Professional Documents
Culture Documents
Wayne Waxman - Kant's Anatomy of The Intelligent Mind-Oxford University
Wayne Waxman - Kant's Anatomy of The Intelligent Mind-Oxford University
Wayne Waxman
OXFORD
UNIVERSITY PRESS
OXFORD
UNIVERSITY PRESS
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
1 3 5 7 9 8 6 4 2
Printed in the United States of America
on acid-free paper
To Alison,
Even More Than Before
This page intentionally left blank
{PREFACE}
According to current philosophical lore, Kant rejected the notion that philosophy
can progress by psychological means. While his reliance on a theory of pure sensible
intuition in his account of a priori knowledge may cast doubt on how successful
he was, Kant is honored for having set psychological philosophy firmly on the road
to extinction and begun the process of strictly demarcating philosophy (includ
ing the philosophy of psychology) from psychology-a process that would eventu
ally see Kant's approach to the a priori eclipsed by advanced analytical techniques
unknown in his day.
The aim of this book is to show that the lore has Kant backward. The Kant pre
sented here is (1) a committed proponent of psychological philosophy, indeed (2) its
preeminent exponent (surpassing even Hume), (3) whose positions should not be
lightly dismissed even today, when psychological philosophy is generally regarded
as beyond the pale.
My case for (1) through (3) has two sides. F irst, I show that on closer inspection,
those passages in Kant's writings that have traditionally been taken as arguments
against the introduction of psychology into philosophy turn out to be directed
against introducing empiricism into psychological philosophy but not against psy
chological philosophy as such. The reason one finds no arguments directed spe
cifically against the latter is that there are none. Instead, Kant's actual view was
that so long as psychology is kept strictly non-empirical-"pure" in his nomencla
ture-it is the ideal vehicle for resolving the issues addressed in the Transcendental
Analytic of the Critique of Pure Reason. For how else, other than in the capacity of
psychological philosopher, could he have both rejected empiricism in all its forms
and unqualifiedly endorsed Hume's insistence on the philosophical priority of psy
chological origins and irrelevance of normative indispensability (in the Preface of
the Prolegomena to any Future Metaphysics)? The lore is conspicuously silent on
that one.
Second, there can be no dispute that Kant set pure self-consciousness, expli
cated as original apperception, at the heart of the doctrine of the Transcendental
Analytic. To this manifestly psychological notion, the doctrine variously assigns the
place of supreme principle, highest unity, original source, and fundamental ground.
Clearly, apperception is not something one can afford to misinterpret and still hope
to understand Kant. Yet I will show that just such a misinterpretation is enshrined
in the lore and universally accepted by scholars: the supposition that Kant regarded
the categories as necessary for apperception and apperception as sufficient for the
categories. I present multiple, mutually reinforcing lines of evidence to show that
viii Preface
Kant's view was the converse-apperception as necessary but not sufficient for the
categories and the categories as sufficient but not necessary for apperception-a
view that, by explaining these (indeed, all) concepts through self-consciousness and
not (as commonly supposed) self-consciousness through these concepts, can only
be characterized as psychologism. I argue further that the Transcendental Analytic
places original apperception in the same psychologizing relation vis a vis space,
time, nature, mathematics, and even logic. By examining each psychologization in
turn, I use the cumulative result to show that the Analytic is not only deeply psy
chological but also essentially so, in conception and execution alike. Acceptance
of the case presented in this book therefore reverses the frame from Kant the
anti-psychological critic of psychological philosophy to Kant the psychological
critic of anti-psychological philosophy. It thus would oblige scholars, students, and
admirers of his philosophy generally to unlearn virtually everything they think they
know about the doctrines of the Transcendental Analytic and their role in the criti
cal philosophy as a whole.
My "psychological" approach to Kant should not be confused with others that
have been similarly labeled (Patricia Kitcher, Andrew Brook, Lorne Falkenstein,
etc.). To my knowledge, none of these challenges the lore's assumption that Kant
used space, time, nature, mathematics, and logic to explain the constitution of our
psychology, not vice versa. Since psychology consequently does not take philo
sophical pride of place in any of these readings, mine should be understood as
"psychological" in a quite different, more fundamental sense, the effect of which is
to set Kant in direct succession to Hume rather than in the line leading to contem
porary nativist scientific psychology (practitioners of which sometimes flag Kant as
a forerunner, e.g ., John O'Keefe). At the same time, my approach suggests entirely
new ways of relating Kant to contemporary philosophy and science of mind, which
I will explore in the conclusion of this book.
Title
plan then was to call the second volume "Kant and the Empiricists: Time Out of
Mind" to indicate its continuity with Understanding Understanding, and I referred
to it as such in the 2005 volume. Although another title change was deemed advis
able with its publication now, it retains its special relationship to Understanding
Understanding, and I would recommend anyone who feels they have learned some
thing from this volume to read its predecessor as well.
My one regret is mistitling the 2005 volume. Because only Kant's name is men
tioned, review editors sent it to Kant specialists, who, not being equipped to com
ment on a specialist work of empiricist scholarship, focused almost exclusively on
the preliminary discussions of Kant in the four-chapter General Introduction and
typically neglected to make clear that my treatment of Kant was not in Kant and
the Empiricists: Understanding Understanding at all but destined for a follow-up
volume. As a result, the book was mistaken for a contribution to Kant scholarship
and never reached its intended audience of empiricist scholars and students. This
could all have been avoided had it been titled, e.g., "Locke, Berkeley, and Hume on
Human Understanding: The Path to Kant".
Secondary Literature
apperception against what seems to be the universal view of Kant scholars that the
categories have priority (indeed, I know of none who has even considered the con
verse, much less identified and weighed up the principal arguments pro and con).
Given the status of apperception as the "supreme principle" and "highest point" of
Kant's theory of the understanding, it will come as no surprise that adopting the
position I do should bring to the fore questions that are peripheral or do not even
arise for those who take the categories to be necessary for apperception and, con
versely, shunt to the margins or render irrelevant questions of the first importance
for those who prioritize the categories. Because this means that there are few points
of contact between my interpretation and others (and those few relate to compara
tively superficial matters), there is neither the occasion nor the need to inflate the
book with extensive discussion of secondary literature (nevertheless, I have com
mented extensively on secondary literature in other publications, some of which are
referenced here).
There is, however, one author with whose interpretation mine has a close affin
ity: Beatrice Longuenesse. She and I developed our readings of Kant side by side
during the 1980s (I am the grateful dedicatee of the original French version of
her Kant and the Capacity to Judge). Our influence on one another's readings dur
ing that period was extensive and, for me at least, hugely productive. Although we
approached Kant from a different starting point-she from the logical functions of
judgment, I from sensibility and its psychology-the resulting interpretations of
the Critique of Pure Reason agree on most fundamentals and are certainly closer to
one another than either is to others. The very pervasiveness of our influence on one
another during that time, however, makes it difficult to single out particular debts.
For that reason, while I mention her work only seldom in this volume, that should
be seen as the inverse of the magnitude of my esteem for it.
http://waynewaxman.com
{ACKNOWLEDGMENTS}
I have been very fortunate in having had the opportunity to work with Peter Ohlin,
the philosophy editor at Oxford University Press, New York. He is a consummate
professional without whose unstinting support and encouragement this project,
comprising both the present book and Kant and the Empiricists: Understanding
Understanding (2005), might never have seen the light of day. I am especially grate
ful to the Press because no other university press would agree even to look at a book
on Kant above 300 pages (this includes a press that takes its name from another
famous English university and touts itself as the leading publisher of Kant mono
graphs). It is painful to think of all of the magnificent books on scholarly topics
that would have been either mutilated or never published at all had this policy been
in effect previously.
I wish to thank the anonymous referees for many helpful indications for
improvements in the manuscript. I would also like to thank colleagues and students
who, over the years, have given me the benefit of their insights into Kant, especially
Arthur Melnick, who first introduced me to Kant, Hubert Schwyzer, my disserta
tion director, Richard Aquila, Jason Potter, Rolf Horstmann, Alison McCulloch,
Klaus Steigleder, Beatrice Longuenesse, Quassim Cassam, Chris Chetland, Zenaida
Beatson, Colin Marshall, and Allen Wood.
Portions of Chapter 2 appeared as "Kant's Debt to the Empiricists" in The
Blackwell Companion to Kant, edited by Graham Bird (2006). Chapter 8 was
adapted from "Kant's Refutation of Berkeleyan Idealism" in Jdealismus als Theorie
der Repriisentation? (Paderborn: mentis Verlag GmbH, 2000). Chapters 9 and 17
contain ideas first presented in "What Are Kant's Analogies About?," Review of
Metaphysics (Sept. 1993), and "Kant on the Possibility of Thought," Review of
Metaphysics (June 1995). And the psychologism theme pervading this book made
its first appearance in two articles, "Kant's Psychologism, Pt. I," Kantian Review
(1999) and "Kant's Psychologism., Pt. II," Kantian Review (2000).
This page intentionally left blank
{CONTENTS}
Bibliography 565
Index 567
This page intentionally left blank
{ REFERENCE SCHEME AND ABBREVIATIONS }
When several citations from the same work occur successively in the same para
graph, I have opted not to repeat the title of the work. Internal references to chap
ters and sections take the form "ch 12-D" or "chs 14-C and 15-D." Translations
are my own unless otherwise indicated. References to secondary literature, with
the exception of my own previous books, are given in footnotes. Citations from
primary sources are from the following editions:
Kant. All references are to, or derived from, critical editions based on the
Prussian Academy edition of the Gesammelte Schriften, begun in 1901, and gener
ally are cited as the Akademie Ausgabe, abbreviated "A A." For Kant's published
works, I mostly use the standard abbreviation scheme: "A--/B--" (Critique of Pure
Reason), "CPrR" (Critique of Practical Reason), "CJ" (Critique of Judgment), "ID"
(Inaugural Dissertation), "PFM" (Prolegomena to any Future Metaphysics), and
"MFPNS" (Metaphysical First Principles of Natural Science). Some works I refer
to by name: Discovery (On a Discovery whereby any new Critique of Pure Reason
is to be made superfluous by an older one), Progress (What progress has there been
in Metaphysics since the time of Leibniz and Wolff?), Anthropology (Anthropology
from a Pragmatic Viewpoint), and Logic(as published by Jasche). Letters are cited
by date and recipient.
Locke. The edition of the Essay concerning Human Understanding I employ is
the 1975 Oxford-Clarendon, edited by Peter H. Nidditch, abbreviated "ECHU."
Berkeley. All references are to Philosophical works, including the works on vision,
edited by Michael R. Ayers(London: J.M. Dent, 1975). References will conform to
the following practices: "PHK Intr. § 3" for the third section of the Introduction to
A Treatise concerning the Principles of Human Knowledge, "lDl 79" for the first of
the Three Dialogues between Hylas and Philonous together with the pagination of
the T. E. Jessop and A. A. Luce edition of the complete Works of George Berkeley
(1946-1957); "V" for the New Theory of Vision; "VV" for A New Theory of Vision
Vindicated; and "C" for the Commonplace Books.
Hume. Citations of A Treatise of Human Nature are from the Oxford Clarendon
1978 edition by Peter H. Nidditch, revising the 1888 edition by L. A. Selby-Bigge,
abbreviated "T HN." After a slash("/"), the corresponding page in the new edition
by David Fate Norton and Mary J. Norton is provided(Oxford: Oxford University
Press, 2000). Citations from An Enquiry concerning Human Understanding are
from the 1999 Oxford University Press edition by Tom L. Beauchamp, abbreviated
"EHU IV/i if 3," for the third paragraph(editor's numbering) of the first part of the
fourth section.
xx Reference Scheme and Abbreviations
The purpose of this memo is to signal what is new and noteworthy in the book, and
to provide guidance for comprehending and evaluating it.
I. General Remarks
1 It should be kept in mind throu ghout that I do not claim that Kant's en gagement with clas
sical ("dogmatic") metaphysics can be adequately explained as a consequence of his engagement
with psycholog y. Precisely for that reason, pure speculative reason and its transcendental ideas
fall outside the scope of this book. My focus is confined exclusively to Kant's theories of self
and understandin g, which alone, in my view, are not only essentially psychological but also psy
chological in essentially the same sense as the theories of Locke, Berkeley, and, above all, Hume.
Memo to Readers 5
II. Overview
I make two general claims against which the success or failure of this volume should
be measured. First, I maintain that the categories are not necessary conditions for
apperception as such, merely for apperception in its properly cognitive guise as the
objective unity of experience. There is no concept more important to Kant's theory
of the understanding than apperception; indeed, as the most fundamental level,
apperception is the understanding. Hence, the success or failure of this volume may
best be judged by the effectiveness of the case I make in Chapters 3-5, 9-11, and
13-18 that there are indeed several noncategoriaF and nondiscursive (transcendental
aesthetic) guises of apperception and that these are presupposed by the categories.
Part I recapitulates the basic theses and apparatus of UU, providing a bridge from
Hume, where UU left off, to Kant.
I argued in the UU General Introduction that the seemingly huge differences
that distinguish Kant from the British empiricists are the result of what is, at its
root, a very slight divergence: Kant identified a source of representational content
in the mind that neither Hume nor anyone before him ever thought of, basing his
psychological philosophy on a doctrine of representations that are at once sensible
and pure (i.e., purely a priori). Part I of this volume is intended to establish the same
point in a different way. In Chapter 1, I review numerous texts to show that there is
indeed something in Kant's treatment of the understanding that can and should be
construed as the "psychological a priori"; and in Chapter 2, I present evidence that
Kant derived his conception of the psychological from the empiricists, especially
Hume, and hence owed more to them than to anyone else, past or contemporary,
and certainly a great deal more than empiricist scholars and Kant scholars alike
tend to suppose. Since what most fundamentally separated Kant from the empiri
cists is his doctrine that sensibility enables us to intuit a priori as well as a posteriori,
I have coined the term sensibilism to preserve their underlying affinity, signifying the
thesis that no purely intellectual representational content is possible because (pure
and/or empirical) contributions from sensibility are essential to the constitution of
everything we mentally represent (which, in Kant's case, includes the logical: chs
9-C, 10-E, and 14-E). T he key notion that emerges from these considerations is psy
chologism. I use the term to designate Kant's Hume-derived method of explicating
Memo to Readers 7
the content and delimiting the scope of application of concepts at the heart of
age-old philosophical disputes by tracing them to their origin as representations in
the mind, with an eye to determining whether the psychological operations respon
sible for forming those representations also contribute essential elements of their
content. T he common commitment to sensibilist psychologism, empirical in the
case of the one and transcendental in that of the other, is the basis for my claim that
Hume and Kant were philosophical birds of a feather, the vast differences between
them virtually all stemming from what, at the root, is merely a slight divergence in
their views regarding the psychological sources of mental representations (viz. it
never occurred to Hume that the senses might intuit a priori: UU chs 2-3). I further
claim that, in neglecting Kant's psychologism, interpreters are acknowledging only
half of Kant's debt to Hume-the debt for what he called "Hume's problem" (how
are synthetic a priori judgments possible?)-while overlooking his equally impor
tant debt to Hume for the method of solving it.
Part II (chapters 3--8) focuses on Kant's theory of sensibility and related issues
such as idealism and the thing in itself
Chapters 3-5 are concerned with the contribution of the unity of apperception
to pure intuitions of space and time. Under Kant's novel, transformed conception
of the understanding as the capacity for apperception, this faculty is essentially
defined by pure self-consciousness rather than by discursivity ( = representation by
means of universals, be it in judgments or inferences), thus opening the way for it to
perform a prediscursive role in the constitution of pure space and time.
Chapters 3 and 4 are perhaps the most important in the book inasmuch as every
thing in Parts III-V presupposes the unity of sensibility described in them. All too
often the boundary separating transcendental aesthetic (the doctrine of pure sen
sibility) from transcendental logic is disregarded so that the motivation for Kant's
revolutionary account of sensibility is obscured, and even the fact that he had such
an account is downplayed or ignored. By contrast, I approach Kant's explication of
space and time as pure intuitions of sensibility purely as a doctrine of sensibility,
prior to and independently of his theory of discursive understanding.
My focus in Chapter 3 is on the elements of Kant's doctrine of the unity of sen
sibility: sensation, intuition, and appearance. I consider what pure intuition is and
why it is necessary to sensibility, prior to and independently of the conceptual and
cognitive uses to which it may be put by means of discursive understanding. To this
end, I argue that given sensations (sense affections) alone, nothing like a unified sen
sibility would be possible, i.e., there would be no capacity to represent all possible
affections of sense as a single homogeneous manifold, united in one and the same
consciousness, ahead of all discursivity. Pure intuition is therefore just as essential
to sensibility as sense affections because only it can provide for such a unity of sen
sibility, a unity that proves to be nothing else but a prediscursive expression of the
original synthetic unity of apperception.
In Chapter 4, I apply the account of pure intuition as essential to the unity of
sensibility to the particular cases of space and time. In addition to detailing the
8 Kant's Anatomy of the Intelligent Mind
Part III (Chapters 9-1 I) deals with Kant's metaphysical deduction of the catego
ries, the focus here being as much on pure general logic as on transcendental logic.
Chapter 9 is among the most important in the book because it deals with Kant's
answer to his subjective transcendental deduction (cf. Axvi-xvii) question of how
thought itself is possible, irrespective of its content, and so without regard to
whether it is analytic or synthetic, subjective or objective, immanent or transcen
dent, theoretical or practical, empirical or mathematical or transcendental, etc. At
its core is his explanation of how concepts are possible, i.e., the logical universality
that constitutes their form irrespective of their matter (the object thought in them).
Historically, Kant's account of concepts is important because empiricist psycholo
gism explicated universality in terms of custom, thereby expelling genuinely logical
universality from the mind to language, which, for Hume in particular, is just to say
that it is dependent on convention (i.e., human artifice rather than human nature)
(cf. D D chs 4-C, 8, 10, 14-E, and 18-A and -B). What Kant did-and I take this to
be among his greatest, most original, yet most neglected innovations-was to expli
cate logical universality in terms of the representation of the identity of the I think
(pure self-consciousness as analytic unity of apperception [ADA]), thus restoring
genuine concepts (with properly logical form, i.e., logical universality) from the
public sphere of language to the individual human understanding considered in
isolation (hence the chapter's title: "concepts in mind"). Equally significantly, by
psychologizing pure general logic's most fundamental representation, Kant became
the first sensibilist to subject this science as a whole to the same psychologistic
analysis to which previous sensibilists had subjected objective cognition.
In addition to my concern with Kant's explanation of the possibility of con
cepts in Chapter 9, I argue that the synthetic a priori unity of apperception said to
make ADA possible (Bl33--4) can be none other than the prediscursive, precatego
rial original synthetic unity of sensibility produced ahead of all thought (discursiv
ity) by pure space and time. For how can the categories, as pure concepts of the
understanding, be involved in making possible the unity that all concepts as such
presuppose? Thus, the interpretation of Chapter 9 merges with that of Chapters 3-5
to pave the way for the interpretation of the Transcendental Deduction of the
Categories in Chapters 13-15.
Chapter 10 is a defense of Kant's table of judgments based on the recognition
that, once concepts (ADA-universals) are present in the mind, a means is required
to put them to representational use, starting with the ability to combine them with
other concepts. Since it is the very nature of logically general representations not
to be combinable by purely aesthetic means Uuxtaposition or succession, by color,
smell, etc.), putting them to representational use demands a set of forms of a com
pletely different character than space and time: logical functions of judgment.
These innate functions permit the understanding to combine any ADA-universal
with any other in a synthetic unity that Kant termed "judgment," and then combine
the resulting judgments hypothetically or disjunctively to form complex judgments
and inferences. The result is a further noncategorial (but still discursive) original
10 Kant's Anatomy of the Intelligent Mind
forms in Kant's table of judgments individually, arguing that there can be no other
logical forms than these because there seem to be no other relations of concepts
or judgments that pertain specifically and solely to the logical form conferred on
representations by the analytic unity of apperception.
Chapter 11 deals with the metaphysical deduction of the categories proper: the
explication of the categories as pure concepts of the understanding derived from
logical functions of judgment and the exhibition of these concepts as pure concepts
of an object of sensible intuition in general (be it our type of sensible intuition-via
pure space and time-or any other). There are two texts at the heart of my inter
pretation: B128-9 and PFM 324. At B128-9, Kant provided an example of the
derivation of a category from a logical function and then said, "and so on with the
others." The body of the chapter is my attempt to continue from Kant's "and so
on" (no easy task!). I do so in such a way as to leave open Kant's other applications
of pure concepts of the understanding, in particular the table of categories in the
Critique of Practical Reason, which likewise derives from the table of judgments.
The question is: what is common and what is different between the two tables of
categories? I maintain that there is a step from pure concepts of the understanding
in the abstract to pure concepts of the understanding as concepts of an object of
sensible intuition in general that must ultimately be understood in terms of the syn
thesis intellectualis of the Transcendental Deduction of the Categories (the focus
of ch 14-B). Nevertheless, the metaphysical deduction suffices to solve the first
conceptual part of what Kant termed "Hume's problem": by exhibiting each pure
concept of the understanding as the universal representation of a necessary syn
thetic unity of the manifold, it proves the existence of precisely the kind of concept
Hume denied-concepts of the necessary relation of the distinct (Kant's solution
to the second, epistemological part of Hume's problem is the focus of Chapters 13
and 17).
Part IV (Chapters 12-15) focuses on the heart of Kant's theory of the
understanding, the Transcendental Deduction of the Categories, the purpose of
which is to demonstrate that the categories, explicated as pure concepts of the
understanding in the metaphysical deduction, have a priori validity with respect
to objects.
Chapter 12 asks what a transcendental deduction is. Many answer by empha
sizing the distinction in the opening paragraph of the Deduction's first chapter
between a quid Juris (the title, or right, to apply a concept to objects) and a quid
facti (no need to prove title because the concept derives directly from the object
empirically). These interpreters take Kant's comparison of the Deduction with a
juridical quid Juris to demonstrate that it is anti-psychological and normative in
character. Moreover, given the centrality of the Deduction to Kant's thought, they
tend to hold it up as proof, or something close to it, of the anti-psychological,
Memo to Readers 11
normative character of Kant's entire philosophical project. I argue that this places
far too much weight on an inconclusive introductory text and that what actually
transpires in the Deduction itself is a far more reliable guide to the sort of inquiry
it is. In the remainder of the chapter, I focus on the few texts in which Kant himself
described the Deduction: the distinction between subjective and objective transcen
dental deductions in the A edition Preface of the Critique of Pure Reason and the
huge footnote in the Preface to the Metaphysical First Principles of Natural Science.
My analysis shows that the Transcendental Deduction of the Categories not only is
an exercise in a priori psychology, but is also the single most psychological compo
nent of Kant's entire transcendental philosophy.
Chapter 13 is an examination of the 1781 ("A" edition) version of the
Transcendental Deduction. Insofar as the notions of experience and experience-based
association are at its heart, I relate it to Hume. Kant raised a question about the
possibility of experience-bred customary association that Hume needed to address
but never did: how is it possible to represent the objectively ordered succession
of perceptions implicated in our ability to take cognizance of the constancy and
frequency of conjunctions of perceptions? I argue that this problem needs to be
folded into a yet deeper issue, specific to Kant's system: since sensibility contributes
neither order nor relation of any kind to the manifolds of space and time, it can
do nothing to uniquely differentiate and completely determine each space and each
time with respect to all of the (infinitely many) other spaces and times possible in
those manifolds. In the case of time, this means that sensibility by itself is insuf
ficient to yield the objective succession of appearances requisite to experience the
frequency and constancy of the conjunctions of perceptions and so form associa
tions between them. It is consequently left to the understanding to fill the gap, and
to do so entirely a priori, by means of its pure concepts.
In the remainder of Chapter 13, I argue that Kant's proof for the a priori objec
tive validity of the categories consists in showing them to be the basis of a transcen
dental synthesis of imagination whereby appearances are necessarily related to one
another in such a way as to spatially and temporally differentiate and determine
each with respect to every other. Its starting point is the question, what is meant
by "an object corresponding to, and so distinct from, our representations"? Since
nothing except representations (appearances) are ever apprehended in intuition, the
object can only be an unknown something in general = x; and since this means
that the relation of appearances to an object can only consist in their standing in
a necessary relation to one another, the source of that necessity and the object cor
responding to appearances must be one and the same. Yet because appearances are
apprehended in intuition devoid of all order and relation ("scattered and single
in themselves"), the necessity whereby they represent an object can have no other
source than the understanding and its pure concepts. One way of saying this is that
the object is nothing else than that the concept of which brings about the necessary
synthetic unity of a manifold of appearances; and since this in turn is just to say that
an object and the necessary synthetic unity of the manifold in one consciousness
12 Kant's Anatomy of the Intelligent Mind
(objective unity of apperception) are one and indistinguishable, it means that con
cepts are constitutive of objects insofar as these concepts bring about objective
unity of apperception. However, no necessity of the kind requisite for objective
unity of apperception is possible unless and until the undifferentiated, indetermi
nate appearances apprehended in intuition are subjected to transcendental synthe
sis. Transcendental synthesis must therefore be recognized as making possible an
original, all-encompassing objective unity of apperception that precedes and makes
possible all others, be they objects constructed in pure intuition in conformity with
mathematical concepts or objects met with in experience by means of synthesis of
recognition in a concept. And since transcendental synthesis is an expression of the
categories, nothing more is required to demonstrate that these concepts are condi
tions of the possibility both of cognition and its objects-which is precisely what
Kant set out to prove in the Transcendental Deduction.
Chapter 14 focuses on the 1787 ("B" edition) version of the Deduction. It opens
with a defense of Kant's claim that there is no substantive difference between the
A and B versions. I then pinpoint their expository difference in the tendency of the
earlier version to fold the intellectual synthesis (synthesis intellectualis) of the cate
gories, which holds for any sensible intuition in general, into the figurative synthesis,
or synthesis speciosa, which applies the categories to objects of our sensible intu
ition in particular (spatial and temporal appearances). T he 1787 version improves
on this by dealing with the two syntheses separately and sequentially, thereby per
mitting Kant to highlight the role of the logical functions and categories in the con
stitution of experience and its objects more effectively than he did in the original
version. T he chapter concludes with an appendix that revisits Kant's conception of
pure general logic in the light of the analysis of the Transcendental Deduction in
Chapters 13 and 14, giving final form to my account of why, on Kant's criteria of
the logical, post-Fregean mathematical logic must be classed as mathematics rather
than logic.
Chapter 15 is a category-by-category elucidation of how transcendental syn
thesis speciosa introduces objective unity of apperception into the manifold of
pure-formal space and time. It is, in effect, a chance to see the categories in action
in a way Kant himself never quite managed, at least not in this detail, and certainly
not in respect to the pure formal manifold of space and time presupposed not only
by experiential cognition but by pure mathematics as well (the Schematism, I argue,
concerns not the pure-formal but the empirical-material manifold of appearances).
T his is the only chapter of the book that is more a supplement to Kant's text than
an exposition. My justification for this liberty is the need to address a problem
confronting Kant's theory of the understanding that can be solved in no other way.
I term it the heterogeneity problem: if pure space and time are purely sensible and
the categories purely intellectual, how can they possibly be combined to yield cog
nition without committing the subreption Kant termed "transcendental amphib
oly" (sensibilizing the intellectual a la Locke or intellectualizing the sensible a la
Leibniz)? Although the only text I know of in which Kant expressly acknowledged
Memo to Readers 13
the problem is a draft of a letter to Tieftrunk, it is one that must be solved if his
system is not to unravel at the foundations. The chapter is also important in other
ways. For example, I show in detail that and how Kant's transcendental concept of
space is capable of admitting any topology and number of dimensions mathemat
ics and science may require and compare transcendental synthesis speciosa with
Berkeley's theory of vision to argue that, despite its complexity, it too should be
understood as part of what the infant understanding can and must do in order to
realize its cognitive potential.
Part V (Chapters 16-18) deals with transcendental schematism as an extension
of the transcendental synthesis speciosa of the manifold of pure-formal intuition elu
cidated in Chapter 15 to the empirical-materia/3 manifold of realities apprehended in
perception. It then proceeds to the objective and subjective principles of pure under
standing (transcendental judgments) that predicate the schemata of appearances by
means of the threefold empirical synthesis of the Transcendental Deduction.
My interpretation of transcendental schematism in Chapter 16 takes its start
from what I term the problem of the imperceptibility of space and time. Because they
are pure, and therefore imperceptible, space and time can have no reality for the
perceptible realities present in the field of appearances (where Kant understood
the reality corresponding to sensation, and so too to the matter of appearances,
as "the transcendental matter of all objects as things in themselves," Al43/B182).
Although the supporting texts are drawn mainly from the principles chapter, it must
be remembered that the schemata play the role of predicates in these synthetic a pri
ori transcendental judgments (A180-1/B223--4). Thus, I argue that in being predi
cated of the matter of appearances (via Kant's threefold empirical synthesis), the
schemata determine empirical realities so that they have all the objective features
they would have ifthere were a space and time in the field of appearances for them
to occupy and interact within.
Chapter 17 deals with Kant's system of principles of pure understanding, focus
ing particularly on the three Analogies of Experience. My consideration of the first
two Analogies is framed so as to bring Kant's accounts of substance and cause and
effect into direct relation with Hume, taking the linear time series of perceptions
presupposed by experience and customary association examined in Chapter 13 as
my focus. Since the Hume portions of the discussion rely heavily on my analysis of
his account of personal identity in UU chs 3 and 17, this background is recalled in
order to show, first, why nothing less than a true permanent will do for the principle
of the First Analogy and, second, how the principle of the Second Analogy makes
possible a Kantian account of continuants comparable to Hume's. The chapter
concludes with an examination of the Third Analogy centered on Kant's distinction
between community as commercium and as communio (local community).
3 When I use 'material' to contrast with 'formal', I am throughout referring not to corporeal
reality but utilizing Kant's distinction between the form and the matter of a representation.
14 Kant's Anatomy of the Intelligent Mind
Transcendental Philosophy
Psychologized
uous with, indeed the culmination of, the psychologization of philosophy initiated
by Locke, advanced by Berkeley, and developed to its empirical outrance by Hume.
The consensus view takes Kant to have investigated the essential structures of
thought and knowledge independently of consciousness and to have then argued
that conscious thought2 and knowledge3 would be impossible if our minds were
not somehow innately "hardwired" to operate with these structures. Since this
is just to say that neither thought nor knowledge are, in their essential nature,
phenomena of consciousness at all, their psychological investigation, on this
conception, becomes at best a distraction from logic and epistemology and at
worst a source of psychologistic fallacies. T he remedy imputed to Kant is to
restrict consciousness to the role of explanandum in all philosophical and sci
entific inquiry, thereby preventing it from ever entering into the explanans. His
"hard problem" of consciousness thus becomes effectively indistinguishable
from today's hard problem of explaining qualia,4 which some regard as the only
conscious phenomena left unexplained after science, linguistics, and logic have
had their say.
T he most obvious and intractable obstacle facing proponents of the standard
view is textual: Kant's writings on the understanding are replete with psycho
logical notions and explanations, including psychological explanations of the
structures of thought and knowledge themselves. Of these, none poses a more
formidable challenge than apperception, the "supreme principle" and "highest
point" (Al17n, B134n, B135, and B136) of Kant's theory of the understanding
in the Transcendental Analytic of the Critique of Pure Reason. "Apperception"
was Kant's preferred term for self-consciousness, pure most particularly, but
also empirical, and concerns the unity of representations in one consciousness
effected by certain kinds of synthesis (synthetic unity of apperception) as well
as the identity of that consciousness in respect to those representations (the
representation I think (die Vorstellung ich denke), or analytic unity of appercep
tion). Its importance, particularly for Kant's theory of the understanding, is
indicated by its status as the supreme principle of synthetic a priori judgments
(A158/B197), the possibility of which was the problem the Critique of Pure
Reason was written to address (B19-28, Al0-14, PFM 276-80). In Kant's view,
4 A common way of characterizing qualia is in terms of "what it's like," e.g., what it's like
to taste a banana, see yellow, be annoyed by a fly, feel embarrassed, or understand a sentence
in Urdu.
Transcendental Philosophy Psychologized 19
any single textual question can be said to rise above all others in the field of
Kant interpretation, it is this: does the established view of his theory of the
understanding get it right when it takes him to have regarded the categories as
necessary for apperception?
The view stands or falls with the response. For if Kant can be shown to have held
that the categories presuppose and are not presupposed by apperception, that would
suffice to prove not only that the categories are not necessary for apperception but
also that these fundamental structures of knowledge are themselves grounded on
self-consciousness, the most quintessentially psychological of all representations.
Over the course of this book, I will argue that there is not just one but several
regards in which the categories presuppose unity of apperception, and that in the
only one where apperception does presuppose the categories, it equally presupposes
all of the other regards in which it is presupposed by the categories. What precisely
this amounts to will become clear over the course of Parts II through V. Suffice it to
say that I will present multiple convergent lines of evidence that together leave little
doubt that Kant did indeed psychologistically explicate the categories in terms of
precategorial apperception. In view of the centrality of apperception to his theory
of the understanding, this means that should my thesis be conceded, the conse
quence is nothing less than a fundamental transformation in the way the theory
needs to be understood.
The standard view of Kant's theory of the understanding is certainly correct
to the extent that the theory banishes empirical psychology from logic and epis
temology. But on the reading presented here, this is only because Kant traced the
essential structures of thought and knowledge not to empirical but to pure con
sciousness. For example, we will see that, for Kant, in the absence of a certain
a priori knowable structure of sensible consciousness, an essential ingredient for
propositional thought is lacking, and not only do our minds become psychologi
cally incapable of propositional thought but also such thought itself becomes logi
cally impossible. That ingredient is logical universality, which Kant explained as
the special contribution of a pure consciousness of self (analytic unity appercep
tion, Bl33--4) that depends in turn on the presence, "ahead of all thought" (Bl32),
of a certain structure in pure sensible consciousness (synthetic unity of appercep
tion, Bl33-4). Since logical universality is the fundamental notion of proposi
tional thought and its logic, explicating it in terms of self-consciousness renders
it psychological and, by extension, psychologizes logic itself. Whether explaining
logic in terms of consciousness rather than consciousness in terms of logic (as
generally supposed) commits a psychologistic fallacy is not my present concern
(see Chapter 9 and especially UU ch 2-E-l for discussion). I mention it here only
because the derivation of a basic logical structure of thought from a structure of
pure self-consciousness proves to be the rule, not the exception, in Kant's theory of
the understanding. Indeed, not only are essential features of thought and knowl
edge explained in much the same manner, but so too are essential features of space,
20 Kant's Anatomy of the Intelligent Mind
time, physical realities, nature and its laws, as well as all things mathematical. And
if that implies that they are incapable of existing independently of pure conscious
ness, and pure apperception in particular, it should come as no surprise that, in
these regards, Kant professed himself to be "of one confession" (PFM 374) with
Berkeley and idealists generally.
However great the challenge of explaining empirical consciousness may be,
that posed by pure consciousness presented Kant with a much harder "hard
problem." The former is a temporal phenomenon systematically correlated to
a physical human nervous system embedded, along with the rest of the body,
in the physical world (a view I think Kant would endorse: ch 18). But if the
physical world owes its most fundamental and essential structures to the struc
tures of pure consciousness, nothing from that world-not space, not time, not
matter, not the laws and constants of nature-can enter into the explanation of
such consciousness (e.g. "the subject in which the representation of time origi
nally has its ground cannot determine its own existence in time by means of
that representation," B422). Nor can it be explained by what exists prior to and
independently of the structures of pure consciousness, Kant's thing in itself. The
thing in itself is unknowable a priori by means of these structures. It is equally
unknowable a posteriori because nothing can be known empirically except by
means of these same structures, which are constitutive not only of empirical
thought and knowledge themselves but of all objects cognizable by their means.
And finally, it cannot be known independently of pure consciousness, since apart
from the structures only such consciousness can provide, knowledge of any kind,
indeed thought itself, becomes impossible. Pure consciousness is therefore left
with nothing to explain it but itself.
How Kant addressed the challenge of the hard problem of pure conscious
ness is my primary focus in this book. I am not so much concerned to assess the
truth or contemporary interest of the theory he developed to meet it as to get
the details right and present the evidence he adduced for it in the most effective
manner I can, leaving assessment of the philosophy that emerges to the reader.
In other words, my objectives in this work are purely interpretive: to exhibit
Kant's theory of the understanding as an a priori psychologism continuous
in all essentials with the a posteriori psychologism of Hume with which UU
concluded. To that end, the first chapter provides a preliminary textual tour of
Kant's a priori psychology and the various uses to which he put it in his theory
of the understanding. The purpose is to acquaint the reader with the lineaments
of Kantian a priori psychology so that even those who judge the very notion
unintelligible will at least be obliged to concede that there is a strong prima facie
case that Kant himself did not agree with that judgment. In Chapter 2, I use
Kant's own assessments of the philosophers who came before him to make the
case that his greatest debt by far was owed to Hume, for bequeathing him not
only the problem his theory of the understanding was crafted to solve but also
Transcendental Philosophy Psychologized 21
the psychologistic method he used to solve it. For while I do not want to under
state Kant's debts to others, especially to Locke and Berkeley (also considered in
Chapter 2), his own testimony seems to leave no doubt that the path to compre
hending his transcendental analytic of the understanding leads through Hume's
psychologistic science of human nature.5
5 Because I apply the term "psychological" to Kant's theory of the understanding, my inter
pretation is often grouped with those of Patricia Kitcher, Andrew Brook, Lorne Falkenstein,
and others who have applied the term (nonpejoratively) to the theory. This is understandable but
unfortunate since there is nothing beyond the word that unites us. Far from reinterpreting Kant's
theory as an a priori extension of Hume's psychologism in opposition to the anti-psychological
construal of the Kantian a priori that has long been established in Kant interpretation, their
interpretations are in basic agreement with it. These "psychological" readings of Kant are fully
consonant with the notion that none of the essential contents of the fundamental elements of
knowledge (space, time, and the categories) derive from the psychological operations responsible
for forming representations of them, as well as the notion that the fundamental structures of
thought (logical universality and forms of judgments) can be understood completely indepen
dently of the psychological operations that implement them in cognition and thinking generally.
Nor, above all, do they challenge the notion that the categories are necessary conditions for
apperception. Accordingly, for these interpreters, just as for those who minimize the role of psy
chology in Kant's theory of the understanding, it is psychology that dances to the tune laid down
by epistemology and logic, not vice versa, resulting in the kind of nativist construal that rein
forces the supposition that Kant regarded features of Euclidean space and Newtonian physics
as innate to the human mind. Emphasizing the relevance to Kant's theory of the understanding
to contemporary cognitive psychology is, indeed, all to the good, but it can only aid our under
standing of Kant if it first gets the role of psychology in his theory of the understanding righ t .
This page intentionally left blank
{ 1 }
Newcomers to the Critique of Pure Reason are understandably perplexed when they
turn to the experts and find them generally in agreement that Kantian transcendental
philosophy is not the psychological philosophy it seems to be on almost every page.
The Transcendental Analytic, in particular, is saturated with talk of consciousness,
the senses and their manifold of data, imagination and its synthesis of these data,
the unity of consciousness and self-consciousness that the understanding brings to
these syntheses, and so on. The objects discussed there are, in classic psychologi
cal theory of ideas/idealist fashion, equated with representations (by contrast with
mind-independent things in themselves), where "representation" (Vorstellung), in its
most general sense (A319-20/B376-7), was expressly chosen by Kant to signify what
the term "idea" had been used to signify in previous early modern theories of ideas
(so as to restore 'idea' to something like its original Platonic sense). Although Kant
did, to be sure, repeatedly insist that his transcendental philosophy should on no
account be confounded with psychology, this was never qua psychology per se, but
always qua empirical psychology (UU ch 2-E-1). For the certainty-universal, neces
sary validity-that empirical psychology necessarily lacks, he claimed to be present
in transcendental philosophy insofar as its psychology is concerned exclusively with
a priori representations of sense, imagination, and understanding. Thus, it should
come as no surprise that Kant saw fit to classify transcendental philosophy and
empirical psychology as distinct species-differentiated as a priori and a posteriori
of a single genus of investigation: the inquiry into the origin of mental representa
tions originated by Locke, developed by Berkeley, and perfected by Hume (ch 2).
Claims by Kant commentators notwithstanding, Kant himself left no doubt that
the origin of representations in the psyche and the psychological operations respon
sible for producing them are the true, ineliminable subject matter at the heart of his
critical, or transcendental, philosophy.1 It is an "inquiry into the first sources of our
1 The upshot of Kant's various characterizations of critical philosophy and its relation to tran
scendental philosophy and metaphysics is that critique is part of transcendental philosophy (B25-
6, A l 3-14/B27-8, AA 18 § 5667), and transcendental philosophy a propadeutic to metaphysics
(Axx-xxi, Bxxii, Bxxxvi A840-l/B868-9, A850/B878, MFPNS 469-70 and 478, Progress 272-3,
AA 17 § 4446, AA 18 §§ 4889, 4895, 5603). Since the differences between critical and transcen
dental philosophy do not matter in the present context, I shall here treat them as interchangeable.
24 Kant's Anatomy of the Intelligent Mind
2 Transcendental logic is a species of special logic, i.e., logics concerned with the content of
discursive (i.e., universal) representations. This Kant contrasted with general logic, which ignores
the content of representation and is completely indifferent to the origin of that content ("[it] has
nothing to do with the possibility of cognition in regard to its content but merely with its form
insofar as it is a discursive cognition, whereas investigation of the origin of a priori cognition
of objects must be left exclusively to transcendental philosophy") (Discovery 244; also AS0-7/
B74-82). See Part III and ch 14-E for discussion and analysis.
3 Kant applied the term "pure" to representations into which no empirical content enters and
no empirical data are presupposed. This will be explained more fully below and in Chapter 3.
The Psychological A Priori 25
contents it confers on them, and (iii) that the categories of traditional metaphysics
substance, cause and effect, necessity, et al.-must be explicated through them. Its result
is therefore quintessentially psychological: proof that the categories "are not of empiri
cal origin but rather have their seat and source a priori in pure understanding" (CPrR
141; also Bl59).
That Kantian transcendental philosophy is best understood as a priori psychol
ogy is never clearer than in the following notation dating from the late 1770s:
Transcendental philosophy does not consider objects but rather the human
mind according to the sources in it from which a priori cognition stems, and
its bounds. Thus, pure mathematics is not part of transcendental philosophy,
whereas the sources in the mind from which it springs indeed are. (AA 18 § 4873)4
Similarly: "In transcendental science, everything must be derived from the subject"
(AA 18 § 5008 [late 1770s]; also 5002), and "In transcendental cognitions only one
single proof is possible, namely from the concept of the subject" (AA 18 § 5003 [late
1770s]; also Al48-9/Bl88). Since Hume could have said the very same thing from
his standpoint in empirical psychology, what remains to distinguish the focus on
the psyche (sensibility, understanding) in Kant's transcendental philosophy from
Humean psychology, other than its being a priori?
Of course, many Kant scholars will object that being a priori makes all the dif
ference in the world, and, in particular, suffices to block any interpretation that
would assimilate Kantian transcendental philosophy to psychology, particularly the
Humean variety. This resistance, it seems to me, stems mainly from the tendency to
extend contemporary notions of the a priori and psychological philosophy to Kant.
Today the "a priori" tends to be equated with a focus on concepts and their role
in our representation to the exclusion of their causes, particularly the psychologi
cal factors involved in their formation, while notions of psychological philosophy
are typically derived from the introspectionist variety favored by philosophers like
William James in the nineteenth century and sense-datumists like C. D. Broad in the
twentieth. Yet, it is precisely when one goes back to Kant and the eighteenth-century
context in which he philosophized that such assumptions cease to hold. In particu
lar, the psychological philosophies of Locke Berkeley and Hume, as documented in
UU, do not place great trust in introspection and depend on it little, if at all. This
is due primarily to the fact that they almost invariably focus on ideas that, for all
cognitive and conative intents and purposes, are continuously employed-causation,
identity, self, substance, space, time, etc.-and on principles relied on implicitly in
4 This passage would appear to be an earlier version of the contrast Kant drew between
philosophy and mathematics late in the Critique of Pure Reason: "From what source the con
cepts of space and time (as the only original quanta) they [mathematicians] occupy themselves
with may originate (herkommen) is, for them, a matter of complete indifference, and it seems to
them equally pointless to inquire into the origin of the pure concepts of the understanding and
therewith the scope of their validity, rather than simply to put them to use" (A725/B753). This
passage, in turn, anticipates PFM 258-9, a crucial text that will be examined in the next chapter.
See also UU ch 2-E-l.
26 Kant's Anatomy of the Intelligent Mind
virtually every act of reasoning-that every beginning of existence has a cause, the
causal uniformity of nature, that actions are exercises of powers, that powers depend
on the internal constitution of complex individuals, and so on. Where what is at
issue are ideas and principles so essential to "all our thoughts and actions ...that
upon their removal human nature must immediately perish and go to ruin" (THN
225/148; also EHU VIII/I if 5), the specific contents present to consciousness at any
given moment, their accessibility to the attentive gaze, and the possibility of describ
ing them in purely observational terms are matters of all but complete indifference.
Hume, for example, was quite explicit about the drawbacks of introspective
methods:
[S]hould I endeavour to clear up in the same manner [as in the natural sciences]
any doubt .. .by placing myself in the same case with that which I consider,
'tis evident this reflection and premeditation would so disturb the operation
of my natural principles, as must render it impossible to form any just conclu
sion from the pha:nomenon. We must therefore glean up our experiments in
this science from a cautious observation of human life, and take them as they
appear in the common course of the world. (THN xix/6)
Most of the mental operations that Hume and his empiricist predecessors
deemed essential to consciousness of a world of enduring, causally interacting,
law-governed bodies and minds occur so rapidly and are so obscured by habit that
nothing could be more futile than to attempt to identify or analyze them by focus
ing the spotlight of attention on them (e.g., ECHU II/ix/§§ 9-10, V if 1590, and
EHU IV/I if 8; cf. MFPNS 471). Indeed, viewed retrospectively through a Kantian
prism, the principal warrant for deeming the necessary conditions for specific kinds
of representations posited by their psychologies empirical at all is Kant's insistence
that associable perceptions cannot be apprehended in sensibility unless preceded by
unqualifiedly pure intuitions (chs 3--4).
The apparent gap between early modern British empiricist psychology and
Kant's transcendentalism narrows still further as soon it is acknowledged that the
latter, far from being a purely abstract affair, is very much concerned with concrete
data of consciousness. The primary datum of Kant's transcendental philosophy
is the a priori self-consciousness signified by the I think (das ich denke). Far from
being anything abstract, distilled from analysis, the I think
Nor is this immediate "feeling of an existence without the least concept" (PFM
334n) the only a priori datum without which Kantian transcendental philosophy
would be impossible. Any action of the mind, whether of understanding or imagi
nation, is perceived in the same indeterminate manner, termed spontaneity (B157-
87n; chs 4, 8-G, and 13). Similarly, receptivity, the complement of spontaneity, is an
equally indeterminate awareness of the affection of our (outer or inner) senses in
sensation prior to and independently of the apprehension of these sensations as a
manifold in perception (for which the spontaneity of imagination is requisite: A97,
A99-100, and A120). These and other data of consciousness-intuitions, concepts,
and cognitions-count as a priori for much the same reason the ideas and prin
ciples considered in the psychologies of Locke, Berkeley, and Hume very nearly
do: they represent what is common to all representations, or at least to all of a
certain kind (perception, thought, cognition), irrespective of all that differentiates
them in regard to sensation (perceptible qualities, relations, and actuality) (ch 3-A).
This is not to deny that the difference between Kantian transcendental representa
tions and principles and those of British empiricist psychology is more than simply
one of degree (of generality and indifference to the contents of sense perception).
It is merely to say that being a priori is not a reason for supposing that the former
are any less essentially psychological than the latter.
One of the principal features that distinguishes Kant's transcendental philoso
phy from previous psychological philosophies is that it brackets out all that is spe
cifically human in the human psyche. This is not to say that pre-Kantians restricted
the scope of the fundamental ideas and principles of human psychology to our
species. Hume, for example, found sufficiently strong analogies between human and
animal psychology to warrant extending his account of the former to the latter
(EHU IX and THN I/iii/§ 16, 11/i/§ 12, II/ii/§ 12, and 448/286). Yet, for the same
reason, he would be prevented from widening its scope to include the psychology
of intelligent creatures whose cognitive and conative operations bore little or no
analogy with our own. Similarly terrestriocentric is Berkeley's account of visual
spatial perception: customary association between nonspatial visible and properly
spatial tangible objects may confer a spatial meaning on the former in beings whose
senses and spatial-recognitive psychology are constituted similarly to our own, but
no pretense can be made to extend the same account to all creatures endowed with
senses and capable of spatial recognition (UU ch 14). And Locke took pains to
emphasize how narrow the sphere of human understanding is, with the implication
that our insight not only into the nature of things but also into the nature of other
species of sensate intelligence is hardly less limited than that of a worm shut up in
a drawer (UU ch 7-C and 9-C).
By contrast, Kant's transcendental philosophy concerns representations and
principles that he held to be true of all intelligent minds, howsoever different oth
erwise, just so long as their intellection is conditioned by the presence of sensation
(sensibly conditioned, non-intuitive understanding). Its foundation is the quintes
sentially subjective-psychological datum, the pure consciousness that I think. The
28 Kant's Anatomy of the Intelligent Mind
because it precedes and makes possible all other concepts of objects (mathematical
and empirical); and a principle of transcendental judgment is necessary because
it precedes and makes possible experience itself (BS). Being a necessary condition
not just for some but for absolutely all representations of the same kind, empirical
included, was, in Kant's view, sufficient to prove a representation to be a priori.
And since "kind," for him, is entirely a function of the representations' faculty of
origin-sensibility, understanding, or judgment-it follows that the necessity of a
priori representations can reflect nothing other than the nature and workings of the
psyche originally responsible for producing them.6
Of course, Hume's associationist psychology likewise purports to show that our
minds contribute contents necessary to all objective representations of a particular
kind. Without association, the impressions of the sense and the ideas resembling
them in thought would be nothing but a kaleidoscopic flux of fleeting existents,
none related to any other. Only insofar as associative imagination supplies the
impression originals of our ideas of necessary connection, identity, complex indi
viduality, space, time, etc., can pre-associative impressions and ideas take on the
conceptual and cognitive value of causes, persons, bodies, etc. (UU chs 16-18). And
only insofar as the associative imagination conceives and believes beyond any pos
sibility of doubting such principles as the general causal maxim and the uniformity
of nature is any instance of empirical reasoning possible at all (UU chs 19-20). The
question therefore arises whether these impressions and principles should not be
regarded as a priori in Kant's sense, given that in Humean psychology they count as
universal and necessary conditions of representations of certain kinds.
For Kant, representations are unqualifiedly a priori if and only if they "take
place independently not of this or that experience but of absolutely all experience"
(B2-3). No previous psychology, including Hume's, could meet Kant's standard of
a consciousness pure of everything empirical. This is because all took for granted
that there is nothing one must already be conscious of before one can apprehend
the pre-associative flux of appearances (UU ch 3, KMM ch 4). Consequently, how
ever universal and necessary some of the contents contributed by the imagination
may be in pre-Kantian psychologies, they are nevertheless invariably preceded by
empirically apprehended appearances and so do not "take place independently ... of
absolutely all experience." This changed with pure intuitions of sensibility. Pure
intuition, according to Kant, is a universal and necessary condition of the flux
of appearances apprehended in inner sense itself, and so of empirical sensibility
as such (chs 3-5). The same strict sense of necessity and unrestricted universality
extends to the pure concepts of the understanding insofar as the manifold of pure
intuition and its pure synthesis in imagination enters directly into their content
("pure synthesis, universally represented, gives the pure concept of the understand
ing," A78/B104). Incorporating into their content the pure intuition that precedes
and makes possible empirical intuition extends the scope of these concepts beyond
associations (reproductive imagination) all the way down to the appearances them
selves as apprehended in empirical intuition (A125 and B164; ch 13). Accordingly,
the principles of pure understanding formed from pure concepts of the understand
ing by the transcendental faculty of judgment ( Urteilskraft) count as universally
and necessarily valid in a sense undreamed of in any previous psychology (only
things in themselves are excluded)-a sense indicative, first and foremost, of the
origin of their contents:
7 Kantian metaphysical cognitions are therefore similar to the accounts of knowledge devel
oped by Locke (UU ch 9-A) and Hume (UU ch 18-A).
Critique of Pure Reason provides, and was intended to provide, a
8 This is not to deny that the
foundation for science. I am only denying the premise of the many interpretations that suppose it
to be inseparably bound up with Newtonian doctrines or those of any other science (mathematics
included: see chs 6, 15, and 17; also UU chs 4-A and -B).
{ 2}
ment of the sensible world points forward to the Transcendental Aesthetic of the
Critique of Pure Reason, with which its theses agree in all essential respects: (i) there
are pure intuitions that precede and make possible all sense perceptions; (ii) pure
intuitions are not just formed by sensibility but derive their content from the con
stitution of this faculty; (iii) as bound up by content with the constitution of sensi
bility, pure intuitions can apply only to sensible appearances, not to things as they
are independently of their appearance to our senses; (iv) that space and time are
the pure intuitions we find in ourselves does not mean that all sensibly conditioned
intelligent beings are similarly constituted; and (v) space and time are "not some
thing objective and real, nor ...a substance, nor an accident, nor a relation," but
"rather subjective and ideal" (ID 403).
The treatment of the intelligible world in the Dissertation, by contrast, points
backward to the dogmatic rationalism that Kant was soon to eschew. This is evident
1 "I agree to your suggestion of a collection and publication of my small writings; still I would
prefer it if nothing earlier than 1770 were included so that it began with my Dissertation: de
mundi sensibilis et intelligibilis" (to Tieftrunk, October 13, 1797). In this book, I shall follow
Kant's lead and ignore the comparatively unremarkable body of philosophical work antedating
the Dissertation.
34 Kant's Anatomy of the Intelligent Mind
above all in the assumptions underlying his account of the origin of metaphysical
concepts:
[T]he concepts met with in metaphysics are not to be sought in the senses
but in the very nature of the pure understanding, and that not as innate
concepts but as concepts abstracted from the laws inherent in the mind
(by attending to its actions on the occasion of experience), and therefore
as acquired concepts. To this genus belong possibility, existence, necessity,
substance, cause, etc., together with their opposites or correlates. Such
concepts never enter into any sensory representations as parts, and thus
they could not be abstracted from such a representation in any way at all.
(ID 395).
Metaphysical concepts are here supposed to derive all their content from the
understanding and none from the pure or empirical intuition of the senses. This
Kant thought sufficient to justify ascribing unrestricted universal and necessary
validity to them so that their scope includes not only sensible appearances (via
their determination of pure intuitions, much as in the transcendental synthesis
speciosa of imagination of the Critique) but also things in general and in them
selves. With a validity that transcends any possible experience, he thus felt secure
in employing them in the traditional metaphysical business of determining the
ultimate nature of reality-a reality "objective" in the sense of being prior to
and independent of the constitution and operations of our sensibly conditioned
minds, which themselves are to be viewed as simply one among many species of
finite substances, mental and nonmental alike, all subject to the same laws of this
unqualifiedly objective, metaphysical reality.
Two years later, Kant confessed to a correspondent that he found the account
of the origin of metaphysical concepts in the Dissertation defective:
One cannot, without feeling a certain pain, behold how entirely every one of
his opponents-Reid, Oswald, Beattie, and lastly Priestly-missed the point
of his problem ...It was not the question whether the concept of cause is
correct, serviceable, and in respect of the whole of our cognition of nature
indispensable, for this Hume never doubted. Rather, it was the question
whether the concept is thought through reason a priori and in this way has
an inner truth independent of all experience and therefore also a far more
extended employment, not limited to objects of experience: here is where
Hume expected a breakthrough (Eroffnung). It was indeed only the issue of
2 That the problem had to do with the origin of the pure concepts of the understanding is con
firmed in a letter written shortly after the publication of the Critique: "In the year 1770, I could
distinguish the sensibility of our cognition quite well from the intellectual through determinate
boundary markers (Grenzzeichen), of which the principal features (jumbled together with many
that I would no longer acknowledge) [were] in the Dissertation .... But then the origin of the intel
lectual in our cognition created unforeseen difficulty" (to Johann Bernoulli, November 16, 1781).
36 Kant's Anatomy of the Intelligent Mind
the origin of this concept, not its indispensability in use: if only the former
were ascertained, then everything concerning the conditions of its use and
the sphere in which it can be valid would already of itself have been given.
(PFM 258-9)
For Kant, just as for Hume, the issue is not the normative correctness or indispens
ability of concepts like cause and effect to objective representation. On the con
trary, it is precisely their paramount importance to such representation that makes
it imperative for philosophers "to investigate the origin of the pure concepts of
understanding and in so doing determine the extent of their validity" (A725/B753).
Since experience is incapable of supplying us with pure concepts of objects, the only
options are to trace them to an origin in pure understanding or to renounce them
altogether. In the latter case, one would then either have to embrace something
along the lines of Humean associationist surrogates or concede that there is noth
ing present a priori in the psyche to underwrite "cause and effect," "substance,"
necessity," etc., "and that these words are absolutely without any meaning, when
employed either in philosophical reasonings, or common life" (EHU VII/ii if 26;
UU ch 2). Thus, for Kant, the only way to meet the challenge of Humean skepti
cism was to adopt Hume's own psychological method to show that (1) the under
standing is indeed a source of pure concepts, (2) pure concepts so originating can
have a priori validity as pure concepts of objects, and (3) their objective validity is
limited to sensible objects.
Before we can understand Kant's Humean method properly, however, we first
need to review the reasoning that spawned it, beginning with Locke's sensibilism
and propositional subjectivism and proceeding to Berkeley's anti-abstractionist
separability principle.
Locke was the first great philosopher to dedicate his magnum opus to the topic
of human understanding. He did so because he "suspected that we began at the
wrong end . ..[when] we let loose our Thoughts into the vast Ocean of Being, as
if all that boundless Extent, were the natural, and undoubted Possession of our
Understandings" (ECHU I/i/§7). Instead, we must first prove title to its possession.
To do this, however, it is not enough simply to define our concepts with a clarity
and distinctness comparable to that demanded by mathematicians.We must instead
look at the ideas present in our understandings behind our discourse.For all effort
at definition will be in vain if there is nothing-no object (notion, idea)-present
to our consciousness capable of underwriting our verbal formulx (UU ch 8). The
philosopher's preeminent order of business must therefore be to "inquire into the
original of those ideas, notions, or whatever else you please to call them, which a
man observes, and is conscious to himself he has in his mind; and the ways whereby
Kant's Debt to British Empiricism 37
the understanding comes to be furnished with them" (ECHU Intro. § 3). That is,
only a psychological investigation of the sources of the materials available to our
understandings can reveal any limitations these origins may impose on their scope
of application-limits on what is and is not possible to think or know by their
means that inevitably remain hidden on any strictly definitional, language-bound
approach to concepts. Thus did Locke set out to employ the psychology of ide
ational origins to penetrate the cognitive and ontological mirages of language and
safeguard himself against any temptation to claim knowledge where clear title to
it is lacking.
Kant acknowledged his debt to Locke for this insight: "Locke's excellence was
that since he did not cognize intellectual ideas (intellectualia) as innate (connata), he
sought their origin" (AA 18 § 4894 [late 1770s]). Why does the rejection of innate
ideas (i.e., innate contents of thought, as distinct from innate faculties, propensi
ties, etc.) confer such great importance on the psychological question of origins?
Innatism is one explanation, Malebranchian illuminationism another, for what
I termed in UU the intellectualist thesis that ideas preexist their presence to con
sciousness in inner or outer sense. If ideational contents are in any sense present in
us prior to or independently of the sensory and reflexive operations whereby they
are brought to consciousness in (clear or obscure, distinct or confused) perceptions,
then these operations are incapable of contributing anything essential to those ideas.
For this reason, intellectualist philosophers relegated to the margins of the theory
of the understanding the psychological question concerning the origin of the per
ceptions in which ideas first present themselves to consciousness. Indeed, intellec
tualist reliance on analytical methods geared to producing precise definitions as the
primary means to determine the contents of ideas stemmed in no small part from
their utility in eliminating any and all sensory-psychological overlay imposed on
ideas by the perceiving subject, thus leaving in their wake purely intellectual per
ceptions confined to the contents proper to the idea in question (clear and distinct
perceptions of ideas).
Since for Locke, by contrast, "having Ideas, and Perception [are] the same thing"
(ECHU II/i/§9), there are, and can be, no ideas prior to or independently of their
presence to consciousness in sensation or refiexion.3 Once perceived, ideas may then
be considered in new ways, "each of which Considerations is a new Idea" (xiv/§14),
by being combined, separated, related, abstracted, or otherwise acted upon by the
understanding. Accordingly, to explicate an idea, rather than merely enumerating
its constituents in definitions, we must first endeavor to discover which of the var
ied and often complex (or multiply complex) psychological operations went into
3 As in UU, I use the spelling "reflexion" to refer not to pondering, reckoning, analysis, or
the like (i.e., operations that Kant denominated "discursive"), but to what Locke, Berkeley, and
Hume considered to be an inner sense analogous to the outer senses: just as the latter provide
visual, olfactory, etc., sensations, the former is our capacity for "that notice which the Mind takes
of its own Operations, and the manner of them, by reason whereof, there come to be Ideas of
these Operations in the Understanding" (ECHU 11/i/§4).
38 Kant's Anatomy of the Intelligent Mind
the idea's formation by the mind; for only by identifying and comprehending these
operations-their nature, workings, and inherent limitations-can we hope to
demarcate the proper cognitive sphere of application of the ideas they generate.
To be sure, even while singling out Locke's theory of the understanding as the
one "most similar" to his own because it "concerns every employment of the under
standing in general" (letter to Garve, August 7, 1783), Kant rejected his predeces
sor's strictly empirical approach (A86-7/Bll8-19), charging him with "the error
of taking the occasion for obtaining these concepts, namely experience, for their
source" (AA 18 4866 [late 1770s]). But this was not, as often supposed, because
Kant, as transcendental philosopher, was in a different line of business than Locke.
On the contrary, "[t]he origin of concepts from sensible representations or the
understanding belongs to psychology and transcendental philosophy" (AA 15 §
1697 [1770s]). Kant as transcendental philosopher was just as committed as empiri
cal psychologists to the Lockean method of tracing representations to their sources
in the faculties of the mind as the optimal way of ascertaining their contents. He
was also committed to Locke's sensibilist principle that ideas (the objects pres
ent to consciousness) are nothing prior to or independently of the senses and the
psychological operations the mind performs on their data (UU ch 1). He diverged
from Locke and other sensibilist psychologists only when it came to the question
of whether the senses are a source uniquely of empirical data. For Kant was saying
nothing less than the truth when he claimed to be the first to entertain the possibil
ity that "the senses also may be supposed to intuit a priori" (PFM 375n).
This certainly includes Locke, who consequently overlooked the possibility of
an a priori yet still sensibilist theory of the understanding. For where the senses
supply a manifold of a priori intuition, the potential is created for the understand
ing to act upon that manifold to generate further representations completely a pri
ori, thereby opening the way to a theory of pure understanding, or transcendental
logic (A55-6/B79-80 and A76-7/Bl02), to complement and underpin the empirical
logic of Locke and others:
priori manifolds of outer and inner sensible intuition. Instead of a strictly empirical
consciousness of the unity of the relation of the manifold in complex ideas, Kant pos
its a pure consciousness of the unity of the pure synthesis of the pure manifold: first,
in the form of original apperception (pure self-consciousness) (chs 3-5 and 9); second,
the universal representation in pure concepts of the understanding (categories) of the
unity that results when pure synthesis in imagination of the pure manifold of sense is
determined conformably to the logical functions of judgment (transcendental synthe
sis intellectualis: chs 11 and 14); and third, the universal representation of the a priori
Kant's emphasis on proving the legitimacy of our title to employ the catego
ries does indeed express a normative concern. However, one must be careful not
to confound the normative implications of a theory of the understanding with the
normative character of the theory itself. For example, when Hume traced the origin
of the idea of necessary connection to customary association-a psychological pro
ceeding no one would equate with normative epistemology-he nevertheless drew
normative conclusions from it regarding cause and effect (e.g., "Rules by which
to judge of causes and effects," title, THN I/iii/§15), including a constraint on the
scope of application of causal concepts similar to Kant's restriction of the catego
ries to objects of possible experience:
Such a discovery ...that this [causal] connexion, tie, or energy lies merely in
ourselves, and ...is acquir'd by custom ...not only cuts off all hope of ever
attaining satisfaction, but even prevents our very wishes; since it appears, that
when we say we desire to know the ultimate and operating principle, as some
thing, which resides in the external object, we either contradict ourselves, or
talk without a meaning. (THN 266-7/173)
Against this, the sensibilist holds that the idea of a triangle contains only those
contents that the judging subject actually thinks in it. To be sure, some of these may
be thought in it only confusedly so that a risk of confusing the idea with other ideas
arises, while others may be so obscure as to escape attentive discernment altogether
(because, e.g., of the rapidity of the actions of the mind and/or the concealing
influence of custom: ECHU II/ix/§10, discussed in UU ch 6; see also UU chs 14,
17, and 19). Nevertheless, for the sensibilist, the equality of the sum of its angles to
two right angles is no more a constituent of the idea of a triangle than the equal
ity of the number of its sides to the cube root of 81 or the number of crew on the
1:33p.m. Wellington to Tauranga flight on October 28, 2026. For none are thought
in the idea of a figure formed from three intersecting straight lines in a plane-not
confusedly, not even obscurely. And, in general, propositional thought (mental, not
verbal: UU ch 8-D) has to be understood in terms of subjects and predicates that
contain only so much content as the judging subject actually thinks in them.
It was Locke's espousal of a subjectivist conception of propositional thought
that led him to regard the necessary relations of quantity dealt with in mathematics
as instructive rather than merely explicative:
[W]e can know the Truth, and so may be certain in Propositions, which affirm
something of another, which is a necessary consequence of its precise com
plex Idea, but not contained in it. As that the external Angle of all Triangles,
is bigger than either of the opposite internal Angles; which relation of the out
ward Angle, to either of the opposite internal Angles, making no part of the
complex Idea, signified by the name Triangle, this is a real Truth, and conveys
with it instructive real Knowledge. (ECHU IV/viii/§8)
Since for the sensibilist subjectivist the idea of a triangle contains nothing-has no
reality, meaning, content-other than what appears immediately to consciousness
in perception, geometrical demonstration does not merely clarify and make dis
tinct necessary quantitative relations already implicit in this idea, but rather forges
those relations itself T hese relations are intrinsically bound up with the actions
of the mind in comparing and considering the ideas, and are nothing prior to or
independently of the sequence of propositions (i.e. comparisons of ideas) whereby
we become sensible (perceive) that the ideas are necessarily conjoined in them. Or,
as Hume put the same point: "the necessity, which makes two times two equal to
four, or three angles of a triangle equal to two right ones, lies only in the act of the
understanding, by which we consider and compare these ideas" (THN 166/112), not
in the ideas themselves (UU ch 18-A). T hus, for Locke and his successors, even if
nothing contained in our ideas remained confused or obscure, the simplest math
ematical equations would still remain objectively undetermined unless and until we
subjectively perform the comparisons requisite to forge them.
C onfirmation of Kant's sensibilist credentials can be found in the unmistakably
subjectivist conception of propositional thought evident in his distinction between
analytic and synthetic judgments:
Kant's Debt to British Empiricism 43
beside the point, for in analytic propositions the question is only whether I actu
ally think the predicate in the representation of the subject" (Al64/B205; also
Progress 322-3).
To be sure, the distinction between analytic and synthetic judgments turns on
the content of the concepts related in it (PFM 266): a categorical (subject-predicate)
judgment is analytic just in case the predicate-that is, the predicated concept
together with the way it belongs to the subject (extension as a property of body, a as
a lesser portion of the total quantity represented by a+ b, etc.)-is contained in the
content of the subject as a constituent mark (Merkmal) and so is logically identical
with it; otherwise, it is synthetic. But what determines which contents do and do
not belong to a given subject or predicate? For a sensibilist, the criterion is strictly
subjective: it depends simply and solely on what the judging subject actually thinks
in them. It will not matter in the least if the subject and predicate are connected by
a necessary objective relation such as quantitative equality. What is decisive for the
distinction between analytic and synthetic is not whether they are identical in the
object but subjectively in the consciousness of the judger. We should therefore not
be surprised to find Kant, the committed sensibilist, distinguishing subjective from
objective identities in precisely this way:
I can form a concept of one and the same quantity by means of multifarious
modes of composition and separation (though each, as well as addition and
subtraction, is a synthesis), which is objectively identical (as in every equa
tion). Subjectively, however, according to the mode of composition I think
in order to arrive at (gelangen um) the concept, it is very different, so that
the judgment must certainly go beyond the concept which I have from the
synthesis, because it sets a different mode of composition (which is simpler
44 Kant's Anatomy of the Intelligent Mind
and better suited to the construction) in place of the first, which neverthe
less always determines the object in precisely the same way. (letter to Schultz
November 25, 1788)5
The conclusion that Kant was as much a subjectivist regarding judgment as Locke
remains impossible for many commentators to accept. Yet no one can dispute that
Kant classified mathematical judgments as synthetic a priori, arithmetical and alge
braic not excepted. Since this means that, for Kant, a judgment like 2 + 2 = 97 is
noncontradictory and in good logical order, it is therefore incumbent on anyone who
rejects a subjectivist interpretation to propose an alternative. For if the contents actu
ally thought by the judger are irrelevant, yet 2 + 2 = 97 is to count as false with the
same a priori necessity as 2 + 2 = 4 counts as true, how can either judgment still be
5 Perhaps the simplest, most direct proof of Kant's allegiance to Lockean subjectivism is his
insistence that "analysis, which seems to be the opposite of synthesis, yet always presupposes it"
(B130; also A77/Bl03 and ch 13-A). Because synthesis "is an act of self-activity," it "is the only
thing among all representations that cannot be given through objects but only by the subject
itself" (Bl30). But since this is just to say that "we can represent nothing as combined in the
object without having ourselves previously combined it," analysis can never reveal anything in a
representation not put there by the synthesizing subject.
6 Kant credited Locke with coming the closest to recognizing that there are synthetic a priori
judgments that need to be distinguished from analytic judgments (PFM 270 and AA 18 § 5066).
Curiously, this was due not to Locke's thesis that mathematics is instructive but to his recognition
that certain truths regarding the coexistence of things admit of being known (in Locke's strong
sense of "know"-a species of insight that yields certainty beyond that of even the most highly
probable judgment).
Kant's Debt to British Empiricism 45
may therefore seem to be compelling sensibilist reasons for regarding Kant's distinc
tion between analytic and synthetic a priori as having at most verbal validity.
The response to this objection goes straight to the heart of Kant's most important
innovation in sensibilist theory of the understanding: establishing the autonomy of
understanding as a distinct faculty over and above imagination. Although Locke did
not discuss the difference between imagination and understanding (as against Hume
who effectively denied it: UU ch 2), in practice, he treated the former as a faculty by
which the mind separates and combines data of the (internal and/or external) senses.
Its combinatory power is crucial to the formation of complex ideas of substances,
modes, and relations. However, before its combinatory act can result in an actual
complex idea, there must first be a consciousness of the combination in such a way
as to consider it one idea (UU ch 5-C). This considering-as-one-or, in Kantian par
lance, consciousness of the unity of the synthesis of a manifold-does seem different
enough from the kind of operation Locke tended to ascribe to imagination to warrant
attributing it to a distinct faculty. But even if that is granted, Lockean understanding
can represent no complex (manifold) that the senses or imagination have not previ
ously (synthetically) set up for it to consider as one. Since the negations of necessary
judgments like 2 + 2 = 4 are representable neither in sense nor imagination, we can be
quite certain that Locke and his successors would have regarded them as intellectually
unthinkable as well: impossible, self-contradictory, and so analytically false.
For Kant, by contrast, the understanding does more than simply represent syn
thesized manifolds of sense (apprehension) or imagination (reproduction) as one,
complex representation. In the first place, understanding, in its guise as analytic
unity of apperception, is constitutive of logical universality itself, without which
concepts, and so too logic in general (transcendental included), would be impos
sible (B 131 and B33--4n; ch 9). This is just to say that, for Kant, there is a species of
mental representation unknown in sensibilist empiricism but distinctive to under
standing: genuine universals rather than merely abstracted ideas (Locke: UU ch
8-B and 10-A) or habitual resemblance relations (Berkeley and Hume: UU ch 10-B
and 18-B). Second, the understanding complements logical universality with logi
cal functions of judgments for forming distinct concepts into judgments (without
which concepts have no representational use: A68/B93; ch 10). Third, and most
important where cognitions and their objects are concerned, the understanding
supplies pure concepts of the understanding that define a synthesis intellectualis
(chs 11-C and 14-B). These three uniquely intellectual operations permit the under
standing to frame mathematically impossible (i.e. objectively necessarily false) judg
ments such as "2 + 2 = 97" which, as synthetic, are free of contradiction and so
logically in good order. For, according to Kant, mathematics requires, in addition
to these intellectual operations, the synthesis speciosa of the imagination to define
its judgments in relation to the pure intuitions of sensibility. Only then, and then
only in this one sensible regard, can the logically well-ordered (subjectively possible)
synthetic proposition "2 + 2 = 97" be determined extra-logically as necessarily false
(objectively impossible). Thus, on Kant's version of sensibilism, not only does the
46 Kant's Anatomy of the Intelligent Mind
understanding count as an autonomous faculty over and above sense and imagina
tion, it also confers the ability to think the unimaginable.
To be sure, the three intellectual operations aforementioned suffice only to
enable the understanding to think synthetic a priori judgments, not to cognize them
as true or false. If we are to become conscious of the necessity of the relation of
nonsubjectively identical concepts, something = X that is not itself a concept or
judgment must be added in virtue of which relation can be so cognized (AS, A9/
B13). In empirical judgments, that Xis experience, while in synthetic a priori judg
ments, the Xis the a priori manifold of sense and its pure synthesis in imagination
(A155/B194; chs 3-6). Accordingly, the a priori necessary falsehood of judgments
like 2 + 2 = 97 is grounded not on understanding alone, and so not on logic as
such (pure general logic), but on sense and imagination as well. This, however, does
not prevent such judgments from being not only sayable in language but thinkable
by the understanding, and so no more logically contradictory in themselves (ana
lytically false) than any other judgment that turns out to be false when the X of
synthetic judgments is factored in: "the Alaskan winter is balmier than Hawaii's,"
"Water is NaCl," or "the shortest distance between two points in a Euclidean plane
is a zigzag." Thus, by vindicating understanding as an autonomous faculty along
side sense and imagination and restricting logical laws, properly so called, to the
first, considered in isolation from its relation to other faculties, Kant was able to
distinguish analytic from synthetic a priori judgments in a manner fully consonant
with sensibilism.
Most of those who embrace this conclusion are too circumspect to claim that
Kant himself had a clear recognition that the natural world is perfectly transpar
ent to the judging subject. Nevertheless, wittingly or not, Kant is deemed to have
produced a philosophy that is perfectly capable of dispensing with theory-of-ideas
style psychology and being recast, with only minor amendments, in the image of
one of the holistic models of thought prevalent in analytic ontology and epistemol
ogy today. Among other things, this means that Kant's iterated insistence that pure
sensible intuition is indispensable to the solution to the problem of the possibility
of synthetic a priori judgment7 must be regarded as mistaken, a consequence of his
inability to completely break free of early modern philosophical forms even after he
himself had denuded them of substance. And is that not after all to be expected of
a great visionary, groping toward something so completely new that the rest of the
philosophical world would take more than a century to catch up and overtake him?
Yet however striking one finds the similarities between Kantian synthetic a pri
ori and contemporary analytic holistic conceptions of thought, Kant's manifest
reliance on theory-of-ideas psychology, both in framing the problem of the syn
thetic a priori and in endeavoring to solve it, should make one wary of such claims.
The notion that no significant loss of content results when "translating" Kant's
transcendental philosophy from the language of mental representation to the lan
guage of contemporary normative epistemology seems spurious on its face: if true,
would it not equally follow that Quine, Strawson, Kripke, and so forth could be
translated back into theory-of-ideas terms with just as little loss? Anything can be
made to seem like anything else if abstraction from differences is pressed sufficiently
far. And if the task is not historical appropriation for contemporary purposes but
merely to interpret-to discover and, so far as possible, recapture what Kant, the
historical personage, thought and why-it stretches credulity, in my view, to sup
pose that the psychological language in which he couched his analysis of the pos
sibility of synthetic a priori cognition is really so dispensable as some profess (ch 1).
In this connection, it is useful once again to recall Locke. As remarked previ
ously, Locke raised the theory of the understanding to philosophical preeminence
by replacing the method of clarifying concepts by definition with the psychological
method of tracing them to their origin as ideas in the mind. A definition, however
precise, is ontologically worthless from a Lockean perspective if there is no idea
present to our consciousness capable of underwriting it. This is because the only
objects we are ever acquainted with are our own ideas, and the only way any other
reality can come within the purview of human knowledge is through its causal and
resemblance relations to ideas (UU chs 5, 7, and 9). Consequently, no concept,
7 That Kant deemed pure intuition and the ideality of space and time essential to the solu
tion to the problem of how synthetic a priori judgments are possible is quite clear: e.g., A39/B56,
B73, PFM 377, Discovery 245, letter to Tieftrunk December 11, 1797, AA 18 §§ 5552, 5637, 5927,
6353, 6355, and AA 22 4. Synthetic a priori practical judgments and synthetic a priori judgments
in matters of taste do, however, require a different X than sensible intuition (CPrR 31, 45-6, and
CJ 288-9).
48 Kant's Anatomy of the Intelligent Mind
intellection is essentially bound up with the unity of the synthesis of the manifold
in one consciousness a priori (synthetic unity of apperception), the I think it makes
possible (analytic unity of apperception), and so too with consciousness and its
sensible and intellectual forms of unity. Thus, it was only if, and insofar as, a sign is
underpinned not only by the rules of language but by psychology as well that it can
be accorded the properly intellectual and objective worth required to earn it a place
in Kant's "science of ontology as immanent thinking" (letter to Beck, 20.01.1792;
ch 18-C and conclusion).
The route from sensibilism to propositional subjectivism is, as we have seen,
both short and direct. What should now also be clear is that necessary relations
expressed by verbal propositions count as synthetic a priori for Kant only to the
extent they are mediated by mental propositions Uudgments) made possible by
pure intuitions and the other pure representations such intuitions make possible
(analytic unity of apperception, transcendental syntheses, and schemata) . To insist
that such formulre are referentially transparent, and so to bracket out their psycho
logical mediation, is ipso facto to negate any claim to their being synthetic a priori.
Thus, just as with Locke, a psychological grounding, while by no means necessary
for semantic worth or normative indispensability (for which convention suffices),
is indispensable to according any verbal expression, whether ordinary, scientific,
or mathematical, the least objective validity and ontological worth in Kant's tran
scendental philosophy. 8
Once it is recognized that Kant's insistence that only "what carries with it a cor
respondent in intuition to our words" applies to synthetic a priori no less than to
synthetic a posteriori judgments, the irrelevance of holistic conceptions of thought
to the Kantian synthetic a priori becomes manifest . For how can an a priori objec
tive identity between subjectively non-identical concepts possibly be forged merely
by nesting one's judgments among other a priori judgments that, in the absence of a
psychological grounding in pure sensibility, are equally incapable of effecting such
identities? From Kant's point of view, no less than from Locke's or Hume's, any
strictly analytical approach that refuses to venture beyond the confines of thought,
indeed of language, and so ignores psychological origins, can never amount to any
thing but mere wordplay. Only sensible intuition can effect objective identities; and
where the objective identity is a priori, and so necessary and universal, there simply is
no other means than pure sensible intuition capable of doing this.
8 While synthetic a priori practical cognition does not involve pure intuition (see note 7 above),
Kant made quite clear that it presupposes other such cognitions that do. For example, there could
be no positive idea of freedom if not for the negative idea of freedom that emerges from Kant's
transcendental idealist solution to the Third Antinomy, which itself is grounded on the psycho
logical doctrines of the Transcendental Aesthetic and Analytic. This presupposition is perhaps
less evident in the case of the synthetic a priori judgments of taste analyzed in the third Critique,
but it is still operative, both for the general solution to the problem of the possibility of synthetic
a priori judgments of any kind laid out in the first Critique and for the specific relation of pure
understanding to the schematizing productive imagination that Kant adapted to explain how
judgments of taste can still be a priori even though they are grounded in a species of pleasure.
50 Kant's Anatomy of the Intelligent Mind
9 Another example of how an analytic philosophical approach to Kant risks leading one
astray is the widespread practice of construing Kant's distinction between analytic and synthetic
judgment as if it were essentially the same as the distinction targeted by Quine in "Two Dogmas
of Empiricism" and discussed in undergraduate philosophy of language courses ever since. The
practice is justifiable to the extent that Kant's initial presentation of the distinction via explicative
and ampliative judgment allows one to construe analytic judgment purely logically in terms of
concept identity and synthetic judgments equally logically but purely negatively as non-analytic
(i.e. where identity of the predicate with the subject concept is lacking). However, the question
Kant's Debt to British Empiricism 51
The question of Kant's debt to Berkeley poses particular difficulty because the his
torical record is a virtual blank . Kant's reading knowledge of English may have
been minimal or even inexistent, and translations of Berkeley's works into lan
guages we can be certain that Kant knew were few. Even accounts of Berkeley's
thought in other writers were rare during Kant's pre-Critical formative period. He
may have had some acquaintance with Berkeley's views via his many Anglophone
friends in Konigsberg, orally or in unpublished translations. Still, there is no men
tion of Berkeley in the first edition of the Critique of Pure Reason (1781). Indeed,
before the Prolegomena, the only recorded reference (in student lecture notes) is
to Sirus, a treatise on the medicinal effects of tar water from which the German
translators had excised nearly all of its (quite substantial) philosophical content.
There is, to be sure, indirect evidence in the 1781 Critique of Pure Reason that Kant
was familiar with Berkeley's views: the dogmatic idealism that denies the existence
of matter and would later be expressly identified with Berkeley is discussed in the
Fourth Paralogism (A350); Kant's notion that in proving something to be the case
is whether construing them this way is adequate for the interpretation of Kant, since doing so
obliges one to discount, or altogether ignore, the essential subjectivity of analytic and synthetic
judgments (or indeed of judgment itself, which even in the analytic case involves the synthesis
of one consciousness with another: B13ln). After all, Kant did not decide to designate them as
he did for no reason (e.g. "that something outside the given concept must still be added as sub
strate which makes it possible to go beyond the concept with my predicates is clearly indicated
by the expression 'synthesis,' and consequently investigation is directed toward the possibility of
a synthesis of representations for the sake of cognition in general, which must soon turn to intu
ition, pure intuition must be acknowledged as the unavoidable condition for a priori cogni
while
tion," Discovery 244-5). In subsequent parts, we will see that Kant's notion of synthetic judgment
rests on a theory of synthesis, logical (judgment) and aesthetic (imagination) alike, the ultimate
ground of which is the quintessentially subjective psychological notion of the unity of the mani
fold of sense effected through the imagination's productive synthesis, or original synthetic unity
of apperception. Similarly, we shall find that his notion of analytic judgment derives from a con
ception of analysis as the process whereby given representations are transformed into concepts
(A76/B102 and A78/Bl04), which in turn has its foundation in the no less quintessentially subjec
tive psychological representation I think (die Vorstellung ich denke), or original analytic unity of
apperception (Bl33-4). Crucially, this latter unity presupposes the former analytically (Bl35 and
B138; also Al13 and Al29-30), with the consequence that the transcendental notion of the origi
nal synthetic unity of apperception necessarily underlies all employments of the understanding
regardless of the contents involved in its judgments (aesthetic or intellectual, theoretical or prac
tical, etc.) or, as with pure general logic, even if abstraction is made from all content (Bl33-4n;
Part III). This should not surprise us. After all, Kant made no secret of the Lockean sensibilist
subjectivist character of the analytic-synthetic distinction itself, or of his purpose in using it to
frame a problem deriving directly from Hume's treatment of the necessary relation (synthesis)
of distinct perceptions (representations), i.e. the question Kant devised his critical philosophy
to answer (deeming the analytic-synthetic distinction to be otherwise useless: PFM 270). Thus,
although it is technically not incorrect to construe the distinction between analytic and synthetic
judgment in terms of concept identity, it is clearly inadequate if one's purpose is to understand
either the problem Kant addressed in the Critique of Pure Reason or his sensibilist psychological
approach to solving it.
52 Kant's Anatomy of the Intelligent Mind
of all triangles we omit from our conception of the figure all the features indifferent
to the theorem (A 713--4/B741-2) is strongly reminiscent of Berkeley's analysis (UU
ch 10); and Kant may well have had Berkeley uppermost in mind when he claimed
to have refuted empiricist arguments against infinite divisibility with his principle
of Axioms of Intuition (Al65-6/B206). This, together with Kant's evident familiar
ity with Berkeley's anti-abstractionism (PFM 289) and idealism (293 and 374-5) in
the Prolegomena, makes it difficult not to believe that Kant was knowledgeable of
Berkeley's principal positions by 1781.10
What we can be sure of is that sometime before 1781, Kant came to appreciate
the significance of Berkeley's preeminent principle, the separability principle (UU
ch 10-B), even if only via Hume's application of it in his analysis of causation.11 The
separability principle transformed sensibilism from the form in which Locke left it
by setting strict limits to the abstractive powers of the mind. Locke had supposed
that our minds equip us to distinguish features in the objects (ideas) present to us
that cannot exist independently of one another in sense perception or imagination.
For example, even though the visible shape of a triangle and its particular light and
color are one and indistinguishable in perception, he still regarded them as distinct
ideas on the ground that we can immediately discriminate one from the other by
abstraction. According to Berkeley, however, the power to abstract "extends only to
the conceiving separately such objects, as it is possible may really exist or be actually
perceived asunder," and "does not extend beyond the possibility of real existence or
perception" (PHK I§ 5; also Intr.§ 10). Thus, whereas the trunk of a human body
and its limbs, or the rose and its scent, qualify as distinct ideas because either can be
met with in perception in the absence of the other, the visible shape and color of a
triangle are not distinct because visible shape straightaway becomes invisible in the
absence of light and color.
How are we nonetheless able to represent and reason about visible shape without
taking light and color into consideration at all? According to Berkeley, such dis
tinctions consist not in distinctions between ideas but in different significative uses
of one and the same idea. Significative uses derive from the various ways in which
ideas are found to resemble one another: their sensible quality (the red of a tomato
10 Alciphron was translated into German in 1737 and the Dialogues in 1756 (French trans
lations date from 1734 and 1750 respectively). Although Italian translations of the Principles
and the Theory of Vision existed, there is no evidence that Kant knew of them or could have
read them even if he did. Kant would not have needed a translation of the Latin De Motu, but
there is no evidence pre-1781 that he knew this or any other of Berkeley's writings besidesSirus.
Eugen Stabler, in George Berkeleys Auffassung und Wirkung in der Deutschen Philosophie bis
Hegel (Zeulenroda: Bernhard Sporn, 1935), makes the case that Kant had no direct knowledge
of Berkeley's writings, arguing that he depended instead entirely on a single secondary source.
However, Stabler's analysis is premised on a reading of Kant's treatment of space and time that
seems to me questionable.
ll
Hume employed the separability principle countless times in his own philosophizing. He
acknowledged his debt to Berkeley in the most admiring terms in the Treatise, at the outset of
"Of abstract ideas" (THN I/i/§7), and reaffirmed it at EHU XII/i 'l]'l] 15-16.
Kant's Debt to British Empiricism 53
and the red of a bell pepper), the manner in which they are received (red and blue
resemble not in quality but in being both sensed by the eyes), as causes (manmade
things), as effects (emetics), and in the circumstances of their occurrence (e.g., the
visual, tactual, olfactory, etc., sensible qualities of an apple resemble in respect of
always being found to occur together) (V § 128 and VV § 39). Thus, the difference
between visible shape and visible color is founded on two of the manifold external
relations of resemblance that visual ideas have to other visual ideas rather than on
anything internally distinguishable in any single, isolated visual idea, prior to or
independently of its comparison to others. These and other resemblances can then
be put to significative use, especially (though by no means exclusively) for purposes
of general signification: one and the same visual idea can be used to denote all red
things indifferently, all triangular things indifferently, all colored things indiffer
ently, all polygons indifferently, all shiny things indifferently, and so on, according
to which resemblance relation the imagination is operating with at a given moment
(UU ch 10).
