REFERENCES 

Kepler, Johannes. 1604. Ad Vitellionem Paralipomena, Quibus Astronomiae

Pars Optica Traditur. Frankfurt: Claudius Marnius and the Heirs of
Joannes Aubrius, 1604.
Kepler, Johannes. 2000. Optics: Paralipomena to Witelo & Optical Part of
Astronomy, trans. W.H. Donahue. Santa Fe: Green Lion Press.
Kremer, Elmar. 1994. ‘Arnauld’s Philosophical Notion of an Idea.’ In The
Great Arnauld, ed. E. Kremer. Toronto: University of Toronto Press.
Kremer, Elmar. 2012. ‘Antoine Arnauld.’ In The Stanford Encyclopedia
of Philosophy (Fall 2012 edition), ed. E.N. Zalta. Available at: https://plato.
stanford.edu/archives/fall2012/entries/arnauld (last accessed: December 18,
2016).
La Forge, Louis de. 1666. Traité de l’Esprit de l’Homme, de ses facultez et
fonctions, et de son union avec le corps. Suivant les Principes de René
Descartes, Paris: Theodore Girard.
La Forge, Louis de. 1997. Treatise on the Human Mind, trans. Desmond
Clarke. Dordrecht: Kluwer.
Lear, Jonathan. 1988. Aristotle: The Desire to Understand. Cambridge: Cam-
bridge University Press.
Lennon, Thomas. 1993. The Battle of the Gods and Giants: The Legacies of
Descartes and Gassendi 1655–1715. Princeton, NJ: Princeton University
Press.
Lennon, Thomas. 2007. ‘The Eleatic Descartes.’ Journal of the History of
Philosophy 45, 1: 29–45.
Levine, Joseph. 1983. ‘Materialism and Qualia: The Explanatory Gap.’ Pacific
Philosophical Quarterly 64, 4: 354–61.
Lindberg, Douglas C. 1976. Theories of vision from al-Kindi to Kepler.
Chicago: University of Chicago Press.
Locke, John. 1823. The Works of John Locke. 10 vols. London: Thomas Tegg.
Locke, John. 1975. An Essay concerning Human Understanding, ed. P. H. Nidditch.
Oxford: Clarendon Press.
Machamer, P. and J.E. McGuire. 2009. Descartes’s Changing Mind. Prince-
ton, NJ: Princeton University Press.
MacKenzie, A.W. 1989. ‘Descartes on Life and Sense.’ Canadian Journal of
Philosophy 19, 2: 163–92.
Malebranche, Nicolas. 1674. De la Recherche de la Verité. Paris: André
Pralard. (Books 1–3 only).
Malebranche, Nicolas. 1675. De la Recherche de la Verité. Paris: André
Paralard. (Books 4–6 only).
Malebranche, Nicolas. 1678. De la Recherche de la Verité. Paris: André
Pralard. (Includes the Elucidations).
 REFERENCES

Malebranche, Nicolas. 1958–67. Oeuvres Complètes de Malebranche. Dir.

A. Robinet. 20 vols. Paris: J. Vrin.
Malebranche, Nicolas. 1997a. The Search after Truth, trans. T. Lennon and
P. Olscamp. Cambridge: Cambridge University Press.
Malebranche, Nicolas. 1997b. Dialogues on Metaphysics and on Religion,
trans D. Scott, ed. N. Jolley. Cambridge: Cambridge University Press.
Marmodoro, Anna. 2011. ‘Aristotle on Complex Perceptual Content.’ Philo-
sophical Inquiry 34, 1–2: 15–65.
Marmodoro, Anna. 2014a. Review of Johansen, The Powers of Aristotle’s
Soul. British Journal for the History of Philosophy 22, 1: 174–91.
Marmodoro, Anna. 2014b. Aristotle on Perceiving Objects. Oxford: Oxford
University Press.
Maull, Nancy. 1978. ‘Cartesian Optics and the Geometrization of Nature.’
The Review of Metaphysics 32, 2: 253–73.
McCracken, Charles. 1983. Malebranche and British Philosophy. Oxford:
Oxford University Press.
Meyering, Theo C. 1989. Historical Roots of Cognitive Science. New York:
Springer.
Miller, Timothy D. 2008. ‘Desgabets on Cartesian Minds.’ British Journal for
the History of Philosophy 16, 4: 723–45.
Millikan, Ruth G. 1987. Language, Thought, and Other Biological Categories.
Cambridge, MA: MIT Press.
Millikan, Ruth G. 1989. ‘Biosemantics.’ Journal of Philosophy 86, 8: 281–97.
Morris, John. 1969. ‘Pattern Recognition in Descartes’ Automata.’ Isis 60
(1969): 451–60.
Muehlmann, Robert. 1992. Berkeley’s Ontology. Indianapolis: Hackett.
Nadler, Steven. 1989. Arnauld and the Cartesian Philosophy of Ideas. Prince-
ton, NJ: Princeton University Press.
Nadler, Steven. 1992. Malebranche and Ideas. New York: Oxford University
Press.
Nadler, Steven, ed. 2000. The Cambridge Companion to Malebranche. Cam-
bridge: Cambridge University Press.
Ndiaye, Aloyse Raymond. 1991. La Philosophie d’Antoine Arnauld. Paris:
J. Vrin.
Nelson, Alan. 1996. ‘The Falsity in Sensory Ideas: Descartes and Arnauld.’
In Interpreting Arnauld, ed. E. Kremer. Toronto: University of Toronto
Press.
Newman, Lex. 2001. ‘Unmasking Descartes’s Case for the Bête Machine
Doctrine.’ Canadian Journal of Philosophy 31 (3): 389–425.
Nolan, Lawrence. 2011. ‘Descartes on “What We Call Color.”’ In Nolan, ed., 2011.
 REFERENCES 