The philosophically crucial implication of this analysis for Berkeley and his
anti-abstractionist successors is that a semantic difference of denotations does not
imply a real difference of ideas. For example, I can apply my idea of existence to
houses, trees, the earth, my body, or any other sensible object without supposing,
even tacitly, that these objects are perceived by a mind. This semantic independence
does not, however, suffice to infer that the ontological independence of these objects
from minds is even so much as intelligible, much less a real possibility. For that
to be the case, according to Berkeley, there would have to be a real distinction of
ideas under the separability principle since ideas are the only objects ever present
to our minds, the only realities with which we are ever acquainted. And however
useful or even indispensable it may be to distinguish the existence of sensible things
from their being perceived for the purposes of ordinary and scientific thought and
action, the failure of this distinction to pass muster with the separability principle
is, in Berkeley's view, sufficient to render ontological materialism not just false but
unintelligible (ch 8-A and UU ch 11).
Because Berkeley shared Locke's sensibilism12-the thesis that the contents of
human understanding all originate in sensation and reflexion or the actions the
understanding performs upon them-nearly all of the differences between their
theories of the understanding can be traced to Berkeley's anti-abstractionist sep
arability principle. The most important of these is the principle that the contents
of thought have ontological (rather than merely significative) application only to
such objects (ideas, notions)13 as they may originally have been acquired from: those
12
For a convmcmg defense of this occasionally contested point, see Michael Ayers,
"Was Berkeley an Empiricist or a Rationalist?," in The Cambridge Companion to Berkeley
(Cambridge: Cambridge University Press, 2005).
13
In Berkeley's preferred terminology, "ideas" are sensations and their images in memory and
imagination, and "notions" are internally perceived data including mental actions, passions, rela
tions, and their reproductions in memory and thought.
54 Kant's Anatomy of the Intelligent Mind
Like Locke and Berkeley, Hume was a sensibilist theorist of human understand
ing who held that the best way to clarify ideas is not analysis with an eye to for
mulating precise definitions but by a psychological investigation into their origins
in sense and imagination (THN 157/106 and EHU VII/i if 4). Hume concentrated
on cause and effect because of its unique importance in the economy of human
understanding, contending that, in the absence of this idea, "Inference and reason
ing concerning the operations of nature would, from that moment, be at an end;
and the memory and senses remain the only canals, by which the knowledge of any
real existence could possibly have access to the mind" (EHU VIII/i if 5; also THN
73-4/52-3). Attributing its cognitive and conative preeminence to its principal con
stituent, the idea of necessary connection, Hume made the impression original of
this idea the focus of his theory of the understanding.
To say that X is necessarily connected to Y as cause to effect is to say that Y can
not exist if X does not and must (cannot not) exist if X does (HTC ch 5-E and UU
ch 19-E). This means that, strictly speaking, enduring things are not causes because
a cause cannot begin to exist without its effect immediately coming into existence
as well. In such cases, the causality of the cause is instead a new state or activity of
an enduring thing that, as soon it obtains or operates, enables the thing to immedi
ately produce the necessarily connected effect (THN 174/117 #8). Since there may
be different states or actions of the same or different things that have this effect, the
cause is, more precisely, what all of these states or actions share in common (THN
174/117 #5). Similarly, where there are different states or actions of the same or
different things that have the same cause, the effect is what they have in common
(THN 174/117 #6). Thus, "[t]he same cause always produces the same effect, and
the same effect never arises but from the same cause" (THN 173/117 #4).
The insight into the idea of necessary connection that led Kant to pronounce
Hume's analysis the most important event in the history of metaphysics (PFM
257) is that the relation concerned in it is restricted to existents distinct in the sense
specified by the Berkeleian separability principle (CPrR 51). We cannot, for exam
ple, conceive mountains and valleys to be related as cause and effect because they
are not separable in the relevant sense: valleys cannot be conceived to exist in the
absence of mountains and vice versa. Fire and smoke, by contrast, are relatable
as cause and effect because each can be conceived to exist in the absence of the
other. But there lies the rub. If to conceive cause and effect as distinct is to be
able to conceive the existence of the one to be possible in the absence of the other,
whereas to conceive them as necessarily connected is for the existence of the effect
to be impossible unless its cause exists (and for the non-existence of the effect to be
impossible given the existence of its cause), distinctness and necessary connection
are evidently logically incompatible. Forced to choose, Hume opted to supplant the
genuine but impossible concept of cause and effect-an objectively real necessary
56 Kant's Anatomy of the Intelligent Mind
When we run over libraries, persuaded of these principles, what havoc must
we make? If we take in our hand any volume; of divinity or school metaphys
ics, for instance; let us ask, Does it contain any abstract reasoning concerning
quantity or number? No. Does it contain any experimental reasoning con
cerning matter of fact and existence? No. Commit it then to the flames: for it
can contain nothing but sophistry and illusion. (EHU XII/iii if 30)
that it is also possible in reality that something may begin to exist without a cause.
(3) And, because all that is requisite to show that the general causal principle is nei
ther intuitively nor demonstrative certain is evidence that their separation in reality
implies no contradiction or absurdity, Hume boldly concluded that "'tis impossible
to demonstrate the necessity of a cause."
Hume was nevertheless careful to emphasize that his conclusion does not imply
that the general causal maxim is false, doubtful, or even dubitable. Quite the con
trary, in endeavoring to show "why a cause is always necessary" (title of THN I/
iii/§3), his purpose was not to challenge the certainty of the principle but only the
consensus presumption regarding the nature of that certainty. Hume broke new
ground with his insistence that its certainty is not intuitive-not a purely intel
lectual affair of the relations of ideas alone-but rather something else entirely,
involving the sensate, feeling part of our minds no less essentially than the con
ceiving part. Being a committed empiricist in his sensibilism, however, he saw no
alternative but to trace its certainty to "observation and experience" (82/58), and
ultimately to the very same source to which he traced ideas of necessary connec
tion: customary association (HTC ch 5 and UU ch 19-D). Thus, the challenge
Hume bequeathed to his successors was twofold: a conceptual challenge to dem
onstrate that genuine, objective concepts of necessary relations between distinct
existents are possible and an epistemological challenge to demonstrate that objects
of experience are subject not merely inductively but necessarily and universally to
these concepts.
In dedicating his magnum opus to Hume's skeptical challenge ("my work in the
Critique of Pure Reason was occasioned by Hume's skeptical doctrine," PFM 52;
also PFM Preface), Kant's first step was to determine whether its scope extended
further than Hume realized. One area in which he thought it did is that of necessary
relations between distinct quantitative determinations, considered independently of
matters of fact and real existence. Relations of equality and inequality are a case
in point: although abstract and indifferent to matters of fact and real existence,
they are distinct by the separability principle criterion essential to Hume's skeptical
reasoning regarding causal connections (CPrR 50-2). For example, in equating the
addition of five to seven with twelve, I conjoin in a necessary relation quantita
tive determinations that are as conceptually distinct as fire and smoke: I can think
twelve without conceiving the addition of five to seven just as easily as I can think
either without conceiving the difference between 3 1 and 19, the square root of 144,
or the cube root of 1728. All are subjectively non-identical in the sense described
in Section A, making any judgment equating them synthetic. Since this is just to
say that the necessity of the relation cannot lie in the concepts whereby the objects
concerned in such relations are thought, whence does it derive? In Kant's view, the
same reasons that, in the case of causal understanding, led Hume to treat objective
necessity as an illusion and set the subjective necessity of customary association in
its stead apply with equal force to mathematical judgments (UU ch 2).
58 Kant's Anatomy of the Intelligent Mind
14 Since Hume's consideration of necessary connections between the distinct in the Enquiry
was largely confined to causes and effects, Kant's belief that he was extending Hume's doubt to
other metaphysical concepts tends to confirm the common assumption that he had little or no
familiarity with the Treatise. Apart from brief extracts, the latter was not translated into German
until the 1790s, and Kant may well have been unaware of the range and systematic character of
Hume's earlier book during the gestation period of his critical philosophy. Certainly, at no point
did Kant betray any awareness that Hume pointed out inconsistencies in the standard definitions
of substance, body, self, and other concepts, or that he sought to explicate them just as he did
causal connections: by tracing them to associationist origins (UU ch 17 and 18). Kant's knowl
edge of Hume's skepticism may therefore have been confined almost entirely to the first Enquiry
and the Dialogues concerning Natural Religion.
Kant's Debt to British Empiricism 59
This complete ...solution of the Humean problem thus rescues the a priori
origin of the pure concepts of the understanding as well as the validity of the
universal principles of nature as laws of the understanding, yet in such a way
as to limit their use to experience, because their possibility depends solely
on the relation of the understanding to experience, but with a completely
reversed kind of connection that never occurred to Hume: they are not
derived from experience but rather experience is derived from them. (PFM
313; also B127, A112-13, A122, and A782-3/B810-11)
Although Kant's debt for what he called "Hume's problem" (the possibility of syn
thetic a priori judgments) is well recognized, the fact that his solution to it could
with equal justice be termed "Hume's solution" is not . The full extent of Kant's
debt only becomes clear when one recalls that it was Hume who convinced him
that normative indispensability is irrelevant to the question of the possibility and
objective validity of the categories, and that instead, everything hinges on their ori
gin as representations in the mind: "if only that were ascertained, then everything
concerning the conditions of [their] use and the sphere in which [they] can be valid
would already of itself have been given" (PFM 259). The reasoning, more precisely,
is this: if the psychological account of the origin of such concepts as representations
Kant's Debt to British Empiricism 61
in the mind reveals that the mental operations involved in their formation at the same
time contribute essential elements of their content, these concepts must be acknowl
edged to be no less bound up with consciousness than pleasure and pain are, both in
content and scope of application (their sphere being restricted to the purview of a
suitably constituted psyche). As in UU, I term this method of addressing questions
concerning the possibility and objective validity of concepts at the heart of age-old
philosophical disputes psychologism. As exemplified by Kant's accounts of the origins
of space and time as pure intuitions of sensibility (Part 11), the categories as pure
concepts of the understanding (Part III), experience and its objects as expressions of
transcendental apperception (Part IV), and nature as the objective unity of appercep
tion constituted by the material side of appearances (Part V), Kant was able both to
secure the objective meaning and validity of these concepts from Hume's conceptual
and epistemological skepticism and, at the same time, to show that they have, and can
have, no application except to appearances that in themselves are nothing more than
representations generated by the mind (transcendental idealism).
Though psychologism can be said to have originated in Locke's account of seeing
in three dimensions (UU ch 6), and it was Berkeley who first extended the method to
such primary philosophical concepts as existence, causality, substance, and sense-divide
transcending objects (UU chs 11, 13, and 14), Kant was certainly correct to credit it
to Hume. A simple comparison of Berkeley's account of causality and substance with
Hume's makes this clear. Berkeley traced these notions to an origin in the internally
perceived operations of one's own thinking: the mind is a substance that "supports"
its ideas and notions by perceiving them and, in its capacities of imagination and voli
tion, "causes" them to exist or enter into certain relations (UU ch 13). However, their
origin in internal perception places no limitation on their scope. For, according to
Berkeley, a person's mind is merely one efficacious spiritual substance among others
in a wider sphere of reality; and while finite spirits are incapable of causal interaction,
this is due solely to their lack of the power to do so, not because their interaction is
unintelligible. God (infinite spirit) makes good this want by, for example, causing me
to hear the words God causes you to hear and feel yourself uttering when you speak
to me. For Berkeley, then, the notions of substance and causation are not limited in
scope to the ideas and notions of the mind that forms them. Quite the contrary, they
apply reflexively to that mind itself as well as (potentially) to infinitely many others,
and their scope is in no way restricted by the operations concerned in their formation.
Indeed, since Berkeley did not preclude the possibility of beings other than spirits
(3D232), there is nothing in his psychologism to prevent the scope of these notions
from embracing non-spirits as well in a single, mind-independent causal nexus cen
tered on the deity.15
15 The only restriction Berkeley imposed is that we lack the conceptual means to conceive any
such "third nature" as either a substance or as causally efficacious because our originals of these
notions are spirits and their actions, and the separability principle prevents us from abstracting
them from their originals (3D239-40).
62 Kant's Anatomy of the Intelligent Mind
extending to the farthest reaches of space and time-exists only in and for associa
tive imagination (108/75 and 662/416--7).
Kant's psychologism is of precisely the same nature as Hume's except that the
contents indispensable to conceiving substances, causal relations, space, time, exis
tence, necessity, number, and more are contributed not by empirical but pure imagi
nation and understanding (which contributes the most subjective content of all, the
I think explicated as original apperception). The scope of these representations is
consequently limited to the imaginative and intellectual consciousness of the indi
vidual isolated self-conscious psyche. They have no application to the psyche that
images and conceives them (any more than the dreamer can be a figment of his own
dreaming), nor a fortiori to psyches generally (minds in themselves), and still less to
mind-independent reality (things in themselves).
Psychologistic explication functions like the traditional method of definition
to the extent that it licenses one to substitute the explicating concept for the expli
cated one. The crucial difference is that the former, in distinguishing the components
of a concept according to their faculty origin, makes it possible to reconstruct the
concept one layer at a time: the transcendental contributions of sensibility (if any);
atop that, the transcendental contributions of imagination (if any); higher still, the
transcendental contributions of understanding (if any); and, most superficially,
the non-transcendental contributions of these faculties. Its virtue is to fend off the
introduction of free riders into concepts, preventing one from substituting one's own
empirically or mathematically derived conceptions for Kant's transcendental ones,
or substituting what one takes to be a transcendental contribution of sensibility for
what is in fact a transcendental contribution of understanding. Take for example
the transcendental philosophical concept of space. If the contents we attribute to
it include its presence in the sensory modalities of vision and touch, three dimen
sions, and/or features of Newtonian space, its psychologistic explication functions to
assign these contents to their proper faculty of origin and determine how it produces
them. In particular, it reveals, first, that insofar as the transcendental representation
of space originates in sensibility, it is not present in any sensory modality (vision,
touch, or any other) at all but rather in pure intuition. Second, it discloses that pure
intuition can account only for the manifoldness of space (the juxtaposed), but not for
the differentiation and determination of the manifold spaces within it characteristic
of objective space, which is instead a transcendental contribution of the understand
ing (chs 13-14) and its synthesis speciosa, both in its application to the manifold of
pure-formal intuition (ch 15) and its application to the manifold of empirical-material
intuition (the matter of appearances corresponding to sensation: ch 16). In so doing,
however, it also shows, third, that features of geometrical space like metric and num
ber of dimensions can only be determined intuitively by mathematical construction,
not discursively by transcendental synthesis, and so do not form part of the properly
transcendental concept of space at all. Similarly, the features of physical space (met
ric, dimensionality, etc.) must be determined by physics (Newtonian, Einsteinian,
or some other), not transcendental philosophy. To include physical or geometrical
64 Kant's Anatomy of the Intelligent Mind
[S]pace and time, including all the appearances in them, are nothing existent in
themselves and outside my representations but themselves only modes of repre
sentation, and it is patently contradictory to say that a mere mode of representa
tion also exists outside our representation. Thus the objects of the senses exist
only in experience; whereas to accord to them a self-subsistent existence apart
from or prior to experience is as much as to represent the actuality of experience
apart from or prior to experience. (PFM 341-2)
In Kant's eyes, even Hume was guilty of this absurdity (CPrR 52-3), presumably by
treating the succession of perceptions as transcendentally real and so as requiring
a transcendentally real ground of their unity in one consciousness (plunging him
straight into his quandary concerning personal identity: chs 3, 4 17, and UU ch 3).16
For many, transcendental psychologism is too high a price to pay for refuting
Humean skepticism. It is not merely the psychological idealism that inevitably
results when space, time, and everything in them-material objects as well as tem
porally enduring thinking beings (empirical selves)-are supposed to exist and be
able to exist nowhere outside the (pure) consciousness of the individual isolated
psyche that represents them. It is also that transcendental psychologism implies
that the subject in whose sensibility time has its ground cannot itself be in time and
consequently can no more intelligibly be conceived to act or be affected in or at a
time than it can be supposed to be occupy or contain space ("the subject in which
the representation of time originally has its ground cannot determine its own exis
tence in time by means of that representation," B422). Nor did Kant shy away from
the implication of this conception that if the mental life each of us inwardly intuits
as a continuous succession of representations were to be perceived by a mind with
a differently constituted sensibility or by one that intuits intellectually (without
depending on sensibility at all), "the very same determinations we now represent as
[successive] alterations would yield a cognition in which the representation of time,
and so too that of alteration, would not occur at all" (A37/B54). There are no doubt
many who simply cannot accept that any view so irreconcilable with their notions
of philosophical sense and nonsense can possibly have been held by the person who
ranks, in the words of the Guardian editorial cited at the outset of this part, as "the
undefeated heavyweight philosophy champion of the world."17 Be that as it may, it
cannot excuse us from following up on the clues provided by Kant's endorsement
and apriorist adaptation of Hume's sensibilist psychologism, even to the point of
nature; like the custom of the Romans (he might have added, the Greeks), in their saturnalia,
who make the slaves the masters. ..." (letter to John Home, September 20, 1775; see analysis in
UU ch 4-C). Words have no sense of their own but owe their sense entirely to the ideas the under
standing is able to bring to them, while ideas lack meaning since meanings must be "instituted
by men" and so can never be private. The lack of an idea is therefore not epistemic but linguistic
ignorance: one makes no judgment and so commits no error, but rather, even while speaking
meaningfully (if one is acquainted with the linguistic convention), cannot give one's utterance
the sense others give it in the privacy of their minds; it is thus lack of understanding, not lack
of knowledge.
17 Peter Strawson, for example, argued that "[i]f it is truly the case that the affecting relation
at the supersensible level gives rise to appearances in outer and inner sense, if things really do
appear to us in spatial and temporal guise, then it seems that the verb 'appear' itself must here
bear a temporal construction, that these appearings must really occur in time. The alternative
would be to say that it non-temporally appears to be the case, to the transcendental subject,
that it enjoys a series of temporally ordered states. But this is strictly unintelligible-indeed
nonsense.... In that case, the difference between Kant's idealism and that of Berkeley is not,
after all, as great as he supposed." Entity and Identity: And Other Essays (New York: Oxford
University Press, 1997) 250-51. Of course, what seems unintelligible nonsense to a philosopher
operating in a tradition that analyzes thought without regard to goings on in the psyche and
rejects the theory of ideas outright may seem quite the opposite to a philosopher operating in a
hitherto neglected variant of sensibilist theory of ideas. We should let Kant's own writings and
the internal logic of his positions determine the answer, not notions of philosophical sense and
nonsense current in our time (which to philosophers of a later time may seem as or more defec
tive than early modern notions).
66 Kant's Anatomy of the Intelligent Mind
bringing temporal succession within its remit (as Hume never did). For it could yet
turn out that in the framework of Kant's philosophy, perhaps uniquely, there is nei
ther inconsistency nor incoherence but genuine philosophical interest in the thesis
that the mind that produces time neither acts nor is affected in time.
In emphasizing Kant's debts to Locke, Berkeley, and above all Hume, I will be thought
by some to be underestimating his debts to the ancients, seventeenth-century ratio
nalists, and/or certain of his German contemporaries. Yet in the case of Hume's
influence in particular, anyone tempted to treat it as simply one among many, and
possibly not even the most important, is obliged to explain why Kant singled him
out as the one who awakened him from his dogmatic slumber (PFM 260; also A764/
B792)18 by confronting metaphysics with the greatest crisis in its history (PFM 257),
but credited neither Tetens, Mendelssohn, Crusius, Wolff, Leibniz, Plato, nor any
one else with anything remotely comparable. They must also explain why Kant said
of Hume but no one else that he came so close to anticipating what is most dis
tinctive in Kant's philosophy that he required only to recognize the syntheticity of
mathematical judgments to have been "drawn into considerations which must needs
have been similar to those which now occupy us" (PFM 273; also AA 18 § 5040; UU
ch 2).19 F inally, I would challenge anyone who would discount or disregard Kant's
own assessment of his debt to Hume to do what I once did: comb through Kant's
collected writings, assemble all of his remarks about past and present philosophers
that relate to the theory of the understanding, and then see whether you can still
rest content with the notion that Hume's influence on Kant was not fundamentally
and importantly superior to that of all others. I certainly could not after I per
formed the exercise. I found that Kant not only devoted far more space and care to
the discussion of Hume's theory of the understanding than to anyone else's, but,
18 In a letter to Garve of September 21, 1798, Kant spoke of being awakened from his dog
matic slumber by the realization of the conflict of reason with itself described in the Antinomies
chapter of the Critique of Pure Reason (see also A407/B433-4). Since the most important such
conflict concerns causality (natural necessity versus freedom), it is not difficult to see how the
1783 statement in PFM crediting Hume and the remark in the letter to Garve could both be true.
There is, however, insufficient information to say anything more precise.
19 In On A Discovery, Kant did present Leibniz in a somewhat similar light. Yet that work is
more polemical than substantive, an attempt to counter Kant's Leibnizian critic Eberhard by
portraying the critical philosophy as truer to the root tendency and spirit of Leibniz's philosophy
than Eberhard's more literal attachment to Leibnizian doctrine. But once the presentation of
Leibniz in that work is weighed against the predominantly negative treatments of Leibniz's posi
tions in Kant's principal writings, it becomes clear that Leibniz is the last philosopher Kant could
credit for awakening him from dogmatic slumber or being able to anticipate him in the creation
of critical philosophy simply by revising a single detail of his philosophy.
Kant's Debt to British Empiricism 67
most tellingly, also spoke of its key tenets in singularly approving terms. From a
philosopher who, in this sphere especially, almost invariably saw only error in the
works of predecessors and contemporaries, or, at best, truth dimly glimpsed, his
express and iterated endorsement of specific Humean doctrines surely demands
special explanation.20
The explanation offered here is the unique importance for Kant's own philos
ophy of his borrowings from Hume. First, he endorsed Hume's insight into the
special problem posed by the concept of a necessary connection between distinct
existents-so much so, in fact, that he used Hume's own principles and procedures
to expand the problem to include synthetic a priori cognition generally (necessary
relations between the distinct in general), mathematics not excepted (CPrR 50-3;
section D and UU ch 2). This, the problem to which Kant dedicated his magnum
opus, the Critique of Pure Reason, informs his critical philosophy as a whole as
well as its major adjuncts (metaphysics of nature, metaphysics of morals, etc.), and
it is of the highest significance that he saw fit to accord Hume the highest honor
he could by terming it "Hume's problem" (PFM 261 and 313). Second, and still
more important, Kant adopted Hume's method of solving the problem as his own.
The method disregards the normative indispensability to judgment and reasoning
of concepts of necessary connection such as cause and effect and focuses instead
on their origin as representations in the mind, with an eye to determining whether
the psychological processes involved in their formation also contribute essential
elements of their content, thereby limiting their scope of application to the con
scious purview of the psyche in essentially the same way that pains exist only in
being perceived (PFM 258-9; section E). This commonality of method makes it
possible to trace virtually all of the many and varied differences between their theo
ries of the understanding to a single, almost trivial, divergence at the root: Kant's
identification of a source of representational content the existence of which Hume
seems never to have suspected. And while Kant owed the discovery of pure sen
sible intuition to no one, it provided the simple, direct bridge from empirical to
pure psychology he needed to extend Hume's psychologistic method from empirical
to pure imagination (the source of a priori synthesis) and from empirical to pure
understanding (the source of a priori synthetic and analytic unity). Is it therefore
any wonder that Kant, owing both his problem and the method of its solution to
Hume, accorded a unique position to Hume, one that entirely eclipses his debt to
any other philosopher, past or contemporary?
Some Kant scholars nevertheless challenge the veracity of Kant's claim to have
owed so much to Hume. The justification most commonly cited is that Kant's
20 Although Kant's presentation of Hume in the first Critique shows how seriously he took
Hume's skeptical challenge, the emphasis there is on distancing critical from skeptical philos
ophy, not highlighting their affinities. Only when that work's considerations of Hume are set
together with those in PFM, CPrR, and Progress can one fully appreciate the nature and extent
of those affinities.
68 Kant's Anatomy of the Intelligent Mind
knowledge of Hume's philosophical corpus was too limited for it to be true. They
point to the Treatise in particular, a work not translated into German until 1790,
after the three Critiques were written.21 But even if Kant was ignorant of the
Treatise, does that mean he could not have derived his problem and the method
of solving it from Hume's Enquiry concerning Human Understanding alone, a work
we can be certain Kant knew and admired from the 1750s onward? The treatment
of the necessary connection between cause and effect in that work seems to me to
agree in every significant respect with its treatment in the Treatise. This includes
both Hume's exposition of the problems involved in conceiving necessary connec
tions between items presupposed as distinct (EHU IV-V) and his endeavor to trace
concepts of necessary connection to their origin as representations in the mind in
order to clarify their content and delimit their scope application (EHU VII). The
only significant omissions are the absence from the Enquiry of counterparts to the
discussions in the Treatise of the general causal maxim and the role of causal rela
tions in personal identity. Yet even these potential gaps in Kant's knowledge could
have been filled in by the extended citations from the Treatise in a book Kant would
almost surely have encountered in or soon after 1772: the German translation of
James Beattie's On the Nature and Immutability of Truth (1771).22 Thus, the claim
that Kant's knowledge of Hume's writings was insufficient for him to have derived
either his problem or his method of solving it directly from Hume simply does not
stand up to scrutiny.
If some interpreters persist in questioning the veracity of Kant's estimation of his
debt to Hume, it is probably due, more than anything else, to their assumption that
Kant's theory of the understanding is fundamentally anti-psychological. Certainly,
one would be hard pressed to find anyone who regards it as fundamentally psycho
logical in anything like the sense of "psychological" accorded to it in this book but
rejects Kant's estimation of his debt to Hume. Yet if one's assessment of a text such
as PFM 258-9 (cited earlier in the chapter), where Kant most explicitly and unequiv
ocally endorsed Hume's psychologistic method as the proper method for solving
the problem of the possibility of synthetic a priori judgments ("Hume's problem"),
merely reflects the assumptions one brings to it, shouldn't one at least consider
being guided by the text instead? After all, it cuts both ways: if the assumption that
21
It is also sometimes suggested that Kant obtained his high opinion of Hume secondhand,
through the writings of Tetens, Hamann, or some other German contemporary who was heav
ily influenced by the Hume of the Treatise. But then why question Kant's debt to Hume? If his
German sources channeled Hume reliably, Kant would still have been right to credit Hume as the
originator of the philosophy to which his own was so deeply indebted. If they did not, and Kant's
true debt was not to Hume but to someone else's distortions and misrepresentations of Hume,
I have y et to see any remotely convincing proof offered to that effect.
22
For a complete list of citations from the Treatise in Beattie's book, see UU 22n3 (I can
not, however, vouch for the completeness of the German translation). Kant also had access to
translations of parts of the Treatise that may have circulated in his circle of anglophone friends
in Konigsberg, at least one of which is known to have been published: a 1772 translation of the
extremely important concluding section of the first book of the Treatise.
Kant's Debt to British Empiricism 69
23 Most Kant specialists are initially attracted to the field by the standard portrait of him
as a fundamentally anti-psychological, hence anti-Humean, thinker; and, indeed, when Kant
is understood in such a manner, he is bound to seem closer to his rationalist predecessors and
near-contemporaries than to Hume and Hume's empiricist predecessors. This was certainly the
view I had at the time I published my first book on Kant. It was not until my trust in Kant grew
to where I could take it on faith that he knew best whose philosophy he was most indebted to
that I resolved to take a second, more serious look at Hume. When I did, I was disappointed to
find that Hume scholars seemed almost as bent on minimizing everything in Hume that smacked
of psychologism and Berkeleian idealism as Kant scholars are where their philosopher is con
cerned. I was thus obliged to work out Hume's theory of the understanding for myself, from the
ground up (what perceptions are, what vivacity is, what associative relation consists in, etc.), until
I comprehended the theory well enough to be able to expound it as a systematic psychologism in
HTC and UU part III. Only then could I appreciate not only how great an innovator Hume was
but also to what extent Kant was engaged in the same enterprise. My reading of Kant's discus
sions of Hume has consequently changed 180 degrees: I now regard his own characterizations of
his debt to Hume as entirely accurate and accept that it does indeed eclipse his debts to Leibniz,
Wolff, Crusius, Tetens, and every other philosopher. My hope is that readers of this volume who
take it on themselves to master the interpretation of Hume presented in HTC and UU will come
to appreciate just how strong the reasons are for regarding Hume and Kant as birds of the same
psychologistic feather.
This page intentionally left blank
{PART II}
[Tjhe concept of time does not apply to anything else, not even to
the subject of time itself
AA 23 E LXXV, p. 31
[Tjake away the thinking subject and the entire corporeal world must cease
to exist, as the latter is nothing but an appearance in the sensibility of our
subject and a mode of its representations.
A383
This page intentionally left blank
No one can doubt that Kant regarded the doctrine of pure sensible intuition as
not only wholly original (PFM 375n) but also as absolutely essential to his critical
philosophy. It is essential both to solve the problem of the possibility of synthetic a
priori judgments and to open up a place for the freedom presupposed by moral law.1
Although synthetic a priori judgments most obviously require pure sensible intu
ition because the understanding is in need of some X (A9/B13) whereby to combine
non-analytically related concepts or judgments completely a priori, its importance
goes well beyond that. Each of the elements of Kant's account of the possibility of
synthetic a priori judgments involves representations that incorporate or presuppose
pure intuition and its manifold. A pure concept of the understanding (category) is
the universal representation of the pure synthesis in imagination of this manifold
(A78-9/B104). The transcendental synthesis of imagination requisite for the catego
ries to apply to the kind of appearances our sensibility produces (spatial and tem
poral) also involves the manifold of pure sensible intuition (B150-2, B160-1). The
schematism whereby transcendental synthesis is extended to the empirical realities
given via affections of the senses (sensations, self-affections) is possible only given
the mediation of pure intuition (Al 38-9/Bl 77-8). Above all, original apperception,
the supreme principle of synthetic a priori judgment (A158/B197), consists in (1) the
representation of the unity of the manifold of pure intuition in one consciousness
a priori (synthetic unity of apperception) and (2) the representation of the identity
of this consciousness in respect of that manifold (analytic unity of apperception)
(B131-4). One reason the manifold of pure sensible intuition had to be incorpo
rated into these and other essential elements of Kant's account is that it includes not
just actual but possible intuitions and does so both universally and necessarily. Pure
intuition thus gives the notions that incorporate it the scope Kant required in order
to solve Hume's problem once he had extended it from judgments that attribute nec
essary connections to distinct existents to those concerned with necessary relations
between distinct determinations of any kind, mathematical judgments not excepted
(ch 2-D). Clearly, therefore, any attempt to understand Kant's solution that rests on
an inadequate understanding of the notion of the manifold of pure intuition at its
heart, much less attempts to discount or disregard it, cannot succeed.
Pure sensible intuition is also essential to Kant's doctrine of freedom and the
broader program of his critical philosophy to establish the bounds of intellectually
and rationally determinable reality because it underlies his demarcation of real
ity into phenomena and noumena. Appearances analytically entail something that
appears (Bxxvi-vii, A249, A251-2, A538/B566; ch 8-D). On the ground that pure
1 Regarding the synthetic a priori, see Chapter 2n7. Regarding pure intuition (transcenden
tal idealism) and its presupposition by freedom, see A536-7/B564-5, A543/B571, CPrR 6n, 95,
101-2, Progress 311, AA 18 §§ 6343-4, 6349, and 6353.
74 Kant's Anatomy of the Intelligent Mind
that pure intuitions of some kind, be their content spatial, temporal, or something
else wholly unknown to us, are an essential prerequisite for even so much as a bare
manifold of appearances to present itself to consciousness, prior to and indepen
dently of ordering and relating these appearances by means of association, concep
tion, judgment, and inference. Indeed, I find it hard to escape the impression that
the vast majority of commentators suppose Kant to have posited pure intuitions
not because sensibility has need of them but because of the higher functions of
the mind invoked in his epistemology (e.g., a presupposed X whereby subjectively
distinct concepts can be connected in synthetic a priori judgments). And such an
outlook has naturally led many to the further supposition that if only Kant had
been acquainted with the great improvements in purely conceptual techniques of
analysis made by philosophers in our own time, he would have been happy to dis
pense entirely with the figment of pure sensible intuition.
In a sense, I agree. If pure intuitions cannot pay their way as necessary condi
tions of the possibility of sensibility itself, then they are rightly regarded as excess
theoretical baggage in Kant's analysis of experience. Accordingly, in this part of
the book , I shall focus on Kant's theory of sensibility with an eye to determining
whether, and if so why, he staked his entire philosophy on the thesis that pure intu
itions are necessary to sensible representation as such, prior to and independently
of higher functions of the mind. In Chapters 3 and 4, corresponding roughly to the
metaphysical expositions of the Transcendental Aesthetic, it will emerge that Kant
deemed pure intuitions necessary for the unity of sensibility implicitly assumed by
his predecessors, sensibilist and intellectualist alike. Chapter 3 focuses on the trans
formative conception of appearances involved in this unity, where instead of being
equated with sensations, appearances have an a priori form as well a matter, and
even that matter, as corresponding to rather than identical with sensations, must be
conceived as existentially and qualitatively distinct from them. The principal thesis
of the chapter is that given sensations alone, sensibility, conceived as a unity of con
sciousness comprising the data of all of the senses (inner as well as outer), is impos
sible. In additions to sensations, sensibility requires pure intuitions; and since pure
intuitions are possible only through a synthesis of apprehension that imagination
alone is capable of supplying, and a unity of synthesis in one consciousness (apper
ception) that only the understanding is capable of providing, sensibility depends on
not only the synopsis of sense (receptivity) but also the synthesis of imagination
and unity of apperception as well (spontaneity).
Chapter 4 provides a detailed elaboration of the distinct yet complementary
contributions of pure space and pure time to the unity of sensibility. The main
thesis of the chapter is that pure space is the basis of the unity of outer sensibility
(sensations, properly so called) in beings constituted like ourselves, and time the
unity of inner sensibility (self-affections). Since the former is comprised within the
latter, the synthetic unity of the manifold made possible by pure time suffices for
the unity of sensibility presupposed by both discursive understanding (representa
tion by means of universals) and cognitive understanding (where universals are
76 Kant's Anatomy of the Intelligent Mind
combined with intuitions to yield cognitive experience). At the same time, the undif
ferentiated, indeterminate character of the manifold of pure space and time shows
them to be inadequate for cognitive experience-a want that can only be made good
by discursive understanding through pure concepts (the categories).
Chapters 5 through 8 confirm and extend these results. Chapter 5 builds a textual
case entirely from the Transcendental Deduction of the Categories for the thesis of
Chapter 4 that pure space and time produce a prediscursive, purely aesthetic expres
sion of the original synthetic unity of apperception and shows how this provided
Kant with the basis for a new understanding of the understanding. I focus on the
transcendental expositions of space and time in Chapter 6, showing first that Kant
was nothing like the Euclidean dogmatist he is commonly portrayed to be and, sec
ond, how the role he accorded to space and time as conditions of the unity of sensi
bility enabled him to explain the possibility not only of geometry but even the most
abstract, purely symbolic varieties of mathematics, among which post-Fregean
mathematical logic should probably be included. Chapter 7 utilizes the distinction
between appearances and sensations of Chapter 3 to elucidate Kant's typology of
idealism and realism, especially his claim that transcendental idealism always pairs
with empirical realism and transcendental realism with empirical idealism. Finally,
in Chapter 8, I take up the issue of things in themselves and connect it with Kant's
affirmation of a robust esse is NOT percipi realism that trumps Berkeley's esse is
percipi idealism.
{ 3 }
The Transcendental Aesthetic divides into three parts. In § 1 (by the 1787 edition
numbering), Kant distinguished the principal faculties of the mind and the differ
ent kinds of representation for which each is responsible in order to single out pure
intuition as the a priori contribution of sensibility. In§§ 2-7, he made the case for
regarding space and time as examples of pure sensible intuition and, on that basis,
treating them, together with everything occupying or containing them, as empiri
cally real but transcendentally ideal. Finally, in§ 8, he offered a number of general
observations on the science of transcendental aesthetic, concluding with the most
important so far as transcendental philosophy is concerned: "we have here, namely,
in the pure a priori intuitions space and time, two of the elements requisite for solv
ing the general problem of transcendental philosophy, how are synthetic a priori
propositions possible?" (B73)
§ 1 is like the opening of a film, zooming in from a wide-angle view of the entire
cognitive faculty to a close-up of its capacity for pure intuition. As representations
that bring us into immediate relation to objects, intuitions are that to which all thought
of objects-i.e., discursive representation or representation by means of universals
(concepts and judgments}-is directed as the sole means whereby it can acquire rela
tion to objects. Objects can only be intuited, however, insofar as representations of
them are given (received), and the capacity to receive them-receptivity-is called
sensibility insofar as they can only be given if the mind is affected (A19/B34). The
cognitive faculty that complements sensibility is the understanding: through under
standing an object given in intuition is thought, that is, its intuition is conceptualized,
and the resulting concepts employed in judgments that, under conditions specified in
the Transcendental Logic, represent what is or is not the case objectively (indepen
dently of how things stand with the judging subject).1 Thus, thought can relate to
objects in no other way than by means of the intuitions afforded by sensibility, and
only through sensibility can objects be given to the understanding.
1 Judgments with objective validity (cognitive worth) stand under the categories and are
elsewhere termed judgments of experience in contrast with judgments of perception, which give
expression merely to the psychological state of the judging subject. See chs 13, 14, and 16. 77
78 Kant's Anatomy of the Intelligent Mind
The focus is then narrowed exclusively to intuitions. Sensations are the represen
tational effects of objects on sensibility (in what follows, I shall use "sensations"2
to refer to all and only affections of the outer senses, i.e., the data of sight, touch,
hearing, smell, and taste). An intuition is empirical if it relates to its object through
sensation, and its indeterminate object-by which Kant presumably meant the
object as sensibly given but not yet determined conformably to a concept (thought,
judged}-is called appearance.
An appearance has a matter and a form . Its matter is that in it which corresponds
to sensation. Its form is what makes the matter of appearances capable of being
ordered and related. Forms of appearances are pure because "nothing belonging
to sensation is found in them;" and since the "pure form of sensible intuitions in
general in which all the manifold of appearances is intuited in certain relations must
be met with in the mind a priori," this "pure form of sensibility itself is also called
pure intuition" (A20/B34-5).
These distinctions raise a number of questions. To what precisely in sensations
does the matter of appearance correspond-sensational qualities, their mere dif
ferentiatedness (manifoldness), and/or the reality of mind-independent things in
themselves? How can a transcendental philosophy dedicated to the analysis of pure
cognition of objects have a place for sensations in it at all given that, for Kant, the
very presence of anything belonging to sensation in a representation is sufficient for
it to count as empirical? And perhaps most important of all: if nothing belonging
to sensation can be found in pure intuition, what justification is there for deeming
it sensible intuition at all?
It is natural to focus on the last of these questions because the notion that a
representation can be pure and yet still be sensible has an undeniable air of para
dox about it. Why can it not, with equal justification, be called nonsensible, or even
intellectual intuition? Certainly, if such intuition is to be equated with space and
time, there seems to be no more than the name to distinguish it from the intellec
tual intuitions (e.g., of space and time) supposed to operate in pure and applied
mathematics by intellectualists like Descartes. Why did Kant nonetheless insist
on the sensible character of pure intuition? And, above all, why should sensibility
be supposed to require such intuitions prior to and independently of all higher
operations of the mind? In this chapter, I propose to zoom in as tightly as possible
on the matter and form of appearances with a view to answering these and related
questions.
28 E LX), but this may have been due to his inclusion of data of the outer senses, via their
perception, among the data of inner sense (see A34/B50-l). In any case, to avoid confusion,
I shall reserve "sensations" for data of the outer senses and refer to data of inner sense as
"self-affections".
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 79
In addition to showing how intuitions and concepts can be pure despite being con
ditioned by sensations, this suggests how sensations themselves can be understood
to have a place in a transcendental investigation. So long as we narrow our consid
eration of them to what is true of all sensations as such, without regard to their
particular content (quality) or the particular existence indicated by the presence of
that content in perception (i.e., "sensation whereof one is conscious," A225/B272),
the purity of the investigation is in no way compromised.3 And there seem to be
two features of sensations that satisfy Kant's criteria for inclusion in transcendental
philosophy: the real existence of something indicated by the presence of any content
in sensation (in which Kant's notion of the matter of appearance is rooted) and the
manifoldness of sensations (to which his notion of the form of appearance relates).
Insofar as sensation essentially involves affection, it confers on perception a rela
tion to an affecting reality distinct from sensibility: "that which corresponds to
sensation in these objects [appearances] is the transcendental matter of all objects
as things in themselves (facticity, reality)" (Al43/B182) (see ch 8 for detailed consid
eration of this relation). Sensation thus "expresses the matter (real) of outer things
whereby something existent is given" (CJ 189}-"that whereby an object is repre
sented as given according to its existence" (Progress 276)-and so constitutes "the
real of perception" (CJ 291), and that whereby "perception is, in the first instance,
the representation of an actuality" (A374). Since there is no reason to doubt that
Kant held this to be true of all sensations simply as such, regardless of their qualita
tive content, one should not think it strange to find characterizations such as these
within the confines of his transcendental philosophy. It would be transformed into
empirical philosophy only if one sought to utilize sensation as evidence for the exis
tence of this or that particular object, for that would require one to distinguish one
sensation (or kind of sensation) from others, which is possible only empirically (e.g.
the impenetrability definitive of body must be given empirically or not at all: BS,
PFM 295). Thus, insofar as the indication of the real existence (actuality) of what
ever is present in it is a universal feature of sensations, Kant may be presumed to
have deemed it a legitimate datum of transcendental investigation.
The same seems to have been true of the manifoldness of sensation. As con
ceived by Kant, sensibility is capable of producing sensations differentiated in
quality in myriad ways. Data of visual sensation, for example, are distinguished
by hue, brightness, glossiness, saturation, in addition to other features that may be
perceived without being discerned or even discernible. For the differentiation pres
ent in sensation but only obscurely perceived constitutes a field so "immeasurably
3 The a priori character of this kind of consideration of sensation is evident from Kant's
inclusion of concepts of an object of sensation in general among the predicables: "I was justified
in calling them by their old name of categories, while I reserved for myself the liberty of adding,
under the title of predicables, a complete list of all the concepts deducible from them by connec
tion with one another, or with the pure form of appearance (space and time), or with the matter
of appearance insofar it is not yet empirically determinate (object of sensation in general)" (PFM
323-4).
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 81
great" that our sensibility may be compared to "a great map on which only a few
places are illuminated" (Anthropology 135-6; also B414-5n and PFM 306-7). Their
manifoldness even extends beyond actual sensations to possible, such as the deli
cious taste, unknown to present-day earthlings, of a fruit growing on a planet
orbiting a distant star. It is the manifoldness of sensation only, not the particular
contents that happen to comprise it, that falls within the remit of transcendental
philosophy. It would be of no interest to the transcendental philosopher if sensa
tional differences that, with the present constitution of our senses, are expressed
by variations of, say, a certain color (the particular shade, brightness, glossiness,
etc.) were instead expressed by variations of a different color, by tactual differences
(of smooth/rough, soft/hard, dry/wet, etc.), olfactory ones, or even variations of
a kind of sensible quality wholly unknown to humans. All that matters to the
manifoldness of sensations is the fact of their difference, not the particularities of
those differences. As a universal feature of all sensations simply as such, possible
no less than actual, this inexhaustible manifoldness is therefore a datum free of all
dependence on acquaintance with any particular datum of sense (which is possible
only a posteriori) and so may legitimately be utilized in transcendental philosophy.
And it is certainly at least part of what Kant had in mind by "the synopsis of the
manifold a priori through sense" (A94), ahead of all synthesis in imagination and
understanding (A97; KMM ch 6-A).
Though there seems no reason to doubt that the form of appearances relates to
the manifoldness of sensation, it is not immediately evident whether their matter
"corresponds" to sensations insofar as they indicate existence, have the sensational
quality they do, or both. The textual evidence proves that Kant certainly included
the first:
The evidence that the matter of appearances corresponds to the quality of sensa
tions as well is less unambiguous. What seems clear is that Kant did not believe that
appearances, the objects of intuition given through sensations, are bearers of the
82 Kant's Anatomy of the Intelligent Mind
same qualities met with in sensations themselves, and so are not blue (colored), loud
(audible), smelly, tasty, painful, and so forth:
The taste of a wine does not belong to the objective determinations of the
wine, not even if by the wine as an object we mean the wine as appearance,
but to the particular constitution of sense in the subject that tastes it. Colors
are not constitutive properties of the bodies to the intuition of which they are
attached but only modifications of the sense of sight that is affected by light
in a certain manner. Space, on the contrary, as condition of outer objects,
belongs of necessity to their appearance or intuition. Tastes and colors are
not necessary conditions under which alone objects can be for us objects of
the senses. They are combined with the appearance only as contingently con
tributed effects of the particular organization [of the human body]. (A28-9)
[E].g. sight, hearing, touch, through sensations of colors, sounds, and
heat, do not allow cognition of any object, least of all a priori, because they
are merely sensations and not intuitions. (B44)
[S]ensation, which designates the real of intuitions, is not subsumed
directly under the concept of magnitude because sensation is not an intuition
that contains space or time, although it posits the object corresponding to it
in both (PFM 306)
[W]hat concerns the constitution of sensible intuition in regard to its mat
ter, namely sensation, e.g. bodies in light as colors, in vibration as sounds, in
salts as sour flavors, is merely subjective and yields no cognition of objects,
and, consequently, what is subjective cannot provide examples of representa
tions in empirical intuition that are valid for everyone, for it contains no data
for a priori cognitions, as do space and time, and in general can never be
counted as cognition of objects. (Progress 269)
Intuition is related to the object, sensation merely to the subject. (AA 23
E XI 15 at A20/B34)4
The first capacity of the human psyche and the condition for the rest is
sense, whereby the psyche receives representations as effects of the presence
of the object and not produced by itself. The representation of the senses
as something belonging to the state of the subject is called sensation; but as
something that is related to an object, appearance. There are sensations with
out noticeable (merklich) appearances and appearances without noticeable
sensation; yet the two are always together (beisammen). (AA 15 § 620, 1769)
None of this is to say that the quality of sensation makes no difference to the
matter of the corresponding appearance. Indeed, Kant's transcendental principle
of Anticipations of Perception relates quite specifically to this aspect of sensations
(ch 16-C-5). But what texts such as these make clear is that Kant did not conceive
4 Citations from AA 23 are Kant's marginal notations in his personal copy of the Critique of
Pure Reason and so nearly always relate to a specific passage.
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 83
tain relations the form of appearance. Since that wherein sensations can alone
be ordered and placed in a certain form cannot itself again be sensation,
while the matter of all appearance is indeed only given to us a posteriori, its
form must lie ready for them a priori in the mind, and therefore be capable of
being considered separately from all sensation. (A20/B34)
The way this is worded, Kant may seem to be equating sensations with the manifold
matter of appearances that is being "ordered and placed in a certain form." But if
this were what he intended, it would imply what he elsewhere denied explicitly (as at
PFM 306}-that sensations themselves have spatial and temporal form-which not
only would erase any distinction between sensations and appearances, but would
imply that olfactory sensations might have triangular shape and auditory sensa
tions a left half, that gustatory sensations can be situated in this location rather than
that, that one sensation of thirst can take up more space than another, and so on.
Yet Kant was quite clear that this is not so: "sensation is in itself not an objective
representation at all and intuition neither of space nor of time is to be found in it"
(B207-8), for "although it posits the object corresponding to it in space and time,
sensation is not an intuition that contains space or time" (PFM 306) and "does not
itself occupy any part of space or time" (309).
Fortunately, there is a way of reading A20/34 that does not involve equating
sensations with the manifold matter of appearances that is "ordered and placed in a
certain form." The characterization of the form of appearances in the first sentence
of the earlier citation from A20/B34 is preceded by this characterization of their
matter: "I call that in appearance which corresponds to sensation its matter." Since
correspondence is a relation and implies the distinctness of the items concerned
84 Kant's Anatomy of the Intelligent Mind
in it, the most natural reading of the sentence is that the matter equated with the
"manifold of appearances," which their form allows "to be ordered in certain rela
tions," corresponds to sensations and so cannot be identical with them. And so it
is with the problematic sentence that follows: it is not sensations themselves that are
ordered and placed in a certain form but the existences indicated by these sensations
and, more particularly, the realities (qualities cognizable as intensive magnitudes
under the Anticipations principle) in the undetermined objects of intuition-the
matter of these appearances-that correspond to the subjective sensational quali
ties in which intuited realities manifest themselves to beings with faculties of sensa
tion constituted like ours. And, in general, where Kant was concerned to be precise,
he was careful to make clear that the matter of appearances corresponds to but is
not identical to sensations (e.g., Al43/B182, A166, and A723/B751).
On this reading, the form of appearance does "lie ready for sensations a priori
in the mind," but only so that the matter of the appearances corresponding to
these sensations may be ordered and related in accordance with that form-a
matter that, like sensations themselves, can "only be given to us a posteriori"
and never admits "of being considered separately from all sensation." Similarly,
when a content originally given in sensation is reproduced in imagination, as
with Humean idea-copies, the appearance has a matter only in the attenuated
sense of no longer indicating an existence but still exhibiting an image of the
reality (the Humean impression-originals) corresponding to the sensation being
reproduced.
Kant's denial of objective status to sensation in the texts cited earlier might seem
to be contradicted at CJ 206, which contrasts sensations such as the green color of
a meadow as "objective sensation, ...the perception of an object of sense" with the
purely subjective data of feeling (pleasure and pain). Yet there is no suggestion that
this objectivity consists in the empirical object (the meadow that appears) being
itself, objectively, green. Rather, sensation is objective, by contrast with the feeling
of pleasure its color produces in one, because it corresponds to the matter of outer
appearances, not because it is identical with it, as an earlier passage in CJ makes
clear: "sensation ...expresses the merely subjective of our representations of things
outside us, but more properly expresses their material (real) (whereby something is
given as existent)," and, to that extent, unlike feeling, "sensation too is employed
in the cognition of objects outside of us" (CJ 189). In the transcendental sense of
objects existing outside us in themselves (A37 3; ch 4-A), sensation is the subjective
expression of "the transcendental matter of things in themselves (facticity, real
ity)" (Al43/B182; ch 8-E). In the empirical-cognitive sense of objects (appearances)
outside us in space, sensation gives subjective expression to their reality insofar as
its variable intensity makes them representable as intensive magnitudes under (the
schema of) the categories of quality. Neither CJ 206 nor any other text I know of
warrants the supposition that objects, be they transcendental (things in themselves,
noumena) or empirical (appearances, phenomena), are themselves ever colored,
smelly, tasty, or the like.
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 85
The most likely cause of the widespread confusion on this score is the tendency
to confound "appearance" in its ordinary empirical sense with the special tran
scendental sense in which Kant employed it . He illustrated the difference with
exceptional clarity in his analysis of the relation of a rainbow to raindrops at
A45-6/B62-3. In the ordinary empirical sense, the colors making up the rainbow
are "appearances" of raindrop "things in themselves." In Kant's special transcen
dental sense, however, where the raindrops themselves are mere appearances of an
unknown transcendental object= X, the qualities of the raindrops "in themselves"
(in the non-transcendental sense) are limited to the properties they posses indepen
dently of differences of both spatial standpoint and the particular organization
of this or that sense. Since this is true neither of color nor any other sensational
quality ("They are combined with the appearance only as contingently contributed
effects of the particular organization [of the human body]," A28-9), it follows that
raindrops, meadows, and physical objects generally, albeit appearances in Kant's
transcendental sense (phenomena, not noumena), are never "in themselves" (in the
non-transcendental sense) bearers of color, smell, taste, or any other sensational
quality. Instead, the qualities of physical objects are limited to those specific to
their intuition.5 Where the matter of appearances (in Kant's transcendental sense
of the term) is concerned, this means that their reality is restricted to those quali
ties concerned with the filling (occupying and containing) of space and time, which
sensations themselves (visual and tactual included) never do.6
The difficulty raised by the difference between the empirical and transcendental
senses of "appearance" is related to the problem many have with Locke's notion
of a secondary quality. It was never Locke's intent to deny that physical objects
have secondary qualities. His point in distinguishing them from primary qualities
was simply to say that ideas of these qualities in the sensing mind, unlike those of
primary qualities, bear no qualitative resemblance, even the most generic, to the
5 Strictly speaking, the space and time in the field of appearances of the raindrops of Kant's
example at A45-6/B62-3 are not the pure space and time of the Transcendental Aesthetic (which,
as pure, are imperceptible: ch 16-B), but are the space and time that result from the determination
of the matter (reality) of appearances in accordance with the principles of pure understanding
(see Part V).
6 It is helpful in this regard to put together what Kant says at B207-that "as objects of
perception, appearances contain ... the matter for some object in general whereby something
is represented as existing in space or time" -with his statement at PFM 306 that "sensation
is not an intuition that contains space or time, although it puts (setzt) the object correspond
ing to it in both." Juxtaposing the two statements makes clear that the matter of appearances
determined by the schema of the categories of quality as reality filling (occupying, containing)
space or time corresponds to, but is in no sense identical with, sensation (sensational quality),
and that sensation plays an essential role in objective cognition only as that which puts some
object in space or time without itself either containing or occupying (PFM 309) space or time.
Consider also the A edition formulation of the Anticipations of Perception principle, "In all
appearances, sensation, and the real in the object corresponding to it (realitas phamomenon),
has an intensive magnitude" (Al66 ): the object here in question is an appearance (in the tran
scendental sense), the matter of which is the realitas phcenomenon that corresponds to, but is
not identical with, sensation.
86 Kant's Anatomy of the Intelligent Mind
ity (Al68-9/B210-ll), is easily extended to color. For just as the moments of grav
ity of the load I bear cause in me a sensation of back-breaking heaviness without
their resembling that sensation in the least, so too the causal moments whereby the
intensive magnitude of the cinnabar causes a sensation of red in me bear no resem
blance to that sensational quality. T hus, so long as we are careful to understand
"appearance" in the transcendental sense Kant intended, there can be no danger of
confounding the sensational qualities that "appear" in the ordinary, empirical sense
with the qualities that constitute the matter of the appearances intuited through sen
sations and exist in them independently of one's standpoint and the species-specific
(e.g., human) organization of one's faculty of sensation.
To appreciate just how important it is to understanding Kant that one never
confound the matter of appearances with sensations, consider what would follow
if the categorial determination of this matter as filling space and having inten
sive magnitude-that is, the matter of the Schematism and Principles of Pure
Understanding-were construed as applying, in whole or in part, to sensational
qualities. Under the Second Postulate of Empirical T hought (ch 18-D), the matter
that fills space has objective existence (actuality) in both its intensive and extensive
magnitude (the two constitutive properties of appearances: Al80/B222-3; ch 17-A)
even if it is not present to consciousness in sensation, provided that it is directly or
indirectly connected with something that is present in accordance with the rules
of connection (Verkniipfung) specified in the T hree Analogies of Experience. Now,
consider the following question: if Kant supposed that sensational qualities are
wholly or even partly identical with the matter of appearances, would the Second
Postulate not have obliged him to accord objective existence not only to physical
objects, forces, energy, etc., when they are not present to consciousness in sensation
but also to colors, sounds, smells, feels, tastes, itches, hunger, thirst, and even plea
sures and pains? To answer yes is to say that he would attribute objective existence
to the green of a meadow that no one sees, the sourness of a lemon no one tastes,
the fragrance of a rose no one smells, and even the itchiness of a skin rash of some
one under sedation, and that he did so notwithstanding his express and iterated
denial that sensations are objective in the same sense as the physical matter that
occupies and contains space and time (A20-l/B35, A28-30/B44-5, A45-6/B62-3,
Al68-9/B210-l Progress 269 , AA 18 § 6355). To answer no, however, is to commit
oneself to reading Kant as denying that sensational qualities enter in any way into
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 87
which individually may be far too weak to occasion (stimulate) any corresponding sensa
tion. In other words, not identifying the matter of appearance with sensation (1) allows
objects to have intensive magnitudes beneath the threshold of sensation, (2) allows these
magnitudes to be composed (zusammengesetz) by coalition so that their sum constitutes
an intensive magnitude capable of occasioning sensations, and (3) allows sensations of
divergent quality to be equated quantitatively insofar as the intensive magnitudes of the
objects (appearances) corresponding to them are equal (e.g., the quantitative equality
between the intensive magnitude of sunlight and the composite formed when the inten
sive magnitude of moonlight is multiplied 200,000 times: Al 79/B221).
Turning now to inner appearances, there is no more textual basis for supposing
that Kant identified their matter with inner affections (self-affections) than for sup
posing that he identified the matter of outer appearances with outer affections (sensa
tions). Given his iterated insistence that his theory of sensibility treats the inner exactly
as it does the outer, I therefore see no alternative but to conclude that the matter of
inner appearances is distinct from self-affection and corresponds to it in the same
sense the matter of outer appearances corresponds to sensation. This, however, raises
the question of what precisely the matter of inner appearances is and, more particu
larly, whether inner appearances have their own matter distinct from that of outer or
whether their only difference from outer appearances is formal, not material. Although
I am aware of no text in which Kant addressed the issue directly, it seems significant
in this regard that he insisted that "in order to understand the possibility of things in
accordance with the categories, and so to demonstrate the objective reality of the lat
ter, we require not merely intuitions but always outer intuitions" (B291). This suggests
that if inner appearances had a matter different from that of outer appearances, a
matter that consequently was impossible to apprehend through outer intuition, inner
appearances could not be understood as things in accordance with the categories,
thereby restricting the objective reality of the categories to outer appearances alone.
Since Kant clearly believed that he could and did demonstrate their objective reality
in relation to all appearances, inner no less than outer, there is strong reason to think
he regarded the matter of inner appearances as being identical to that of outer. Given
that the matter of any species of appearances is, qua matter, formless and so properly
neither outer (spatial) nor inner (temporal), there was nothing to prevent him from
adopting this view. And textual confirmation can be found in his assertion with regard
to inner intuition that "the representations of the outer senses constitute the actual
material (Stojj) with which we occupy our mind" (B67).7
7 What would it mean to cash out the matter of an inner appearance corresponding to an
inner affection such as an emotion in terms of physical reality? While purely temporal in form,
the appearance would have as its matter the intensive magnitude of a reality in an outer appear
ance such as the brain (more particularly, the amygdala in the case of an emotion). Insofar as the
degrees of that reality were causal moments, the emotion could then be cognized as its effect in
the same way sensations of heaviness are cognized as effects of gravitational reality and sensa
tions of color as effects of electromagnetic reality. I will return to this theme in Chapter 18, where
I will argue that Kant regarded the objects apprehended a posteriori in inner intuition as one and
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 89
What seems to me to clinch the case regarding the matter of inner appearances
1s Kant's account of substance. Substance, as the permanent substratum of all
change, is the final subject (letzte Subjekt) of all objective predicates, qualitative no
less than quantitative, as well the locus of all causal agency (A204-5/B249-50; ch
17-H). This means that the final subjects of all determinations of intensive magni
tude are substances; they are their true bearers. Now, according to Kant, "in order
to give something permanent in intuition corresponding to the concept of substance
(and thereby establish the objective reality of this concept), we require an intuition
in space (of matter), since space alone is determined as permanent, while time, and
so all that is in inner sense, constantly flows" (B291; ch 17-G). Where intensive mag
nitude is concerned, this is just to say that that the ultimate bearers of these quanti
ties are always material substances and that intensive magnitude is never anything
other than the measure of the reality with which they fill (extensive magnitudes of)
space. The consequence in respect to the matter of inner appearances should be
clear: since the locus of intensive magnitude is the matter of appearances, the only
objective meaning and validity the matter of inner appearances can possess is that
conferred on them by material substances. Thus, the matter of inner appearances
must be the same as that of outer appearances.
To call attention to the strict demarcation of the matter of outer and inner
appearances from the affections of outer and inner sense, I shall henceforth speak
of the appearances intuited through affections as exhibiting these affections in intu
ition through their matter. 8 I will also sometimes refer to the reciprocal mediation
of affections and appearances, meaning thereby that appearances are mediated by
the affections through which they are intuited ("intuition . . . is related to the object
through sensation," A20/B34), while those affections are in turn mediated by the
(matter of the) appearances through which alone they can be exhibited in intuition
(and thereby made capable of entering into cognitive experience). Having in this
section considered the former mediation, I turn now to the latter.
Where Kant's conception of appearances departs most radically from the concep
tions of his sensibilist predecessors-who invariably equated sensations with what
all accidents of material (spatial) substance (which, as mere substantia phamomenon, is neverthe
less compatible with his transcendental idealism).
8 I have chosen the term "exhibit" with the exhibition (Darstellung) of figurative synthesis and
schematism in mind (chs 15-16). Since the purpose of the latter is to exhibit concepts (includ
ing transcendental concepts by means of transcendental figurative synthesis) in intuition, it is a
natural extension of the term to speak of appearances as exhibiting the sense affections through
which they are intuited-a terminological choice further supported by Kant's determination
that intuition is a product of a synthesis of apprehension of the imagination (Section E and
Chapter 5).
90 Kant's Anatomy of the Intelligent Mind
appears (ch 7)-is his assertion that something remains to intuit even in the absence
of their matter. For the form of appearance makes it possible to intuit appear
ances "separately from all sensation" and so as the purely formal manifold of a
pure intuition:
I entitle all representations pure (in the transcendental sense) in which noth
ing is met with that belongs to sensation. Accordingly, the pure form of sen
sible intuitions in general, in which the manifold of appearances is intuited in
certain relations, is met with in the mind a priori. This pure form of sensibil
ity itself will also be called pure intuition. (A20/B34-5)
The metaphysical expositions of space and time of the Transcendental Aesthetic (as
Kant termed them in the 1787 edition of the Critique of Pure Reason) are intended
to explicate these concepts as concepts of pure intuitions of sensibility. They are,
more particularly, explications of what might be termed second-order concepts,9 for
a key component of the concepts of space and time expounded in the Aesthetic is
that neither space nor time are in any regard discursive, that is, representations in
any way dependent on universals (A24-5/B39, A31/B47, A68/B93, L 58, Progress
325), and so a denial that their representations are first-order concepts. Instead, as
pure intuitions, they are present in the mind prediscursively, through the operations
of sensibility alone, prior to and independently of all concepts of space and time, be
they ordinary, geometric, or scientific. The metaphysical expositions of the concepts
of space and time thus do double duty as expositions of the second-order concept
of pure intuition itself, with Kant's focus throughout confined almost exclusively to
those features space and time share that make them intuitions rather than concepts
and pure rather than empirical.
Subsequent to the expositions proper, it becomes clear that pure intuition is not
simply a more general concept than the concepts of space and time, but a concept
that, independently of how it is instantiated, holds a quite particular place in Kant's
transcendental aesthetic theory. This is indicated by his iterated insistence that it is
impossible to preclude the possibility of pure intuitions other than space and time
(A42/B59, B72, Bl39, Bl46, A230/B283, A613--4/B641-2, PFM 350-1, Progress
267, Anthropology 399, AA 18§ 5056 and§ 6312). Since a feature common to space
and time might very well be missing in other (perhaps possible) pure intuitions, one
cannot straightaway assume that their common features are ipso facto common to
all pure intuitions as such. In order to understand the metaphysical expositions
correctly, one must therefore abstract from everything distinctively spatial and dis
tinctively temporal to consider only those features that remain, for these, and these
alone, will be proper to pure intuition as such. And since this is just to say that pure
intuition can and must be conceivable independently of space and time, it must thus
9 This second-order notion of conception is rare in Kant. It is applied to space and time also
at A85/Bl 18, A88/Bl20, Al07, Al 56/Bl95, and AA 15 § 2967.
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 91
be recognized as being not just more general than space and time but theoretically
prior as well.
When this is done, it emerges that a pure intuition of a certain form, as the con
dition that precedes and makes possible all appearances of that form, 10 contains
these appearances within it as its manifold, does so completely a priori (hence uni
versally and necessarily), and thereby serves to unite all possible appearances of that
form in one and the same consciousness a priori. For the metaphysical expositions
characterize space and time as wholes that (1) can contain an inexhaustible infinity
of spaces and times within them (not under them as with instances of concepts),
(2) precede their parts (which are possible only via the limitation of these presup
posed infinite wholes), and (3) are essentially one in the sense that no space or time
is possible outside or independently of them (#3-#4 at A25/B39--40 and #4-#5 at
A31-2/B47-8)." Since, for Kant, spaces and times are merely appearances whose
constitutive forms derive from pure space or pure time, and since appearances are
mere representations, this portion of the metaphysical expositions makes clear that
he conceived of pure intuition-be it space, time, or anything else-as an individual
representation that contains all its manifold (possible as well as actual) within it
and so serves to necessarily unite all appearances of its form in one and the same
consciousness entirely a priori
T he remainder of the metaphysical expositions of space and time are concerned
with what remains after carrying out the abstractions to which the Transcendental
Aesthetic is devoted, one to establish that pure intuition really is an intuition (pre
discursive) and not a concept, the other that it really is pure (absolutely a priori)
and in no way empirical:
T he abstraction from sensation is limited to its contents, which can only be given
empirically in perception and so are out of bounds to transcendental philoso
phy: sensational qualities such as colors, sounds, tactual feels, smells, and tastes, as
well as the distinct existences and realities corresponding to such qualities (Section
A). Given that space and time have their origin in pure intuition and pertain exclu
sively to the form of appearances, the implication of prescinding from all that is
10
The status of pure intuition as a condition of all possible appearances of its form is
described and affirmed at A23/B38-9, A30-l /B46, A34/B50, A90/Bl 22-3, A93/Bl25, Al lO,
Al 11, Bl36, Bl48, Al49/Bl 88, A429/B457n, A441/B469, PFM 284, and ID 402.
u Space and time are not in every sense impossible to multiply instantiate since, as pure intu
itions, the space and time of my sensibility are different from the space and time of yours: see
A362-3.
92 Kant's Anatomy of the Intelligent Mind
manifold of empirically apprehended sensory data that, with the present constitu
tion of sensibility, makes its appearance in empirical inner intuition as a succession
would not appear temporal at all but quite otherwise. For if time, and so too all
of its modes (succession, simultaneity, and duration), pertain only to the form of
appearances, and appearances are distinct (both in form and matter) from sensa
tions and data of affection generally (section B), then the inescapable implication is
that sensations are not in themselves successive and are therefore so entirely devoid
of all temporal character that no amount of empirical scrutiny could succeed in
extracting the least temporal content (quality or relation) from them.
The case is the same with space. Since space pertains solely to the form of appear
ances, and appearances are distinct in matter no less than in form from sensations
(Section B), the implication of Kant's explication of space as a pure intuition of
sensibility is that space, properly so called, is never to be met with in visual, tactual,
or any other kind of sensation, and that if we had a pure intuition other than space,
no amount of empirical scrutiny could discover the least spatial content (quality
or relation) in sensation. It is merely an accident of human psychology that outer
intuition is primarily a visual and tactual affair. For if our pure intuition was not
space but Z, these same sensations would be pressed into service for the intuition
of Z-ial appearances, and we would be just as convinced of the inherent Z-iality of
what we see and touch as we presently are of its inherent spatiality. Similarly, such
inherently species-specific psychological limitations as the human inability to visu
alize outer appearances in more than three dimensions cannot be supposed to limit
outer appearances themselves to three dimensions. For just as the constitution of
these appearances is completely indifferent to the particular qualities of the (visual,
etc.) sensations through which they happen to be intuited, it is in no way concerned
in the inherent limitations of the empirical psychological capacities whereby those
sensations are utilized for outer intuition.
What then does the abstraction from sensation called for by transcendental aes
thetic tell us about pure intuition generally, rather than just space and time in par
ticular? Its manifold's complete independence of the contents of sensation and the
species-specific limits of human psychology opens the way to precisely the kind of
a priori yet fully objective determination (A28/B44) that Kant postulated in the first
part of his "Copernican" hypothesis: "If intuition must conform to the constitution
of the objects, then I do not see how anything can be known of them a priori; but if
the object (as object of senses) conforms to our faculty of intuition, I can represent
this possibility to myself quite well" (Bxvii). If the object conformed to our faculty
of sensation, then it would be subjective and not an object at all (Section B). Only
if the object can conform to our faculty of intuition without ipso facto having to
conform to any faculty of sensation (touch, vision, etc.) does the possibility exist of
determining that object completely a priori (namely, by determining the manifold
of pure intuition through transcendental synthesis or mathematical construction).
The other part of the abstraction required in transcendental aesthetic involves
"separating out everything that the understanding thinks in [sensibility] by means
94 Kant's Anatomy of the Intelligent Mind
contained in them (it being part of their metaphysical exposition that they are
essentially one, so that no space or time-no appearance having their form-is
possible outside or independently of them) (ch 4-C). Otherwise, the appearances
apprehended in intuition by their means are devoid of all order and relation (they
are "met with in the mind scattered and single in themselves," Al20). Nevertheless,
as pure intuitions that contain all their manifold in one representation, united in
its consciousness, prediscursive space and time supply precisely what the higher
relating and ordering functions of imagination and understanding presuppose. For
if appearances, possible as well as actual, could not all be apprehended a priori as
the manifold of a single intuition, all immediately together in one and the same
consciousness a priori, it would be impossible to compare, associate, combine, or
otherwise synthetically relate them, much less confer on them the a priori neces
sary synthetic unity that Kant deemed essential in order to cognize them as objects
in experience or construct them in pure intuition. T hus, by abstracting from the
contributions not only of sensation but also of discursive understanding, the meta
physical expositions of transcendental aesthetic serve to make clear precisely what
kind of unity of sensibility a mind capable of these operations must have.
It is this unity, I believe, that provides the answer to the question posed ear
lier: why a pure intuition of some sort, be it space, time, or any other, is an essential
requirement of sensibility itself, prior to and independently of all higher operations
of the mind. I shall therefore devote the remainder of the chapter and the entirety
of the next to examining the unity of sensibility and why only pure intuition is
capable of bringing it about.
Pre-Kantian philosophers took for granted the unity of the data of all the senses,
inner no less than outer, in one and the same consciousness, typically in the guise
of a "perpetual flux and movement" (THN 252/165) of "internal and perishing
existences" that "appear as such" (194/129). 12 To this scattered manifold of data, as
input, higher operations of the mind are applied to copy, discern, separate, com
pare, relate, combine, and otherwise organize them until they are sufficiently well
ordered to yield, as output, ordinary and scientific consciousness of a world of
causally interacting, enduring substances, subject to universal laws of nature.
Whereas Kant's transcendental analytic of the understanding was directed to
explaining the same output, the radically original trail he blazed to get there starts
from a radically transformed conception of the input. It is not simply that he was
the first to distinguish appearances from affections of sense by positing a form and
12 Similar conceptions can be found in Descartes (letter to Arnauld, June 4, 1648), Locke
(ECHU 11/vii/§9 and xiv/§3), and Berkeley (PHK I § 98).
96 Kant's Anatomy of the Intelligent Mind
matter in the former that is absent from the latter. For this, I shall endeavor to show,
is simply a consequence of the deeper recognition that one cannot simply take for
granted but must instead explain the unity of all possible sense affections in one
and the same consciousness, ahead of all association and propositional thought. It
is this problem, I believe, that obliged Kant to rethink the nature of sensibility from
the ground up and led eventually to the notion of a pure yet sensible intuition as
the only means of solving it.
Kant's focus on pure intuition, in the Critique of Pure Reason and elsewhere, was
almost entirely on the indispensable role it plays in effecting his "Copernican revolu
tion" in metaphysics, together with its consequence, transcendental idealism as the
inescapable of price of making synthetic a priori judgments of any kind possible
at all (whether in transcendental philosophy, metaphysics of nature and morals, or
mathematics). What tended to be given short shrift was sensibility's own immanent
need for pure intuition, prior to and independently of the question of the possibil
ity of such judgments. Yet unless this is understood, I do not see how one can hope
to fully and precisely comprehend what pure intuition is and, in particular, why it
not only can but must be understood as no less a product of sensibility than sensa
tion itself, ahead of all propositional thought (judgment, discursivity). We must
therefore attempt to extrapolate from the concept of pure intuition expounded in
the Transcendental Aesthetic and related texts precisely what the explanatory need
is that necessitates its incorporation into the theory of sensibility.
The data of vision are as good a place as any to begin. The unity of the data
of this sense, often characterized as afie ld, is simply the consciousness in which all
colors are capable of standing immediately together, represented as a single mani
fold: reds immediately together with greens, greens with blues, and so forth. Yet,
although I am able to perceive a green immediately beside a red and a yellow imme
diately beside a green in the visual field, I cannot set a smell, sound, or tactual feel
immediately beside that yellow in it. The unity of visual sensibility thus has no place
for any but visual data. Where then is there a field-a prediscursive, purely sensible
consciousness-in which these nonvisual data can be set immediately together with
visual, a field in which all data of the senses are immediately contained, represent
able as a single homogeneous manifold of sensible appearances?
None of the other sensory fields seem any better equipped to meet this need
than vision. In the auditory field, sounds can stand immediately together with other
sounds, ready to be blended and combined synchronically to form such things as
chords and diachronically to form such things as melodies. But does this unity of
auditory sensibility extend to data of other senses to enable us to blend sounds with
colors, smells, tastes, and tactual feels to form multisensory "chords" or "melo
dies"? Clearly not. Or again: one tactual datum can be felt immediately together
with any other in a single unity of tactual sensibility and so can blend with or
gradually morph into any other in ways they cannot with data of the other senses;
and so too for smells with other smells in the olfactory field but not non-olfactory
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 97
data, and flavors with other flavors in the gustatory field but not non-gustatory
data. Since unity of sensibility calls for the data of all the senses to be represented
as a single homogeneous manifold, all contained immediately together in one and
same prediscursive, purely sensible consciousness, it is therefore necessary to look
beyond any particular sensory field for the requisite "super-field."
Before conceding that any such super-field "cannot itself again be sensation"
but "must lie ready for them a priori in the mind" (A20/B34), one might instead
maintain that unity of sensibility is simply a consequence of the co-occurrence
(1) since pure intuition is essentially one, all appearances bearing its form are
ipso facto contained within it as its manifold, and so united in one and the
same intuitive consciousness a priori (universally, necessarily);
(2) pure intuition is as inexhaustibly differentiable as sensation itself, so that
nothing can ever be present in sense, as affection, that could not also be
exhibited in intuition as appearance; and
(3) no matter how much appearances may differ on their material side (as
exhibitions of the qualitatively heterogeneous sensations to which they
correspond), they are uniform in form, and so always admit of being
represented as a single formally homogeneous manifold.
The upshot is that even if sensations can never be all immediately together as the
manifold of a single representation, much less be represented as such a manifold
completely a priori, the appearances corresponding to them can and must be. Thus,
unity of sensibility is a consequence of the reciprocal mediation of appearances and
sensations: sensations, however incommensurable in quality, are united in one con
sciousness by means of the a priori unity in one representation-be it space, time,
or anything comparable-of the formally homogeneous appearances intuited by
means of these same sensations.
If this interpretation is correct, it means that without pure intuition, and so
apart from homogeneous appearances to exhibit heterogeneous sense affection, no
consciousness is possible in which all affections could be represented as a single
manifold. There could then be no unity of the manifold in sensibility itself, prior to
and independently of all higher functions of the mind, and nothing would prevent
the representations given in the a priori synopsis of sense-the only operation Kant
attributed uniquely to receptivity ahead of all synthesis of imagination and unity of
the understanding-from being "completely alien, isolated as it were, and separate
from other representations" (A97). Indeed, in the absence of the unity that pure
intuition alone can confer on the synoptic manifold, it seems doubtful that one
could any longer speak of "sensibility" at all. Sensibility, in pre-Kantian sensibilist
and intellectualist theory alike, is the most elementary level of consciousness, the
source of the raw, unassociated, unthought, yet-to-be-ordered-or-related inputs on
which the imagination and understanding were supposed to operate. At the same
time, sensibility is, and must be, a multisensory consciousness in which data of all
the senses are represented as given immediately together, a single homogeneous
manifold of data, completely a priori. Only when apprehended in this preassocia
tive, prediscursive unity of consciousness can these data become available to the
imagination in the form of a single manifold required by that faculty if it is then to
perform any of the operations credited to it by pre-Kantians: separately reproduce
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 99
any representations that are apprehended together and reproduce together any
representations that are apprehended separately so as to compare or associate any
with any other (ch 13). Similarly, the unity of sensibility must already be in place
if the data of sense are to be made available to the understanding in the form of a
single manifold that it too requires (ch 9) in order to perform the operations that
pre-Kantians credited to it: derive concepts from these data by means of reflection,
comparison, and abstraction-concepts that can then be combined in judgments
that themselves can be related in inferences.
If one grants that the unity of sensibility presupposed by the imagination and
understanding is impossible without pure intuition, then the question confront
ing one ceases to be whether there is any justification for classifying pure intuition
as sensible notwithstanding its purity and instead becomes the question whether
there is any justification for classifying sensations (affections of sense in synopsis)
as representations of sensibility in the absence of pure intuition. For implicit in
the notion of "sensibility" is not simply that a manifold of affections be given in
the synopsis of sense but, in addition, that there be a means of representing that
manifold as a manifold (A99). According to Kant, pure intuition alone can enable
us to apprehend all data of sense (affections, sense impressions) as the manifold
of a single representation, all contained in the consciousness of that representa
tion, such that, apart from this prediscursive unity of sensibility, they could not
be represented as a manifold but would instead be "completely alien, isolated as it
were, and separate from other representations" (A97). It is the two together, then,
that constitute sensibility: the a priori synopsis through which a manifold of senses
affections is given and the pure synthesis of apprehension whereby that manifold is
exhibited as a homogeneous manifold of appearances in pure intuition. And since
this is just to say that without either sensibility is impossible, it should now be clear
why, on Kant's account, pure intuition, notwithstanding its purity, has just as great
a claim to be accounted a representation of sensibility as sensations do. 13
Some readers may object that my interpretation is too speculative. To this, I have
four responses. (1) Even if it were speculative, we still have to ask whether it is
avoidable speculation; and there is compelling reason, in my view, to think that it
is not. Simply put, the prediscurive manifold of sensibility enters into every notion
of importance in the Transcendental Analytic: apperception as the synthetic unity
of that manifold, imagination as the source of the transcendental synthesis of that
manifold, the categories as universal representations of the pure synthesis of that
manifold, etc. If one has not already adequately understood what the prediscursive
13 When I was a student, I was taught that Kant was not an innovator in the theory of ideas
but rather took his straight off the Lockean-Leibnizian rack . But if that were true, Kant's distinc
tion between synopsis and apprehension had no business being in his theory. To find out more,
I scoured the literature but found that the few commentators who even took notice of the distinc
tion did not recognize its significance, much less do it justice. T hus, I was set on the course that
eventually led me here.
100 Kant's Anatomy of the Intelligent Mind
manifold is purely in terms of sensibility itself, ahead of all matters properly logi
cal and epistemological, how can one possibly hope to approach Kant's theory of
synthetic a priori and a posteriori cognition without risking major gaps and mis
understandings? It is the essence of Kant's psychologistic method to determine not
only what the faculties of the mind contribute to the content of representations but
also which faculties contribute what and in which psychological sequence (ch 2-E).
Clearly, where the manifold of space and time is concerned, pure sensibility comes
first; and if we have not previously determined what it contributes prediscursively,
we cannot possibly hope to determine what the categories contribute via transcen
dental synthesis (or, for that matter, what mathematical concepts contribute via
constructive synthesis and empirical concepts via reproductive synthesis). In my
view, any interpretation that starts from an inadequate or mistaken understand
ing of Kant's doctrine of pure sensibility is effectively guaranteed to result in an
inadequate or mistaken understanding of everything in Kant's theory of discursive
understanding that incorporates or depends on it. And since there are only so many
things the unity effected by pure intuition (the essential oneness of the prediscur
sive manifold of sensibility) can be, I see no way for a responsible interpreter to
avoid taking a position, even if it involves a measure of conjecture (i.e., informed,
well-thought-out speculation).
(2) The textual evidence regarding the unity effected through pure intuition may
or may not be conclusive, but it is by no means insubstantial. We have already seen
some of it and will consider a good deal more in the following section and the next
two chapters.
(3) There are considerations of fit with Kant's general and transcendental logical
doctrines that need to be taken into account. Even if one thinks the direct evidence
is inconclusive, there is surely only one reading of the unity of pure intuition that
attains optimal fit with these other doctrines, or at least distinguishes itself from
alternatives that, by comparison, lead to inconsistent, incoherent renderings of the
theory elaborated by Kant in the Transcendental Analytic. And it is my contention
that the interpretation of that unity developed here has the advantage over every
rival of a well-nigh seamless fit with Kant's theory of discursive understanding in
both its general and transcendental logical developments (chs 5-G, 9-B, 10-E, 11-C,
14-E, 15, and 16).
(4) Finally, the Critique of Pure Reason is not a sacred text locked away in some
shrine, but the exposition of a philosophy. Kant provided his readers with prin
ciples, doctrines, and methods not so that they would do nothing with them but
rather so that they would put them to use to frame new problems and illuminate
remaining obscurities in order to extend and perfect that philosophy. Consequently,
whatever can be proven to follow from it to the exclusion of all alternatives may,
and probably should, be counted as part of it, even if it is not found in the writings
Kant bequeathed to us but instead depends on informed conjecture. By contrast,
whatever requires the importation of principles and methods devised centuries later
by others, such as the borrowings from analytic philosophy now so commonly used
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 101
One should not assume that the sensible nature of pure intuition implies that it is
a passive given of sense like sensations. Kant's view was not that sense (receptivity)
furnishes two types of data, heterogeneous sensations and homogeneous appear
ances. On the contrary, the active faculty of imagination (spontaneity) needs to
generate a formally homogeneous manifold precisely because sense, in its original
receptivity (a priori synopsis), suffices only for sense affections but not for unity of
sensibility. For the latter, a pure synthesis of apprehension in intuition to generate
pure intuitions is necessary as well; and since this operation cannot be attributed to
receptivity, Kant credited it to the imagination (A99-100). T hus, his transformed
conception of sensibility has this further consequence: the activity of imagination
must be reconceived to extend beyond the reproduction and association of the man
ifold all the way down to the manifold apprehended in intuition itself (A120 + n).
To convince ourselves of the necessity of this new reconceptualizion, recall
what emerged in the preceding sections: there is no one consciousness in which
all sensations are originally given; such a consciousness is possible only through
the mediation of pure intuition and the homogeneous manifold of appearances it
makes possible. What else can this mean but that unity of sensibility is impossible
given sense alone in its original receptivity? Yet to concede that it is possible only
through the mediation of pure intuition is as much as to say that this consciousness
has to be made and is in no sense a given. And since its production is the immediate
result of intuiting, through heterogeneous sensations, a manifold of homogeneous
appearances contained in a pure intuition that is "essentially one" (A25/B39), this
can only mean that both pure intuition itself and the appearances it makes possible
are products of synthesis in imagination rather than givens of sense.
If this analysis is correct (and, as soon will become clear, the textual evidence in
its favor is strong), Kant's fundamental insight regarding sensibility went beyond
the recognition that even if sensations themselves cannot all be represented as the
manifold of a single representation, they can still be united in one consciousness by
representing something through them that can be represented as such a manifold.
102 Kant's Anatomy of the Intelligent Mind
It also included the realization that this unity can be effected even if the formally
homogeneous manifold of appearances that does so is generated in imagination
and incapable of existing outside or independently of it. For what effects unity of
sensibility is the essential oneness of the pure intuition that ensures that all possible
appearances intuited through sensations will be apprehended as the manifold of a
single representation (Section C). If this oneness is possible only as a synthetic unity,
dependent on spontaneity as much as on receptivity, then one will simply have to
reconceive imagination as "a capacity for a priori intuitions" (CJ 190), "productive
and self-active as creator of arbitrary forms of possible intuitions" (CJ 240), and so
"a necessary ingredient of perception itself " (A120n).14
That pure intuition needs to be understood as an expression of the spontane
ity of the mind, and not merely its receptivity, follows also from its role in Kant's
system as a condition not for having sensations but for being able to represent
an object through them. For example, one cannot form the representation of an
apple without combining (synthesizing) the data of various senses into a single
complex representation. However, before they can be combined, they must be rep
resentable immediately together as a single, homogeneous manifold, united in one
and the same consciousness. Given the incommensurability of the smell, taste,
feel, look, and (signature crunch) sound of an apple-such that each datum falls
within a distinct, isolated sensory field not immediately combinable with any of
the others-their representation as a complex, sense-divide transcending object
of a single, unified outer sense would otherwise be impossible. The mind must
consequently be equipped to synthesize representations that correspond to the
data of sensation given in synopsis but, unlike them, admit of being represented
immediately together as a single manifold. The resulting appearances, although
no less "scattered and single in themselves" (A120) than the data to which they
correspond, can then be subjected to the combinative syntheses of imagination
and understanding that transform them into the representation of an objective
unity ("complex idea"). It is thus not directly but via the appearances that exhibit
them that the imagination and understanding are able to combine the data of
vision, touch, taste, smell, and sound into the image of an apple. Or, to put the
same point in more Kantian terms: if distinct, as yet unrelated perceptions are
to be associated in reproductive imagination and combined in discursive under
standing, the imagination must first "take impressions up into its activity, that is,
apprehend them" (A120) in accordance with the prediscursive synthetic unity of
the manifold in pure intuition (B 160).
That a synthesis of apprehension in intuition is requisite to make perceptions
accessible to reproductive imagination was not a novel notion. What was original
14 See also A99-100, A107, B136n, B140, B160-ln, CJ 232, 287, 292, CJ 287, Anthropology
§ 28, and AA 15 §§ 341-2. My earlier book on Kant, KMM, focuses on the role of imagination
in intuition.
Unity of Sensibility ( 1): Sensation, Intuition, and Appearance 103
to Kant was the attribution of a synthesis with this function to imagination rather
than sense:
Kant situated apprehension between the synopsis of the senses, which supply a vari
ety of impressions (visual, tactual, etc. sensations) without bringing them together
as the manifold of a single representation, and the synthesis of reproduction, which
associates apprehended perceptions (intuited appearances) that, while contained in
a single intuition, have yet to be ordered or related in any way. The product of
the synthesis of apprehension-the bare manifold of homogeneous appearances
(corresponding to sensational qualities) that have yet to be associated and recog
nized (e.g., as an apple)-thus corresponds to what previous psychologists equated
with the given of the senses and regarded as the input required by imagination and
understanding, namely, "distinct perceptions met with in the mind scattered and
single in themselves" (A120). 16 Since, for Kant uniquely, this "given" is not in fact
given at all but synthetically made, his attribution of apprehension to imagination
requires a transformation in one's notion of imagination that enables one to under
stand it-as Hume or anyone before him could not-as extending all the way down
to perceptions themselves.
One should not suppose that Kant's claim to be the first to recognize that appre
hension of a manifold in intuition is the doing of imagination was based on any
thing he introspectively observed in human mentation that somehow escaped his
predecessors. It derives not from empirical scrutiny but transcendental reflection
on the a priori conditions of the possibility of a consciousness in which sensations
are united, insofar as such a consciousness is required for the cognition of com
plex, sense-divide transcending objects through sensations (e.g., the apple cognized
through visual, tactual, and other data). In particular, the possibility of cognition
that appearances must be synthesized to exhibit them because there can be no unity
of sensibility by means of the synopsis of sense (sensations, sense affections) alone.
By exhibiting sensations in the matter of appearances, the synthesis of apprehension
therefore makes it possible for sensations to participate mediately, at one remove, in
the unity of sensibility made possible by the form of these same appearances (this
is the second part of the reciprocal mediation first mentioned in Section B and con
sidered without regard to synthesis of imagination in Section D). Given that it has
now been established that the imagination's synthesis of apprehension is the means
whereby the unity of sensibility is effected, the exhibition of sensations via the mat
ter of appearances must consequently be understood as one of the imagination's
original and primary functions in this synthesis (i.e., the imagination synthesizes
appearances as regards their matter no less than their form). For though synthesis of
apprehension is likewise responsible for our ability to represent the form of appear
ances independently of their matter in pure formal intuition, the need actually to do
so does not arise originally in intuition, but only subsequently, in connection with
the higher syntheses of imagination and judgment responsible for converting the
"scattered and single" manifold generated by the synthesis of apprehension into the
objectively ordered and related manifold on which the possibility of cognition and
its objects depends (Parts IV-V). The upshot is that sensations and the matter of
appearances are two completely different, independent manifolds in a deeper sense
than was evident in Section B: the latter is a product of imagination and expression
of spontaneity, the former a given of receptivity. Accordingly, it should now be even
clearer that the qualities of the one (determinable objectively as intensive magnitude
and modal actuality) cannot be qualities of the other (visual, tactual, etc., qualities
present in subjective perception). All thought of the (objectively determinable) mat
ter of appearances must therefore proceed via concepts derived not from sensation
but the synthesis of apprehension in intuition (Part V).
In addition to immediacy, Kant commonly distinguished intuitions from other
representations by their individuality. Since discursivity, for Kant, is representation
by means of universals, his insistence in the Transcendental Aesthetic that pure
space and time are prediscursive (A24-5/B39 and A31-2/B47) means that they,
along with everything in them (all appearances in respect to form) do not involve
universals in their intuition. So, if individuality is understood merely negatively, as
representations without universality, then all intuitions and appearances ipso facto
count as individuals in apprehension (Kant's account of the advent of universality
will be examined in Chapter 9). They also count as individuals in the positive sense
of a representation that contains its manifold within it, as parts, rather than under
it, as instances (#3-#4 at A25/B39--40 and #4-#5 at A31-2/B47-8; also B136n, dis
cussed in ch 5-B). What intuitions and appearances are not, given only the imagina
tion's purely aesthetic, prediscursive synthesis of apprehension, are individuals in
the senses defined by the logical functions and categories of quantity (chs 10-B and
11-A) or the properly objective senses defined by transcendental synthesis intellec
tualis (chs 11-C and 14-B) and synthesis speciosa (chs 13-H, 14-D, 15, and 16-C).
{ 4}
Pure space differs from pure time not only in quality, as juxtaposition (nebenein
ander, iuxta se invicem positorum) from succession (nacheinander, post se invicem
positorum), but also with respect to the data through which the appearances it
makes possible are intuited: space is the pure intuition whereby appearances cor
responding to the data of sensation are synthesized, and so is the basis for the
intuitive representation of objects outside the mind (the form of outer sense), while
time is the pure intuition responsible for the synthesis of appearances through data
of self-affection, and so the basis for the intuitive representation of mentation itself
(the form of inner sense). In this chapter, I will examine how the difference in the
data pure space and time exhibit in intuition both distinguishes the contributions
each makes to the unity of sensibility and shows how the two are yet able to merge
to constitute a single sensible consciousness in which all possible data of the senses,
self-affections no less than sensations, are united.
That pure space, the form of outer sense, and its manifold are presupposed by
pure time, the form of inner sense, and its manifold seems clear. As shown in
Chapter 3-D, if there were not already a synchronic unity of consciousness produced
by exhibiting the data present in sense as a homogeneous manifold of appearances,
all contained in a single representation, there could be no diachronic unity of con
sciousness encompassing a succession of such manifolds; and since "simultaneity is
not a mode of time itself, as in time no parts whatever are simultaneous, all are suc
cessive" (A183/B226), synchronic unity of consciousness is possible only by means
of a non-temporal form of sensibility which, in beings constituted like ourselves,
is pure space. Correspondingly, pure time is the form of sensibility that enables
us to unite all possible such synchronic unities exhibiting data of sensation (i.e.,
all syntheses of apprehension) as the homogeneous manifold of a single intuitive
consciousness a priori (along with data exclusive to inner sense: Sections C and
D). The result is a diachronic unity of sensibility in which all synchronic unities of
Unity of Sensibility (2 ): Space and Time 107
Further respects in which time depends on space include: (1) space makes it possible
to form an image of time (A33/B50, Bl54, B156, B292), to construct time (AA 18 §
6312 [1790), and to represent time as an object (B291-3, ID 405); (2) without space,
we could never understand the conformity of objects to the categories (B291-3),
especially in respect to their existence in time ("The space that represents the outer
precedes the possibility of time determination," AA 18 § 6312 [1790]); (3) space is
presupposed by the object of inner cognitive experience, the temporally enduring
self (B275-9); and (4) space is presupposed in the representation of number by suc
cessive counting of units (ID 397-8, AA 17 § 4629, and AA 18 § 6314).
Yet in all of these cases, the role that Kant accorded to space involves transitions
of thought (relations) and so presupposes a succession of representations, whereas
our interest in space at present is strictly limited to the unity of sensibility that pre
cedes and makes possible such succession: the manifold united in pure space prior
to and independently of pure time, a manifold not yet temporal in any way, succes
sion included. In other words, the question here is not whether sensibility requires
space as well as time, but rather the reverse: why time (succession) is required as
well as space (juxtaposition). For even if one grants that cognitive experience and its
objects are impossible given space alone, this does not tell us why sensibility itself,
and in particular its prediscursive unity, requires succession. And the absence of any
compelling, purely transcendental aesthetic reason for this would, at the very least,
put in question Kant's claim that time is a constitutive form of all appearances as
such, merely as given (apprehended) in intuition, and so prior to and independently
of the syntheses of these appearances by the higher operations of imagination and
understanding responsible for conferring temporal meaning on the categories and
concepts of objects generally.2
The reason succession is essential to the unity of sensibility may be thought
obvious: because the mind is not a spatial object, its actions and passions can only
be given in intuition as a succession, and so only in time. But is this not a matter
of empirical fact, and so inadmissible as a transcendental basis for admitting time
as a second pure form of sensible intuition? Even if its transcendental bona fides
are conceded, matters are still neither so simple nor so straightforward. Since the
mind ( Gemut), considered as a thing in itself, is, for Kant, no more temporal than
it is spatial, its nature in no way necessitates that its appearance to sensibility be
apprehensible in inner intuition conformably to time rather than in outer intuition
conformably to space or intuition of some other kind ("that something which, con
sidered as noumenon (or better, as transcendental object), underlies outer appear
ances and so affects our sense that it obtains representations of space, matter,
shape, etc.-might yet at the same time also be the subject of thoughts," A358).3
Nor is there anything perceived in the affections of inner sense that precludes their
exhibition in intuition by outer appearances apprehended in pure space rather than
inner appearances in pure time. Since space "cannot itself be sensation" (A20/B34),
all qualitative resemblance is severed between the form of spatial appearances
and the quality of the affections of the outer senses to which the matter of these
2 T hat Kant accorded transcendental and not just empirical or mathematical validity to the
distinction between space and time is clear: (1) time is singled out from space for a special role in
the Transcendental Logic because it holds a priori not just of some appearances but all (A98-9
and Al38-9/B 177-8); (2) the sensible difference of space from time is essential to the Refutation
of Idealism (B274-9); and (3) the General Note on the System of Principles makes clear that time
must be supplemented by space if any of the categories are to have objective validity (B291-3).
My question here is simply whether there is any purely aesthetic transcendental philosophical
basis for supplementing space (juxtaposition) with time (succession), which, in the present con
text, is to ask why the solution to the problem of the unity of sensibility, ahead of all higher
operations of imagination and understanding, requires succession in addition to juxtaposition.
3 See also A34-6/B51-2, B157n, A276-7/B332-3, A278/B334, B422, B422-3n, and AA 23 31
E LXXV (marginal notation at A277).
Unity of Sensibility (2 ): Space and Time 109
appearances corresponds (ch 3-B and -E). This means not only that there is nothing
intrinsically spatial about visual or tactual sensations and that these sensations bear
no more qualitative resemblance to outer appearances than auditory, gustatory, and
olfactory sensations do, but also that nothing pertaining to the quality of the affec
tions of inner sense (affections coinciding with the spontaneity of imagination and
thought, emotions, desires, etc.) precludes the possibility of their exhibition in intu
ition by means of outer appearances. For if the absence of all resemblance to space
in the quality of gustatory and olfactory sensation does not prevent them from
counting as data through which objects in space can be cognized (Anthropology §§
16 and 20), it is by no means obvious why the same could not be true of the data
that furnish the raw material from which cognition of thinking beings and their
states is obtained. On what transcendental ground, then, did Kant preclude the
possibility of exhibiting the affections of inner sense through the homogeneous
manifold of spatial appearances in outer intuition?
The key to answering this, I believe, is Kant's view that "the representations of
the outer senses constitute the actual material with which we occupy our mind"
(B67) so that "without objects in space, there would be no empirical representa
tion at all" (A492/B520). This suggests a distinction between, on the one hand, the
affections of the outer senses (sensations) together with the actions of imagination
(syntheses of apprehension) that exhibit these data in outer intuitions as spatial
appearances and, on the other hand, the actions directed upon these appearances,
taken as given, that reflect on them in various ways (discernment, recollection,
comparison, etc.), together with the self-affections that coincide with these exer
cises of reflective spontaneity. In other words, affections of the outer senses can
be distinguished from those of inner sense as the prerefiective manifold given for,
and so presupposed by, reflective spontaneity from the manifold that arises in and
through exercises of reflective spontaneity. And since the former cannot be given
for reflection unless and until a pure intuition of space produced by the imagina
tion's pure synthesis of apprehension makes it possible to exhibit these data as a
single, homogeneous manifold united in one consciousness, the prereflective unity
of sensibility constituted by pure space must be regarded as preceding and making
possible reflective spontaneity in general and, more particularly, the pure intuition
of time whereby the self-affections coinciding with these acts of exhibition are, in
their turn, exhibited as a homogeneous manifold of inner appearances. Thus, at
the very most elementary level of sensible representation, space, the pure intuition
responsible for the unity of sensibility ahead of all reflection, must precede time,
the pure intuition responsible for the unity of sensibility comprising the affections
given in and through reflection on the prereflective spatial manifold "with which we
occupy our mind."
If this way of understanding Kant's transcendental distinction between space
and time at the most basic, purely aesthetic level is correct, it follows that, in beings
constituted like ourselves, space must be understood as the form of outer intuition
responsible for uniting affections of sense that count as "outer" because they relate
110 Kant's Anatomy of the Intelligent Mind
to what exists prior to and independently of the reflecting subject, and time as
the form of inner intuition responsible for uniting affections of senses that count
as "inner" because they arise in and through the representational operations of the
reflecting subject. The two senses of "outer" distinguished by Kant at A373 say it
all: "the expression 'outside us' brings with it unavoidable ambiguity, in that it may
refer to what exists distinctly from us as thing in itself and to what belongs merely to
outer appearance," and needs to be disambiguated by distinguishing "empirically outer
objects from those that are outer in what might be termed the transcendental sense by
explicitly terming them [i.e., empirically outer objects] things that are found in space"
(A373). In the present context, this means that data of the outer senses, considered
apart from their exhibition by outer (i.e., spatially formed) appearances, are "outer"
because they involve only affection by things in themselves (ch 8) and so are present in
sense prior to and independently of the self-affections coinciding with acts of sponta
neity (including those responsible for synthesizing the spatial appearances themselves).
If the appearances coinciding with these data (sensations) were synthesized in confor
mity with a form other than space, they would still be "outer" in the transcendental
sense but not the empirical, properly spatial sense ("things that are found in space").4
And from this the following emerges as the fundamental transcendental aesthetic
ground of not only the difference between space and time but also the different tran
scendental aesthetic requirements to which each is a response: space exemplifies pure
intuition of the kind specific to affections capable of presenting themselves in the mind
prior to and independently of reflective spontaneity (data of the outer senses, sensa
tions), while time exemplifies the kind specific to affections whose presence in the mind
is occasioned by spontaneity itself (data of the inner senses, self-affections).
Before utilizing this proposal to elucidate the complementary contributions of
space and time to the unity of sensibility, it may be helpful to consider it in more
detail. The existence given with affections of outer sense is the thing in itself; it
is this that ultimately corresponds to the matter of outer appearances and is the
reason they count as outer in the transcendental sense ("that which corresponds
to sensation in [appearances] is the transcendental matter of all objects as things
in themselves (facticity [Sachheit], reality),"5 Al43/Bl82). The existence given with
affections of inner sense is the mind ( Gemut), or subject, in itself ("the subject in
itself . . . must not merely perceive sensations within itself but act upon and connect
4 Because "inner" is not a synonym for "temporal" as "outer" is for "spatial," this ambigu
ity is avoided in the case of data of inner sense; they are "inner" in the transcendental sense of
involving self-affection and so remain transcendentally "inner" whether the appearances that
exhibit them in intuition are temporal, as in beings constituted like ourselves, or something else
(cf. A36-8/B53-4).
5 In the Kemp Smith translation of the Critique, a "not" is inserted at A143, reversing the
sense. There is no textual justification for this, nor is the point expressed in the unaltered version
unique in Kant's writings, e.g., "That (in representation) which is related to the object of the
senses in itself ( Gegenstand der Sinne an sich) is sensation" (AA 18 § 6314 [1790-1 ). See Chapter 8
for detailed analysis of the thing in itself.
Unity of Sensibility (2 ): Space and Time 111
Though reflection of a very primitive kind might still be possible in the absence of
the unity of outer sensibility effected through pure intuition, it would fall far short of
the kind that can eventuate in cognitive experience. The latter involves a capacity to
represent a single sense-divide transcending object and determine it through data of
6 For an examination of Kant's notions of the representing subject, the subject in itself, the
the five senses, all functioning together as a unified multisensory outer sense.9 But no
such object could be represented without pure intuition since there would then be no
formally homogeneous manifold of appearances apprehensible in intuition through
all possible sensations, however heterogeneous (ch 3-D). For example, without homo
geneous appearances to exhibit their data, no amount of comparison of the qualities
present to sight and touch could reveal anything common between them, such that one
and the same sense-divide transcending object might be given through them. Nor can
the thing in itself make good the want for such an object since what cannot be given in
representation at all, through sensation intuition or concept, is necessarily "nothing to
us" (A105; ch 13). Hence, given sensations alone, without the condition whereby alone
homogeneous manifolds of appearances can be apprehended immediately together in
intuition through them, the outer senses could not operate as diverse modes of a single
outer intuition in which one and the same sense-divide transcending object is able to
present itself, and so none of the kinds of reflection employed in the cognition of such
an object would be possible at all.10
Reflective spontaneity can eventuate in cognition only if preceded by a pure
intuition specific to data of the outer senses that can bring about the unity of all
possible sensations in one and the same consciousness completely a priori. In beings
constituted like ourselves, the pure intuition that brings about the unity of outer
sensibility is space. Recalling that Kant deemed receptivity by itself incapable of
yielding such a unity, it should also be clear that spontaneity must contribute to its
production by synthesizing both pure space (through pure synthesis of apprehen
sion in intuition) and the appearances met with in it (through empirical synthesis
of apprehension in intuition) (ch 3-E). This means that, strictly speaking, neither
9 In humans, vision and touch are more useful for this purpose than the other senses
(Anthropology§§ 16--21), but from a transcendental perspective, there is no reason there could
not be creatures capable of cognitive experience that, like bats, depended more on sound than
vision or touch or that depended on senses not found in any terrestrial creature, extinct or extant.
10
Once one opts to treat Kant's transcendental aesthetic as a theory of sensibility as such,
without regard to its determination by categorial understanding, the natural place to begin is
with the need to explain how the various faculties of sensation are able to function as a single uni
fied multisensory outer sense through which sense-divide transcending objects can be perceived
(apprehended) and experienced (cognized). T his purely aesthetic explanatory demand arises as
a direct result of Kant's endorsement of the anti-abstractionist separability principle he received
from Berkeley via Hume (ch 2-C). I urge readers to (re-)familiarize themselves with my treatment
of the topic in UU chs 14 and 18. To recap briefly, Berkeley used the principle to show that visual
and tactual data have absolutely nothing in common to qualify them as spatial senses and so
nothing to unite them as a single outer sense with a single sense-divide transcending object. More
generally, the separability principle shows that there is nothing in our visual, tactual, olfactory,
gustatory, and auditory sensations of apples that "bundles" them together as a unified multisen
sory outer experience of a single sense-divide transcending spatial object. Berkeley addressed the
problem by developing the first philosophically sophisticated, explanatorily potent association
ism. Hume then perfected the account by further shifting the burden of spatial representation
from sensation (touch) to imagination. Where the associationist account of outer sense and its
spatial objects falls short is its inability to explain how the data of all of the senses can be repre
sented immediately together as a single homogenous manifold right in sensibility itself, which is
a precondition for their association.
Unity of Sensibility (2 ): Space and Time 113
space nor spatial appearances are ever present to us in perception (providing the
raison d'etre of transcendental judgment: ch 16-B). They exist only in and for our
consciousness of the intrinsically non-spatial data of the outer senses (sensations)
insofar as the myriad differences met with in them are exhibited in imagination and
(prediscursive) understanding under the form of space as a homogeneously differen
tiated manifold of appearances.
Imagination for Kant is "the capacity of intuitions or exhibitions" (CJ 287; also
190, 232, and 292) and is as much a part of sensibility as the outer and inner senses
themselves: "the synthesis of imagination, although exercised a priori, is neverthe
less in itself always sensible" (Al24) so that "[s]ensibility in the cognitive capacity
(the capacity of representations in intuition) contains two components: sense and
imagination" (Anthropology§ 15; also A78/B104, AA 15 § 229, AA 28 473). More
particularly, imagination is "the capacity to represent an object even without its
presence in intuition" (BISI). As an object that is never present in sensation, space
and its manifold (so too time and its manifold) are ipso facto present in intuition
only insofar as they are synthesized in imagination, which consequently is, in this
unique regard, productive and not merely reproductive:
u Also: "One can and must concede that space and time are mere thought entities and crea
tures of the imagination. But because they are the essential form of our sensibility and the
receptivity of its intuitions whereby in general objects are given to us, and because the universal
conditions of sensibility must at the same time necessarily be a priori conditions of the possibil
ity of all objects of the senses as appearances and so agree with these, they are not fictitiously
invented by the imagination but underlie all its compositions and creations" (Discovery 203).
And: "Space and time are of course not objects of intuition but merely its subjective forms. They
do not exist apart from representations and are given only in the subject; i.e. their representa
tion is an act of the subject itself and a product of the imagination for the sense of the subject.
Yet, their representation is the cause of the object in appearance (phamomenon), not derived
(reprcesentatio derivativa) but original (originaria)" (AA 22 76; also 473). See KMM Part I for
analysis of these and other texts with similar purport. Evidence from the Critique of Pure Reason
itself will be my focus in the next chapter.
114 Kant's Anatomy of the Intelligent Mind
represented that are not both pure and original is necessarily produced by that fac
ulty, not reproduced by it. It is thus in terms of a species of imagination that no one
previous to Kant had identified that one must understand his account of the origin
of pure intuitions generally and pure space and time in particular.
What needs to be accounted for in the case of pure space is the "the mere pos
sibility of outer appearances so far as they either exist in themselves or can still be
added to given appearances" (A429/B457n). More precisely, pure space is the "mere
possibility of compresence (Beisammenseins)" (A374--5n) and so "the intuitively
given possibility of universal coordination" (ID 407). And more precisely still, it
"is not something existing as an object of intuition (no more than time is) but
the mere form for putting together (Zusammenstellung) the manifold next to . . . one
another" (AA 22 435) and so "as not merely distinct but in distinct places" (A23/
B38). Finally, as purely formal, and so exclusive of all consciousness of sensations,
its pure intuition is a consciousness solely of the synthesis whereby outer (i.e., spa
tial) appearances are generated ("Space is itself a synthesis a priori," AA 18 § 5876
[1783-4]).
Since this synthesis of outer appearances is without limit (A25 #5 and B39-40
#4), pure space suffices for the exhibition (in the form of a homogeneous mani
fold of spatial appearances) of the inexhaustible infinity of sensational differences
whereof receptivity is capable, however heterogeneous in quality and irrespective
of the limits inherent in a creature's species-specific capacity to discern, attend to,
or recollect sensational differences ("Space is infinite because the capacity to admit
many impressions of outer things, or receptivity, has no limits in itself," AA 17 §
4373 [1773-5]). For example, although human visual and tactual powers of dis
cernment vastly exceed our other senses, they still remain insufficient to discern all
of the myriad super-fine-grain differences present in visual and tactual affection,
which accordingly remain irremediably obscure. Yet obscure consciousness is still
consciousness, and, in this obscurity, many sensational differences may exist:
sensations (ch 3-B and -D), spatial appearances ensure that all possible data of all
possible outer senses belong a priori to one and the same original a priori synthetic
unity of outer sensibility.
Of course, unlike the Newtonian sensorium, there is nothing redolent of divinity
about Kantian pure space. This product of pure apprehension in imagination not
only is not objective in itself, it also is not objective in the sense required for math
ematics, science, or cognition generally. Like the perceptions "met with in the mind
scattered and single in themselves" (Al20) of empirical apprehension, the manifold
of pure space is devoid of all differentiation and determination:
To understand the sense in which space (and, by extension, time as well) is so "uni
form" and "indeterminate" as to be altogether devoid of that "wealth of meaning"
commonly supposed to lie in it, one must be careful to abide by the strictures of
Kant's psychologism (ch 2-E) and not attribute to this space anything more than
sensibility alone-prior to and independently of "the conditions of the synthetic
12
So too for geometrical axioms: "The principle that a straight line is the shortest between
two points presupposes that the line is subsumed under the concept of magnitude which is cer
tainly no mere intuition but has its seat in the understanding" (PFM 301). See also Bl53-4 dis
cussed in ch l 4n9.
Unity of Sensibility (2 ): Space and Time 117
can I conceive how anyone could. The unity of sensibility clearly does not require
that pure space be determinately fiat hyperbolic or elliptical, three-dimensional or
ten-dimensional or any other number of dimensions, Ricci-fiat or Ricci-curved,
etc. Indeed, I do not see why the prediscursive operations that Kant credited with
producing this unity would or could incorporate any mathematical, categorial, or
empirical differentiation or determination of any kind. If present in his theory of
sensibility, where aesthetic grounds are the only kind admissible, they could only
be there as unjustifiable free riders (ch 2-E). Thus, apart from supplying "the con
dition of the possibility and of the manifoldness of [outer] intuitions," the pure
space yielded by sensibility must be treated as completely undifferentiated ("uni
form") and indeterminate ("in respect of all particular properties"); and it falls to
the understanding to make good this want.
The general Kantian principle operative here is this: if the psychologistic
method he adapted from Hume means anything at all, it is that only so much and
no more content may be ascribed to a representation than its originating faculty is
capable of supplying. We need to let ourselves be guided by this principle toward
a conception of the pure space of sensibility as supplying a "manifold" only in
the barest sense of word and nothing more: no order, no connectedness, no parts
making up wholes, no space objectively and determinately above or below another,
inside or outside another, here or there, no objective determinate spatial relations
of any kind-just sheer, purely aesthetic spatial outsideness Uuxtaposition) as
such. Recognizing this is, in my view, absolutely essential to comprehending the
significance of the Transcendental Aesthetic for the remainder of the Critique of
Pure Reason. For only by grasping how very little Kantian sensibility is able to set
in place ahead of the understanding can one hope to appreciate the hugeness of
the void the latter is obliged to fill by means of its pure concepts. The role of the
categories in constituting the fully differentiated, determinate pure (formal) space
presupposed in pure mathematics will be examined in Chapters 13-15, while their
role in constituting the fully differentiated, determinate (material) space (and time)
of experience will be the principal topic of Part V. Here, it suffices to recognize that
the problem Kant's theory of the understanding was crafted to address is entirely
self-created: because space is a pure intuition originating in sensibility sufficient
only for a bare manifold of representations ("scattered and single in themselves"),
the understanding, through its pure concepts, must become "itself the creator of
the experience in which its objects are found" (B 127).13
13 I have adapted the formula "uniquely differentiated from and completely determined in
relation to" from A58 l-2/B609-l 0, a text in the Transcendental Ideal chapter that Kant expressly
related to the Transcendental Analytic. Since the principle of complete determination in it relates
most directly to the category of community (the category underlying the Transcendental Ideal),
I consider the cited text in the course of my treatment of that category's application to the mani
fold of sensibility (ch 15-D).
Unity of Sensibility (2 ): Space and Time 119
If Kant was clear about anything, it was the importance of not confounding
the activity of spontaneity with the inner intuition (perception, appearance) of this
activity as temporally unfolding (B68-9, Bl52--4, Anthropology 396--9, AA 18 § 6349
and § 6354). The former is not present to consciousness in any intuition at all but
only as an "indeterminate perception [which] here means only something real that
is given, given indeed only to thought in general, and so neither as appearance
nor as a thing (Sache) in itself (noumenon)" (B422-3n). Or, again: "I do not have
another self-intuition that gives the determinative in me," but instead have to make
do with an awareness "only of its spontaneity," and so must always take care not
to mistake the manifold of temporally distinguishable reflective acts apprehended
in inner intuition for such an intuition (of the determinative in me): "the manifold
belonging to [my existence] ...belongs to a self-intuition that has an underlying
form, time, given a priori, which is sensible and belongs to the receptivity of the
determinable" (B 157-Sn). The consciousness of ourselves as spontaneity is a mere
thought in which no manifold is apprehended whatsoever.14 And the implication
of this is clear: given no pure intuition but space, plus the lack of any intuition of
the activity responsible for synthesizing that intuition and its manifold, the acts of
apprehension responsible for producing outer appearances could not themselves be
apprehended as a manifold in intuition at all, and it would be impossible even to
represent them as a manifold, much less subordinate them in the manner requisite
to order them.
To understand how sensibility is able to make good this want, one must start
with self-affection. Like his sensibilist predecessors, Kant held that sensibility is
equipped to be affected not only by realities (transcendentally) external to the mind
but by the mind (das Gemiit) itself as well. Unlike them, however, Kant regarded
imagination as a necessary ingredient of perception itself (Al20n). Since this is just
to say that perception is as much an act (the spontaneity of the synthesis of appre
hension in intuition) as it is an affection (the receptivity of the synopsis of sense), it
is self-affecting. Exercises of the synthesis of apprehension in outer intuition thus
coincide with data of inner sense, including the pure synthesis of apprehension
responsible for generating pure space: "Space and time are products (but primitive
products) of our own imagination; hence, they are intuitions created by the subject
affecting itself" (AA 22 37).
As purely receptive in nature, these self-affections are distinct both from the
spontaneous acts of apprehension with which they coincide and the spatial appear
ances produced by those acts. They consequently fall entirely outside the scope of the
unity of sensibility effected by pure space. Nor can they be supposed to define a new
14 Kant stated on numerous occasions that spontaneity is a consciousness that offers no mani
fold and so no material from which to form the concepts we would need in order to fashion judg
ments about the mental acts that represent them as they are and not merely as they appear: B135,
B l38-9, A340/B398, A345-6/B404, A354-5, A381-2, B422-3n, A443/B471, A784-5/B812-13,
Anthropology§ 7 (including deleted portion at AA 8 396-9), letter to Tieftrunk April 5, 1798.
Unity of Sensibility (2 ): Space and Time 121
unity of inner sensibility in their own right. Interpreters15 who suppose otherwise
most likely do so in the belief that, like his predecessors, Kant ascribed existence in
time directly to the data of inner sense, holding that self-affections-indeed, affec
tions generally-are themselves received in inner sense successively (rather than
merely the inner appearances produced to exhibit them in inner intuition). One obvi
ous problem with this assumption is how Kant could have supposed self-affections
to exist successively in time without also supposing the same of the acts whereby the
subject affects itself. Yet he could hardly have been more explicit in this regard: "the
subject in which the representation of time originally has its ground cannot deter
mine its own existence in time by means of that representation" (B422).16 No intui
tive determination of myself as active means just that: "if I do not have still another
self-intuition that gives the determinative in me (I am conscious only of its spontane
ity) prior even to the act of determining, as time gives the determinable, then I cannot
determine my existence as that of a self-active being but only represent the spontane
ity of my thought, i.e. the determining, and my existence always remains only sensi
bly determinable, i.e. as the existence of an appearance" (Bl57-8n). Moreover, any
supposition that either self-affections and/or the acts with which they coincide exist
successively in time seems incompatible with any straightforward reading of Kant's
insistence that "if I could intuit myself, or another being intuit me, without this
condition of sensibility [i.e., the pure time intuition], the very same determinations
we now represent as [successive] alterations would yield a cognition in which the
representation of time, and so too that of alteration, would not occur at all" (A37/
B54). Although Kant could not affirm the possibility of a differently constituted
sensibility, one in which inner intuition was preceded and made possible by a pure
intuition other than time, he could not deny it either (A42/B59, B72, Bl39, Bl46,
A230/B283, A613--4/B641-2, PFM 350-1, Progress 267, Anthropology 399, AA 18 §
5056 and§ 6312). And that alone is enough to show his doctrine of sensibility treats
succession in time as intrinsic neither to the acts performed by the subject nor the
affections present in its receptivity, but instead as an appearance it synthesizes in
order to exhibit them as a manifold in inner intuition-just as spatial appearances
are synthesized to exhibit sensations in outer intuition.
The view that the acts and affects of the subject do not, and indeed cannot, exist
successively in time is regarded by many as patent nonsense on the semantic ground
that the only way anything can be understood either as an "act" or as a sensory
"affection" is by representing it as an occurrence (action, event) in time; and since
no philosopher of the caliber of Kant would build his philosophy on obvious non
sense, they deem any interpretation of his words that avoids this consequence, however
15 See KMM Part I for discussions of examples. There does not seem to be any significant
change in the consensus view since KMM was written.
16 Also: "The principle of permanence does not concern things in themselves, hence the sub
ject of the representations of things as itself, i.e. apperception, but only appearances. For the
concept of time does not apply to anything else, not even to the subject of time itself" (AA 23 E
LXXV, p. 31, at A277).
122 Kant's Anatomy of the Intelligent Mind
stretched, preferable to one that does not, however well supported by the text. In the
face of such resistance, it is particularly important to be clear about what is and is not
here at issue. There is no dispute that Kant held that inner appearances, as given imme
diately (apprehended) in intuition, are successive. Nor is there disagreement that synthe
ses of imagination and understanding higher than apprehension take this succession for
granted (no one questions, for example, that the succession of mental acts and affects
apprehended in inner intuition is the subjective succession that Kant contrasted with the
objective modes of time at issue in the Analogies of Experience). Where I part company
from most interpreters is on a purely aesthetic matter specific to Kant's doctrine of sen
sibility as such, prior to and independently of those operations of reflective spontane
ity that take the outputs of the synthesis of apprehension in inner intuition-temporal
appearances-as their input. The question is whether Kant held that mental acts (the
determinative in me) and/or the self-affections that coincide with these acts exist in time
in themselves, and are in themselves successive or whether, on the contrary, he held that
these intrinsically non-temporal elements of representation are merely exhibited by a
formally homogeneous manifold of successive appearances synthesized (apprehended)
by the imagination in pure and empirical inner intuition. In other words, it is the ques
tion whether or not he distinguished the matter of the appearances of inner intuition
from self-affections (data of inner sense) in the same way he distinguished the matter
of spatial appearances from sensations (data of outer sense). And it seems to me that
the same strictly aesthetic considerations that oblige one to conclude that Kant did not
equate the matter of spatial appearances with sensations, and so did not objectify olfac
tory, etc., data as occupants and containers of space, also lead to the conclusion that
he identified the matter of temporal appearances neither with self-affections nor the
acts responsible for them, and so did not treat either of the latter (any more than he did
sensations) as occupants and containers of time (ch 3-B).17
17 The contrary view partakes of the same incredulity that led Kant's contemporaries to object
in the manner described at A36-7/B53-4. Although Kant repeatedly insisted that there is precise
parity in his theory between time and space-the temporal and the spatial (e.g., A38/B55, A370- l ,
A492/B520, PFM 337, cited i n c h 7-A}-it seems t o many t o fl y i n the face o f logic. Kant was
well aware of the problem, attributing it to a failure to distinguish appearance (Erscheinuung),
in its transcendental sense (ch 3-B), from illusion (Schein): "When I say that intuition of outer
objects as well as self-intuition of the mind represent both in time and space, it does not mean
that these objects are a mere illusion. For, in appearance, objects, nay even the constituent quali
ties (Beschaffenheiten) we ascribe to them, are always to be regarded as something actually given,
only that, insofar as this constitution depends only on the mode of intuition of the subject in the
relation of the given intuition to the subject, this object as appearance is to be distinguished from
itself as object in itself. I thus do not say that bodies merely seem (scheinen) to exist outwardly to
me, or that my psyche only seems to be given in my self-consciousness, when I maintain that the
quality of space and time in conformity to which, as condition of their existence, I posit both bod
ies and my soul lies in my mode of intuition and not in these objects in themselves. It would be my
own fault if I made what should be accounted appearance mere illusion" (B69; also Anthropology
398 and AA 18 § 6349). I will return to the topic of Kant's insistence that treating the succession
apprehended in intuition as mere appearance and nothing real in itself be distinguished from treat
ing it as an illusion, with the former being the very antithesis of skepticism, in Chapters 7 and 8.
Unity of Sensibility (2 ): Space and Time 123
In the case of sensations, unity of sensibility was found to require a pure intu
ition of space whereby sensations, howsoever heterogeneous, can be exhibited by
formally homogeneous spatial appearances, all capable of standing immediately
together within the unity of pure space. Are self-affections any less heterogeneous
than sensations? Given that they coincide with the full range of mental actions and
passions, they do indeed seem to be as myriad in their variety as sensations: being
affected by anger is qualitatively different from being affected by fear, being affected
by fear different from amusement, amusement from lust, lust from envy, envy from
recollection, recollection from a shift of attention, a shift of attention from making
a comparison, comparison from association, association from concept formation,
concept formation from judgment, judgment from inference, inference from feeling
respect for the moral law, respect for the moral law from aesthetic pleasure, and so
on and on. Moreover, their independent variability would seem to give us just as
much reason to speak of several inner sensory fields as the independent variability
of touch, vision, etc. gives us to speak of several outer sensory fields (and Kant did
indeed assert, on at least one occasion, that "we have several inner senses," AA 15 §
224 [1783--4]).18 Since homogeneous differentiation is therefore no more to be met
with (perceived) in self-affections a posteriori than it is among sensations of differ
ent sensory fields (visual, tactual, auditory, etc.), inner sense clearly stands in the
same need as outer for a means of representing all of its data, possible no less than
actual and regardless of human species-specific sensory and psychological limita
tions, as a single homogeneous manifold a priori, all contained in one and the same
purely aesthetic consciousness, or unity of inner sensibility.
As with outer sense, this demand can be met in no other way than by an appre
hending imagination equipped to produce a pure formal intuition in and through
which all self-affections can be exhibited by a manifold of homogeneous appear
ances immediately within that intuition (as an individual, rather than under it
as with a universal). For if self-affections are to be apprehended as a manifold,
the same kind of reciprocal mediation is required as that whereby pure space
enables the data of the outer senses to be united in one consciousness (ch 3-B
and -D): self-affections united in one consciousness by means of the unity in a
single intuitive representation of the appearances apprehended in inner intu
ition by means of these same affections. In beings constituted like ourselves, the
pure intuition responsible is of course time. The features pertaining to it in this
capacity are no different from those of space. Time enables us to apprehend the
18
T his may seem to contradict Kant's assertion in Anthropology § 24 that there is only one
inner sense, but in the latter text, he seems simply to be claiming that there are not distinct bodily
organs of inner sense as there are distinct organs of outer senses, whereas AA 15 § 224 is con
cerned not with bodily organs but differences among the data of inner sense themselves. Today,
and still more so in the coming decades and centuries, a Kantian could point to regions of the
brain and/or subsystems of the connectome as distinct inner senses in much the way Kant him
self could point to eyes, nose, tongue, ears, and skin as distinct outer sense organs.
124 Kant's Anatomy of the Intelligent Mind
19 Also: "Time is in itself a series (and the formal condition of all series)" (A41 l/B438).
Unity of Sensibility (2 ): Space and Time 125
Pure time is the basis of Kant's psychology of outer and inner experience, both empiri
cal and transcendental (A98-9). It starts with the synthesis of apprehension, including
that whereby space-synthesizing acts of apprehension can themselves be represented
as a manifold and united in one consciousness by means of pure time. The following
text is of special importance in the latter respect because in it, Kant takes space and its
manifold as given and describes what happens from there:
Every intuition contains within it a manifold that yet would not be represented
as such if the mind did not distinguish the time in the succession of impressions
one after the next. For, as contained in an instant, every representation is never
anything else than an absolute unity. So, if unity of intuition is to emerge (werde)
from this manifold (as, say, in the representation of space), it is necessary to
run through this manifoldness first and then take this running through together,
which action I call synthesis of apprehension because it is directed immedi
ately upon intuition, which does indeed offer a manifold but cannot effectuate
(bewirken) it as such, and contain it in one representation, unless a synthesis takes
place. (A99)
To understand empirical apprehension correctly, one must not confound the running
through (Durchlaufen) and taking together (Zusammenhemung) described here with
the entirety of the threefold empirical synthesis of apprehension-reproduction-recog
nition that complements synopsis (A97, A99-104). For example, in order to reproduce
or recognize a manifold in space as a line, one must not only successively apprehend
the manifold but also relate predecessors to successors, whether subjectively (purely
aesthetically, nondiscursively), by preferentially combining (associating) the manifold
in the figure of a line in reproductive imagination alone, or objectively, by means of
the recognitive concept of a line that confers the necessity of a rule on the reproduc
tive synthesis (ch 13). By contrast, the mere apprehension of the manifold of the line
consists solely in running through a succession of inner appearances (whose matter
corresponds to the inner affections coinciding with the acts responsible for synthesiz
ing the spatial manifold of the line) and taking these inner appearances together as
a succession, without relating predecessors to successors through association or con
cepts, much less setting them in objectively necessary relations (through the schema
tized categories). Thus, in synthesizing inner appearances of its own outer-intuition
synthesizing activity, the apprehending imagination yields a bare manifold of inner
appearances immediately together in one consciousness and nothing more. 20
20
The confusion of apprehension with the entire threefold synthesis has a long history. In
anglophone Kant scholarship, it is exemplified at least as far back as the (still) highly influential
commentaries of Norman Kemp Smith and H.J. Paton (see KMM 17-18, 193-7). It tends to be
coupled with the notion that apprehension without reproduction is instantaneous forgetting-a
somewhat factitious reading of Kant's assertion that without reproduction, we would "lose the
126 Kant's Anatomy of the Intelligent Mind
would not appear as a succession in inner intuition and so, a fortiori, would not
yield cognition of successively altering states of consciousness). The subject in itself
that produces the time in and through which its mental activity is apprehended in
inner intuition as a temporal appearance can no more be identical with anything
the empirical self included-occupying or containing that time than a fantasizing
imagination can be its own figment (i.e., one of the fantasies it creates). The empiri
cal reality that Kant accorded to outer and inner appearances alike means nothing
more in the context of transcendental aesthetic than that they do not merely seem
to but really do occupy and contain space and time (A27-8/B43-4, A35-6B51-3,
B69-70, A374-5, Progress 269; ch 7).21 This "reality" is perfectly consistent with
Kant's insistence that (1) the space and time in which all appearances are contained
are themselves mere representations (A373-5), (2) these representations are prod
ucts of a synthesis of apprehension in pure intuition (A99-100), (3) as pure yet
original representations they could not originate otherwise than through produc
tive imagination (Anthropology§ 28), and (4) the unity of the manifold in them is
both synthetic and original (A107, B136n, B140, B160-ln; ch 5). So, while from
the perspective of empirical psychology the unfolding life of the mind apprehended
in inner intuition is no illusion and should be taken at full face value (A545-6/
B573--4 and Progress 269-70), regarded transcendentally, inner temporal intuition,
no less than outer spatial intuition, must be understood simply as the synthesized
product of the non-temporal, self-affecting activity required if our minds are to
satisfy the fundamental aesthetic precondition of cognitive experience: the a priori
preceding representations from the thought" (A l 02). Apart from Kant's explicit denial ("The
capacity to combine the sequence of its past states in one consciousness [is] memoria, not repro
ductive imagination," AA 15 § 339 [1780s]), one needs to remember that reproductive synthesis is
essentially the equivalent of Humean association, so that "not losing preceding representations
from the thought" means that there must be something that relates the perceptions concerned in
transitions of thought, and we do not merely advance from one to the next (apprehension) but
(either psychologically or via a concept-rule) relate them. Apprehended perceptions, by contrast,
"are met with in the mind scattered and single in themselves" (Al 20}-which of course is not to
say that they are forgotten! See also Chapter 13-B and the analysis of Hume on associative rela
tion in UU Chapters 16-18.
21
Any claim to the contrary would, it seems to me, fall foul of Kant's assertion that "the
subject in which the representation of time originally has its ground cannot determine its own
existence in time by means of that representation" (B422). This same consequence is implicit, in
my view, at A362-3 as well. See also note 16.
Unity of Sensibility (2 ): Space and Time 127
unity of all affections without exception in one and the same consciousness (unity
of sensibility as prediscursive synthetic unity of apperception).
It may be thought that this way of interpreting Kant's account of time and the
unity of sensibility implies an infinite (albeit non-temporal) regress insofar as acts
of time-synthesis result in affections that require new acts to apprehend them in
inner intuition, which in turn issue in new self-affections that require still further
acts to apprehend these in inner intuition, which result in new self-affections, and so
on. But even if that were so, so what? There is no psychological necessity to appre
hend every inner affection in intuition. For Kant, the focus of cognitive experience
is knowledge by means of the outer senses, not knowledge of this knowing via
inner sense ("the representations of the outer senses constitute the actual material
with which we occupy our mind," B67). Inner intuition is cognitively parturient
only insofar as it enables us to apprehend acts of spatial synthesis as a non-spatial
manifold of successive perceptions of the spatial, as the presupposition not only
for association in reproductive imagination and recognition in concepts but also
for their a priori counterparts, transcendental synthesis and the objective unity of
apperception. It would therefore not be long into the regress before the cognitive
loss resulting from a failure to exhibit data of inner sense as temporal appearances
became negligible.22
Pure time equips the imagination to apprehend series of every kind, infinite
regresses not excepted ("Time is in itself a series and the formal condition of all
series," A41 l/B438). There is no limit to the ability it gives us to exhibit new times
before, after, or between times already posited. As with spatial apprehension, one
must be careful not to confuse the limits of temporal apprehension with the thresh
olds of discernibility, attention, and recollection specific to human psychology.
Considered from the side of spontaneity, and so prior to and independently of all
temporal intuition, there is nothing to preclude an infinity either of apprehensive
views, attending to this or that spatial or temporal manifold, or the inner affections
consequent upon them. Thanks to the infinitude of pure time (A32/B47-8), each
and every one of these acts and affections can be apprehended in intuition as a suc
cession of appearances, even if that succession is so rapid as to transpire in a time
beneath the human threshold of temporal discernibility, as if the whole were but
a single, instantaneous appearance. Indeed, the time of pure sensibility is in itself
not only as inexhaustibly differentiable as the present human complement of inner
senses but also as differentiable as those senses would be if they were subject to no
limits of any kind, human or any other, if the data present in them were all perfectly
discernible (clear, not obscure, in the B414n sense) and if, in addition to our human
inner senses, we had perfectly discernible inputs from all other possible inner senses,
22
Since unexhibited data would not be represented within the a priori sy nthetic unity of con
sciousness presupposed by the I think, the I think would not be able to accompany them, with
the result that they would, quite literally, be "nothing to me" (B 131 ) . I will take up the issues of
the scope of the I think in Chapter 9.
128 Kant's Anatomy of the Intelligent Mind
none of which were subject to creaturely limits of any kind. There is therefore no
reason to suppose that Kant would deny-and, in view of his many affirmations
of obscure representation, much reason to believe he would affirm-that there are,
or at least could be, inner temporal appearances sufficiently complex to exhibit an
infinity of mental actions and affections packed into every discernible instant (times
within times within times without end).
If this was indeed Kant's view, it would not be unprecedented. Locke, for
example, maintained that it is common not to discern, or even be able to discern,
complex mental activity because of "how very quick the actions of the Mind are
performed: For as it self is thought to take up no space, to have no extension; so
its actions seem to require no time, but many of them seem to be crouded into an
Instant" (ECHU II/ix/§10). It is true that he did not apply this analysis to cases like
the perception of the movement of a bullet, which he described not as a succes
sion of internal perceptions too rapid to discern individually but as taking "up the
time of only one Idea in our Minds, without the Succession of another" (xiv/§10).
Yet rather than evidence of inconsistency, the semblance of conflict disappears as
soon as one takes into account the difference in context between ECHU II/ix
which concerns the nature of perception itself and how we confuse ideas that in fact
involve complex mental activity with simple, passive ideas of perception-and II/
xiv, in which the bullet example was part of Locke's case that succession and dura
tion are ideas of reflexion rather than sensation (UU ch 13-C). Since proving the
reflexive character of these ideas did not require that he establish the existence of a
succession of indiscernible perceptions of the bullet's motion in internal sense, the
introduction of such an analysis into ECHU II/xiv would only have complicated
matters unnecessarily, leading him to opt instead for the simpler, more straightfor
ward approach (no succession of ideas at all), without thereby precluding, much
less contradicting, the indiscernibility analysis of II/ix (UU ch 6). Granting, then,
that Locke never foreswore the view that a succession of mental acts can take place
so rapidly as to be indiscernible from one another, he was in a position, like Kant
after him, to allow that even in the briefest discernible time ("discernment" in the
II/xi sense), an indiscernible host of successive mental actions might be present
in internal perception ("perception" in the II/ix sense). And given that Lockean
abstraction is nowhere implied by this supposition, I see no reason to think that
Berkeley, Hume, or any other sensibilist would disagree.23
The real difference between Kant and empiricist sensibilist psychologists is that
on Kant's account of the inner, it is neither mental actions nor the inner affec
tions coinciding with these acts but only the imagination-generated appearances
(be they discernible or not) corresponding to them that are perceived (apprehended
in intuition) as a succession. The psychological significance of this becomes clearer
in light of its implications for the notion of a sensory field. In ch 3-D, I used this
notion to introduce the problem of the unity of sensibility: the visual, tactual, etc.,
fields, culminating in the need for a "super-field" in which the data of all the senses
can be represented as standing immediately together, the manifold of a single rep
resentation, united in one and the same consciousness. To do this, I deliberately
did not focus on the temporal dimension implicit in the notion of a sensory field,
although it is evident enough. Visual sensations, for example, can only be conceived
to constitute a "field" insofar as we suppose that the manifold of these sensations
changes, or at least can change, with time: where there is red now there may be
blue a moment's hence (say, by looking skyward), green after that (by shifting one's
gaze groundward), and so on. But if conceiving of a sensory field means being able
to represent sensations as a determinable manifold capable of varying in quality
over time, how must it be reconceived in light of the involvement of spontaneity
in the time pure intuition and the temporal appearances it precedes and makes
possible? The involvement of spontaneity implies that visual and other kinds of
sensations are no more a succession in themselves, prior to and independently of
synthesis of apprehension in inner intuition, than acts of reflective spontaneity,
self-affections, and things in themselves are-or, indeed, the spatial appearances
apprehended in outer intuition, ahead of the self-affections occasioned by these
apprehending acts and the exhibition of these affections by temporal appearances.
Since this precludes conceiving the existence of sensations in the synopsis of sense
as a coming to be in a sensory field, occupying positions in it formerly occupied by
others or changing their relative situations, the very notion of a field and, with it,
that of a diachronic unity of sensibility of any kind whatever, however attenuated,
cannot be supposed to have any application prior to or independently of the acts
of apprehension in productive imagination responsible for synthesizing temporal
appearances of succession.24
Since Kant's psychologistic principle requires that we ascribe only so much con
tent to a representation as its origin is capable of supplying (Section B and ch 2-E),
it is vital not to attribute more to pure time and its manifold than purely aesthetic
grounds of the unity of sensibility alone can warrant, prior to and independently
of "the conditions of the synthetic unity in which the concepts of the understand
ing collectively issue" (PFM 321). The appearances apprehended in pure time are
"met with in the mind scattered and single in themselves" (Al20) and so are as
uniform (undifferentiated) and devoid of all particular properties (indeterminate)
as the manifold of pure space. There is thus no warrant for crediting sensibility with
the objectification of pure time as a universal, all-encompassing order in which the
time of each appearance is uniquely differentiated from and completely determined
in relation to that of every other (it takes categorial synthesis speciosa to make good
this want: chs 13 and 15).
24 This recognition furnishes the key to resolving Hume's quandary concerning personal iden
tity in a manner consistent with Hume's-and Kant's--fundamental principles: UU ch 3. The
cognitive part of Kant's solution to the quandary will be considered in Chapter 17.
130 Kant's Anatomy of the Intelligent Mind
T his is not to deny that Kant conceived of sensible time as imposing a "neces
sary law" of subordinative order on all apprehended appearances ("I cannot arrive
at the following elsewise than through the preceding," Al99/B244). Yet we must be
careful not to read more into this purely aesthetic law than the demand for unity
of sensibility can warrant. Merely to perceive one appearance to follow another we
do not require that each be determined as situated in a time differentiated from the
infinitely many other times possible within pure time, much less that it be deter
mined as situated precisely so much (and neither more nor less) before or after each
of these others. To remedy the problem that arises insofar as unity of sensibility
is founded on a pure intuition that suffices only for coordinative synthesis of the
manifold, a subordinative synthesis supplying intuition of the acts of apprehension
responsible for synthesizing outer appearances is all that is required. For example,
such a synthesis enables sensibility to distinguish spatial manifolds apprehended in
distinct intuitions set together (by juxtaposition) as coordinate spaces from a single
complex, intuition encompassing the entire coordinated space given separately in
each (Section C). But this prediscursive, purely aesthetic subordinative ordering of
(temporal appearances of) acts of spatial apprehension in inner intuition is suffi
cient for one thing and one only: the unity of all possible appearances, outer no less
than inner, in sensibility. It thus falls to understanding and its pure concepts to dif
ferentiate and determine temporal appearances in the manner requisite for objects
to be given through them, mathematical no less than empirical.
{ 5}
In the preceding chapter, I argued that the unity of sensibility made possible by
pure space and time is an original synthetic unity of all the manifold of sensibility
in one consciousness a priori, prior to and independent of discursive understanding
in general and the categories in particular. Characterized in this way, it is natural to
ask whether Kant regarded this unity as a purely aesthetic, prediscursive expression
of the original synthetic unity of apperception he deemed the supreme principle of
the understanding in relation to the sensible manifold (B136). One might well think
so, owing to his insistence that the original synthetic unity of apperception in intu
ition itself, "ahead of all thought" (B132), is an essential condition of the possibility
of all discursive thought as such, whether cognitive or not, and so "is the highest
point to which all employment of the understanding, even the whole of logic and,
in accordance with it, transcendental philosophy, must be attached" (Bl33--4n).
Yet, I think it fair to say that virtually all interpreters would deny that any such
prediscursive synthetic unity of apperception is present in Kant's philosophy. This,
in my view, is due not to lack of textual evidence but to the mistaken belief that
Kant regarded the categories as necessary conditions of unity of apperception. If
that belief were true, it would indeed follow that Kant could not have treated the
unity of the manifold in pure space and time as an expression of the original syn
thetic unity of apperception because of its incompatibility with their purely aes
thetic, nondiscursive character as products of sensibility, prior to and independently
of all concepts, including the categories (A24-5/B39, B40, A31-2/B47, Bl60-ln).
However, what seldom, if ever, is considered is that although the categories may
indeed be necessary conditions of the objective unity of apperception involved in
the possibility of cognition and its objects, this does not make them necessary con
ditions of apperception as such, in all its guises, including the necessary synthetic
unity of pure space and time (unity of sensibility: chs 3 and 4) and the analytic
unity of apperception (the source of logical universality, the constitutive form of
all concepts as such, whether cognitive or not: ch 9).
132 Kant's Anatomy of the Intelligent Mind
Certainly, there is no shortage of texts that might be cited to support the standard
view. Yet the same is true of the contrary proposition: that original apperception,
the a priori unity of all the manifold in one and the same consciousness, underlies
(grounds) and is presupposed by the categories. In the B edition Transcendental
Deduction of the Categories, for example, Kant introduced the unity of appercep
tion this way:
This unity, which precedes all concepts of combination a priori, is not the
category of unity (§10), for all categories are grounded on logical functions in
judgments, and combination, and so too unity of given concepts, is already
thought in the latter. That unity, as qualitative (§12), must therefore be sought
still higher, namely in that which itself contains the ground of the unity of
distinct concepts in judgments, and consequently of the possibility of the
understanding even in its logical employment. (B131)
We will return to this text when we consider its place in the Transcendental Analytic.
Here, my concern is which view of the relation of the unity of apperception to the
categories accords with it best. If they are supposed to be its necessary conditions,
then I find it difficult to understand why Kant would say that this unity "is already
thought in" and must be "sought still higher" than both the unity effected by the
categories and the unity effected by the logical functions on which the categories
are grounded. Could the unity of apperception be so high as to "ground ...the
possibility of the understanding even if in its logical employment" if, as generally
supposed, its own possibility presupposed the categories and logical functions? By
contrast, the claim that the unity of apperception is higher than the unities effected
through categories and logical functions makes perfect sense if unity of appercep
tion is understood not as their consequence but their ground. Nor are these the
only unities it is higher than. For, on the interpretation advanced here, the original
synthetic unity of apperception underlies not only the categories and logical func
tions but the unity of sensibility effected by pure intuitions as well, and so consti
tutes the apex of Kantian transcendental psychology: the unique point at which the
fundamental a priori representations of sensibility and understanding converge,
each a distinct expression of this apperception, the one prediscursive and the oth
ers discursive. Surely, construing apperception as presupposed by, rather than pre
supposing, the categories and logical functions fits best with its designation as the
highest point and supreme principle of Kant's theory of the understanding and its
synthetic a priori judgments.
Kant's most extended affirmation of the priority of apperception with respect to
the categories is probably the following:
Apperception is itself the ground of the categories, which on their side repre
sent nothing else but the synthesis of the manifold of intuitions insofar as the
manifold has unity in apperception. Self-consciousness in general is therefore
the representation of that which is the condition of all unity and yet is itself
A New Understanding of Understanding 133
unconditioned. One can therefore say of the thinking I (soul) that thinks of
itself as substance, simple, numerically identical in all times, and the correlate
of all existence from which every other existence must be inferred, that it
does not so much cognize itself through the categories as cognize the catego
ries, and through them all objects, in the absolute unity of apperception, and
so through itself Now it is quite evident that I cannot cognize as an object
that which I must presuppose in order to cognize an object at all, and that
the determinative self (the thinking) is distinct from the determinable self
(the thinking subject) as cognition is from the object. Nevertheless, nothing is
more natural or misleading than the illusion which holds unity in the synthe
sis of thoughts to be a perceived unity in the subject of these thoughts. It may
be termed the subreption of the hypostatized consciousness (apperceptiones
substaniata;). (A401-2)
Although the place of this text in the context of the Paralogisms is not here my
concern, Kant could not have made it clearer than he did here that the true relation
of the unity of apperception to the categories is that of ground to grounded and of
the original to the derived, and that any assumption to the contrary leads inevita
bly to the illusion that the categories can be applied to apperception to determine
it as an object, a subreption that hypostatizes it, Cartesian style, as an efficacious
substance with qualities as well as (intensive) quantitative determinateness. Nor can
there be any doubt that Kant is asserting here the same as he did at Bl30-l and
elsewhere: that apperception, or self-consciousness in general, is the unconditioned
condition of all unity in representation, including the objective unity thought in
and through the categories. One must therefore ask again how the categories can
be necessary conditions of self-consciousness in general and so be presupposed
by apperception, if apperception is the "unity of consciousness that underlies the
categories" (B421)?
The priority of apperception with respect to the categories is also evident in
Kant's thesis that unity of apperception is not only implicit in but also constitutive
of logical universality: "the analytic unity of consciousness which makes [a repre
sentation] into a conceptus communis . . . attaches to all common concepts as such"
(Bl34n; ch 9-C). Apperception "is the vehicle of all concepts in general, and so
of transcendental concepts as well, and thus is always conceived along with these
latter, and therefore is just as transcendental as they are, but can have no specific
title because it serves only to present all things as belonging to consciousness"
(A341/B399-400) so that "the mere apperception I think . . . makes possible even
all transcendental concepts, in which it means that I think substance, cause, etc."
(A342-3/B400-401; also A348/B406).1 Since the categories are concepts, anyone
who supposes them to be necessary conditions of the unity of apperception must
1 And also: "the apperception that necessarily underlies understanding and all thought" (PFM
318) and "we have concepts only through the unity of consciousness" (AA 18 § 5650 [1785-8]).
134 Kant's Anatomy of the Intelligent Mind
explain how the latter can be posterior to (presupposed, grounded on, condi
tioned by) the categories and yet, at the same time, be a condition of the general
logical possibility of all concepts as such (representations universal in form) and
"the highest point to which all employment of the understanding, even the whole
of logic, and in accordance with it, transcendental philosophy, must be attached"
(B133-4n).
The difficulties with the standard view that the categories are necessary condi
tions for unity of apperception do not stop with texts that display Kant's pro
pensity for contrary formulations.That view seems to imply that there can be no
unity of consciousness without the categories. But what about pure space and
time? Their essential oneness (A25/B39) ensures that all their manifold is con
tained within them as individuals (not under them as universals) and so in one
and the same intuitive consciousness, completely a priori. Hence, if the categories
are supposed to be necessary conditions of any such all-encompassing a priori
unity of the manifold in one consciousness, would it not follow that the categories
are necessary conditions of pure space and time themselves? Yet how could that
be reconciled with Kant's iterated denial that there is any discursive (conceptual)
component involved in the pure intuition of space and time (A24-5/B39, A31/
B4 7) and, more particularly, his express rejection of the notion that the unity of
the space and time of transcendental aesthetic "belongs ...to the pure concept of
the understanding" (B161n)?
This difficulty may be evaded by insisting, first, that the unity of the manifold
in one consciousness effected by pure space and time, contrary to the preponder
ance of evidence to the contrary, is not a synthetic unity and, second, that the
categories are necessary conditions only of the synthetic unity of apperception.
But even if that were true, the existence of an a priori all-encompassing unity of
the manifold of intuition in one consciousness ahead of all thought, synthetic or
not, would seem to render the synthetic unity of consciousness produced by means
of the categories redundant. For with an original a priori unity of the manifold
in one consciousness already in place, prior to and independently of all synthesis
conformably to the categories, what need is there for the categories to provide the
manifold with unity?
One might reply that the unity of consciousness effected by the categories is the
one required for the possibility of experience and its objects. I would not disagree.
But if one is going to distinguish different all-encompassing a priori unities of the
manifold in one consciousness according to whether or not they suffice for the pos
sibility of cognitive experience, what reason is there for rejecting the view developed
here, according to which the purely aesthetic unity of sensibility made possible
by pure space and time is a prediscursive manifestation of the same original syn
thetic unity of apperception to which the categories give discursive expression but
with this difference: the former does not suffice to make possible experience and
its objects and so must be supplemented by the latter if their possibility is to be
secured?
A New Understanding of Understanding 135
If one concedes that unity of apperception grounds the categories and not vice
versa, what follows is a radical transformation in the way understanding needs to
be conceived. Instead of being essentially and fundamentally a discursive faculty,
defined in terms of an operation such as judgment, understanding becomes first
and foremost the faculty of apperception, responsible for all unity of consciousness,
even the purely aesthetic, prediscursive unity of sensibility effected by pure space and
time. In Chapters 3 and 4, we considered numerous texts that support this reading,
drawn from Kant's entire corpus. But to show that this was also Kant's view in the
Critique of Pure Reason, I will devote the remainder of the present chapter primar
ily to elucidating texts from that work, which, I shall argue, are best read the same
way. T he focus, not surprisingly, will be on those passages of the Transcendental
Deduction that pertain to its subjective side (the subjective transcendental deduc
tion of the categories: ch 12-B). T hese considerations should make clear that an
understanding not limited to or even defined primarily in terms of discursive repre
sentation is as essential to the doctrine of the Critique as a sensibility not limited to
or defined primarily in terms of sensation.
What would a text need to say to count as explicit confirmation of the thesis that
pure space and time are prediscursive expressions of the original synthetic unity
of apperception? An unambiguous statement that their unity is both original and
synthetic, where there can be no doubt that unity of apperception is meant, would
do well. T he text should also make clear that the space and time in question are
those of the Transcendental Aesthetic and that the manifold-unifying individuality
at issue is the same as that exhibited in the metaphysical expositions (ch 3-C). And
it should occur nowhere but in the course of the subjective transcendental deduction
of the categories since this is the one area of Kant's philosophy concerned not with
this or that special employment of the understanding (transcendental included) but
with "the pure understanding itself, according to its possibility and the cognitive
faculties on which it rest," the question being: "how is the capacity to think itself
possible?" (Axvi-xvii).
We find just such a text in § 17 of the B edition Transcendental Deduction, a
section marked out as part of the subjective deduction by its focus on "the very
possibility of the understanding" (B 137). If the significance of the text is so sel
dom appreciated, it is perhaps because it is preceded by what seems like a contrast
between space and time and the original synthetic unity of apperception: the for
mer are principles of the possibility of intuition in relation to sensibility, whereas
the latter is the principle of the possibility of intuition in relation to the under
standing. Evidently concerned with preventing his readers from misconstruing this
as denying that space and time themselves involve unity of apperception in their
136 Kant's Anatomy of the Intelligent Mind
Space and time and all their parts are intuitions, hence individual repre
sentations with the manifold they contain in them (see the Transcendental
Aesthetic), not mere concepts through which one and the same conscious
ness is contained in many representations but many contained in one and in
the consciousness of that one, and thus as composite; consequently, unity
of consciousness that is synthetic yet also original is met with in them. T his
individuality of space and time is important in application (see§ 262). (Bl36n)
2 Though the original refers the reader to § 25, I have altered it for reasons that will be made
clear shortly.
3 For a detailed discussion of commentary on this and other texts examined in this chapter,
see KMM part I. The difference between my treatment of these texts here and in my earlier book
is that my present focus is the role of understanding (apperception) rather than imagination
(synthesis) in the production of pure space and time.
A New Understanding of Understanding 137
that he understood apperception to encompass not only the unity thought through
the categories but also the prediscursive unity apprehended in intuition through
pure space and time ahead of all thought, categorial or otherwise.
Implications such as this, whether consciously recognized or not, most likely
account for the tendency of commentators to treat 136n as irremediably prob
lematic, draw a line around it, and thereafter neglect it. Rather than prejudice or
preconception, their reaction seems to me emblematic of the failure to meet the
challenges posed by Kant's conception of prediscursive sensibility. For it is only
by surmounting them that one can hope to understand why Kant, coming after
Berkeley and Hume, might have found it necessary to treat the prediscursive unity
of sensibility as no less a construct of the active (imagining, apperceiving) mind
than the unity of material and thinking beings in nature that Kant ascribed to dis
cursive understanding and its pure concepts.4
As soon as this is recognized, one cannot help but remark a number of other pas
sages in both versions of the Transcendental Deduction of similar import to Bl36n.
A pure synthesis of apprehension in imagination is essential because "without it we
could have the representations neither of space nor of time a priori" (A99). Space and
time require a synthesis and a unity not belonging to sense merely in order to be "given
as intuitions" (Bl60n). Unity of apperception enters into "the pure form of intuition
in time, merely as intuition in general that contains a given manifold" (B140), for
4 Some interpreters of Kant, notably John McDowell, dispense with the problem of the unity
of sensibility by ascribing to Kant a denial of prediscursive sensibility altogether, with the impli
cation that all consciousness, however primitive and unreflective, is permeated by the logical
space founded on the categories. In one form or another, this way of viewing Kant goes back at
least to nineteenth-century Neo-Kantians like Hermann Cohen and accounts for Kant's reputa
tion as an intellectualist. For further discussion, see conclusion and KMM introduction.
5 This parenthesis could refer to space and time in addition to a priori concepts (the catego
ries), or it could be another instance of the second-order concepts of space and time described
in ch 3-D: an enumeration of the marks that pertain to the pure intuitions of the metaphysical
expositions whereby they are understood as prediscursive, wholly nonconceptual representations.
In either case, the implication is that space and time are possible only through the relation of
intuitions to pure apperception. The objective unity of space and time will be discussed below,
and the objective unity of their manifold will be considered in connection with the Analytic of
Concepts and the Analytic of Principles in Parts IV and V.
138 Kant's Anatomy of the Intelligent Mind
In B136n, Kant tells us that the individuality of space and time, understood as an
original synthetic unity of their manifolds in one consciousness, will be "important
in application"; and although he refers to§ 25, it seems overwhelmingly likely that
he in fact meant§ 26 since the manifold-unifying individuality of space and time
plays no role in§ 25 but is central in§ 26 (most likely, at some point in in the later
stages of preparing the 1787 edition of the Critique, Kant decided to divide an
overlong section-possibly § 24---and, having split it in two, neglected to correct
the reference in B136n).6 T he portion of§ 26 that concerns the nature of space and
time begins with the remark that they are represented a priori not merely as forms
of sensible intuition but as themselves intuitions "with the determination of the
unity of this manifold" (Bl60). To elucidate the distinction between mere forms of
intuition and actual a priori intuitions of space and time, Kant appended the fol
lowing footnote:
that are essentially one. (2) In the Aesthetic, it was sufficient for Kant's purposes
to treat the pure formal intuitions of space and time-space and time as individual
unities of their manifolds-as if they had the same source (the receptivity of sense)
and the same nature (sensible) as the forms of intuition space and time-the mani
folds on which they confer their unity (formal intuition is also described as sensible
at PFM 287, AA 18 § 5928 and§ 5649). (3) Evidently concerned that this way of
understanding the unity of space and time might hinder readers' comprehension
of the reasoning in§ 26 of the Transcendental Deduction, Kant chose this section
to make explicit what had been left implicit in the Aesthetic: that this unity derives
from the understanding's determination of sensibility by means of the figurative
synthesis (synthesis speciosa) of productive imagination described in§ 24. (4) T his
is not to say, however, that the formal intuitions of space and time derive their unity
from any determination of sensibility by means of concepts of space and time; on
the contrary, such concepts are first made possible by them. (5) T his includes pure
concepts of the understanding: the unity of formal intuitions cannot be equated
with the unity that belongs to the categories because it is the unity through which
space and time "are first given as intuitions." (6) Although their being given as intu
itions depends on productive imagination's synthesis speciosa, it is evidently not the
transcendental synthesis speciosa mentioned at B151-2 since the unity of the cat
egories unquestionably does belong to that synthesis. (7) Thus, the unity of individ
uality that pure space and time confer on their manifold is purely aesthetic and so
properly ascribed to sensibility, even though it is not, as Kant's treatment of it in the
Transcendental Aesthetic might lead one to suppose, a product purely of receptivity
but also involves both the imagination's synthesis speciosa and the understanding's
original unity of apperception.
A principal objective of the footnote is to justify Kant's procedure in the
Transcendental Aesthetic of ascribing to sensibility a unity that only in the
Transcendental Deduction of the Categories is revealed to have its source in the
understanding. T he elucidation would have to be adjudged a failure, however, if
"understanding" were here to signify discursive understanding (understanding as a
faculty of representation by means of universals) , since the space and time of the
Aesthetic are explicitly said to be nondiscursive, and virtually everything in Kant's
doctrine of pure yet sensible intuitions depends on that being so. T his means that
the nondiscursive character of the unity of the space and time of the Aesthetic can
be preserved only if nothing involving the logical functions of judgment-which,
for Kant, define discursivity (understanding as Vermogen zu urteilen: A68-9/B93-
4) , the categories included (A80-1/B106}-is in any way concerned in the action
whereby the understanding produces formal intuitions. In particular, the justifica
tion given at B160n for Kant's proceeding as he did in the Aesthetic only holds
water if, on his conception of the understanding, it is just as possible for this faculty
to confer synthetic unity on the manifold prediscursively as discursively.
To recognize that this was indeed the case, one only needs to take to heart Kant's
insistence that the unity of apperception he equated with "the understanding itself"
(B133-4n) is higher than the unity belonging to the categories and logical functions
(B131) since, as noted in Section A, this opens up the possibility that it is also higher
than the prediscursive unity of space and time as individuals. In § 17 at B136n,
Kant affirmed this in the clearest possible terms: the individuality of the space and
time of the Transcendental Aesthetic-the a priori unity of the manifold contained
within their essential oneness-is none other than the original synthetic unity of
apperception. T hen, in § 26 at B160n, as if to banish any residual doubts, Kant
140 Kant's Anatomy of the Intelligent Mind
proceeded to reaffirm it, insisting that the space and time of the Aesthetic "are first
given as intuitions" only when understanding has synthetically determined sensibil
ity. Since this is just to say that given receptivity alone, operating in accordance
with forms of intuition, any intuition of space or time is just as impossible as any
concept of them, the clear implication is that understanding is no less essential to
intuitions than to concepts. Thus, the only way to read Bl60 + n so as to preserve
the nondiscursive character essential to the space and time of the Aesthetic is to
suppose that, in producing the formal intuitions in which they are first given, the
understanding confers a priori unity on the manifold prediscurisvely, that is, with
out the aid of universals, logical functions, or anything else that would compromise
the nondiscursive character of its action.
Although the foregoing reading of the footnotes at Bl36 and Bl60 is the most
straightforward, the overwhelming majority of interpreters nevertheless resist it
because it seems to make nonsense of the notion of understanding, not only in
Kant's theory but in general, to suppose that it can ever be anything other than dis
cursive. Any supposition to the contrary unquestionably conflicts with Kant's initial
explication of understanding as the capacity to judge (A69/B94), as well as with each
of its subsequent characterizations (such as those at Al26), with the single possible
exception of its equation with the capacity for apperception. This last characteriza
tion of understanding is unique in that unlike all of the others, it relates not to any
discursive representation or operation but to pure self-consciousness: the a priori
unity of all the manifold in one consciousness together with the representation it
makes possible of the identity of that consciousness in respect to all the manifold
(the synthetic and analytic unity of apperception of Bl33-4, analyzed in ch 9-B).
Among the most salient and original features of Kant's theory of understanding
is his endeavor to bring pure self-consciousness together with discursivity in a single
indissoluble union. But the question facing us is which he deemed most fundamen
tal to the nature of the understanding. Where most interpreters go wrong, I believe,
is in taking discursivity, as defined by logical functions of judgments, to be defini
tive of understanding for Kant and self-consciousness as something secondary and
derivative that results when logical functions of judgment (in their guise as pure
concepts of the understanding) are applied to the a priori manifold of sensibility.
It is not just that he regarded apperception as a higher unity than that of either the
categories or the logical functions and made quite clear that categories presuppose
apperception as their ground (Section A). It is also his claim that apperception, in
its guise as analytic unity, is constitutive of logical universality itself (Bl33-4n),
with the implication that the original synthetic unity that precedes and first makes
this analytic unity possible (Bl33-4) must already be in place, right in intuition
itself, "ahead of all thought" (Bl32), and so before discursive understanding is
A New Understanding of Understanding 141
even possible: "the synthetic unity of apperception is the highest point to which all
employment of the understanding, even the whole of logic and , in accordance with
it, transcendental philosophy, must be attached" (B133--4n).
Yet even if one concedes, at least provisionally, the priority of understanding
qua capacity for apperception over understanding qua discursive capacity, one may
still balk at accepting the implication that understanding, in producing formal intu
itions of space and time, is not a discursive faculty of concepts and judgments at all
but a nondiscursive faculty of intuitions. How can apperception be nondiscursive?
And if it can, why should this not instead be taken as proof that apperception is
not, after all, the province of the understanding?
The synthetic unity of the manifold in one consciousness involves three distinct
components, each of which Kant attributed to a different faculty: the manifold
of sense, the synthesis of imagination, and the unity of understanding (B 130-1).
From a sensibilist point of view (ch 2, UU ch 1), it has unmistakable affinities with
the notion of a complex idea developed by Locke, where given ideas (a "mani
fold"-which may be genuine givens of sense or previously formed complex ideas)
are brought together ("synthesis") and the resulting combination considered as a
single complex idea ("synthetic unity of the manifold"), be it of a substance, mode,
or relation (UU ch 5-C). Thus, complex ideas of objects always involve the per
formance of two distinct actions on the ideas concerned in them: in addition to
bringing them together in one consciousness, the ideas must also be represented as
the manifold of one and the same representation (as Kant put it, "by taking sev
eral things together, you achieve without difficulty a whole of representation (totum
repraesentationis) but you do not, in virtue of that, arrive at the representation of a
whole (repraesentationem totius)," ID 390).
For Kant, the objective unity that makes the difference between an object in the
full cognitive sense and objects of this or that particular person's own subjective
representation (Al08-9) is discursive in nature: "an object is that in the concept of
which the manifold of a given intuition is united" (B137, my emphasis; also A104-
6, A191/B236, and A494/B522). Since this unification of representations "demands
unity of consciousness in the synthesis of them" (B137; also A78/Bl04, A105-6,
Bl45, and A326/B382-3), a Kantian object, just like a Lockean complex idea,
involves three distinct components, each with its own distinctive faculty source: the
manifold of sense, its synthesis in imagination, and the concept that, by determin
ing this synthesis, unites the manifold in the representation of a single (complex yet
unitary) object in the understanding (A103-4, analyzed in ch 13). But on Kant's
analysis, all objective unity of consciousness in the synthesis of representations
through concepts (synthesis of recognition) rests ultimately on the necessary syn
thetic unity of apperception, which therefore constitutes "an objective condition
of all cognition ...[,] a condition under which every intuition must stand in order
to become an object for me" (B138). Thus, concepts can produce objects (objective
unity) from the synthesis of the manifold only insofar as they partake of this objec
tive (necessary synthetic) unity of apperception, and the categories, as the conditions
142 Kant's Anatomy of the Intelligent Mind
that make such unity possible, thereby acquire the status of pure concepts of an object
of sensible intuition in general (which is what Kant set out to prove in their transcen
dental deduction: chs 11-C, 13, and 14).
Now, while Kant's goal in analyzing objectivity may have been to establish that con
cepts can partake of the unity of apperception and thereby acquire cognitive value only
if they are subordinated to the categories (§§ 19-21, analyzed in ch 14), one of the
intermediate steps toward that goal was the equation of the original synthetic unity
of apperception with "the objective unity of self-consciousness" (title of § 18 [B139]).
This includes the thesis that "the pure form of intuition in time-merely as intuition in
general which contains a given manifold-stands under the original unity of conscious
ness simply through the necessary relation of the manifold of intuition to one I think,
hence through the pure synthesis of the understanding that underlies the empirical a
priori" (B 140; at A 107, cited at the end of section B, the same point is made with regard
to the "objective unity" of space as well as time). Given the essentially nondiscursive
nature of pure space and time, this raises the question of how the objective unity of
self-consciousness involved in these intuitions can obtain prediscursively, i.e. prior to
and independently of all concepts the categories included (as Sections B and C make
clear it must). A possible answer emerges with the recognition that pure intuition of
space and time involves the same three components found in discursive (concept-based)
objective unities: the manifold, its synthesis, and the consciousness of that synthesis as
a unity. For without the addition of the latter, neither pure space nor pure time could
be represented as a single complex object (individual unity) in which all their manifold
is contained. Since Kant invariably treated the difference between these components as
sufficiently great to warrant attributing each to a different faculty-sense, imagination,
and understanding-we should not therefore be surprised that he seems also to have
seen fit to attribute the prediscursive objective unity of pure space and time to these
three faculties as well so that in the case of pure space and time uniquely, the two basic
features of understanding-apperception and discursivity--diverge.
I noted earlier that the issue of whether Kant, in attributing the objective unity
of pure space and time of the Aesthetic to the understanding, was conceiving of its
operation as nondiscursive turns on which of the two basic features of this faculty is
regarded as more fundamental. Since the evidence suggests that no unity was more
fundamental for him than apperception and that apperception must be conceived as
an action just as distinct from the aesthetic synthesis of imagination as from any dis
cursive action (conception, judgment, inference), it should now be clear not only that
there is nothing extraordinary, much less absurd, in Kant's treatment of understanding
as the source of the nondiscursive objective unity of space and time, but also that it is
precisely what he should have held in the sole instance where the apperceptual signifi
cation of understanding fails to coincide with the discursive.7
7 That Kant found nothing odd about characterizing the understanding as a faculty of intu
itions is clear from the following: "It is readily evident that if the capacity of cognition in general
is to be called understanding (in the most general sense of the word), the latter must contain a
A New Understanding of Understanding 143
The B160 footnote should be read in conjunction with Kant's discussion of objec
tive unity in§§ 17-18 because its topic is the representation of space and time as
objects. The argument of § 26 required its insertion because it turns on the rec
ognition that space and time are not only forms of intuition but intuitions them
selves that, as such, already involve unity of apperception. But what in that case are
forms of intuition? Both the footnote and the text to which it is attached imply that
they are not intuitions and, in particular, that space and time cannot be given as
intuitions by their means alone (space and time are "first given as intuitions" only
through the synthetic determination of sensibility by the understanding). Kant's
lone positive characterization is that "form of intuition merely gives a manifold"
(B160n). But a manifold of what? It seems unlikely that Kant had appearances in
mind: though spatial and temporal in form, appearances are never given in intu
ition except as preceded, made possible, and contained within the pure intuitions of
space and time, and Kant equates "intuitions that contain a manifold" (Bl60) with
formal intuitions. Affections in synopsis (data of outer and inner sense) are neither
intuitions nor contained in them; yet, having no form, it seems just as improbable
that Kant can have regarded them as the manifold given in accordance with space
and time in their capacity as forms of intuition. But what other option is there?
capacity for the apprehension (attentio) of given representations in order to produce (hervorzu
bringen) intuition, a capacity for abstraction of that which is common to several (abstractio) in
order to produce a concept, and a capacity for reflexion (rejlexio) in order to produce a cognition
of the object" (Anthropology 138).
8 Given Kant's general practice of grounding objectivity on concepts and the categories in
particular, it is strange that he would characterize prediscursive space and time as instances of the
objective unity of apperception at A107 and B140. It is possible that these passages actually refer
to space and time insofar as their manifold has already been determined by categorical synthesis
as a necessary synthetic unity. Or it may be that Kant was using "objective" in a broader sense
than he did elsewhere. A synthesis is "objective" insofar as it is necessitated in some manner;
and since the same synthesis of apprehension whereby appearances are given also necessitates
that they fall within the synthetic unity of space and time, he may have found this sufficient for
regarding space and time as prediscursive expressions of the objective unity of apperception,
even though the manifold in them lacks the necessary interrelation in which discursive, properly
cognitive objectivity consists (ch 13).
144 Kant's Anatomy of the Intelligent Mind
Such evidence as there is suggests that space and time, considered merely as forms
of intuition, prior to and independently of the synthetic determination of sensibility
by understanding through which space and time are first given as intuitions, are not
representations of any kind at all but are mere capacities to exhibit the manifold data
of sense as outside or after one another and so differ from one another in the same
sense the capacity of sight differs from that of hearing as the source of sensations
of color and light rather than sounds. It is not just Kant's insistence in B 160n that,
given forms of intuition alone, neither intuitions nor concepts of space and time
are possible that indicates this. In the only other passage in the Critique where both
between the mere capacity to intuit spatial and temporal appearances, which involves
receptivity alone, and actual intuition of these appearances, which is a unity of sensi
bility effected through the synthetic determination of sensibility by the understanding
(different forms of intuition-a possibility that Kant held cannot be denied-would
be a capacity to intuit appearances that were not spatial or temporal but of a form
unimaginable by beings constituted like ourselves and would yield actual synthetic
unitary formal intuitions that preceded and made possible those appearances).
Elsewhere in the Critique, Kant used "formal intuition" as a parenthetical clari
fication of pure intuition (B207) and to indicate the sense in which he was using
"form of outer intuition" (A429/B457n). But the text that most closely echoes
A267-8/B323--4 is a passage in Kant's response to Eberhard, a Leibnizian critic:
This first formal ground alone, e.g. of the possibility of a space intuition, is
innate, not the space representation itself. For it always requires impressions
in order first to determine the cognitive faculty to the representation of an
object (which is always its own act). Thus arises the formal intuition termed
space as an originally acquired representation (the form of outer objects
generally), the ground of which (as sheer receptivity) nevertheless is innate
and the acquisition of which long precedes determinate concepts of things
that conform to this form. . .. Concerning this signification of the ground of
the possibility of a pure sensory intuition no one can be in doubt save he
who wanders through the Critique with the aid of a dictionary but has not
thought it through. (Discovery 221-3)
Like A267-8/B323--4, this passage sheds light on the B160 footnote by contrasting a
faculty ground with space considered as an actual a priori representation. The latter
is termed a "formal intuition" and, as resulting from an act of the cognitive faculty,
is contrasted with the "sheer receptivity" of its innate "formal ground." To contrast
with receptivity, formal intuition must at least involve a synthesis of imagination,
which the use of "act" would seem to confirm (Kant seldom if ever used "Handlung"
except in reference to spontaneity, whether syntheses of imagination, apperception,
or the properly discursive operations of the understanding). Specifying it as an act
of the cognitive faculty also implicates the understanding in the production of for
mal intuition, for Kant always, to my knowledge, employed the term "cognitive fac
ulty" (Erkenntniskraft) so as to include the understanding in its capacity as faculty of
apperception (e.g., Al 14, B137-8). Since there is therefore every reason to equate the
formal intuition described in the response to Eberhard with the formal intuitions of
B160n, the contrast in the former between the actual (formal) intuition of space and
time and the innate formal ground of its possibility serves to confirm my reading of
the forms of intuition of B160n as mere capacities to intuitively represent manifolds
of spatial and temporal appearances and not actual intuitions of these appearances.9
9 See KMM ch 2 for a more detailed discussion of the topics discussed in this section and
the next.
146 Kant's Anatomy of the Intelligent Mind
If the unity of formal intuitions precedes and first makes possible concepts of space
and time, then the supposition that formal intuitions are intuitions constructed in
accordance with mathematical concepts is viciously circular. To avoid the circle, one
must instead understand the formal intuition of space "required by geometry" as
the prediscursive transcendental space presupposed by geometrical concepts, that is,
as identical with the space of the metaphysical and transcendental expositions of
10
See AA 18 § 4873, cited in Chapter 1.
148 Kant's Anatomy of the Intelligent Mind
just as essential to the space and time of the Transcendental Aesthetic as the
receptivity of sense. 11
Even in the more limited context of § 26 itself, the categorial construal of for
mal intuition is more problematic than it initially seems. For whatever attraction
one finds in a reading that allows the conformity of synthesis of apprehension of
the manifold of appearances to the categories to be inferred from its conformity
to categorial formal intuition, it disappears as soon as one considers the difficulty
in demonstrating what it presupposes: the conformity of formal intuitions to the
categories. Where is this conformity demonstrated? Where did Kant even claim to
have established it? What he actually said is this:
[S]pace and time are represented a priori not merely as forms of sensible intu
ition but as themselves intuitions (which contain a manifold), and thus with
the determination of the unity of this manifold in them (see Transcendental
Aesthetic). T hus, unity of the synthesis of the manifold, without or within us,
hence also a combination to which everything that is to be represented deter
minately in space or time must conform, is at the same time already given as
a condition of the synthesis of all apprehension with (not in) these intuitions.
But, this synthetic unity can be none other than that of the combination of the
manifold of a given intuition in general in an original consciousness in confor
mity to the categories, only applied to our sensible intuition. Consequently,
all synthesis whereby perception itself becomes possible stands under the cat
egories, and since experience is cognition through connected perceptions, the
categories are conditions of the possibility of experience, and therefore hold
a priori also of all objects of experience. (Bl60-l)
T he key to a correct reading of this text, in my view, is giving proper due to the
qualifying phrase "with (not in) these intuitions." Were the parenthetical phrase
absent, Kant would indeed most naturally be read as claiming that a combination
to which everything that is to be represented determinately in space and time-a
combination that, as the next sentence makes clear, is none other than categorial
combination-is already given in the formal intuitions of space and time. But what
needed to be said to vindicate the categorial construal of formal intuition and what
actually was said are two different things. It therefore is all the more striking that
Kant took the trouble to add the parenthetical "not in," for it was evidently with
the intent of preventing his readers from interpreting the categorial combination
that goes "with these intuitions" as a categorial combination of the manifold in
these intuitions. And although what Kant was affirming in the cited passage may be
u Although critical of Kant for not making this clear in the Transcendental Aesthetic, Hegel
recognized that "in the deduction of the categories, .. . space and time are conceived as synthetic
unities, and the productive imagination, i.e. spontaneity and absolute synthetic activity, as the
principle of sensibility which previously had been characterized only as a receptivity" (Faith and
Knowledge 297). Whether Hegel also appreciated that synthetic unitary space and time do not
depend on the categories I do not know.
150 Kant's Anatomy of the Intelligent Mind
less than pellucid, we can at least be sure what he used the parentheses to deny: the
categorial nature of formally intuited space and time themselves (which, as noted
earlier, is made explicit in the B160-1 footnote: "the unity of this a priori intuition
belongs to space and time and not to the concept of the understanding").
The significance of the parenthetical insertion can be appreciated only if one
recognizes that the necessary unity of the manifold of all possible appearances in
the pure space and time of the Transcendental Aesthetic still leaves these appear
ances a bare manifold, devoid of all spatial and temporal differentiation and deter
mination ("scattered and single in themselves," A120). This was the source of the
self-created problem considered in ch 4-B and -D: since pure sensibility is incapable
of supplying the differentiated, determinate manifold on which the possibility of
cognitive experience and its objects depends, pure understanding must make good
this want by means of its pure concepts. This they do by means of a transcendental
synthesis speciosa that brings the manifold of formally intuited space and time into
conformity with the objective unity of apperception; and since a synthesis based
on the categories is capable of yielding only philosophical-discursive synthetic a
priori cognition, never the mathematical-intuitive kind (A712-38/B740-66), it does
indeed determine their manifold "with," not "in," these intuitions. Consideration
of precisely how this works will, however, have to postponed until Part IV, when we
turn to the Transcendental Deduction of the Categories and are finally in a position
to provide a category-by-category elucidation of the of the transcendental synthesis
speciosa of the manifold of pure formal intuition.
My thesis that prediscursive apperception is part of, indeed central to, Kant's con
ception of the understanding is both novel and transformative and so is likely to be
resisted by many readers. Since the evidence in its support is dispersed throughout
the book, it may prove helpful to assessing it to (i) define the issue more precisely,
(ii) provide a clear formulation of my position, (iii) summarize the multiple conver
gent strands of evidence I adduce in its favor in one place, and (iv) indicate where
detailed consideration of each strand can be found.
The issue, as I see it, is twofold: first, whether apperception essentially incorpo
rates the categories in its representation, as generally supposed but which I deny
and, second, whether the categories essentially incorporate apperception, which
I affirm. In respect to the former, my thesis is that unity of apperception incorpo
rates the categories into its representation only insofar as its unity is to count as objec
tive in the sense requisite to ground experience and its objects; that is, it is not unity of
apperception as such that is categorial but only the objective unity of apperception. In
the latter regard, my thesis is that the categories incorporate apperception into their
representation both because the analytic unity of apperception enters into all concepts
A New Understanding of Understanding 151
(universals) as such, the categories included, and because they essentially involve the
representation of the synthetic unity of the judging subject insofar as logical functions
of judgment enter into their content. Moreover, insofar as the categories incorporate
these two discursive precategorial guises of apperception into their representation and
both alike are preceded and made possible by an original synthetic unity of appercep
tion right in the manifold of intuition, ahead of all discursivity, this prediscursive unity
of apperception must be recognized as preceding and making possible the categories
themselves as well. Finally, I hold that this prediscursive original synthetic unity of
apperception can be none other than the unity of sensibility effected by pure intuitions.
The evidence:
1. There are several more or less explicit statements of the thesis that apperception
grounds and is not grounded by the categories, such as B131, B133-4n,
A341/B399-400, and, above all, A401-2, all cited in Section A. Because such
passages leave the reasoning underlying apperception's priority somewhat
unclear, I rely on them more to confirm than prove the thesis.
2. Kant singled out the definition of cognitive understanding as the faculty of rules
from the others he provided as the one that comes closest to its essence (A126),
while its definition as the capacity to judge or think (A69/B94 and AS 1/B106)
is truest to its purely formal general logical character. Clearly, if either of these
were truly Kant's most fundamental definition of the understanding, there
would indeed be no place for a prediscursive understanding in his theory.
However, as we have seen, there is yet another definition more fundamental
than either of these that captures the transcendental psychological essence of
the understanding: the faculty of apperception (self-consciousness). It is in this
latter character that the understanding admits of prediscursive employment.