Nolan, Lawrence, ed. 2011. Primary and Secondary Qualities. New York:
Oxford University Press.
Nolan, Lawrence. 2012. ‘Malebranche on Sensory Cognition and “Seeing
As.” ’ Journal of the History of Philosophy 50, 1: 21–52.
Normore, Calvin. 1986. ‘Meaning and Objective Being: Descartes and His
Sources.’ In Essays on Descartes’ Meditations, ed. A.O. Rorty. Berkeley:
University of California Press.
Ott, Walter. 2002. ‘Propositional Attitudes in Modern Philosophy.’ Dialogue
41, 3: 1–18.
Ott, Walter. 2004. Locke’s Philosophy of Language. Cambridge: Cambridge
University Press.
Ott, Walter. 2006. ‘Descartes and Berkeley on Mind.’ British Journal for the
History of Philosophy 14, 3: 437–50.
Ott, Walter. 2009. Causation and Laws of Nature in Early Modern Philoso-
phy. Oxford: Oxford University Press.
Ott, Walter. 2014. ‘Malebranche and the Riddle of Sensation.’ Philosophy and
Phenomenological Research 88, 3: 689–712.
Owens, Joseph. 1982. ‘Aristotle on Common Sensibles and Incidental
Perception.’ Phoenix 36: 215–36.
Panaccio, Claude. 2001. ‘Aquinas on Intellectual Representation.’ In Perler
ed., 2001.
Pasnau, Robert. 1997. Theories of Cognition in the Later Middle Ages. New
York: Cambridge University Press.
Pasnau, Robert. 2002. Thomas Aquinas on Human Nature. Cambridge:
Cambridge University Press.
Peppers-Bates, Susan. 2005. ‘Does Malebranche Need Efficacious Ideas?’
Journal of the History of Philosophy 43, 1: 83–105.
Perler, Domink, ed. 2001. Ancient and Medieval Theories of Intentionality.
London: Brill.
Pessin, Andrew. 2009. ‘Mental Transparency, Direct Sensation, and the
Unity of the Cartesian Mind.’ Studies in the History of Philosophy of Mind
9: 1–37.
Pitcher, George. 1977. Berkeley. London: Routledge.
Polansky, Ronald. 2007. Aristotle’s De Anima. Cambridge: Cambridge
University Press.
Pyle, Andrew. 2003. Malebranche. London: Routledge.
Radner, Daisie. 1978. Malebranche. Amsterdam: Van Gorcum Assen.
Radner, Daisie. 1994. ‘Malebranche and the Individuation of Perceptual
Objects.’ In Individuation and Identity in Early Modern Philosophy, eds.
K.F. Barber and J.J.E. Gracia. Albany: State University of New York Press.
 REFERENCES