3. The passages examined in the present chapter show that Kant disambiguated the
doctrine of the Transcendental Aesthetic in the Transcendental Deduction of
the Categories to make clear that the unity of the manifold in sensibility made
possible by prediscursive pure space and time is in fact an original synthetic
unity of that manifold in one consciousness. The main obstacle-perhaps
the only one-that prevents us from taking texts such as B136n at face value,
as meaning what they say, 12 is the extent to which Kant scholars have grown
accustomed to treating the relation of categories to apperception as conditions
12
It is true that B136n is not as explicit as A401-2. Instead of Kant stating that space and
time are instances of prediscursive apperception, we are obliged to put together two statements
to infer it: the statement in the Aesthetic that pure space and time are not discursive and the
statement at B136n that the space and time of the Aesthetic are original synthetic unities of their
manifold in one consciousness. But just as two plus two cannot go together any other way than
to yield four as their sum, the pure space and time of the Aesthetic cannot be both nondiscursive
and original synthetic unities of their manifold in one consciousness unless apperception has a
prediscursive as well as precategorial and categorial discursive guises. And since the understand
ing is the faculty of apperception, this is just to say that it is a faculty that has prediscursive no
less than discursive employment.
152 Kant's Anatomy of the Intelligent Mind
to conditioned. So try an experiment: set habit aside, reread this chapter if you
are not yet fully convinced, and then approach the rest of the book with an
open mind as to whether more (and better) sense can be made of their relation
by regarding it in the way I propose.
4. For Kant, thinking (judgment, propositional thought) starts with concepts
(e.g., only the concept derived form the sensation of red can function as
the subject or predicate of a judgment, not the sensation of red itself).
Concepts, however, derive their constitutive form, logical universality (L
91), from the analytic unity of apperception, which in turn is preceded and
made possible by the synthetic unity of apperception. B133-4n in particular
could not be clearer on this score: concepts derive their constitutive
logical form from pure self-consciousness. Because thinking is impossible
without concepts, there seems no way to escape the conclusion that all
concepts, qua universals, are ineluctably bound up with psychology. Now,
the categories are concepts: "pure synthesis [in imagination of the a priori
manifold of intuition] represented universally gives the pure concept of the
understanding" (A78/Bl04). More particularly, they are concepts of that
synthesis as determined conformably to logical functions of judgment
(A78-9/B104-5, A321/B378, A401). So how can the categories be supposed
to precede and make possible the analytic unity of apperception that must
first attach to any representation before it can become a concept?13 This
shows that the analytic unity of apperception is precategorial and that the
synthetic unity of apperception it presupposes as being in place "ahead of
all thought" is prediscursive. This argument will be presented in detailed
form, together with supporting evidence, in Chapter 9.
5. It might be argued that the circularity just described could be avoided if one
were to suppose that the logical functions from which the categories derive,
rather than the categories themselves, were presupposed by apperception.
The problem with this suggestion is that logical functions cannot operate
but must instead lie dormant in the mind until concepts are provided to
occupy the logical places of subject and predicate in categorical judgments;
e.g., only the concept of red formed by attaching the analytic unity of
apperception to the sensation (B133-4n), not the sensation itself, can
function as the subject of a categorical judgment. Thus, logical functions
of judgment are just as dependent on the analytic unity of apperception
and the synthetic unity of apperception it presupposes as the categories.
This argument will be presented in detailed form, together with supporting
evidence, over the course of Chapters 9 and 10.
13 In other words, either the categories are pure concepts of the understanding that, as con
cepts, incorporate analytic unity of apperception into their representation (Bl33-4n and A341/
B300-400), and so to that extent cannot be presupposed by that unity, or they are not concepts.
But if they are not concepts, what are they?
A New Understanding of Understanding 153
One reason for taking so seriously the task of identifying Kant's purely aesthetic
raison d'etre for positing pure intuitions of space and time was to be able to put the
transcendental expositions of space and time of the Transcendental Aesthetic into
proper perspective. Because that task is so seldom undertaken, one typically finds
interpreters highlighting the transcendental expositions as Kant's chief justification
for treating space and time as pure intuitions of sensibility-as if our senses have
the constitution they do only so that a select few might devote themselves to the
arcana of mathematics and natural science once the cultural conditions for doing
so were in place. T his is particularly true in the case of geometry. Many interpret
ers find Kant's claim that geometry presupposes a pure intuition of space plausible
given the state of the science in his time. But geometry has changed so much since
then that few would follow him in holding pure intuition to be essential to present
and future geometry, much less endorse his claims that arithmetic and even alge
bra would be impossible without pure time. Kant's claims may well be false. But,
in my view, the reasons most interpreters deem them to be false stem from basing
their understanding of pure space and time less on Kant's doctrine of pure sensi
bility than on the exogenous notions of space and time they bring to Kant, and it
is mainly these notions they find incapable of sustaining Kant's claims. To guard
against this, one must situate his claims in their proper psychologistic context so
as to determine whether the features of space and time Kant is supposed to have
attributed to sensibility do in fact derive from that faculty and are not instead prod
ucts of the understanding, either in its transcendental, mathematical, or empirical
employments. Doing so brings this added bonus: the doctrines of the transcenden
tal expositions emerge as the exact opposite, in a variety of crucial respects, of what
they are usually taken to be.
without ipso facto thinking the other(s)-that is, incorporating the latter into the
contents thought in the former-then the judgment is analytic (ch 2-A). Because
they relate to the contents thought in concepts and judgments rather than to any
thing outside the thought that may or may not correspond to its contents, analytic
judgments never have objective validity, and their value is confined to helping to
prevent conceptual confusion by making explicit contents that may previously have
been thought only obscurely. There is therefore no need to look beyond the com
ponents of an analytic judgment to determine its truth: to grasp their relation is, in
1 Although Locke and Hume were aware that judgments in which we are sensible that a
necessary relation between nonlogically identical components exists, they did not recognize the
insufficiency of the principle of contradiction to ground them and so never descried the crucial
importance of the question of how such judgments are possible (see ch 2; also UU chs 2-D, 8-D
and 18-A).
Mathematics and the Unity of Sensibility 157
Pure sensible intuition is the most fundamental of the various X-factors, the
one presupposed by all the others. (1) Without the synthetic a priori judgments of
pure understanding that pure intuition makes possible, there would be no objective
world of cognitive experience: no natural laws, no material and thinking beings,
no necessary connections of any kind at all. Under such circumstances, the very
notion of empirical subjects and their agency would be unintelligible; no maxims
of actions could be formulated; and there would be nothing to which the moral
law could be applied. Indeed, the moral law itself would be meaningless, for, where
no objective "is" is possible, and so neither fact nor truth, how can there be an
"ought"? (2) Without the world of experience made possible through the synthetic
a priori judgments founded on pure intuition, there would be no nature in which
to find beauty (or sublimity or adaptive purposiveness). Moreover, given its depen
dence on the interplay of cognitive faculties, aesthetic pleasure itself could never
be felt if the conditions for the cognitive exercise of those faculties failed to be
met. (3) And insofar as pure intuition is essential to the original synthetic unity of
apperception that precedes and makes possible the analytic unity of apperception
essential to all concepts as such (ch 9-C), it underlies synthetic a priori judgments
of every kind, indeed all judgment as such, since even analytic judgments involve a
synthesis of the consciousness of the predicate with the consciousness of the subject
(B131).2 Kant's general question as to the possibility of synthetic a priori judg
ments, framed in terms of an unspecified a priori X-factor, thus resolves itself into
the more specific question of how pure intuition enters into all a priori relations
between nonlogically identical discursive representations in judgments, be their
content theoretical, practical, aesthetic, or teleological.3
2 Did Kant overlook a species of synthetic a priori judgment when he ignored immediately
perceptible qualitative relations? Both Locke and Hume insisted that we can sometimes appre
hend the necessity of such relations with the same intuitive or demonstrative certainty we per
ceive quantitative relations in mathematics (ECHU IV/ii/§13 and THN 70/50). Since being lighter
than purple does not enter into the content of the visual idea of yellow, the relation is obviously
not analytic (there is no contradiction in the thought that one could have a sense of sight that
enabled one to see yellow but not purple); and since judgments do not, in and of themselves,
require acquaintance with objects of experience (material and thinking beings), judgments con
cerning necessary qualitative relations might seem to contradict Kant's belief that synthetic a
priori judgments involving pure intuition have pride of place. I do not know of any text in which
Kant addressed the issue. If he had taken it up, he could have responded by noting that universals
(concepts and judgments) are possible only given conditions for the I think to take place, that is,
the purely aesthetic synthetic unity of apperception that forms out of intuition (not sensation,
i.e., affection) antecedently to all thought (Bl32-4, ch 9-B); thus, no judgments of any kind,
including subjective judgments of perception concerning sensational qualities, are possible prior
to and independently of the synthesis of the manifold of intuition in one consciousness made
possible by pure space and time (see chs 5, 9, and UU ch 3-E).
3 See, for example, Discovery 244-5: "[T]hat something outside the given concept must still
be added as substrate which makes it possible to go beyond the concept with my predicates is
clearly indicated by the expression 'synthesis,' and consequently investigation is directed toward
the possibility of a synthesis of representations for the sake of cognition in general, which must
soon turn to intuition, while pure intuition must be acknowledged as the unavoidable condition
for a priori cognition."
158 Kant's Anatomy of the Intelligent Mind
Yet one's answer to this question can be only as good as one's understand
ing of pure intuition and its manifold is accurate and sufficiently comprehensive.
Commentators confronted by the paradoxical-seeming notion of a pure yet sen
sible intuition tend to fall back on more accessible, pictorial notions of sensible
spatiality and temporality, particularly when attempting to elucidate the role Kant
accorded to the pure intuition of space in the synthetic a priori propositions of
geometry. This is fine so long as one is careful to distinguish seeing or visualizing
a figure and various auxiliary lines in the course of a geometrical demonstration
from the intuitive exhibition of that figure in a space that, as pure, has no more to
do with visual sensation than auditory, olfactory, or any other kind. Unfortunately,
this seldom happens. Whether it is the widespread assumption that Kant was a
dogmatic Euclidean that misleads so many to construe pure space imagistically or
the widespread tendency to construe pure space imagistically that misleads so many
into supposing that he was a dogmatic Euclidean, these two misconceptions make
it virtually impossible to comprehend the role Kant accorded to pure intuition in
geometry and synthetic a priori mathematical judgments generally.
I see no reason to doubt that human imaginations lack the capacity to visualize
non-Euclidean spaces or figures, spaces or figures occupying dimensions beyond
the third, and so on. The question is: are species-specific psychological limitations
grounded in the nature of a particular faculty of sensation at all relevant to Kant's
conception of the pure space intuition and its manifold of juxtaposed appear
ances as a presupposition of mathematical cognition? If the interpretation of pure
intuition in Chapters 3-5 is correct, the answer is no. In pure space, nothing is
represented but the prediscursive (purely aesthetic) unity of a synthesis (by juxta
position) whereby sensations are exhibited by a homogeneous manifold of appear
ances devoid of all sensational quality, representable as the manifold of a single
representation, all contained in one and the same consciousness. To be sure, inso
far as the pure space intuition functions as a "schema that is always related to the
reproductive imagination that calls up the objects of experience" (Al56/B195; also
ID 403), it combines with sight (or internal visualization) and touch (or tactual
imagination) to yield visual and tactual perceptions of position, shape, distance,
direction, movement, dimension, perspective, and much else besides (the posture
of a body, the expression on a face, etc.). But as pure, and so free of any depen
dence on the sensational contents supplied by vision, touch, or any other capacity
for sensation, it is likewise independent of the psychology of these capacities and
so free of their species-specific limitations as well. Indeed, the manifold made pos
sible through spatial synthesis is not only as inexhaustibly differentiable as data
of existing human outer senses but also as differentiable as vision, touch, smell,
etc. would be (1) if they were subject no limits whatsoever, human or any other;
(2) if the data present in them were all perfectly discernible (nothing obscure);
and (3) if, in addition to our five senses, we had perfectly discernible inputs from
all other possible outer senses, none of which were subject to creaturely limits of
any kind (ch 4-B). There is therefore no reason to suppose that the pure space
Mathematics and the Unity of Sensibility 159
intuition cannot be utilized to apprehend spaces in pure intuition that, for human
beings, are quite literally unvisualizable, including spaces as yet uncontemplated
by geometers.
It cannot be emphasized too strongly that, on Kant's conception of space as a
pure intuition, nothing spatial can be perceived in sensation, visual and tactile not
excepted (Chapter 3). For example, the border formed by different colors in the
visual field is, in itself, no more a source for the representation of a line than the
clarinet glissando that opens "Rhapsody in Blue" or a continuously intensifying
odor of ammonia. Nor is the principal reason for this that one must, in addition to
seeing a line, trace it in thought (imagination and discursive understanding) so as
to "synthetically bring into being a determinate combination of the given manifold
such that the unity of this action is at the same time the unity of the conscious
ness (in the concept of a line), whereby an object (a determinate space) is first cog
nized" (B137-8; also A162-3/B203 and AA 18 § 5090). It is instead that, on Kant's
doctrine of pure sensibility, the action of juxtaposing a manifold, and so too the
formation of a concept of a line or any spatial concept, would be impossible in the
absence of the pure space formal intuition. If performed on the same visual datum
(the border formed by different colors) or on a comparable tactual datum (e.g., run
ning a finger along the edge of a table), but in accordance with a different form of
sensibility than space, the same psychological operation would yield a non-spatial
appearance and a concept formally incommensurable with our spatial appearance
and concept of a line.
The form of sensibility is thus the crucial factor in spatial representation, not
concepts or a creature's imagizing psychology. In humans, the capacity to make
super-fine discriminations visually and tactually accounts for the preeminence of
these senses in spatial imaging. But once it is acknowledged that unity of sensibil
ity-the ability to exhibit the data of all of the senses as the manifold of a single
representation-is, for Kant, the raison d'etre of pure space (ch 4-B), the primacy
of vision and touch in human outer experience must be recognized as nothing more
than a human peculiarity, with no relevance to spatial representation properly so
called. Other creatures, with the same form of intuition, might rely primarily on
olfaction, gustation, or some sense unknown to us to generate spatial images and
concepts. Equally, there might be creatures none of whose senses were of use to
generate spatial images. Would this prevent them from doing geometry? There is
no reason to think so since they could still have recourse to symbolic constructions
of the kind that figure in algebra and reign supreme in geometry and topology
today (A717/B745, A734-5/B762-3, CJ 251, 351-2, and letter to Rehberg [before]
September 25, 1790). For spatial images in Kantian theory are spatial not because
of the quality of the (visual, tactual, etc.) sensation through which outer appear
ances are apprehended but the particular mode in which sensational differences,
regardless of their particular quality and associated species-specific psychological
capacities, are exhibited in the unity of sensibility (prediscursive synthetic unity of
apperception).
160 Kant's Anatomy of the Intelligent Mind
What then of Kant's many affirmations of the necessity of the three dimen
sions of space and the necessity of the propositions (axioms, postulates, theorems)
of Euclidean geometry? Do such assertions not definitively shut the door on any
attempt to accommodate Kant's theory of sensibility to post-Kantian mathematical
innovations such as hyperbolic, elliptic, projective, and differential geometry, espe
cially Riemann's mathematically rigorous notion of n-dimensional manifolds and
his formula for representing their curvature intrinsically (i.e., without presupposing
a containing space)?4 To appreciate why the answer is no, one must first consider the
kind of view Kant was principally concerned to refute in the transcendental exposi
tions of space and time and related texts. It all starts from Kant's radical break with
precedent by construing mathematics to be not analytic but synthetic. For as soon
as this is done, empiricism immediately becomes the main threat to the necessary,
universal character of mathematical judgments (ch 2-A),5 particularly when forti
fied by the recognition that Hume's skeptical reasoning regarding causality extends
to mathematics when the latter is conceived as synthetic and so puts mathematical
necessity and universality as much in question as the causal variety (ch 2-D and
UU ch 2). Kant therefore saw it as his task to show that mathematical cognition,
though synthetic, is nonetheless a priori, and so necessary and universal; and the
only means of doing so, he believed, is on the supposition that space and time are
intuited completely a priori ("pure intuition must be acknowledged as the unavoid
able condition for a priori cognition," Discovery 245).
A case in point is Kant's assertion at B41 that "geometrical propositions are one
and all apodeictic, i.e. combined with the consciousness of their necessity, e.g. that
space has only three dimensions." Taken out of context, this looks like a dogmatic
Euclidean denial that space is capable of having more than three dimensions. Yet
the sequel-"such propositions cannot be empirical, or judgments of experience,
nor inferred from them"-makes clear that his point was simply that the necessity
of this or any synthetic proposition regarding space could never be recognized if
our only representation of space were empirical.
Equally important to the sense of B41 is Kant's statement that the proposition
limiting space to three dimensions is an example of a geometrical synthetic a priori
proposition (also A239/B298-9). T he synthetic a priori propositions of the meta
physical and transcendental expositions of space in the Transcendental Aesthetic
are philosophical-discursive rather than intuitive-mathematical in character,
4 A helpful account of post-Euclidean geometrical developments for the general reader can be
found in Ian Stewart's Taming the Infinite: T he Story of Mathematics (London: Quercus, 2008).
5 E.g.: "If this representation of space were a concept acquired a posteriori, created from
general outer experience, then the first principles of mathematical determination would be noth
ing but perceptions. They would thus all have the contingency of perception, and it would not
even be necessary that between two points there is only one straight line, but rather what experi
ence always teaches. What is derived from experience has only comparative universality, namely
through induction. One would therefore only be able to say that up to the present time no space
has been found that has more than three dimensions" (A 24 ). See also PFM 284 and Progress 278.
Mathematics and the Unity of Sensibility 161
the necessity of propositions of Euclidean geometry. But that guarantee was always
in principle extendable: provided geometers could devise ostensibly or symbolically
constructible non-Euclidean concepts of space, the purity of space could preserve
the necessity of the resulting axioms, definitions, etc. against empiricist skepticism
as well. Thus, statements like that at B41 should be read as intended to immunize
Euclidean propositions not against contingency of every kind, least of all against
future mathematical developments, but solely against the sort of contingency that
pertains to propositions that are dependent in any way, shape, or form on empirical
consciousness.
The same result can be obtained via a slightly different course of reasoning. The
charge that Kant was a dogmatic Euclidean can be made to stick only by showing
that he held the propositions of geometry-axioms, definitions, postulates, demon
strations, and theorems-to be derivable analytically from the concept of space at
the focus of the metaphysical and transcendental expositions of the Transcendental
Aesthetic.7 But if this were true, construction (synthesis) would not then be needed
to cognize the connections between geometrical concepts, and geometry would be
analytic, contrary to Kant's iterated insistence that it is synthetic. We must there
fore grant that for Kant, no amount of analysis of the Transcendental Aesthetic's
concept of space is capable of yielding any geometrical proposition whatsoever and
that geometrical concepts must be exhibited in intuition before any connections
between them can be cognized in axioms, definitions, and so forth. Why did he
nevertheless advertise his system as the only one capable of preserving the necessity
of Euclidean propositions? The synthetic character he attributed to them left him
open to challenge from the quarter of empiricism: if the connections between geo
metrical concepts are not analytic to the representation of space, how can they be
forged synthetically except by means of experience, and so at the price of their nec
essary, universal validity? In response, Kant maintained that their a priori validity
can be preserved only through his theses that space is a pure intuition of sensibility
and that the understanding is equipped with pure concepts (categories) capable of
determining pure space transcendentally in such a way that geometrical concepts
can be exhibited in it completely a priori. But these theses serve only to ensure
that whatever propositions geometers succeed in establishing by means of ostensive
and symbolic construction will be valid of space completely a priori, and so with
strict necessity and universality; they do not, additionally, serve to specify which
propositions these are-Euclid's, Riemann's, Yau's, or any other. For, again, what
must never be overlooked is that the transcendental synthetic a priori cognition of
pure space obtainable by means of the categories is philosophical-discursive, not
mathematical-intuitive, and so can never imply, entail, anticipate, constrain, contra
dict, or affect in any way the findings of properly geometrical cognition. Thus, while
Kant was within his rights to insist that his transcendental philosophical account of
mathematics assures the necessity and universality of Euclidean geometry against
the challenge of empiricism, he never, to my knowledge, sought to defend it against
a priori, properly mathematical-intuitive innovations of the sort that eventuated in
Euclidean geometry being superseded by non-Euclidean.
Behind these considerations is a key result of Kant's psychologistic explication
of the pure space supplied by sensibility (ch 4-B): prior to its determination by
the understanding, pure space is so utterly uniform (undifferentiated) and devoid
of particular properties (indeterminate) that it suffices for no more than a bare
manifold of outer appearances, "scattered and single in themselves" (Al20). As
such, sensibility does not itself in any way limit determinations of its manifold in
synthetic a priori cognition-not to any set of axioms, any number of dimensions,
any group theoretical invariants, or any topological continuities. Its import for cog
nition is confined to one thing and one thing only: whichever determinations the
understanding may effect in pure space, they will, thanks to the purity of the space
intuition, be a priori and so necessary and universal. Since this is just to say that
the question of whether Kantian pure space is or is not capable of accommodating
hyperbolic geometry, more than three dimensions, etc. does not concern his theory
of sensibility at all, it therefore needs to be redirected towards the understanding.
The intellectual question regarding space is the issue whether transcendental
synthesis in conformity with the categories implies or entails the kind of dogmatic
Euclideanism associated with Kant or whether the objective space it synthesizes
(ch 15) is too indeterminate to constrain properly geometrical synthesis in any way.
I argued in UU ch 4-A that it has no geometrical entailments whatsoever. To be
sure, when the Critique of Pure Reason was written, Euclidean geometry was the
only kind in existence, and it was natural for Kant to promote his transcendental
theory of the understanding as the only one capable of preserving the a priori neces
sity and universality of Euclidean geometry. But, as noted earlier, such claims can
all be read as saying no more than that his is the only theory capable of preserving
Mathematics and the Unity of Sensibility 163
them against empiricist skepticism once the synthetic character of a priori geometrical
cognition is granted. In no sense does it represent a commitment to a geometrical coun
terpart to fiat-eartherism that sets up Euclidean geometry as necessary and universal
in a sense that precludes the possibility of its synthetically produced necessities being
made conditional on more fundamental synthetic determinations of pure space that
enable its theses to be accommodated to hyperbolic geometry, n-dimensional manifolds,
and all the rest. Admittedly, if these innovations genuinely contradicted or otherwise
falsified Euclidean geometry, it would pose a serious problem for Kant's theory. But
since non-Euclidean geometry is most plausibly construed not as falsifying Euclidean
geometry but as subsuming it as a special case, the necessity and universality ascribed to
the latter under Kant's principles is simply transferred to the former (and from it to any
subsequent geometry that may subsume today's). Kant's transcendental theory of the
understanding merely provides a foundation for the a priori (necessary, universal) valid
ity of whatever spaces geometers of a given era succeed in constructing. It does not, nor
could it, offer any basis for selecting either among different kinds of geometry or differ
ent methods of geometrical construction (i.e., ostensive versus symbolic). And, so far
as I am aware, there is no textual or other evidence that Kant ever supposed otherwise.
Similar considerations pertain to applied geometry (physical space) under the Axioms
of Intuition principle of pure understanding. As a purely philosophical-discursive prin
ciple that is neither itself an axiom nor in any way mathematical (A732-4/B760-2), it
merely affirms that whatever geometers (or mathematicians generally) may succeed in
constructing in pure formal space will ipso facto also be applicable to physical space. It
does not, nor could it, determine in addition which of the many possible geometrically
constructible spaces will actually be true of physical space and so is just as compatible
with its being hyperbolic as Euclidean, or having four, twelve, or any other number of
dimensions instead of only three. For in the end, all that Kant's transcendental theory
of physical space does, or seeks to do, is provide the means to reckon with any empiri
cist skeptical challenge to the a priori applicability of geometry to physical space.
The one feature of pure intuition indispensable to all of the varieties of synthetic a
priori cognition it makes possible, transcendental no less than mathematical, is its
manifold-unifying individuality (the essential oneness of prediscursive original syn
thetic unity of apperception), since without it, no objective individuals answering to
the concepts contained in such cognition would be possible at all. What principally
distinguishes one variety of mathematical synthetic a priori cognition from another
are the other features of pure intuition from which it does and does not abstract.
Geometry, at least in the form Kant knew it,8 abstracts the least, taking space just as
8 My concern in the remainder of this chapter, unless otherwise indicated, is with geometry,
arithmetic, and algebra as Kant knew them.
164 Kant's Anatomy of the Intelligent Mind
not the particular image that results from it, was termed by Kant the construction
of the object corresponding to the concept9:
9 To exhibit an empirical concept, by contrast, is to form an image of an object from the sen
sational matter thought in it as well as its intuitive form and so to exemplify it by the reproductive
imagining of an actual corresponding existent: Progress 325.
166 Kant's Anatomy of the Intelligent Mind
for the construction of his concept, and so knew that he could determine its
magnitude as closely to that of the object itself as he wished, and could give
it in intuition in accordance with the concept, and thereby demonstrate the
reality of the rule itself, and likewise that of this concept for the use of the
imagination. (Discovery 212; also letter to Reinhold May 19, 1789)
A light went on in the mind of the first person (be he Thales or another) who
demonstrated the isosceles10 triangle, for he found that it did not consist in
what he saw in the figure or in what he could detect (nachspiiren) in its concept
and, as it were, learn from this what the properties of the figure are. Instead,
it consists in producing the figure through what, conformably to his concepts,
his own thought put in (hineindachte) and exhibited (through construction)
a priori, and that in order to know anything with certainty a priori he had to
ascribe nothing to the thing except what necessarily followed from what he
himself had put into it conformably to his own concept. (Bxi-xii)
10
The original reads "equilateral," but Kant corrected it to "isosceles" in a letter to Schutz
dated June 25, 1787.
Mathematics and the Unity of Sensibility 167
them as parts of a greater space. Either way, the result is the construction of a single
extensive magnitude composed, 11 wholly or in part, from these lesser spaces, thus
synthetically uniting the objects corresponding to the constructed concepts a priori
as the concepts by themselves could never be (pure space is, to be sure, only part of
what is required for mathematical construction; the mathematical determination of
its manifold as an extensive magnitude presupposes categorial synthesis speciosa as
well: ch 15-E and -F ).12
In the case of the Euclidean theorem equating the sum of the angles of a tri
angle to two right angles, the first step in its demonstration is to unfold the concept
of a triangle in pure intuition by delimiting pure space conformably to the rule of
pure synthesis specified by the concept ("mathematicians cannot make the slight
est claim about any object without displaying (darzulegen) it in intuition," letter to
Reinhold May 19,1789). Next, further delimitations of pure space are constructed
from the pure intuition of the triangle by setting the exhibitions of other, distinct
concepts in immediate juxtaposition with it: lines drawn parallel to sides of the
triangle, extending the sides to form exterior angles, etc. If done in the appropriate
way, the internal angles of the triangle are not only replicated outside the triangle
but also added together as parts to form a single angle in such a way that the mag
nitudes of the component angles make no alteration in the magnitude of the whole.
That the composite angle synthesized according to this procedure in a Euclidean
plane is evidently equal to two right angles, and equally evidently cannot fail to be
equal to them, thus serves to demonstrate that this relation necessarily holds for any
triangle, regardless of the size of its angles.
Although geometers will (rightly) have eyes only for the parts of their construc
tion directly concerned in the result-the straight line constructed by replicating the
internal angles of a triangle from a single vertex-the transcendental philosopher is
concerned with the construction in its entirety, including the parts geometers ignore
after making inferential use of them (the auxiliary lines, the angles not involved in
forming the straight line, etc.). For what is distinctive about a constructive infer
ence is that each new step comes about by delimiting a new space in juxtaposition
with those already delimited and cognizing their relations in respect of the newly
constructed total composite space. Further spaces can be added to whatever degree
of complexity is needed to cognize the relation requisite to demonstrate a proposed
ll Magnitudes count as extensive if they "can be cognized only through successive synthesis
(of part to part)," that is, "intuited as aggregates (quantities) of previously given parts" (A162-3/
B203-4).
12
Since the concepts of spaces exhibited in pure intuition can become the source of fur
ther rules by means of demonstrations, Kant sometimes used the higher designation of "prin
ciple": "The figure of a circle is an intuition that is determined by the understanding according to
a principle: the unity of this principle which I arbitrarily assume and, as an underlying concept,
apply to a form of intuition (space) that is likewise found in me merely as a representation and
indeed a priori, makes the unity of many rules yielded from the construction of that concept"
(CJ 364).
168 Kant's Anatomy of the Intelligent Mind
Synthetic a priori cognition is possible because there are two intuitus a pri
ori: space and time, in which a synthesis of composition is possible a priori.
The two objects are quanta and indeed originaria. All concepts of quanta may
be constructed in them, i.e. given a priori in intuition, so too all concepts of
quantity, i.e. number, which requires time as well as space. The universale is
here given in the singulari in intuition, and the universal of synthesis consid
ered in the singulari. (AA 18 § 5593 [1778-83])
13 Another difference is that concepts of the spatial constitution and relations of objects of
experience have to be given (derived from these objects), whereas pure concepts of space are
always our own arbitrary invention (L 141). Similarly, because the objects of experience have an
existence of their own independently of their presence to us in sensation (A225/B272; chs 17 and
18), their spatial relations must be given in experience, whereas the relations cognized a priori
in pure geometry do not exist until we exhibit our pure concepts of space in pure intuition and
construct a composite space in which they are united in certain a priori relations. Still, on Kant's
account, neither the arbitrariness of the concepts nor the constructed character of the spaces in
which these concepts are exhibited makes the relations between the parts of these spaces any less
factual (objective, real) than the spatial relations of objects of experience.
Mathematics and the Unity of Sensibility 169
minds constituted like our own, its validity can extend no further than the represen
tations of such minds (transcendental idealism: ch 7).
Pure time is a principle of the unity of sensibility just like pure space: its parts
exist only through limitation, and distinct delimitations of time can all be set into
direct temporal relation by delimiting the intervening time. Nevertheless, taken by
itself, apart from space, pure time is useless for synthetic a priori cognition. Given
that times are subordinate and never coordinate, without pure space to make pos
sible the simultaneous representation of the distinct (Al83/B226, AA 17 § 4171, and
AA 18 § 6313; chs 3-D and 4-A), no two parts of time could ever be represented
together, and each would be in itself at once a first beginning and an absolute end.
Since this would make it impossible to unify the exhibitions in pure time of distinct
pure temporal concepts as the component parts of a single, greater time, no propo
sition containing these concepts could ever be demonstrated. Space makes good
this want by enabling us to represent time by means of a continuously lengthening
line ("time, though not at all an object of outer intuition, cannot be made repre
sentable otherwise than under the image of a line insofar as we draw it, and apart
from this mode of exhibition we could not cognize the oneness of its dimension
at all," B156; also A33/B50, Bl54, B291-3, ID 405, and AA 18 §§ 5653, 6312, and
6359). Of course, since the geometry of such a line is far too one-dimensional to be
of scientific interest, pure time comes into its own as a source of synthetic a priori
cognition only when treated as an extra dimension to those of pure space. Pure
time thus proves its worth as a basis for synthetic a priori cognition by extending
geometry to encompass a general doctrine of motion (B48-9 and B155n) and a
pure mechanics (ID 397-8 and PFM 283).
When abstraction is made not from the imagability of pure space and time but their
character as !imitable quanta, the kind of mathematical synthetic a priori cognition
that continues to be possible is arithmetic (that numbers can be presented in images
is stated at Al40/Bl 79). What then is considered in the original synthetic unities of
space and time is simply the manifoldness constituted by the infinite differentiabil
ity of pure space by juxtaposition and pure time by succession, but without the
possibility of fixing points or instants in them, lines or durations connecting them,
figures bounded by lines, solids bounded by surfaces, etc. Since determinations of
their manifold can no longer be effected by delimitation, the exhibition of concepts
in pure space or time ceases to consist in the production of determinate spaces or
times and demonstration no longer proceeds by combining these different spaces or
times into a single composite space or time. It also means that, strictly speaking, no
concepts of spaces or of times ever enter into arithmetical judgments ("Concepts of
number ...do not presuppose the concept of time," AA 18 § 5656). For although the
uniformly differentiated pure manifolds that constitute their sole and entire sensibly
170 Kant's Anatomy of the Intelligent Mind
14 "We have 3 quanta. Space, time, and sensation (motion, reality). The first has a positive for
its boundary which is a quantum; the second what is not a quantum, the third nothing positive at
all and not bounds but limits" (AA 18 § 5582 [late 1770s]). None of this is to say that arithmetic
cannot be applied to spaces and times, just as it can to apples and oranges or anything else that
can be counted.
15 Since the delimitation of a time is as much an action as the delimitation of space, there are
inner affections here too and so the prospect of a regress. See ch 4-C for discussion.
6
1 This is not to deny that it is still possible to cognize a space as an indeterminate quantum
independently of successive synthesis: "We can intuit an indeterminate quantum as a whole if
it is enclosed within bounds without being obliged to construct its totality by measurement, i.e.
successive synthesis of its parts). For the bounds already determine the completeness by cutting
it off from everything more (sie alles Mehreres abschneiden)" (A428/B455-6n). Cf ID 406, A99,
B136-8, B202-3, and A l 62-3/B203-4.
Mathematics and the Unity of Sensibility 171
17
Magnitude in general is the synthesis of the homogeneous: 'The magnitude of a thing is
the unity that can be produced through the mere repetition of one and the same" (AA 18 § 5727
[1780s]; also A723-4/B751-2). See ch 15-E for an analysis of the transcendental synthesis spe
ciosa that constitutes the intellectual basis of magnitude in pure intuition.
18
Both continuous and discrete magnitudes presuppose categorial synthesis speciosa of the
manifolds of pure space and time as well. The contribution of this intellectual synthesis is my
focus in ch 15-F.
172 Kant's Anatomy of the Intelligent Mind
19 The category of unity is equated with unit of measure at PFM 303, presumably with its
application in experience in mind via its schema and the Axioms of Intuition principle. Without
objective units of measure, there would be nothing to determine when an amount of anything
given empirically is, say, objectively equal and not just equal in the general way it appears from
a particular point of view, nor is there anything to prescribe what does and does not resemble
the thing being quantified sufficiently closely to count as an increase or decrease in its amount
(rather than another, unrelated amount-apples and oranges). Yet even if the lack of objective
measures denudes quantitative representations of all cognitive worth, this in no way prevents us
from following the dictates of our imaginations and, however arbitrarily and subjectively, con
ceptualizing things quantitatively (i.e., aesthetic quantities discursively expressed).
Mathematics and the Unity of Sensibility 173
Thus does Kant's psychologistic method lead inexorably to the conclusion that
time, the form of inner sensibility in beings constituted like ourselves, is implicit in
all concepts of number, with the consequence that creatures with any other mode
174 Kant's Anatomy of the Intelligent Mind
Number requires space and time for its intuitive representation. (AA 17 §
4629 [1771-3]; also AA 18 § 5593 and B293)
The space required by number is not the space needed to exhibit time as a quantum
(letter to Rehberg before September 25, 1790, AA 18 § 5656) or to express it in
an image (B l 54-6). Number concepts contain no sensible content whatsoever-no
temporal, spatial, or sensational quality. But just as their exhibition for purposes
of cognition depends on pure time as the (synthetic unitary) form of all succession
(succession within and outside succession ad infinitum), it likewise requires pure
space as the (synthetic unitary) form of juxtaposition Uuxtaposition within and
outside juxtaposition ad infinitum). It suffices to represent a manifold as simul
taneous that space be considered apart from all delimitative synthesis, for, in the
absence of limits (points, lines, surfaces), nothing spatial that could compromise
the purely intellectual character of number concepts is or could be apprehended
through pure synthesis of juxtaposition. Accordingly, while concepts of number are
no more concepts of spaces than of times, or the concept of the number series a con
cept of spatial juxtaposition, the space synthesis itself (coordination), considered in
abstraction from the limits that make it possible to cognize its products as extensive
magnitudes of space, is nevertheless essential to their content. The apprehension
of number is impossible by means of succession alone since the units must be rep
resented simultaneously together, as coordinate, and space is the condition of all
apprehension of coordinates ("Space contains the form of all coordination in intu
ition, time of all subordination," AA 18 § 5886; also ID 407; ch 4-B and -C). Thus,
by disclosing that all concepts of number are bound up by content with both the
subordinative synthesis of pure time and the coordinative synthesis of pure space,
Kant's psychologistic explication implies that the scope of the validity of synthetic
a priori arithmetical judgments is restricted to the unity of sensibility demarcated
by these prediscursive, purely aesthetic syntheses.
One of the striking features of Kant's account of the origin of number con
cepts is that the manifold of numbers can be apprehended in intuition simply by
abstracting from the limits (points, instants, etc.) imposed by the a priori synthe
ses responsible for producing spatial and temporal quanta continua. Consequently,
neither oral nor written signs for designating individual numbers are necessary to
apprehend them, and it is even possible, at least in principle, for there to be crea
tures capable of quite advanced arithmetical cognition who make do without signs
altogether. Human psychology, however, is different. Our powers of numerical dis
cernment and memory (reproductive imagination) are so feeble that arithmetical
cognition would be almost entirely beyond our reach without the assistance of lin
guistic representation and, most especially, specially crafted schemes of notation.
By enabling us to express any number we may have occasion to mention, in what
ever form suits us (fractions, decimals, etc.), as well as any operation upon or rela
tion between numbers, however complex and variegated, they magnify our ability
to form arithmetical concepts and exhibit them a priori in intuition immeasurably.
176 Kant's Anatomy of the Intelligent Mind
20
The reference to the understanding should not be construed, as it often is, to imply that the
concept of v2 is not inextricably bound up with sensible forms (succession and juxtaposition), for
there is ample evidence that, in Kant's view, all concepts of number are: B293, AA 17 § 4629, AA
18 §§ 5593, 5656, and 6314. To describe a concept as "intellectual" only means that it has no sen
sible content, which, as shown above, is the result of abstracting from spatial and temporal limits.
Thus, as Kant explained earlier in the Rehberg letter, the understanding has recourse to succes
sion (considered quantitatively rather than as a quantum): "the understanding that arbitrarily
forms the concept v2 cannot also generate the complete numerical concept, namely by its rational
relation to unity, but must rather submit to adopting an infinite approach to the number in this
determination, guided by another capacity as it were, grounded on time: successive advance as
the form of all counting and numerical magnitude, the underlying condition of this production
of the quantity." In the same letter, Kant characterizes the square root of a negative number
as represented in algebra as "impossible" and "contradictory," though it is unclear whether he
means logically impossible (literally unthinkable) or mathematically impossible (unconstruc
table). Either way, it seems that Kant denied that imaginary numbers (and, a fortiori, complex
numbers) are numbers at all. What he would have made of William Rowan Hamilton's analysis
of imaginary numbers is hard to say, but I see no reason to think that his transcendental concepts
of space and time could not accommodate it.
Mathematics and the Unity of Sensibility 177
21 According to Michael Friedman, the Rehberg letter discussed earlier confirms that Kant
worked with a distinction between algebra and arithmetic such that the former deals with irra
tional as well as rational numbers while the latter is confined to the rationals: Kant and the Exact
Sciences (Cambridge, Mass.: Harvard University Press, 1992) 110-12. Yet the only passage in the
latter explicitly contrasting arithmetic with algebra suggests a contrary conclusion: "To be sure,
the mere concept of a square root from the positive quantity �a, as algebra represents it, does not
require any synthesis in time at all . ... But as soon as, instead ofa, the number for which it is the
sign is given, so as not merely to designate (bezeichnen) its root, as in algebra, but also to find it, as
in arithmetic, the condition of all production of quantity, time, becomes the indispensable basis,
and indeed as pure intuition, in which we learn not only the given numerical quantity but also its
root, whether this can be found as a whole number or, if this is not possible, only through a series
of fractions decreasing ad infinitum, and so as an irrational number." Here it seems that the shift
from the generalized indeterminate quantity signified by "�a" to a particular irrational number
like �2 (Kant's example in the letter) coincides with the shift from algebra ("universal arithme
tic") to (un-universalized) arithmetic, with no indication whatever that the latter does not deal
with irrationals. To be sure, Kant holds that neither algebra nor arithmetic is capable (at Kant's
time of writing) of giving an adequate intuitive exhibition of an irrational number and that only
geometrical considerations can demonstrate the reality of such numbers. But the introduction
of space here is not, it should be remembered, the first: in Section C above, it emerged that Kant
regarded juxtaposition (pure space) as no less essential to the intuitive exhibition of numbers as
succession (pure time). The only novelty in the case of the exhibition of irrational numbers is that
the spatial component of exhibition, as geometrical, does not abstract from limits.
178 Kant's Anatomy of the Intelligent Mind
universalized magnitudes themselves, the variables and symbols that stand for them
admit of combination by juxtaposition and succession and so enable us to manipu
late algebraic magnitudes indirectly by performing intuition-grounded quantitative
transformations on their symbols. In this way, according to Kant, we can construct
equalities and other relations between fully universalized (abstracted) algebraic
magnitudes that carry the same a priori evidentiary force as ostensive constructions.
The method of symbolic construction that distinguishes algebra from geom
etry and arithmetic opens up wholly new possibilities for mathematical cognition
("the remaining parts of mathematics (Mathesis) expect their growth for the most
part from the extension of that universal science of quantities," letter to Schultz,
November 25, 1788).23 These include algebraic definitions (symbolic exhibitions)
of transcendental and complex numbers, non-arithmetic algebraic operations and
relations (abstract algebras, e.g., of rings, fields, and groups), as well as algebras
that find their concrete realization not in numbers or geometrical magnitudes but
in magnitudes of entirely new kinds (differential operations, functional equations,
etc.). Yet no matter how general and abstract, these and all future developments in
algebra are possible, from a Kantian perspective, only insofar as the symbols relate
to an a priori manifold that, no matter how generally and abstractly considered,
remains the manifold encompassed in the unity of sensibility made possible by the
pure formal synthetic unitary space and time intuitions.
species-specific human psychology that prevent us from advancing a great deal further in arith
metic without depending on signs-language, notation-of any kind at all). Friedman does,
however, reason consistently from his dubious premise, concluding from the fact that both arith
metic and algebra deal with quantitas that the latter, just like the former, has "no room for general
laws . . . corresponding to general laws of geometry" (109n24), and then proceeds (109-14) to
conjecture that Kant termed algebra "universal arithmetic" because it deals with a more general
class of magnitudes, including irrational numbers (which, as I argued in the preceding footnote,
does not seem to have been Kant's reason).
23 Since Kant never made clear where he believed the boundaries of algebra lie, he may have
been working with a different conception than ours, encompassing fields that historians of
eighteenth-century algebra might not consider part of algebra at all, e.g., analysis and number theory.
180 Kant's Anatomy of the Intelligent Mind
operators. Yet from a Kantian standpoint, the completely abstract objects stipu
lated in the symbolic constructions of quantification theory constitute a manifold
in precisely the same sense the objects of geometry, arithmetic, and algebra do and
consequently would seem to be no less fully dependent than they on synthetic uni
tary pure space and time as conditions for a single universe of objects constituting
the domain over which quantifiers range (a domain that therefore cannot exceed the
bounds of the unity of sensibility). Moreover, the objects in this universe are each
supposed to be uniquely differentiated from and determined in relation to every
other (such that each can be picked out from all the others and determinately falls
or does not fall within the extension of any given n-place predicate). Since such
differentiation and determination is precisely the sort that depends on categorial
synthesis speciosa of the manifold of pure formal intuition (chs 13-15), there thus
seems ample reason to conclude that Kant would not have regarded quantification
theory as part of pure general logic.
T hat still leaves the question of whether Kant would have classified quantifica
tion theory as transcendental, like categorial synthesis speciosa itself, or as math
ematical, like algebra. Yet to recognize that he would have grouped it with the latter,
we need only remember that categorial synthesis counts as philosophical rather than
mathematical synthetic a priori cognition because it is purely discursive, that is, its
concepts are not constructible in any fashion, symbolically no less than ostensively
(ch 15). Quantification theory, by contrast, is replete with symbolic constructions
that have an obvious kinship to algebraic and other kinds of symbolic mathemati
cal construction. If Kant was ready to count algebra as intuitive-mathematical syn
thetic a priori cognition even though its constructions are all symbolic, I therefore
see no reason to think he would not have done the same with quantification theory
or any other symbolic science, including set theory, modal logic, and metalogic.
Since my concern in this chapter has been with the transcendental expositions
of space and time, the only point I hope to have established thus far is that Kant's
view that pure intuition and its manifold are implicated in the whole of math
ematics, no matter how symbolic (non-ostensive) its constructions, could easily
extend to include all of what today is considered logic rather than mathematics.
But before this question can be pursued further, we will first need to consider what
logic was for Kant and determine whether any of contemporary logic, down to
and including truth functional logic, would count as logic by his criteria (Part
111). When that is done, we shall then need to consider both general logic and
mathematics from the point of view of their transcendental underpinnings as set
forth in the Transcendental Deduction of the Categories (ch 14-E). Finally, we
will have to examine the transcendental synthesis speciosa category by category
in order to understand fully the kind of differentiation and determination of the
manifold Kant seems to have had in mind when he posited that all objective cogni
tion depends equally on pure intuitions and pure concepts of the understanding,
mathematics not excepted (ch 15).
{ 7}
The metaphysical and transcendental expositions of space and time show them to
be empirically real but transcendentally ideal. Since Kant used this pairing to dis
tinguish his philosophy from every other, it is essential to comprehend what they
mean. To do so, one must first situate them in the broader framework of Kant's
conceptions of idealism and realism generally, determine how they divide into tran
scendental and empirical varieties, and then explain why the former always goes
with its opposite: transcendental idealism with empirical realism and transcenden
tal realism with empirical idealism. At the same time, one must also solve the puzzle
of why Kant seemed to preclude non-idealism-being neither a transcendental nor
an empirical idealist-as a possible philosophical position. Those are my objectives
in the present chapter. In the next, I will turn to the question of what distinguishes
Kant's view from Berkeley's, not as regards space and time, but things in them
selves, which I shall present as Kant's realist alternative to Berkeley's esse is percipi
idealism.
Idealism has been defined in myriad ways. Perhaps the best and most useful
in relation to the period from the pre-Socratics to Berkeley is the denial of the
mind-independent reality of the sensible world. The principal virtue of this defini
tion is that it does not conflate idealism with the denial of the reality of the material
world. Intellectualists (ch 2-A, UU chs 1 and 15), from Plato to the precritical Kant,
can be classed as idealists in the former sense without this by itself committing
them to anti-materialism. Early modern Rationalists, in particular, operating with a
distinction between sensuous imagery (sensations together with their reproduction
in imagination) and ideas of pure intellect (especially those developed by math
ematical means), took the contents of the former to reflect nothing but the sensible
constitution of our minds and the latter to reflect the nature of mind-independent
reality, including the material objects that affect our senses. These philosophers
were then generally able to adduce reasons for holding that matter, in this purely
intellectual sense, does in fact exist, whether as the cause of sensations (Descartes)
Idealism and Realism 183
or so that the ideas we form at the promptings of sensation will be true to the extent
of having actual material objects corresponding to them (Spinoza, Malebrache).
Even Leibniz, who is often classified as an idealist in respect to the material world,
should be counted as a materialist in this sense. No one insisted more strongly than
he on the need to distinguish ideas, as the objects of pure intellect, from sensuous
imagery. For however useful, even indispensable, the latter may be to stimulate the
intellect to form ideas of the material world, the senses contribute nothing to the con
tent of those ideas, and everything they do contribute must be eliminated before we
can arrive at ideas that truly (clearly and distinctly if not adequately and intuitively)
reflect the nature of the material reality that exists independently of the senses (UU
ch 15-A).1 Up to this point, Leibniz was as much an idealist about the sensible world
and a realist about the material world as other intellectualists. He parted company
with them only on the further issue of whether, in the final ontological reckoning
of reality, matter can be conceived of as a species of substance. F inding the notion
of a self-subsistent material entity riddled with contradiction and paradox, Leibniz
insisted that space and the spatial as well as time and the temporal are merely con
fused representations of what, in ultimate metaphysical truth, are purely intelligible
relations of coordination and subordination of a non-spatial, non-temporal char
acter. Thus, in the final analysis, Leibniz treated the material world of spatially and
temporally related beings as nothing more than a well-founded phenomenon of an
underlying reality comprised exclusively of spiritual substances (monads).2
The definition of idealism as the denial of the mind-independent reality of the
sensible world works equally well for idealism in the sensibilist tradition. Its expo
nents agreed with intellectualists that the sensible world is ideal but rejected their
distinction between ideas of pure intellect and sensory imagery, and so too their
distinction between a mind-independent intelligible world comprised of entities
that reflect the contents of ideas and a mind-dependent world comprised of entities
that reflect only the constitution of the sensing mind. Their materialism was instead
premised on the belief that the sensational contents supplied by the senses can be
employed to form concepts of a mind-independent reality that, in some measure
at least, resembles mind-dependent sensations in such features as spatial exten
sion (being composed of distinguishable parts, shape, position, etc.) and temporal
succession (change, motion, etc.). Locke, for example, held (1) that such ideas of
diverse senses as space and time could be considered in abstraction from ideas like
color and heat that are specific to a single sense and incapable of pertaining to any
thing other than its data3 and (2) that, so considered, their contents can be used to
frame the idea of corporeal realities capable of existing outside and independently
of any sensing mind. He then proceeded to infer, on probabilistic grounds, that
these beings are the causes of sensations in us and, on that ground, can be regarded
as actually existent (UU ch 7-D and 9-B).
And then came Berkeley. His anti-abstractionism promised to sweep away both
varieties of materialist idealism at a single stroke by nullifying any attempt to distin
guish contents of thought that can be utilized to frame notions of mind-independent
matter from sensational data incapable of reflecting anything beyond the sensory
constitution of the human mind. While granting that the distinctions posited by
materialists are possible semantically as indifferent denotations formed on the basis
of resemblances among ideas, Berkeley insisted that they fail to correspond to a
genuine distinction of ideas, with the implication that any argument for the ideality
of the sensible world is ipso facto an argument for the ideality of the material world
as well (ch 2-C and UU chs 10-12). Since materialists found compelling reasons
to treat sensations like heat, color, smell, flavor, and headache as essentially bound
up with their perception by a conscious mind, the same must then apply, by parity
of reason, to the contents necessary to conceive material things that, according to
Berkeley, are impossible to abstract from them-including the idea of their actu
ality ("I might as easily divide a thing from itself," PHK I § 5). Thus, material
ism emerges in the wake of Berkeley's anti-abstractionism as not just false but also
unintelligible (ch 8-A).
Yet as extreme as Berkeley's idealism was, there were still ways it could be
taken further. In the first place, although Berkeley regarded tangible space as
mind-dependent, he still treated it as a genuine, hence real, feature of tactual
sensation and not, like visible space, an illusory product of imagination (UU
ch 14). Second, temporal succession, though evidently regarded by Berkeley as
mind-dependent, was also treated as an intrinsic and therefore real feature of men
tal activity (UU ch 13-C). And third, Berkeley treated the mind as an unqualifiedly
real temporally enduring spiritual substance with equally unqualifiedly real causal
powers (volition, imagination, etc.) and capacities to be affected with sensations by
the efficacy of another unqualifiedly real (divine) substance distinct from it (UU
chs 12-C and 13-D).
Hume's associationism idealizes the third element of Berkeley's realism by
replacing the unqualifiedly real substantial bonds within complex objects and
unqualifiedly real causal bonds between distinct objects with associative "qualities,
which can give ideas an union in the imagination," and so count merely as "uniting
principles in the ideal world ...the very essence of [which] consists in their produc
ing an easy transition of ideas" (THN 260/169). In the case of the external sensible
world, this leads to an idealism not unlike Berkeley's: bodies-and complex individ
uals existing at or over time generally-are inconceivable except by means of ideal
uniting principles and the various fictions of imagination they spawn (UU ch 17-B,
HTC ch 7); the causal relations that connect one body to another are inconceivable
except by means of ideal easy transitions of thought that are likewise projected
Idealism and Realism 185
onto the objects concerned in them (UU ch 16-F, -G, 17-D, and HTC ch 5); and
the real existence (actuality) accorded to both the bodies present to us in sensation
and the causal relations found by experience to connect them is likewise conceiv
able only by means of an inward feeling-"some sensation or peculiar manner of
conception" (THN 184/123; also EHU Viii§ 12)-that, again, is illusorily projected
onto them (UU ch 17-C and -D). Where Hume went farthest beyond Berkeley,
however, was in extending his idealizing associationism to the inner world of the
mind: the mind itself (one's self or person) is inconceivable except by means of the
same ideal relations and fictions that make external complex individuals (bodies)
conceivable; exercises of volitions and other causal powers of the mind are incon
ceivable except by means of the same illusorily projected ideal associative relations
whereby causal relations between bodies are conceived; and the reality one accords
to one's mind as well as to its actions and passions is the same inward feeling pro
jected onto the objects of internal sense whereby the real existence (actuality) of
bodies (objects of external sense) is distinguished from mere fictions (UU chs 4-A,
16-17, and HTC ch 6).4
The result is a root and branch transformation of idealism. For if mind, body,
and causal independence/dependence must all be regarded as ideal, ideality can no
longer be conceived in terms of mind-dependence or indeed causal or substantial
dependence of any kind. Reconceived in associationist terms, an object counts as
ideal if it is impossible to form an idea of it that does not include in its content facil
ity affect (the easy transitions of thought Hume deemed the essence of associative
relation), vivacity affect (the feeling that constitutes the difference between regard
ing something as real and regarding it as a fiction), and/or the projective illusion
that induces us to attribute to the objects we are conscious of what in fact exists only
in and through affects immanent to our consciousness of them. Correspondingly,
reality, insofar as it presents itself to human understanding, consists of the fleetingly
existent perceptions (sensations, reflexions, and thoughts) of the present instant,
their juxtaposition (in the case of visual and tactual perceptions: UU ch 18-C) and
succession, and the consciousness in which successive perceptions are united as the
manifold: "Without this quality, by which the mind enlivens some ideas beyond
others ... we cou'd only admit of those perceptions, which are immediately present
to our consciousness, nor cou'd those lively images, with which the memory pres
ents us, be ever receiv'd as true pictures of past perceptions" (THN 265/173).
Yet as radical as Hume's idealism was, Kant's concept of pure yet sensible intu
itions effects an even more fundamental transformation of idealism. Its nature and
magnitude can best be appreciated by directly contrasting Kant's idealism with
4 Exponents of so-called "New Hume" interpretations read him as a realist in all of these
respects, while others, most notably Don Garrett, portray him as an exponent of a species of
philosophical double existence view. These interpreters all take Kant to have gotten Hume fun
damentally wrong. In HTC and UU Part III , I elaborate an interpretation of Hume that supports
Kant and develop what, in my view, are decisive objections against realist readings of Hume.
186 Kant's Anatomy of the Intelligent Mind
Hume's vestigial realism. In UU ch 3-A and -B, I argued that it was Hume's realism
regarding succession that left him incapable of explaining "the principles, that unite
our successive perceptions in our thought or consciousness" without resorting to
precisely the kind of real principle of connection-substance or causation-pre
cluded by the two fundamental principles of his skepticism he found impossible to
renounce: "that all our distinct perceptions are distinct existences, and that the mind
never perceives any real connexion among distinct existences" (THN 635-6/400).
Kant's idealism, by contrast, avoids Hume's quandary. It posits a succession of
perceptions in inner sense that is only an appearance preceded and made possible
by a pure intuition of time that is as much the product of productive imagination
as of the receptivity of sense. With no real (imagination-independent) succession
of perceptions to explain, Kant's idealism offers no temptation to violate either of
Hume's unrenounceable principles by positing a transcendentally real substrate or
cause to connect and unite them. By thus extending all the way down to the suc
cession and juxtaposition of perceptions apprehended in immediate intuition, the
conception of time and space as pure intuitions idealizes the remnants of realism
that Hume saw no way to avoid affirming. Kant's idealism therefore establishes a
true parity between the spatial and the temporal, between the outwardly sensed
physical world of space and the inwardly sensed psychological world of mentation.
Even those who question the ultimate reality of the external world tend to regard
this parity as a step too far (which includes not only Kant's contemporaries but also
post-Kantian dualists, panpsychists, and protopanpsychists as well). As Kant ana
lyzed their resistance, his opponents hold that
the former [space and the spatial] may be merely an illusion, while the latter
[the time of one's mind and its successive representational states] is undeniably
something actual. What they have failed to recognize, however, is that while their
actuality as representations can never be disputed, both nevertheless belong
only to appearance, which has two sides, the one, where the object is considered
in itself (without regard to the mode of intuiting it-its constitution therefore
remaining always problematic), and the other, where the form of the intuition
of that object is taken into account, a form that is not to be sought in the object
in itself but in the subject to which the object appears and nevertheless belongs
actually and necessarily to the appearance of this object. (A38/B55)
outer things exist as I myself exist, and indeed both on the immediate testi
mony of my self-consciousness, only with this difference: that the represen
tation of myself as thinking subject is related merely to inner sense, while
representations that designate extended being are related to outer sense as
well as inner. Regarding the actuality of outer objects I have just as little need
to infer as I do in respect of the actuality of the object of my inner sense
(of my thoughts), for on both sides they are nothing but representations for
Idealism and Realism 187
Nor could Kant have been clearer as to the idealist implications of the parity
between inner and outer objects, setting both on the same ontological level as mere
representations rather than things in themselves:
[T]his space itself, together with that time and, at the same time, all appear
ances, are not things in themselves but rather nothing else than representations
and can never exist outside our mind, and even the inner and sensible intuition
of our mind (as object of consciousness), the determination of which is repre
sented through the succession of distinct states in time, is also not the genuine
self as it exists in itself-that is, the transcendental subject-but only an appear
ance that is given to the sensibility of this, to us unknown, being. (A492/B520)
[W]hether bodies (as appearances of outer sense) exist as bodies in nature
apart from my thoughts can be denied. But the question whether I myself as
an appearance of inner sense (the soul according to empirical psychology)
exist apart from my faculty of representation in time is an exactly similar
question and must likewise be answered in the negative. And, in this way,
everything, when it is reduced to its true meaning, is decided and certain. The
formal (which I have also called transcendental) abolishes the material, or
Cartesian, idealism. (PFM 337)
In these and similar passages, Kant's affirmations of parity clearly relate to the
objects of outer and inner sense themselves ("myself and my state") and not merely
to our judgments about them. They also leave no room to doubt that denying that
the inner life of the mind is present to inner intuition in the same way the external
world is present to outer intuition is, for Kant, just as definitive of the empirical
idealism he rejected as its affirmation is to the transcendental idealism he espoused.
The innovation that best encapsulates the radicalization of sensibilist ideal
ism wrought by Kant's doctrine of pure sensibility is the distinction it introduces
between appearances and sensations (chs 3-B and -E). Previous sensibilists equated
sensations with the objects that appear immediately in outer sense perception. This
means that sensational qualities-colors, sounds, smells, etc.-were considered
qualities of outer appearances themselves, on all fours with spatial and temporal
qualities-sensational extension, shape, motion, etc. So, too, were the ideas of real
existence, or actuality, derived from sensations (the fact of their presence in us: chs
3-A and 8): most sensibilists equated sensations with the external realities of the
vulgar-sensible objects themselves, outer appearances-and the absence of sensa
tion with fictitiousness (fantasies, dreams, illusions, prefigurements, unrealized pos
sibilities, etc.).5 Even Hume, for whom vivacity and not sensation as such connoted
5 Locke equated ideas of sensation with the idea of real existence, or actuality, as opposed to
their images in thought, dreams, etc. in his discussion of sensitive knowledge (UU ch 9-B), as did
Berkeley in connection with his esse is percipi idealism (ch 8-A and UU ch 11-12).
188 Kant's Anatomy of the Intelligent Mind
actuality, still identified outer appearances with visual and tactual sensations and
so attributed the same (transcendental) reality to the former he did to the latter
("Hume ...took the objects of experience to be things in themselves," CPrR 53).6
For Kant, by contrast, outer appearances are constituted by forms that exist only
in and through productive imagination and prediscursive apperception and so need
to be distinguished from the sensations through which they are apprehended in intu
ition (chs 3-5). In particular, sensations have neither spatial nor temporal quali
ties (B207-8, PFM 306, 309; ch 3-C), while appearances (their matter included) do
not have color, smell, or any other of the qualities met with in sensations (A28-9,
B44, AA 18 § 5058; chs 3-B and 8). Similarly, the fa ct of the presence of sensations
in us has the same transcendental reality that pertains to things in themselves (chs
3-A and 7), while appearances, on their material side, do no more than exhibit that
reality without actually possessing it, just as, on their formal side, they exhibit the
qualities that differentiate sensations without possessing them (ch 3-B, -C, and -E).
The result is a bifurcation within the sensible sphere strikingly similar to the dis
tinction between the subjective sensible and objective corporeal worlds favored by
pre-Berkeleian realists such as Descartes and Locke. Like their sensation-causing
transcendentally real corporeal substances secreted behind the veil of perception,
Kantian outer appearances are distinct from sensations. On their material side, they
have their own reality distinct from sensation, cognizable under the categories of
quality as intensive magnitude. It belongs to them objectively, regardless of the sen
sation through which they are intuited (which owes its quality not to the appearance
but to the degree of sensitivity and constitution of the sense organs of the creature
that perceives them), and even belongs to their matter independently of their pres
ence in intuition (under the Second Postulate of Empirical Thought: chs 3-B and
18-D). On the formal side, appearances are representable exclusively through spatial
and temporal properties (Lockean ideas of primary qualities) and possess none of
the properties proper to sensations qua sensations (Lockean ideas of secondary qual
ities). Indeed, in the manner of an intellectualist, the status of space and time as pure
intuitions precludes (contra Locke) the presence of spatial (and temporal) properties
in sensations at all, rendering them impossible to represent objectively ("Sensations
are not representations but the matter therefor," AA 15 § 177 [1769-72]).
This of course is not to deny that the primary qualities and intensive magni
tude that constitute the objectivity of Kantian appearances are no less inextricably
bound up with the subjective constitution of sensibility than sensational qualities
(PFM 289), that their dependence not only on receptivity but on productive imagi
nation and prediscursive apperception as well makes them even more subjective
6 Hume accorded to sensations the same privileged place in his account of external reality
that earlier sensibilists did but not for the same reason, that is, not because of anything intrinsic
to them, but only because human nature is so constituted as to regard sensations (and reflexions)
rather than thoughts (independently of their associations with sensations or reflexions) with the
maximal degree of vivacity distinctive of sense impressions. See UU chs 16-A, 17-C, and HTC
chs 1 and 7.
Idealism and Realism 189
than sensations (chs 3-E, 5-E, 8-C), or that their material side merely corresponds
to (exhibits, reflects, projects) the transcendental reality that sensations actually
have (chs 3-A, -B, and 8-D). But although appearances are indeed ideal, they are
so only transcendentally, and nothing short of a transcendental critique founded
on the doctrine of pure sensibility detailed in Chapters 3 and 4 is capable of dis
closing their ideality. So if the objects that make up the material world posited by
Descartes, Locke, and others count as transcendentally real because they are both
formally and materially distinct from sensations, then Kantian outer appearances
certainly deserve to be considered empirically real by virtue of their no less real
formal and material distinctness from sensations.
Just as Kant's empirical realism is inseparable from transcendental idealism, it is
equally clear that transcendental realism is bound up with empirical idealism: the
view that sensible objects (appearances) are neither formally nor materially dis
tinct from sensations. Since corporeal objects cannot be identified with sensations,
the empirical idealist supposition that sensible objects and sensations are one and
indistinguishable directly implies the transcendental realist conviction that the
material world must lie beyond the sensible world. T his means that in order to uti
lize sensations to cognize corporeal objects-for example, the sun from the bright
yellow disk-shaped sensation that appears-an inference is required from the given
of sense perception to something distinct from it that can never itself appear to
the senses. Since inferences of this form are an easy target for skepticism (A372),
empirical idealists were obliged to devise strategies to fend off doubt, typically by
invoking a benign, undeceitful deity or bargaining away the formal incommensura
bility between sensible and corporeal realities by conceding at least some degree of
resemblance between them (ideas of primary qualities).
Kant, by contrast, had no need for such expedients: since spatial appearances are
representations as immediately present to consciousness in intuition as sensations
are in synopsis, knowledge of their existence requires no inference of any kind,
much less one from mental representations to mind-independent things in them
selves forever concealed behind the veil of perception. To be sure, these appearances
are not material substances even in Kant's idealist sense (substance as phenom
enon) until determined conformably to the categories. Nevertheless, the pure forms
of sensibility confer on outer appearances the formal and material distinctness
from sensations that renders them accessible to a priori categorial determination.
Still less may we hold appearance and illusion to be the same. For truth or
illusion are not in the object so far as it is intuited but in the judgment about
190 Kant's Anatomy of the Intelligent Mind
the object so far as it is thought. Thus, one may quite correctly say that the
senses do not err, not because they always judge correctly but because they
do not judge at all. (A293/B349-50; also PFM 290-1, Anthropology§ 11, and
AA 18 § 5642)
Strictly speaking, appearance is limited to that which would still be present to our
senses even if all memory of past experience were wiped from our minds and the
manifold now appearing were the very first perception in our lives. Appearances
present themselves regardless of whether the mind discerns one from another,
retains them, or even notices them, much less whether it conceptualizes them and
employs their concepts in judgments. Illusion, by contrast, can arise only after all
of these operations have been performed on the manifold of appearances appre
hended in immediate intuition and consists in something going amiss in the process
of transforming that manifold into objective judgments of experience.
For Kant, the distinction between appearance and illusion goes to the heart of
philosophy's idealism question:
The thesis of all genuine idealists, from the Eleatic School to Bishop Berkeley,
is contained in the formula: " all cognition through sense and experience is
nothing but sheer illusion and only in the ideas of pure understanding and
reason is there truth." The principle that everywhere governs and determines
my idealism is quite the opposite: "All cognition of things from mere pure
understanding or pure reason is nothing but sheer illusion, and only in expe
rience is there truth." ... How did I come to avail myself of that expression for
an entirely opposed viewpoint? Space and time, together with everything they
contain, are not things in themselves or properties of things in themselves
but belong merely to the appearances of such things; up to this point, I am
of one confession with those idealists. But they, and Berkeley in particular,
viewed space, together with all its determinations, as a mere empirical rep
resentation that, like the appearances in it, we are acquainted with only by
means of experience or perception. I, on the contrary, show that space (and
so too time, to which Berkeley did not attend7), together with its determina
tions, can be cognized by us a priori because space, as well as time, dwells in
us prior to all perception or experience as a pure form of our sensibility and
makes possible all intuition of them, and thus all appearances as well. From
this it follows that since truth depends on universal and necessary laws as
its criteria, experience can have no criterion of truth in Berkeley because its
appearances (according to him) have nothing a priori underlying them, from
7 It is unclear whether Kant meant that Berkeley never considered the question of time, which
is false, or whether he failed to develop his analysis of time as far as he did that of space, i.e., that
the same reasoning Berkeley applied to space he should, had he been consistent, have applied to
time as well-in which case he would have recognized the untenability of his version of idealism,
as Kant suggested at B70.
Idealism and Realism 191
which it follows that experience is nothing but sheer illusion, whereas with us
space and time (in combination with the pure concepts of the understanding)
prescribe a priori to all possible experience its law which at the same time
yields a sure criterion distinguishing truth from illusion in it . (PFM 374-5;
also AA 23 57-8)
All idealists agree, in Kant's view, that space, time, and everything in them are mere
appearances of things in themselves and are not themselves such things or their
properties. The issue on which they differ is whether these appearances are partly
pure or wholly empirical. And Kant's contention is that the only way for an idealist
to avoid the implication that "experience is nothing but sheer illusion" is by adopt
ing the view that space and time are pure forms of appearances that, together with
the categories (the a priori application of which to appearances the purity of space
and time makes possible), supply the universal and necessary laws whereby alone,
experience can become a source of truth rather than illusion.
Kant's distinction between appearance and illusion rests on the distinction
between appearances and things in themselves. Since "we can sense only inside, not
outside, ourselves, and our entire self-consciousness thus supplies nothing except
our own representations exclusively" (A379), appearances can be nothing more
than "mere modes of representation that we only ever find within us, and whose
actuality rests on immediate consciousness just like the consciousness of my own
thoughts does" (A372).8 What this means is perhaps best expressed in the following
passage dealing with the space of sensibility (and, by extension, sensible time as
well), which, as I read it, fills in the details of the "confession" that Kant admitted
to sharing with all idealists at PFM 374-5:
One needs to pay close heed to the paradoxical, yet correct thesis that nothing
is in space except what is represented in it. For space is itself nothing else than
a representation, and so what is in it must be contained in the representation,
and nothing whatsoever is in space except insofar as it is actually represented
in it. The thesis that a thing can exist only in the representation of it must cer
tainly seem strange, but it loses its shocking character because the existents
(Sachen) we have here to do with are not things (Dinge) in themselves but only
appearances, that is, representations. (A374-5n)
8 Also: "appearances ... cannot exist in themselves but only in us" (A42/B59); "as appear
ances, they constitute an object which is merely in us, since a mere modification of our sensibility
is not to be found outside of us at all" (A l29); "appearance that, separated from our sensibility,
is nothing" (A370); "appearances are not things in themselves but rather nothing else than rep
resentations and can never exist outside our mind" (A492/B520); "appearances are in themselves
actual only in perception as mere representations" (A493/B521); "appearances ... exist only in
our representations" (A506-7/B5 3 4); and "appearances ... are mere representations which are
always again sensibly conditioned" (A563/B591).
192 Kant's Anatomy of the Intelligent Mind
only inside us, and thence is essentially inseparable from "our entire self-conscious
ness" (i.e., apperception). This, however, leads straight to the question regarding
truth that every idealist must confront: since the judgments we make concerning
chairs, tables, our hands and feet, the sky, the sun, the moon, and everything our
senses present to us impute an external reality to these objects that is independent
of immediate consciousness (the representations of sense), does the idealists' deter
mination that space and everything in it are mere appearances, and so mere repre
sentations bound up with the immediacy of consciousness, not oblige them to treat
all such judgments as not just false but rank illusions?
Kant summed up the answers provided by all previous idealists in the proposi
tion that "all cognition through sense and experience is nothing but sheer illusion
and only in the ideas of pure understanding and reason is there truth." If truth
cannot be had in the sensible world because sensible space is merely a representa
tion inside us, then if it is to be found at all, it would have to be sought in another,
purely intelligible world. Thus Plato had recourse to the intelligible world of forms,
Descartes to true and immutable essences, Leibniz to purely intelligible relations
of coordination and subordination that are confusedly perceived as spatial and
temporal, and Spinoza, Locke, Newton, and Berkeley to God as the transcendent
source of the necessary laws responsible for all the order and relations we find in
outer appearances. The implication is that, on the basis of sense experience alone,
without recourse to the intelligible and the transcendent, no judgment regarding
the sensible world that imputes external reality to the objects met with in it can ever
be anything other than illusion.
Hume alone among idealists seemed to recognize this and make the best of
it. For in denying that there exist nonsensible "ideas . .. of so refin'd and spiritual
a nature, that they fall not under the conception of the fancy, but must be com
prehended by a pure and intellectual view, of which the superior faculties of the
soul are alone capable" (THN 72/52), and confining the scope of sensible ideas to
objects from which they might originally have been derived (i.e., to impressions
of sensation and reflexion), he proceeded to account for such formerly sacrosanct
notions as substance, cause and effect, identity, space, time, and the external world
itself as fictitious imaginings produced from infancy in response to the blind dic
tates of associative propensities rooted in human (and nonhuman animal) nature.
Although Kant may have been "of one confession with those idealists" where
the status of space, time, and everything in them as mere representations is con
cerned, his transcendental idealism enabled him to uphold a thesis directly contrary
to theirs: "All cognition of things from mere pure understanding or pure reason
is nothing but sheer illusion, and only in experience is there truth." The thesis of
Kantian idealism premises the presence of something a priori right in appearance
itself: space as "a pure form of our sensibility [that] makes possible all intuition of
them, and thus all appearances as well" (these being in themselves nothing distinct
from our representation of them in intuition). This obviates the need to employ infer
ence to move beyond the precincts of sensibility to the intelligible or transcendent.
Idealism and Realism 193
For pure intuition provides the understanding with the sensible ground it requires
to determine appearances conformably to its pure concepts entirely a priori, which
thereby become intellectual bases for universal and necessary laws of appearances
of precisely the kind necessary to furnish "a sure criterion for distinguishing truth
from illusion." Thus, far from casting the role of the understanding in the sensible
domain as that of spinner of illusion, Kant's idealism regarding pure space and
time transforms it into "the legislator for nature" (Al26) and the source of the tran
scendental truth whereby alone empirical truth is possible (Al46/Bl85).
Kant endeavored to leave idealists with no tenable option but transcendental ide
alism. For if indeed his critique of pure reason demonstrates that "[a]ll cognition of
things from mere pure understanding or pure reason is nothing but sheer illusion,"
and if the arguments of his predecessors demonstrate that "all cognition through
sense and experience is nothing but sheer illusion," what other choice is there? If
one were to respond by proposing a third option-rejecting idealism entirely-then
Kant presumably would have denied that such an option exists on the ground that
there are only two possible realisms-transcendental and empirical-and these are
inseparably bound up, respectively, with empirical and transcendental idealism.
According to the transcendental realist, space and time are mind-independent real
things or properties of things. This does not simply mean that in all of our empiri
cal judgments regarding sensible objects "we treat those appearances as objects in
themselves without troubling ourselves about the first ground of their possibility
(as appearances)" (A393; also A237-8/B297, A545-6/B573/4, and Progress 269).
For no one-not Kant, Berkeley, Plato, Descartes, Hegel, or anyone else-is an ide
alist in the context of actual experience, where the natural and inevitable assump
tion is that sensible appearances are material and thinking things with their own
nature and existence independent of our intuitions and other representations of
them, and no question can ever arise regarding "their true correlate, i.e. the thing
in itself " (A30/B45). Idealism enters the picture only when one interrogates that
assumption in the appropriate way, which, in the first instance, means asking not
whether it is justified (the epistemological question), but rather how it is possible
even to conceive sensible appearances as independent objects with their own nature
and existence and, moreover, to do so naturally, virtually from the first moments of
conscious life, without needing to be taught to do so (the psychologistic question).
It is this conceptual query that first raises the question whether empirical judg
ments are not all illusions (even those we are epistemically justified in taking to be
true). For if the external objects that appear to our sight, touch, and other senses
are no different from the visual, tactile, etc. sensations through which they are appre
hended, then any notions of "outsideness" and "independence" we extract from these
appearances empirically will merely be misnamed attributes of our own sensations,
that is, of representations inside us, bound up with consciousness, rather than genuine
outsideness and independence. The only way out is to suppose that the latter notions
do not derive from appearances at all but are present in our minds independently of
sensations. But this is to embrace intellectualism. Once we do that, we are obliged to
194 Kant's Anatomy of the Intelligent Mind
suppose that we come by our notions of objective outsideness and independence either
through their being innately implanted in us (Cartesian innatism) or as beneficiaries of
quasi-mystical intellectual insight (Malebranchean illuminationism). By contrast, Kant's
distinction between appearances and sensations neatly avoids this predicament. Here, the
notions of objective outsideness and independence apply exclusively to the appearances
apprehended through sensations rather than to the sensations themselves; and instead of
extracting them from the appearances empirically, pure understanding introduces them
into appearances a priori. For if, as Kant held, the objectivity of appearances can consist
in nothing other than their necessary synthetic unity in one apperception (Pts IV-V),
then the a priori determination of both their form and their matter in conformity to
the categories (as conditions for such unity) suffices to provide them with just the sort
of outsideness and independence to convert epistemically justified empirical judgments
from illusions to truths. Illusion is then confined to the transcendent judgments of pure
reason because they ignore the fundamental distinction between empirical objectivity
(outsideness and independence) and the transcendental kind (ch 4-B).
Kant is often portrayed as the exponent of a quite different type of conceptual
ist reasoning, commonly referred to as "transcendental argument." T he basic idea is
that necessities of our thought regarding sensible objects can validly and soundly be
extended to these objects themselves. Such reasoning is clearly fallacious, since we can
have no reason to suppose that objects are under any constraint to conform to the
ways our minds oblige us to think about them. Fortunately, nothing obliges us to sup
pose that Kant depended on reasoning of this sort, and texts such as the following
remark from Kant's General note on the System of the Principles make clear that he
was well aware of the problem with it:
Clearly, necessities of our thought of objects can never be a ground for inferring a
necessity to which the objects themselves are subject: "If we obtain [a priori concepts]
from ourselves, then what is merely in us cannot determine the constitution of an
object distinct from our representations, i.e. be a reason why there should be a thing
to which something as we have it in thought pertains, and why all of this representa
tion is not rather empty" (A128-9).9 To bridge the gap, it is necessary to show that
9 Also: "The error .. . lies in making the employment of the understanding transcendental,
contrary to its vocation, and [supposing] that the objects, i.e. possible intuitions, have to direct
themselves according to concepts rather than the concepts according to possible intuitions (as
that on which alone the objective validity of the concepts rests)" (A289/B345-6).
Idealism and Realism 195
the objectivity of the objects, their outsideness and independence vis a vis empiri
cal thought of them, is itself a product of transcendental judgments, i.e., that there
are no objects to think or know apart from transcendental syntheses. But this is the
claim that the understanding is the author not merely of necessities of our thought of
objects but of necessities (laws of nature) to which the objects themselves are subject
and indeed is constitutive of their very objectivity. Clearly, it is a far stronger claim
than mere "transcendental argument" could establish. On the contrary, nothing less
than Kant's doctrine of pure sensibility and the transcendental psychology it makes
possible can suffice.10
A passage from the Transcendental Deduction of the Categories makes clear the
principle involved:
The fact that it is quite literally beyond my power, organized as I am, to conceive
independently existing objects in any other way, far from demonstrating the truth
of my concept of cause and effect, is proof merely of the irresistible force of the
illusion wrought by it. This is not to deny that the concept might still correspond to
objective reality. Yet, its doing so cannot justify my application of it to objects since,
as a necessity of thought, I would do so even if it did not correspond to their reality
at all. In short, like a stopped clock that happens to indicate the correct time twice
daily, even if necessities of thought do agree with reality, they nonetheless remain
illusions (tauter Schein), not guides to objective truth.
10
The same considerations that pertain to the objects of empirical judgment also apply to
the objects of mathematical judgment: "if the object (the triangle) were something in itself, apart
from any relation to you, the subject, how could you say that what necessarily exist in you as
subjective conditions for the construction of a triangle, must of necessity belong to the triangle
itself?" (A48/B65).
196 Kant's Anatomy of the Intelligent Mind
Kant is often thought to have overreached himself when he denied that things in
themselves are in any sense spatial or temporal. If these objects are as unknow
able as he claimed, how could he know this? Yet, to discount his express and
iterated denials as mere hyperbole has the quite serious consequence of under
mining the distinction he drew between his own idealism and that epitomized by
Descartes. Kant termed the latter "skeptical" or "problematic" idealism because
it posits a spatio-temporal realm of corporeal objects that, secreted behind the
veil of perception, can never be known with certainty to exist or to be the true
causes of sensations. But because the same veil makes it impossible to know with
certainty that corporeal objects do not exist and are not the causes of sensations,
how could this problematic idealism be distinguished from Kant's transcendental
idealism if the latter is also supposed to be unable to preclude the possibility
that space, time, and their occupants exist beyond the veil, mind-independently
in themselves? Certainly, it would put him in no position to deny that these occu
pants could be corporeal substances and that they could cause sensations, since
the conditions for the application of the categories to appearances-the transcen
dental schemata-could, unbeknown to us, be satisfied by things in themselves as
well. Thus, a decisive test of the worth of any interpretation of Kant's transcen
dental idealism is to determine how he could preclude the possibility that space
and time exist mind-independently in themselves in a manner consistent with his
denial that anything can be known (intuited or conceived) regarding things in
themselves.11
Since many interpreters remain unconvinced that Kant did in fact deny that
things in themselves are spatial and temporal, it is useful to begin with a brief tex
tual review. Among the most explicit in this regard are the following:
We have thus wanted to say: that all our intuition is nothing but the repre
sentation of appearance; that the things we intuit are not in themselves what
we intuit them as being nor are their relations so constituted in themselves as
they appear to us. (A42/B59)
[S]pace and time ...are not determinations attaching to things in them
selves at all but mere determinations of their relation to sensibility.(PFM 284)
This theory can be called the doctrine of the ideality of space and time
because they are represented as something that does not attach at all to things
in themselves. (Progress 268)
[T]he concept of time does not apply to anything else, not even to the sub
ject of time itself. (AA 23 E LXXV, at A277; also B422)
u This is the idea at the core of what is generally known as the neglected alternative objection
Even more compelling are those texts in which Kant gave modal weight to his claim
that things in themselves are neither spatial nor temporal (emphases added):
[S]pace and time ...attach to the form of intuition alone and so to the
subjective constitution of our mind, apart from which these predicates cannot
be attributed to anything at all. (A23/B37-8)
What we mean by this phrase ["transcendental ideality"] is that if we
abstract from the subjective conditions of sensible intuition, time is nothing
and cannot be ascribed to the objects in themselves (apart from their relation
to our intuition). (A35-6/B51-2)
[I]f we were to eliminate our subject, or even merely the subjective con
stitution of the senses in general, the entire constitution [of the things we
intuit]-all relations of the object in space and time, indeed space and time
themselves-would disappear, and as appearances they cannot exist in them
selves but only in us. (A42/B59)
But this [transcendental] something is not extended, not impenetrable, not
composite, because all these predicates concern only sensibility and its intu
ition insofar as we are affected by objects of this kind (otherwise unknown to
us). These expressions do not at all give us cognition of what kind of object
it is but only that these predicates of outer appearances cannot be attached to
such a thing considered in itself, without relation to the outer senses. (A358-9)
The non-sensible cause of these representations is entirely unknown to us
and we therefore cannot intuit it as an object; for such an object would have
(miissen) to be represented neither in space nor in time (as mere conditions of
sensible representation). (A494/B522)
I leave their actuality to the existents (Sachen) we represent through senses
and only limit our sensible intuition of these things so that not even a single
constituent, not even in the pure intuitions of space and time, represents
something more than the mere appearance of those things, but never the con
stitution of those things in themselves. (PFM 293)
[S]pace contains nothing in itself that could be a representation of a thing
in itself or of the relation (of distinct) things in themselves to one another
and, if it is considered as such a determination, it is a non ens as ens imagina
rium. (AA 18 § 6316 [1790-1])
Clearly, there is, on the one hand, compelling reason to believe that transcen
dental idealism must preclude the possibility that things in themselves are spatial or
temporal in any sense, even the most attenuated, if it is not to collapse into prob
lematic idealism, and, on the other hand, abundant textual evidence to suggest that
Kant believed that it can and does preclude it. The question is how. Since nothing
can be known of things in themselves, the only way to secure this result is by show
ing that space and time, when psychologistically explicated as forms of appear
ances, can no more exist apart from consciousness than the pleasure of music or the
pain of a migraine can. Forms of appearances are, of course, not feelings, or even
198 Kant's Anatomy of the Intelligent Mind
sensations, since what "lies ready for them a priori in the mind" and is "capable of
being considered separately from all sensation" "cannot itself be sensation" (A20/
B34). So what precludes space and time from being forms of things in themselves
as well as appearances?
The answer, on any interpretation that supposes Kant to have accorded no
essential role to spontaneity in the constitution of pure space and time, is: noth
ing. For, so construed, the pure space and time that lie ready for sensations in the
mind are intuitions (not mere capacities: ch 5-E) yet are completely independent of
sensations and indifferent to whether they are present in us or not. As such, I see no
alternative but to understand them as innate representational contents, with noth
ing intrinsic that could restrict them to sensations. So, as representations present
in the mind prior to and independently of sensation, what is there to prevent them
from conferring the same unity on things in themselves that they confer on sensa
tions? Indeed, what justification would there be for treating them, together with the
appearances ( t sensations) they make possible, as sensible at all, much less essen
tially so? As innate, nonsensational representational contents, I do not see what
would remain to distinguish Kantian pure space and time from the notion of intel
lectualists such as Descartes that ideas of space and time belong to the receptivity
(innate endowment) of the mind, but, as nonsensational, are part of the receptivity
of intellect rather than of sensibility. It will not do to object that Kantian space and
time cannot be represented in the absence of sensation since intellectualists would
readily agree that human consciousness is never without an admixture of sensation,
no matter how pure the idea it considers may be of all sensational content (God, res
cogitans, res extensa, etc.). This, indeed, is why intellectualists insisted that intrinsi
cally intellectual ideas need not only to be made clear but also distinct before their
true nonsensational character can be recognized for what it is. So until exponents
of receptivity-only readings of the origin of Kantian space and time come up with
a convincing justification for restricting the origin of space and time themselves to
sensibility (rather than merely our confused perceptions of them), I see no more rea
son for supposing that they could not apply to things in themselves than any of the
other ideas to which intellectualists credited such an application: substance, cause
and effect, number, and so forth.
How can according a role to spontaneity in the constitution of pure intuitions
of space and time change things? If the psychological operations responsible for
forming a representation can be shown to contribute contents essential to that rep
resentation, it becomes inconceivable that the representation could correspond to
anything in mind-independent transcendental reality. In Chapter 2, I termed this
explanatory model psychologism and presented Hume's account of the impres
sion original of the idea of necessary connection as a paradigm. F inding no such
impression either in the data of outer and inner sense (impressions) or their copies
in thought (ideas), Hume located its source in the consciousness that regards these
objects, tracing it to an origin in the actions and affects immanent to imagination
on the occasion of transitions from impressions to customarily associated ideas
Idealism and Realism 199
(UU chs 16--17). He then proceeded to argue that ideas copied from this source
are capable of performing all of the roles in thought and action that had previ
ously been thought to require an idea of cause and effect whose contents are wholly
independent of our psychology (UU chs 18-19). The only exception is metaphysics,
in which these ideas are applied in ways incompatible with their being bound up
by content with conscious imagination (EHU XII/iii, especially if 34). For with
out pretending or needing to know anything whatsoever about things as they are
in themselves, Hume could contend that if indeed all ideas of cause and effect
are bound up with consciousness in this way, "we either contradict ourselves, or
talk without a meaning" (THN 267/173) when we suppose that the objects them
selves-be they perceptions (sensations, reflexions, and thoughts) or unperceived
realities-stand in causal relations independently of our consciousness of them
in associative imagination. In this way, it becomes possible to convert a disguised
absurdity into a patent one: causal relations can no more be intelligibly conceived
to exist mind-independently than the affects and operations immanent to custom
ary transitions of thought are possible independently of thought. And all of this
is recognized to be so without our needing to know the first thing about things in
themselves or even whether there are such things.
Kant's reasoning regarding pure space and time precisely echoes Hume's regard
ing cause and effect:
[S]pace and time, including all the appearances in them, are nothing existent
in themselves and outside my representations but themselves only modes of
representation, and it is patently contradictory to say that a mere mode of
representation also exists outside our representation. Thus the objects of the
senses exist only in experience; whereas to accord to them a self-subsistent
existence apart from or prior to experience is as much as to represent the
actuality of experience apart from or prior to experience. (PFM 341-2)
By arguing that pure space and time-and so too all appearances in respect to
form-are in themselves mere representations, with the implication that their very
notion is bound up by content with being "mere modes of representation," Kant
could claim, without needing to know anything at all about mind-independent things
in themselves, that the supposition that these things are in any respect spatial or
temporal is simply a disguised expression of the patent absurdity that "the actuality
of experience can be represented apart from or prior to experience."
This is the psychologistic conclusion that follows from the explication of space
and time as pure intuitions of sensibility. Our question was how Kant would have
had to conceive pure intuition in order to draw it. Since the receptivity-only reading
of pure space and time cannot preclude the possibility that things and themselves
are spatial and temporal, it seems to me that there is no alternative but to suppose
that his idealism is of the same general form as Hume's: as pure intuitions, space
and time are incapable of corresponding to the constitution of things in themselves
because they depend as much on spontaneity as on receptivity. In particular, thanks
200 Kant's Anatomy of the Intelligent Mind
Kant's refutation of Cartesian idealism is well known since he saw fit to add an
argument so denominated as a corollary to the Second Postulate of Empirical
Thought in the 1787 edition of the Critique of Pure Reason. In it, he criticizes what
he took to be Descartes's key premise: the claim that inner experience gives us a
species of infallible certainty regarding our own existence that is lacking in respect
to the existence of the external world (ch 7-C). However, we search in vain for
a comparable refutation of the principle of Berkeley's idealism: that the esse of
sensible things is perc ip i - their existing and their being perceived are one and indis
tinguishable. Instead, Kant merely states that he has nullified the ground (B274) of
Berkeley's idealism in the Transcendental Aesthetic, without specifying what that
ground is or how it had been nullified. This would pose no problem if it were obvi
ous where and how the principle of Berkeley's idealism is refuted in the Aesthetic.
Yet it contains no argument expressly dedicated to establishing, or even so much
as an explicit endorsement of the realist thesis, that the esse of sensible things is
NOT percipi.1 Quite the contrary, one finds iterated affirmations of the view that
space and time, together with everything in them, are, and can be (ch 7-C), noth
ing other than representations in a mind, with the clear implication that anything
existing in itself, independently of our minds and their representations, can have
any character except that of a material object: "the transcendental idealist coun
tenances this matter [corporeal substance] and even its inner possibility merely as
1 In the A edition at A377, Kant claimed in the Antinomy chapter to remove the difficulty
posed by the dogmatic idealism he elsewhere identified with Berkeley. Since the solution to the
antinomies depends on the transcendental idealism established in the Aesthetic, the inconsistency
is only apparent. Indeed, since it was only in the Antinomies that Kant made a point of show
ing that his transcendental idealist conception of space and time avoids the kind of contradic
tions Berkeley claimed to be inherent in space and time in PHK, De Motu, and other writings,
the A edition reference is perhaps more justified than that in B: the Aesthetic does, to be sure,
establish the ground of Kant's response to Berkeley's debunking analysis of space and time but
only in the Antinomies does Kant dot the i's and cross the t's on this point. What is missing in
both editions of the Critique, as well as the rest of Kant's writings, is an explicit refutation of the
principle of Berkeley's idealism: an argument for the realist thesis that the esse of sensible things
is not percipi.
202 Kant's Anatomy of the Intelligent Mind
Before attempting to dispel this mystery, it will be helpful to recall the outlines of
Berkeley's idealism from UU chs 10-11.
1. Berkeley asked "what is meant by the term exist when applied to sensible
things" (PHK I§ 3). Elsewhere, he declared "'tis on the Discovering of the nature &
meaning & import of Existence that I chiefly insist .... Let it not be said that I take
away Existence. I onely declare the meaning of the Word so far as I can comprehend
it" (C 491, 593; also 408).
2 Kant alternately branded Berkeley's idealism dogmatic, mystical, and fantastical: see
Transcendental Aesthetic, Refutation of Idealism, and Fourth Paralogism of the Critique of
Pure Reason, as well as the Remarks to Part I and Appendix of the Prolegomena to any Future
Metaphysics; also A l65-6/B206-7. Regarding Kant's familiarity with Berkeley's views, see
Chapter 2n10.
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 203
3 Although Berkeley denied that existence can be attributed to any sensible object not present
to us in sensation, he did believe that sensations occur according to laws laid down by God and
that science can give us insight into these laws. But even if science can disclose which sensible
objects will succeed which others, these objects, properly speaking, are nothing but sensations
that exist only in being perceived (UU chs 11-12).
204 Kant's Anatomy of the Intelligent Mind
but of associative imagination, since it is only in imagination, not in sense, that the
various sensible qualities that compose them-visual, tactual, olfactory, et al.-can
be united-considered as one object-and accorded a (for him) fictitious spatiality
(UU ch 14). Thus, the existence of sensible things and the presence of individually
distinct, fleetingly existent sensations to consciousness are, for Berkeley, quite liter
ally one and indistinguishable, and contrast with ideas of memory and imagina
tion that, although present to consciousness like sensations, lack existence in their
sense.4 And, in general, Berkeley argued that any attempt to conceive the reality of
sensible things without including in the conception the content sensation alone is
capable of contributing is to denude the idea of precisely that on which its mean
ing most fundamentally depends, with the consequence that our discourse loses the
ideational underpinning that is alone capable of conferring ontological worth on it.
5. Since the difference between real existence and fiction signified by the fact
of the presence or absence of sensation concerns not the thing contemplated but
only the mode of its presence to consciousness, it is best regarded as a modality of
existence (adapting this term from Kant, both because it fits Berkeley and by antici
pation of what follows).5
4 This means that if all traces of past sensation were suddenly wiped from our minds, we
would immediately become incapable of conceiving the real existence of sensible things, at least
in anything remotely approaching the way we presently do, where seeing is believing and merely
imagining doesn't make it so.
5 The difference between modalities of thought-in particular, memory and imagination
must then be understood in conformity with the modality of sensation, indeed as parasitic on
it. In particular, something present to my consciousness in thought has the status of a memory
and not a fantasy if the idea of it is accompanied by the idea that it was once present to my con
sciousness in sensation. Imagination, too, is parasitic on the modality of sensation insofar as, on
Berkeley's conception of it, "nothing enters the imagination which from the nature of the thing
cannot be perceived by sense, since indeed the imagination is nothing else than the faculty which
represents sensible things either actually existing or at least possible" (DM § 5). Insofar as any
thing produced in the imagination is itself copied from sensation or consists of simpler elements,
all of which have been copied from sensations, the status of everything we imagine as a possible
sensible object is again derivative from the modality of actuality specific to sensation. Thus, the
relevance of any thought to sensible reality, whether memory, fantasy, or hypothesis, is parasitic
on the modality of sensation: only by adding in thought the idea of existence acquired from
sensation can we conceive of things present to us in thought as having been existent, or as possibly
being in existence, or as actually being in existence because of their relation to what is present to
us now in sensation (as the smoke we see implies the existence of a fire not present to our senses if
fire is taken to be the cause of smoke). For a detailed consideration of these points, see UU ch 11.
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 205
termed it) all things that resemble it in respect of (what we call) its pitch or, in
another indifferently signifying denotation of the same idea, all things that resem
ble it in respect of (what we call) its timbre.6
8. Berkeley explained all ideational generality in terms of indifferent denotations
such as these, founded on the various kinds of resemblances that emerge when we
compare ideas that differ by the criterion of the separability principle: resemblances
not only of sensible quality but also in the manner in which the ideas are received
(red and blue resemble not in quality but in being both sensed by the eyes), the causes
to which they are referred (man-made things), their effects (emetics), their constant
co-occurrence (the spatial meaning that intrinsically non-spatial visual data acquire
through customary association because of their constant co-occurrence with genu
inely spatial data of touch), and so on (V § 128 and VV § 39).
I. The two kinds of distinction, ideational and semantic, each perfectly legiti
mate in its own sphere, must never be confounded: (i) differences that do conform to
the separability principle and so constitute a real difference of ideas and (ii) differ
ences that do not and so constitute distinctions only in the signifying uses to which
one and the same idea may be put.
2. Only the first sort of distinction did Berkeley regard as ontologically signifi
cant. The use of an idea to form different general ideas-to denote the members
of different resemblance classes indifferently-has in and of itself only semantic
worth. Not until it is subjected to the separability principle test can we determine
whether any such semantic difference cashes out as a real difference relating to the
only objects ever present to us in experience: ideas of sensation.
3. Berkeley's rationale for his ontological criterion is simple enough: no one
can mistake acquaintance with the significative uses of ideas founded on custom
ary resemblance relations for an expanded acquaintance with reality; and although
there may well be entities whose nature prevents them from ever being present to
our consciousness, since such things obviously can never be anything to us, reality
for us consists, and can consist, only of those beings that are capable of present
ing themselves to consciousness in sensation, memory, or imagination. It is these
objects that Berkeley held to be subject to the separability principle and to which
he applied the term "idea." Accordingly, he ascribed objective validity, in the full
ontological sense of concern to philosophers, only to those distinctions that satisfy
6 Another example: the general terms "sphere" and "white" have their origin not in anything
we could in principle detect by scrutinizing a single visible object (sensation, idea) in isolation but
rather in the multiple significations (indifferent denotations) of which one and the same visible
object admits; a white billiard ball is found to resemble non-white billiard balls, bowling balls,
and meatballs in the respect we call "shape," and to resemble cotton, snow, clouds, and the sclera
of human eyes in the respect we call "white." Consequently, the idea of that ball can be used to
indifferently denote all things that resemble it in respect either of its shape or its color.
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 207
the separability principle criterion and so cash out as differences among ideas-the
only things that can legitimately be deemed actual or possible existents (according
to whether they are present in sensation or imagination). Since semantic differ
ences, insofar as they are more than conventional stipulations (regarding which the
ontological question cannot even arise), tend to reflect distinctions not among ideas
but only their significative uses, Berkeley refused to accord them ontological worth
unless and until they could be shown to satisfy the separability principle criterion.
Otherwise, however useful or even indispensable they may be to human thought
1. For reasons that will become apparent when we turn to Kant in the next sec
tion, two applications of the separability principle to "existence" need to be distin
guished in Berkeley's esse is percipi thesis:
(i) the existence of sensible things is inseparable from the existence of sensations,
i.e., the presence of sensible things in sensation (rather than thought); and
208 Kant's Anatomy of the Intelligent Mind
(ii) the existence of sensations is inseparable from the existence of the mind's
perceiving of them, i.e., the presence of sensations in consciousness (the
sensibility of a mind).
talk of the existence of sensible things without talking of their presence in sensation
is one thing; to form the idea of the one without also forming the idea of the other is
quite another. For, according to Berkeley, in idea, they are one and the same so that
"I might as easily divide a thing from itself " as distinguish two ideas where there is
in truth only one idea with distinct, resemblance-relation-based significative uses.
3. Respecting (ii), we do not even talk of the existence of sensation without
implying that the sensation is perceived; it is part of the grammar of the indifferent
denotation "sensation." But even if we had a use for a semantic distinction between
them, we could still not separate them in idea. The existence of sensation and the
perceiving of it are one and indistinguishable, no more possible to separate into
distinct ideas than to divide a thing from itself.
4. It is in respect to the second component of esse is percipi idealism that an
opening can be found that leads to a refutation of the first, that is, proof that
sensation does indeed provide us with a means of conceiving the existence of
mind-independent things in themselves, and so proof of the intelligibility of the
realist thesis that the esse of sensible things is not their percipi.
and treated its products as the passively received inputs of imagination properly
so called. In particular, imagination, as traditionally conceived, starts with the
reproduction of the data of perception (appearances: Al19-20, A493/B521) as
images in thought (copies), which thereafter remain in readiness to be called up and
set alongside any perception that prompts the imagination to do so (in accordance
with subjective principles of preferential reproduction-Al21-2-such as Hume's
fundamental associative relations of resemblance, contiguity in space and time, and
cause and effect: ch 13 and UU ch 18). Kant broke with this tradition by insisting
that in order to be able apprehend perceptions together in one and the same con
sciousness at all, even if only as a jumble (scattered and single in themselves), the
mind must be equipped to represent them as a manifold, contained in one and the
same intuitive representation, which, in Kant's view, is only possible if the recep
tivity of the senses is supplemented by spontaneity (A97, A99-100; ch 3-E). In
particular, apprehension must conform to pure intuitions of space and time7 that
"are first given as intuitions" when "the understanding determines the sensibil
ity" by means of "a synthesis that does not belong to the senses" (Bl60n; ch 5-C).
Without pure intuitions to enable us to (synthetically) unite all affections in one and
the same consciousness, sensibility properly so called cannot exist at all (chs 3-4).
Sensation, for Kant, is thus only a necessary, not a sufficient, condition for the kind
of sense-perceptual consciousness that Berkeley had in mind when he claimed the
esse of sensation to be percipi (a consciousness in which the manifold of sensations
can be apprehended as a single manifold, preliminary to their combination into
sensible objects such as "the table I write on, ...houses, mountains, [and] rivers").
Kant's account of the possibility of perceptual sensibility as predicated on
spontaneity no less than receptivity implies the falsity of the second component of
Berkeley's esse is percipi idealism: the thesis that the existence of sensations and the
existence of the mind's perceiving of them are one and indistinguishable. For what
could be easier than to separate one thing from another if the existence of the first
is not subject to a condition requisite to the existence of the second? If perception,
insofar as Berkeley deemed it essentially to involve apprehension of sensations as
a manifold in one consciousness (unity of sensibility), depends on spontaneity but
the sensations themselves do not, then, pace Berkeley, it is by no means impossible
to conceive the latter to exist in the absence of the former: one need only suppose
that something intervenes to prevent imagination and prediscursive understanding
from contributing their synthetic unity to the manifold of sense. So even if one
concedes Berkeley's thesis that "the things we see and feel ...are ...but so many
sensations, notions, ideas or impressions of the sense," on Kant's account of the
possibility of sense perception (synthesis of apprehension in intuition), the esse of
these things is precisely not percipi. And is this not already in itself enough to refute
Berkeley's idealism?
7 "[P]ure intuition underlies all perception (in respect to the status of perceptions as represen
tations, the form of inner intuition, time, is their basis)" (Al 15-16).
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 211
Still, one may balk at the notion that Kant's account of the synthesis of apprehen
sion responsible for perception suffices to refute Berkeley's idealism. This idealism is
generally represented as distinguishing existence into two kinds, mind-independent
and mind-dependent, such that whatever is mind-independent is real, and what
ever is mind-dependent ideal. Since by this criterion sensations are just as ideal as
perceptions, no realist gain can be made over Berkeley's idealism by distinguishing
them as Kant does, and it then seems otiose to treat Kant's novel psychology of
perception as yielding a realist esse is not percipi refutation of Berkeley's idealism.
The first problem with this objection is that it misrepresents the nature of
Berkeley's idealism. Berkeley's idealism is neither epistemological nor metaphysical
but psychologistic. He questioned the materialists' very notion of mind-independent
existence as an ontological category-its consistency and meaningfulness. If sensa
tion is ineluctably part of the answer we must give to the question "what is meant
by the term exist when applied to sensible things" (PHK Pt. I § 3), and if sen
sation is indistinguishable from perception (other than semantically), then sensa
tion cannot possibly furnish us with any notion of mind-independent existence.
Since the only other source from which one can derive concepts of real existence
is the internal awareness of the activity of one's own mind (the reflexive notion of
existence specific to cogito ergo sum), and since this patently cannot be conceived
separately from mind (again, only semantic distinctness is possible), it follows that
neither external nor internal sense are capable of providing the understanding
with the wherewithal to conceive mind-independent existence in any manner at all.
Consequently, a realist refutation of Berkeley's idealism consistent with the prin
ciples of anti-abstractionist sensibilism must minimally be able to demonstrate that
sensation is the source of a genuine notion of mind-independent existence-or at
least emerges as such once we cease to confound the existence of sensation with the
existence of its perception!
Kant's attribution of perception to spontaneity as well as receptivity positioned
him to take the critique of Berkeley's idealism a step further than had previously
been possible, so as to distinguish the core conceptual question it poses-is sensa
tion capable of furnishing us with a representation of existence sufficiently robust
to enable us to conceive of things as existing mind-independently in themselves?
from a quite different subsidiary question that otherwise gets confounded with it
can the mind-independent things in themselves sensation enables us to conceive
be further conceived as corporeal in character? Since, according to Kant, every
thing essential to the constitution of bodies is intrinsically bound up with pure
self-consciousness (apperception), his answer to the second question is fully as ide
alist as Berkeley's: "if the thinking subject were removed, the entire corporeal world
must cease to exist, as it is nothing but an appearance in our subject's sensibility
and a mode of its representations" (A383). Bodies, transcendentally considered,
212 Kant's Anatomy of the Intelligent Mind
are therefore incapable either of existing as things in themselves (A357 and A370-l)
or of resembling things in themselves (PFM 289-90) (ch 7-C, KMM introduction
and Part I).
It is instead to Kant's response to the core question that one must look to
understand how his idealism differs from Berkeley's. Kant's idealism is limited to
the forms of sensible things-the formal constitution they derive from space and
time as conditions of the possibility of perception and from the categories as con
ditions of the possibility of experience-but, unlike Berkeley's, does not extend to
their matter, which, in this context, has nothing to do with corporeal being (physi
cal reality) but instead specifically concerns that in sensible objects (appearances)
which corresponds to sensation (ch 3-B). Because sensations (sense affections), by
contrast with what appears to us empirically in perception and experience (A493/
B521), are not predicated on pure intuitions of space and time (which depend on
imagination and apperception: chs 3-5), they must be regarded as both prior to
and independent of the formal intuitive conditions of perception and, a fortiori,
the formal conceptual conditions of experience as well. Since it was precisely the
presupposition of these formal conditions that led Kant to regard bodies, (empiri
cal) selves, and, in general, everything in any way dependent on space or time as
transcendentally ideal (ch 7-A), it follows that sensations are not that element of
our representation on account of which he denominated corporeal and thinking
beings ideal ("mere sensations such as colors, sounds, and heat . . . have, strictly
speaking, no ideality, although . . . they belong to the subjective constitution of our
species of sensibility," B44; also A28-9, PFM 290-2, CJ 189, AA 17 § 5058, and
AA 20 268-9).
To understand what it means to be able to consider sensations apart from per
ception insofar as the latter is conditioned by space and time (B 160), one must recall
what sensibility would be like without the unity that space and time make possible.
In Chapters 3 and 4, we saw that the unity of the manifold effected by pure space
and time is as essential to sensibility as the presence of a manifold of sensations (the
synopsis of a manifold a priori through sense), since, without pure space and time, it
would be impossible to represent data of the senses as a manifold, that is, contained
in a single representation and united in one and the same consciousness (unity of
sensibility). It is true that sensibility alone, without the categories, suffices only for
the synthesis of a spatial and temporal manifold devoid of all differentiation and
determination, so that appearances are "met with in the mind scattered and single
in themselves" (Al20). Yet it is unity enough for the higher functions of the mind,
reproduction and recognition (ch 13), to set about ordering and relating this mani
fold, eventuating in an objective unity of nature in accordance with laws founded
on the categories (Parts IV and V). But, most important for present purposes, it is
precisely the unity implicit in and so presupposed by Berkeley's conception of the
mind as an enduring substance through which manifold ideas of sensation continu
ally succeed, together with its successive acts of thought directed towards that man
ifold (its volitions, passions, desires, and other reflexive perceptions: UU ch 13-C).
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 213
It is also the unity "of our successive perceptions in our thought or consciousness"
(THN 636/401) from which Hume's quandary concerning personal identity springs
(chs 4-D, 7-A, and UU ch 3-B). In other words, the unity of sensibility presupposed
in the theories of Berkeley and Hume, and explained as the result of the synthetic
action of imagination and apperception (prediscursive understanding) in Kant's
theory of pure sensibility, is part of the core notion of mind (enduring mental sub
stance) implicit in the percipi part of Berkeley's esse is percipi idealism and which
Hume could neither explain nor explain away by associationist principles alone.
Thus, to separate sensations from perceptions, as Kant's theory alone permits,
is, in effect, to separate them from conscious mind itself, as Kant's predecessors
conceived it.
This is not true, however, on Kant's own conception of the mind ( Gemiit). Since
on this conception time derives from mind, the mind cannot exist in time and so,
unlike that of his sensibilist predecessors, Kantian mind cannot be conceived in
terms of the unity of successive perceptions in consciousness. In particular, because
succession arises with the imagination's synthesis of apprehension in intuition,
one can abstract from succession without abstracting altogether from mind, since
this still leaves its purely receptive capacity for the synopsis of a manifold a priori
through sense (ch 3-A). To be sure, a mind lacking apprehension would be a mind
without the unity of consciousness that alone can warrant attributing a sensibil
ity to it. Its mental life would consist of no more than a manifold of qualitatively
differentiated sensations that could not be represented as a manifold, contained
in a single representation, and united in the consciousness of that representation.
Sensations would simply be present rather than absent, and present in this quality
rather than that, but no more present in succession (before or after, or now or then)
than in juxtaposition (here or there). There is thus in Kant an ontologically real
difference, rather than merely a semantic one, between the existence of sensations
and their being perceived and so a real basis, consistent with Berkeley's separability
principle, on which to reject the second component of his esse is percipi idealism.
With the denial of the second component of Berkeley's esse is percipi idealism and
the emergence of the non-temporal subjectless mind ( Gemiit), the way is opened to
the denial of the first component-the distinction of the existence of sensible things
from the existence of sensations-and the emergence of subjectively unconditioned
things in themselves. To be sure, comprehending how Kant got from the exclusion
of sensations from the scope of transcendental ideal conditions of the unity of sen
sibility to sensations being the source of a notion of existence adequate to conceive
and affirm transcendentally real existents is no easy task , and, before undertaking
it, it is vital to assure ourselves that he in fact took this step.
214 Kant's Anatomy of the Intelligent Mind
The best place to begin is with Kant's own characterizations of the difference
between his idealism and Berkeley's in terms of whether idealism extends beyond
the form to the matter of sensible objects:
The opinion of Eberhard and Garve that Berkeley's idealism is identical with
critical idealism, which I could better term the principle of the ideality of
space and time, does not deserve the slightest attention: for I speak of ideality
in regard to the form of representation, while they make it out to be ideality
in respect of the matter, i.e. the ideality of the object and its existence. (letter
to Beck, December 4, 1792)
[T]his so-called idealism of mine concerns not the existence of things
(Existenz der Sachen) (the doubting of which however actually constitutes
idealism in the received signification), for doubting it never entered my head;
[it concerns] merely the sensible representation of things to which space and
time especially belong and shows that these, and so too all appearances in
general, are neither things nor determinations pertaining to things in them
selves but mere modes of representation. (PFM 293; also A358, A367-94,
B427-8, CPrR 13n, AA 18 § 6316, AA 23 57-8)
Accordingly, Kant's idealism encompasses all and only the (imagination- and
understanding-dependent)forma/ features by virtue of which sensible things count
as physical objects: their intuitable spatial and temporal properties and relations
together with their category-based mathematical and dynamical character as indi
vidual substances existing in the law-governed system of material nature, The only
feature of sensible things that eludes this idealism is their matter. For although their
matter is itself a product of imagination (ch 3-E), its sole and entire raison d'etre
is to exhibit sensations that are purely the product of receptivity; and sensations,
in their turn (as affections), correspond to things in themselves (i.e,, the matter
of appearances corresponds at one remove to "the object of the senses in itself").
When, in addition, we prescind from the qualitative side of sensations, dependent
as it is on the special constitution of our senses ("related merely to the subject
(according to its quality)"), and consider only the bare fact of their presence in us,
we are, according to Kant, already in the non-spatial, non-temporal, noncatego
rial realm of subjectively unconditioned transcendental reality ("the transcendental
matter of all objects as things in themselves (facticity, reality)"). Thus, once the
existence of sensation is detached from the existence of the unified consciousness
in which it is present as a manifold of perceptions (the denial of the second com
ponent of esse is percipi idealism), it emerges as the source of a genuinely tran
scendental notion of existence with which we can both conceive and affirm the
existence of a transcendentally real something in genera? wholly distinct from our
representations and their a priori forms (the denial of the first component). And,
on this psychologistic basis, Kant could legitimately claim to have defined a species
of idealism with a mind-independently real component that, albeit noncorporeal,
is still sufficiently robust to distinguish it from Berkeley's psychological idealism as
esse is NOT percipi realism.
From the first appearance of the Critique of Pure Reason, philosophers have puz
zled over Kant's affirmation of transcendentally real things in themselves. Things
in themselves can never be given in experience (A30/B45, A393, A545-6/B573--4),
yet according to the Critique, our concepts have no application to any objects
other than those that can be given in experience, and this restriction holds even
with respect to concepts that are a priori and purely intellectual in content such as
cause and effect and substance and accident (Bl48-9, Al46-7/Bl85-7, A235-60/
B294-315). So how could Kant possibly infer the existence of things in them
selves from the mere fact of the presence of sensations in us without transgress
ing the bounds of legitimate (cognitively meaningful) inference fixed by his own
philosophy? As a causal inference from the existence of affections of the senses
to thing-in-itself sense-affectors, it would violate the restriction of the scope of
objective causality to the sphere of appearances. As an inference to a substantial
substrate, it would violate the same restriction.10 And as an inference from the
existence of appearances to the existence of something that appears, it looks like
mere verbal sleight of hand. But if Kant did not infer the existence of things in
themselves in any of these ways, on what basis did he? And is it a basis that can
pass critical philosophical muster?
It must be conceded straightaway that there is no shortage of textual evidence
one could adduce to show that Kant did in fact infer the existence of things in
themselves in each of the three ways just enumerated. Not once but on several
occasions Kant characterized them as the cause of sensations (A288/B344, A387,
A391, A393, A494/B522, A496/B524, A546/B574). They are also often referred
to as the substrate of appearances (A251, A350, A359). And whenever sensible
objects in space and time are designated "appearances," Kant seems to pull tran
scendental realities out of his hat as the presupposed non-spatial, non-temporal,
noncategorial "things in themselves that appear" (Bxxvi-vii, A249, A251-2,
A538/B566).
Perplexed by Kant's apparently blithe disregard for boundaries he elsewhere
insisted on the importance of never transgressing, many of his defenders have
tended to downplay things in themselves, treating them either as a vestige from his
pre-critical period that somehow slipped into the Critique or as ontologically innoc
uous posits that careless phraseology makes seem like something more, and a few
have even gone so far as to deny that it was ever Kant's intention to affirm things
in themselves at all.11 Yet texts such as the following make it difficult to believe that
Kant felt the slightest qualm about affirming their existence or conceived of them
with anything less than full ontological robustness:
10
As Hume remarked, since "all our perceptions are different from each other, and from
every thing else in the universe, they are also distinct and separable, and may be considered as
separately existent, and may exist separately, and have no need of any thing else to support their
existence. They are, therefore, substances, as far as this definition explains a substance" (THN
233/153). Kant endorsed the elements of Hume's reasoning together with the separability prin
ciple on which it is founded (e.g., CPrR 51-3), and from this concluded that the principle that
everything that exists must either be a substance or inhere in one is synthetic a priori and valid
only within the field of possible experience.
u The first group is too numerous to list. Probably the best-known exponent of the view that
things in themselves, properly understood, are ontologically innocuous is Gerald Prauss in Kant
und das Problem der Dinge an sich (Bonn: Bouvier Verlag Herbert Grundmann, 1974). Nicholas
Rescher is one of the better-known commentators to aver that it was never Kant's intention to
go so far as to affirm things in themselves at all: see The Primacy of Practice; Essays Towards a
Pragmatically Kantian Theory of Empirical Knowledge (Oxford: Blackwell, 1973).
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 217
[T]o sensible entities intelligible entities do, of course, correspond, and there
may also be intelligible entities to which our sensibility has no relation at alL
(B308-9)
In fact, if, as is right (billig), we regard the objects of the senses as mere
appearances, we thereby concede at the same time that a thing in itself under
lies them, though we do not know (kennen) this thing as it is constituted but
only its appearances, i.e. the way our senses are affected by this unknown
something. Hence, the understanding, by granting appearances, thereby
also concedes the existence of things in themselves and thus far we can say
that the representation of such beings underlying appearances, and so mere
intelligible entities, are not only admissible but unavoidable. (PFM 314-15;
also 318)
[Eberhard writes:] "choose what we may, we come to things in themselves."
Now, this is just what the constant assertion of the Critique is ...It says: the
objects as things in themselves give the material (Stoff) to empirical intuitions
(they contain the ground of the determination of the faculty of represen
tation in conformity with its sensibility), but are not their materiaL (On a
Discovery 215)
Such texts make it plain not only that Kant inferred the existence of things in
themselves as ontological conditions of sense affection, but also that, in his view,
a critical (transcendental) idealist has no other choice. Consequently, the only real
issue for his interpreter is whether and how he could provide a critical-philosophical
grounding for so doing since otherwise his claim that transcendental philosophy
contains a nullification of idealisms such as Berkeley's would have to be deemed a
failure even on his own terms.
The issue can be formulated more precisely. According to Kant's theory of
understanding, a judgment relating determinations or existents that are conceiv
able independently is synthetic: the predicate does not merely explicate what we
already are obliged to think merely in conceiving the subject-concept but also
introduces something new, amplifies our thought beyond the subject, and so has
the potential to increase our knowledge if the judgment is true (ch 2-A, UU ch 2).
Since sensations, as mere modifications of the mind dependent on its subjective
constitution, are without question distinct in existence from mind-independent
things in themselves, it seems that any judgment connecting the existence of sen
sations to things in themselves must be synthetic. Now, according to Kant, the
objective validity of any synthetic judgment, whether a priori or a posteriori,
depends on its conformity to the a priori formal conditions of intuition (pure
space and time) and experience (the categories), apart from which it can carry nei
ther objective sense nor signification but instead collapses into empty words that
say nothing about any possible existent. Since these conditions are in themselves
nothing more than modes of conscious representation, and since "it is patently
contradictory to say that a mere mode of representation also exists outside our
218 Kant's Anatomy of the Intelligent Mind
representation" (PFM 341-2), it follows that the requirements for legitimate syn
thetic judgments are impossible to satisfy in any sy nthetic-ampliative judgment
that proceeds from the matter of transcendentally ideal appearances and phe
nomena (which are dependent on the formal conditions of intuition and expe
rience) to transcendentally real things in themselves (to which space, time, and
the categories have no application) as the ground or condition of the affection
of the senses (i.e., of the presence of sensation in the mind). Consequently, a
sy nthetic judgment relating sensations to things in themselves is clearly illegiti
mate by the criteria specified in Kant's own account of the possibility of such
judgments.
A successful Kantian response to this objection must show that, while a synthetic
inference from sensations to things in themselves is impermissible, none is required.
To do this, it must furthermore be shown that we can get to mind-independent
things in themselves analytically and so without ever having to move beyond sen
sations, notwithstanding their distinction from things in themselves in respect of
mind-dependence. And while Kant may never have assembled the elements of his
reasoning in the form of a premise-by-premise argument, it is not difficult to deter
mine how such an argument would go:
1. The fact of affection in sensation
For Kant, the only feature that entitles sensation to be classed as objective,
and so to count as the matter of cognition, is that it "expresses the material
(real) whereby something existent is given" (CJ 189; ch 3-B). More particularly,
though the green of a meadow counts as "objective sensation, . . . the perception
of an object of sense" (CJ 206), it is not, strictly speaking, the sensible qual
ity green that warrants Kant's claim that sensation expresses a given reality but
simply this: the fact of this sensation's presence in one, one's affection in
sensation.
2. Ambiguity in the notion of fact of affection in sensation
with another genuinely distinct from it, as is the case with causal relations.12 It is the
fact of affection itself-its "facticity, reality"-that coincides with "the transcen
dental matter of all objects as things in themselves" (Al43/B182), not any relation
it may have to something else, which underpins the Kantian refutation of Berkeley's
esse is percipi idealism. For Kant's question, like Berkeley's, was neither metaphysi
cal nor epistemological but psychologistic: what concept of existence is afforded
by the presence of sensation in us?, that is, what kind of entity does it make intel
ligible to us? And he answered it by the same method Berkeley did: psychologistic
explication.
We have already examined one of the reasons Kant's analysis leads to a conclu
sion diametrically opposed to Berkeley's idealism: insofar as perceptions depend
on imagination but sensations do not, the esse of the latter is precisely NOT percipi
(Sections B and C). To this distinction can now be added a second, likewise based
on a difference of subjective conditions: if the fact of affection is prior to and inde
pendent of the subjective conditions that determine whether and in which sensible
qualities a manifold of sensations will coincide with it in a priori synopsis, then that
esse is subjectively unconditioned. As such, the fact of affection emerges from Kant's
analysis of sensation as the source of a concept of real existence (esse) sufficiently
robust to render intelligible the mind-independent transcendental reality of things
in themselves (beings that exist completely independently of the representational
faculties of the human mind). Nor should it go unremarked that Kant's analysis of
12
Quite apart from the lack of objective intelligibility, on Kantian principles, of any appli
cation of the concept of cause and effect to things in themselves, such an inference would be
vulnerable to a skeptical objection similar to the one Descartes raised against the supposition
that bodies are the causes of sensation: how, if sensations are all that is ever present to us, can
we be certain that corporeal beings cause sensations rather than something else-God, or some
other immaterial entity, or even our own mind through some faculty we are unaware of possess
ing? Since Kant's transcendental idealism relegates bodies to the status of mere representations
(so too the mind insofar as it is determined conformably to pure time or the categories), the
relevant skeptical doubt here would be how one can determine that a mind-independently exist
ing thing in itself is the cause of one's sensations rather than some (atemporal, transcendentally
real) mind, one's own or any other. And here we can be quite certain that Kant would have
maintained that if such a doubt could be intelligibly raised, it could never be allayed since, on
his principles, there is no knowing whether the transcendental object ground of bodies (outer
appearances) and the transcendental subject underlying inner experience are the same or differ
ent (see Section G and ch 4-A).
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 221
affection yields this result and thereby refutes Berkeley, not despite but precisely by
following Berkeley in equating the esse of sensible things with the fact of affection
in sensation (Section A), but now conceivable as subjectively unconditioned, Kant
was thus in a position to claim to have turned Berkeley's own method of reasoning
against him with as much warrant as he did in the case of Descartes in his refutation
of Cartesian idealism (B276).
6. The refutation is not abstractionist
Although we can only conjecture as to how Berkeley might have responded to
this refutation of his idealism, Kant's distinction between the form and matter
of appearances may well have been a target of criticism. For even while insist
ing that they are distinct in a certain sense, Kant admitted that in subjective con
sciousness, form and matter of appearances are one and indistinguishable (A429/
B457n and A452/B480n). So Berkeley could fairly ask how they could be distinct,
in any save a semantic sense, without resorting to abstractive considerations of
precisely the kind proscribed by the separability principle to which both philoso
phers subscribed.
Kant did indeed employ abstraction to distinguish the contribution of pure
intuition to appearances from those of sensation and the understanding (A20-1/
B35-6, A27/B43, and AA 8 240; ch 3). But far from being the kind that conflicts
with Berkeley's separability principle, Kant distinguished sensation, as the source
of the matter of appearances, from their form in the same manner Berkeley's
theory of vision distinguishes tactual space from visual space: by their different
subjective conditions. In Berkeley's case, tactual space is a pre-imaginative given
of sense (perception, in Locke's sense: Section B), whereas visual space is the ficti
tious product of customary association with tactual data in imagination (UU ch
14). Similarly, for Kant, sensation involves only the receptivity of sense (synopsis)
while the spatial and temporal forms of appearances depend also on spontane
ity (imagination and apperception as conditions of the unity of sensibility: chs
3-5). And it was the same principle of separation at work when he extracted a
concept of transcendentally real existence from the fact of affection: whereas both
sensational qualities and the formal manifoldness of appearances are subjectively
conditioned, the facticity of affection in sensation is not.13 Thus, the distinctions
required for a Kantian refutation of the two inseparability theses that together
make up Berkeley's esse is percipi idealism (Section A) turn out to accord perfectly
with Berkeley's anti-abstractionism.
13 Berkeley himself held that sensation, regardless of quality, is the source of the idea of the
modality of existence whereby we distinguish the real existence of sensible things from their
merely fictive presence in thought. And while he did not make the ground of the distinction as
clear as he might have, it seems evident from his treatment of sensation as something that requires
a cause outside the mind that he operated with a distinction between the subjectively grounded
quality of sensation and the objectively grounded fact of its existence in us (enabling him to argue
that God is the cause of the existence of sensations in us).
222 Kant's Anatomy of the Intelligent Mind
negation, neither plurality nor unity, and so forth affirms a something indistin
guishable from a nothing, Yet abstractionist indirect realism of the kind advocated
by Locke (UU ch 7), where at least the possibility remains open that things in them
selves resemble the corporeal appearances of outer sense and the thinking, feeling
appearances of inner sense, was not an option for sensibilists like Kant who sub
scribed to Berkeley's anti-abstractionist separability principle. In that case, the only
alternative to things in themselves is the one that adds the ideality of the existence
of sensible objects to the ideality of their spatial, temporal, and categorial forms,
i.e., esse is percipi idealism.
Of the various grounds for objecting to the foregoing refutation of Berkeley's ideal
ism, perhaps the most natural and compelling relates to step 4, the claim that exis
tence in the guise of the fact of affection in sensation remains to be conceived and
affirmed even when the subjective conditions necessary for sensation fail to be met.
The target is the supposition that a change in the character of a sensation, whether
of intensity or quality, is continuous with a change from the presence of sensation
to its absence: are these contingencies not different in kind, or at least sufficiently
distinct, to block the conclusion that the fact of affection and the existence it con
notes remain even in the absence of actual sensation?
The response emerges as naturally as the objection. Since the conclusion does
not posit a difference between the quality of sensation and the fact of affection per
se, but between the subjectively conditioned character of the former in contrast
to the subjectively unconditioned character of the latter, its validity turns on how
"subject" is construed. Kant made quite clear that no concept can be applied to
the subject of representation that has its ground in the constitution of the subject's
own faculty of representation, thus preventing it from being conceived in terms of
the categories or space and time (A346/B404, A402, B422). And, as noted earlier,
this means that the subject cannot be thought of as in any sense the cause of the
existence marked by the fact of affection in sensation.14
Only two possibilities remain: either sensations exist completely independently
of the representing subject or they depend on it only insofar as its constitution
determines their quality, intensity, and variegation-"depend," that is, in the
14 Because Descartes, like other pre-Humeans, attributed unrestricted scope to the idea of
cause and the principle that every beginning of existence must have a cause, he could not (without
invoking a non-deceiving God) preclude the possibility that the subject is the cause of its own
sensations. In the context of Kant's philosophy, where the principle of cause and effect is limited
in scope to objects of possible experience, the subject responsible for constituting experience and
its objects cannot be conceived as efficacious (a cause of anything's existence), thereby rendering
the possibility that the subject is the cause of its sensations not just false but unintelligible (for the
same reason, nothing else-God included�an be conceived as the transcendentally real cause
of the existence of sensations).
224 Kant's Anatomy of the Intelligent Mind
same sense that appearances depend on the representing subject as the source
of pure intuitions or that phenomena (objects of experience) depend on it as
the source of pure concepts of the understanding. In the first case, the objection
to the Kantian refutation of Berkeley's idealism would fall straightaway because
sensations existing independently of the representing subject would ipso facto be
the source of a concept of a subjectively unconditioned existence that could then
be used to conceive the existence of other transcendentally real things in them
selves.15 In the second case, there is indeed this difference: the constitution of the
15 This indeed is how Hume conceived of all perceptions. However, for reasons that will be
made clear shortly, this conception cannot be sustained even on his own principles.
Things In Themselves: A Kantian Refutation of Berkeley's Idealism 225
only visual and tactual impressions but also smells, itches, aches, fears, yearnings,
fantasy images, resolves to act, et aL-count as self-subsistent existents: "since
all our perceptions are different from each other, and from everything else in the
universe, they are also distinct and separable, and may be consider'd as sepa
rately existent, and may exist separately, and have no need of any thing else to
support their existence" (233/153; also 234/153-4, 244/160, 252/164-5, and
636/400).
However, as Hume subsequently came to recognize, this position is unten
able: "having thus loosen'd all our particular perceptions ...all my hopes vanish,
when I come to explain the principles, that unite our successive perceptions in our
thought or consciousness" (THN 635-6/400). Since necessary connections cannot
be conceived pre-associatively, he could have recourse neither to causal dependence
nor inherence to explain this unity. Yet Hume also could not explain it away since
the facile transitions of thought constitutive of the associative relations premised
everywhere in his philosophy presuppose that the perceptions from and to which
these transitions are made are present to one and the same temporally continuous,
pre-associative consciousness. Faced with this quandary, he had no option but to
"plead the privilege of a sceptic, and confess, that this difficulty is too hard for my
understanding" (636/400; UU ch 3-B).
Hume's quandary demonstrates that the subject of representations is less dis
pensable than he originally supposed. It was, however, left to Kant to discover a way
of conceiving the subject that does not depend on concepts of necessary connection
between the distinct at alL The first step is to both deepen and widen the problem
by extending it from the question of the unity of successive perceptions in the same
consciousness to one of the unity of any manifold of representations, whether in
the form of a succession, juxtaposition, or anything else. This, the problem of the
unity of sensibility, begins with the unity of sensations originating in the different
senses in one and the same consciousness (ch 3-D); and the solution Kant proposed
was that these data can be exhibited as the manifold of a single representation,
all contained in the consciousness of that representation, by means of the pure
intuition of space (ch 4-B). Succession only enters the picture when a second pure
intuition must be posited to explain how distinct apprehensions of manifolds of
juxtaposed appearances can be apprehended immediately together while retaining
their distinctness (ch 4-C and UU ch 3).16
It is pure space and time that furnish the basis for a way of conceiving the rep
resenting subject that Hume never envisaged. For, far from being givens like sensa
tions, these intuitions, together with the appearances exhibited by their means, exist
only in and through such quintessentially mental acts as synthesis in imagination
and consciousness of the unity of that synthesis in understanding (prediscursive
16
The Humean quandary will be revisited in connection with the objective succession of the
Second Analogy of Experience in Chapter 17, and a consideration of the relation of the pure
representing subject to the empirical subject features in Chapter 18.
226 Kant's Anatomy of the Intelligent Mind
given, given indeed only to thought in general, and so neither as appearance nor as
an existent (Sache) in itself (noumenon), but as something that in fact exists and in
the proposition I think is designated as such an existent" (B422-3n). It is therefore
"prior to any experience which is supposed to determine the object of perception in
respect of time by means of the categories, and the existence here is not yet the cate
gory, as the category of existence does not have relation to an indeterminately given
object but only to such an object as one does have a concept of and seeks to know
whether or not it is also posited outside this concept." Indeed, not only must the
17 Also: "[T]he representation I am which expresses the consciousness that can accompany all
thinking is what immediately includes the existence of the self, but still no cognition thereof, and
so no empirical cognition, i.e. experience; for there belongs to the latter, outside the thought of
something existent, intuition as well, and here inner intuition, in respect of which, i.e. of time, the
subject must be determined" (B276-7).
228 Kant's Anatomy of the Intelligent Mind
want of any "self-intuition that gives the determinative in me" (B157-8n), can only
enter into its representation if there is a synopsis of the manifold of self-affections
a priori, which, as simply another mode of representing, must also be regarded as
subjectively conditioned (e.g., "without any empirical representation to supply the
material for thought, the act (Aktus) I think would not take place," B423n; ch 3-A).
It is with the removal of a priori synopsis, the most elementary of all subjective con
ditions of representations and so the complete loss of the manifold of sense itself,
that affection ceases to be possible, and one is left only with the fact of the affection
that would occur if this condition were in place. This last element of representation
alone, arrived at purely analytically, is subjectively unconditioned and is none other
than that something= X, which Kant alternately described as thing (object, exis
tent) in itself, transcendental object (or subject), and negative noumenon.
What about the mind ( Gemiit) and its relation to the thing in itself? As I under
stand Kant, mind and representing subject are the same. Their existence is appre
hended through the indeterminate perception given to thought itself (B422-3n),
which, as devoid of representable content (i.e. an apprehension simplex), consists
simply and solely in the fact of conscious representation and so, at its core, is iden
tical with the thinking thing (res cogitans) of Descartes' cogito ergo sum, but shorn
of all intellectualist epistemological and metaphysical baggage (Conclusion-B). The
transcendental reality of this fact is either the subject in itself or the transcendental
object corresponding to affection in sensation. As such, however, it is no more intel
ligible to distinguish the mind from the thing in itself, or transcendental object, than
it is to equate them:
another relation, whose thoughts we cannot indeed intuit, but the signs of
which in appearance we can. (A358-9, emphasis added; also A383, A390-1)
Kant's refutation of esse is percipi idealism reveals just how robust the realism at
the heart of his idealism actually is. It is true that he regarded nature as a tran
scendentally ideal realm of appearances: space and time, the corporeal and think
ing beings that occupy and contain them, and the transcendental principles that
constitute their objectivity ensure their mathematizability and make possible all
of the causal and other laws of nature to which they are subject-all have, and
can have, meaning and validity only for appearances, not for things in themselves.
The worlds of noumena and phenomena may thus seem to be two mutually exclu
sive realms with nothing whatsoever to relate them. Yet as soon as one shifts one's
regard from the form to the matter of appearances, the aspect alters radically. For
the matter of outer and inner appearances, though distinct from sensations and
self-affections, exhibits them so that their reality carries over into appearances (ch
3-B), and thereby into the otherwise ideal realm of material nature (ch 18-D). This
means, first, that the corporeal beings generated through categorial synthesis spe
ciosa of outer appearances have the same reality the sensations through which these
appearances are apprehended do, and second, that the thinking being (empirical
self) generated through the application of that synthesis to inner appearances has
the same reality the self-affections through which these appearances are appre
hended do. Having now determined that the reality of sense affections analytically
includes "the transcendental matter of all objects as things in themselves (facticity
[Sachheit], reality)" (Al43/B182), we get the result that the existence exhibited in
corporeal and thinking beings is none other than (analytically includes) the exis
tence of things in themselves. So however ideal these phenomena may be in all that
concerns them as objects, they thus emerge, in the bare fact of their existence, as one
and indistinguishable from noumena.
This existential identity makes everything present in our minds a proxy, as it
were, for noumenal beings we cannot even so much as conceive, much less know.
This seems to be the basis of Kant's notion of a representation ( Vorstellung). For he
did not deem mentation as such representational. It becomes so only insofar as it is
related to something outside the mind for which it goes proxy:
The Kantian refutation of Schulze's skepticism is the same reasoning that overturns
Berkeley's: the esse is not percipi proof whereby the existence of subjectively uncon
ditioned things in themselves can be conceived and affirmed by analytic means
alone. All empirical representations as such relate, via the matter of appearance,
to a non-representational subjectively unconditioned existence for which they go
proxy in all perception and judgment. This relation to a transcendentally real object
is not, to be sure, sufficient for cognition since the object ("the transcendental mat
ter of all objects as things in themselves," Al43/Bl82) is not given as a manifold
in intuition, hence does not admit of being conceived, and without concepts, it is
impossible to form judgments about it (not even false judgments).19 Nevertheless,
this noncognitive relation of representations to an object in itself is folded into
the relation of properly cognitive representations to their objects (corporeal and
thinking beings), thereby extending to the latter the status of representations of
subjectively unconditioned transcendental reality as well (chs 16-C and 18-D). And,
with this, we at last arrive at Kant's fundamental ontological dichotomization of
all objects into representations (appearances, phenomena) and things in themselves
(transcendental objects, noumena).
18 Also: "impressions are still not representations, for the latter must be related to something
else which is an action" (AA 15§ 413 [early 1770s]) and "repr(J!sentatio . ..is that determination
of the soul which is related to other things" (AA 15§ 1676 [1750s]). Apart from this feature,
Kant's notion of a representation is otherwise similar in scope and purpose to Locke's notion of
an idea (UU ch 5-B) and Hume's notion of a perception (UU ch 16-A): it designates everything
that is or can ever be present to consciousness, be it in sensation, reflexion, intuition, concept,
judgment, etc., as a representation (A319-20/B376-7, L 64-5. AA 15§ 426 and§ 1705). Since
Locke treated every idea and Hume every perception as likewise an object, it should come as
no surprise that Kant did so as well: since "[a]ll representations, as representations, have their
object, and can themselves in turn be objects of other representations" (Al08-9), "everything,
even every representation, can be termed an object insofar as one is conscious of it" (Al89/B234),
while consciousness itself is simply "a representation that another representation is in me" (L
33; Hume seems to have conceived of consciousness similarly: "consciousness is nothing but
a reflected thought or perception," THN 635/400). Again, like his predecessors, Kant included
obscure perceptions within the scope of representation: "The notion of representation .. . com
prehends obscuras within it, representations we never know that we have" (AA 15 1677 [early
1750s]; ch 4-B). Accordingly, the only objects that are not representations are those that can never
(unlike sensations and self-affections) present themselves to consciousness and which are there
fore nothing to us anyway. In light of Kant's conception of the representing subject in terms of
constitutive modes of conscious representation (pure intuition, pure concepts, etc.), this means
that the only non-representations are unknowable (unintuitable, inconceivable) objects in them
selves: "appearances are not things in themselves but are themselves only representations which
in turn have their object, which object can no longer be intuited by us and may therefore be called
the non-empirical, i.e. transcendental object= X" (Al09).
19 Representations are prior to and independent of cognition: see, for example, A320/B376-7,
We now come to a concept that was not classified above in the universal
list of transcendental concepts and nevertheless must be counted among
them without thereby in the least altering or revealing any defect in that
table. This is the concept or, if one prefers, the judgment: I think. It is
readily evident that it is the vehicle of all concepts in general and so of
transcendental concepts as well, and is therefore always conceived along
with these latter, and so is just as transcendental as they are.
A341/B399
because it opens the way not just to the formation of empirical concepts (empirical
discursive understanding) but pure concepts as well (pure discursive understand
ing). And since this includes the explication of the traditional categories of meta
physics as pure concepts of the understanding in the metaphysical deduction of the
categories, I conclude the chapter by showing how Kant's psychologistic explica
tion of logical universality provided the indispensable foundation for responding to
Hume's skeptical challenge to the very possibility of such concepts.
The explication of logical universality is, however, only the beginning of Kant's
account of the possibility of discursive understanding. Once concepts are available
to it, there must be some means of combining them if they are to be put to repre
sentational use. Since it concerns the combination of representations specifically
with regard to their logical form as universals, this need cannot be met by aesthetic
modes of relation Uuxtaposition, succession, color, sound, etc.). What are required
instead are specially dedicated forms of logical relation; and since concepts have no
representative use except when formed into judgments (A68-9/B93-4), these logi
cal modes of relation must be forms of judgment or, as Kant termed them, logical
functions of judgment. Logical functions not only make it possible to synthesize
any concept with any other to form a judgment but also enable the understand
ing to synthesize any judgment they produce with any other to form either a more
complex judgment or a series of logically related judgments (inference). The result
is a new, purely logical synthetic unity of the thinking subject (unity of discursive
understanding) that extends the prediscursive, purely aesthetic synthetic unity of
the intuiting subject (unity of sensibility) yielded by pure space and time.
In Chapter 10, I will analyze each of the logical functions from the standpoint of
the need to be able to relate the discursive representations that result from attaching
the analytic unity of apperception to nondiscursive representations. The great virtue
of this perspective is that it enables one to disentangle the analysis of thought-the
logical relatability in the discursive synthetic unity of judgment of any representa
tion with any other insofar as they have been transformed into concepts by means
of the analytic unity of the I think-from the logical analysis of language (broadly
construed to include mathematical formulae), be it in the Fregean tradition or any
other. As a result, a far stronger case can be made than is otherwise possible for the
much maligned table of judgments that Kant deemed essential to the character of
his philosophy as a system and so a primary source of its evidentiary force.
Finally, in Chapter 11, I will show how Kant's table of categories emerges from
the logical functions. Insofar as the concepts made possible by the analytic unity
of apperception have no representational role except through their synthesis in
judgments in conformity to the logical functions, these functions confer a logical
determinability on concepts as subjects or predicates that are related affirmatively
or negatively (or infinitely) and universally or particularly (or singularly), as well
as a logical determinability on the resulting judgments as grounds or consequences
or as disjuncts of a logically divided judgment. They thus open the way to the
formation of pure concepts of the understanding, devoid of all sensible content,
Kant's Psychologistic Explication of the Possibility and Forms of Thought 235
that consist simply and solely in the determination of concepts as either always
subject and never predicate or always predicate in relation to such a subject, as
either always affirmative and never negative or always negative and never affirma
tive, etc. In addition, they yield pure concepts that consist simply and solely in the
determination of the relation of judgments so that one is always ground and never
consequent in respect to another, or always and only in disjunction with others,
never not, etc. For it was as concepts with these meanings that Kant believed he
could meet Hume's challenge to demonstrate the existence of concepts of the neces
Concepts in Mind
T hought, as Kant conceived it, involves concepts, that is, representation by means
of universals, and so is essentially discursive. It is cognitive if it incorporates intu
ition and does so in a manner that conforms to the conditions of possible experi
ence. But representation by means of universals remains possible even if it fails
to satisfy these conditions and derives its contents not from intuition but from
sensations, feelings of pleasure and pain, morality, religion, or, as in pure general
logic, it abstracts entirely from content. To account for the possibility of thought
is therefore, in the first instance, to account for logical universality, without regard
to content.
Interpreters tend to focus so single-mindedly on Kant's account of the possibil
ity of cognitive thought (experience) that they mostly do not even notice that he
provided an account of thought as such as well. Since the latter is as important and
original as anything in Kant's philosophy, its all but universal neglect goes a long
way toward explaining why his account of the former has so often been taken for an
exercise in anti-psychological intellectualism. In this chapter, I propose to correct
the record by focusing on Kant's explication of logical universality in terms of the
analytic unity of the I think. By supplying otherwise purely sensible representations
with the logical form requisite for propositional thought, it opened the way for him
to psychologize both judgment and the categories, and so free them of any essential
dependence either on intersubjective meaning (linguistic convention) or objectiv
ized meaning (Cartesian true and immutable ideas, Fregean sense and reference,
language of thought algorithms, etc.). No less important for Kant's purposes in the
Transcendental Analytic, the logical meaning of the I think is an indispensable ele
ment of the transcendental syntheses (intellectualis and speciosa) at the heart of his
account of cognitive thought. For it is not until the I think is added to the manifold
of pure formal and empirical material intuition, converting it into a manifold of
concepts, that its determination in conformity to the logical functions of judgment
and the categories, and so too to cognitive experience, first becomes possible. T hus,
in order to fully comprehend Kant's account of the possibility of cognitive thought,
one must first master his account of the possibility of thought as such.
238 Kant's Anatomy of the Intelligent Mind
If ideas be particular in their nature, and at the same time finite in their num
ber, 'tis only by custom they can become general in their representation, and
contain an infinite number of other ideas under them .... Nay so entire is the
custom, that the very same idea may be annext to several different words,
and may be employ'd in different reasonings, without any danger of mistake.
Thus the idea of an equilateral triangle of an inch perpendicular may serve
us in talking of a figure, of a rectilinear figure, of a regular figure, of a tri
angle, and of an equilateral triangle. All these terms, therefore, are in this case
attended with the same idea; but as they are wont to be apply'd in a greater or
lesser compass, they excite their particular habits, and thereby keep the mind
in a readiness to observe, that no conclusion be form'd contrary to any ideas,
which are usually compriz'd under them. (THN 24/21 and 21/20).
2 See UU 260n5. Many would also maintain that theists like Locke and Berkeley could not
entirely escape the Platonic archetype model of logical universality on the ground that the intel
lect of God knows things as they really are, including universals such as essences and laws, with
out in any way relying on affection in sensation, resemblance relations and habit, or anything else
specific to the sensory psychology of finite minds.
240 Kant's Anatomy of the Intelligent Mind
from individuating circumstances, and custom. Otherwise, they are just conven
tional contrivances, useful or even indispensable to human communication, but
with nothing to anchor them in what, for sensibilists, is the only reality that can ever
present itself to us: the sensible objects that appear to the outer and inner senses.
What is true of linguistic universality would seem to be true of propositional
structure as well. For present purposes, propositional structure can be understood
as the special kind of unity that results when logical and grammatical components
are combined in such a way as to have a truth value if asserted. If sensibilist ideas
(perceptions for Hume, representations for Kant) were isomorphic with propo
sitional structure, they would divide into action ideas (verbs), agent and patient
ideas (nouns), and ideas that combine the two to yield mental propositions of
varying degrees of structural complexity (connectives, prepositions, quantifiers,
etc.). Though Locke (UU 205-6, 217, 222) and Berkeley (UU 270-1) seem to have
affirmed some such isomorphism, there is little if any evidence of it in Hume, the
sensibilist closest to Kant. There is no suggestion in Hume's account of the com
bination and relation of perceptions of anything corresponding to subjects and
predicates. Relations generally (including composition) consist in (if natural) or
are founded on (if philosophical) the facility felt in transitions from impressions to
ideas (UU ch 17-B). That the resulting relations and ideas of complex individuals
have nothing whatever in common with subject-predicate form or anything else spe
cific to the logico-grammatical structure of verbal propositions seems clear on the
face of it but is further confirmed by Hume's contention that they "operate in the
same manner upon beasts as upon human creatures" (THN 327/212-13). Similarly,
on Hume's account, belief does not essentially involve the asserting or denying of a
proposition but is instead a feeling of "force and vivacity" immanent to the manner
in which a sensation or reflexion is apprehended or a thought conceived, and so is
also well within the capacity of a broad spectrum of language-less animal minds
(UU ch 17-C). Accordingly, in the mind of the judger, all judgment in matters of
fact and real existence reduces to nothing more than a transition of thought that,
to the extent it is facile, confers a feeling of vivacity on the perception to which
the transition is made proportionate to the intensity of the facility feeling and the
intensity of the vivacity of the perception from which the transition is made, while
(empirical) inferences are just sequences of such transitions that consequently
"resolve themselves into the first, and are nothing but particular ways of conceiving
our objects" (THN 97n/67n). Thus, if one is to speak of "mental propositions" in
connection with Hume at all, they must be understood as at best remote, purely aes
thetic (sensible) analogues of logico-grammatically structured verbal propositions.
A standard objection to theories of ideas of all kinds is that they equate the
mental accompaniments of discourse with its meaning, or at least reduce the
non-conventional content of the latter to the former, resulting in accounts of
meaning that are easily shown to be woefully inadequate to capture the seman
tics of natural language. This criticism does, in my view, retain considerable force
against an approach such as that of William James, for whom the meaning of,
Concepts in Mind 241
say, "but" is given by a "but" -feeling in the mind. But I think it is a mistake to
extend this conception of the relation between language and mentation, as so
often is done, to early modern sensibilists such as Hume. They were not in the
business of developing theories of meaning, whether of the private language
variety or any other kind, but rather, far more modestly, of devising an onto
logical check on what we say in relation to the only reality they believed is ever
given to us: the kaleidoscopic flux of fleetingly existent sensations, reflexions,
and thoughts, present immediately in perception. Even if all, or nearly all, of
At first sight, Kant's view of how language is anchored in consciousness may not
seem very different from those of earlier sensibilists, at least when considered from
the standpoint of pure general logic. Like them, Kant held that "we can under
stand nothing except what carries with it a correspondent in intuition to our words"
(A277/B333). How one advances from individual intuitions to concepts, with uni
versal form, he explained in similar fashion:
Comparison consists in discerning the features of each sensible object and noting
how they differ from those of other objects; reflection isolates those features in
which the objects compared resemble; and, when abstraction is made from the dif
ferences, the resemblances that remain are united as constituent marks (Merkmale)
in a concept that can be used to sort, classify, and rank the objects one experi
ences. By expanding the scope of one's comparison, more general concepts can be
produced in the same way: I compare trees with shrubs, flowering plants, lichen,
etc.; reflect on what they have in common; and abstract from the differences to
fashion the concept of a plant (which I can afterward modify in the light of new
information, e.g., an understanding of photosynthesis). Or, by altering the scope of
the comparison, I can fashion different concepts from the same objects: I compare
trees with rocky outcroppings, cliff faces, and ledges; reflect on what they have in
common; abstract from the differences; and thereby fashion the concept of avian
nesting sites. And, presumably, Kant would have maintained that our very first,
most primitive concepts, formed while still in our cradles, are produced by the same,
essentially psychological processes.
Reflection, comparison, and abstraction, however, do not concern the form of
concepts as universals but rather their matter, the objects thought or not thought
in them ("the matter of a concept is the object, its form universality," L 91). They
enable us to eliminate extraneous detail, fashion only those concepts that most
accurately reflect the patterns frequently and constantly encountered in experience,
and ease the burden of memory by retaining only the sortals that best serve our
cognitive and conative needs. Where the form of concepts is concerned, the most
striking feature of Kant's account in relation to his sensibilist empiricist predeces
sors is the absence of custom. In pre-Kantian sensibilist accounts, custom is the
psychological basis of universal scope ("If ideas be particular in their nature, and
at the same time finite in their number, 'tis only by custom they can become gen
eral in their representation, and contain an infinite number of other ideas under
them," THN 24/21). In its place, Kant set the analytic unity of apperception: the
representation of the identity of the I think in respect of all the manifold. Since this
conception of logical universality has no real precedent, either in sensibilist or intel
lectualist logic, it deserves extended scrutiny. And since the only text in which Kant
elaborated it is§ 16 of the B edition Transcendental Deduction of the Categories,
I will consider his explication of the I think as the analytic unity of apperception
in the remainder of this section and devote the next section to the implications he
drew from it regarding logical universality in the footnote at Bl33--4.
Section 16 begins with Kant's famous formulation of the principle that "The I
then present in sense and thought ("consciousness" being "a representation that
another representation is in me," L 33). For that consciousness cannot make the
consciousnesses of the contents that immediately precede and succeed it mine, nor
those of the contents that precede and succeed these, nor those that precede and
succeed them, and so on for all preceding and succeeding consciousnesses of the
contents of sense and thought. As distinct in their time of occurrence, if I were
to equate them, individually or collectively, with my self, then "I should have as
many-colored and diverse a self as I have representations of which I am conscious"
(Bl34). So the question is: what makes these and all other such particular con
sciousnesses, past, present, future, or merely possible, mine-my consciousnesses,
my representations?
Kant's answer proceeds from the recognition that the consciousness that is to
accompany each and every one of the individual consciousnesses of each moment's
representational contents must be a distinct consciousness in its own right-a dif
ferent representation from all these others-if it is to be used to represent them
as all my representations. This accompanying consciousness cannot, however, be
empirical, for then it too would be in time and would dissolve into as many distinct
consciousnesses as there were distinct times. Yet if it cannot be in time, then it can
not be given in inner intuition and so can never be an object of the senses (or of
sensible cognition). This means that the identity of the consciousness that accom
panies all my representations cannot be understood as identical with respect to
time, as is the case with the identity of the self, or person, with which Locke, Hume,
and other empiricist sensibilists concerned themselves. Instead, its identity must be
pure (not empirical), purely intellectual (not sensible), and purely discursive (not
aesthetic), that is, a thought identity rather than an intuited one.
How is such an identity of consciousness to be conceived? As the Paralogisms
chapter of the Transcendental Dialectic makes clear, Kant rejected the Cartesian
approach, which consists in explicating the identity of the I think in terms of tran
scendental concepts as a thinking substance endowed with various (causal) powers,
such as the capacity to have visual sensations, the capacity to retain in thought what
was present in sensation, the capacity to imagine, and so on (see also Conclusion-B).
Instead, Kant pioneered a new approach made possible by his explication of time
as a pure intuition of sensibility. For if, according to the latter, time is the product
of the representing subject, then the representing subject must exist outside of time
("the subject in which the representation of time has its original ground cannot
thereby determine its own existence in time," B422; chs 4-C and 8-F). This impli
cation becomes especially clear when one also recognizes that, for Kant, time, as
the product of an act of pure synthesis and the representation of that synthesis as
a unity, cannot exist outside or independently of the imagination and prediscur
sive understanding (chs 3-5 and 7-C). For just as a fantasized world can only be
inhabited by fantasized beings, not by the being that fantasizes it, the subject that
synthesizes time in its imagination and understanding cannot itself be one of the
synthesized beings present in that time (and so can be neither the empirical self nor
244 Kant's Anatomy of the Intelligent Mind
a human body or anything else within the synthesized realm of material nature: chs
7-8 and 17-18). There is, to be sure, a distinction of levels to be taken into account
as well; for, while the empirical subject-your self as it exists in time (the phenom
enon of thinking being experienced through inner sense) -is real relative to the
world of its fantasies, this same subject, viewed from the transcendental point
of view of pure intuitions of sensibility and pure concepts of the understanding,
reveals itself to be just another transcendentally ideal product of synthesis like any
other empirical object (B155-6; ch 18) so that the non-empirical subject responsible
for synthesizing the empirical subject is excluded from the empirically real world
in which the existence of the empirical subject temporally unfolds (a world that is
itself synthesized by the non-empirical subject: Pts IV-V). Thus, the identity of the
consciousness that must be able to accompany all our representations cannot be
given in intuition at all, but merely as "something real that is given, given indeed
only to thought in general, and so neither as appearance nor as an existent (Sache)
in itself (noumenon)" (B422-3n) (chs 4-C and 8-F).
What it means to represent this merely thinkable I as an identical consciousness
capable of accompanying all of my representations becomes clearer in light of the
second part of Kant's thesis in § 16: that any representation not capable of being
accompanied by the I think "would either be impossible or at least be nothing for
me." It is not difficult to see why. To represent something as my representation is
to represent it as accompanied by a consciousness that can be represented equipol
lently as accompanying every other possible consciousness of the contents present
in outer and inner sense and so as identical with respect to all the other representa
tions united in sensibility. This, indeed, is what defines it as my sensibility and the
representations contained in its unity as my representations. Any representation
that could not be represented as accompanied by this consciousness, assuming such
a representation were possible at all, would therefore not be representable as mine;
and since this is just to say that I cannot even so much as think it, it would ipso facto
be nothing to me, and so, as far as I am concerned, would be indistinguishable from
perfect nonbeing.
Assuming such a representation were possible (which Kant does not exclude),
would there be anything to prevent its being accompanied by another conscious
ness, and so belonging, together with others that my I think cannot accompany, to
another identity of consciousness, another I, in the same representing subject (i.e.,
mind/Gemiit: ch 8-F)? Though admittedly arcane even for Kant, it is an important
possibility to consider because it highlights the difference between the I think as
a fundamental transcendental principle to which all possible representations are
necessarily subject (at least where any not subject to it are precluded from belong
ing to a different I think), and the I think as a merely contingent, local identity of
consciousness, compatible with other such identities coexisting in the same repre
senting subject. In other words, it will help us to pinpoint the meaning of one of the
key concepts in the Kantian canon, especially as concerns apperception: original
(urspriinglich).
Concepts in Mind 245
When I turn to the analysis of the theory of objectivity elaborated in the tran
scendental deduction of the categories in Chapter 13, it will emerge that contingent,
local identities of consciousness are apperceptions corresponding to particular con
cepts functioning as rules in synthesis of recognition (A103-6), and are preceded
and made possible by (i.e., derived from) original apperception. My concern here
is exclusively with the latter, the I think in the capacity of a fundamental transcen
dental principle of unrestricted representational scope that precludes the possibil
ity of another I think in the same representing subject. More particularly, what
makes this representation of the identity of consciousness original is that it is in
no way dependent on and/or derivable from the representation of any other iden
tity of consciousness, while all other representations of identities of consciousness,
whether temporal (in pure or empirical inner intuition) or purely intellectual (e.g.,
algebraic concepts), depend on and derive from it (B132).
Universality of scope and necessary validity of such an order are only attain
able a priori and, more particularly, only insofar as the condition for forming the
thought of the identity of the I think is satisfied already in intuition itself, prior to and
independently of all thought (discursivity, representation by means of universals):
That representation which can be given ahead of all thought is entitled intu
ition. All the manifold of intuition has, therefore, a necessary relation to
the "I think" in the same subject in which this manifold is found.... I call it
pure apperception to distinguish it from empirical, or again original apper
ception because it is that self-consciousness which can be accompanied by
no further self-consciousness because it produces the representation I think
that must be able to accompany all others and is one and the same in all
consciousness. (B132)
As the sequel makes clear, the conditions intuitions must satisfy in order to make
possible an I think that has the status of an original identity of consciousness are
those responsible for bringing their manifold together in an original synthetic unity
of apperception (urspriinglich-synthetischen Einheit der Apperzeption). I want to
temporarily set aside the fact that this unity depends on synthesis in order to focus
here on the core claim: that the representation of the original identity of conscious
ness in respect of all the manifold-the representation of the I think as being able
to accompany all representations-is preceded and made possible by the original
unity of that manifold in one and the same consciousness:
self-consciousness, and even if this thought is not yet itself the consciousness
246 Kant's Anatomy of the Intelligent Mind
3 This passage, together with the accompanying footnote at B133-4, is the only occasion
I know of that Kant refers to the a priori representation of the identity of the I think as the
analytic unity of apperception. The use of 'analytic' in this regard seems to be related to its use
at A76/B102 to signify the pure-general logical process whereby representations are transformed
into concepts under which other representations fall, i.e. the production of a common concept
(conceptus communis) (also A78/Bl04 and Bl33-4n). When a conceptual consciousness is com
mon to other representations, there is an identity of consciousness in all these representations
(albeit a merely logical identity, not a real one, much less a temporal one). It thus may have
seemed to Kant natural to extend this notion of analysis to designate judgments in which the
consciousness of the predicate is logically identical to the consciousness of the subject in the
representation I think. See also Section C and Chapters 2n9, 10 and 13.
Concepts in Mind 247
It is not simply that the unity of sensibility effected by synthetic unitary pure space
and time meets the description of just such a universal self-consciousness perfectly.
When the further requirement that this unity be in place right in intuition itself,
ahead of all thought (discursivity), is taken into account, it becomes clear that the
universal self-consciousness in question at Bl32-3 can be nothing else.
What is the alternative? Were one to propose the synthetic unity of appercep
tion grounded on the categories, a vicious circle would result. For if the "analytic
unity of consciousness attaches to all common concepts as such" (Bl33-4), it must
also attach to the categories as pure concepts of the understanding ("The form of
judgments transformed into a concept of the synthesis of intuitions [brings] forth
categories that guide all employment of the understanding in experience," A321/
B378).4 But in that case, how could the categories be in any way responsible for the
synthetic unity of apperception that must already be in place in intuition, "ahead
of all thought" (Bl32), as a condition of the analytic unity of apperception that
is essential to all concepts as such, the categories included? The original synthetic
unity of apperception that precedes and makes possible the very analytic unity of
apperception that when added to a (nondiscursive) representation ipso facto "makes
it into a conceptus communis" (Bl33-4n) cannot, without circularity, be supposed
to involve concepts in its production. It must therefore be purely aesthetic and pre
discursive. And the only original synthetic unity of apperception that answers to
this specification is the original unity of sensibility constituted through pure space
and time.
One might seek to avoid this implication by supposing that the categories, as pure
concepts of the understanding that do not involve sensible intuition in their repre
sentation, cannot owe their logical universality to the analytic unity of appercep
tion and so must not, like it, presuppose the original synthetic unity of apperception
4 Also: "pure synthesis, universally represented, gives the pure concept of the understanding"
(A78/B104).
Concepts in Mind 249
that is in place in intuition ahead of all thought . Certainly, anyone who regards the
categories as conditions for the unity of apperception and not vice versa (ch 5-A
and -G) is committed to such a reading. But then how can it be distinguished from
the supposition that the categories are innate concepts? If they in no way depend
on sensibility and its unity, I see no way to justify regarding them as any less innate
than the ideas of true and immutable natures that Descartes supposed to be innate
to intellect (Section A, chs 7-D and 11-C). Yet Kant could hardly have been clearer
in his rejection of innatism. For him, the categories are original concepts, underiv
able from any others, yet acquired and not innate (Discovery 221-3; also Bl67-8,
CPrR 141, ID 395 and AA 18 §§ 4851, 4894, 5637). And, in any case, Kant made the
dependence of the categories on the analytic unity of the I think quite clear: "the
mere apperception I think ...makes possible all transcendental concepts" (A343/
B401) for, as "the vehicle of all concepts in general, and so of transcendental con
cepts as well" (A341/B399), "the proposition I think ...accompanies all categories
as their vehicle" (A348/B406).
If not the categories, might the logical functions of judgment from which these
concepts derive underlie the original synthetic unity of the manifold presupposed by
the analytic unity of apperception? Logical functions are indeed innate and so are
present in us prior to and independently of consciousness and its unity. Moreover,
since Kant's anti-innatism is directed not against innate faculties but innate rep
resentations, the innateness of logical functions posed no problems for him since
they are not representations at all but merely the particular forms that define the
understanding as a capacity to judge ("which is the same as the capacity to think,"
A81/Bl06) in beings constituted like ourselves (A69/B94). Yet this proposal too
seems a nonstarter. Logical functions are the very essence of discursivity. Their
inputs must be able to assume the form of subjects and predicates that can relate to
one another universally, particularly, or singularly, and affirmatively, negatively, or
infinitely. This means that logical functions can only apply to representations that
already have logical form, whether that of a concept (logical universality conferred
by the analytic unity of apperception) or that of a judgment (conferred by the logi
cal functions themselves). Yet logical form is precisely what representations lack
insofar as they are merely given in sensibility ahead of all thought as affections or
appearances, making them completely unsuitable as inputs for the logical functions.
Their nondiscursive unity can therefore only come about by means of purely aes
thetic modes of synthesis. And since the only aesthetic modes of synthesis capable
of producing an original synthetic unity of apperception right in intuition itself,
ahead of all thought (Bl32), are pure space and time, these, and these alone, can be
the basis of the unity presupposed by the original analytic unity of apperception,
not logical functions of judgment .
If the unity of sensibility is indeed the original synthetic unity of apperception
presupposed by the original analytic unity, how then must we understand Kant's
claim that a representation that the I think could not accompany "would either be
impossible or at least be nothing for me" (Bl31-2)? On this reading, the only way
250 Kant's Anatomy of the Intelligent Mind
out any one red thing or selection of red things from other red things. It thus counts
as logically universal in two senses: (1) its scope is universal because it relates to the
totality (universe) of possible representations, serving to divide that universe into
two mutually exclusive totalities, one comprising all things that are red and the
other all things that are not red; and (2) its reference is universal because it relates to
all red things indiscriminately, without singling any out from others (singular refer
ence is a use of concepts that are in themselves universal made possible by the logi
cal function of singular judgment: ch 10-B). Kant's thesis in the B133--4 footnote is
that logical universality is simultaneously explicated in both these aspects by the a
priori psychological notion of the analytic unity of apperception: the a priori repre
sentation of the original identity of consciousness (the I think) in relation to all the
manifold made possible by the original (synthetic) unity of all the manifold in that
one consciousness. And what is important to bear in mind in examining this claim
is that apperception here, uniquely, must be regarded from a purely general logical
perspective, that is, without regard to the content of concepts and, in particular, the
transcendental logical concern with the objective validity of the categories in pure
and empirical cognition.
A good place to begin is by asking what a representation is insofar as the ana
lytic unity of apperception does not attach to it. Since the addition of this unity is
what makes it into a conceptus communis, apart from this unity it can be anything
except a concept or concept-containing representation (judgment, inference). It
can be an empirical intuition, a pure intuition, a sensation, a desire, a passion,
or any other representation just so long as consciousness of it is possible prior to
and independently of logical universality. In short, it is a representation devoid
of logical form, and so entirely useless for all purposes of thought (judgment,
discursive representation) since, in the absence of logical form, a representation
cannot be slotted into the subject and predicate positions of judgments or have
logical quantity and quality. Thus, before the sensation of red can be employed
in such thoughts as "all red things are visible objects" and "some peppers are not
red," it must first be converted into a concept by means of the analytic unity of
apperception.
252 Kant's Anatomy of the Intelligent Mind
How does the analytic unity of apperception do this? The representation of the
I think is the a priori representation of an original consciousness that is identical in
respect to all possible representations and is therefore a consciousness that can be
represented as accompanying each and every possible representation. Now, to say
that a consciousness can be represented as accompanying each and every possible
representation is just another way of saying that it can be represented as common
to all possible representations. Does this mean that the I think is itself a conceptus
communis? To answer yes seems inadvisable for two reasons. First, all concepts have
both a matter and form; their matter is the object (content) thought in them, and
their form is logical universality (L 91). In the preceding section, we saw that the
I think is not intuitable but only "something real that is given . .. only to thought
in general, and so neither as appearance nor as an existent in itself (noumenon)"
(B422-3n). Since, in the absence of intuition, nothing manifold can be given, there
is consequently no material (content, object) from which to fashion a concept of
the I think.5 Second, to say that the I think "has" the form of a concept, logical
universality, is misleading insofar as it suggests that it is possible to represent the
logical universe (totality) of possible representations prior to and independently of
the I think. As an original representation, necessarily able to accompany all others,
and so the only consciousness logically common to them all, the I think is the one
and only representation with the intrinsic logical value of a universal. The psycho
logical representation of the original identity of consciousness in respect to the
manifold of all possible representations therefore constitutes, and so defines, their
logical universe.
Once this is recognized, it is easy to see why attaching the analytic unity of con
sciousness to a representation that is in itself devoid of logical form straightaway
"makes it into a conceptus communis" (Bl33--4). In thinking it (representing it as
accompanied by the I think), it is thereby related to the logical universe (totality)
of possible representations defined by the I think, and so partakes of its univer
sal logical scope. But because there is now a logical matter (content) in addition
to the formal logical universality defined by the I think, the tandem (the I think
and the sensation, intuition, etc. that I think) can now be thought as a logically
common mark (Merkmal) that divides those things in the logical universe that
resemble what is thought in the representation from those that do not. Moreover, it
does so completely indiscriminately, without singling out anything within the logi
cal sub-universe of things that resemble it from the rest. My thought is therefore
logically universal both in scope and reference. And that, in my view, is what Kant
meant when he asserted that the addition of the analytic unity of apperception to
representations devoid of logical form immediately converts them into common
concepts fit for employment in judgments.
One might object that a logical universe confined to the totality of possible rep
resentations is not truly universal and so is inadequate to capture the unrestricted
logical universality met with in linguistic quantifiers like "all." For one thing, the
scope of mental concepts, on Kant's account, exclude things in themselves (beyond
their mere existence as a something in general = X: ch 8), whereas no such restric
tion occurs in language. For another, since he did not preclude the possibility of
representations that fail to meet the conditions for being able to be accompanied
by the I think but nonetheless exist (insisting only that they would be "nothing for
me," B132), these too would fall outside the scope of mental concepts. Still, this
divergence in scope is a distinction without a difference. How would human under
standing be handicapped if objects and representations that are nothing to it, and
to which it is condemned by the conditions of its possibility to be forever oblivi
ous, were excluded from the scope of language? What possible semantic difference
could it make if the scope of discourse vacuously extends to objects and representa
tions our minds are incapable of ever intuiting or thinking through concepts? Since
everything that can ever be anything to me-the totality of my possible representa
tions-falls within the scope of the original analytic unity of apperception, there is
no reason to suppose that this unity could ever fail to be adequate to the scope of
any contentful discourse.
Kant's psychologistic explication of logical universality in terms of the synthetic
and analytic unities of apperception accords perfectly with his sensibilist princi
ples. Both representations have to be produced, the one synthesized in intuition
prediscursively and the other an originative act of thought made possible by this
prediscursive synthesis. The analytic unity of the I think is a purely intellectual
representation; yet because it (1) presupposes a synthesis of the manifold given
in sensible intuition, (2) has no matter (no manifold), and (3) represents nothing
except the formal relation of logical identity that the I think has to all the manifold
(Section B), it does not raise the specter of intellectualist representational innatism
that Kant rejected. At the same time, its contents are pure, not empirical. This
means that Kant's psychologistic explication of logical universality, in contrast to
the empirical psychological explications offered by Locke, Berkeley, and Hume
(Section A), makes possible not only empirical common concepts but pure ones as
well. Most important where the Transcendental Analytic is concerned, it opens up
a space for pure concepts of the understanding, a space Kant found that he could
fill in a manner consistent with his sensibilist principles by deriving their content
entirely from the logical functions of judgment, which become employable as con
cepts in transcendental judgments as soon as the analytic unity of apperception is
added (chs 11, 13, and 15).
One cannot help but be struck by the parallels between the I think in its rela
tion to common concepts and the relation of pure intuitions to appearances. Pure
intuitions are transcendental principles of appearances because they precede and
make them possible. Similarly, the analytic unity of the I think counts as a tran
scendental principle of concepts because it precedes and makes them possible. Pure
254 Kant's Anatomy of the Intelligent Mind
legitimacy or even indispensability, but whether and to what extent a given lin
guistic concept fails to be underwritten by ideas derived from the only reality ever
given to us: the affects and actions present to consciousness by means of the outer
and inner senses. To the extent that it is not, the concept lacks objective, ontological
content and validity (UU chs 4-C, 8-A, and HTC ch 3-C). No verbal definition of a
concept, however clear and precise, nor any amount of evidence as to its normative
indispensability to thought and action, can rectify this lack ("If we have really no
idea of a power or efficacy in any object, or of any real connexion betwixt causes
and effects, 'twill be to little purpose to prove, that an efficacy is necessary in all
operations," THN 168/113). Clearly, therefore, Hume's skepticism regarding cause
and effect and other fundamental concepts of concern to philosophers-identity
over time, complex individuality, substance, existence, space, time, universals, etc.
is always a question of the mental underpinnings of symbolic discourse, never a
question of linguistic propriety as such.
Once the centrality of the mental underpinnings of discourse to Humean skep
ticism is appreciated, it becomes clear that any supposition that Kant's concern
with it was peripheral can be sustained only by denying the centrality of Hume's
problem to Kant's transcendental philosophy or by denying that Kant understood
it correctly. That fact that Kant saw fit to term "Hume's problem" (PFM 313) the
problem he crafted his philosophy to solve-the possibility of synthetic a priori
judgments-proves the first course a nonstarter. Yet the second does not stand
up to scrutiny any better. For Kant was perfectly correct when he observed that
Hume never questioned the normative indispensability of concepts such as cause
and effect and that his inquiries were instead directed at establishing the origin of
each such concept in the mind, because in that way, "everything concerning the
conditions of its use and the sphere in which it can be valid would already of itself
have been given" (PFM 259). In particular, Kant understood Hume to have "proved
incontrovertibly that it is altogether impossible for reason, a priori and from con
cepts, to think such a combination as contains necessity" (PFM 257). Nor can it
be doubted that Kant recognized and endorsed the conclusion Hume drew from
this, given his insistence that "we can understand nothing except what carries with
it a correspondent in intuition to our words" (A277/B333). Thus, the evidence is
clear and unambiguous: Kant derived not only his problem from Hume but also
the method of solving it as well-psychologistic explication-so that virtually all
the differences between the two, however great they may seem, can be traced to
a single, slight divergence at the foundations: Kant's identification of a source of
representations, pure intuition, that Hume never considered (ch 2 and UU chs 1-3).
What Kant said of Reid et al., and could equally be said of the normativist Kant
presented in much of today's scholarly literature, is therefore not true of Kant him
self: "always taking for granted precisely what [Hume] doubted and demonstrating
with vehemence and often with great presumption what he never took into his head
to doubt" (PFM 258).
Concepts in Mind 257
Logical forms of judgment are at the heart of Kant's proof in the metaphysical
deduction of the categories' that
This list makes clear why it is difficult to exaggerate the importance of the logical
functions of judgments to the Metaphysical Deduction of the Categories. But they
have a wider importance as well. Philosophy, as Kant understood it, differs from
mathematical knowledge as synthetic a priori discursive cognition of objects differs
from synthetic a priori intuitive cognition (i.e., cognition through the construction
of concepts in pure intuition) (A 712-38/B740-66; ch 6). An absolute desidera
tum of any body of discursive a priori knowledge purporting to be philosophi
cal, according to Kant, is that it constitute a system. A body of a priori discursive
cognition is systematic if each and every question that can raised in it is answer
able completely and with demonstrative certainty (Axiii, A12-13/B26-7, A477-81/
B505-9, A762-3/B790-1, PFM 321-2); the whole must be understood functionally,
as the organic unity of its parts, and each of its parts through its function within
1 The metaphysical deduction of the categories is the topic of the first chapter of the Analytic
of Concepts, especially A76-80/Bl02-6. Although the expression "metaphysical deduction" is
actually used only once, at Bl59, it is current among commentators.
A Defense of Kant's Table of Judgments 259
the whole (Bxxxvii-viii, CPrR 10, PFM 261 and 263); and it is establishable beyond
any possibility of doubt that nothing essential is excluded and nothing inessential
included (Axiv, Bxxiii-iv). Showing that each of these conditions is satisfied is pos
sible only on the basis of a system schema rooted in the very nature of discursive
understanding, to which everything entering into the system conforms. And it was
Kant's contention that the logical forms enumerated in his table of judgments at
A70/B95 furnish just such a schema for a system of transcendental logic built on
pure concepts of the understanding ("the categories find their logical schema in
the four functions of all judgments," A406/B432; also A66-7/B91-2, AS0-1/B 106,
PFM 302, 305-6, 322-6, and Progress 271-2).
Kant's aspirations to systematic philosophy thus depend, in the final analysis, on
the systematic character of the table of judgments itself: that judgment is the funda
mental operation of discursive understanding from which all others derive; that the
form of judgments can be fully and precisely classified under precisely four "heads,"
each composed of exactly three "moments"; and that these are none other than the
twelve forms of judgment specified in Kant's table. Yet no tenet of his philosophy
has proven more controversial or is more commonly denigrated than his claim to
have exhaustively enumerated the logical forms of thought. Nor does criticism end
there, since Kant's subsequent steps in the metaphysical deduction have been found
equally questionable: his derivation of pure concepts of the understanding from
logical forms; his attempts to explicate the categories of traditional metaphysics in
terms of these concepts; and, last but by no means least, his contention that this
deduction furnishes a transcendental system schema capable of specifying every
thing from a system of the fundamental laws of nature (B164-5, PFM 302, 305-6,
320) to a system of civil law (letter to Jung-Stilling after March 1, 1789).
In this chapter, I will first attempt to show that the criticism leveled against
Kant's table of judgments loses its sting when (1) the logic of thought, considered
as a mental operation, is distinguished from the logic of its expression in language,
and (2) the basis of Kant's distinction between properly logical and extra-logical a
priori forms of thought, particularly his demarcation of the logical from the math
ematical (and, by extension, from the mathematical logical), is clarified. It will then
emerge that Kant's claim that his table furnishes an exhaustive list of logical forms
is by no means farfetched, if not altogether unassailable. Finally, in Chapter 11,
I will examine what is involved in converting logical forms of judgments into pure
concepts of the understanding capable of serving in the capacity of a transcenden
tal system schema and consider the basis on which Kant accorded these concepts
the status of pure concepts of objects.
language, it is never more necessary, in my view, than in the case of logical forms
of judgment. The fact that Kant himself did not stress the difference is precisely
why it so important for us to do so. For a sensibilist focused on representations
(Vorstellungen) in the mind in much the same sense Locke and Berkeley concen
trated on ideas and Hume on perceptions, it generally went without saying that
language, though a source of valuable clues (PFM 322-3), can never be relied on to
mirror the nature or workings of discursive mentation (UU chs 4-C, 8, 10, 16-G-5,
and 18-B). The risk of misinterpretation is compounded by the fact that a simi
lar silence prevails today for the opposite reason: the psychological consideration
of the individual isolated consciousness is held to be largely, or even completely,
irrelevant to the philosophical examination of propositional thought. Since this
includes logical forms, which tend to be assimilated to language (often as mere
abstractions, with no real separateness), it becomes all the more important, when
examining Kant's account of logical forms, to make explicit and understand the
status he accorded to them as features innate to the constitution of the mind and
so prior to and independent of language.2 At the very least, so proceeding should
help curb the common tendency both to hold the inadequacies of Kant's doctrine
in respect of language and its logical analysis against it and to treat disparities
between them as counterexamples.
Language, especially when understood holistically or, more broadly, as a form
of life continuous with the rest of human existence, defies attempts to identify a
fundamental discursive operation, segregate forms from contents, distinguish logi
cal from other kinds of form, and so on. Even if one ventures to try, its mixture of
the conventional and the innate makes it all but impossible to establish that one's
results are not simply artifacts of inquiry-useful, perhaps, for modeling linguistic
discursivity but less so for identifying anything rooted in its essential constitution,
prior to and independently of all empirical and social determinants. Indeed, the
very notion of an objective formal analysis of language may ultimately prove to be
a chimera. And even if it does not, the prospects of success, given the present state
of knowledge, are not particularly encouraging.
By bypassing language entirely, Kant's psychologistic explication of the formal
features of discursivity in terms of structures of consciousness-its unity, iden
tity, and manifoldness-at least has the virtue of drastically simplifying matters
(UU ch 4-D). Universality, for him, consists in attaching to otherwise ordinary aes
thetic representations (affections, intuitions) the analytic unity of apperception (ch
2 As remarked in the preceding chapter, Kant did not deny faculties innatism, only representa
tional (content) innatism, and the forms of judgment, like the forms of intuition (ch 5-E), ought
to be understood accordingly: Discovery 221-3. As such, they should be recognized as not only
prior to and independent of language but also as capable of characterizing the understanding
of creatures that lack language or even social, conventionally governed norms of any kind. For
Kant clearly supposed that the general lineaments of his analysis of understanding extend to
every sensibly conditioned understanding without exception (Bl45-6) and so without regard to
how radically other intelligent creatures may differ from humans.
A Defense of Kant's Table of Judgments 261
that by virtue of which a mental representation counts as a concept: its contents are
irrelevant as is its use as an indifferent denotation to call to mind any other repre
sentation associated with it and, a fortiori, all the additional significance it acquires
through its relation to the conventions and norms of symbolic language and the
wider context of human forms of life.