Régis, Pierre-Sylvain. 1690. Cours Entier de Philosophie, ou Systeme General

selon les Principes de M. Descartes. Amsterdan: Huguetan.
Reid, Jasper. 2003. ‘Malebranche on Intelligible Extension.’ British Journal
for the History of Philosophy 11, 4: 581–608.
Reid, Jasper. 2014. ‘Descartes and the Individuation of Bodies.’ Archiv für
Geschichte der Philosophie 96, 1: 38–70.
Rodis-Lewis, Geneviève. 1963. Nicolas Malebranche. Paris: Presses Universi-
taires de France.
Sabra, A.I. 1978. ‘Sensation and Inference in Alhacen’s Theory of Visual
Perception.’ In Studies in Perception, eds. P. Machamer and R. Turnbull.
Columbus: Ohio State.
Schmaltz, Tad. 1996. Malebranche’s Theory of the Soul. Oxford: Oxford
University Press.
Schmaltz, Tad. 2000. ‘Malebranche on Ideas and the Vision in God.’ In
Nadler ed., 2000.
Schmaltz, Tad. 2002. Radical Cartesianism. Cambridge: Cambridge Univer-
sity Press.
Scott, David. 2010. ‘Resemblance as a Principle of Representation in Descar-
tes’s Philosophy.’ American Catholic Philosophical Quarterly 84, 3:
483–512.
Secada, Jorge. 2000. Cartesian Metaphysics. Cambridge, UK: Cambridge
University Press.
Sergeant, John. 1984. Solid Philosophy Asserted Against the Fancies of the
Ideists. New York: Garland.
Shapiro, Lionel. 2012. ‘Objective Being and “Ofness” in Descartes.’ Philoso-
phy and Phenomenological Research 84, 2: 378–418.
Shoemaker, Sydney. 2002. ‘Content, Character and Color.’ Whitehead
Lectures, Harvard University.
Simmons, Alison. 1999. ‘Are Cartesian Sensations Representational?’ Noûs
33: 347–69.
Simmons, Alison. 2003. ‘Descartes on the Cognitive Structure of Sensory
Experience.’ Philosophy and Phenomenological Research 67, 3: 549–79.
Simmons, Alison. 2008. ‘Guarding the Body: A Cartesian Phenomenology of
Perception.’ In Contemporary Perspectives on Early Modern Philosophy:
Essays in Honor of Vere Chappell, ed. P. Hoffman, D. Owen, and G. Yaffe.
New York: Broadview Press.
Simmons, Alison. 2009. ‘Sensation in a Malebranchean Mind.’ Studies in the
History of Philosophy of Mind 9: 105–29.
Slezak, Peter. 2000. ‘Descartes’ Startling Doctrine of the Reverse-Sign Rela-
tion.’ In Descartes’ Natural Philosophy, eds. S. Gaukroger, J. Schuster, and
J. Sutton. London: Routledge.
 REFERENCES 

Smith, A. Mark. 2004. ‘What is the History of Medieval Optics Really

About?’ Proceedings of the American Philosophical Society 148, 2: 180–94.
Sorabji, Richard. 1993. Animal Minds and Human Morals. Ithaca: Cornell
University Press.
Sowaal, Alice. 2004. ‘Cartesian Bodies.’ Canadian Journal of Philosophy 34,
2: 217–40.
Stuart, Matthew. 2013. Locke’s Metaphysics. Oxford: Oxford University
Press.
Suárez, Francisco. 1614. Francisci Suárez Metaphysicorum Disputationum.
2 vols. Cologne.
Suárez, Francisco. 1978–1981. Comentaria una cum questionibus in libros
Aristotelis De Anima, ed. and trans. Salvador Castellote. 2 vols. Madrid:
Labor.
Sutton, John. 2007. Philosophy and Memory Traces. Cambridge: Cambridge
University Press.
Symington, Paul. 2011. ‘Thomas Aquinas, Perceptual Resemblance, Categor-
ies, and the Reality of Secondary Qualities.’ Proceedings of the American
Catholic Philosophical Association 85: 237–52.
Thomas, Emily. 2015. ‘In Defense of Real Cartesian Motion: A Reply to
Lennon.’ Journal of the History of Philosophy 53, 4: 747–62.
Thomas, Janice. 2006. ‘Does Descartes Deny Consciousness to Animals?’
Ratio 19, 3: 336–63.
Tye, Michael. 1984. ‘The Adverbial Approach to Visual Experience.’
The Philosophical Review 93, 2: 195–225.
Tye, Michael. 2007. ‘New Troubles for the Qualia Freak.’ In Contemporary
Debates in Philosophy of Mind, eds. B. McLaughlin and J. Cohen. London:
Blackwell.
Vinci, Tom. 2008. ‘Mind-body causation, mind-body union and the “special
mode of thinking” in Descartes.’ British Journal for the History of Philoso-
phy 16, 3: 461–88.
Watson, Richard. 1995. Representational Ideas from Plato to Patricia
Churchland. New York: Springer.
Wee, Cecilia. 2005. ‘Animal Sentience and Descartes’s Dualism.’ British
Journal for the History of Philosophy 13, 4: 611–26.
Wee, Cecilia. 2006. Material Falsity and Error in Descartes’s Meditations.
London: Routledge.
Wells, Norman. 2003. ‘The Conimbricenses, Descartes, Arnauld, and the
Two Ideas of the Sun.’ Modern Schoolman 81, 1: 27–56.
Wilson, Catherine. 2003. Descartes’s Meditations: An Introduction.
Cambridge: Cambridge University Press.
Wilson, Margaret. 1978. Descartes. London: Routledge.
 REFERENCES