Matters are equally straightforward on Kant's account of the discursive employ
ment of concepts. To be able to put representations to use discursively, as con
cepts, they have to be relatable logically, qua universals, and not merely aesthetically.
This, however, the I think alone cannot explain. There must in addition be forms
that enable us to become conscious of relations between representations specifi
cally insofar as the analytic unity of apperception attaches to them (I will refer to
them as ADA-universals to contrast them with both the linguistic kind and the
consciousness-independent kind posited by intellectualists from Plato and Leibniz
to Frege and beyond). Because they govern only the analytic unity (universal scope)
that attaches to all concepts in the mind, these forms are fully as indifferent as that
unity itself to the content of these representations and, a fortiori, to their empirical
use as signs for any representations associated with them or any additional sig
nificance they acquire from becoming embedded in the wider network of linguis
tic and other human forms of life. The logical forms do one thing and one thing
only: enable one ADA-universal to be related to another. Thus, "the form of the
judgment consists in the determination of the way in which distinct representations,
as such, belong to one consciousness" (L 102), while judgment itself, considered
strictly psychologistically, is nothing else than "the representation of the unity of
consciousness of distinct representations, or the representation of their relation so
far as they constitute a concept" (L 101).
When judgment is considered as the exclusive means whereby otherwise iso
lated ADA-universals can alone be synthetically combined and put to represen
tational use, it becomes possible to see why Kant deemed it the most fundamental
of all discursive operations. This, needless to say, in no way implies that judgment
is also the most fundamental linguistic operation or that discursivity, outside and
independently of the individual isolated consciousness, is best explained in terms
of it or even that it can be so explained. For the crucial feature of Kant's logical
explication of judgment is its psychologistic character (ch 2-E, DD ch 1-C). It is
not just that forms of judgment could have no representational role and so would
remain dormant apart from the original synthetic unity of the manifold of intuition
262 Kant's Anatomy of the Intelligent Mind
in pure space and time, ahead of all thought, that alone makes analytic unity of
apperception possible (ch 9-B). The very meaning of these forms for Kant-their
immanent and sole content-consists in relating representations made universal
by the character conferred on them by this analytic unity. It is this that makes logi
cal forms inherently psychological functions of judgments, as fundamental to the
representational employment of concepts (AVA-universals) in discursive mentation
as affections are to the representational employment of intuitions in prediscursive
sensibility (A68/B93). Everything else, including the role of logical forms in cogni
tion as sources of pure concepts of the understanding, concerns not their content
(logical meaning) but their myriad discursive applications.
The psychologistic explication of a form of judgment is not, to be sure, a defini
tion. These forms are as indefinable as the I think itself, and for the same reason: as
constitutive elements of all judgments, and so too of propositions (assertoric
judgments),3 anything proposed as a definition would be itself a proposition that
presupposes forms of judgment and thus would be viciously circular (A245-6).4
Psychologistic explication is more modest . It does no more than trace logical
forms of judgment to their origin in the mind, showing that their content there
consists simply and solely of functions that enable one consciousness universal
ized by means of the analytic unity of apperception to be synthetically united with
another (a synthesis even analytic judgments require: Bl3ln). Yet by stripping out
everything having to do with the application of logical forms-their linguistic and
other conventional extensions and the content of the representations to which the
analytic unity of apperception attaches-psychologistic explication also opens up
the prospect of being able to achieve something that no other mode of philosophi
cal analysis seems able to do: isolate logical forms of thought from everything that
otherwise shrouds them in obscurity and entangles them in irremeable complexity.
The same is true when it comes to specifying logical forms of judgment individu
ally. According to Kant, in beings with understandings constituted like ours,5 there
are three kinds of relational forms of judgment: categorical (subject-predicate),
hypothetical (logical sequence), and disjunctive (logical community). The first is
the primary form of relation because it is the only one that extends to concepts
prior to and independently of their relation in judgments (the others only relate
concepts indirectly insofar as they have already been formed into judgments). The
primacy of this hoary Aristotelian form in Kant's logical analysis of judgment has,
however, been much criticized . One can pick almost any proposition at random
and raise questions about how much of its structure is clarified by analyzing it
into subject and predicate. What are the subject and predicate of "7 + 5 = 12"? Or
"Boys will be boys" and "Mammals are animals"? Are these not better analyzed
as identities (perhaps on the model of objective identity Kant himself provides in
his November 25, 1788, letter to Schultz: ch 2-A)? Alternatively, one might ask
whether there is really only one subject in "Jack and Jill joined hands as they strode
up the hill" and one predicate in "He cleaned the dishes with a sponge rather than
steel wool." Subject-predicate analysis seems even more inadequate in dealing with
embedded structure as in "The man who lived next to the house where Joan showed
the movie that she shot in Milan chased down the cop-killers' hard-to-catch-up
with super-souped-up getaway car." And what is the subject and what the predicate
of the interrogatory "What is the subject and what the predicate?"?
But these and similar objections are damaging only if one supposes that Kant
intended his analysis to be adequate to propositional discourse or that, for his
purposes, it needed to be. Conversely, their force is greatly dissipated if the cat
egorical form of judgment is understood not linguistically but psychologistically
as the means whereby alone aesthetic representations rendered universal by being
accompanied by the representation of the identity of the I think (AVA-universals)
can be synthetically united.6 A psychologistic explication specifies the essential
elements of the content of a representation without pretending to express, much
less exhaust, the additional significations it acquires from its various (e.g., cogni
tive) applications and (e.g. , linguistic) extensions; or, more succinctly, it specifies
its (core) meaning but not its use. To be sure, one can, without shifting from mean
ing to use, abstract from the transcendental content of the representation of the
categorical form of judgment and characterize its content in a non-transcendental
(non-psychologistic), general logical manner as the representation of the relation of
one otherwise isolated universal to another (i.e., without making mention of AUA).
But in so doing, one must be careful never to violate the limitations imposed on its
scope as revealed by its transcendental psychologistic explication ( U U ch 2-E-l). In
particular, since the I think constitutive of universality can only accompany those
representations that conform to conditions of the synthetic unity of apperception,
and since these include pure space and time as constitutive of the unity of sensibil
ity (ch 9-B), the scope of categorical and other logical forms can extend only to
the representations of outer and inner sense but never to the things in themselves
these sensible representations represent (even in the precognitive sense of "repre
sent" specified in ch 8-G). Thus, transcendental psychologism reveals, as general
logic could never do, that logical forms of judgment concern, and can concern, the
relation of categoricals in hypothetical judgments exist prior to and independently of their lin
guistic expression, it cannot be counted as evidence against Kant's table of judgments that lan
guages such as Tagalog and Lezgian lack grammatical subjects (see, e.g., Nicholas Evans and
Stephen Levinson, "The Myth of Language Universals: Language Diversity and Its Importance
for Cognitive Science," Behavioral and Brain Sciences 32 (2009) 440-42), or that there is no overt
conditional in Guugu Yimithirr (443).
264 Kant's Anatomy of the Intelligent Mind
AVA-universal with any other, the resulting judgments can only be synthetically
combined insofar as the two latter relational forms make this possible. This origi
nal synthetic unity of the thinking subject is an extension of the original synthetic
unity of all possible aesthetic representations in the intuiting subject considered in
Chapters 3-5, but only in a formal sense. The two synthetic unities, discursive and
aesthetic, concern the same representational contents, the same universe of repre
sentations, differing only insofar as the former adds discursive form to the aesthetic
form constitutive of the latter. Nevertheless, this formal extension is essential if
Kant's theory of understanding is to explicate cognition, understood as the combi
nation of intuitions with concepts. For it is the purely logical, noncognitive original
synthetic of the thinking subject that the synthesis intellectualis of the categories
transforms into the objective unity of apperception that, according to Kant, makes
possible not only cognitive experience but the very objects (corporeal and thinking
beings) we experience, their laws, and nature itself, as well (chs 11-C and 13-17).
Kant tended to use the expressions logical functions of judgment and logical forms
of judgment interchangeably (e.g., "Logical functions are only forms for the relation
of concepts in thinking," AA 23 E XLII at A78-9/B105). If there is a difference
in meaning, it is that "logical form" is abstract, not a specifically psychological
notion, whereas "logical function," understood as "the unity of the act of order
ing distinct representations under a common one" (A68/B93), is to understanding
what affections are to sensibility: they make it possible to use AVA-universals to
represent apprehended realities discursively ("the understanding can make no other
use of these concepts than judging by means of them," A68/B93), just as affec
tions make it possible to exhibit transcendental realities in intuition (the matter
of appearances corresponding to sensations corresponding to "the transcendental
matter of all objects as things in themselves," Al43/Bl82; ch 8-E and -G). There
also seems to be a subtle difference in Kant's use of the two expressions. Judgment,
for Kant, involves both a vertical dimension-the subsumption of representations,
whether concepts or intuitions, under concepts, understood as common marks
and a horizontal dimension-the subordinative relation of distinct concepts to one
A Defense of Kant's Table of Judgments 265
7 In my view, the best account of the logical forms of judgment and the derivation of the
categories from them can be found in Beatrice Longuenesse, Kant et le pouvoir de juger. The prin
cipal lesson I derived from her is that whenever the categories are at issue, one must always keep
266 Kant's Anatomy of the Intelligent Mind
universality is the form of all concepts as such, which means that, considered purely
logically, without regard to content, any subject of a judgment can also be used as a
predicate and vice versa. Insofar as a judgment is singular in logical quantity, it is not
because of the quantity of its subject, which, qua concept, is always universal, but
rather the quantity conferred on the judgment by the logical form of singular judg
ment (if language, with its grammatical apparatus of definite articles, indexicals, and
the like, is set aside so that only representations (Vorstellungen) are considered-as
it must be when analyzing Kantian general logic (Section A)-intuitions are the only
individuals). The categorical form of judgment must therefore be understood as the
way beings with understandings like ours are constituted to relate one ADA-universal
to any other, where, considered purely logically, it is a matter of complete indifference
which is subject and which is predicate.
Categorical predication can result in a judgment only with the addition of three
further logical determinants: the quality of the predication as affirmative, nega
tive, or infinite; its quantity as universal, particular, or singular; and its modality as
problematic, assertoric, or apodeictic. In affirmative predication, the content of the
predicate to be thought together with the content of the subject concept is added
to it, resulting in the consciousness of a complex unified content,8 while, in nega
tive predication, one is conscious of excluding the content of the predicate from
the thought of the subject concept. Either way, it makes no difference whether the
contents thought in the predicate have already been thought in the subject (they
may even be identical to it: A598/B626) and so whether the judgment is analytic or
synthetic, or even whether it is self-contradictory. For the quality of judgment is
merely a logical form of judgment, completely indifferent to the contents thought in
it or their compatibility in the relation in which they are thought. In other words,
this and other logical forms in Kant's table of judgments are merely formal deter
minants of judgment and so are necessary but not sufficient conditions for a judg
ment to result from a synthesis of concepts. Because the principle of contradiction
concerns the content of the representations related by these forms, it must therefore
be understood as an additional material condition of the possibility of judgments
(analytic no less than synthetic).
Kant anticipated that his inclusion of infinite judgments among the moments of
logical quality might be challenged:
uppermost in mind the logical functions from which they derive. I also found her analyses of the
individual forms extremely helpful and deem her criticisms of alternative analyses such as those
of Klaus Reich, Michael Wolff, and Reinhart Brandt convincing.
8 The inclusion of the content of the predicate in that of the subject reciprocally implies that
the subject falls within the extension of the predicate (L 103-4).
A Defense of Kant's Table of Judgments 267
Still, one may wonder whether this explanation is sufficient. If infinite judgments
are distinguishable from affirmative only insofar as an infinite sphere of being is
conceivable and such a sphere is conceivable only under extra-logical, properly
cognitive conditions that require one to take account of the "worth or content"
of judgments, is Kant not violating his own demarcation between the logical and
nonlogical by distinguishing them in in a purely logical table of forms of judgment
that purports to "abstract from all content of a judgment in general" (A70/B95)?
The question seems to me legitimate: a distinction of general logical forms without
a genuinely logical difference should have no place in a table ostensibly confined to
logical forms. It thus must be shown that a properly formal logical meaning can be
conferred on the distinction without recourse to transcendental logic and its prin
ciple of synthetic unity or any other extra-logical principle.
Although indications of how Kant might respond are present in the passage just
cited, his reasoning is even clearer in the following:
9 All universals are specifiable simply by arbitrarily adding marks (Merkmale) to a given con
cept so as to form new concepts with narrower spheres: see L §§ 8-15.
10
Even if no higher concept than the predicate existed in a particular case (or indeed in all
cases), infinite judgment would still be thinkable as a logical/arm, without regard to the content
of the judgment, which might be empty or unintelligible and so, materially, no judgment at all.
A Defense of Kant's Table of Judgments 269
satisfy his criteria for admission into formal logic, and so the only three logical forms
of quality there are.
The quantity of predication relates the subject and predicate concepts of a cat
egorical judgment according to their specifications. A concept S is a specification of
another concept P and is said to fall within its extension (Umfang) if, in addition to
thinking all the same contents, other contents are thought in S that are not thought in
P but not vice versa (L 103-4). For example, the concept "dog" (analytically) specifies
the concept "mammal," and "human" (synthetically) specifies "creatures susceptible
u Strictly speaking, the analytic/synthetic distinction applies not to concepts but to judg
ments. My formulation should thus be taken as abbreviating the claim that the judgments corre
sponding to these specification relations can be either analytic-if the concept of a mammal is a
constituent element of one's concept of a dog (which it need not be for all thinkers since concepts
outside of language-Lockean "rules of propriety," i.e., semanatic, syntactic, and pragmatic
conventions--are matters of an individual's experience, conceptual needs, and, in some cases,
whim)-or synthetic (the universal affirmative judgment "all humans are susceptible to infection
by retroviruses"). Since the form of judgments takes no account of their content, it does not mat
ter how-by logical identity (the principle of contradiction) or the X of synthetic judgment (A9/
B13)-one concept that is thought to be the specification of another actually relates to the other
(or even whether it in fact relates to it at all).
12
I do not mean to suggest that Kant explicitly and systematically distinguished subordina
tion (Unterordnung) from subsumption (Subsumtion) in this way. Nevertheless, the difference is
certainly present in his thought and accurately reflects the distinction he drew between judgment
as considered from the standpoint of pure formal logic and judgment considered nonformally.
In particular, subsumption typically tracks the difference between a rule and a particular case
of the rule and was used by Kant to distinguish the understanding from the faculty of judgment
( Urteilskraft): "If the understanding in general is explicated as the capacity of rules, then the
faculty of judgment is the capacity to subsume under rules, i.e. to distinguish whether something
does or does not stand under a given rule (casus datae legis)" (A133/Bl 72). In exercising our
faculty of judgment, we focus not on the relation between the universals (concepts or judg
ments) within it but on that outside the judgment to which the relation of those universals in the
270 Kant's Anatomy of the Intelligent Mind
the subject of J are also specifications of its predicate (the possibility that none are
fails to define a distinct logical form because it can be derived simply by negating
an existing logical quantity, be it a negative-particular or a universal-negative judg
ment). What is not thinkable, so long as one is confined to the judgment-to-judgment
horizontal axis, is the possibility that the subject of J has a determinate number of
specifications-one, two, three, a few, many, infinitely many, nearly all, or any other
properly mathematical quantity. For logical specifiability is illimitable: it is always
formally possible to incorporate additional constituent marks (Merkmale) into
a concept (if only capriciously and without regard to whether the contents added
materially contradict any of the marks already thought in the concept).13
With ADA-universality its only datum, the only quantitative distinction possible
in pure general logic is whether all or only some specifications of the subject of a
judgment are also specifications of its predicate. Greater quantitative determinate
ness and, in particular, the (symbolic or ostensive) representation of continuous
and discrete magnitudes only become possible when the context of the subordina
tion of one universal (concept or judgment) to another is left behind for that of
the subsumption of individuals under universals.14 Since Kant held that individu
als can only be given in intuition and that intuition, in beings like ourselves with
discursive understanding, is always sensible and rests on the synthetic unity of the
manifold of appearances in pure space and time, there can be no individuals in pure
general logic and so too no doctrine of their subsumption under universals. This is
why pure general logic, the science of the pure understanding considered by itself,
judgment applies. Logical subordination, by contrast, considers only their relation, without
regard to extra-logical factors: "According to relation judgments are either categorical or hypo
thetical or disjunctive, namely, the given representations in judgments are subordinated one to
one another for unity of consciousness either as predicate to the subject or as consequence to
ground or as member of the division to the divided concept" (L 104). Categorical judgment subor
dinates one concept-the predicate- to another-the subject- insofar as the latter is conceived
of as a ground, analytic or synthetic, for affirming the predicate of it (whereupon the subject is
recognized as a specification of the predicate). Hypothetical judgment subordinates one judg
ment-the consequence- to another-the ground-insofar as the assertion of the ground leads
immediately to the assertion of the consequence (whereupon the subject is affirmed as specifying
the predicate if the consequence is affirmative-universal: see Section C). In disjunctive judgment,
two or more judgments are subordinated to another insofar as the former are formed by the logi
cal division (mutually exclusive, conjointly complete specification) of the subject or the predicate
of the latter (Section C). Thus, in its pure general logical sense, " subordination" seems to be
confined to the relations between universals (concepts or judgments) within judgments, without
regard to their application outside the judgment.
in isolation from its relation to sensibility or any other faculty, must, on Kant's
conception, abstract from all content of concepts and judgments, and so from all
individuality, leaving only their form as ADA-universals or logical relations of such
universals (A52-5/B77-9). Only if, as Descartes, Leibniz, and other intellectual
ists supposed, discursive understanding could supply itself with content, including
ideas of individuals (numbers, the figures considered in geometry, God, etc.), is a
genuinely logical doctrine of subsumption, altogether independent of the relation
of understanding to sensibility, possible. Without purely intellectual individuals of
any kind, however, even mathematical quantity must be conceived as extra-logical
and, in particular, as dependent on the indissoluble union of aesthetic with logical
forms effected in imagination (transcendental and mathematical synthesis speciosa).
Thus, mathematical notions of quantity, including the abstract, purely symbolic
individuals over which algebraic and mathematical logical quantifiers range, fall
outside the scope of what Kant considered (pure general) logical quantity proper
and belong instead to the sensibility-dependent context of pure cognition (A51-2/
B75-6; chs 6 and 15).15
How then is one to understand the formal-general logical character of singular
judgments, which Kant included as a moment of logical quantity alongside univer
sal and particular? It cannot be stressed too strongly that on Kant's conception of
pure general logic, where aesthetic content is excluded from consideration, there
can be no individuals. Since there is no purely formal way, without regard to the
content of concepts, to set any limit to their specifiability, a logical form of singular
judgment does not even seem thinkable, much less something capable of contribut
ing to the logical determinateness of categorical judgments. Recognizing this, Kant
justified its inclusion in his table of logical forms of judgments as follows:
Logicians are correct to say that singular judgments can be treated just like
universal ones in rational inferences. For since they have no extension at all,
The shift from the horizontal axis of pure general logic to the vertical axis of sub
sumptive cognition in order to distinguish singular from universal judgments may
seem only to confirm the suspicion that Kant's table of judgments is defective. For
how, on his conception of logical form, can differences in the quantity of cognition,
where the aesthetic element is no less essential than the logical, be supposed to war
rant or even to make conceivable a strictly logical distinction of quantity, grounded
solely on the understanding in isolation from its relation to other faculties?
Kant's response is evident in the text just cited but even clearer in the following:
Though an I think that could not relate its concepts in judgments could hardly be
accounted a thinking subject at all, an I think that could not relate its judgments to
form compound judgments and inferences is hardly any better. The problem that
confronted Kant in the latter regard arises from the fact that concepts and judg
ments are universals of very different kinds. Unlike concepts, judgments already
have a logical quality: they affirm or deny some relation of concepts and so (assum
ing the concepts do not contradict) have the capacity to be true or false ("Truth and
falsity lies not in concepts but in judgments, and indeed these as propositions," AA
16 § 2259 [early 1790s]; also A293/B350, L 53).16 Since the categorical form of judg
ment relates concepts rather than judgments, it is unsuited to the latter purpose.
Therefore, other logical forms are requisite, specifically adapted to relate distinct
judgments, irrespective of their content and their truth or falsity.
16
Kant treated propositions as a subclass of judgments, though not always in the same
way: "In judgment the relation of distinct representations to the unity of consciousness is
thought merely as problematic; in a proposition, on the contrary, it is thought as assertoric.
A problematic proposition is a contradictio in adiecto.-Before I have a proposition, I must still
first judge" (L 109; also AA23 E XXXVIII at A74). Alternatively: "Judgment and proposition
are distinct from one another. The judgment is the sense; the proposition: how the concepts are
placed in relation, whether categorical or hypothetical or disjunctive. The very same judgment
can be expressed by distinct propositions" (AA 16 § 3100 [early or mid-1770s]). Since, as will
emerge shortly, stand-alone relations are always assertoric, the difference between the two char
acterizations turns out to be only superficial.
274 Kant's Anatomy of the Intelligent Mind
Kant held that in beings constituted like ourselves, there are two, and only two,
logical forms capable of relating judgments. The hypothetical relation of conse
quence (Konsequenz) subordinates one judgment to another by making it possible
to conceive the truth or falsity of the one (termed "consequence") as dependent on
the truth or falsity of the other ("ground" or "antecedent") (A 73/B98, L 105-6).
The disjunctive relation of a community of opposites subordinates the truth or
falsity of two or more judgments to the truth of another by conceiving the sub
jects (or predicates) of the former as specifications of the subject (or predicate) of
the latter in such a way that the spheres of these specifications nowhere overlap
yet jointly coincide with the sphere of the concept they specify (hence, all further
specifications of the latter are specifications of one and only one of its disjunctive
specifications: A 73--4/B98-9, Bl 12, L 106--8). Both judgment-relating forms must
be understood so as to satisfy all the criteria Kant considered essential for some
thing to count as an element of pure general logic, including a complete indiffer
ence to (1) everything in any way concerned with the content of the concepts in the
judgments they relate (the matter, or objects thought in them), (2) the actual truth
or falsity of those judgments, and (3) the presence of singular or infinite judgments
(i.e., nothing in hypothetical or disjunctive relations suffices to distinguish singular
from universal or infinite from affirmative judgments: Section B). And presumably,
in Kant's view, all other relations between judgments fail to satisfy one or more of
these criteria and so cannot be counted as logical relations in the sense recognized
by pure general logic.
Virtually no one today would accept that Kant's list of three relational forms of
judgment is well grounded, much less exhaustive. Why, for example, are bicondi
tionals, conjunctions, and disjunctions that take no account of the relations of the
concepts in the disjuncts excluded? Do equivalences such as that between "if p then
q" and "not-p or q" show that hypothetical form is not primary but derivative? And
do innovations like the Sheffer stroke not expose Kant's list of relations between
judgments to be comparatively superficial, lacking any basis in fundamental logical
principles?
Certainly, Kant cannot be blamed for reflecting the logical thought of his
period. Moreover, many commentators would exonerate him, at least partly,
by confining their view of his list of logical forms to their application in tran
scendental logic as providing the bare-bones, purely intellectual meaning of the
framework of pure concepts of the understanding at the heart of his theory of
cognitive experience. Yet there can be no doubt that Kant himself believed his list
was and, more important, needed to be impeccable from a strictly logical point of
view, prior to and independently of any application-transcendental or empiri
cal, cognitive or noncognitive. So, there seems to me no way to avoid the chal
lenge of determining whether an effective defense, drawn from elements of his
philosophy, can be mounted precisely where his list of logical forms seems most
defective: its restriction of logical relations between judgments to hypothetical
and disjunctive forms.
A Defense of Kant's Table of Judgments 275
The first line of defense focuses on the conception of truth and falsity employed
in defining truth-functional connectives: is it properly logical or does it depend on
considerations that, by Kantian criteria, ought to be counted as mathematical? In
other words, is truth-functional "logic" of the kind we find in De Morgan laws logic
or a calculation technique-a technique that, among many possible applications, is
useful for modeling formal-logical relations within and between propositions but
which should never be confounded with these relations themselves? A situation simi
lar to this has already been met with in the case of logical quantity: by excluding
individuals (manifoldness) from the domain of logic, no matter how abstractly rep
resented (a priori ostensively or symbolically), Kant implicitly also excluded any
thing capable of being counted, ordered l-to-1, or otherwise functionally related,
with the consequence that "all" and "some" cannot be interpreted in the manner
of quantificational logic without transgressing the boundary that for him demar
cates pure general logic from mathematics (chs 6-D, 14-E, and UU ch 4-B). The
question is whether something similar occurs when truth and falsity are conceived
of as truth-values. That the meaning of truth-functional connectives can be exhib
ited by such methods as truth-table analysis raises the question, more particularly,
whether truth and falsity are not being conceived as (abstract) individuals capable
of being counted and related by a species of mathematical function. By substitut
ing 1 and 0 for T and F in the truth table of a TF-connective, the latter emerges
as a function that assigns either a 1 or a 0 for any possible combination of 1 and 0
inputs. But then why limit it to 1 and O? If true and false are viewed as applications
(interpretations) of abstract mathematical functions, the latter emerge as simply
one species of a wider class: functions that take 2, 1, and 0 as arguments and assign
1 or 0 as values; functions that take 3, 2, 1, and 0 as arguments and assign 1 or 0 as
values; functions that take 1 and 0 as arguments and assign 2, 1, and 0 as values;
and so on. Other modes of analysis besides truth tables also invite one to view
truth-functional logic as simply an application of a calculus that, in its uninter
preted form, is as amenable as elementary arithmetic to algebraic generalization
and other forms of mathematical analysis. If so, then the same reasons that oblige
a Kantian to treat techniques of symbolic construction such as arithmetic and alge
bra as mathematics rather than logic (ch 6) would carry over to truth-functional cal
culi.Truth-functional equivalences (De Morgan, Sheffer, et al.), for example, would
count as synthetic objective identities comparable to 2 + 2 = 3�64 (ch 2-A) and
so be, for Kant, mathematically meaningful but logically meaningless. And such
misconceptions would be inevitable right from the very first step: the notion that
truth and falsity are countable-two of something (truth possibilities, truth values)
in a logically significant sense-number being, for Kant, a notion impossible inde
pendently of pure sensibility (succession and juxtaposition) and its transcendental
synthesis speciosa in conformity with the categories of quantity (chs 6-C and 15-E).
Denying truth-functional logic the status of logic does not, of course, make it any
less a source of sound and significant insights, much less imply that it is in any sense
deficient or defective. It is simply to say that from a Kantian standpoint, the insights
276 Kant's Anatomy of the Intelligent Mind
its use provides are of a mathematical rather than a logical nature because they depend
as much on formal aesthetic as formal logical elements.17 I have already considered
reasons for this demarcation from the aesthetic side in Chapter 6. Here, I want to look
at it from the side of logic with specific regard to the logical relation of judgments. As
representations universal in scope, judgments, for Kant, are possible only by means
of the analytic unity of apperception. The latter is a pure intellectual discursive repre
sentation devoid of aesthetic content in which nothing is represented beyond the bare,
formal identity of the I think in respect to all appearances (ch 9-B). As such, the "one"
signified by this identity cannot be construed aesthetically, as something represent
able quantitatively (countable, etc.), but purely logically, as signifying only that the
I is a consciousness of universal scope that is common to every representation insofar
as that representation can be anything to me at all, no matter how much it may differ
from others (ch 9-C). It is true that Kant sometimes spoke of it as a numerical identity
(Al07, Al 13). Yet in the absence of any manifold, and so too the possibility of synthe
ses of succession and juxtaposition, the I is a unity that is neither repeatable, countable,
divisible, nor multipliable, and so must not be confused with the number one, a unit of
measure, or anything else to which number, the synthesis speciosa of the categories of
quantity (ch 15-E), is applicable (Bl31). Neither these nor any categorial or categori
ally based determination can be applied to the I think (A346/B404, A401-2, B422). Its
analytic unity is simply the bare form of logical universality, "numerically one" only
in the purely logical sense that one and the same representation-the I think-is com
mon to every representation that can ever be anything to me (chs 9-B).
To this it might be objected that one can count concepts and judgments just as
surely as one can sensations, intuitions, and representations generally. Yet, while
true in one regard, it is not in the sense requisite for truth-functional logic to be
classified as logic rather than mathematics by Kant's criteria. Attaching the analytic
unity of apperception to representations and combining them according to logical
forms of judgments are indeed mental actions, and so representations. But being
representations does not in and of itself make them intuitions or in any way sensi
ble. Intuitions of these spontaneous acts arise only insofar as inner sense is affected
and the resulting affections are exhibited as homogeneous temporal appearances in
inner intuition in accordance with pure time (ch 4-C). The intuitions belong to the
manifold contained within this synthetic unitary individual and, as such, are suited
for mathematical synthesis-pure as well as empirical, symbolic as well as ostensive,
algebraic and mathematical logical as well as arithmetic and geometrical (ch 6). But
this does not make the ADA-universals and the acts of judgments that combine
them such a manifold as well. On the contrary, their constitutive forms-the ana
lytic unity of apperception and the logical forms of judgment-make possible only
relation by specifiability in the case of concepts (universals), and by assertability
17 Since Kant had a far higher opinion of the synthetic, ampliative results of mathematical
cognition than he did of the nonsynthetic, noncognitive propositions of the science of logic, he
would likely have viewed the classification of T-F logic, quantification, set theory, et al. as math
ematics rather than logic as an elevation in cognitive status.
A Defense of Kant's Table of Judgments 277
conditions (section D) in the case of judgments. So just as the Kantian logical quanti
ties "all" and "some" need mathematical/aesthetic supplementation before they can
take on the significance accorded to them in quantificational logic, the capacity to be
true or false attributed to judgments in Kantian logic must be supplemented by math
ematical/aesthetic underpinning before it can take on the significance accorded to it in
the calculus of truth functions.
The upshot of the considerations in the previous section is that truth-functional connec
tives (including negation) are, in the final Kantian analysis, as ineluctably aesthetic as
geometry, Newtonian mechanics, or anything else that is possible only through juxtapo
sition in space and succession in time. Yet if truth values are excluded, what (pure gen
eral) logical meaning remains for the hypothetical and disjunctive forms of judgment?
And, more particularly, how does this meaning allow otherwise unrelatable judgments
to be combined to form new judgments by logical (purely nonaesthetic) means alone?
The following passage indicates that, as conceived by Kant, the logical meanings of
hypothetical and disjunctive forms of judgment are as ineluctably bound up with the
modalities of the component judgments in relation to the whole as the logical meaning
of categorical judgment is bound up with the quantity and quality of the relation of
the predicate to the subject:
What all three moments of relation have in common is that the relation itself is always
assertoric: "In categorical judgments, nothing is problematic, everything assertoric" (L
105}--except when categoricals enter into hypothetical or disjunctive relation to one
another, in which case they are problematic and only their hypothetical or disjunctive
relation to one another is assertoric (L 105-7, AA 16 § 3012, AA 24 765-6 and 932).18
18 Although Kant does not, to my knowledge, discuss cases of multiply embedded judg
ments-judgments within judgments to two or more levels (the bottommost consisting of
categoricals)---1 see no reason for him not to admit them. My guess is that he would regard the
relation that governs the whole of the judgment as assertoric and all others as problematic, down
to their ultimate categorical constituents.
278 Kant's Anatomy of the Intelligent Mind
This suggests, first, that categorical judgments can only be conceived as relatable if, and
only if, they can be conceived problematically (rather than assertorically) and, second,
that hypothetical and disjunctive forms of judgments are the sole means, in beings with
discursive understandings constituted like ours, whereby categorical judgments are con
ceivable problematically.
This way of understanding the logical meaning of hypothetical and disjunc
tive form rests on the assumption that assertoric modality is the modality of all
stand-alone19 categorical judgments-their default modality as it were-and that
unless there were a way of representing their modality otherwise, they would be
completely unrelatable to one another (with the consequence that there would be
no inferential relations as well). This means that in order for Kant's claim that "In
categorical judgments, nothing is problematic, everything assertoric" (L 105) to
hold true, seemingly stand-alone problematic categorical judgments ("It may rain
today") must be construed as components of implicit hypothetical or disjunctive
judgments (further reason to be wary of using language as a guide to propositional
mentation: Section A).
Why should problematic categoricals be supposed incapable of standing on
their own? Though Kant appears never to have directly addressed this question,
the only answer that seems capable of yielding his thesis that categorical judgments
not embedded in hypothetical or disjunctive judgments are assertoric is that such
judgments are not in fact judgments at all because, as problematic, they fail to actu
ally affirm or deny a predicate. Problematic judgments "give only the matter [i.e.,
subject and predicate concepts: L105] with the possible relation between predicate
and subject" (L109) and so are "accompanied with a consciousness of the mere
possibility of judging" (L 108).2° For a logical relation (judgment) to become actual,
assertion must occur (assertion just is "a consciousness of the actuality of judging,"
L108; also A76/B101). So long as problematic judgments are embedded in larger
relations which themselves are assertoric, the conditions exist for "a free choice
whether to admit such a proposition as valid, a merely discretionary acceptance
of it into the understanding" (A76/B101). But where problematic judgments are
neither explicitly nor implicitly functioning as components of actual (i.e., asser
toric) hypothetical or disjunctive judgments, they are "possible judgments" only
in the sense in which smoke and fire, prior to and independently of experience and
custom, are "possible associates" in reproductive imagination: mere potential. In
any stronger sense, however, the notion of a stand-alone problematic categorical
judgment is, for Kant, a contradiction in terms.
19 That is, categorical judgments that are not components of hypothetical or disjunctive
judgments.
20
The logical form of problematic judgment is, to be sure, insufficient by itself to guarantee
the logical possibility of the relation of the contents in a judgment, because in addition to the
possibility of judging, the concepts themselves must not contradict. See Al50-3/Bl89-93 and
Progress 278.
A Defense of Kant's Table of Judgments 279
21
Nor does the importance of this universal yet purely logical ordering principle end with
solving the problem of how the I that produces categorical judgments is then able to relate them.
For the I that has the power to relate its judgments thereby also gains the freedom to posit any
thing it is capable of thinking without ipso facto asserting it. This freedom, essential to the very
notion of a thinker, derives directly from the logical meaning of hypothetical judgment. But it is
also central to Kant's notion of moral freedom since, in his view, freedom is conceivable only by
means of the category of cause and effect, the principal component of which is none other than
the hypothetical form of judgment. Thus, the freedom of the logical subject (the I think) to assert
or not to assert problematic judgments, far from being merely a precondition for a moral subject's
freedom to choose (will) the maxims of its conduct, is directly continuous with it and neither can
be fully understood independently of the another.
280 Kant's Anatomy of the Intelligent Mind
disjunction consisting of the remaining terms, so that the assertion of the negative
judgment leads immediately to the assertion of that disjunction. Given a sequence
of such hypothetical judgments (inferences), each of which has the effect of elimi
nating one term of the original assertoric disjunction, the result will be the assertion
of the sole remaining disjunct.
This is not to say that reasoners do or should explicitly articulate their inferences
step-by-step in hypothetical form. But it does mean that all reasoning implicitly
utilizes the hypothetical form of judgment and is impossible otherwise, since it is
the only means available to minds constituted like ours to conceive problematically
distinct, stand-alone, otherwise unrelatable assertoric (categorical, hypothetical, and
disjunctive) judgments, as the inferential relation of these judgments demands. To be
sure, logical relations are not the whole story of inference. Propositions can also be
related by unfolding the content of their constituent concepts and either compar
ing them to one another under the law of contradiction (in the case of analytically
related propositions) or exhibiting the objects corresponding to them in intuition,
be it empirically (to include sensation, feeling, or desire) or a priori (mathematical
construction, transcendental synthesis speciosa). But even here, reasoning from one
proposition to another still depends on hypothetical form. In geometrical construc
tion, for example, each step must be guided by the content thought in the concepts
that are to be exhibited a priori in intuition; and since relations between concepts
can only be expressed in judgments, each step in such a construction must consist
in the sensible exhibition of something that cannot be represented otherwise than
in the form of a judgment. But in order for these stand-alone (hence assertoric)
judgments to be related so as together to constitute a course of reasoning-a dem
onstration with premises leading through intermediate consequences to a desired
conclusion-they must first be conceived problematically. For simply exhibiting the
propositions in temporal sequence does not suffice to form them into an inferential
sequence (consequentia: L 121). The latter can be produced only insofar as each of
the intuition-grounded propositions is conceived as a problematic judgment whose
assertion leads immediately to the assertion of the next in the sequence; and this is
possible only by means of the logical form of hypothetical judgment. Thus, if not
in and of itself sufficient for inference, hypothetical form, for Kant, is always its
necessary formal condition.
In its inference-grounding role, hypothetical judgment also makes judgments
with logically apodeictic modality possible. All stand-alone judgments are asser
toric except when, in the course of an inference, their assertion is the consequence
of the assertion of a hypothetically related antecedent judgment, the simplest case
being a modus ponens inference:
Though Kant did not specify which logical laws these are, one is surely the pres
ence of an assertoric consequence-relation (Konsequenz) that serves as a rule (L
121-2) under which the assertion of the antecedent leads not only immediately but
apodeictically to the assertion of the consequence. An assertion of a consequence
that occurs under such a rule-whether the rule takes the form of an assertoric cat
egorical, hypothetical, or disjunctive judgment-will be apodeictic in modality. An
extended course of reasoning may involve such a rule for each step, together with an
overarching rule from which the conclusion follows, so that all of the propositions
inferred during the course of the reasoning will be represented apodeictically. For
example, let the principal rule be the assertion of a disjunctive relation of affirma
tive-particular categoricals (premise 1) and the next assertion (premise 2) be a cat
egorical judgment, one of whose concepts specifies a concept in one of the disjuncts
(i.e., one of the concepts yielded by the non-overlapping yet exhaustive logical divi
sion of the concept on which the disjunction is based). By forming a hypotheti
cal judgment (premise 3) in which the categorical judgment (premise 2) is thought
problematically as the ground and the disjunct it specifies as the consequence, one
now has an intermediate rule from which the consequence follows immediately and
apodeictically (premise 4) from the assertion of its ground (premise 2). The newly
asserted disjunct is then represented problematically as the ground of an assertoric
hypothetical judgment (premise 5) that has the negation of the remaining disjuncts
as its consequence. Under this rule, together with the principal rule (premise 1, the
original disjunction), the negation of the remaining disjuncts now follows immedi
ately and apodeictically (Conclusion).22 Thus, by means of the three logical modali
ties alone, without depending on truth-functional logic or anything else that, by
Kant's criteria, would require aesthetic form (symbolic construction) in addition to
properly logical form, he was able to account for the relation of judgments, both as
components of hypothetical judgments and as steps in inferential reasoning (which,
to be sure, is not to deny that truth-functional logic can accurately model Kantian
pure general-logical inference and, indeed, greatly amplify its power by opening the
way for other computational methods to be employed, effectively mathematizing
reasoning).
22
For example, let the disjunctive judgment that serves as principal rule be "The world exists
either through blind chance or through internal necessity or through an external cause" (A74/
B99) and the premised categorical be "God created the world." From the categorical together
with the subsidiary rule "If God created the world, then the world exists through an external
cause," it follows that "The world exists through an external cause" which, under the principal
rule, y ields the conclusion "The world exists neither through blind chance nor through internal
necessity."
282 Kant's Anatomy of the Intelligent Mind
Kant's inclusion of disjunctive form and no other alongside the two other logi
cal forms of relation (categorical and hypothetical) assumes that this form alone,
among all other candidates one might propose, is neither derivative from categori
cal or hypothetical nor in any way dependent on aesthetic form. Yet it cannot be
emphasized too strongly that this holds true only if the disjunction is formed by
two or more divisions of a concept in such a way that the spheres of each of the
divided concepts reciprocally exclude the spheres of the others and their spheres
together coincide with the sphere of the undivided concept, so that the individual
disjuncts count as logical opposites and the disjuncts collectively constitute a logical
community (A 73--4/B99, Bl 12, and L 106-8). For if, as argued earlier, the exclusion
of aesthetic forms precludes a truth-functional construal of the rule "A or B or C"
(which does not require that either the subjects or the predicates of these proposi
tions divide up the sphere of a single concept), a disjunctive rule "A or B or C"
not built on the division of a concept cannot constitute a distinct logical form in
its own right. T he reason is not that inferences such as that from assertoric "N' to
apodeictic "Not (B or C)" cannot be understood in such a way as to follow strictly
logically from the assertion of the rule "A or B or C," but rather that the assertoric
rule under which they follow is then not, in truth, "A or B or C" but rather one
or more of the following hypothetical judgments: "if A then not (B or C)," "if B
then not (A or C)," "if C then not ( A or B)," "if not A then (B or C)," and so for
each possible hypothetical combination. For suppose that none of these hypotheti
cals had previously been premised in the reasoning. How, in that case, could the
assertion "B or C" follow purely logically (hence non-truth-functionally) from the
mere assertion of "Not N'? If the hypothetical "if not A then (B or C)" is nowhere
premised (asserted) AND none of the constituents logically exclude one another
because their subjects or predicates non-overlappingly divide up the sphere of the
same concept, then nothing at all follows from the assertion of "not N' under the
rule "A or B or C." T hus, the exclusion from pure general logic of truth-functional
connectives implies that when the rule "A or B or C" is not construed in terms of the
non-overlapping division of a concept, it cannot be classified as a logical form in its
own right but is instead merely shorthand for the hypothetical rules just enumerated
so that when the latter are all asserted, there is nothing more for "A or B or C" to
say, i.e., no further properly logical meaning for it to express.
To this some may object that a collection of hypothetical rules is capable of
capturing the logical meaning of "A or B or C" howsoever it is construed, whether
in Kant's narrowly constrained logical manner or any other. To this a Kantian logi
cian would surely respond that hypothetical relations are incapable of expressing
precisely what is most distinctive of disjunctive relations: the reciprocal determina
tion made possible by the logical division of a concept that transforms the indi
vidual disjuncts into logical opposites and the disjunction relating them in a logical
community. Of course, the importance of a logical form that incorporates both
opposition and community into its meaning, like that of singular and infinite forms,
only becomes apparent in the cognitive domain, where the concept of community
A Defense of Kant's Table of Judgments 283
is crucial (chs 15-D, 16-C-3, 17-I, and 18). Nevertheless, because the difference
between disjunctive form, as construed by Kant, and the other logical forms of
relation is definable in strictly (pure general) logical terms of concept specification,
it is just as deserving of a place among the fundamental logical forms of relations
as the third logical forms under the headings of quantity and quality are.
Once the logical meaning of disjunctive judgments is understood, the logical
relation of the subsidiary hypothetical rules employed in disjunctive syllogisms to
the primary disjunctive rule also becomes clear. As noted earlier, when a disjunctive
rule of the form "A or B or C or ..." is understood without regard to the logical
relationships between the constituent concepts, it can have no autonomous logical
meaning but must instead have it conferred by construing it to abbreviate a col
lection of hypothetical rules ("if A then not (B or C or ...)," etc.). But when it
is understood as conforming to Kant's logical form of disjunctive judgment, the
assertion of the combination of logical opposition and community in a disjunctive
relation immediately and apodeically yields the assertion of the various hypothetical
rules (assertoric hypothetical judgments) utilizable in the course of any disjunc
tive syllogistic course of reasoning. In this respect, then, disjunctive syllogisms are
the unique exception to the otherwise universal Kantian principle that the form of
hypothetical judgment is necessary to all inferred conclusions.
Since Kant saw no reason to suppose that the logical forms of judgment we find
in ourselves are not the contingent result of the constitution of our minds (Bl45-6,
A230/B283, Progress 272), a demonstration that the three logical forms of relation
listed in his table are all there are seems no more possible than in the cases of the
forms of quantity, quality, and modality. Indeed, not only can the possibility of
creatures with differently constituted discursive understanding not be precluded,
but even in ourselves, the possibility that new analytical techniques may someday
reveal logical forms that meet all of Kant's criteria of the logical cannot be ruled
out a priori (though Kant was skeptical that anything truly new in logic would ever
be found: Bviii-ix). Suffice it to say that if, as I have maintained, truth-functional
considerations have no place in logic, as Kant distinguished it from mathematics,
then his table of fundamental (original, underived) logical forms seems as complete
as it is likely ever to be.
23 The following texts help illustrate the point (emphases mine): "judgments ... are nothing
other than the unity of consciousness in the relation of concepts in general, indeterminately,
whether that unity is analytic or synthetic" (Progress 272); "A judgment is the representation of
the unity of consciousness of distinct representations, or the representation of their relation so far
as they constitute a concept" (L 101); "the form of the judgment consists in the determination of
the way in which distinct representations, as such, belong to one consciousness" (L 102); "judg
ments are either categorical or hypothetical or disjunctive, namely, the given representations in
judgments are subordinated one to one another for unity of consciousness either as predicate to
the subject or as consequence to ground or as member of the division to the divided concept" (L
104); "The logical function is the act of uniting one and the same consciousness with many repre
sentations" (AA 18 § 5642 [early 1780s]); "In every judgment we represent an analytic unity" (AA
16 § 4273 [early 1770s]); "We can think only through judgments because we have concepts only
through the unity of consciousness" (AA 18 § 5650 [mid-l 780s]); "Judgment is the consciousness
that a concept is contained under another concept, either as its predicate or its ground or as a
member of its division" (AA 16 § 3053 [1780s or 1790s]); and "that action of the understand
ing whereby the manifold of given representations (be they intuitions or concepts) is in general
brought under one apperception is the logical function of judgment" (Bl43).
A Defense of Kant's Table of Judgments 285
I think is therefore, for Kant, the one and only representation with the intrinsic logi
cal value of a universal, the only originally universal representation, and so too the
one from which all others (all concepts as such) derive their universality. And from
this, the psychologistic implication follows that we either contradict ourselves or
talk without a meaning if we suppose that anything can be universal independently
of ADA.24
The second alternative-not limiting logical form to the combination of univer
sals-would have the effect of expanding the scope of the concept of logical form
far beyond the twelve functions listed in Kant's table of judgments. It might then
include algebraic, truth-functional, quantificational, and/or set theoretical opera
tors and connectives. But quite apart from their dependence on the sensible mani
fold (Section C and ch 6), their inclusion would completely undermine Kant's claim
that his table provides a complete list of logical forms of judgment, with "judg
ment" understood as the most fundamental act of (discursive) understanding,
definitive of it as a faculty (the capacity to think defined as capacity to judge: A69/
B94 and A80-1/B106). And since without his original table, there could be no table
of categories or system schema of the kind he required, Kant would certainly have
opposed any attempt to open up the concept of logical form in this manner.
One of my principal goals in the present chapter has been to show that when
the logical functions making up Kant's table of judgments are understood purely
psychologistically, as principles for the combination of ADA-universals, a plausible
case can be made for its completeness. However, showing that they are essentially
psychological does not suffice to prove, conversely, that our psychology is essen
tially characterized by these forms. This, it seems to me, is what most fundamentally
differentiates the psychologistic explication of the logical functions of judgment
from that of logical universality. In Kant's system, no matter how sensibly condi
tioned understandings otherwise differ, all must have some means of attaining ana
lytic unity of apperception if they are to have concepts available for employment in
judgments. By contrast, there is no necessity, at least that Kant could ascertain or
prove, that ADA-universals are combinable into judgments by means solely of the
twelve logical functions in his table. He could show only that understandings consti
tuted like ours are innately determined to produce judgments from ADA-universals
by means of these and no other logical forms, but not that all sensibly conditioned
understandings as such necessarily employ the same forms to this end (B145-6,
A230/B283, Progress 272).
Indeed, the psychologistic explication of logical functions perfectly parallels that
of space and time as regards both indirectness and conditionality. The psycholo
gistic explication of space and time starts from the demand for unity of sensibility.
This, as we saw in Part II, is the necessary synthetic unity of sense affections in
one and the same consciousness that results insofar as pure intuitions enable us
to exhibit even the most heterogeneous affections through homogeneous appear
ances, any (and all) of which admit of being represented immediately together as
the manifold of a single intuition, via the synthesis of apprehension in intuition.
Since without unity of sensibility no higher syntheses would be possible-including
association, judgment, and cognitive experience-all sensibly conditioned under
standings necessarily have pure intuition. The same is not true, however, of pure
intuition of space and time: all that can be claimed is that pure space and time are
the aesthetic means whereby alone sensibly conditioned understandings constituted
like ours can attain unity of sensibility.
The psychologistic explication of logical functions starts with an analogous
demand for unity of discursivity. This, we have seen, is the necessary synthetic
unity of ADA-universals in the judging subject that results insofar as logical func
tions enable us to synthesize every possible concept with any other in a judgment
and to synthesize the resulting judgment with any other in further judgments and
inferences. Since, in the absence of such a unity, discursivity (representation by
means of universals) would be impossible, logical functions of some kind are just
as essential to all sensibly conditioned understanding as such as pure intuitions of
some kind are. The same is not true, however, of the particular logical functions
listed in Kant's table of judgments: all that can be claimed is that subject-predicate,
ground-consequence, etc. are the logical means whereby alone a sensibly condi
tioned understanding constituted like ours can attain unity of discursivity. Thus, just
as Kant's transcendental psychologism posits that the unity of the intuiting subject
must be synthesized from sense affections by means of some pure intuitions but
not necessarily ours, it likewise posits that the unity of the thinking subject must be
synthesized from ADA-universals by means of some functions of judgment but not
necessarily ours.
{ 11 }
Kant's derivation of the table of categories from the table of judgments in the
metaphysical deduction of the categories (MDC) has two sides: the derivation itself
and the status accorded to pure concepts of the understanding as pure concepts of
objects in general.1 It is important to distinguish them when attempting to under
stand the MDC because of the inherently puzzling nature of the claim that empty,
purely logical forms of discursive representation are capable-in and of themselves,
without any contribution from sensibility or any other source-of yielding con
cepts with a "transcendental content" that "relate to objects a priori" (A79/Bl05)
irrespective of how the objects are given (A253-4/B309, CPrR 54, 136, 141, letter to
Beck, January 20, 1791) or even whether they can be given at all (A290-2/B346-9).
Logical functions are nothing other than modes for relating representations insofar
as the analytic unity of apperception attaches to them (chs 9-C and 10-A). Since
this unity-the representation of the I as identical and so common to all the mani
fold-is nothing more than the bare form of logical universality devoid of all con
tent and objective relation, Kant's claim that the understanding is able to extract
concepts of objects from mere logical functions cannot but seem suspect. I shall
therefore focus solely on the derivation of the table of categories from the table of
judgments in the first two sections of this chapter and turn to the more contentious
objectivity claim in the third.
The principle of the derivation of the categories from the logical functions was
never more clearly stated than in PFM § 38, where Kant claimed that Aristotle's
1 See, for example, A50-1/B74-5, A94/Bl25-6, Bl 13, B128, B146, B158, A248/B305, A290/
B346, CPrR 50, 65-6, PFM 324, letter to Beck October 16, 1792, AA! 7 § 4638, AA 18 §§ 5643,
5854, 5930-3.
288 Kant's Anatomy of the Intelligent Mind
2 The expression 'letzte Subjekt' is used in reference to substances at A205/B250. Although the
notion of a final predicate does not figure in Kant's discussion of the categories, the logical func
tion of categorical judgment would seem capable of yielding such a concept as well: predicates
that can only be thought as predicates, never subjects, so that no concepts they can be predicated
of can ever be thought as their predicates. Kant's neglect of it may be due to its lack of cognitive
application, much less a role in the constitution of experience.
290 Kant's Anatomy of the Intelligent Mind
unity of apperception, already universal in form and related to another such uni
versal as subject to predicate-and (2) no possible use except to ensure that these
concepts are always employed either as subjects (= substances) or as predicates of
subjects whose employment has been so determined(= accidents of substances):
Setting aside the objectivity aspect for the time being, this passage makes par
ticularly clear the extent to which the categories, explicated as pure concepts of
the understanding, coincide with the logical functions of judgment. For Kant, the
3 This B edition addition was foreshadowed a year earlier by a footnote in the preface to
the Metaphysical First Principles of Natural Science in which Kant first publicly expressed his
intention to revise the Transcendental Deduction of the Categories by remedying the obscurity
surrounding the notion of judgment in the A edition version: "the table of categories completely
contains all pure concepts of the understanding, and so too all formal actions of the under
standing in judgment, from which these concepts are derived and differ in nothing except that,
through the concept of the understanding, an object is thought as determined in respect of one
or another function of judgment. Thus, for example, in the categorical judgment the stone is
hard, the stone is used as subject and hard as predicate, in such a way that the understanding
remains free to exchange (umzutauschen) the logical function of these concepts and to say that
a hard thing is a stone; by contrast, if I represent it as determined in the object that the stone, in
every possible determination of an object (not of the concept), must be thought only as subject,
and the hardness only as predicate, these same logical functions now becomepure concepts of
the understanding of objects, namely as substance and accident" (MFPNS 475n, analyzed in ch
12-C). Kant is sometimes criticized for mixing up the predicative copula ("The stone is hard")
with the "is" of identity ("A hard thing is a stone") as well as playing fast and loose with language
("A hard thing is a stone" is awkward to say the least). What needs to be kept in mind is that
there is no requirement, nor even a desideratum, that logical functions reflect linguistic propriety.
They are concerned solely with the psychological operation of synthesizing ADA-universals, and
categorical judgment in particular is simply the mode in which psyches constituted like ours
perform this synthesis on their ADA-universalized representations. It does not predict, much
less imply, that all or even any human languages will reflect this function in their grammar, nor
does it entail that users of logical functions even have the capacity to express their propositional
thought symbolically. The temptation to suppose differently, tacitly or otherwise, can easily be
avoided simply by consenting to take Kant's psychologistic method seriously. For however much
one's philosophical training leads one to suppose that categorical judgment, like logical forms
generally, is non-psychological in nature and objectively independent of all mentation, according
to Kant's transcendental logical explication of the possibility of general logic,
it is the psychologi
cal operation of synthesizing AVA-universals in a judgment that explicates categorical judgment,
not vice versa.
The Metaphysical Deduction of the Categories 291
When I say "Some men are individuals," this "some" is a plurality, to be sure,
but it is determined only logically in comparison with the representation . . ..
292 Kant's Anatomy of the Intelligent Mind
But when I represent to myself a being in such a way that a one contains sev
eral, I have the concept of magnitude. For example, "Some men are learned"
and "Some learned individuals are men," I can reverse the proposition. But if
the proposition is determined in such a way that it cannot be reversed, then it
is a magnitude. (AA 2 8 / 1 472)
To understand this correctly, one needs to keep in mind that the notion of mag
nitude employed here can be construed neither in terms of part/whole relations
(composites, extensive magnitudes), number, nor quantitative concepts of the kind
employed in post-Fregean mathematical logic. According to the Kantian concep
tion of the demarcation between the logical and the mathematical, quantitative
concepts of extensive, numerical, and other magnitudes are bound up by content as
much with the forms of sensibility as the forms of judgment (chs 6-E, 10-C, 14-E,
15-E and -F). By contrast, the categories of pure magnitude, as pure concepts of
the understanding, derive their content exclusively from the logical functions of
quantitative judgment and so must be understood simply and solely as determi
nations that serve to fix the logical places of concepts related according to these
functions.
The derivation of the categories of magnitude from the logical functions of
quantity begins with the recognition that, apart from such determination, a con
cept A related to another concept B by quantity, whether universally or particularly,
can equally well be thought of as quantified (specified) by the other (thus "some
learned individuals are men" is just as thinkable as "some men are learned"). So
long as their quantitative relation remains reversible, however, no magnitude can
be conceived through either of these concepts (much as a whole cannot also be
thought as a part of its own part, a number that contains other numbers within it
cannot also be thought of as contained in any one of them, etc.). By fixing their
logical quantitative relation so that it holds in only one direction, the categories of
quantity make good this want.
The pure concept of magnitude is, to be sure, extremely weak since the catego
ries of quantity must be understood through pure concepts of the understanding
that derive their sole and entire content from the logical functions of quantity. It
goes no further than the representation of a concept A as always and only the logi
cal specification of another concept B, with no implication that instances subsumed
under A are parts of (extensive or intensive) wholes constituted by instances sub
sumed under B, that instances subsumed under A are numbers that enter into the
numbers constituted by instances subsumed under B, etc. Kant's explication of the
categories of magnitude entirely in terms of pure concepts of the understanding
formed from the logical functions of quantitative judgment thus denudes them of
all mathematical significance; they are compatible with any mathematical relation
and its negation and would remain significant for pure understanding even if there
were no mathematical (including mathematical logical) relations at all. Whether
this bodes well or ill for their employment in transcendental judgments remains to
The Metaphysical Deduction of the Categories 293
be determined. But what is important here is that it makes strikingly clear how per
fectly, in Kant's scheme, categories of magnitude coincide with the logical functions
of quantitative judgment.
Kant's example of the irreversibility of "Some men are learned" concerns, more
particularly, the formation of the pure concept of the understanding that explicates
the traditional category of plurality. To form it and thereby think "man" as a plu
ral magnitude in respect to "learned individual," nothing more is required than to
conceive the otherwise reversible quantitative relation thought in the judgment to
preclude "learned individual" from ever specifying "man." This, for Kant, is the
sole and entire content of the pure concept of understanding: the logical positions
given with the logical form of particular judgment are thought as purely intellectual
predicates added to the content of concepts related according to this form, thereby
fixing their logical quantitative relation. His claim in the metaphysical deduction of
the categories is therefore this: when all contents extraneous to pure understand
ing are stripped from the category of plurality, this determination of concepts in
relation to the quantitative logical functions of judgment is all that remains to be
thought in the pure concept of the understanding.
The other categories of quantity-unity (measure) and totality (whole)-are
best understood in relation to the concept of magnitude that emerges with plu
rality.4 Unity derives from the logical function of singular judgment,5 which dif
fers logically from universal judgment only insofar as the subject is thought as an
unspecifiable concept (ch 10-B). This means that a concept A thought through this
category agrees with concepts determined by the category of plurality to the extent
that any concept B to which it is related as a constituent magnitude cannot in turn
be thought as a constituent magnitude of A (i.e., the judgment relating them in
logically reverse order is precluded). But whereas concepts determined conformably
to plurality can in turn have other concepts whose relation to them is determined
as that of plural magnitude (A as a magnitude comprised of C's, D's, etc.), no
concept whatsoever can be represented as entering into the magnitude of A once
A is determined as unity (unit, measure). How then does the determination of A by
4 In the presentation of the table of categories in PFM, Kant inserted measure (Maj]) paren
thetically after unity (Einheit), magnitude (GrojJe) after plurality (Vielheit), and whole(Ganze)
after totality (Allheit).
5 Since singular judgment is listed third under the moments of quantity in the table of judg
ments and unity is listed first under the moments of quantity in the table of categories, many
assume that Kant traced unity not to singular but to universal judgment. But it is difficult to
make sense of the notion that unity corresponds to universal logical form and totality to singular.
Kant's trichotomous ordering of the categories under each heading seems to be due not primarily
to the sequence of the corresponding functions of judgment in the table of judgments, but rather
to the fact that the third member of each categorial trichotomy must be understood as resulting
from the combination of the first two (Bll 0--1; cf. CJ 197n) (there is no reason to suppose that he
held that the third logical function of judgment under each heading is a combination of the first
two). For an incisive analysis of the correspondence between the categories of quantity and the
logical functions of quantitative judgment, see Beatrice Longuenesse, Kant et le pouvoir de juger.
294 Kant's Anatomy of the Intelligent Mind
the category of unity differ from the same concept not so determined but occupy
ing the position of subject of a singular categorical judgment? Though the dearth
of textual evidence again obliges one to conjecture, the answer would seem to be
that so long as A is undetermined by the category, one remains free to think it as
the subject (or predicate) of nonsingular judgments as well (general logic, it will
be recalled, must completely ignore the content thought in the concept and so too
whether or not it is, in fact, specifiable). It is only when A is determined conform
ably to the category of unity that it must, in all extra-logical contexts, be thought as
always and only unspecifiable and so is never conceivable otherwise than as a unit
of magnitude (never a plurality).
The category of totality is explicated by a pure concept of the understanding
formed when the logical function of universal judgment is conceived as a deter
mination that fixes the order of concepts related by this function. The function of
universal judgment enables one to think every specification of the subject of a cat
egorical judgment as a specification of the predicate as well (whether analy tically,
in virtue of the content of the predicate, or on the basis of some synthetic a priori
or a posteriori ground: ch 10-B). But whereas the order of the concepts involved in
a universal judgment can be reversed and so do not represent genuine magnitudes,
the addition of the pure concept of the understanding, in fixing the order as irre
versible, determines both the subject and all its specifications as magnitudes that
enter into the predicate. Of course, there is more to the category of totality than
this: since the specification of any concept is capable of proceeding without end, the
only magnitude one can conceive simply by stipulating that every specification of
the subject is also a specification of the predicate is an indeterminate plurality, not
a totality. Before one can conceive the latter, the category of unity must be factored
in as well so that the process of specification (ultimately) terminates in unspecifiable
units. Yet even then, an infinity of units can prevent one conceiving a magnitude as
a totality (Bl 11). Only when a non-infinite "plurality is considered as unity" (Bl 11)
can a magnitude be represented as a totality (whole).
How can two such disparate concepts as unity and plurality be combined? As
pure concepts of the understanding distinguishable only in terms of the purely
logical property of specifiability, the category of totality would seem to require
that one and the same concept be thought as at once specifiable (a plurality) and
unspecifiable (a unit). Although Kant does not spell it out, I can think of only one
way he could have done so in the context of the metaphysical deduction, where all
sources of content besides logical form are excluded (transcendental synthesis spe
ciosa, ostensive and sy mbolic mathematical construction, etc.). Totality can result
from the combination of unity and plurality when the addition of the former to the
latter serves to determine as not further specifiable the concept already determined
as the plurality. For even if the units are finite and the plurality contains them all,
the resulting magnitude cannot be represented as a totality until the possibility of
adding new units is precluded. Where magnitude must be understood entirely in
terms of logical specifiability, it can consist only in the thought ("special act of the
The Metaphysical Deduction of the Categories 295
6 Kant offers number as an example of a totality. Since the schema of the categories of quan
tity is designated number as well, it is not the best example he could have given of what the
unschematized categories of quantity do when they determine concepts. His point seems to be
that, say, five is not just the thought of the unity of five units-which would only be to conceive
them as a plurality-but the thought that no more units can be added, which makes five not sim
ply the aggregate composed of five units but itself a unit. Once five is so determined, the addition
of further units must be conceived either as a plurality composed of five and the new units or as
yet another totality (whole), i.e., a different number.
296 Kant's Anatomy of the Intelligent Mind
non-being," AA 18 § 6338a [1794-5]). This does not mean that one could not still
employ an affirmatively determined concept as the predicate of a negative judgment
or a negatively determined concept as the predicate of an affirmative judgment.
Doing so, however, would in both cases have the extra-logical significance of exclud
ing content from the subject. For just as predicating an affirmatively determined
concept negatively means that something positive, quite properly termed a reality,
is excluded from the subject, predicating a negatively determined concept affirma
tively means that something negative, which excludes a given reality and is quite
properly termed a negation, is added to the subject. The difference between this
and the same judgment considered solely from the perspective of pure general logic
should therefore be clear: the strictly logical functions of affirmation and negation
take no account whatsoever of the content of the predicates concerned in them
and so are completely indifferent to whether they are positive or negative. Thus, the
addition of the categories of quality at once introduces an extra-logical content to
judgment and, because they are pure concepts of the understanding derived exclu
sively from the logical functions of quality, does so purely intellectually, without
depending on contributions from sensibility or any other source external to discur
sive understanding.7
The pure concept of the understanding that psychologistically explicates the cat
egory of limitation derives from the logical form of infinite judgment. A judgment
is infinite in form if its predicate is the negative of a concept P (i.e., non-P) where
(1) P is conceived as a specification of another concept C and (2) the spheres of P
and non-P taken together coincide with the sphere of C (ch 10-B). A pure concept
of the understanding becomes possible when C is conceived as a positive concept
(reality), making both P and non-P positive as well. Its locus is not P, however, since
P's contents suffice by themselves to determine its sphere. Non-P, by contrast, has
no content; insofar as it is determined as a positive concept, and so a reality in its
own right, the understanding is therefore faced with the problem of determining
its sphere. This it does by combining negation and reality in a new concept of the
understanding: the representation of the negation of P within the reality of C, or
limitation ("limitation is nothing else than reality combined with negation," B111).8
For a limit is conceivable only if a reality (quality) is already there for it to limit (cf.
A169-70/B211); if C were not thought under the category of reality, and so was
7 A text that seems to capture this difference is the following: "Logically determine means to
affirm or deny a predicate of a thing (copula in a judgment) regardless of the content; metaphysi
cally to determine means to ascribe to a concept a predicate that contains in itself a being or
excludes such a being. What in itself excludes a being is negation, and only in the opposite can
the real be found." (AA 18 § 5704 [1783-4])
8 There is, to be sure, a positive concept D that specifies C and whose sphere coincides with
that of non-P (for C, like every specifiable universal, is specifiable without limit: ch 10-B). But
since D's contents, like those of P, suffice to determine its sphere independently of C, its consid
eration can shed no light on the procedure of the understanding whereby it delimits the sphere
of reality represented by non-P.
The Metaphysical Deduction of the Categories 297
9 This is what results when one reads back from the schematized category of cause and effect
(objective succession: ch 16-C-2) to the unschematized one. See, for example, AA 18 § 5167 (late
l 770s/early 1780s): "Subjectively, representations A and B can succeed this way or conversely. In
order for the succession to be valid objectively, i.e. serve for experience, it must be determined so
that I cannot reverse them."
10
The same text mentions an exception that will be considered shortly.
The Metaphysical Deduction of the Categories 299
u This does not conflict with Kant's assertion that "the inference from a given effect to a deter
minate cause is always uncertain, since the effect may have originated from more than one cause"
(A368). The focus of my analysis is the meaning of the category of possibility-impossibility itself
as it relates to relations of cause and effect. By contrast, Kant's concern at A368 was with empiri
cal applications of the category of cause and effect, where there is an ineliminable risk not only
of misidentifying the cause of any empirically given effect but also of only partially identifying
it, thus leaving open the possibility that the identified cause may occur without the effect follow
ing because the unidentified part of the cause is lacking. In the transcendental context, where
synthetic a priori knowledge is philosophical-discursive and data of intuition do not come into it
at all, there can as yet be no identification of causes, and so no risk of wholly or partially misiden
tifying them. The meaning of possibility-impossibility in relation to cause and effect concerns the
complete cause of any given effect and says that-whatever it happens to be, whether or not it is
correctly and fully identified or even whether it is empirically identifiable at all-it is unique. So
if different things seem to be causes of the same effect, it is not what makes them different that
is efficacious but some feature that all share which causes that effect and no other, always and
necessarily. Similarly, for the complete effect of an empirically given cause: where the cause seems
to give rise to different effects, there is some feature that all the effects share that is its true effect,
the unique effect which that cause produces always and necessarily. In the transcendental context
where this is true, the only agents and patients are substances, and substances (the permanent
in respect to both past and future time), as it happens, are not identifiable empirically: A205-6/
B250-1 (discussed in ch 17-H).
300 Kant's Anatomy of the Intelligent Mind
a pure concept of cause and effect that relates two judgments to one another in an
irreversible ground-consequence relation that at the same time excludes their being
similarly related to any others. It thus should be clear why Kant would have deemed
this concept sufficient for the psychologistic explication of the category of cause
and effect in terms exclusively of pure understanding, to the exclusion of all other
sources of content, sensible or any other.
The significance of the category of possibility-impossibility for the category of
cause and effect extends also to causal sequences. We have seen that the former
precludes multiple causes of the same effect and multiple effects of the same cause,
making each causal relation unique. In addition to restricting the relation in this
way, it also restricts the order of the terms in it: when a possibility (P1) thought as
ground is determined as cause and the possibility (P) thought as its consequence
is determined as its effect, the P1 --+ P2 order counts as extra-logically possible, and
the P2 --+ P1 order becomes extra-logically impossible. Since the ground of the
ground of a consequence is logically that consequence's ground as well, the same
irreversibility dynamic carries over from proximate to remote causal relations as
well. Accordingly, the causal series P1 --+ P2 --+ P3 --+ P1 is extra-logically impossible
because P1 --+ P2 precludes the possibility of P2 standing in any kind of grounding
relation, indirect no less than direct, with respect to P1• This makes it impossible
for any possibility to recur in the same causal series. Moreover, given a universal
law of causality applicable to all possibilities as such, causal irreversibility has the
further consequence that each cause has one and only one possibility as its effect,
and each effect has one and only one possibility as its cause. And though it may not
seem difficult to cite counterexamples from experience, we need to remember what
is in question here is not an empirical causal law but a transcendental one relating
12
specifically to substances (final subjects) and their determinations (ch 15-C).
Existence-nonexistence, in the pure concept of the understanding, derives from
the logical function of assertoric judgment. As we saw in preceding chapter, asser
toric judgment is the default modality of judgment: to make a judgment-to effect
a categorical, hypothetical, or disjunctive relation in thought-is to assert it. The
logical leeway this leaves, and which the category may be supposed to eliminate,
is the freedom to make or not make (including unmake) judgments from given
12
See note 11 above.
The Metaphysical Deduction of the Categories 301
not; a cause and effect relation exists if the categorially determined hypothetical
judgment that represents it is made but otherwise is non-existent; and a relation
of community exists if the categorially determined disjunctive judgment that rep
resents it is made and is non-existent if it is not. Accordingly, when the category
of existence-nonexistence eliminates the freedom to make or not make a judgment,
it carries with it the thought of something = X external not only to the judgment
but to thought itself that leaves the judging subject no longer extra-logically free to
make or not make the relation (thinking can no longer make it so or not so, however
much I may wish things might be otherwise). Since this Xis simply a place-holder
so long as the category is not related to anything outside pure understanding (it
only becomes sensation when that faculty is related to sensibility), the derivation
(metaphysical deduction) of the category of existence-nonexistence from the logical
function of assertoric judgment excludes all sources of content outside pure under
standing, and so qualifies as a genuine concept of pure understanding.
The necessary connection between cause and effect might be thought to hold
the key to comprehending the modal category of necessity-contingency were it not
for Kant's characterization of necessity as "nothing else than existence that is given
through possibility itself " (B111) and contingency as "something the non-being
of which can be thought" (B290; also A459-60/B487-8). So construed, the neces
sary existence thought in the category seems indistinguishable from the representa
tion of something the nonexistence of which is at once impossible and unthinkable
and so a concept exemplifiable by nothing less than the deity of the ontological
proof (a substance the essence of which is to exist, self-caused existence). Yet such
a reading seems irreconcilable with Kant's general procedure of explicating the
categories via pure concepts of the understanding that draw their sole and entire
content from logical functions of judgment. For example, while the logical func
tion of categorical judgment enables us to form a concept of substance as a final
subject (always subject, never predicate), it does not furnish the additional con
ceptual resources requisite to conceive it as having a nature (essence) from which
existence necessarily follows. Substance, as far as pure understanding is concerned,
is simply one of the two nodes of a categorical relation and nothing prior to or
independently of that relation. Similarly, possibility-impossibility, understood via
the logical function of problematic judgment, always relates to the copula of judg
ments and has nothing to do with their content, which is limited to the quantity,
quality, and categorical relation of their conceptual contents. And so it is, too, with
302 Kant's Anatomy of the Intelligent Mind
13 The modal character of the category of existence is the ground of Kant's critique of the
ontological argument for the existence of God (A592-602/B620--30), especially with respect to
the distinction between logical and real predicates (A598/B626). The only sense in which "exists"
can count as a predicate is as a determination of the copula of a judgment, not its subject; and
only nonsense can result from the attempt to affirm it of (i.e., add it to) the subject. For the same
reason, "God does not exist" is no less unintelligible. Thus, if Kant is to be accounted an atheist,
the ground lies not with the predicate but the subject: his analysis of the concept of God as a
transcendental ideal riddled with subreptions that render it incapable of objective employment,
thus precluding even the thought of God's possibility or existence, much less cognition of them.
Nevertheless, Kant held that the concept, warts and all, still has a use for practical reason as a
postulate.
The Metaphysical Deduction of the Categories 303
to render the ordering of its components (the occupants of its logical positions)
irreversible. What freedom is there in forming apodeictic judgments that the cor
responding pure concept of the understanding annuls? The most likely candidate,
I believe, is the freedom either to subject a judgment to a rule of the understand
ing and think it as the conclusion of an inference, with apodeictic copula, or to
regard it independently of any such rule and so think it with non-apodeictic copula.
The pure concept of the understanding eliminates this freedom either by deter
mining the judgment so that it must always be thought independently of rules of
the understanding, never under them, or by determining it so that it must always
be thought under such a rule. In the former case, whenever existence (or nonexis
tence) pertains to the possibility thought in the judgment, the existence is always
contingent because, in the absence of any rule of the understanding connecting the
existence to something else via cause and effect or community, there is no condition
whose satisfaction can prevent us from thinking the nonexistence (or existence) of
that same possibility (where the contingent is that "the non-being of which can be
thought," B290).14 In the latter case, whenever existence (or nonexistence) pertains
to the possibility thought in the judgment, the existence (or nonexistence) is always
necessary because, under the condition thought in the rule of the understanding
(now irreversibly determined as irremovable), existence (or nonexistence) follows of
necessity whenever that condition is satisfied.
It is in this hypothetical (cf. A228/B280) sense of necessity alone that exis
tence can be thought in a pure concept of the understanding as "given through
possibility itself" (Bl11): insofar as conditionedness is built into the possibil
ity of any substance-accident relation that is also subject to a law of the under
standing, existence is indeed given through possibility. Since the logical function
of apodeictic judgment suffices for the formation of no concept of necessity
other than hypothetical-conditional (or, in one case to be considered shortly,
disjunctive-exclusionary), absolute necessity, by contrast, is beyond the capacity of
pure understanding to think by means of its pure concept (in relation to absolute
necessity, the hypothetical necessity thought by pure understanding counts as a
species of contingency: A458-60/B486-8, A560-l/B588-9). And it is therefore this
pure concept of the understanding that furnishes the category of necessity with the
psychologistic explication it needs to stand up to Humean skeptical scrutiny (chs
17-18).
Community, or reciprocity between agent and patient, is the only category to
elicit from Kant an explicit defense of its coincidence with the corresponding logi
cal function, disjunctive judgment (ch 10-D):
14 The scope of contingency is, however, greatly limited as soon as the categories of cause and
effect and community are determined to be among the conditions for possible experience in the
Transcendental Deduction of the Categories. For this means that nothing can be given in experi
ence, and so exist, without conforming to the rules represented in them (ch 18-D).
304 Kant's Anatomy of the Intelligent Mind
To be certain of their accord, one must observe that ...[t]he same procedure
of the understanding when it represents the sphere of a divided concept is
also observed when it thinks a thing as divisible and, as the terms of the divi
sion in the former exclude one another and y et are combined in one sphere,
the understanding represents the parts of the latter as terms whose existence
(as substances) also belongs to each exclusive of the rest yet as combined in
one whole. (Bl 12-3)
The problem that seems to have concerned Kant is the extent of the disanology
between a disjunction of judgments and a community of things, starting with the
fact that things can be conceived to belong to a community without needing to be
thought of as specifications of the same concept, much less as concepts with non
overlapping spheres that, taken together, coincide with the sphere of the undivided
concept. Indeed, it is by no means evident that things existing in community need to
be thought of as concepts (universals) in any sense at all, in which case they would
have no spheres. Clearly, Kant was right to be concerned that his readers would find
the derivation of the category of community from the logical function of disjunc
tive judgment far from self-evident.
Nevertheless, the category of community, when properly explicated, proves to
have enough in common with the logical function of disjunctive judgment to sup
port Kant's claim that the former derives from the latter. This becomes clear when
one recalls that community, like the third category of each of the other categorial
headings (Bl 10-1; CJ 197n), results from the combination of the other two catego
ries of relation, substance and accident and cause and effect. A substance, in the
pure concept of the understanding, is that which can be conceived only as subject,
never predicate, and otherwise has no content. This means not only that the cat
egory essentially involves universality (even if only the degenerate kind, when its
logical quantity is singular: section A and ch 10-B), but also that no substance can
be subordinate to any other substance in a subject-predicate relation and so none
can ever be the accident of another.To this extent, then, a community of substances
agrees with the logical community of a disjunctive relation in which the members
"are represented as coordinated with, not subordinate to, one another" because
"one cannot be contained under another" (Bl 12).
The coordinate status of substances coexisting in a community extends to their
causal relations as well. A substance could be represented as causally subordinate
to another substance only if the existence of one could be conceived to cause the
existence of another. That this is impossible for pure understanding is, however,
clear from a consideration of the logical functions from which the relevant catego
ries derive their sole and entire content:
Thus, just as disjunctive relations require of their constituents, substances are coor
dinate, not subordinate, not only as bearers of accidents but in their causal relations
as well.
A second property of disjunctive judgments is that the non-overlapping spheres
of the disjuncts jointly constitute a whole coincident with the sphere of the undi
vided concept. Kant argued that a community of substances exhibits this feature as
well insofar as they causally interact:
This is a completely different kind of connection than the one met with in the
mere relation of cause to effect (ground to consequence), in which the conse
quence does not reciprocally determine the ground and so does not constitute
with this other a whole (as the creator does not with the world). (B112)
Despite the obvious difference that the substances which constitute a whole (com
munity) through their interaction do not need to be non-overlapping specifications
of a single concept, Kant's claim that the category of community and the logical
function of disjunctive judgment nevertheless coincide has more in its favor than
one might at first suspect. Substances, even if conceived as universals (Section
A), can never have overlapping spheres for the simple reason that one can never
be thought to subordinate any other substance as its predicate. Each accordingly
demarcates its own sphere, exclusive of every other, which, when combined with the
spheres of the substances with which it is in direct or indirect causal interrelation,
coincides perfectly with the whole constituted by their interrelation.
Nor are the "spheres" of substances and their "exclusivity" only superficially
akin to the exclusive spheres of the concepts from which the judgments related in
a disjunctive judgment are formed. The category of substance applies to concepts,
determining whether they must always be subjects and never predicates of judg
ments. Consequently, substances, whether conceived as universals or individuals,
must, in the first instance, be thought of as concepts and only secondarily, by means
of these concepts, as the objects (if any) represented by them (chs 13-H and 15-B).
From this (the general logical) point of view, an individual substance is simply an
306 Kant's Anatomy of the Intelligent Mind
unspecifiable concept, that is, a universal with a sphere that, by its very nature, is
empty. Since the category of substance has no source but pure general logic from
which to draw its content, substances, even when conceived as individuals, have
spheres; and these spheres are necessarily non-overlapping because, in order for
one substance-concept to subsume another within its sphere, either particularly or
universally, one of them would have to be thinkable in the logical role of predicate,
and this is precisely what the category of substance precludes. For the same reason,
spheres of substance-concepts exclude one another in precisely the same sense the
concepts formed by logical division in a disjunctive relation do. The latter, whether
in the subject or predicate position of the judgments in a disjunction, exclude the
spheres of the other concepts in such a way that the affirmation of any one disjunct
implies the negation of all the others. Similarly, in a community of substances,
the application of the category of existence to any affirmation of an accident of a
substance ipso facto implies the nonexistence of that selfsame determination in any
other substance. It could be otherwise only if one substance could be conceived to
subsume another within its sphere (in which case the accident would exist in both).
Since the subsumption of substances is precluded by the meaning conferred on the
category by the logical function of categorical judgment, one can immediately infer
from the existence of a determination in any one substance its nonexistence in all
others.
In a Kantian disjunctive judgment, the negation of one of the disjuncts implies
the affirmation of one of the others, but only if the judgment containing the undi
vided concept is itself affirmed. Can one similarly infer the existence of a determina
tion in one of the other substances in a community from its nonexistence in one?
One can, but only if that determination has already been posited as existing in that
community. For if it does not exist in the sphere of one substance but does exist
somewhere within its community of substances, then, since no determination can
have its existence in (modify) more than one substance, it follows that the determi
nation must fall within the sphere of some other substance within the same com
munity. Indeed, the inferential structure in this and every other respect not only
parallels that distinctive of disjunctive judgment but is identical with it. And for
Kant, that is all that needs to be recognized in order to demonstrate the origin of
the category of community in this logical function.
The one feature of the category of community that remains to be explained is
the irreversibility concerned in it: how does the logical function of disjunctive judg
ment leave us free to consider its judgments in different ways, and how does the
category annul it? As with all of the other categories save substance and quantity,
Kant seems never to have spelled out the answer and left few if any clues as to his
view in this regard. Still, in all likelihood, the freedom provided by the logical func
tion of disjunctive judgment that is annulled by the category of community is the
freedom to consider a substance as existing inside or outside a community. To apply
the category of community to a substance is to determine it so that it must always
and only be thought as a member of a community and, in particular, as always and
The Metaphysical Deduction of the Categories 307
only causally interacting with the other members of that community (moreover,
coupled with the category of possibility-impossibility, it not only determines sub
stances as belonging to a community but to one and only one community). That
is the only extra-logical sense of the category, even though, like the contents of all
the categories, it derives simply and solely from the corresponding logical func
tion, with no sensible or other extra-intellectual contribution. And it is here that
we find the one exception to the rule that all necessity is hypothetical: insofar as
substances are determined as members of a community, it follows immediately and
15 The exception described here is simply the categorial counterpart to the exception discussed
in connection with disjunctive inference in Chapter I 0-D.
16 Also: "through the relation of the manifold to the unity of apperception there arise
( werden)
concepts that belong (angehoren) to the understanding but can only come into being by means of
the imagination in relation to sensible intuition" (Al24) and "without intuition, no object occurs
in respect to which the logical function could be determined as a category" (MFPNS 474n).
308 Kant's Anatomy of the Intelligent Mind
content between the object and what is thought in its concept is essential. For how
can thoughts have even so much as the potential to be true of objects outside the
understanding if none of their contents are in any way (formally or materially )
homogeneous with those objects?
Kant saw no difficulty where the objectivity of empirical concepts is concerned,
because their contents derive directly from what appears in sense perception.
Similarly, the objective credentials of mathematical concepts, including arithmeti
cal and algebraic concepts, are impeccable from a Kantian standpoint because all
directly incorporate or otherwise depend on juxtaposition, succession, and the for
mal unity of the manifold in space and time (ch 6). By contrast, pure concepts of
the understanding draw their content exclusively from logical functions of judg
ment and so are completely incommensurable with the sensible and everything else
outside discursive understanding (i.e., everything not incorporating or founded on
the analytic unity of apperception and the logical conditions for the relation of
different representations insofar as that unity attaches to them): "pure concepts
of the understanding, in comparison with empirical intuitions (indeed sensible
intuitions in general), are completely heterogeneous and can never be met with in
any intuition . . . since no one will say that, e.g. causality, can be intuited through
sense and be contained in appearance" (A137-8/Bl 76-7). The "objects" thought
in the categories, if these concepts may be said to have any object at all, are simply
concepts and judgments considered solely with regard to their logical positions in
judgments and inferences. Insofar as the categories relate to anything outside the
understanding, it is only indirectly, through the concepts and judgments whose logi
cal positions they fix. If they are to have objective significance in their own right, it
would have to derive entirely from this fixing of positions. But how can objectivity,
in any proper sense, possibly be supposed to attach to the purely formal, objectively
empty ritual of rendering the logical places of discursive representations in judg
ment irreversible?
This is what I call the heterogeneity problem, the solution to which will be
expounded in detail in Chapter 15. Here, I simply want to focus on the question
and the general structure of Kant's response. If one seeks to evade such questions
as introducing inessential complications that are best avoided, it should be remem
bered that Kant's overriding objective in the Critique of Pure Reason was to estab
lish "the thesis that the whole speculative use of our reason never reaches beyond
objects of possible experience" (MFPNS 474n). His proof rests on the twin pillars of
his "theses concerning the sensible character (Sinnlichkeit) of all our intuition and
the sufficiency of the table of the categories as determinations of our conscious
ness borrowed from the logical functions in judgments in general." Demonstrating
the second-the task he set himself in the MDC-required Kant to negotiate a
path between the dogmatic intellectualism of those who supposed intellect to be
in its own right (independently of sensibility ) a source of objective content and
Hume's skepticism regarding the very existence of pure concepts of the understand
ing. Kant believed he could do this with his proof that the categories derive their
The Metaphysical Deduction of the Categories 309
sole and entire content from the logical functions of judgment, for it implies both
that, pace Hume, pure concepts of the understanding do indeed exist and that, pace
intellectualism, they lack all intrinsic objective sense and signification. The latter
they can acquire by means of pure sensibility, but only with respect to objects of
the senses: "the categories are, in their own right, nothing but logical functions,
and as such do not constitute the least concept of an object in itself but require a
foundation in sensible intuition, whereupon they serve only to determine empirical
judgments with respect to the logical functions, which otherwise are undetermined
and indifferent with respect to all these functions, thereby conferring universal
validity on them and making possible judgments of experience in general by their
means" (PFM 324). Kant's overriding objective of showing that pure reason can
never extend beyond objects of possible experience thus depends on showing that
the senses are capable of conferring objective worth on concepts that intrinsically
are devoid of it. But how can one hope to do the latter without also answering "the
follow-on question how the application of the category to the form of intuition is
possible, since categories and form of intuition are heterogeneous" (AA 18 § 6359
[1796-8])? Clearly, the question of how representations so completely heteroge
neous as pure concepts of the understanding, whose sole and entire content derives
from purely discursive logical functions of judgment, and pure intuitions of sensi
bility, completely devoid of both discursive form and content (Part 11), can possibly
combine to y ield pure concepts of objects, far from being ancillary, goes to the very
heart of Kant's critical project.
The solution to the heterogeneity problem is by no means obvious. In my
view, the key to its discovery lies in recognizing how pure concepts of the under
standing considered in isolation differ from these same concepts considered
in relation to sensibility or other nondiscursive faculties such as the faculty of
desire (Begehrungsvermogen). In the previous sections of this chapter, I treated
them exclusively from the first perspective, with an ey e solely to their intrinsic
logical-function-derived content. So regarded, pure concepts of the understanding
are the same in whatever relation they are considered, contributing identical con
tents to the theoretical categories of the first Critique and the practical categories of
the second Critique; and if the human mind were endowed with other nondiscursive
faculties with a priori foundations analogous to those of sensibility and the faculty
of desire, these same contents would be common to the categories of those faculties
as well. Different sets of categories rooted in these same pure concepts emerge only
when the various a priori determinative roles of pure understanding in relation to
nondiscursive faculties are factored in. This happens not through the addition of a
new source of content, for any accretion there would result not in purely intellectual
categories but rather (if the source is sensibility) in schemata ("the sensible concept
of an object in agreement with the category," Al46/B186), types (if the source is
the faculty of desire: CPrR 67-9), or something comparable (for nondiscursive a
priori determinable faculties that humans lack but other creatures with sensibly
conditioned understandings might have). Instead, each set of categories develops
310 Kant's Anatomy of the Intelligent Mind
under the guidance of the principle that underlies the determinative relation of pure
understanding to the relevant faculty: the original synthetic unity of the manifold
in the case of sensibility (Bl31, A401), and the moral law, together with the concept
of positive freedom implicit in it, in the case of the faculty of desire (CPrR 65-6).
Without altering the content of pure concepts, these principles confer on them a
new worth. Thus, in connection with the first Critique principle of the synthetic
unity of the manifold, the result is a table of categories each of which has the value
of a pure concept of an object in general (A79/B105).
The significance of this principle is evident in the portion of section§ 10 leading
up to the table of categories (A76-80/Bl02-5). It opens by contrasting transcen
dental logic, which requires a suitable content if its concepts are not to be empty,
with general logic, which abstracts from all content. Since the concepts dealt with in
a transcendental logic are pure, it requires an a priori source of content. This is why
transcendental logic is possible only in the wake of transcendental aesthetic and its
revolutionary thesis that the senses are equipped to intuit objects entirely a priori
(cf. Al5-6/B29-30 and A55/B79-80). But before concepts can be extracted from
the manifold of pure intuition, it must "first be gone through in a certain manner,
taken up, and combined, in order to make a cognition from it," that is, synthesized;
and, in the case of the concepts at the focus of transcendental logic, this synthesis
must be pure.
Synthesis is in the first instance
synthesizing the manifold to the exclusion of all others (ch 13-C). Accordingly, the
concept of a decade constitutes a "common ground of unity" because it imposes
a unique ordering on the manifold, one that excludes all others and so counts as
a necessary unity of consciousness (ch 13-G). In the same way, a pure concept of
the understanding functions as a common ground for the unity of synthesis in
the imagination by necessitating one among the many possible ways of synthesiz
ing a manifold to the exclusion of all others ("The concepts that give unity to this
pure synthesis ...consist simply in the representation of this necessary synthetic
unity ...and rest on the understanding"; ch 13-H). Since the sole and entire content
of each of these concepts is the corresponding logical function of judgment and
consists solely in irreversibly fixing the logical positions of concepts or judgments
in respect to these logical functions (Sections A and B), the only necessity they
can confer on the synthesis in imagination of the a priori manifold of sense in its
synopsis must stem from their determination of concepts and judgments in this
manner. Thus, the same purely intellectual determinants that confer analytic unity
on concepts in categorical judgments (as when a concept acquires the identity of a
substance by being determined as always subject and never predicate) and on judg
ments in hypothetical and disjunctive judgments also confer a necessary synthetic
unity on the manifold in intuition and thereby acquire transcendental content as
pure concepts of objects of intuition in general:
The objectivity Kant claims for the categories as determinants of the synthe
sis of intuition here, subsequently termed their synthesis intellectualis (Bl50-l, ch
14-B), leads straight back to the heterogeneity problem: how can concepts formed
from logical functions, with no sensible or other nonlogical content and determina
tively applicable to only to representations that have already been made universal
(concepts and judgments), possibly be supposed to apply to objects of the senses
given prediscursively in pure or empirical intuition? Kant's general answer is that
the categories introduce necessary synthetic unity into the manifold of sensibility
a priori. But how is this possible? Because empirical concepts derive their content
directly from sensible intuition (by comparison, reflection, and abstraction: ch 9), it
is easy to understand how they can impose a necessary unity on the manifold of a
posteriori intuition by determining its synthesis so as to exhibit (reproduce) objects
in intuition corresponding to them (ch 13-C). The same is true of mathematical
312 Kant's Anatomy of the Intelligent Mind
Logical functions are only forms for the relation of concepts in thinking.
Categories are concepts whereby certain intuitions are determined in regard
to the synthetic unity of their consciousness as contained under these func
tions; e.g. what must be thought as subject and not as predicate. (AA 23, E
XLII, p. 24).
Here, Kant makes clear that the means whereby the categories determine intuitions
in regard to the synthetic unity of their consciousness (apperception), and thereby
take on the status of pure concepts of objects of sensible intuition in general, is the
determination of the empirical and mathematical concepts in which the manifold of
these intuitions and its synthesis are represented universally as, e.g., always subject
and never predicate or as always predicate in relation to concepts so determined.
That this should be so is not surprising. As noted earlier, Kant's explication of the
categories as pure concepts of the understanding whose sole and entire content
derives from the logical functions of judgment limits their role in representation
exclusively to the determination of given universals (concepts or judgments) so as
to irreversibly fix their logical place in judgments. Their status as pure concepts of
objects must therefore be attained in this manner or not at all. The question is how.
The MDC makes clear what Kant's general strategy for connecting the sole
mode of determination of which the categories are capable with objectivity is: if
objectivity can be resolved into the addition of necessity to the synthesis in imagi
nation of the manifold given in intuition, thereby transforming the question of
objectivity into the question of the possibility of the necessary synthetic unity of
the manifold in one consciousness, then to explicate and validate the status of the
categories as pure concepts of an object of sensible intuition in general, it would
suffice to show that pure concepts of the understanding are essential to all necessary
synthetic unity of the manifold. Insofar as each category is shown to be a ground
of the necessary synthetic unity of the manifold, Kant would thereby have demon
strated the possibility of pure concepts of objects of precisely the sort that Hume's
The Metaphysical Deduction of the Categories 313
skepticism denies. For the case Hume made to show that concepts of the necessary
relation between the distinct are inherently paradoxical and therefore impossible
(UU ch 17-A) would be refuted by proof that the categories, as "pure concepts of
the understanding that relate to objects a priori" (A79/B105), are universal a priori
representations of the necessary synthesis (relation) of the manifold (the distinct).
T he way would then be opened for restoring to metaphysics the pure concepts of
objects that Hume rejected (EHU XII/iii, especially if 134) and proposed to replace
with associationist "bastards of imagination" (PFM 257-8). However, since Kant
reserved for the Transcendental Deduction of the Categories the execution of the
strategy laid down in the MDC, I turn now to it .
This page intentionally left blank
{PART IV}
B138
This page intentionally left blank
The Transcendental Deduction of the Categories is Kant's demonstration that pure
concepts of the understanding have validity as pure concepts of an object of sen
sible intuition in general. Since their objective validity is inferred from a proof that
the categories are conditions of cognitive experience, the Deduction at the same
time restricts their objective employment to objects of experience. For it means
that any object that cannot present itself to the senses in intuition, and so cannot
be given in a possible experience, cannot be cognized through the categories. Since
Kant's transcendental idealism implies that things in themselves are precisely such
objects (chs 7-8), and since the doctrine of pure sensibility (chs 3-6) that yields
this idealism is premised in the proof that the categories are conditions of possible
experience, the Deduction thus has the consequence that no judgment involving the
categories can hold of things in themselves.
Nor is that all. Since the proof that the categories are conditions of possible
experience depends on showing them to be essential components of any judgment
purporting to relate to objects, it implies both that things in themselves are com
pletely beyond our cognitive ken and that the objects cognized in experience which,
in their interconnected totality, constitute the natural world, are never anything
other than our own representations. Thus, while it "is sure to sound very absurd and
strange" (Al 14), the objective validity accorded by the Transcendental Deduction
to the categories as conditions of the possibility of experience and its objects is
therefore at the same time psychologistic proof of the ineluctable subjectivity of
nature itself.
In this part of the book, I shall focus on the notion at the heart of the
Transcendental Deduction, the objective unity of apperception, and show, among
other things, how it enabled Kant to explicate the possibility of the objective thought
determinations whereby the understanding is legislative in respect to nature itself
(the source of nature's most fundamental laws, detailed in Part V). The broad out
lines of his account may already be evident from anticipations in earlier chapters.
The first step is the proof that the necessary synthetic unity of apperception is the
principle of the objective unity of the manifold of sensible intuition and, as such,
the highest principle of cognitive understanding. Next, the categories attain the
status of conditions of the possibility of experience and its objects by proving them
to be the sole means whereby this necessary synthetic unity of apperception can
be introduced into the sensible manifold. Since pure space and time are prediscur
sive embodiments of the same synthetic unity, it further becomes clear how purely
discursive categories can relate to the sensible: as representations not of space and
time per se, but of the original synthetic unity of a manifold of a sensible intuition
in general (synthesis intellectualis), including theirs. In particular, when represented
(in synthetic a priori judgments) as principles determinative of all possible concepts
of intuitions in space and time, the categories define a transcendental figurative 317
318 Kant's Anatomy of the Intelligent Mind
limits is particularly important in the case of the categories, for, as concepts with
no sensible content, they inevitably give the impression of having the potential for
supersensible employment (a potential putatively realized in traditional metaphys
ics, including Kant's own Inaugural Dissertation of 1770: ch 2). And insofar as this
apparent potential carries over to any a priori sensible concept with which they
may be combined, it creates the need for transcendental deductions of those con
cepts as well, such as those provided for pure space and time in the Transcendental
Aesthetic (A87-8/Bl 19-21).
From the time of Hermann Cohen and his Neo-Kantian successors to the analytic
neo-Kantians of recent times, the opening paragraphs of§ 13, "The Principles of
any Transcendental Deduction," have been regarded as proof that Kant held that
an inquiry cannot be both a priori and psychological:
Kant presumably had in mind the kind of cases of disputed title that were dealt
with by the legal authorities of his day. Thus, the quidfacti of one's birth (who one's
parents are, one's date of birth, etc.) is incapable by itself of settling a disputed
line of succession to a duchy. Only in conjunction with societal norms, whether
in the form of established law or unwritten custom, can biological facts prove or
Interpreting the Transcendental Deduction of the Categories 321
disprove title and thereby settle the question, quid Juris? Similarly, many interpret
ers take Kant to be saying that the birth of the categories in pure understanding
is incapable by itself of settling the issue of their application to objects. For, by
contrast with empirical concepts, neither empirical nor transcendental objects enter
into the account of their origin as pure concepts of the understanding. To prove
their title to objective employment, therefore, it is just as necessary to invoke logical
and epistemic norms over and above psychological fact as it is to invoke societal
norms over and above biological fact in order to establish title to a duchy. By this
sort of reasoning, then, interpreters frequently invoke A84-5/Bl 16-7 as proof that
psychology is inessential to Kant's project in the Transcendental Deduction of the
Categories (TDC).
Such interpreters draw a more general lesson as well regarding the meaning of
Kant's term "transcendental." The question quid Juris? can only be posed if the
matter in philosophical dispute remains unsettled even after all issues of fact have
been settled. Since this is not the case with an empirically deducible concept, they
take the fact that matters are otherwise with the categories to be precisely what
justifies terming their deduction "transcendental" rather than "empirical." In other
words, given that facts must be established empirically or not at all, any question
that is not a quid facti is ipso facto non-empirical; and since "non-empirical" in rela
tion to the objective validity of concepts originating entirely a priori qualifies as a
transcendental question in Kant's sense (A87/Bl 19-20), these interpreters tend to
regard A84-5/B116-7 as evidence of the normative character of the transcendental
generally. And indeed, if they are right, it obliges one to take seriously the claim
that the transcendental deduction must, in the final analysis, be understood as an
exercise not in psychology but normative epistemology.1
I remain unconvinced. In assessing whether the normativist reading A84-5/
B116-7 truly is superior to a psychologistic one, the first thing one needs to con
sider is whether, with the admission of pure intuition and its manifold, the field
of psychological fact remains limited to the a posteriori (ch 1). Accounts of the
origin of a priori representations in pure sensibility and other a priori faculties of
the mind seem to me to have as great a claim, from a Kantian perspective, to being
matters of fact as accounts of the origin of a posteriori representations in empirical
sensibility and other faculties. What distinguishes them from a posteriori matters
of fact is that they rely not on particular perceptions (consciousness of the contents
of particular sensations: ch 3-A), but solely on data of indeterminate perceptions
given to thought alone plus the fact of affection (chs 4-C, 8-E and -F). The quid
Juris/quid facti issue consequently turns on whether Kant meant to exclude psy
chological facts generally or merely the empirical sort. Thus, in order to establish
their construal of the significance of Kant's characterization of the transcendental
1 For a consideration of some of the relevant literature, particularly the much-discussed paper
by Dieter Henrich, see my two-part article "Kant's Psychologism," published in the 1999 and
2000 editions of Kantian Review.
322 Kant's Anatomy of the Intelligent Mind
Objects for concepts whose objective reality can be proved-be it through pure
reason or through experience (and in the former case, from theoretical or practi
cal data of reason), but in all cases by means of an intuition corresponding to
them-are facts (Tatsachen) (resfacti).* These include the mathematical prop
erties of magnitudes (in geometry) because they are capable of a priori exhibi
tion for the theoretical employment of reason. Facts further include things or
properties of things that can be established (dargetan) by means of experience
(one's own or that of others via testimony).-Particularly noteworthy, however,
is that an idea of reason (which is in itself incapable of exhibition in intuition
and so too of any theoretical proof of its possibility) is found among the facts,
and that is the idea of freedom whose reality, as a particular kind of causality
(the concept of which would be a transgression from a theoretical perspective),
is proven by practical laws of pure reason and, in conformity to these, in actual
actions, hence, in experience.-The sole idea among all the ideas of pure reason
whose object is a fact and must be reckoned among knowable matters (scibilia).
(CJ 468)
*I here extend, as seems warranted, the concept of a fact beyond the usual
meaning of this word. For it is not necessary, or even feasible, to restrict this
expression merely to factual experience if the relation of things to our cogni
tive capacity is at issue since a merely possible experience is already sufficient
for speaking of them merely as objects of a determinate mode of cognition.
(CJ 468n)
Anything a mathematician can construct a priori counts as a fact for Kant (see also
Progress 323) as does the transcendental idea of freedom in its relation to a pure
will. Indeed, as the footnote to CJ 468 makes clear, Kant treated as factual every
synthetic a priori proposition demonstrable through "a merely possible experience,"
which presumably includes everything in the proof in the TDC that the categories are
conditions of possible experience. Given that Kant also believed that he could vali
date synthetic a priori propositions regarding the constitution and workings of the
mind insofar as it contributes intuitions and concepts essential to the possibility of
experience and its objects completely a priori, is there any reason to think that he had
any principled opposition to countenancing the admission of a priori (as opposed to
empirical) psychological facts into transcendental philosophy? Quite the contrary, it
seems to me.
Normativist interpreters might still deny that the TDC, as distinct from other parts
of transcendental philosophy, is psychological simply because Kant characterized its
question as a quid Juris rather than a quid facti. But how much explanatory weight
can a legal analogy offered in an introductory paragraph and never mentioned again
Interpreting the Transcendental Deduction of the Categories 323
can only result in nonsense ("we either contradict ourselves, or talk without a mean
ing," THN 267/173). Kant, we have seen, employed essentially the same psychologistic
reasoning in respect to space and time when he argued that once they are recognized
as being nothing more than mere modes of representation, it becomes as "patently
contradictory" to accord them or anything in them "a subsistent existence apart from
or prior to experience" as it is "to represent the actuality of experience apart from or
prior to experience" (PFM 341-2; chs 2-E and 7-C). Here he is drawing a manifestly
normative conclusion about the scope of objective meaning and validity of concepts of
space, time, and everything they contain that is grounded in his non-normative, a priori
psychological account of their origin in the mind as mere modes of representation (UU
ch 2-E-2). It should therefore be clear that Kant's legal analogy quite simply lacks suffi
cient force to warrant the conclusion that the TDC is not a psychological inquiry merely
because it yields the answer to a quidJuris rather than a quidfacti. 3
In approaching the question of what is and is not implied by Kant's characteriza
tion of the question addressed in the TDC as a quid Juris, one should bear in mind
that the TDC premises the MDC (metaphysical deduction of the categories). To the
extent that the MDC exhibits the categories as pure concepts of the understanding
whose sole and entire content consists in logical functions of judgment, a case can
be made that there is nothing inherently psychologistic about it. But insofar as it
explicates the categories as pure concepts of objects by specifying them as universal
representations of the determination of the pure synthesis in imagination of the a
priori manifold of sensibility conformably to logical functions of judgments (ch
11-C) , the MDC proves the categories to be constitutively bound up with sensibility
and imagination and is therefore ineluctably sensibilist and psychologistic. So, the
answer to whether the TDC is or is not an inherently psychological inquiry may be
2 The only other texts known to me where the legal terms quidjuris and qua;stio juris occur are
Progress 275 and AA 18 § 5636. They contribute little, in my view, to understanding what is said
in the Critique and are certainly not dispositive for the issue presently under discussion.
3 The fact of reason is perhaps the clearest illustration of Kant's readiness to ground the
normative on the psychological. In identifying this as the ground that alone can establish the
objective reality of the moral law, he based this most normative of propositions on something
as quintessentially factual and non-normative as the immediacy of one's own individual, private
consciousness (CPrR 31 and 46-7)-a fact, moreover, that may fairly be described as a priori
psychological (for, as shown in ch 3-A, being conditioned by the presence of sensation does not,
of itself, make a representation a posteriori).
324 Kant's Anatomy of the Intelligent Mind
supposed to turn on whether it is concerned with the categories merely as pure con
cepts of the understanding (i.e., the meaning explicated at Bl28-9) or with the cat
egories as pure concepts of objects (as explicated at A78-9/B 104-5). That it is the
latter, not the former, is something that no one should wish to contest. Even as early
as the A edition preface, Kant made clear that the primary purpose of the TDC is
to provide an objective transcendental deduction of the categories, which "relates
to the objects of pure understanding and is intended to expound and render intel
ligible the objective validity of its a priori concepts" (Axvi). Thus, once again, we
are driven to the conclusion that the TDC should be understood as an exercise
in a priori psychology directed to normative ends (and even in their supposedly
non-psychological, root meaning as pure concepts of the understanding, the cat
egories are still inherently psychological insofar as the logical functions from which
they derive are essentially bound up with the analytic unity of the I think: ch 10-E).
Indeed, one may well wonder whether there is really any reason to consider
the TDC normative at all, at least in the sense interpreters commonly suppose. Its
constant focus is consciousness and self-consciousness, imagination, and the syn
thesis in imagination of the manifold offered by sense. The a priori functions of
mind dealt with in it are systematically correlated to empirical operations that are
manifestly psychological in character. The natural conclusion, one would think,
is that the TDC is an a priori psychological inquiry (ch 1). The only obstacle is its
normativist-seeming characterization as a quid Juris. The question is how much
weight can be attached to what is, after all, merely an analogy. Issues of right
before courts of law are normative because they are concerned with facts only
within a framework of established conventions and institutions. But how much
of this socially grounded framework, if any, can be supposed to extend to the
Transcendental Deduction? Certainly, the admission of social concepts and expe
riences into the explanatory framework of the Transcendental Deduction would
fatally compromise Kant's claims to have relied exclusively on a priori proofs,
involving none but pure representations and in which no empirical evidence is
posited or premised. Yet with everything in any way reliant on anything originat
ing in norms of human social interaction precluded, it is by no means obvious
how the Deduction is supposed to be normative in a sense "like" a juridical quid
Juris. Anti-psychological interpreters may think they correctly grasp Kant's anal
ogy because they are wont to characterize the normative epistemology of their
own time as "a priori" and so may not see what else Kant could have meant.
But Kant philosophized centuries ago and said virtually nothing to help one
understand precisely what, beyond not being a matter of empirical fact, carries
over from the juridical to the transcendental context. So rather than risk reading
something into his analogy that may not be there, a better approach, it seems to
me, is to turn to the body of the TDC itself (Sections 2 and 3 in the A edition, §§
15-27 in the B edition) and determine from what actually transpires in it just how
anti-psychological its justification of our "title" to employ the categories really
is. And when this is done, as we shall see in succeeding chapters, it becomes clear
Interpreting the Transcendental Deduction of the Categories 325
of his philosophy can be detached from the chimera of a priori psychology those
elements that retain most value in relation to the present-which is usually deemed
to be the Transcendental Deduction understood as an attempt to establish a nor
mative framework for objective judgment by means of so-called "transcendental
arguments" (ch 7-B).
Entering into the question of whether the bifurcation between a priori phi
losophy and psychology is well founded or merely the parochial expression of
a currently reigning dogmatism would take me beyond the scope of the pres
ent work. Instead, I shall confine myself to remarking that applying this bifurca
tion to Kant's transcendental philosophy makes it all but impossible for one to
take even its most manifestly psychological parts seriously as a priori psychology.
Kant's doctrine of pure yet sensible intuition is a good example. If taken at face
value as an essay in a priori psychology, then the best anti-psychologically minded
interpreters can do for Kant is argue that pure sensible intuition, notwithstanding
his iterated insistence to the contrary, is inessential to the achievement of his a
priori philosophical objectives and thereafter factor it out wherever he factored
it in (i.e., virtually everywhere). If not taken at face value (if only on the ground
that a philosopher of Kant's caliber cannot have meant to embrace anything so
absurd as a priori psychology), then the way is open to domesticating it, as it were,
by assimilating it to a normative epistemological paradigm. Either way, a philo
sophical preconception, whether true or not, prevents one from understanding
Kant's theory of a priori sensibility as being just what he presented it as being: a
theory of sensibility.
Admittedly, the obvious, straightforward approach is by no means the easiest,
especially at this distance of time. Yet one should not be too quick to conclude
from the elusiveness of a doctrine that it is impenetrable, incoherent, or unim
portant, much less that its deficiencies, real or imagined, leave one free to recast
it as something altogether different that one does understand. If nothing else,
I would hope to have shown in Part II that it is possible to make sense of Kant's
theory of pure sensibility as both a continuation of earlier sensibilist psychology
and, as a priori, a radical break from it. Even if I am wholly or partly mistaken
in thinking that the raison d'etre of pure sensible intuition is to explain the unity
of sensibility, the need to come to grips with pure intuition on its own distinctly
psychological terms seems to me inescapable for anyone seriously concerned to
understand what Kant, the historical personage, thought and why. For no one
326 Kant's Anatomy of the Intelligent Mind
can deny that the concept of an a priori sensible manifold enters essentially into
Kant's characterizations of virtually every major component of the theory of
understanding he developed in the metaphysical and transcendental deductions
of the categories: the pure productive synthesis of this manifold in the imagi
nation and its specifically transcendental incarnation (transcendental synthesis
speciosa); apperception as the original synthetic unity of this manifold ahead
of all thought; pure concepts of the understanding as the basis for universal
representations of the determination of the pure synthesis of this manifold in
imagination in conformity to the logical functions of judgment (pure concepts
of an object of sensible intuition in general); and so on. How can one hope to
come to grips with any of these notions without first expending the time and
effort necessary to make sense of the a priori psychological elements of Kant's
philosophy on their own terms, as a response to the challenges bequeathed by
the most original and important psychologists of his age, Hume above all? We do
neither Kant nor ourselves any favor by restricting ourselves to articulating the
views we think he should have held before we have succeeded in fully understand
ing those he actually did.
What then of quid Juris? Since the categories (in the Bl28-9 sense) have no
sensible or other extra-logical content, they relate, potentially at least, to sen
sible and nonsensible objects alike ("since they do not speak of objects through
predicates of intuition and sensibility but of pure a priori thought, they relate
to objects universally, apart from any conditions of sensibility," A88/B 120).
T here is therefore reason to query whether these concepts can legitimately be
applied to all of the objects to which they seem to have the potential to apply.
It is a quid Juris because it is a demand to justify our right to apply concepts
whose content consists simply and solely of logical forms of judgment to
either or both of the two types of objects the Transcendental Aesthetic enti
tles us to distinguish: objects as they appear in intuition and objects as existing
mind-independently in themselves.
To deal with this objective question, Kant deemed it necessary to sink it into a
different, subjective question: to explain how the understanding in which the cat
egories originate itself is possible. Which faculties are requisite merely in order to
think Uudge) at all? Which are requisite to form concepts and judgments that enable
us to think objects a priori? Do these faculties suffice not only to think objects but
to cognize them as well and, if so, do they give our thought a priori relation merely
to sensible objects or to supersensible ones as well? T he normative result of this
inquiry (the TDC) is both the a priori justification of the application of our a priori
concepts of objects to sensible objects and proof that they are never applicable to
supersensible ones. But the thing that all interpreters of Kant should acknowledge
is that the inquiry that vindicates our title to apply the categories to sensible objects
and no others is, at its core, fully as subjective and psychological as the analyses of
understanding offered by Locke and Hume. And it is to this, the subjective side of
the TDC, that I now turn.
Interpreting the Transcendental Deduction of the Categories 327
In the preface to the A edition of the Critique of Pure Reason (1781), Kant dis
tinguished two strands in the inquiry pursued in the TDC:
made it his business to explain in the STDC is the result established in the OTDC
so that the OTDC's conclusion is the STDC's explanandum.4
4 A reader of the manuscript version of this book wondered whether "Waxman has not over
looked a principal difference between the two arguments which would undermine his claim of a
continuity in method between them. The subjective deduction is 'a search for the cause of a given
effect,' [Axvii] and so presumes the validity of the concept of cause and effect in its investigation
of human psychology, and it is precisely because it makes use of this concept (or its predicable,
'power') with respect to inner appearances that it is psychological (though, because the con
cept of causality does not admit genuine application to inner appearances, the results are merely
'hypothetical' as Kant stresses [at Axvii]). By contrast, the objective deduction makes no such
presumption of the legitimacy of this concept but expressly seeks to justify its use. Accordingly,
Waxman's normativist opponent might claim there is a basis to draw a clear distinction between
the psychological method employed by the subjective deduction, and the properly normative
enterprise that is the objective deduction." This is an ingenious reading that would pose a serious
problem for my interpretation if it did not depend on several errors that, when corrected, render
it untenable. First, the OTDC premises both the Transcendental Aesthetic and MDC and so is
thoroughgoingly psychological in its own right, with or without the STDC, and so is a priori
psychological in a sense that does not presuppose cause and effect or any of the other categories.
Second, Kant did not say that the subjective deduction is a search for the cause of a given effect,
much less explicitly invoke the category of cause and effect. His insertion of gleichsam, which
translates as "so to speak" or "as it were,'' prevents us from taking it as a straight assertion, but
instead obliges us to construe his words as saying that the subjective deduction bears a certain
(unspecified) resemblance to a search for the cause of a given effect. And there is an obvious
nonresemblance that my critic ignores: a cause explains a given actuality, whereas the STDC is
concerned not with the actuality of pure understanding, but with its possibility. Its possibility is
like(" the same as) a given effect insofar as it is given with the OTDC result that pure concepts of
the understanding that are valid a priori of objects are necessary conditions of possible experi
ence, and the search for "the cognitive faculties on which [pure understanding] rests" is like ("
the same as) the search for its cause except, of course, that these faculties (pure sensibility and
productive imagination are presumably meant) do not cause pure understanding to exist, but are
merely conditions of its possibility. There is therefore no evidence to support, and strong reason
to reject, the theses that the STDC "presumes the validity of the concept of cause and effect"
and that it is "because it makes use of this concept ... that it is psychological." The third error my
critic makes is taking Kant to have asserted that the subjective deduction is merely hypothetical.
For although its resemblance to a search for the cause of a given effect might make it seem hypo
thetical, Kant made quite clear that this semblance is false: "as I will show elsewhere, it is not in
fact so." Kant clearly intended to make good on his promise not in the preface, where such con
siderations have no place, but in the TDC itself. The upshot is that however like a search for the
cause of a given effect the STDC may seem, it is not one, hence does not take the validity of the
category of cause and effect for granted, and so is not in the least hypothetical but fully as certain
(albeit not as essential) as the OTDC. That certainty, we shall see, derives from the a priori psy
chological principle at the heart of the STDC: that all consciousness has a necessary relation to
original apperception (A l l 7n-restated in the B edition as the thesis that the I think must be able
to accompany all my representations: B 131 ). Fourth, my critic evidently supposes that the STDC
is concerned with empirical psychology (cause and effect and the other categories can apply only
to the empirical, not the a priori psyche: A346/B404, A402, and B422-3). But again, there is no
reason to think this is true and every reason to think it false: it is stated explicitly at Axvi-xvii that
the STDC is a study of pure understanding, not empirical, and so, by implication, "the cognitive
faculties on which pure understanding rests" (i.e., that explain its possibility) must likewise be
pure, not empirical. Finally, the supposition that Kant did not take the category of cause and
effect to "genuinely apply" to inner appearances has no textual basis at Axvi-xvii, where "inner
appearances" are neither mentioned nor alluded to (indeed, they are not mentioned or discussed
Interpreting the Transcendental Deduction of the Categories 329
The OTDC's conclusion and the reason Kant deemed it essential to the Critique's
overarching purpose of determining what and how much understanding and rea
son can cognize a priori, independently of all experience, is made clear in the pas
sage referenced at Axvii:
[T]here are two conditions under which alone cognition of an object is possi
ble: first, an intuition, whereby the object is given but only as appearance, and,
second, a concept corresponding to this intuition through which the object
is thought. However, it is clear from the [the Transcendental Aesthetic] that
the first condition, the one under which alone objects can be intuited, in fact
underlies the objects, regarding their form, a priori in the mind. All appear
ances therefore necessarily agree with this formal condition of sensibility
because it is only by means of this condition that they are able to appear,
that is, be empirically intuited and given. The question now arises whether
concepts do not also precede a priori as conditions under which alone some
thing can be, though not intuited, yet thought as an object in general, since,
in that case, all empirical cognition of objects is necessarily in conformity to
such concepts because, apart from their presupposition, nothing is possible
as object of experience. But now all experience contains, besides the intuition
of the senses whereby something is given, a concept of an object given in
intuition, or which appears, with the consequence that concepts will underlie
objects in general as a priori conditions of experiential cognition. The objec
tive validity of the categories as a priori concepts will consequently rest on
this: through them alone is experience (according to the form of thought)
possible. For they would then relate necessarily and a priori to objects of
experience, because only by their means can any object of experience what
soever be thought.
The transcendental deduction of all concepts a priori thus has a principle
on which the entire investigation must be grounded, namely: that they must
be cognized as a priori conditions of the possibility of experience. (A92-4)
The possibility of experience is the principle under which alone the categories can
relate to objects a priori and so too the keystone of their objective transcendental
deduction. Since this principle suffices only for cognitive experience of objects inso
far as they do or can present themselves to the senses and so conform to the formal
in the A preface at all). Nor does it stand up to scrutiny in relation to the Critique as a whole. For
although Kant did hold that outer as well as inner intuition is needed to demonstrate the objec
tive reality of things in conformity of the categories (B291), nowhere did he suggest, much less
assert, that outer intuition by itself, without inner, suffices therefor. Indeed, if cause and effect or
any other category did not apply to all inner appearances without exception (including spatially
unintuitable appearances such as those that exhibit emotions, desires, and psychological activity),
it would contradict Kant's iterated insistence that the categories are valid of appearances a priori
and so universally as well as necessarily (it is precisely because they must be shown to genuinely
apply to all appearances that time and the inner, not space and the outer, were his focus in the
Analytic of Principles: A l55/Bl94).
330 Kant's Anatomy of the Intelligent Mind
conditions of sensible intuition, the OTDC at the same time establishes that these
concepts have no objective meaning or validity whatsoever in respect to objects as
they are in themselves, independently of their appearance in sensible intuition. And
in thus setting bounds to the cognitive employment of pure understanding and pure
reason, the OTDC does indeed prove to be essential to Kant's overarching purpose
in the Critique.
Since the STDC is simply an explanation of the result established independently
in the OTDC, it cannot be said to be essential in the same way. Does this mean, as
interpreters of a normativist bent sometimes maintain, that Kant was in a position
to disavow the subjective, psychological side of his inquiry without compromising
his purpose? Not at all. Even if one takes the view that rejecting the STDC does
not compromise the objectivity of the result of the OTDC, the OTDC is hardly
less a priori psychological than its STDC explanation. T his seems clear from its
presupposition of the doctrine at the foundation of Kant's a priori psychology: the
conformity of appearances to pure sensible intuitions. Without this a priori ground,
it would be impossible to represent appearances a priori at all and so, a fortiori,
impossible to represent their conformity to the categories a priori as well. T he
OTDC is also beholden to Kant's a priori psychological theory of sensibility for the
pure concepts of the understanding themselves since their metaphysical deduction
as pure concepts of objects would be impossible if there were no pure synthesis in
imagination of the pure synthesis of the manifold of sense for the logical functions
to determine (ch 11-C). Indeed, the only evident difference between the metaphysi
cal deduction and the OTDC is that the latter serves to extend the result of the for
mer from the pure manifold of intuition and its pure synthesis (A78-9/B 104-5) to
the manifold of appearances given in empirical intuition and its empirical synthesis
(A92-4). So, rather than freeing Kant's transcendental philosophy of any taint of a
priori psychology, the only consequence of discarding the STDC would be to leave
the psychology premised in the OTDC unexplained.
To understand why Kant feared that the lack of an explanation would leave
the OTDC dependent on a version of preestablished harmony, one first needs to
determine more precisely where the OTDC leaves off and the STDC begins. Kant's
statement in the A preface that what is said on A92-3 (cited earlier) "may alone
be sufficient" to give the OTDC "its full strength" is certainly suggestive but not
by itself conclusive. Taken at face value, it says that if one is content to accept its
result without explanation, the principle of the possibility of experience articulated
at A92-4 already is the OTDC, that is, sufficient warrant for employing the catego
ries in relation to appearances (their quid Juris). Most interpreters, however, would
probably reject this proposal. At the mention of the "Transcendental Deduction,"
the text that invariably comes to mind is Section 2 of Chapter 2 of the Analytic of
Concepts (A95-Al30). Section 1 (A84-95) is generally viewed as a series of pre
liminary remarks intended to explain what a Transcendental Deduction is and why,
in the case of the categories, it is indispensable. Its purely introductory character
would seem to be confirmed by Kant's decision in the 1787 B edition Critique to
Interpreting the Transcendental Deduction of the Categories 331
place A92--4 (carried over essentially verbatim into the new edition) under the head
ing, "Transition to the Transcendental Deduction of the Categories" (Bl24). Thus,
the notion that the "essential" OTDC task of proving the objective validity of the
categories is complete even before the start of Section 2, and that the latter, which
almost everyone views as the crowning achievement of Kant's theoretical philoso
phy, is in fact nothing more than the "inessential" STDC investigation into the pos
sibility of pure understanding itself is likely to seem a nonstarter.
Yet when one takes the trouble to compare Section 2 with the description of the
STDC in the A preface, the evidence in favor of drawing the line separating OTDC
from STDC at A94 appears quite strong. Section 1 concludes in precisely the way
one would expect if Kant had just completed the objective side of his inquiry at
A92-3 and had now to set the stage for its subjective side: he lists the "three origi
nal sources (capacities or faculties) of the psyche that contain the conditions of
the possibility of all experience" (A94) and states that, having already examined
sense in the Aesthetic, he will now, in Section 2 of the Deduction, be considering
the other two sources, imagination and apperception. Section 2 itself opens with a
restatement of the principle of the objective deduction: "if there are pure concepts
a priori, they can of course contain nothing empirical; yet, they must nevertheless
be nothing but a priori conditions of a possible experience as that on which alone
their objective reality can rest" (A95). Kant next proceeds to characterize the focus
of the inquiry on which he is embarking as "how pure concepts of the understand
ing are possible" (A95), that is, concepts that express the "formal and objective
condition[s] of experience universally and sufficiently" (A96). Though proof of
possibility is not the same as a proof of objective validity, he follows this up with
the claim that the categories are the concepts in question and that "it is already a
sufficient deduction of them, and justification of their objective validity, if it can be
proved that by their means alone an object can be thought" (A96-7).
If Kant had stopped there, one could reasonably take him to be announcing
that the OTDC is to follow (contrary to the suggestion in the A preface that A92-3
already suffices). Yet its immediate sequel calls attention to something that gives a
decidedly subjective turn to the investigation of the thesis that the categories make
possible the thought of an object:
[B]ecause in such thought more than only the capacity to think (Vermogen
zu denken), viz. the understanding, is employed, and since the understanding
itself, as a cognitive capacity that is supposed to relate to objects, likewise
requires elucidation as regards the possibility of this relation, we must first
consider the subjective sources that constitute the a priori foundation of the
possibility of experience, not according to their empirical but their transcen
dental constitution. (A97)
into "the pure understanding itself according to its possibility and the cognitive
faculties on which it rests," with the aim of solving the problem, "how is the capac
ity to think (Vermi:igen zu denken) itself possible?" Certainly, in the investigation
that follows and in the conclusions to each of its three expositions into which
Section 2 is divided (A98-Al14, A115-19, and A119-129), subjective consider
ations are everywhere paramount: empirical syntheses of apprehension, reproduc
tion, and recognition, along with their a priori subjective sources in the a priori
synthesis of apprehension, pure productive imagination, and a priori synthetic
unity of apperception. Equating any of these expositions with the OTDC seems
impossible since, as the OTDC is described in the A preface, such considerations
are extraneous to it, and introducing them would have defeated Kant's aim of
preserving the OTDC from the kind of subjective considerations that might give
it the appearance of being hypothetical (by contrast, A92-3 is free of any mention
of such STDC fixtures as pure and empirical syntheses of imagination and pure
and empirical unity of consciousness). Since internal textual evidence thus sug
gests that Section 2 (A95-130) is not the OTDC but the STDC, nothing could be
more wrongheaded than the endeavor, de rigeur among normativist interpreters,
to bracket everything that smacks of a priori psychology from their reading of
Section 2.
There is one piece of evidence that speaks against too rigid a demarcation between
the OTDC and the STDC: Kant's view of them as two sides of a single inquiry. It
is, after all, only in the A preface that the two are expressly distinguished. Since
prefaces tend to be among the last parts of a book to be written, one may con
jecture that Kant became concerned that the predominantly subjective character
of Section 2 of the TDC might put off some of his readers and so formulated a
fallback position in the preface in the hope of minimizing the risk of their reject
ing the Critique wholesale, including the OTDC's indispensable proof that the pure
concepts of the understanding have no cognitive application to the supersensible.
In particular, by making clear that the result of the OTDC is self-sufficient and
can stand on its own even without the explanation provided in the STDC, he could
hope to preserve that result in the face of resistance to his a priori psychological
explanation, much as Newton's theory of gravitation remains solid with or without
an explanation of how gravitational effects are caused.
Kant articulated essentially the same fall back position, even invoking Newton in
its defense, in a lengthy footnote inserted into the preface of the Metaphysical First
Principles of Natural Science. It is Kant's response to the contention of an anony
mous reviewer of a book by a certain Professor Ulrich that "without a completely
clear and adequate deduction of the categories the system of the Critique of Pure
Reason would totter on its foundation." It begins with a reaffirmation of the claim
Interpreting the Transcendental Deduction of the Categories 333
of the A preface that the basic thesis of the Critique regarding the limits of a priori
cognition can be established even without a complete deduction:
Against this, I maintain that for those who subscribe to my theses concern
ing the sensibility of all our intuition and the sufficiency of the table of the
categories as determinations of our consciousness borrowed from the logical
functions in judging in general (as the reviewer does), the system of the Critique
must carry with it apodeictic certainty because it is erected on the thesis that
the whole speculative use of our reason never reaches beyond objects of possible
experience. For if it can be proven that the categories which reason must make
use of in all its cognition can have no employment whatever other than in rela
tion to objects of experience (in that they make possible the form of thought in
this experience), then the response to the question how they make it possible is
indeed sufficiently important in order to complete this deduction where possible,
but, in relation to the main purpose of the system, namely, the determination of
the bounds of pure reason, is in no sense necessary, merely meritorious. For this
purpose, the deduction has already been carried sufficiently far if it shows that
the categories are nothing else than mere forms of judgment insofar as these
are applied to intuitions (which in us are always merely sensible), whereby, how
ever, objects are gotten for the first time and become cognitions: because this
already is sufficient to ground the entire system of critique proper with complete
certainty. Thus Newton's system of universal gravitation remains secure even
though it carries with it the difficulty that one cannot explain how attraction at
a distance is possible; but difficulties are not doubts. (474n)
Here, as in the A preface, Kant distinguishes between what is and is not essential to the
task of determining the bounds of a priori cognition. His claim is that the TDC does
not need to be "complete" so long as it is adequate to prove "that the categories which
reason must make use of in all its cognition can have no employment whatever other
than in relation to objects of experience (in that they make possible the form of thought
in this experience)." This is an almost verbatim repetition of the thesis expressed at
A93-the text Kant described in the A preface as sufficient for purposes of the OTDC
and so essential to achieving the primary purpose of the Critique: "the objective validity
of the categories as a priori concepts will rest on this: through them alone is experience
(according to the form of thought) possible." The other MFPNS footnote formulation
of the reason a Transcendental Deduction is necessary-that the categories are that
"whereby objects are gotten for the first time and become cognitions"-is likewise a
nearly exact match to the A93 statement that "apart from their presupposition, nothing
is possible as object of experience . . . with the consequence that concepts will underlie
objects in general as a priori conditions of experiential cognition."5
5 The most obvious difference between the two texts is that the there is nothing in A92-3 that
corresponds to the MFPNS statement that the categories are "nothing else than mere forms of
judgment insofar as these are applied to intuitions." But this is easily explained by the fact that in
334 Kant's Anatomy of the Intelligent Mind
Nor do the resemblances stop there. The part of the TDC Kant characterizes
as inessential in the MFPNS footnote starts from the result of the essential part
" that the categories . . . can have no employment whatever other than in relation
to objects of experience (in that they make possible the form of thought in this
experience)"-and endeavors to explain "how they make this possible." This repro
duces the pattern in the A preface, where the STDC takes its start from the conclu
sion of the OTDC-the principle that the pure concepts of the understanding are
objectively valid a priori as conditions of possible experience-and endeavors to
explain how an understanding is possible that can be the source of concepts such
as these. Both texts accordingly ascribe the same form of explanation to the ines
sential part of the Deduction ("a search for the cause of a given effect,"6 i.e., how
an understanding with this effect is possible); both identify its explanandum (the
"given effect") as the result of the essential part; and both describe the result of
the essential part in essentially the same terms. With correlations so close and so
numerous, it seems safe to conclude that the distinction in the two texts between
the essential and inessential sides of the Deduction is the same and that Kant, far
from discarding or forgetting the distinction between OTDC and STDC drawn in
the A preface as some suggest, retained it in both the MFPNS preface and the B
edition Critique of Pure Reason.
If that indeed is the case, then a subsequent portion of the MFPNS preface
footnote can be regarded as confirmation of the reading of the A preface as affirm
ing that TDC Section 1, especially A92-3, already of itself suffices to establish the
result of the OTDC and equating Section 2 (A95-130) with the STDC. For after
announcing his intention to produce a completely new version to replace a portion
of the A edition, Kant unmistakably identified the part of the TDC he will replace
with its inessential side, identified in the A preface with the STDC:
It therefore follows that all employment of pure reason can never relate
to anything but objects of experience, and because in a priori principles
( Grundsiitzen) nothing empirical can be the condition, they cannot by any
thing more than principles (Principien) of the possibility of experience in gen
eral. This alone is the true and sufficient foundation of the determination of
the boundary of pure reason, but it is not the solution to the problem how
experience is possible by means of these categories and only by their means.
This latter task, although the building stands firm without it, nevertheless has
the Critique, Kant had no need to repeat what he had just explained a few pages earlier, in the
metaphysical deduction of the categories (A 76-83): that pure concepts of the understanding
are universal representations of the determination of the pure synthesis in imagination of the a
priori manifold of sense in conformity to the logical functions of judgment (a result alluded to
at A90 ["functions of the understanding"]). Because the MFPNS text, by contrast, is an isolated
footnote with only a tangential relation to the text to which it is attached or to the MFPNS as a
whole, Kant rightly regarded a reminder of the contents of the MDC as warranted.
6 As remarked in note 4, the STDC is only like the search for a cause of a given effect (glei
chsam, so to speak), but not actually such a search.
Interpreting the Transcendental Deduction of the Categories 335
enormous importance and, as I now realize, equally great ease, since it can be
carried out almost by a single inference from the precisely determined defini
tion of a judgment in general (an act through which given representations first
become cognitions of an object). The obscurity that attaches to my previ
ous treatment in this part of the deduction, and which I do not disclaim, is
attributable to the usual fate of the understanding in inquiry, that the shortest
route is not the first that comes to notice. I shall take the next opportunity
to repair this defect (which concerns only the manner of exposition, not the
ground of explanation, which was already correctly given there). (MFPNS
475-6n)
Given that Kant replaced Section 2 (Bl29-69 in place of A95-130) when he pub
lished the second edition of the Critique one y ear later while retaining Section 1
with virtually no change (the only significant difference being the Bl27-9 addition),
there can be no doubt that Section 2-the text everyone has uppermost in mind
when they speak of "the Transcendental Deduction of the Categories"-is the text
consecrated to the inessential how question. The MFPNS footnote thus serves to
confirm the reading of the A preface that equates Section 1 of the TDC with the
OTDC and Section 2 with the STDC. For the A preface characterization of the lat
ter as an inessential explanation of the result of the former that takes a form similar
to that of a search for the cause of a given effect (Axvii) coincides perfectly with the
MFPNS characterization of the text to be replaced in the B edition as concerned
with an inessential how question that also, in light of the comparison with Newton,
must be understood as having a form akin to that of an explanation of the cause
(cf. the mechanism whereby attraction at a distance operates) for a given effect (cf.
a force inversely proportional to the square of the distance between the attracting
bodies).7
Kant's suggestion in the MFPNS preface footnote that accepting the result of
the OTDC while rejecting its explanation in the STDC is no more problematic
than accepting that Newtonian attraction at a distance exists without being able to
explain its mechanism raises the question why he nonetheless ascribed "enormous
importance" to the STDC. The reason is given immediately after the preceding cita
tion, where he asserts that failure to explain "how experience is possible by means
of these categories and only by their means" would
Here Kant makes clear that the only way objective necessity can be conferred on the
categories-which must occur a priori or not at all-is "from underlying a priori
principles of the possibility of thought itself" Since answering the question, "how
is the capacity to think itself possible?" (Axvii), is precisely the task Kant allotted to
the STDC, it seems clear that the underlying principles in question must be subjec
tive in nature and differ from Hume's associationist principles as a priori psychol
ogy differs from empirical (pure imagination and pure apperception founded on
pure sensibility from empirical imagination and empirical apperception founded
on empirical sensibility). Consequently, if one cannot tolerate the presence of psy
chology in a priori philosophy, or at least more than is found in the Transcendental
Aesthetic and the MDC, then one must be prepared to rest content with the won
derful but inexplicable fact that appearances comport themselves in perfect har
mony with the categories. To do so, however, is to deprive the categories of even
so much as the subjectively grounded objective necessity (Al49/Bl88) that accrues
to them as concepts that make experience itself possible (which alone, in Kant's
view, can suffice to rebut Hume: "it never entered his mind that the understanding,
through these concepts themselves, is perhaps itself the creator of this experience in
which its objects are found," Bl27; also PFM 313). T he categories will then be nei
ther necessary nor objective. For even if objects do comport themselves in perfect
harmony with the categories, this could only be known a posteriori, through actual
experience, and so as a matter of contingency and inductive generalization, not
strict necessity and universality. Indeed, the only "necessity" and "universality" that
could be accorded to such a harmony would be the subjective kind, characteristic
of Humean custom. As Kant put it in the B edition Deduction:
Before proceeding, it is worth noting that the distinction Kant drew-not only
here but also in the MFPNS footnote and elsewhere (e.g., Bl27, PFM 313)
between the categories as subjective conditions of thought that are also objective
because, through them, the understanding creates the very objects it experiences
(termed epigenesis at Bl67), and the categories as mere subjective conditions on
which the objects of experience in no way depend (preestablished harmony) poses
a fundamental challenge to normativist construals of the TDC and its quid Juris.
For to treat the categories as nothing more than a normative framework that con
fers objective worth on otherwise subjective judgments (which some have argued
is what the OTDC becomes without the STDC), while excluding, at least from a
final accounting of the TDC, the subjective, a priori psychological factors that were
Kant's only basis for treating the categories as constitutive not only of our thought
of objects but also of the objects themselves seems to transform the TDC into pre
cisely the kind of preestablishmentarian nonexplanation of the a priori agreement
of the categories with object that Kant deemed "just what the skeptic most desires."
So-called "transcendental arguments" do nothing to remedy matters. For in
pretending to prove that being normatively necessitated to think objects one way
(the categorial way) rather than another makes it impossible-incoherent, unintel
ligible-to doubt that they are that way in fact, one leaves oneself wide open to the
kind of objection critics such as Barry Stroud have raised against the normativist
version of Kant:
What calls into question the validity of the last step of would-be transcen
dental arguments from the way we think to the way things are is the appar
ently simple logical observation that something's being so does not follow
from its being thought or believed to be so. Something's being so does not
follow from everyone in the world's believing it to be so, from everyone's fully
reasonably believing it, even from every reasonable person's being completely
unable to avoid believing it. . . . That we all think things are a certain way is
one thing; that things are a certain way is another.8
Whether or not one is persuaded by Kant's argument that the STDC is necessary
to the Transcendental Deduction because it confers on the result of the OTDC
the objective necessity (albeit subjectively grounded) that enables it to stand up to
skeptical challenge, one might still contend that the proofs found in the Principles
of Pure Understanding suffice to the same end and so render the STDC redun
dant, with the added benefit of greatly reducing the role of a priori psychology
in the argument of the Critique as a whole. For the essential dependence of these
proofs on the two subjective "sources (faculties, capacities)" (A95) at the core of
the STDC, the necessary synthetic unity of apperception and the transcendental
synthesis speciosa of imagination, is by no means obvious and suggests that, had
Kant so desired, the principles could have been formulated and demonstrated inde
pendently of the STDC.10
It is true that Kant maintained that the transcendental schemata produced by
the transcendental synthesis speciosa of the imagination are presupposed in the
principles (Al45/Bl85). But the question is whether an explanation of their origin
in the mind is essential to establishing their role as sensible concepts of objects in
agreement with the categories (Al46/Bl86). For what difference does it make in
that regard whether the categories originate in transcendental imagination, Platonic
heaven, as Leibnizian innate ideas, Fregean Bedeutungen, or however else? Even
if one grants that the schemata are just as much in need of explanation as the
result of the OTDC, why suppose that Kant's Humean psychologistic search for
origins was the only kind of explanation he could (or should) have given? Why
not come to Kant's aid by introducing into their explanation the kind of analytic
techniques that were perfected only in the twentieth century, such as Strawsonian
connective analysis? T he same may be asked of Kant's characterization of the result
9 Admittedly, texts such as A2 l 3/B259-60 and A346-7 exhibit a contrary tendency. Yet when
subjected to scrutiny, it becomes clear that in none of these passages did Kant draw conclusions
of the kind he deemed "synthetic a priori" solely on the basis of conceptual analysis, without
also factoring in one's immediate, preconceptual consciousness of the manifold of sense in pure
intuition-the first, and indispensable, condition under which synthetic a priori judgments are
possible (B73).
10
T his seems to be the impetus behind Paul Guyer's interpretation of the Transcendental
Analytic in Kant and the Claims of Knowledge (Cambridge: Cambridge University Press, 1987).
Interpreting the Transcendental Deduction of the Categories 339
both editions) are one and the same. For once their identity is acknowledged, we
can no longer doubt that the Deduction, far from being an exercise in normative
philosophy more akin to analytic philosophy than to psychological philosophy as
practiced by Berkeley and Hume, is the most quintessentially psychological com
ponent of Kant's entire philosophy.
{ 13}
T he A Edition Transcendental
Deduction: Objects as Concepts of the
Necessary Synthetic Unity of the Manifold
In this and similar passages, Kant affirms that the very objects we experience, their
most fundamental laws and nature itself as their systematic totality, are products
of the understanding, at least in the sense that none of these are possible except
through the necessary synthetic unity of apperception this faculty confers on
appearances apprehended in intuition through its pure concepts.
It is crucial to distinguish this from the far more modest normativist thesis
commonly attributed to Kant: that the objects present to our senses can only be
represented as objects insofar as they are thought through certain concepts and
judgments, which together constitute the discursive framework (conceptual scheme)
1 My terminology in the preceding chapter reflected the emphasis there on the distinction
between objective and subjective transcendental deductions of the categories. Since my concern
now is not to determine what kind of inquiry is pursued in Section 2 of Chapter 2 of the Analytic
of Concepts but to interpret its contents, I shall revert to the common practice of referring to it
simply as the "Transcendental Deduction of the Categories," unless otherwise warranted. I also
return to my primary focus, within Kant's a priori psychology, on the psychologistic explication
of concepts, the method he adapted from Hume (ch 2-E).
342 Kant's Anatomy of the Intelligent Mind
[B]y synthesis in the most general sense I understand the action of putting
together distinct representations and conceiving their manifold in a cogni
tion.2 Such a synthesis is pure if the manifold is not given empirically but a
2 In his personal copy of the A edition of the Critique, Kant altered this sentence as fol
lows: "by synthesis I understand the action through which synthetic judgments come to be, in the
general sense...." (AA 23, E XXXIX)
344 Kant's Anatomy of the Intelligent Mind
priori (like that of space and time). Prior to all analysis of our representa
tions, these representations must first be given, and no concepts, in respect of
their content, can originate analytically. However, the synthesis of a manifold,
whether given empirically or a priori, first gives rise to a cognition, which
may indeed at the outset be raw and confused, and thus in need of analysis;
yet, synthesis is what actually collects the elements for cognition in the first
place and unites them in a certain content; it is thus the first thing we need to
attend to if we are to form a judgment about the first origin of our cognition.
(A77-8/B102-3)
What has not been combined cannot be taken apart, and, according to Kant,
no combination can be met with in the mind prior to and independently of the
spontaneity of imagination and judgment (i.e., aesthetic and discursive synthe
sis). If combination is never given, never a datum of the senses, never an affair of
receptivity alone, then the subject must synthesize its representations before it can
analyze them.3
To this one may object that one cannot collect and unify (synthetically unite)
the elements for cognition unless those elements are already present separately in
consciousness (analyzed), so that there is just as much reason to hold that synthesis
presupposes analysis as the opposite, or rather than each presupposes the other. To
understand why Kant assigned priority to synthesis, the A edition Deduction is the
natural place to look since it is there that one finds his most extended treatment of
the topic of synthesis. Kant explained the need for it as follows:
Apprehension is synthesis in its simplest, most elementary form, and takes its start
from the synopsis of the manifold a priori through sense. As an act of imagination
that applies immediately to the manifold given of sense, one might take apprehen
sion to be an example of a synthesis that presupposes an analysis, at least in the
4 Nor is it less absurd to posit a different order in the case of the a priori counterparts of
apprehension and reproduction. Before the manifold of pure space or time can be arrayed,
ordered, or otherwise related in pure imagination (figurative synthesis), be it conformably to the
The A Edition Transcendental Deduction 349
is still more evident. Kant stated on numerous occasions that appearances can be given
in intuition without any help from concepts, including the categories (B67, A89-91/
B122-3, Alll, Al24, B145, A253/B309). Indeed, concepts cannot even be formed
until the analytic unity of apperception confers logical universality on already appre
hended intuitions (ch 9-C).5
Just as with apprehension in relation to reproduction and recognition, any
attempt to construe Kant's assertion that reproduction is "inseparably combined"
with recognition in a concept as a claim that the former unconditionally presupposes
categories (transcendental synthesis speciosa) or not, the a priori manifold of pure space and time
must already have been taken up into the activity of pure imagination (synthesis of apprehension
in intuition), and this always requires a "synthesis of the manifold that sensibility offers in its
original receptivity," that is, "a pure synthesis of apprehension" (AIOO): see chs 3-5.
Reproductive synthesis can take the manifold, apprehended any which way, and set
it in an order fixed in accordance with customary association and other psychologi
cal influences. If, for example, I should happen to apprehend the fifth in a habitual
sequence before the third and fail to apprehend the fourth because, say, I happened to
be looking elsewhere, the fifth nevertheless induces me to expect the imminent appear
ance of the sixth. But even though the quantum being produced in my imagination
arithmetically increases as my thought advances from one unit to the next, I can have
no consciousness of this successive increase so long as I merely blindly repeat the same
operation of adding one but do not at the same time recognize the act of having pro
duced precisely this number as having been necessitated by what came before and as
itself necessitating which number is to come next. For merely advancing in thought
from a third to a fourth to a fifth and so on is not yet to recognize that the fourth, and
only the fourth, may succeed the third, that after the fourth no other than the fifth may
follow, and so on. Only by means of a consciousness of the principle of unity of the
synthesis as a rule (A105) that necessitates all the steps in the sequence can each of its
individual steps (the final step included) be recognized as the arithmetical sum of all
the preceding. Instead, one is aware only that the transition from the third to the fourth
in the sequence is preferred because it feels easiest, that a transition to any other mem
ber of the sequence or to anything outside the sequence is disfavored because it feels
difficult, so that there is never, at any point in the sequence, a consciousness of the total
(sum) being produced (continuously increased) with each transition from predecessors
to successors.
T here is, however, a seemingly insuperable obstacle to incorporating syntheses
of recognition into the psychology of experience that helps to explain their absence
from pre-Kantian sensibilism. While Hume, in particular, could admit "philosoph
ical relations" and "general rules," they are essentially analogical extensions of the
same natural relations that owe their influence on thought and action entirely to
their transition-facilitating associative character.8 To prescribe what must follow
8 Although Hume deemed facility affect essential (THN 99169, 204/135, 220/145, 260/169)
to those natural relations "by which two ideas are connected together in the imagination, and
the one naturally introduces the other, after than manner above-explained [in 1.1.4, 'Of the con
nexion or association of ideas']" (13/14), philosophical relations remain possible in its absence.
Nevertheless, in all cases where philosophical relations are not necessary in the strict sense (95/66),
Hume seems to have regarded them as parasitic on natural, most notably in the case of causal
relations: "tho' causation be a philosophical relation, as implying contiguity, succession, and con
stant conjunction, yet 'tis only so far as it is a natural relation, and produces an union among
our ideas, that we are able to reason upon it, or draw any inference from it" (94/65). Why are
philosophical relations incapable of serving as a basis for reasoning unless there is a correspond
ing natural relation? In THN 1.3.6 ff., Hume makes clear that inference always proceeds from
something believed to something not believed, where the vivacity of the first is communicated
to the second by means of their relation (vivacity being, for Hume, constitutive of belief: UU ch
17-C and HTC chs 1 and 2). So if philosophical relations unsupported by natural relations are
unable to serve as a basis for reasoning, it must be because they lack something essential whereby
the vivacity of one perception related in them can be communicated to another perception that
lacks it. Since facility is the feature that distinguishes the two kinds of relation, the presence of
this affect must be what is crucial to the ability of relations to communicate the vivacity of one
352 Kant's Anatomy of the Intelligent Mind
what in the synthesis of the manifold, with complete indifference to how the tran
sition feels, one must be able to produce rules of a kind not explicable by associa
tive psychological means alone. For rather than blindly repeating or extending to
different but closely resembling circumstances the same sequence of synthesizing
operations, one must be conscious (1) of each operation in a synthesizing sequence
as necessitated by its predecessor and necessitating its successor, (2) of each new
operation as the (synthetic) unity of all the preceding (e.g., the sum generated at
any point in the count, the total space occupied at any point in describing a line),
(3) of the synthesis as a whole in each of its constituent operations and so as identi
cal throughout the entire sequence (the rule as analytic unity of consciousness),
(4) of the same identity as present whenever the same operations are or can be
performed in the same sequence, including any and all apprehended manifolds in
which the sequence is reproducible (universal scope), and (5) of these as instances
of the same analytic unity of consciousness. Clearly, no such consciousness of
reproductive synthesis is attainable by means of anything less than a full-fledged
concept, that is, a representation of the synthesis that is logically universal in form
(ch 9-C) and so amenable to logical combination with other concepts injudgments
(ch 10). Humean ersatz "universals" therefore fall short because, as themselves
nothing more than blind associations responsive only to what is or is not felt in
transitions of thought in accordance with the "rule" of (human) natural resem
blance relations (ch 9-A, UU ch 18-B), they are quite incapable of yielding a con
sciousness of reproductive synthesis adequate to any, much less all, of the facets
of consciousness (1)-(5).
Nor can this psychological want be made good by supplementing universals of
customary association with linguistic universals: far from being in any sense repre
sentations (consciousnesses) of sequences of synthesizing operations performed in
reproductive imagination, the rules specific to linguistic universals are mere conven
tions governing the use of signs that have their meaning completely independently
of mentation and its contents (although they derive objective meaning and ontolog
ical worth from the latter: chs 9-A, -B, and UU 4-C). Only logically universal mental
concepts of the kind unique to Kant's sensibilism, formed by attaching the analytic
unity of apperception to purely aesthetic representations of associations and other
reproductive syntheses,9 are capable of yielding the rules he deemed requisite for
perception (the believed "premise") to another (the otherwise unbelieved "conclusion"). In other
words, only insofar as the transition from an impression or vivid idea X to an idea Y feels easy
will the vivacity of the former be naturally extended to the latter and X constitute a reason for
Y. In order for philosophical relations to be usable as supports for reasoning, they therefore
must first be recognized as analogous to some species of transition-facilitating natural relation
(via resemblance, a general rule, or some other circumstance). See UU 486n23 and HTC chs
2-D and 5.
synthesis of recognition, and so too for cognitive experience. For since it should
now be clear that Kant's conception of association as reproductive syntheses in
which "a representation enters into combination in the imagination rather with this
one than another" (Al21) derives from Hume, he may be presumed to have under
stood them as Hume did: "principles which are permanent, irresistible, and univer
sal .. .so that upon their removal human nature must immediately perish and go to
ruin" (THN 225/148; Kant too speaks of the "rules" of preferential reproduction at
Al21). As "hard-wired" into our species-specific psychology (and much nonhuman
animal nature as well: THN 327/212-3), rules of preferential reproduction can also
be presumed to operate virtually from the beginning of conscious life, producing
the associations from which recognitive concepts can be formed. Since, for Kant,
realizing this potential in actual concepts requires nothing more than the addition
of the analytic unity of the I think to these associative rules (ch 9-C), synthesis of
recognition may likewise be presumed to take place long before we are out of our
cribs. And, indeed, Kant stated clearly that in his view, the representation of the
identity of the I think exists even in the minds of infants long before they learn the
words "I think" or any other conventions governing the use of signs, and so prior to
and independently of the entire socio-normative framework in which the universals
of human languages are embedded:
That man can have the I in his representation raises him infinitely above all
other living things on earth. Because of this he is a person and, by virtue of
the unity of consciousness in all alterations that may befall him, one and the
same person, i.e. a being entirely different from things (Sachen), such as the
animals that lack reason whereof he is master through his quality and worth,
even when he cannot yet speak the I, since he still has it in thought: just as all
languages, when they discourse in the first person, must still think it even if
they do not express this 1-ness through a particular word. For this capacity
(namely, to think) is the understanding. (Anthropology 127)
The I think presupposes nothing but the prediscursive synthetic unity of the mani
fold in pure space and time (unity of sensibility) (ch 9-B), the product of pure
synthesis of apprehension, which precedes and makes possible not only the I think
but also the entire threefold synthesis as well (A98-9), and so must be in place
virtually from the beginning of conscious life. And since Kant intended his model
of the mind to apply not just to humans but to all sensibly conditioned under
standings as such, even apperceptive creatures whose existence was devoid of any
semiotic or socio-normative dimension would still be able to fashion recognitive
concepts from their reproductive syntheses (for which reason it is imperative to
comprehend the conceptual component of recognitive synthesis, and cognition
generally, not normatively but psychologically, in relation to the analytic unity of
apperception: ch 9-C).
To be convinced that concepts formed in this way yield a consciousness capable
of satisfying all of (l)-(5) above, one needs to recognize that a recognitive concept
354 Kant's Anatomy of the Intelligent Mind
10 Presumably, Kant would have agreed with Locke that even when a complex mental opera
tion has become habitual, each of its constituent acts must still be performed individually and in
sequence if we are to become conscious of their complex product as a whole (synthetic unity),
notwithstanding the fact that the automatic character and rapidity of habitual performances
obscures the inner perception of all the mind is doing: "Nor need we wonder, that this is done
with so little notice, if we consider, how very quick the actions of the Mind are performed: For as
it self is thought to take up no space, to have no extension; so its actions seem to require no time,
The A Edition Transcendental Deduction 355
apprehended in intuition (space and/or time), are completely devoid of order and rela
tion ("scattered and single in themselves," A120), all order and relation that we take
to be "found" in intuition in fact have to be put there ("we can represent nothing as
combined in the object without having ourselves previously combined it," B130; also
A97-8, B155). This means that strictly speaking, no images of objects in space and
time are ever apprehended (A120n); they must always be synthesized by reproductive
imagination, so that an image produced by this synthesis, such as that of a line, can
only be recognized via a consciousness of the rule of its successive generation. And
such consciousness not only requires a concept of the object, it is the concept of that
object (A103-4).
Reproduction is thus only the start of the process of transforming apprehended
appearances into objects fit for cognitive experience. In the absence of empirical
psychological principles of preferential reproduction (association), apprehended
manifolds would be "reproduced just as they fall together, without distinction from
one another, [and] no determinate interconnection among them could arise, merely
chance aggregations (regelose Haufen)" (A121). In the human case, these princi
ples make transitions to contiguous perceptions easier than to noncontiguous, an
association that becomes far stronger when, with frequent repetition, it becomes
customary (UU ch 18-C). So far, so Humean. What Kantian psychology adds is
a capacity of the mind to convert strong associations of this kind into recognitive
concepts by uniting the associative rule (as the matter of a recognitive concept) with
the consciousness of the identity of the I think (the form of all concepts, logical
universality, as analytic unity of apperception).
In the case of the image of a line, the addition of logical form confers a neces
sary unity on the synthesis that it lacks in its original purely aesthetic incarnation
as a sequence of felt transitions in imagination. As we saw in Chapter 9, the rep
resentation I think, as a consciousness that "must be able to accompany all my
representations" (B131), is ipso facto a consciousness common to them all and so
is universal in scope (any representation to which the I think is not common is, for
that reason, "nothing for me," B132; ch 9-B). But since the I think is a conscious
ness with no manifold (simple, empty) (chs 2n14, 8-F and 9-B), it is only insofar
as its representation is combined with a representation provided by the senses that
an actual concept, with a matter as well as a form, can emerge (ch 9-C). In Kant's
example at B133-4n, the concept of red is formed by attaching the analytic unity
but many of them seem to be crouded into an Instant. ... How, as it were in an instant, do our
Minds, with one glance, see all the parts of a demonstration, which may very well be called a long
one, if we consider the time it will require to put it into words, and step by step shew it another"
(ECHU II/ix/§10; UU chs 4-D and 5-B). Hume, too, took this view: see UU ch 19-B (this is one of
the more important findings in UU because it goes against the way David Owen and others have
portrayed the operation of customary association in human understanding). The only difference
with Kant is that the synthesizing actions coincide with inner affections that the imagination
exhibits in a succession of appearances (ch 4-C), and it is this succession that seems crowded into
an instant and makes us overlook its character as a synthesized manifold.
356 Kant's Anatomy of the Intelligent Mind
of the I think to a sensation of red (or, rather, its reproduction in thought), mak
ing this otherwise purely aesthetic consciousness logical in two respects: through
the form of the concept, we are conscious of it in relation to the logical totality
of possible representations (universal scope), and, through its matter, we are con
scious of this totality as logically divided into the red and the non-red, i.e., instances
and non-instances, without singling out any instance from any other (generality
of reference). And the same is presumably true when concepts of associative rules
are formed: by attaching the I think to the sequence of operations comprising the
synthesis of, say, a line, we add to our consciousness of it a relation to the logical
totality of possible representations as well as a consciousness of this totality as logi
cally divided into lines and non-lines, without singling out any line from any other.11
What makes the formation of the concept of a line in the manner just described
so crucial for Kant's purposes in the Transcendental Deduction of the Categories is
that it is through the concept of the synthesis that we are, for the first time, conscious
of the sequence of synthesizing operations (Humean transitions of thought) as a
unity. For what apart from the analytic unity of the I think is perceived merely as
a succession of distinct acts whose only association with one another are the facile
transitions between them becomes, with its addition, something we are conscious of
relating as a circumscribed, integrated whole to the totality of representations and
of dividing up that totality into lines and non-lines according to the associative rule
as such a whole. Moreover, with its conversion into a concept, the possibility arises
of using it as a rule in subsequent reproductions in an entirely new sense. Instead of
proceeding blindly, guided only by how transitions of thought feel, the reproduction
of the sequence of acts is now accompanied at every step by a consciousness of the
whole in the concept. As a consciousness, however weak, of the whole sequence,
recognition in a concept makes us aware of each action's place in it so that in per
forming it, we are at the same time conscious of performing it in consequence of acts
previously performed, and so as necessitated by its predecessor and necessitating its
successor under its concept-rule. T his, in turn, makes us conscious of each act as
successively generated by all of those that preceded it, and, as such, their (synthetic)
unity, and so on for each subsequent act until, with the act that completes the syn
thesis, we are conscious of the necessary synthetic unity of the synthesis as a whole.
And since that also includes the apprehended manifold synthesized within it, the
concept of this whole IS the object ("this object is nothing more than that something
the concept of which express such a necessity of synthesis," Al06; also Bl37).
In the context of the argument of the Transcendental Deduction, the unity rep
resented in recognitive concepts is best understood as a derivative expression, in
empirical or mathematical microcosm, of the all-encompassing original analytic
u While assessing in individual cases which representations are and are not instances of the
concept of a line is the work of comparison and resemblance association, the division between
instances and non-instances is itself purely logical, hence indifferent to whether either or both
divisions are instantiated at all, and so independent of the exercise of reproductive imagination.
The A Edition Transcendental Deduction 357
and synthetic unities of apperception (A106-7; ch 9-B). For just as the original
analytic unity of apperception is the representation of the identity of conscious
ness in respect of all the manifold, a recognitive concept is the representation of
the identity of the consciousness of the whole reproductive synthesis in relation to
the consciousness of each in the sequence of its constituent operations-"a con
sciousness that what I am thinking is the same as what I thought a moment before"
(A103)-and so, in respect to the manifold of the synthesis, the concept counts
as the analytic unity of its apperception. Similarly, just as the original synthetic
unity of apperception is the necessary synthetic unity of all the manifold in one
consciousness, a recognitive concept is a consciousness of each step in a reproduc
tive synthesis as the synthetic unity generated, and so necessitated, by the totality
of its predecessors: "a consciousness that unites the manifold successively intuited,
then reproduced, in one representation ... so that we connect it [not] only with the
outcome but ... with the action itself, that is, immediately with the production of
the representation,.. represents in any given appearances the necessary reproduc
tion of their manifold, hence the synthetic unity in their consciousness" (Al 03-6).12
The effect of this necessity of reproduction is, to be sure, to render the ease felt in
transitions redundant for a consciousness of the relation of what follows to what
precedes, freeing the synthesis from continuing dependence on its origin in associa
tive imagination and thereby opening the way to objective judgments of experience
("representations are combined in the object, i.e. with indifference to the state of
the subject," B142). Nevertheless, it would be a mistake to conclude that this frees
recognitive synthesis from any further dependence on reproductive association, thus
proving psychology to be, in the final analysis, dispensable. On the contrary, synthe
sis of recognition could never suffice for objective judgments of experience without
reproductions grounded in the affinity of the manifold. So it is to this, and the wider
question of the nature and grounds of objectivity as such, that I now turn.
12 My elision between Al04 and A106 does not result in any distortion of Kant's mean
ing. The synthetic unity of apperception produced by recognitive concepts is, of course, deriva
tive and, as will emerge in subsequent sections, depends as much on the original synthetic unity
founded on the categories as on the prediscursive original synthetic unity founded on pure space
and time.
358 Kant's Anatomy of the Intelligent Mind
to reproductive syntheses and so, even in their capacity as rules for generating the
necessary synthetic unity of a manifold, have no more than subjective, psychological
worth. This limitation can to some extent be overcome in the cases of counting, draw
ing a line, and other reproductive syntheses from which everything empirical can be
excluded, thus converting the concept from a rule of empirical reproductive synthesis
to an arbitrary rule of pure productive synthesis of the purely formal manifold of
pure intuition (all mathematical syntheses are arbitrary: A729-31/B757-9 and L 141).
For these concepts are not essentially dependent on association: they could, in prin
ciple, have been acquired directly from the manifold of pure intuition by attaching the
analytic unity of the I think to figurative syntheses in pure productive imagination.
But this is not the case with those recognitive concepts that are as much representa
tions of the matter as the form of apprehended appearances, including all concepts
of objects of experience. Such concepts first become possible through a consciousness
of sensation-what Kant termed perception ("sensation whereof one is conscious,"
A225/B272; also B147, B207, B233-4, A373-4, Progress 276 and 279)-and so are
ineluctably bound up with empirical syntheses of apprehension and reproduction.
Since this is just to say that recognitive concepts of this kind depend essentially on the
reproductive syntheses to which they give discursive expression, the question arises
how it is possible for them ever to result in a consciousness of genuine objects (mate
rial and thinking things) rather than simply acts of imagination proceeding in blind
conformity to the principles of associative imagination.
To comprehend the problem more fully, consider what Kant deemed requisite
if empirical syntheses of recognition (cognitions) are to have relation to an object:
[O]ur thought of the relation of all cognition to its objects carries with it
something of necessity, since this object is regarded as what prevents our
cognitions from being random or arbitrary but instead determined a priori a
certain way. For since they are supposed to relate to an object, they necessar
ily agree with one another in relation to this object, i.e. they must have that
unity which constitutes the concept of an object. (A104)
Yet it is difficult to see how the necessity conferred on the reproductive synthesis
of a manifold by a concept functioning as a rule of necessary synthetic unity can
relate the manifold to an object when that concept simply replicates the ordering
of representations that results through customary association founded on frequent
encounters with constantly conjoined appearances. Since, in Kant's view, necessary
relations between distinct appearances are never given empirically, and since associa
tion is always purely empirical, it is impossible for any necessity objects might confer
a priori on the manifold apprehended in intuition to be conveyed to the understand
ing by means of reproductive imagination. Association is precisely not necessary
but merely preferential reproduction (Al 12, A121). As Hume put it, it is not some
thing "that really binds our several perceptions together" but "only associates their
ideas in imagination," not "some real bond .. . we observe" but one we "only feel
among the ideas we form of them" (THN 259/169), and so merely "a gentle force,
The A Edition Transcendental Deduction 359
which commonly prevails" rather than "an inseparable connexion" (10/12-13). Nor
are reproductive preferences in any way grounded in the manifold apprehended in
intuition. Since the latter consist of "distinct perceptions met with in the mind scat
tered and single in themselves" (Al20), order or relation cannot be derived from
apprehension but must instead be added by the subject in reproductive imagination
(Sections A and B). In Hume's associationism, this happens when a transition to a
certain perception feels easier than to any other. But however it occurs, associative
propensities are all that prevent apprehended appearances from being "reproduced
just as they fall together, without distinction from one another" so that "no determi
nate interconnection among them could arise, merely chance aggregations" (Al21).
What this means is that apart from the associative propensities of human nature,
reproductive imagination confronts the understanding with as many possibilities for
forming recognitive concepts as there are possible ways to reproduce an apprehended
manifold. Given that apprehension is not subject to the limitations of our psycho
logical capacity to discern (the minima of Berkeley and Hume) but contains succes
sive and juxtaposed manifold within manifold without end (ch 4), the reproductive
possibilities offered by each manifold apprehended in empirical intuition are quite
literally infinite, each of which admits of being converted into a concept of the nec
essary synthetic unity of the apprehended manifold simply by attaching the analytic
unity of the I think to it.13 Since only one such recognitive concept can reflect the
necessary synthetic unity conferred on the manifold by the object a priori (Al04),
there is only one fact of the matter to be cognized. How then is the understanding
to sift through the plethora of recognitive possibilities, eliminate all of those that
yield a necessary synthetic unity of appearances different from the a priori necessary
synthetic unity the object actually imposed on them, leaving only the one concept,
the one rule of reproductive synthesis, that actually corresponds to the object? Since
"we have to do only with the manifold of our representations and that X which cor
responds to them (the object), because it is supposed to be something distinct from
all our representations, is for us nothing" (Al05), the question seems impossible
to answer. And doesn't this mean that Hume was right after all? For even with the
addition of synthesis of recognition in concepts and the necessary synthetic unity
its concepts impose on the reproduction of apprehended manifolds, the understand
ing seems incapable of transcending its dependence on the ineluctably subjective,
species-specific psychological deliverances of associative imagination.
The key to solving the objectivity problem lies in appreciating that it goes beyond
a plethora of recognitive possibilities and puts the possibility of association itself
into question:
[E]xperience ... does indeed teach us that something else customarily fol
lows an appearance but not that it must necessarily follow, nor that it can
13 Even if restricted to minima, the ability to repeat the same apprehended appearance in
reproductive imagination as often as one chooses yields infinitely many reproductive possibilities.
360 Kant's Anatomy of the Intelligent Mind
action, or motion) happens that is not preceded by something else upon which
it always follows.14 Kant's query concerns what must be the case if the unifor
mity principle is to count as a law of nature, that is, a rule to which perceived
appearances are themselves subject, prior to and independently of the concepts
we form of them in experience. Since the principle is founded on customary
association, if it is to have the status of a law, it presupposes that appearances
are "in themselves ...associable" (A121-2). And so the question becomes how
this is possible.
What does it mean for appearances to be "associable in themselves"? It is
through customary association that we learn from past experience. Such customs
are instilled through the observation of resemblances between distinct percep
tions, as concerns not only their quality but their co-occurrence and sequence
(temporal contiguity) as well. But before experience of patterns (regularities) in
the occurrence of perceptions can issue in associative customs, the observing sub
ject needs to be able to recognize new co-occurrences and sequences as recurrences
of a previously experienced pattern of conjunction and, moreover, to distinguish
their conjunction as more or less frequent and constant. For this to be possible,
however, the observing subject must have the capacity to situate all occurrences
of perceptions along a single, linear, ordered succession of time, with determinate
before and determinate after; otherwise, it would be impossible to distinguish the
one-hundredth encounter from the first or to distinguish 100/100 recurrences of
perceptions in the same pattern from 10/100 recurrences. This, however, is not
14 The uniformity principle asserts "that the future will resemble the past" (EHU IV/ii if 21),
i.e., when "such an object has always been attended with such an effect ... other objects, which
are in appearance, similar, will be attended with similar effects" (if 16). Its validity thus precludes
"that the course of nature may change, and that an object, seemingly like those which we have
experienced, may be attended with different or contrary effects" (if 18). According to Hume, our
unshakeable conviction in its validity is produced not by reason but custom, and it constitutes a
principle of probable reason because if we were not perfectly convinced of it, or were convinced
of its contrary, no such thing as probable reasoning could ever take place (if 19). It is thus a con
viction that is, and that human nature ensures will inevitably be, "perfectly familiar to me, long
before I [am] out of my cradle" (if 23). See UU ch 19 and HTC ch 4-C.
The A Edition Transcendental Deduction 361
how appearances are apprehended in inner intuition, where they are met with
"scattered and single in themselves" and without the least temporal differentia
tion and determination (ch 4). For although appearances in inner intuition do, of
necessity, conform to the order of time itself, where "the preceding time neces
sarily determines the following (because I cannot arrive at the following elsewise
than through the preceding)" (A199/B244), the pure time of sensibility, as the
product merely of a prediscursive (nonconceptual, non-categorial) synthesis and
apperception, is quite incapable of situating inner appearances within a univer
sal, all-encompassing temporal order in which each appearance is uniquely dif
ferentiated from and completely determined in relation to every other (ch 4-D).
There is, accordingly, nothing in inner intuition remotely adequate to the ordered
linear succession presupposed for the kind of experience that permits patterns of
temporal occurrence (constant conjunctions) to be observed. And it is precisely
this non-associability of appearances as apprehended in intuition, I believe, that
Kant had in mind when he asserted that cognitive experience, pace Hume, can
never result from association alone but always requires, in addition, the affinity
of appearances:
[H]ow is this association itself possible? The ground of the possibility of the
association of the manifold, insofar as it lies in the object, I call the affinity of
the manifold. I therefore am asking how you make comprehensible to your
selves the thoroughgoing affinity of appearances whereby they stand under
constant laws and must belong under them. (Al 13)
In the absence of the affinity of appearances under constant laws, the experi
ence of regularities that issue in customary association would be impossible, and
"perceptions would belong to no experience, would thus be without any object,
and be nothing but a blind play of representations, i.e. less than a dream" (Al 12).
For although Hume believed that customary association does not presuppose but
rather is presupposed by causal connection, the identity of complex individuals at
and over time, and universal natural laws (UU chs 18-C, -D, and 19-D), his asso
ciationism depends entirely on the assumption that all occurrences of perceptions
fall along the kind of linear, ordered succession of time, with a determinate before
and a determinate after, that enables us to represent the frequency and constancy of
their conjunctions. If, as Kant maintained, this affinity of appearances as regards
their existence in time turns out not to be something that can simply be taken for
granted and indeed is altogether lacking at the level of apprehension in intuition,
then the possibility of cognitive experience requires that this want be made good
at a higher level of synthesis, since otherwise there could be no such things as the
"frequency" and "constancy" of the occurrence of perceptions at all. As this would
preclude the kind of experience requisite for inculcating customs of associating
distinct perceptions, Kant was not exaggerating when he characterized conscious
ness in the absence of affinity as "nothing but a blind play of representations, less
than a dream."
362 Kant's Anatomy of the Intelligent Mind
Along with reducing conscious experience to a "blind play less than a dream," Kant
claimed that the absence of affinity among appearances would preclude the unity
of the subject of that experience:
[I]f ...association did not have an objective ground such that it was impos
sible for appearances to be apprehended otherwise by the imagination than
under the condition of a possible synthetic unity of this apprehension, then
it would be entirely contingent that appearances fit into an interconnected
whole (Zusammenhang) of human cognition. For though we had a capacity
to associate perceptions, it would nevertheless remain entirely undetermined
and contingent whether they were also in themselves associable; and in case
they were not, numerous perceptions, and indeed an entire sensibility, would
be possible in which much empirical consciousness would be found, but sepa
rately, and without belonging to a consciousness of myself. (A121-2)
If perceptions lacked affinity, and so did not all exist "in themselves" (objectively)
in a single necessary order of temporal succession, they could not all belong to a
single experience and would not all be associable with one another. Though some
might remain associable insofar as they happened to fall together in inner appre
hension, there would be nothing to necessitate that all others fall within that succes
sion; and if there were perceptions that fell outside it, they would not belong to the
same experience and so could not be related to those in the other through custom
ary association. Indeed, since the others could not belong to the identical empirical
self whose consciousness that experienced temporal succession demarcates, the way
would be open for there to be as many distinct empirical subjects, each with its own
sensibility, as there were disconnected successions of perceptions in a mind. This
Kant rightly declares to be impossible (A122). He thus concludes that the affinity
responsible for making appearances associable in themselves is essential if we are to
have a single interconnected experience that not only is more than a dream but also
rises all the way to genuinely objective cognition.
Kant's claim that without affinity there could be multiple selves, each with "an
entire sensibility" (A122), might be thought to contradict my interpretation of his
doctrine of sensibility as affirming the unity of sensibility on the basis of prediscur
sive, purely aesthetic operations of sense imagination and apperception (chs 3-4).
For if that were Kant's view, would it not follow that he had already, on purely aes
thetic grounds, precluded the possibility of multiple selves, each with its own unity
of sensibility? Yet he could not have been clearer in the Transcendental Deduction
(1) that they remain possible in the absence of a transcendental affinity among
appearances, (2) that this affinity consists in a necessary synthetic unity of appear
ances in one apperception effected by their subordination a priori to constant laws,
and (3) that these laws depend not merely on space and time but on the categories as
The A Edition Transcendental Deduction 363
well (Al 11-4, A123-5). Since this is just to say that the possibility of multiple selves
and sensibilities can be precluded only by means of the categories, the interpreta
tion of pure space and time as ensuring the unity of sensibility thus seems liable to
a potentially fatal objection.
Rather than requiring a reply, the objection is best viewed as a demand for clari
fication. The question is how the two original synthetic unities of apperception pos
ited in my interpretation, precategorial and categorial, relate. In Chapter 4, I argued
that the key to understanding Kant's problem in the Transcendent Deduction is to
recognize just how little sensibility is able to set in place ahead of the understand
ing and, correspondingly, how great the void the latter is obliged to fill by means
of transcendental synthesis in conformity to its pure concepts. In particular, to be
"in" the space or time of sensibility is not to be spatially or temporally situated
relative to other appearances. For neither the pure space nor the pure time of sensi
bility may be conceived as a universal, all-encompassing order in which each space
and time (i.e., each outer or inner appearance considered formally, without regard
to its matter) is uniquely differentiated from and determinately related to every
other. The pure space and time of sensibility are sufficient for one thing and one
thing only: a mode Uuxtaposition or succession) of representing a homogeneous
manifold as all standing immediately together in one and the same consciousness.
Otherwise, the appearances in them are "met with in the mind scattered and single
in themselves" (A120), devoid of all relation among themselves and, in particular,
that necessary interrelation (affinity) which Kant deemed essential to objectivity
(A104-8, A122-3).
Once this is grasped, it becomes clear that the unity of sensibility effected pre
discursively by pure space and time answers perfectly to Kant's characterization of
what consciousness would be like in the absence of categorial affinity as a "blind
play of representations, i.e. less even than a dream" (Al 12). For the prediscursive
original synthetic unity of apperception founded on pure space and time is com
pletely indifferent to whether the appearances concerned in these unities are related
to one another, including whether they exhibit transcendental affinity (i.e., are or are
not associable in themselves). That they are "scattered and single in themselves" and
so devoid of all spatial and temporal differentiation and determination, therefore
does nothing to prevent them from all being represented as the manifold of a single
representation, contained in the consciousness of that representation (ch 4), and so
suffices for the representation of the identity of this consciousness in all perceptions,
that is, the analytic unity of the I think (ch 9). But this unity/identity of the subject
extends only so far as the most elementary level of representation (apprehension).
Customary association is possible only on the basis of experience of sequences and
co-occurrences of perceptions as (more or less) frequent and constant; and no such
experience is possible without a consciousness of all perceptions as occurring in lin
ear succession, each uniquely situated in relation to the others, not just as preced
ing or succeeding this or that inner appearance in intuition (apprehension), but as
determinately before or determinately after each and every other, and determinately
364 Kant's Anatomy of the Intelligent Mind
so much before or after it. Where categorial affinity is lacking, nothing approaching
the single integrated interconnection (Zusammenhang) of appearances requisite for
experience is possible. Since this includes self-experience (cognition of an object in
which all inner appearances are necessarily united), it follows that sensibility can have
no objective unity, nothing that necessarily unites syntheses of the manifold of inner
appearances. In other words, in the absence of affinity, the manifold of inner intu
ition, far from being able to support empirical self-consciousness (the representation
of an identity in inner appearances at and over a span of time), could yield nothing
more than chance aggregations of perceptions (A121). Thus, the subjective unity of
sensibility yielded by synthesis of apprehension (chs 3-4) must be complemented
by an objective unity of sensibility constituted by the determination of appearances
conformably to the categories through the subsumption of the threefold synthesis
under transcendental schemata by transcendental judgment (chs 17-18).
The considerations lead naturally to Kant's definition of experience as the syn
thetic unity of the manifold of perceptions according to laws or, alternatively, of
appearances according to concepts (indicating the discursive character of this syn
thetic unity):
take place. For if we speak of different experiences, these are only so many
perceptions belonging to one and the same universal experience. The thor
oughgoing and synthetic unity of perceptions constitutes the very form of
experience, and this form is nothing else than the synthetic unity of appear
ances according to concepts. (Al 10)
(discursive) conditions of a possible experience that make good this want, and they
become associable in themselves, the integrity of the subject of intuition as subject
of experience is established. For while the subject of intuition may not suffice for
the necessary synthetic unity of perceptions according to laws that Kant equated
with experience, the representations specific to this experience are, qua representa
tions, appearances apprehended in inner intuition and so belong to the unity of
sensibility of the intuiting subject. In particular, when the conditions of possible
experience are met , and these appearances become events determinately situated
within a single linear succession of time (transcendental affinity), the representation
of their frequency and constancy of occurrence becomes possible (empirical affin
ity), thereby extending the scope of the I think from the bare manifold of appear
ances in intuition to the "thoroughgoing and lawful interconnection" of these same
appearances in one experience. Clearly, therefore, the identity of the I think made
possible by the synthetic unity of experience, far from being a distinct subject, is
simply the representation of the subject of intuition as being also a subject of expe
rience. And at the same time the unity of experience enables the subject to cognize
objects through outer appearances, it becomes possible for it to cognize itself as the
object through inner appearances (the empirical self) (chs 17-18).
The affinity of appearances thus gives the subject of experience its special cognitive
worth. When this subject attaches its analytic unity to representations, they do not
simply acquire the logical form of a universal to become concepts, as when combined
with the I think of the subject of intuition (ch 9-C). Instead, it relates these representa
tions to a universe in the literal sense: the thoroughgoing and lawful interconnection
of appearances that spatially and temporally differentiates and determines each in
relation to every other. Again, it is the same subject that converts representations into
concepts in the logical sense and confers on them the value of concepts of objects of
experience, the difference consisting in its deploying pure concepts of the understand
ing in addition to pure intuitions, analytic unity of apperception, and logical functions
of judgment. For example, when the subject, in its capacity as subject of experience,
attaches its analytic unity to the sensation of red (Kant's example at B133-4n dis
cussed in ch 9-C), it does not just convert it into a concept of red but into a concept of
something with a reality corresponding to the sensation that has extensive and inten
sive magnitude, stands in causal relations, and inheres in a substance. Thus does the
subject apply its cognitive faculties to "spell out appearances conformably to synthetic
unity in order to be able to read them as experience" (A314/B370-1).