Wilson, Margaret. 1990. ‘Descartes on the Representationality of Sensation.’

In Central Themes in Early Modern Philosophy, eds. M. Kulstad and
J. Cover. Indianapolis: Hackett.
Wilson, Margaret. 1993. ‘Descartes on the Perception of Primary Qualities.’
In Essays on the Philosophy and Science of René Descartes, ed. S. Voss.
New York: Oxford.
Wilson, Timothy D. 2002. Strangers to Ourselves: Discovering the Adaptive
Unconscious. Harvard, MA: Harvard University Press.
Winkler, Kenneth. 1989. Berkeley: An Interpretation. Oxford: Oxford
University Press.
Wittgenstein, Ludwig. 1953. Philosophical Investigations, trans. G.E.M.
Anscombe. New York: Macmillan.
Wittgenstein, Ludwig. 1960. The Blue and Brown Books, trans. G.E.M.
Anscombe. New York: Harper.
Wolf-Devine, Celia. 1993. Descartes on Seeing. Carbondale: Southern Illinois
University Press.
Yolton, John. 2000. ‘Response to my Fellow Symposiasts.’ In Descartes’
Natural Philosophy, eds. S. Gaukroger, J. Schuster, and J. Sutton. London:
Routledge.
Index
abstraction 20–1, 52–3, 128, 172, 196,
see also phantasms
adverbial theory of perception 161–6
animals 20–1, 28, 30, 32–5, 37, 40, 43,
54, 59–61, 86, 109, 219–20
animal spirits 27–8, 31, 35–6, 38, 40,
58, 84
Alquié, Ferdinand 158, 170, 196 n.6,
205 n.29
Aquinas, St. Thomas 20–6, 51–3, 73
n.32, 171–2, 208, 210 n.42
Aristotle 7, 20–6, 52–3, 186, 196
Arnauld, Antoine 4, 34–5, 81, 138,
150–6, 221, see also Malebranche,
exchanges with Arnauld
Augustine, St. 12, 158–9, 188, 208 n.34
Averroes 52 n.59
Bacon, Roger 27 n.58, 34 n.11, 59–60
Baconian synthesis 1, 26–9, 35, 59–60,
110, 128, 219
Bailey, D.T.J. 77 n.48
Berkeley, George 167 n.25, 215–17, 222
bête machine doctrine 33–5, 220
Block, Ned 19
Bolton, Martha 215–17
Boyle, Robert 17–18
brain images 31, 38–42, 47–57, 62, 65–6,
84–90, 92–3, 98, 100–5, 128–9,
145–7, 219, see also ideas,
corporeal; selection argument
Bréhier, Émile 173
Brown, Deborah 125–6, 188–9
Chignell, Andrew 55–6
Clarke, Desmond 44 n.32
color 22, 24–5, 61–2, 118–22, 145, 155,
172–3, 189–90
constancy of 103, 105, 121, 187
in Aristotelianism 24–5
in brain images 41
see also qualities, sensible; répandre
(spreading)
conversio doctrine 2, 21–2, 31, 47–57, 87,
97–8, 110–11, 128, 129–30, 219
corporeal species see brain images
Cottingham, John 33 n.6, 161 n.8
De Rosa, Raffaella 67, 77, 89, 93, 95 n.99
Descartes, René
Dioptrique 57–8, 100–11, 115, 129
Discours 33–4
L’Homme 30–57, 119
La Lumiere 44–7, 122–3
Meditations
First 64, 90
Second 68, 89–90, 91–2
Third 68–72, 79–84
Fourth 61 n.84, 120
Fifth 82, 131
Sixth 51, 53, 64–5, 84, 86–99
Notae 46, 118 n.13, 124, 131
Passions 38 n.18, 122–7, 129, 143
Principles 11, 14, 16–17, 83, 106–7,
124–7, 129, 198 n.14
Regulae 36–7, 39 n.23, 41, 102, 115
Replies
First 71–2
Second 70, 85–6
Third 131–2
Fourth 81–2
Fifth 78–9, 88
Sixth 113–22, 129, 142, 148–9
Desgabets, Robert 4, 12–13, 40, 49, 74,
83, 137–9, 144–8, 150, 192, 221
direct realism 65, 144–8, 150–4, 170 n.37
distance 42, 57–62, 107–11, see also size
constancy
ecological properties 95
efficacious ideas 158, 170 n.35, 194 n.1,
199–202
Epstein, William 28–9, 59 n.77, 98
Erisman, Theodor and Ivo Kohler 57
extension 81–2, 95 n.99, 114–19, see also
intelligible extension; qualities,
geometrical
externalism 9, 94, see also teleosemantics
Foucher, Simon 40, 75–6, 215, 219
 INDEX
Galilei, Galileo 14–16, 128
Gallois, André 216 n.54
Gassendi, Pierre 78–9, 88
Gaukroger, Stephen 32 n.2, 57 n.72, 60,
114 n.2, 155 n.44, 196 n.9
Goldman, Alvin 9
Hatfield, Gary 10 n.6, 18 n.31, 28–9,
59 n.77, 98
Hobbes, Thomas 12, 131
Hoffman, Paul 151 n.32
Hume, David 13, 222
ideas
corporeal 35–42, 62, 89, 93 n.92, 219,
see also brain images
definition of in Descartes 70–3, 203
descriptive account of 77
essentially representational 69
innate 66–7, see also selection
argument; sensations, fusion
account of
objective being of 69–73, 151–3,
166–7, 203–17, 221
roles in Malebranche 169–73
strict and loose sense in Descartes 68,
77–9
see also direct realism; efficacious
ideas; intelligible extension;
material falsity; representation;
resemblance; sensations, fusion
account of
illusions 25, 103, 114, 152, 183,
see also size constancy
imagination 20–2, 36, 47–8, 52–3, 85,
87–90, 120, 145–9
individuation 10, 127–30, 142, 147,
159, 172–3, 175, 188–9, 202,
222–3
see also qualities, geometrical
inference see judgment
infinity, idea of 204–6
intelligible extension 5, 155, 158–9,
171–3, 182 n.2, 194–200, 203
not a Cartesian idea 206–14, 221–2
Jolley, Nicholas 68 n.14, 70 n.21, 135
n.47, 158 n.4, 162, 200 n.19, 202,
203 n.22
judgment
in Aristotelianism 22–6
in Descartes 117–18
in Malebranche 179–80, 183–4
see also natural geometry, natural
judgments
Kant, Immanuel 196, 222
Keating, Laura 67 n.9, 97 n.103
Kepler, Johannes 1, 28–9, 31, 37–9, 58
n.76, 128, 219
Kremer, Elmar 150 n.79, 154 n.42
Kripke, Saul 19
La Forge, Louis de 4, 49–50, 120,
138–43, 220–1
Lennon, Thomas 171, 195, 213 n.46
Levine, Joseph 18
Lindberg, Douglas C. 28 n.62
localization 10, 61–2, 110, 127–30,
141–3, 147–8, 155, 175, 192, 202,
222–3
see also qualities, sensible
Locke, John 13, 17 n.27, 21, 49 n.46,
72–3, 185 n.9
McCracken, Charles 217 n.57
Mackenzie, Ann Wilbur 67 n.9, 97 n.103
Malebranche, Nicolas
and L’Homme 49
Conversations Chrétiennes 201
Dialogues 168–9, 197–8, 204
Elucidations 165, 182 n.2, 195–9, 204,
210, see also intelligible extension
Méditations Chrétiennes et
Métaphysiques 209
The Search After Truth
1674 edition 176–81, 185, 225
see also natural judgments,
répandre, sensations as
instrumental to preserving the
mind-body union
exchanges with Arnauld 151, 155–6,
172–3, 205–6, 212
exchanges with Régis 172–3, 205
see also efficacious ideas; ideas, innate;
imagination; intelligible
extension; selection argument
Marmodoro, Anna 23
material falsity 66, 79–84, 130, 220
memory 20, 31, 36–42, 44 n.32, 52, 147
Meyering, Theo C. 185 n.8
Millikan, Ruth Garrett 8 n.1

Nadler, Steven 150 n.29, 163, 170, 173,
205, 207 n.32
Nagel, Thomas 18–19
natural geometry 2–3, 31, 43, 57–62, 65,
97, 99, 101, 107–9, 115, 128–9, 178
natural judgments
anemic 183–93, 194, 198–9
robust 176–81, 187
two layers of 177–8
see also répandre; sensations, compound
Ndiaye, Aloyse Raymond 153
Newman, Lex 32 n.4, 33 n.6
Newton, Isaac 15 n.21, 218 n.1
Nolan, Lawrence 158 n.3, 161 n.8, 165,
188–9, 195, 199–202
overlap thesis 30–5, 43, 97, 101, 130
Owens, Joseph 23
Panaccio, Claude 72 n.27, 73 n.30, 75 n.39
participation 210–11, see also intelligible
extension
Pasnau, Robert 73 n.32
Peppers-Bates, Susan 200 n.18, 209 n.40
Pessin, Andrew 72 n.26
phantasms 20–6, 51–4, see also brain
images; conversio doctrine;
imagination; memory
phenomenal character 18
pineal gland 2, 7, 36–42, 47–8, 54, 60–1
Pitcher, George 216 n.53
positioning 10, 127–30, 147, 159, 189,
202, 222–3
see also individuation; qualities
qualities
geometrical (common sensibles)
and judgment (in
Aristotelianism) 22–6
defined 13–16
sensible (proper sensibles)
defined 11–19
dispositional treatments of 17
ejection from the mind-
independent world 12–13
eliminativism about 16
not phenomenal character 18
see also color; sensations
Radner, Daisie 135 n. 48, 196 n.6, 205
‘rainbow-colored soul’ 145–6, 166–7
INDEX 
rapporter (referring) 122–7, 136, 143,
148, 188–9
see also judgments, localization
real distinction (between mind and
body) 90–1
referring see rapporter
Régis, Pierre-Sylvain 4, 137–8, 148–50,
221, see also Malebranche,
exchanges with Régis
Reid, Jasper 208, 214
répandre (spreading) 155, 197, 200, 209,
see also natural judgments
representation
corporeal 35, 39–40, 42, 51, 86, 130,
200, see also brain images;
pineal gland and phantasms 21
internalist vs. externalist senses
of 8–10
see also ideas, objective being of;
intelligible extension, not a
Cartesian idea; resemblance;
sensations, not representations
resemblance
entitative vs. intentional 40, 73–5,
78–9, 84, 130, 145, 204–6, 215
intentional resemblance in La
Forge 139–43, 146–7
of mental ideas in Descartes 73–9,
96–7
see also brain images; representation,
corporeal
rete mirabile 38
Rodis-Lewis, Geneviève 157 n.1, 195 n.2
Sabra, A.I. 58 n.76
Schmaltz, Tad 91 n.88, 131, 137 n.2, 139
n.5, 146, 161 n.8, 162 n.10, 171
n.39, 200, 202
Scott, David 74
Secada, Jorge 63 n.2, 70, 117 n.12
selection argument (Malebranche) 3–4,
131–6, 149–50, 157, 181, 221
sensations
as opposed to ideas of
sensations 81–5
as preserving the mind-body
union 93–4, 168–9, 174
cannot be related to each other
180–1, 191
compound 158, 185–91
defined 15–16
 INDEX
sensations (cont.)
fusion account of 67, 89, 93
not representations 10, 65, 67–79, 96–7,
117, 130, 141, 159, 163–6, 220
see also material falsity
semiotic vs. causal account of 44–7,
105–7, 122–3
see also color; individuation; qualities,
sensible
sensory core 98–9, 110, 116, 178
Sergeant, John 72–3
signs and signification 31, 39, 44–7,
93–5, 106–7, see also sensations,
semiotic vs. causal account of
Simmons, Alison 15 n.22, 43 n.31, 66,
95–6, 114 n.2, 115 n.6, 116, 120,
165, 202, see also sensations,
fusion account of; teleosemantics
size constancy 103–5, 195–6, 198,
see also positioning; qualities,
geometrical
skepticism 90, 212
Sorabji, Richard 100 n.1
Suárez, Francisco 70–1, 130
substance/mode distinction 166–7, 215
teleosemantics 9, 94–7, see also
externalism
Voss, Steven 123 n.25, 125–6
Wilson, Margaret 58 n.75, 67 n.10,
69 n.19, 120 n.17
Winkler, Kenneth 216 n.54
Wolf-Devine, Celia 26–7, 37 n.17, 41
n.27, 103 n.7, 109–10, 119