Kant's primary focus in the Transcendental Deduction of the Categories was not
on the empirical truth of the judgments we form regarding the lawful intercon
nection of appearances but on this interconnection itself and the conditions of its
366 Kant's Anatomy of the Intelligent Mind
possibility. This in turn resolved into a focus on the conditions of the possibility
of experience, understood as the necessary synthetic unity of perceptions in one
apperception a priori. For while the vast majority of laws connecting appearances
are specific to their empirically given matter (the laws on which Kant focused in the
Critique of Judgment), the regularities through which these laws express themselves
are possible only in experience; and since there could be neither experience nor a
subject of experience unless the necessary synthetic unity of perceptions in one
apperception were already in place a priori, all lawful interconnection of appear
the unity of apperception" these concepts make possible so that apart from them,
"there would be no nature at all."
The metaphysical deduction of the categories (ch 11) makes clear that the
categories can have only such content as is derivable from logical functions of
judgment and no use except to fix the logical positions of concepts in judgments.
Insofar as cognitions in experience are expressed in judgments, it is possible to
see how the categories can relate to it and even how they could prove to be condi
tions of the possibility of judgments that relate to the objects of experience. Yet
it is one thing to maintain that the categories are conditions of the possibility of
judgments of experience but quite another to claim that they are "at the same time
conditions of the possibility of objects of experience" (Al 11), "the source of the
laws of nature, hence of the formal unity of nature" (A127), and, "as the original
ground of its necessary conformity to law (as naturaformaliter spectata), [that] on
which nature (considered merely as nature in general) depends" (B165). If Kant
had refrained from extending his claims for the categories to the objects we meet
with in experience, their laws, and nature itself, their Transcendental Deduction
would be no more than an epistemological argument to the effect that without
these concepts, we would lack the conceptual scheme (normative framework)
necessary to learn anything about the objects that appear to our senses through
our perceptions of them. But the textual evidence that Kant did not stop at that
is overwhelming. Time and again, one finds claims that, prior to all experience,
the categories not only make possible our cognition of objects but these objects
themselves; that they do not just enable us to represent regularities involving these
objects but prescribe actual laws to these objects; and that they do not merely
equip us to conceive objects as constituting a systematic totality insofar as they
stand under laws but themselves, collectively, constitute nature itself as a single
universal lawful whole comprising all objective realities. Kant did, to be sure, limit
the object-constituting role of the categories to the formal (spatial and tempo
ral) differentiation and determination of the existents we encounter in experience,
thereby excluding everything that requires acquaintance with the matter of what
appears in actual experience. But even if the categories are constitutive of objects,
laws, and nature only in their formal a priori regard as specified by his theory of
pure sensibility (space, time, and the formal side of everything in them), Kant left
no doubt that in his view, this regard comprises all that is essential to objectivity
as such. For at the heart of the transcendental deduction of the categories is his
thesis that objectivity can be fully explicated by the necessary synthetic unity of the
manifold in one apperception (A104-8, Al24-7, B163-5).
This phase of the argument of the transcendental deduction is quintessentially
psychologistic. It consists in establishing that it is impossible even so much as to
conceive an object of experience, law of nature, or nature itself except by means of
the transcendental notion of experience as the necessary synthetic unity of percep
tions (Section E), and so too the contents essential to its conception, including
the categories as universal representations of the synthesis in imagination of the
368 Kant's Anatomy of the Intelligent Mind
manifold of intuition in conformity to the logical functions (ch 11-C). For to prove
that any notion of objectivity is otherwise empty or unintelligible is already of itself
sufficient, first, to restrict the meaning and scope of application of "object," "law of
nature," and "nature" to the sphere of representation demarcated by categories as
conditions of the possibility of the (synthetic and analytic unity) of experience, and,
second, to demonstrate that the categories, as the basis of constitutive principles of
the possibility of experience are, for this very reason, constitutive also of the objects
of experience, its laws, and nature itself Kant's psychologism thus resolves the task
of demonstrating the objective validity of the categories completely a priori-their
transcendental truth (A221-2/B269) with respect to every existent encounterable
in experience-into the purely subjective endeavor of showing them to be a priori
conditions of the experience in which these objects are encountered:
and determination of the manifolds of both pure and empirical intuition that,
given pure space and time alone (prediscursive synthetic unity of apperception), is
impossible? Clearly, where all such differentiation and determination are lacking,
objects of experience, their laws, and nature itself are no more conceivable than
experience itself. Indeed, showing the categories to be the sole and entire basis of
the differentiation and determination of the manifolds of space and time is not
just the best but the only possible way of demonstrating both their own objectivity
and their status as that through which alone an object can be thought through any
other concept.
our thought of the relation of all cognition to its object carries with it some
thing of necessity since this object is regarded as what stops our cognitions
from being random or arbitrary but instead determined a certain way a
priori. For since they are supposed to relate to an object, they necessarily
agree with one another in relation to this object, i.e. they must have that unity
which constitutes the concept of an object. (A104-5)
What does this mean? In the first instance, it implies that the only representa
tions that can enable us to think an object for our representations are concepts,
since the necessary relation of appearances can never be given in intuition, only in
thought. But Kant did not stop there. To get to the point where he could equate the
object of appearances with that in the concept of which the synthesis of the mani
fold is united, he argued as follows:
[I]t is clear that since we have to do only with the manifold of our represen
tations, and since that X which corresponds to them (the object), because
it is supposed to be something distinct from all our representations, is for
us nothing, the unity that the object makes necessary can be nothing else
than the formal unity of consciousness in the synthesis of the manifold of
representations. We then say that we cognize the object if we have brought
about synthetic unity in the manifold of intuition. But this is impossible if the
intuition could not be produced through such a function of synthesis accord
ing to a rule that makes the reproduction of the manifold necessary a priori
and makes possible a concept in which this manifold is united. Thus we think
a triangle as an object in that we are conscious of the composition of three
straight lines according to a rule, in conformity to which such an intuition
can always be exhibited. Now, this unity of rule determines all the manifold
and limits it to conditions that make possible the unity of apperception, and
the concept of this unity is the representation of the object = X that I think
through the thought of the predicates of a triangle. (Al05)
determinations of other such subjects (3) in such a way that all directly or indirectly
interact (community). And how else can transcendental judgments do this, and do
so entirely a priori, except by bringing the categories into a determinative relation to
the one element in apprehended appearances contributed solely by the subject-the
manifolds of pure space and time? These concepts must therefore be supposed to
define a transcendental synthesis of imagination that serves to make good the com
plete want of differentiation and determination in the appearances apprehended in
intuition (the self-created problem posed by Kant's account of the unity of sensibil
ity)-at least to the extent necessary for this, their transcendental determination,
afterward to be carried to completion non-transcendentally by means of figurative
synthesis in conformity to mathematical and empirical concepts of space and time.
It is thus through transcendental synthesis that the categories attain objectivity in
the full, unattenuated sense of constituting the objects we experience and not just
our experience of them.
To this it may be objected: how can determinations founded on the categories,
which exist only in and through thought (the transcendental judgments predicating
these concepts), possibly be supposed to be constitutive of the objects met with in
experience, their laws, and nature itself ? I have argued that Kant's sole and entire
warrant for the claim that the categories are constitutive of objectivity in this unat
tenuated sense is that they, and they alone, are capable of conferring the spatial and
temporal differentiation and determination on appearances without which appear
ances in intuition are no more than a "blind play of representations, less than a
dream." The problem is that Kant made quite clear that the presence of appear
ances in sensibility in no way depends on the categories and that appearances
could still be given even in the absence of these concepts (B67, A89-91/Bl22-3,
Al 11, Bl45-6, AA 17 § 4679). So if appearances, as originally given in intuition,
are devoid of all spatial and temporal differentiation and determination, how can
mere thinking rectify this, whether through categorial predicates, mathematics, or
anything else?
This objection is apt to seem unanswerable to anyone who confounds Kantian
appearances with sensations (affections) and so supposes them to be present in sen
sibility through receptivity alone, prior to and independently of all spontaneity. But
we have found that for Kant, not even the matter of appearances can be equated
with sensations (ch 3-B), and that appearances, in both form and matter, are prod
ucts of a synthesis of apprehension in intuition that involves the spontaneity of pure
imagination and original apperception no less than the receptivity of sense (ch 3-E).
Appearances are, to be sure, anchored in transcendental reality insofar as their mat
ter corresponds to sensation and, through this, to things in themselves (ch 8). Yet,
in and of themselves, they are nothing more than a homogeneous manifold synthe
sized in imagination and united in apperception so as to exhibit the heterogeneous
manifold of a priori synopsis as the manifold of a single representation a priori,
all contained in the consciousness of that representation (ch 3-D and -E). The sub
ordination of appearances to the categories constitutive of objects of experience,
The A Edition Transcendental Deduction 373
their laws, and nature itself consequently consists simply in the addition of an extra
layer of a priori psychological synthesis: a categorial synthesis speciosa of produc
tive imagination overlaid atop the same faculty's pure synthesis of apprehension.
Whereas pure apprehension can yield nothing by itself but a scattered manifold
of spatial and temporal appearances in empirical intuition, when supplemented
by categorial synthesis, the form of every appearance apprehensible in empirical
intuition is determined a priori to become a spatio-temporally differentiated, deter
minate object (in the thought-the transcendental judgment-of the same subject
that intuits it: Section E).15 With both precategorial and categorial expressions of
the same spontaneity of imagination and apperception in place ahead of all actual
experience, the threefold empirical synthesis can serve its proper cognitive purpose
of revealing patterns in the co-occurrence and sequence of perceptions-which are
precisely the materials required by empirical understanding to form the judgments
(of experience), both ordinary and scientific, that determine ever more specifically
and precisely the objects that we already know to be appearing to us transcenden
tally, in a maximally general purely formal sense.
And here it should be remarked that from the transcendental standpoint delin
eated by Kant in the Analytic of Concepts, it is mere groundless illusion to suppose
that, a priori and "in themselves," the objects we cognize in actual experience are
vastly more determinate than our experience reveals, or can reveal, them to be. As
originally constituted by the transcendental synthesis founded on the categories, 16
these objects, their laws, and their systematic unity (community) in nature are, on
the contrary, far less determinate than they are in our empirical and mathematical
judgments. They have only the barest, most general, purely formal differentiation
and determination. Even matters such as the number of dimensions of space can
not be fixed until transcendental synthesis is supplemented by mathematical and
empirical synthesis (chs 6-A and 15-C). For all that is required of the categories'
transcendental synthesis is to bridge the gap between the completely undifferen
tiated, indeterminate manifold of pure apprehension and the differentiation and
determinateness requisite to make properly cognitive empirical and mathematical
spatial and temporal synthesis possible. It is therefore as the ground of this bridg
ing transcendental synthesis that Kant could claim that the understanding itself,
15 Alison McCulloch explicates the moral objectivity of the Critique of Practical Reason as
an additional layer set atop the purely intellectual objectivity superadded by the categories to the
imagination's synthesis of apprehension in intuition. This imaginative, well-argued thesis can be
found in "Freedom and Reason: Kant's Construction of Morality" (PhD dissertation, University
of Colorado, 2003).
16 Here and in the remainder of the chapter (unless otherwise specified), I am using "transcen
dental synthesis" in its broadest sense to include (I) the categorial synthesis intellectualis of the
manifold of a sensible intuition in general (chs 11-C and 14-B), (2) categorial synthesis speciosa in
relation to the manifold of pure formal intuition (chs 14-D and 15), (3) categorial synthesis spe
ciosa of empirical material intuition via transcendental schematism (ch 16), and (4) the synthesis
of transcendental judgment whereby categorial schemata are predicated of appearances via the
threefold empirical synthesis in principles of pure understanding (chs 17 and 18).
374 Kant's Anatomy of the Intelligent Mind
through its pure concepts, is the author not only of experience but its objects, their
laws, and nature itself, and mean what he said, no matter how "bold," "exagger
ated," "strange," or even "absurd" it may strike one.
The strategy adopted by Kant in the Transcendental Deduction for proving the
categories to be constitutive of experience and its objects should now be clear. His
analysis of the unity of sensibility, together with his commitment to a Humean-style
psychologism that permits nothing to be ascribed to the content of a representation
that its originating faculty is incapable of supplying (chs 2-E and 4-B), suffice to
make clear that the possibility of experience and cognition generally (mathemat
ics included) depends on a transcendental synthesis directed upon the otherwise
undifferentiated, indeterminate manifold of the space and time generated in pure
synthesis of apprehension. To prove that the categories, and they alone, are the
indispensable conceptual foundation for transcendental synthesis, Kant needed to
identify a middle term connecting them. This he found in the maximally general,
content-indifferent principle of the necessary synthetic unity in one appercep
tion of all the manifold according to laws. The only restrictions on the scope of
this unity are that the intuition be sensible and the laws of its necessary synthetic
unity be founded on sensibly conditioned intellection (i.e., discursive, not intuitive,
understanding: Bl38-9, Bl45-6). Otherwise, it is a matter of complete indiffer
ence whether appearances are synthesized in accordance with space and time or
other pure intuitions of a kind completely unknown to us, and whether the laws
are founded on the functions of judgments and categories listed in Kant's tables or
other logical functions and categories of a kind completely unknown to us (the pos
sibility of which can no more be denied than affirmed: Bl45-6, A230-l/B282-3).
And in order for the necessary synthetic unity in one apperception of all the mani
fold according to laws to function as the intermediary between the categories and
transcendental synthesis in Kant's proof of the objectivity of these concepts, he had
to establish that this subjective principle of self-consciousness is at the same time the
supreme principle of objectivity in all sensible intuition:
its representations, and indeed do so a priori, if it did not have before its eyes
the identity of its action which subjects all synthesis of apprehension (which
is empirical) to a transcendental unity, and first makes possible its intercon
nection according to rules. (Al06-8)
Kant's reasoning makes use of his thesis that the one constant through all
subjective variation in beings whose intuition is sensible and whose intellection is
sensibly conditioned is the I think: the recognition of the identity of conscious
ness in all of one's representations (termed the "analytic unity of apperception"
in the B Deduction). As a purely formal representation, devoid of all intuitive and
conceptual content, 17 the I think is indifferent to whether the representations con
cerned are apprehended in accordance with pure space and time or other forms of
intuition, and to whether the grounds of their necessary reproducibility and the
necessary interconnection that determines them as cognitions of an object are the
categories of substance, cause and effect, etc., founded on the logical functions of
categorical, hypothetical, etc. judgment or different pure concepts of understand
ing and logical functions. But it is not so much the I think itself that was crucial
for Kant's reasoning as "apperception as a capacity" (All 7n). For quite apart from
whether the identity of consciousness in all representations (the analytic unity of
the I think) is actually represented, not being able to represent it in the case of any
representations would be as much as to say that they "are for us nothing and do not
in the least concern us" (All6) (ch 9-B).
Now, the capacity to represent the identity of consciousness in all representa
tions does, to be sure, entail the synthetic unity of these representations in that
consciousness (for which unity of sensibility suffices: Section E). But if the repre
sentations united in my consciousness are at the same time to be cognitions of some
object, as is required for experience (Section F ), they must in addition exhibit that
transcendental affinity which is possible only insofar as "all appearances stand in
determination of the concepts or judgments to which they are applied with respect
to these functions. In the case of the logical functions characteristic of human
understanding, the determination of a recognitive concept as the concept of, e.g., a
substance means that its relation to its predicate is made irreversible: it necessarily
is always subject in relation to other concepts in categorical judgments, never predi
cate (ch 11-A). With its determination by this category, in other words, the concept
itself now necessitates its relation to other recognitive concepts: those to which it
can be related in judgment (which it has substance/accident affinity with) and those
to which it cannot (which it lacks this affinity with). And so too with the other cat
egories: the categorial determination(s) each recognitive concept or judgment has
necessitates its relation in judgments to every other similarly determined recognitive
concept or judgment (causal affinity, quantitative affinity, etc.).
Of course, a determination that confers necessary relation only with respect to
logical positions in judgments is a purely intellectual synthesis that leaves everything
else undetermined, including whether the relation is or even can be objectively valid
in experience or pure intuition What is required is not merely an intellectual affinity
of concepts and judgments but a transcendental affinity of appearances adequate
for experience and its objects. In particular, the system of relations produced by the
former brings together concepts and judgments with complete indifference to their
content and so does not seem !imitable in scope to those capable of functioning as
a basis for synthesis of recognition (to the exclusion of, e.g., moral, formal logical,
and hedonic concepts and judgments). The question therefore becomes this: how,
on transcendental grounds alone, can the scope of categorial determination be
restricted to recognitive concepts and judgments even while remaining indifferent
(and so oblivious) to their non-transcendental content (as is essential if the deter
mination is to qualify as transcendental), and moreover do so in such a way as to
introduce the transcendental affinity required for experience and its objects into the
manifold of apprehended appearances reproduced in accordance with empirical
and mathematical recognitive concepts and judgments?
The answer lies in the prior application of the categories to pure concepts
derived from the manifold apprehended in pure intuition a priori-the application
whereby alone pure concepts of the understanding can acquire the additional prop
erly transcendental validity of pure concepts of an object of a sensible intuition in
general ("fundamental concepts for thinking objects in general for appearances,"
Alll; ch 11-C). Through the simple act of attaching the original analytic unity of
apperception made possible by pure intuition (ch 9-C), every appearance that can
ever be apprehended becomes the source of a concept of its form. In the case of
beings with sensibility constituted like ours, attaching this analytic unity to them
gives rise to concepts of juxtaposition and succession. Because of the absence of all
differentiation and determination in the manifold of pure intuition, the concept of
juxtaposition or succession derived from any one appearance is so indeterminate as
to be indistinguishable from that derivable from any other. While this means that
the generic concepts of juxtaposition and succession that capture all and only what
The A Edition Transcendental Deduction 379
which it can produce neither judgments nor any other kind of representation (ch
10). Their systematically interconnected, unsunderable unity enters into every act
of judgment as its constitutive form. This is not to say that each individual judg
ment employs all twelve forms at once, since each is reciprocally exclusive with
respect to certain others. Nevertheless, given the formal constitution of our capac
ity to judge ( Vermiigen zu urteilen), we are conscious, even if only obscurely, that
any categorical judgment admits of qualitative and quantitative alteration (ch
10-B) and is combinable with any other in a hypothetical or disjunctive judgment,
and that assertoric judgment is alterable into a problematic or apodeictic judgment
(ch 10-C). The capacity we employ every time we judge thus serves to situate each
and every judgment we form within the complete systematic structure exhibited in
Kant's Table of Judgments.
Since pure concepts of the understanding are nothing else than the selfsame logi
cal functions of judgment conceptualized as determinations whereby to fix the logi
cal positions of concepts or judgments in judgments (ch 11-A), the systematically
interconnected, organically unified character of the Table of Judgments carries over
to the Table of Categories and so extends to everything into which these concepts
enter, including the imagination's transcendental synthesis speciosa. Accordingly,
whenever any act of transcendental synthesis is present in a consciousness, the
entire categorial structure of systematic interconnection is present as well. In the
case of beings with understandings constituted like ours, this means that no sub
stance can be represented except in conjunction with the representation of acci
dents, no accidents except in conjunction with a quantity, quality, and modality,
and no substance with accidents except as a member of a community comprising
all other substances constituted by direct and indirect causal connections linking
their accidents. Since their community is nothing other than the necessary synthetic
unity of apperception that Kant equated with nature (A216/B263), and since this
pure self-consciousness is present in every judgment of experience we form, the rep
resentation of the entire system of nature constituted by the categories as a total
ity (rather than as an aggregate) is contained, albeit only obscurely, in every such
judgment.
This, then, is how the categories enable our minds to represent nature itself
as the systematic totality constituted by the necessary synthetic unity of appear
ances interconnected according to laws. And just as pure understanding is defined
logically in accordance with the systematic interconnection of all logical forms as
the capacity to judge, the same faculty can be defined transcendentally through
its capacity for consciousness of the systematic interconnection of appearances in
imagination by means of the categories:
T he B Edition Transcendental
Deduction: Objective Unity of Apperception
and Transcendental Synthesis
T he primary difference between the 1781 A and 1787 B edition of the Transcendental
Deduction of the Categories is the latter's treatment of judgment and the clarifica
tion of its relation to the unity of apperception in§§ 19-21 (B140-6). Up to that
point, the 1787 exposition deals with the same topics in essentially the same way,
though in a clearer, more orderly fashion than the 1781 edition:
his decision the previous y ear in the MFPNS preface footnote considered in
Chapter 12-C:
In the preface to the new edition Critique of Pure Reason in 1787, Kant reaffirmed
that the revision concerns only the manner of exposition but "alters absolutely noth
ing fundamental as regards the theses and their grounds of proof " (xlii). Indeed,
far from disavowing the A edition version, he advised B edition readers to consult
it wherever lack of space prevented him from treating topics as fully in the revision.
The natural questions to pose regarding Kant's new version of the Transcendental
Deduction of the Categories are therefore these: If indeed the omission of any
explicit consideration of judgment analogous to§§ 19-21 of the B edition is neither
a defect nor an error in the 1781 exposition, how does its incorporation into the
1787 exposition reduce Kant's explanation of how the categories make experience
possible to almost a single inference? Does this reduction significantly alter our
picture of the Deduction as a whole? And is it true that there are no substantive
changes in the revision?
1 Kant seems to deny that judgments of perception can be expressed by means of the copula
in the Logic. But although the Logic is the later work, it was adapted from lecture notes and is
consequently less authoritative than PFM. Moreover, Kant had plenty of time to revise PFM if,
in the interval between PFM and the B edition Critique, he had changed his view in this regard.
The B Edition Transcendental Deduction 387
in the Critique of Judgment: estimates of beauty and ugliness are ineluctably sub
jective, as indeed are all reflective judgments inasmuch as the relation thought in
them is grounded not on the objective unity of apperception and the categories
of theoretical understanding but on a transcendental principle of adaptedness2
(ZweckmiifJigkeit) rooted in the (nondeterminative, merely reflective) faculty of
judgment itself, which is a source of no pure concepts of objects at all. Other non
objective judgments include analytic judgments, which concern concepts and their
content rather than the objects thought through them; the judgments of general
logic, which ignore content altogether and concern only the forms of concepts and
judgments; and those judgments of transcendental philosophy that relate not to
objects but to the subjective sources of their cognition (such as the judgments mak
ing up the subjective transcendental deduction account of the possibility of under
standing itself). It thus seems certain that Kant deemed any relation of discursive
representations a judgment, irrespective of whether or not the relation brings them
to the objective unity of apperception.
How then should one understand the B Deduction thesis that "a judgment is
nothing other than the mode of bringing given cognitions to the objective unity of
apperception"? The first thing to note is that it is the consequent of a complicated
conditional whose antecedent suggests another way it might be read. The clause "if
we investigate ...the relation of given cognitions in any judgment and distinguish
that relation as belonging to the understanding from relation in accordance with
laws of reproductive imagination" may point to a distinction not between judg
ments and nonjudgments but between two types of judgments: judgments that
belong to the understanding and bring given cognitions to the objective unity of
apperception and judgments that belong to reproductive imagination and merely
give discursive expression to customary associations between cognitions but with
out bringing them to the objective unity of apperception. Kant's grammar would
be at fault if this is the reading he intended, for he could certainly have been clearer.
But since the alternative runs contrary to everything he said about judgments else
where, it has to be the preferred reading.3
2 I here follow Ernst Mayr's suggested translation, which, on the whole, seems to me prefer
able to the more standard "purposiveness."
3 Against reading Bl41-2 as distinguishing kinds of judgments, one can cite a statement
later in the paragraph that certainly makes it look as if Kant's concern was to distinguish judg
ments from nonjudgments: "Only thus does a judgment arise from this relation, i.e. a relation
that is objectively valid, and is sufficiently distinguished from a relation of the very same repre
sentations in which there was merely subjective validity, e.g. conformably to laws of association"
(Bl42). Since taking this at face value contradicts virtually everything Kant said of judgment
elsewhere, I see no option but to construe it otherwise. One can do this by noting that judgments
of experience "belong to understanding" in a sense not true of judgments of perception: whereas
reproduction imagination operates independently of discursivity and does not require that its
deliverances be given discursive expression in judgments of perception, judgments of experience
are never possible nondiscursively because they depend on the categories, and the categories
directly apply only to concepts and judgments (as determinations of their logical positions in
judgments: ch 11).
388 Kant's Anatomy of the Intelligent Mind
If this is granted, the next step is to determine what precisely it means for a
judgment to "belong to the understanding." The explanation of the categories
Kant appended to§ 14 of Section 1 of theB edition Transcendental Deduction
here seems key: "They are concepts of an object in general whereby its intuition is
regarded as determined by one of the logicalfunctions of judgment," as, for example,
"by bringing the concept of body under the category substance, it is thereby deter
mined that its empirical intuition must always be considered only as a subject, never
as a mere predicate" (Bl28-9). Although I utilized this text to help flesh out the
metaphysical deduction of the categories in Chapter 11, the purpose for which Kant
seems principally to have intended it was to prepare the reader for the treatment of
judgment in § 19 of Section 2, where it helps to make clear that his thesis relating
judgment to the objective unity of apperception at Bl41-2 concerns neither judg
ments generally and as such nor a particular species of judgments Gudgments of
experience) in distinction from others Gudgments of perception), but the logical func
tions of judgment themselves. This is confirmed not only by the title of§ 19-"the
logical form of all judgments consists in the objective unity of the apperception of
the concepts they contain" -but also by the distinction between judgments "belong
ing to the understanding" and judgments that "conform to laws of reproductive
imagination" used to set up the thesis: both alike must, qua judgments, conform to
the logical functions that define the understanding as a capacity to judge since judg
ments, for Kant, are possible in no other way; but the only judgments truly belonging
to the understanding are those that result when their contents are determined con
formably with these logical functions by means of the pure concepts that have their
source in pure understanding and that faculty alone, as, for example, the use of the
category of substance to determine the subject concept of a judgment to be always
subject and never predicate.
With its true focus disclosed, the thesis of§ 19 emerges as the principle that cog
nitions in a judgment are brought to the objective unity of apperception by being
determined conformably with the logical functions of judgment. What clinches this
reading, in my view, is Kant's employment of this thesis in§ 20 to introduce the cat
egories into the argument ( their first mention in theB Deduction since the introduc
tory remarks in§ 15) to provide the "transcendental deduction in a single inference"
promised in the MFPNS preface footnote:
The manifold given in a sensible intuition necessarily belongs under the origi
nal synthetic unity of apperception because through the latter alone is unity of
intuition possible (§ 17).But that act of the understanding through which the
manifold of given representations, be they intuitions or concepts, is brought
under an apperception in general is the logical function of judgment (§ 19).
Thus, all the manifold, insofar as it is given in one empirical intuition, is deter
mined in respect to one of the logical functions, whereby it is brought to a
consciousness in general. Now the categories are nothing else than just these
functions of judging insofar as the manifold of a given intuition is determined
The B Edition Transcendental Deduction 389
nothing is thought in these concepts but logical functions of judgment, and their
only use is to irreversibly fix the logical positions of the components of judgments
(ch 11). Yet one finds nothing in the A Deduction like the clear distinction Kant
drew in B between the understanding's imagination-independent synthesis intellec
tualis and the synthesis speciosa of productive imagination that counts as transcen
dental "if it pertains merely to the original synthetic unity of apperception, i.e. that
transcendental unity which is thought in the category" (Bl51). He seems instead to
have elected to downplay the difference, in effect folding the former into the latter.
For example, on the one occasion that he correlated functions of synthesis to cat
egories, he presented the latter in the character of time determinations (Al11-2),
which, strictly speaking, is true not of the categories themselves, in the intellectual
purity of their synthesis intellectualis, but only of the corresponding schemata gen
erated by the imagination's transcendental synthesis speciosa (Al42-5/Bl81-4).
The transcendental synthesis of the imagination figures prominently in Section
3 of the A Deduction as the means whereby alone appearances can be brought
together with the categories in their capacity as conditions for synthetic unity of
apperception (AllS-9, Al23-5). It is the action of this synthesis that Kant des
ignated a function in Section 3 (Al23, and Al24),5 never the action of categorial
understanding in isolation. Thus, the overall impression one gets from the A edi
tion exposition is that Kant was so single-mindedly focused on the aspect of the
Deduction concerned with the possibility of subsuming empirically given appear
ances under the categories, thereby preparing the way for the Analytic of Principles
(the principles of the subsumption of appearances under the categories), that he
4 I have altered this from§ 13 because Kant only mentions functions in that section in order
to say that appearances stand in no need of them to be given in intuition, whereas B128-9 of
§ 14 relates directly to the premise in the argument of B143 that references Section 1 of the
Transcendental Deduction. Evidently, Kant (or his ty pesetters) either wrote the wrong number
or decided at a later point to divide§ 13 but failed to correct the reference at B143. This could
also explain Kant's reference to§ 25 at B136n when§ 26 seems far more probable (see ch 5-C).
5 "Function" is also used in reference to the action of apprehending imagination at A120n.
390 Kant's Anatomy of the Intelligent Mind
left its most crucial premise somewhat obscure: that and how the categories, merely
as intellectual modes of combination (synthesis intellectualis), make possible a gen
uinely objective synthetic unity of apperception right in sensibility itself. And to
the extent this remains unclear, everything in Kant's philosophy that depends on
it-the possibility of experience and its objects, nature and its laws, mathematics, and
synthetic a priori judgments generally-becomes obscure as well.
The B edition version of the Transcendental Deduction remedies the obscurity in
A but not in so radical a manner as to invalidate Kant's contention that the need for
revision lay merely in the manner of exposition rather than the substance of the proof.
In both§§ 19-20 of the B Deduction and the portions of Section 2 of the A Deduction
concerned with association and its need for an objective ground (affinity), the focus
is on what differentiates an objective from a merely subjective relation of cognitions.
Insofar as cognition (synthesis of recognition) is impossible without a concept, the
combination of distinct cognitions is treated as essentially a logical affair, since the
only way one concept can be combined with another is in a judgment (in A, Kant
seems to take this for granted, as an evident consequence of the metaphysical deduc
tion, whereas in B, he makes it explicit that his concern is "the relation of given cogni
tions in any judgment"). Each concept is combinable with any other in categorical
judgments, and the resulting judgments can be conjoined in hypothetical or disjunc
tive judgments, without thereby fixing their logical relations to one another: concepts
related as subject and predicate in one judgment can be related as predicate and sub
ject in another; their quantities and qualities are equally reversible; and the resulting
judgments can in turn be related as antecedent and consequent in one hypothetical
judgment and as consequent and antecedent in another (ch 10-B and -C). What the
explanation appended to Section 1 of the B edition Transcendental Deduction (B128-
9) makes explicit that was left implicit in A is that the categories, albeit concepts of
objects as given through (the necessary connections objects confer on) appearances,
have no immediate application to appearances at all, but instead serve only to elimi
nate the logical indeterminacy of discursive representations (universals) by fixing their
logical position in judgments (this, and this alone, is the necessity the categories add to
the synthetic unity of the manifold generated by the threefold synthesis-the necessity
Kant equated with objectivity).
The distinction between the immediate and indirect application of the categories
lies at the root of the distinction between pure understanding's synthesis intellectua
lis and the imagination's synthesis speciosa (figurative synthesis) around which Kant
structured much of the B Deduction, with the former the focus of §§ 19-23 and the
latter the focus of §§ 24-26. The distinction takes on transcendental significance, as
relating to the possibility of experience and its objects, insofar as the scope of the pro
cedure described at B128-9 is restricted to all and only concepts employable in sensible
syntheses of recognition (ch 13-H). It is under this restriction that the synthesis intellec
tualis emerges as the basis of the objective unity of apperception whereby relations of
recognitive concepts in judgment alone can acquire the element of necessity requisite
to elevate them from mere discursive expressions of the deliverances of associative
The B Edition Transcendental Deduction 391
As noted in the previous section, the implication Kant would have us draw from
his claim in § 19 that "a judgment is nothing other than the mode of bringing
given cognitions to the objective unity of apperception" (Bl41) is made clear by its
one-sentence summation in§ 20 as a premise in the argument aimed at providing
at least the synthesis intellectualis of the categories with its transcendental deduc
tion: "that act of the understanding through which the manifold of given represen
tations, be they intuitions or concepts, is brought under an apperception in general
is the logical function of judgment" (Bl43). As I understand it, this refers to the
role performed by logical functions examined in Chapters 10 and 11. In Chapter 10,
it emerged that the analytic unity of the I think suffices only for any sensible rep
resentation to be converted into a universal, but not for the means necessary to
relate one universal to another.The logical functions enumerated in Kant's Table of
Judgments make it possible for every possible concept to be combined (synthesized)
with any other to form a (quantified, qualified, assertoric categorical) judgment
(ch 10-B), and for each of the resulting judgments to be combined (synthesized)
with any other to form a (hypothetical or disjunctive) judgment (with problematic
components and the possibility of assertoric modality) (ch 10-D). Insofar as the
I think necessarily attaches to each of these representations and all are also united
6 The only objective unity not due to the categories and that does not take discursive form
is that of pure space and time (ch 5-D). But since the appearances within the objective unity of
space and time are objectively indeterminate ("scattered and single in themselves"), the unity of
pure space and time is not an exception to the rule that all objectivity within the field of appear
ances derives from pure concepts of the understanding.
392 Kant's Anatomy of the Intelligent Mind
in the (prediscursive) original synthetic unity that the I think presupposes (ch 9-B),
judgment counts as a synthesis through which every discursive representation (con
cept or judgment) can be united with any other in one and the same consciousness
a priori and so constitutes a synthetic unity of apperception specific to discursive
representations (the synthetic unity of the thinking subject that even analytic judg
ments presuppose: B13ln).Moreover, insofar as these representations represent the
sensible manifold and so are capable of functioning as rules in recognitive synthe
sis, this same synthetic unity of apperception extends to intuitions as well. And
it is in this sense, at least in the first instance, that logical functions of judgment
collectively bring "the manifold of given representations, be they intuitions or con
cepts, ...under an apperception in general" (Al43).
This synthetic unity of apperception is, however, not yet the objective synthetic
unity of apperception referred to in § 19. Synthesis by means of the logical func
tions of judgment serves only to relate discursive representations (concepts and
judgments), but does not confer necessity on their relation. For as Kant reminded
us in § 14 (Bl28-9), they can equally well be set in the reverse relation, whereas
necessity is precisely what any synthesis of the manifold must have in order to count
as objective, or, in other words, only the necessary synthetic unity of all the mani
fold in one apperception qualifies as an objective unity of apperception (§§ 17-18).
As emerged in Chapter 11, this want can directly be made good only if the relation
of discursive representations to one another in judgments is determined conform
ably to the logical functions themselves in the manner described at Bl28-9 (as well
as PFM 324, another post-A Deduction text of similar purport). For determining
a concept so that it can only function as subject of judgments, never as predicate,
is to add precisely that element of necessity which otherwise is lacking from its
relation to other concepts. Insofar as these determinations are applied systemati
cally to all possible discursive representations, the synthetic unity of apperception
those representations constitute by virtue of the logical function of judgment alone,
without these determinations, will be transformed into a necessary synthetic unity
of apperception of precisely the kind required for objective unity of apperception.
Moreover, insofar as the discursive representations comprised in this appercep
tion are recognitive, their objective unity will extend to the manifold synthesized
through them (as rules of synthetic unity), producing the necessary synthetic unity
of all the manifold in one apperception (chs 13-H and 15). And since the metaphysi
cal deduction had shown the categories to be "nothing else than just these functions
of judging insofar as the manifold of a given intuition is determined in respect to
them," the conclusion follows in § 20 that "the manifold in a given intuition neces
sarily stands under the categories" (Bl43).
What is most notable about this proof is that it is specific to the synthesis intel
lectualis of the categories. In the A Deduction, Kant seems to have been so focused
on the goal of showing that the categories have sensible sense and meaning that
instead of emphasizing the distinction between the purely intellectual synthesis
they immediately define and the transcendental synthesis of imagination by means
The B Edition Transcendental Deduction 393
shown from the mode in which empirical intuition is given in sensibility that
the unity of intuition is none other than that the category prescribes to the
manifold of a given intuition in general according to § 20 preceding. In that
way, the goal of the deduction will for the first time be attained, in that its a
priori validity in respect of all objects of our senses is explained. (Bl44-5)
Before proceeding to the second part of the B Deduction, mention should be made
of Kant's use of the example "Bodies are heavy" to illustrate the necessity the cat
egories confer on judgments. He made a point of emphasizing that determination
conformably to the categories does not make the relation of "heavy" to "body"
necessary. The only necessity the categories impose is the kind that determines that
"body" be used in judgments only as subject, never as predicate, and that "heavy"
must always be treated as predicate in relation to subjects so determined. Categorial
necessity does no more than ensure that whenever the judgment "Bodies are heavy"
is made, it partakes of the objective unity of apperception and thereby has relation
(signified by the copula "is") to a reality capable of determining it as true or false
completely independently of the empirical psychology of the judging subject ("Only
394 Kant's Anatomy of the Intelligent Mind
thus does a judgment arise from this relation, i.e. a relation that is objectively valid,
and is sufficiently distinguished from a relation of the very same representations in
which there was merely subjective validity, e.g. conformably to laws of association,"
B142). Everything else is left contingent. For it is one thing for a judgment to par
take of a necessary relation of appearances that confers on it relation to objective
reality, but quite another for the connection of concepts or judgments thought in it
to actually hold of that reality, much less do so of necessity. Empirical truth cannot
be judged through mere transcendental predicates and their synthesis intellectualis.
The actuality and necessity of relations such as that thought in "Bodies are heavy"
must be assessed non-transcendentally or not at all.7
A similar point can be made with regard to necessary relations of concepts in
judgments such as "2 + 2 = 4" that owe their necessity to constructions in pure intu
ition (ch 6) and not to the categories. By applying the synthesis intellectualis of the
categories of quantity to this judgment, one merely converts it into an irreversible
plurality relation such that (1) four can never be thought as plural in relation to two,
(2) two must always be thought as plural in relation to four, and (3) the sum of two
and two must be thought as plural (or not) in relation to all and only those (logical)
quantities to which four has (or lacks) these relations (ch 11-A). It is in this sense
that the categories of quantity bring the judgment to the objective unity of apper
ception. They do not, however, necessitate that four will actually result from adding
two to two, much less that it is the necessary result of performing this addition.
Indeed, since addition, counting, and all matters numerical are, for Kant, bound up
by content with juxtaposition and succession (chs 6-B and 15-E) and so with a priori
sensible synthesis (synthesis speciosa), the necessity founded on the synthesis intel
lectualis of the categories of quantity leaves the actuality and necessity of numeri
cal relations completely undetermined. All it does is necessitate that whenever the
judgment "2 + 2 = 4" is made, it will partake of the objective unity of apperception
and so have a relation to reality that is completely independent of memory, asso
ciation, human senses, and everything else pertaining to the judging subject apart
from objective apperception (although, in the case of pure mathematics, only the
formal side of objective reality is ever in play). Thus, like the heaviness of bodies,
the actuality and necessity of the relation between four and the sum of two and two
must be determined sensibly or not at all (indeed, mathematically-intuitively, by
7 Kant's contributions to the metaphysical foundations of the science of nature show that
he did not rule out the possibility of necessary relations pertaining to objects that can only be
cognized through empirical concepts as well as certain "fundamental experiences" (Al 71-2/
B213; also A847-8/B875-6, CJ 181, MFNS 469-70 and 478; ch 1). These determinations do
not count as transcendental, however, for as that which precedes and makes experience and its
objects (objective unity of apperception) possible, the transcendental cannot incorporate the
least empirical content or build on actual experiences in any way. So, whether or not "Bodies are
heavy" can be judged metaphysically and so qualify as a necessary relation between its concepts,
we can be certain that neither the actuality nor the necessity of the relation admits of being
assessed transcendentally.
The B Edition Transcendental Deduction 395
Space and time are valid as conditions of how it is possible for objects to
be given to us and no further than for objects of the senses, and so are valid
only for objects of experience. Beyond these bounds, they represent nothing,
for they exist only in the senses and have no actuality outside of them. The
pure concepts of the understanding are free from this limitation and extend
to objects of intuition in general, be that intuition like ours or not, so long
as it is sensible and not intellectual. However, this wider extension beyond
our sensible intuition is no help to us. For they are then empty concepts of
objects through which we can never judge whether or not the objects are
possible, mere forms of thought without objective reality because we have
no intuition at hand to which the synthetic unity of apperception that these
concepts alone contain could be applied and an object thus determined. Only
our sensible and empirical intuition can confer sense and meaning (Sinn und
Bedeutung) on them. (Bl48-9)
The most striking presentational difference between the treatments of the transcen
dental synthesis of imagination in the A and B edition Deductions is the far greater
prominence given to pure space and time in the latter. Indeed, of the three passages
where pure space and/or time figure in the A Deduction, none deal with transcen
dental synthesis. The first is the "general observation" at A98-9 that because all
representations belong to inner sense, all cognitive representations are subject ulti
mately to time, the form of inner sense, and so it is in time that all cognitive represen
tations have to be ordered, connected, and brought into relation. The second comes
two paragraphs later when Kant argues that without the spontaneity of an a priori
synthesis of apprehension of imagination, we could have a priori representations
396 Kant's Anatomy of the Intelligent Mind
neither of space nor of time (A99-100). Since Kant did not mention transcendental
synthesis in this connection and never again mentioned a priori synthesis of appre
hension in the Deduction (or anywhere else to my knowledge), there is no reason to
identify them. Indeed, there is every reason not to, given that Kant's first mention of
the transcendental synthesis of imagination relates it not to apprehension but to the
reproducibility of appearances (A102), and whose product he characterized as the
categorially determinate, necessary synthetic unity of appearances (Al 18-9), which
is the very antithesis of the "scattered and single" manifold of appearances that
issues from synthesis of apprehension (A120). Finally, no mention of or allusion
to transcendental synthesis is evident in the third passage in which space and time
figure prominently, where they are characterized as the "purest objective unity" and
claimed to be "possible only through the relation of intuitions to transcendental
apperception" (A107). And since transcendental apperception is treated as distinct
from the transcendental synthesis of imagination later in the Deduction (Al 17-19),
it seems reasonable to conclude that the claim at A107 has nothing directly to do
with the latter.
Things are different in the B Deduction. The transcendental synthesis of the
imagination is introduced in § 24 as synthesis speciosa (figurative synthesis) inso
far as it "pertains merely to the original synthetic unity of apperception, i.e. that
transcendental unity which is thought in the category" (BISI). This synthesis is the
principal focus of the B Deduction from§ 24 to B165 in§ 26, and it is during the
course of its discussion, at the culminating point of the second and final stage of
the transcendental deduction of the categories, that Kant introduces the pure space
and time of the Transcendental Aesthetic into the argument (Bl 59-63). Their intro
duction includes a significant amendment at B160n to the account given of them
in the Transcendental Aesthetic. There, Kant treated the unity of the manifold in
space and time as if the intuitions in which space and time are first given are prod
ucts purely of the senses; as newly amended, this unity belongs not to the form of
intuition but to the formal intuition in which pure space and time are "first given as
intuitions" by means of a synthesis of the manifold offered by sense that "does not
belong to the senses" (B161n). Any temptation to identify the synthetic unity of the
formal intuitions of space and time with transcendental synthesis speciosa should,
however, vanish in the face of Kant's insistence that "the unity of this a priori
intuition belongs to space and time and not to the concept of the understanding,"
for he everywhere expressly characterized the unity of transcendental synthesis as
deriving from the synthesis intellectualis of the categories (ch 5-C).8 Thus, of all that
distinguishes these two original synthetic unities of intuition, the most essential
difference is that transcendental synthesis speciosa incorporates the synthesis intel
lectualis of the categories, whereas formal intuition does not.
8 The categories can effect unity of the manifold only indirectly and discursively, via the deter
For this as well as all of the other reasons laid out in Chapters 3-5 and 7, the
evidence seems overwhelmingly in favor of equating the formal intuitions of space
and time of§ 26 with the pure space and time ofthe Aesthetic and against equating
them with the transcendental synthesis speciosa responsible for yielding "an intu
ition corresponding to the categories" (Bl51). How then do the two relate? In the
first part of the B Deduction(§ 20), Kant demonstrated that there can be no objec
tive unity of apperception of the manifold of intuition apart from the categories. It
is precisely this unity that is lacking from the "scattered and single" representations
comprised within pure (formally intuited) space and time produced by the pure
synthesis of apprehension in intuition. Not only do these representations lack the
necessary relations that Kant showed to be essential to objectivity (§ 17), they are
devoid of relation of any kind and so are both completely undifferentiated from
and undetermined in relation to one another(ch 4). And this suggests the following
relation between the formal intuitions of space and time and transcendental syn
thesis speciosa: the latter introduces into the manifold of the former precisely that
ries. It may be recalled from the previous section that judgment, for Kant, is a syn
thesis which ensures that any discursive representation(representation to which the
9 A suggestive passage in this regard is the following: "[U]nder the title of a transcendental
synthesis of imagination, the imagination exercises this action upon the passive subject whose
capacity it is, whereupon we justifiably say that inner sense is thereby affected. So far are apper
ception and its synthetic unity from being the same with inner sense that the former, as the source
of all combination pertaining to the manifold of intuitions in general under the name of the cate
gories, instead precedes all sensible intuition of objects in general, whereas inner sense is the mere
form of intuition, but without combination of the manifold in it, and so containing no determi
nate intuition at all, which is possible only through consciousness of the determination of the
manifold through the transcendental action of the imagination (synthetic influence of the under
standing on inner sense) which I have termed figurative synthesis" (Bl53-4). Although Kant
characterizes sensible intuition as passive in contrast to the spontaneity of understanding and
even terms it a form of intuition (the notion of formal intuition having not yet been introduced),
this is most probably only a comparative passivity. "Understanding" here is expressly specified
as categorial, and there is abundant evidence, including in the B Deduction itself (B136n, Bl40,
and Bl60-ln), that the nondiscursive pure space and time of the Transcendental Aesthetic to
which synthesis speciosa is applied are products not only of receptivity but also of the sponta
neity of imagination and the prediscursive unity of apperception (chs 3-5). Pure intuitions are
therefore to be understood as "passive" only in the sense indicated in the text: they are devoid of
all categorial determination and, in particular, that combination (Verbindung) requisite to deter
mine their manifolds conformably with the objective unity of apperception . It is true that, on
my reading, one has to understand the unity and combination (Zusammenfassung) attributed to
formal intuitions in the Bl60-l footnote and the determination of sensibility by the understand
ing mentioned there as prediscursive and therefore distinct from the unity, combination, and
determination of sensibility by categorial understanding of Bl53-4 (also Bl51-2). By the end of
the next chapter, however, it should be sufficiently clear how precisely the input of transcendental
synthesis speciosa differs from its output to permit the reader to comprehend the doctrine that is
otherwise apt to be obscured by Kant's shifts in terminology and levels of explanation (especially
the shifts from text to footnotes at B133-4n, B136n, and Bl60-ln).
398 Kant's Anatomy of the Intelligent Mind
analytic unity of the I think attaches) can be united with any other in one and the
same consciousness and so results in a discursive synthetic a priori unity of apper
ception that comprises not only all concepts and judgments but also, by extension,
everything in intuition thinkable through them . This synthetic unity cannot, how
ever, qualify as an objective unity of apperception since the latter is possible only
through the necessary relation of the representations concerned in it, and logical
functions are incapable of imposing the least necessity on the synthesis of concepts
or judgments through them (a concept related to another as subject to predicate
can always relate to it in the reverse fashion, as predicate to subject; a judgment
related to another as ground to consequence can always related to it in the reverse
fashion, as consequence to ground; and so on for the remaining logical functions).
By necessitating that concepts or judgments relate to one another in only one of the
ways judgment permits, the categories convert the nonobjective synthetic unity of
consciousness yielded by logical functions into an objective synthetic unity.
As with the logical functions of judgment, the prediscursive a priori synthetic
unity of the manifold in one apperception that results from pure space and time is
not an objective unity of apperception because this manifold is devoid of all rela
tion, including the necessary relations constitutive of objectivity.1° Categorial synthe
sis speciosa makes good this want by subjecting the inexhaustible infinity of spaces
and times possible within formally intuited space and time to a system of necessary
relations whereby each is uniquely differentiated from and completely determined in
relation to every other. By incorporating the resulting objective unity of appercep
tion comprising all spaces and times into ordinary recognitive concepts and judg
ments, the latter can then serve as a vehicle for introducing into the appearances
synthesized conformably with them precisely the necessary interrelations requisite to
bestow relation to an object on these appearances (the transcendental affinity of the
A Deduction). Thus, it is by conferring objective unity of apperception on the mani
fold of pure (formally intuited) space and time in transcendental synthesis speciosa
(ch 15), and then extending this unity to manifold of empirical intuition by means of
transcendental schemata (ch 16), that the categories are able to earn the status Kant
accorded to them as conditions not only of our experience of objects but also of the
objects themselves we experience, their laws, and nature itself.
In his influential 2002 article comparing Kant's conception of logic with Frege's,
John MacFarlane argued that generality was a defining feature of logic for Kant,
but formality not (or, as Kant might have put it, the former is analytic to the concept
10 Although space and time themselves count as objective unities of apperception, the mani
fold in them, as undifferentiated and determinate, does not (see note 6 above). It is this want that
the transcendental synthesis speciosa makes good.
The B Edition Transcendental Deduction 399
u "Frege, Kant, and the Logic in Logicism," Philosophical Review 111, no. 1 (2002): 25-65.
Although MacFarlane does not use the analytic-synthetic distinction, it is clear that he has in
mind the kind of explicative versus amplificative proposition that Kant would have character
ized in terms of the distinction, e.g.: "Generality is [not just] a part of Kant's characterization of
logic, . . . it is the whole," while formality "is for Kant merely a consequence of logic's Generality,
not an independent defining feature" (44). Also: "Kant's claim that general logic is Formal is a
substantive thesis, not an attempt at 'persuasive definition."' (48).
12
"On Kant's view, 'the only use which the understanding can make of these concepts is to
judge by means of them' (A68/B93). A 'concept' that could not be used in any possible judgment
would have no objective significance, no semantic content, at all." Ibid., 50.
13
"For Kant, what distinguishes a judgment (which is capable of being true or false) from a
mere subjective association of representations (which is not) is that in a judgment, the representa
tions are claimed to be 'combined in the object' (B142)." Ibid., 50.
400 Kant's Anatomy of the Intelligent Mind
way of understanding how Kant's conception of logic differs from Frege's, namely,
the latter presupposes objective unity of apperception, whereas the former does not.
To understand why, note first that according to MacFarlane, Frege's concep
tion of logic counts as general yet non-formal because it is a logic that includes
such purely general, abstract objects (semantic contents) as concept extensions and
the True. As I read Kant, it is not the notion that these are objects that he would
find problematic, but rather the claim that they are logical objects and, more par
ticularly, that they have a properly logical generality. Kant was sufficiently familiar
with higher mathematics to appreciate that its symbolic constructions were likely
to become ever more abstract, general, and purely intellectual as time went on, and
so ever farther from the ostensive constructions of Euclidean geometry and simple
arithmetic (ch 6). Yet according to the analysis of objectivity in the Transcendental
Deduction, no matter how remote mathematics may come to seem from everything
sensible, however independent of the sensible, it can only have objects and demon
strate synthetic a priori objective identities (chs 2-A and 10-C) if there is objective
unity of apperception. T his means that mathematics, by contrast with what Kant
deemed logic properly so called, is subject to the conditions that make such apper
ception possible: the philosophical-discursive determinations of the manifold of
pure sensible intuition effected through categorial synthesis speciosa (Sections B
and D). T hus, from the Kantian point of view, Fregean mathematical logic, pre
cisely because it does deal with genuine symbolically constructible objects and is
able to employ its constructions to expand cognition synthetically yet a priori, nec
essarily belongs to mathematics rather than to logic.
Logic, as Kant conceived it, has no objects and yields no synthetic cognition14
because it deals with discursive understanding in abstraction from the conditions
for objective unity of apperception (ch 6-E). Yet contrary to the usual supposition,
this does not mean that general logic has no grounding in original apperception
at all. For although it is true that it is prior to and independent of the categories,
apperception, as we have seen, is not intrinsically objective and so does not itself
always presuppose the categories. Indeed, on Kant's account, pure general logic
incorporates two distinct noncategorial, nonobjective guises of apperception into
its representations: (i) the analytic unity of the I think made possible by pure space
and time as conditions of the prediscursive unity of sensibility (ch 9-B); and (ii) the
nonobjective, noncognitive synthetic unity of thought itself (the judging subject)
that results insofar as the logical functions of judgment make it possible to com
bine any concept with any other in a judgment, and any judgment with any other
in hypothetical and disjunctive judgments as well as inferences (ch 10). T hus, in the
14 This, however, is not to say that logic is analytic. Analytic judgment is as much concerned
with the content of concepts as synthetic (ch 2-A) and so has no more place in formal general
logic than synthetic judgment does. Moreover, analysis presupposes synthesis, and where there is
no manifold and can be no synthesis, there also can be no contents to analyze (ch 13-A)-all of
which is just another way of saying that general logic, for Kant, is essentially formal.
The B Edition Transcendental Deduction 401
final Kantian analysis, the reason Fregean mathematical logic cannot rightly be
termed "logic" comes down to where it falls in the development of original apper
ception from (1) mere unity of sensibility (the prediscursive synthetic unity of the
intuiting subject) to (2) the analytic unity of concepts to (3) the nonobjective, non
cognitive unity of thought effected through logical functions of judgment (the syn
thetic unity of the thinking subject), and finally to (4) the objective unity effected
through the categories (the synthetic unity of the experiencing subject): mathemat
ics of all kinds, including mathematical logic, involves all four, whereas general
logic, as Kant conceived it, is prior to and independent of the fourth.
So what of the generality Frege ascribed to mathematical logic? From a Kantian
point of view, it is not maximally general in the sense the properly logical needs to
be because it does not (cannot) abstract from the transcendental synthesis speciosa
in productive imagination of the sensible manifold whereby the objective unity of
apperception is brought to the manifolds of pure space and time. Only that can
count as logical which permits that abstraction, with the consequence that logic,
as conceived by Kant, is more general than Fregean mathematical logic-indeed,
so general as to altogether exclude the relation of the understanding either to the
sensible manifold or to the representations furnished by any other faculty. Since
Kant denied that the understanding is ever, by itself, a source of representable con
tent (the matter for thought, over and above mere logical form), this is just to say
that Kantian logic is by its very nature (analytically) a purely formal science and
indeed the only such science possible. Thus, Kant's science of pure general logic is
no more comparable asformal logic with Fregean mathematical logic than it is with
geometry or algebra.15
MacFarlane's study of the relation between Kant and Frege overlooks Kant's
assertion that the possibility of logic, like that of mathematics and natural science,
falls within the remit of the (subjective) transcendental deduction of the categories
(B131, B133--4n). Instead, he takes his cue from Frege, for whom objectivity must
be understood as excluding subjectivity and logic is objectivity at its maximally
general. But is it irrelevant that these propositions too are thoughts that have to be
understood, that understanding requires an intelligent mind, and so a conscious
subject capable of thinking them? The Kantian response, as I understand it, seems
quite measured: how can we claim competence to answer this question unless we
have first investigated the anatomy of the intelligent mind itself, in a transcendental
psychological Analytic of Concepts? His inquiry has no more striking result than
its explication of objectivity as the objective unity of apperception, a notion that
cannot but seem paradoxical because it unites objectivity with its antithesis: pure
self-consciousness, the most quintessentially subjective notion of all. Frege never,
to my knowledge, addressed the question of whether the objectivity of logic might
15 Since there is little if anything to say about these forms that had not already been said by
Aristotle, it is no surprise that Kant regarded the science of logic, as he understood it, as long
since complete.
402 Kant's Anatomy of the Intelligent Mind
be compatible with at least this one species of subjectivity. Nor, a fortiori, did he
consider whether apperception might offer a more fundamental way of demarcat
ing logic from mathematics than the criteria he saw fit to employ. To do so, how
ever, is to confront the possibility that far from superseding Aristotelian logic as
Riemannian geometry superseded Euclidean (ch 6-A), Fregean mathematical logic
is simply a new domain of mathematics (objective synthetic a priori cognition),
altogether different from logic properly so called, which transcendental analysis
exhibits as maximally general yet nonobjective, hence purely formal by definition,
and so prior to the distinction between synthetic and analytic.16
16
Does this mean that Kant would fault mathematical logicians for ignoring their debt to
transcendental philosophy? No more than he faulted other mathematicians for not making use
of the transcendental investigation of the possibility of their science. Since much that is essen
tial to the representation I think in the context of transcendental logic is clearly inessential to
the business of logicians, Kant would probably not only countenance but encourage them to
strip away everything from pure self-consciousness that transcendental logicians cannot, so as
to arrive at the fully abstracted, maximally general, strictly formal notions required by their sci
ence. Yet their exclusion of psychology from logic must then be recognized as predicated on an
implicit quid pro quo: logicians must defer to the practitioner of transcendental psychologism to
explicate the fundamental representations their science presupposes. For Kant would surely have
rejected any inference from the fact that logic is not beholden to the theory of self-consciousness
that self-consciousness itself is not an essential presupposition of that science. See UU ch 2-E-l
for a fuller discussion.
{ 15 }
A Category-by-Category Elucidation of
the Transcendental Synthesis Speciosa of
Pure-Formal Intuition
1 This use of 'material' relates to the empirically apprehended matter of appearance and con
trasts with 'formal'. It does not concern corporeality, which, as spatial, involves the form of
appearances as well as their matter.
2 Assuming the expression "sensible intuitions in general" encompasses pure intuitions, Kant
made essentially the same point in the Critique: "pure concepts of the understanding, in com
parison with empirical intuitions (indeed sensible intuitions in general), are completely heteroge
neous and can never be met with in any intuition ...since no one will say that, e.g. causality, can
be intuited through sense and be contained in appearance" (Al37-8/Bl 76-7).
A Category-by-Category Elucidation 405
deduction without its being necessary to explain how each functions in transcen
dental synthesis speciosa?
The attempt to answer is complicated by a related question: how does the tran
scendental synthesis speciosa of imagination of the Transcendental Deduction of
the Categories relate to the transcendental schematism of the Analytic of Principles?
Though Kant attributed the schematism to the transcendental synthesis of imagi
nation (Al45/Bl85), it does not follow that he had the action of schematism in
mind in the Deduction. Neither schematism nor schemata are ever mentioned in the
Deduction, and while it may seem legitimate to equate schematism with transcen
If all our synthetic judgments are analyzed so far as they are objectively valid,
it will be found that they never, as commonly supposed, consist of mere intu
itions connected in a judgment through comparison, but that they would be
impossible were a pure concept of the understanding not added to the con
cepts abstracted from intuition, under which those concepts are subsumed,
and so connected for the first time in an objectively valid judgment. Even the
judgments of pure mathematics in their simplest axioms are not exempted
A Category-by-Category Elucidation 407
from this condition. The principle that a straight line is the shortest between
two points presupposes that the line is subsumed under the concept of magni
tude which is certainly no mere intuition but has its seat in the understanding,
and serves to determine the intuition (of the line) in regard to the judgments
which can be made about it in respect of its quantity, namely of plurality (as
iudicia plurativa), in that under them it is understood that a plural homoge
neous is contained in a given intuition. (PFM 301-2; also A733/B761)
Kant's frequent use of examples drawn from pure mathematics in the Transcendental
Deduction (including in discussions relating to transcendental synthesis speci
osa: Bl54--5) would have been counterproductive, even absurd, if mathematical
constructions that involve only the manifold of pure-formal intuition were not
grounded on categorial synthesis speciosa. Nor could he have claimed that the cat
egories are conditions of sensible cognition generally, not merely the empirical sort,
if transcendental synthesis speciosa did not relate to the manifold spaces and times
of formal intuition as well as the manifold realities of empirical intuition.
Indeed, how else could the categories determine the manifold realities appre
hended in empirical-material intuition if not via the determination of the mani
fold spaces and times apprehended in pure-formal intuition? Given Kant's view
that the categories could not determine empirical realities if appearances did not
have a formal side that could be represented and sy nthetically determined entirely
a priori (Al38-9/Bl77-8), I do not see how to avoid the implication that a tran
scendental synthesis speciosa of the manifold of formal intuition is presupposed
by transcendental schematism, understood as the determination of the manifold
realities apprehensible only in empirical-material intuition. It is not just that the
realities given through sensation have to be represented a posteriori or not at all,
while pure intuition, as involving only the formal side of appearances, is appre
hended completely a priori. There is a radical heterogeneity between the catego
ries, which have the logical form of universality, contain no content except that
derived from logical functions of judgment, and are representable only a priori
and the matter of appearances, which derives exclusively from sensation, has nei
ther logical form nor content, and is representable only a posteriori. This gulf
can only be bridged by means of pure space and time, to which appearances must
conform as the source of their constitutive form, which can and must be repre
sented completely a priori and which, despite being intuitions devoid of both
logical form and content, are at least homogeneous with the categories by virtue
of giving prediscursive expression to the same original sy nthetic unity of apper
ception expressed discursively by means of the logical functions of judgment and
objectively by means of the synthesis intellectualis of the categories ("The time
intuition is not homogeneous with the categories, but rather time determination,
the unity of representations in sy nthesis (composition) of the given intuition,"
AA 16 § 6359 [1796--8]). We thus arrive at the same conclusion from theoretical
considerations that we did earlier on purely textual grounds: the determination
408 Kant's Anatomy of the Intelligent Mind
of the manifold of formally intuited space and time conformably to the objective
unity of apperception thought in the synthesis intellectualis of the categories by
means of transcendental synthesis speciosa precedes and makes possible the tran
scendental synthesis of imagination of the matter of appearances that equips the
categories with their schemata .
With the foregoing in mind, Kant's reason for not including a category-by-category
elucidation of transcendental synthesis speciosa in the Transcendental Deduction is
not difficult to divine. T he Deduction is an attempt to prove the categories to be objec
tively valid a priori by showing them to stand in a determinative relation to appear
ances as apprehended empirically in intuition, that is, a determinative relation that
includes the matter of appearances. Proving this to be the case relates the categories
via the sensation matter of appearances to "the transcendental matter of all objects
as things in themselves (facticity, reality)" (Al43/Bl82). For even if they cannot deter
mine realities as they are in themselves, the a priori relation of their schemata to things
in themselves via their a priori determination of the matter of appearances sets a priori
categorial determinations on the same plane of empirical reality as the earth's gravity,
the heat emanating from the sun, and everything else objectively existent in nature.
Clearly, this kind of objective validity could never be established simply by showing
the categories to hold merely with respect to the form of appearances in pure space
and time ("whether there can be things that must be intuited in this form remains
to be discovered" since " [tjhings in space and time are ...given only through empiri
cal representation," Bl47). Accordingly, a category-by-category elucidation of the
determination of pure (formal) intuition by transcendental synthesis speciosa would be
completely beside the point for Kant's purposes in the Deduction, where everything is
geared toward showing the categories to be valid of realities that can only be met with
a posteriori in empirical intuition by means of the matter of appearances correspond
ing to sensation.
Why then is there no category-by-category elucidation of transcendental schemata
in the Transcendental Deduction if, as maintained earlier, it concerns the material,
real side of intuition? Why did Kant instead situate the schematism in the Analytic
of Principles? Since the transcendental doctrine of judgment will be considered
in the next chapter, I will here confine myself merely to the observation that no
such elucidation is needed in order to succeed in meeting the demand in§ 14 (Bl24-9)
for a transcendental deduction for the categories. Providing the categories with their
transcendental deduction requires only that it be shown that the objective unity of
apperception (transcendental affinity) is essential to the possibility of experience and
its objects, and that this unity is possible in no other way than through the determina
tion of all synthesis of the manifold of intuition, from recognition all the way down
to apprehension, conformably with the categories (Al25, Bl64). In this way, pure
concepts of the understanding take on the worth of pure concepts of the objects of our
sensible intuition, which is the value they require if the synthetic a priori judgments
founded on them are to count as genuinely objective principles of intuition, percep
tion, and experience itself (respectively, the Axioms, Anticipations, and Analogies
A Category-by-Category Elucidation 409
be given. At the same time, since the categories can never be used to do anything
other than determine concepts and judgments with respect to the logical functions
of judgment, they can never directly apply to anything except that which is already
universal in form. Only universals-concepts and judgments-admit of being
combined in judgments, and so only they can have their behavior in judgments
shaped by the categories. Since intuitions are representations to which the analytic
unity of the I think is not attached, they are not universal in form and so are not
directly determinable conformably to the categories (intuitions can never them
selves occupy the logical positions of subject or predicate of judgments). Hence, if
intuitions are to be subjected to the categories, it must take place by means of inter
vening concepts on which the categories can be employed, hence entirely in thought
Uudgment), as synthetic a priori philosophical-discursive cognition rather than the
mathematical-intuitive sort (A708-38/B736--66). It is therefore clear that both tran
scendental syntheses, synthesis intellectualis and synthesis speciosa, have the same
essential nature: the use of transcendental concepts to determine the manifold of
appearances apprehended in intuition by means of the determination of the role
in judgments of the (non-transcendental) concepts and judgments that represent
these appearances.
This is also clear from Kant's characterization of the schema of a category not
as an intuition but as "the sensible concept of an object in agreement with the
category" (Al46/Bl86), and, more particularly, the kind of concept that, by con
trast with empirical and mathematical concepts, "can never be met with in any
intuition," "be itself intuited through the senses," be "contained in appearance"
(Al37-8/Bl76--7), or ever "be brought into any image whatsoever" (Al42/Bl81).
What is true of transcendental schematism may also be presumed true of the
transcendental synthesis speciosa of the manifold of pure-formal intuition it pre
supposes: its products are not themselves concepts of intuitions but concepts for
determining given concepts of intuitions (and given judgments containing them)
with respect to the logical functions. Indeed, the only thing that distinguishes tran
scendental synthesis speciosa from transcendental synthesis intellectualis is that in
the former, the rule thought in the category is applied to concepts of intuitions that
have a particular sensible form (Al46/Bl85), while in the latter, the rule is applied
to concepts of sensible intuitions in general, irrespective of their form, just so long
as they are sensible (Bl48, Bl50-l, Bl53-4).
In extrapolating the transcendental synthesis of formal intuition from the
transcendental synthesis of material intuition (schematism), it will be helpful to
modify Kant's terminology somewhat in order to avoid misunderstanding. The
schema of the category of substance, for example, is "the permanence of the
real in time, i.e. the representation of the real as the substratum of empirical
time-determination in general, which thus remains while everything else changes"
(Al43/Bl83). Since formal intuition has nothing to do with the real given
through sensation, and since Kant's use of the term "substance" in the Analytic
of Principles always signifies a reality of this kind, it is better to characterize the
A Category-by-Category Elucidation 411
4 Kant used the expression "final subject" (letzte Subjekt) in reference to substance at
A205/B250, but its sense carries over more easily to the context of formal intuition than
"substance" does.
412 Kant's Anatomy of the Intelligent Mind
Pure formally intuited space and time make possible an inexhaustible infinity of
appearances that can be apprehended in juxtaposition or succession in intuition. By
attaching the analytic unity of apperception to this manifold, its representation is con
verted into the representation of an equally inexhaustible infinity of concepts of these
appearances and their juxtaposition or succession with each and every other (ch 9-C).
Utilizing the logical form of categorical judgment to represent each of these (concepts
of) appearances as subject and (the concepts of) their juxtaposition or succession with
some other appearance as predicate, the stage is set for the synthesis speciosa of the
category of substance and accident. The synthesis determines the subjects of these
judgments as final subjects-concepts that can never be predicated of anything else
and so as the ultimate substrata of all possible spatial and temporal determinations of
appearances. Correspondingly, it determines the predicates of each final subject-its
juxtaposition or succession with another final subject-as never anything other than
its predicates, that is, as incapable of obtaining independently of their final subject
substrate. And since the number of final subjects and their determinations possible in
pure-formal space and time is inexhaustibly infinite, the same is true of the output of
the synthesis speciosa of substance and accident: to every appearance possible in pure
space and time there corresponds a (transcendental synthetic) final subject, and to
every possible juxtaposition or succession of that appearance with another there cor
responds a (transcendental synthetic) determination predicated of that final subject.
Insofar as the synthesis speciosa of substance and accident effects a necessary cor
relation of each final subject, as given in a concept, to each of the spaces and times pre
discursively present in the infinite manifolds of the pure-formal intuitions of space and
time, it goes together with the synthesis speciosa of the modal category of existence. The
two representations necessarily correlated in accordance with the synthesis of substance
and accident, intuitions and the concepts that represent them, are completely identical
in content, differing only by the addition (or not) of universal form (the analytic unity
of the I think). This means that the transcendental worth conferred on the concepts by
synthesis speciosa as representations of final subjects extends immediately to their intu
ition correlates, determining those intuitions as actually existent final subjects (thereby
making the sensible non-amphibolously intellectual), and, conversely, conferring objec
tive reality on the contents thought in their conceptual correlates, including their tran
scendental worth as final subjects (thereby making the intellectual non-amphibolously
sensible). For if indeed, according to Kant's theory of universals, the matter of the con
cept, apart from the form of logical universality it acquires when the analytic unity of
apperception is attached to it, literally is the intuition,5 then to categorically determine
5 Just as the matter (i.e., the object) of the concept of red formed by adding the analytic unity of
apperception to a sensation of red in Kant's example at Bl 33-4n literally is the sensation: ch 9-C.
A Category-by-Category Elucidation 413
the concepts representing the manifold of pure, formally intuited space and time as final
subjects is ipso facto to determine the manifolds of spatial and temporal appearances
themselves as manifolds of final subjects. And since this is just to say that under the
synthesis speciosa of the category of existence-nonexistence, the final subjects of deter
mination yielded by the synthesis speciosa of substance and accident are objectively
present in pure-formal intuition, these subjects must be accorded existence outside their
concepts in (pure) intuition.
The determinations of juxtaposition and succession predicated of final subjects
in substance-accident and existence-nonexistence synthesis speciosa must not yet,
however, be thought of as relating or ordering them, much less as yielding the kind
of necessary interconnection to which Kant traced the relation of representations
to an object (ch 13-H). Since the manifolds of pure formally intuited space and
time are completely undifferentiated and indeterminate, the final subjects posited
in pure space and time through substance-accident and existence synthesis speciosa
can be no less undifferentiated and indeterminate. Nevertheless, unlike the spatial
and temporal manifolds conceptualized in them, final subjects and the determina
tions predicated of them are precisely the conceptual "raw material" to which alone
the transcendental synthesis speciosa of the remaining categories can be applied to
make good their want of differentiation and determination and thereby (mediately)
make good the same want in the corresponding manifolds of formally intuited
space and time themselves.
Before the transcendental synthesis speciosa of the remaining categories can be
considered, however, the spatial and temporal significance of final subjects and
their determinations needs to be specified more precisely. One of the features of
the prediscursive pure space and time of the Transcendental Aesthetic is the impos
sibility of any holes or gaps in them (an absence of space surrounded by spaces or
an absence of time between times: ch 3). To put the same point another way, pure
space does not make possible an outer intuition of the nonbeing of any space, nor
does pure time make possible an inner intuition of the nonbeing of time. T he ques
tion is how this essential feature of all possible spaces and times can be captured
in the corresponding concepts (formed by incorporating the analytic unity of the
I think into their intuition). For in general logical terms (ch 10), even a concept of
something so positive that its nonbeing is impossible to represent can be employed
as a predicate in judgments and related to its subject negatively (S is not P). To
preclude this, it is necessary to determine the concept as a final subject, one that
can never be predicated, positively or negatively, of anything else in judgments, by
means of the synthesis speciosa of substance and accident. Only then is the concept
fully able to capture the essential character of the space or time it represents: not
being predicable, it cannot be used privatively to conceive an absence (a final subject
S that might have had space but instead lacks it). And while a final subject's affirma
tion is equally inconceivable, this makes no difference if, as is the case here, it is a
pure concept of juxtaposition or succession that is identical in content to something
the nonbeing of which is unintuitable, and so impossible, in pure space or time.
414 Kant's Anatomy of the Intelligent Mind
There is, however, a genuinely insuperable obstacle that prevents times (the
manifold of formally intuited pure time) from being conceived as final subjects.
For although in one sense the nonbeing of time anywhere in pure time is not a
possible representation, there is another sense in which nothing is easier: different
times are all successive, none simultaneous, and so are the very antithesis of the
permanent ("simultaneity is not a mode of time itself, as in time no parts whatever
are simultaneous, all are successive," Al83/Bl16).6 Matters are quite different when
it comes to spaces, where adjacency goes with simultaneity with the same necessity
adjacency in time goes with succession. Since formally intuited space necessarily
has no "holes," and every space in it is consequently adjacent to other spaces, all
spaces-the totality of spaces-must be simultaneous and remain so at all times.
Indeed, given that spaces in pure space are distinguishable only by juxtaposition
(one outside another), never by succession (one preceding or succeeding another),
these simultaneous spaces cannot participate in temporal succession at all but must
instead be represented-via their concepts, determined as final subjects-as the
"permanent with which the successive is simultaneous" (AA 18 § 6313 [1790--1]),
that is, as the permanent substratum of all succession7 Although merely an aes
thetic (nondiscursive, precategorial) representation of the manifold of formally
intuited space as the permanent substratum of succession, it nevertheless suffices
to preclude the kind of representation of nonbeing that prevents the manifold of
pure time from being conceived as final subjects under the synthesis speciosa of sub
stance. Thus, by determining the inexhaustible infinity of concepts of intuitions jux
taposable in formally intuited space conformably with the category of substance,
synthesis speciosa transforms them into representations of the manifold spaces
apprehended in the pure-formal intuition of space as themselves final subjects of
6 Also: "there is only one time in which all distinct times must be placed, not simultaneously
says. The permanence of the form in our mind is not the same (for the form of time is just as per
manent), rather as a representation of something outside us that we set under (unterlegen) all time
determination and therefore represent (as) permanent" (AA 18 § 5653 (p. 308) [mid-late 1780s]).
To represent the manifold of formally intuited space as a succession would require nothing less
than a capacity to transform it "into a representation of time, i.e .. .. to represent space itself as
a time (according to one dimension) and that is contradictory" (AA 18 § 6311 [1790]). There can
be little doubt that it is precisely this merely formal concept of permanent final subjects in pure
space that underlies the material concept of permanence produced by the synthesis speciosa that
schematizes the category of substance and enters into "the principle of the permanence of sub
stance" (A182/B224) of the First Analogy of Experience. For the material permanence needed
to demonstrate the objective validity of the category of substance in the field of appearance
"time as the constant correlate of all existence of appearances, all change and all concomitance"
(A l 83/B226)--can be "given in intuition" only through "an intuition in space" because "space
alone determines permanently, whereas time, and so everything in inner sense, constantly flows"
(B291). Thus, nothing more is required to derive material permanence from the formal perma
nence generated by the synthesis speciosa of formally intuited space than the addition of the kind
of filling of space that only the matter of appearance, corresponding to sensation, can supply.
See chs 16-C-1 and 17-D.
A Category-by-Category Elucidation 415
determination, the nonbeing of which can neither be intuited nor conceived (for
again, these spaces differ from their transcendental-synthetically determined con
cepts solely by their lack of the form of logical universality contributed by the
analytic unity of apperception).
The upshot is that despite being subjectively (analytically) distinct notions,
under substance-accident and existence synthesis speciosa, final subjects of deter
mination and distinct spaces become objectively (a priori synthetically) one and
the same: distinct final subjects of determination are different spaces, and different
spaces are distinct final subjects. 8 This objective identity has a consequence of the
first importance for Kant's system: it creates the possibility of uniquely differentiat
ing each space from and determining its relation to every other space by means of
categorial connections between the determinations ("accidents") of distinct final
subjects. If, for example, determinations of two final subjects, A and B, are imme
diately related by a categorial determination capable of connecting distinct deter
minations of distinct final subjects, then the spaces corresponding to A and B must
be conceived to be immediately related as well, that is, adjacent. Since categorial
relations are always irreversible, however, in order to conceive their relation to be
reciprocal (B to be adjacent to A as well as A adjacent to B), the reverse relation
must be effected by different determinations, irreversible in the other direction, i.e. ,
the adjacency of A to B by the irreversible relation of determination d of A ("A is d"
in judgment form) to determination e of B, the adjacency of B to A effected by the
irreversible relation off of B to g of A. If, in addition, the categorial relation that
does this is subordinative (rather than coordinative), then determinations imme
diately related by it would have to be conceived to be immediately subordinated
as well, and so as adjacent in time (the subordinated determination the immediate
successor of its subordinative determination). This means that if d, e, f, and g of
final subjects A and B are each in turn immediately related in this way, the result
must be conceived as successively passing from the space of A immediately into the
space of B and then immediately back again; and if g of A is similarly immediately
subordinately related to h of B, the "motion"9 will continue immediately back to
B, with the potential for perpetual oscillation so long as (always) different deter
minations of A and B are related in the same manner. If other determinations of
A and B stand in the same relation but via determinations of other final subjects
rather than immediately, these relations have to be conceived as passing through
the spaces of the intervening final subjects C, D, E, etc . , and so as roundabout
routes from the space of A to the space of B. Since there is no limitation to the
number of determinations that can be conceived to inhere in final subjects, it is pos
sible in this manner to pass from any space to any of the infinitely many other spaces
either directly (adjacency) or via others, and do so in any and all sequences. All that
8 See Chapter 2-A for a discussion of the distinction Kant drew in a letter to Schultz between
in time. And this, as soon will become apparent, is all that is requisite to uniquely dif
ferentiate and completely determine each space and each time in the inexhaustibly
infinite manifolds of formally intuited pure space and time in respect to every other.
While formal juxtaposition within formally intuited pure space is sufficient to dem
onstrate the non-identity of the spaces concerned (as being outside one another), it
leaves the identity of each as the unique space it is undetermined because there is as
yet no principle of order to determine for every final subject in space which others
are adjacent, which are situated at a remove from it, and, in the case of the latter,
which other final subjects (spaces) intervene. The synthesis speciosa of substance
and accident must therefore be supplemented by that of some other category or
categories before the final subjects produced by the former can acquire an identity
sufficient to uniquely differentiate each from, and completely determine its relations
to, all others.
The synthesis speciosa chiefly responsible for conferring such an identity on final
subjects of determination in space is cause and effect. This synthesis cannot, how
ever, connect final subjects directly but only via their determinations. For whereas
the category of substance and accident serves to irreversibly fix the categorical rela
tion of subject and predicate concepts, that of cause and effect irreversibly fixes the
hypothetical relation of judgments. Its synthesis speciosa therefore relates to the
predication of determinations of final subjects rather than to these subjects them
selves (chs 10-D and 11-B). It functions to determine each of the infinite totality of
these predications as conditional upon another predication. Since to be conditional
is for a determination not to be able to be predicated of its subject until another
determination is predicated of its subject, none of the predications of determina
tions of final subjects can originally be actual. All must originally be merely pos
sible and remain so until the actualization of their immediate condition necessitates
their own actualization. Thus, like the categories themselves (ch 11-B), the synthesis
speciosa of cause and effect and possibility-impossibility go inseparably together.
The synthesis speciosa of cause and effect and possibility-impossibility does two
the possibility of all the rest is precluded. It is because the relation of judgments is
fixed by the synthesis in this manner that each determination of each final subject
must have one determination as its unique necessitating ground and another one as
its unique necessitated consequent-not two, not ten, not none, not the sum-total
of all the rest. Each "cause" in formally intuited space and time has its own unique
"effect," and each "effect" its own unique "cause." And the same holds true of
the ground's own ground, the consequence's own consequence, the ground of its
ground's ground, the consequence of its consequence's consequence, and so on: all
have a unique necessitating ground and a unique necessitated consequent; none
have multiple grounds or multiple consequences. Thus, the synthesis speciosa of
cause and effect and possibility-impossibility creates perfectly linear causal lineages
linking each determination of every final subject of determination in space.
Because this synthesis uniquely determines the order of the members of the
series as well as which members it has, it not only precludes any branching in them
(multiple grounds or multiple consequences of any individual member of the series)
but also prevents any determination from ever recurring as well. For just as the
tenth member of a series determined conformably to the synthesis speciosa of cause
and effect must be immediately preceded by the ninth, its immediate ground, and
nothing else, so too the ninth can be approached in no other order than through the
eighth, its unique immediate ground, and so on back to the second, which can be
approached through no other than its immediate ground, the first. Accordingly, the
entire ordered series up to the tenth is just as much the tenth member's unique neces
sitating ground, and the tenth the entire ordered series' unique consequence, as the
ninth is its unique immediate necessitating ground and the tenth the ninth's unique
immediate consequence. And so too for the eleventh, the twelfth, and arbitrarily
far beyond. But if for each member of a series the particular ordered series leading
up to it, taken in its entirety, is its unique necessitating ground and it is the unique
consequence necessitated by that series, then each determination of final subjects
in space that the series contains is determined to a unique point in the series such
that it not only cannot occur before then but also cannot recur after it. This applies
even to its first member. For it to recur later in the series, some other member of
the series would have to be its immediate necessitating ground. But since the entire
ordered series terminating in that member, including the first, is also its necessitat
ing ground, the first, or any other member, could only recur in the series by ground
ing itself, which is impossible. For, like every determination of final subjects, the
first member is irreversibly determined by the synthesis speciosa of the categories
of cause and effect and possibility-impossibility as conditional upon some deter
mination other than itself and so can never ground itself. It thus cannot have any
member of the series as its ground (and so can never recur in the series) but must
instead be grounded in a determination of a final subject situated outside the series
("the things themselves that, as effects, presuppose other things as causes cannot at
the same time reciprocally be causes of these causes," CJ 372).
A Category-by-Category Elucidation 419
(I) the causal lineages linking the infinitely many determinations predicable
of each final subject are all perfectly linear, never branching;
420 Kant's Anatomy of the Intelligent Mind
subjects, this means that the subordination of the infinite totality of its determina
tions to the synthesis speciosa of the category of possibility-impossibility puts all
originally on the same modal footing: in themselves merely possible, dependent
for their actualization on the actualization of some external condition. Insofar as
these merely possible determinations are at the same time determined as condi
tional upon some other by the synthesis speciosa of cause and effect, it follows that
at least one of any final subject's determinations must have its immediate ground in
the determination of some other final subject. For if all of its own determinations
A Category-by-Category Elucidation 421
are originally determined as merely possible, then none can be actualized until a
determination actualized in some other final subject, related to one of its own as
immediate ground, actualizes that determination.
Although this leads to a vicious circle that only the synthesis speciosa of the
category of community can resolve (Section D), it also shows how the irreversible
ground-consequence relations y ielded by the synthesis speciosa of cause and effect
can order spaces in pure intuition. Given a series in which all of the members are
determinations of different spaces (final subjects), in passing from one to the next
in time, the corresponding spaces are determined as contiguous: the space of the
ground necessitates that the space of its immediate consequence comes immediately
after it, and so is next to it, adjacent, in space. Thus begins the process of dif
ferentiating and determining the undifferentiated, indeterminate manifold of the
pure-formal intuition of space: the first space in the series determines the second
as next to it in the order of spaces within pure space, and thereby singles out that
space from all the others; the second space does the same with respect to the third,
the third the fourth, and so on through to the space at which the series concludes.
Although contiguity in space is not a unidirectional relation like irreversible
ground-consequence relations, it can be exhibited by the synthesis speciosa of cause
and effect and possibility-impossibility by reciprocal causal relations. This is possible
because each final subject in space has an inexhaustible infinity of determinations
(Section B) and each of these determinations has its own linear, non-intersecting
causal lineage.10 Accordingly, if a final subject has (at least) two determinations
immediately related to the same number of determinations in the other, (at least)
one as ground and the other as consequence, they are determined as occupants of
adjacent spaces. The same applies for whole series of spaces: one causal lineage of
their determinations may order spaces (final subjects) so that space S is contiguous
to space T, T to U, U to V, and V to W, while another causal lineage orders the same
spaces in the reverse direction, from W to V to U to T to S.
This is still only a further step along the way to the complete differentiation
and determination of the inexhaustibly infinite manifold of spaces within formally
intuited space. By contrast, the ordering of spaces by contiguity under the syn
thesis speciosa of cause and effect does suffice to explicate temporal simultaneity.
We saw in the previous section that the synthesis speciosa of substance determines
all spaces (final subjects) as simultaneous and permanent. Their determinations,
however, can be either successive or simultaneous and that synthesis did not suf
fice to determine them one way or the other. By contrast, cause and effect synthesis
speciosa, together with the synthesis speciosa of possibility-impossibility, deter
mines the totality of any final subject's determinations as merely possible until
10 Causal lineages cannot intersect without violating the first rule specified earlier: no determi
nation in a causally determinate series can have multiple grounds or multiple consequences. Since
two lineages converging would result in the same determination having two immediate grounds,
the category's synthesis speciosa precludes it.
422 Kant's Anatomy of the Intelligent Mind
such time as their conditions in other final subjects are actualized. If only one of a
final subject S's determinations, d, is actualized by an external determination, then
none of its others are actualized simultaneously with it, i.e., all are successors to d
in the order of time. If there are other actualized determinations of S that, like d,
have unactualized consequences, then, given rules 1-111 above, they are determined
as occupying identical positions in different lineages, and as such count as simulta
neous with d. With identity of position across lineages thus established, d and all
determinations of S co-actualized with d will now count as occupants of a particu
lar unique position in the series of pure times within formally intuited pure time,
and so as simultaneous not only with respect to S but with respect to the order
of pure time itself.11 Similarly, the immediate consequences of these simultaneous
determinations of S, be they determinations of S or some other final subject, will
ipso facto count as simultaneous occupants of the (ordinally) next time, as will
their immediate consequences, and so on through the time series. So too in the
other temporal direction: the immediate grounds of the determinations actualized
in S simultaneously with d will be defined across lineages as occupants of the time
immediately prior to that of their consequences, as will their immediate grounds,
and so on backward through the time series, regardless of whether the determina
tions concerned belong to the same or different final subjects (spaces). Finally,
since every determination of final subjects in space that does not have its immedi
ate ground in the determination of another subject must remotely follow one in its
causal lineage that does, every determination of every final subject is determined by
transcendental synthesis speciosa as either simultaneous with or at such-and-such
determinate remove in the order of time from every other. This serial order is none
other than that fully objectified pure time whose manifold comprises an inexhaust
ible infinity of times that are each, without exception, uniquely differentiated from
and completely determined in every relation to (i.e., at such-or-such remove from)
every other time. And in this way objective unity of apperception is brought to
the manifold of formally intuited pure time by the synthesis speciosa of cause and
effect and possibility-impossibility.
In what sense is it 'objective'? Under Kant's transcendental criterion of objectiv
ity-necessary interconnection of the manifold (ch 13-H)-the succession of times
yielded by the synthesis speciosa of cause and effect and possibility-impossibility
qualifies as an objective (albeit merely formal) succession because it is grounded
entirely on series of irreversible ground-consequence relations in which the
u To avoid confusion, it is important to not to forget that the grounds and consequences con
sidered here are judgments, w ith as yet no extra-logical quantitative determination at all, least of
all spatial or temporal quantities such as variable durations. Indeed, given the completely undif
ferentiated, indeterm inate character of the manifolds of pure-formal space and time, I see no way
for there to be a transcendental synthesis speciosa for the categories of quantity (or quality) other
than on the foundation provided by the purely ordinal nexus of spaces and times established
through the synthesis speciosa of the categor ies of relation and modality: see Sections E and F.
A Category-by-Category Elucidation 423
13 Kant's treatment of incongruent counterparts (PFM 286) can be understood on these prin
ciples, e.g., the spaces included in the figures of a left hand and a right hand in three dimensional
space. Considered from a transcendental standpoint, the surfaces constituting both configura
tions mathematically actualize (construct) the identical series of irreversible ground-consequence
relations and thereby order their component spaces (final subjects) identically. But this is just to
say that they (e.g., left and right hands) are congruent insofar as abstraction is made from the dimen
sion that includes these spaces together with the spaces (final subjects) surrounding them. As soon
as the external spaces are factored in, together with the series of irreversible ground-consequence
relations that link them to (different) determinations of the spaces composing the figures, differ
ences in the series immediately emerge sufficient to demonstrate the impossibility of congruence.
A Category-by-Category Elucidation 425
dimension, et al.), and so provide cognition, mathematical and empirical alike, with
the objective, necessary ordering of spaces on which their possibility depends.
The synthesis speciosa of cause and effect and possibility-impossibility accounts
for (ordinal) distance between spaces in terms of the number14 of grounds and con
sequences involved in the different causal lineages connecting determinations of
any space A (final subject) with any other space B (different final subject). The
distance between the spaces is objectively determined by the least of all of these lin
eages. If the least is an immediate ground-consequence relation, then the two spaces
are adjacent, that is, no space separates them Gust as times can be distinct without
there needing to be any time intervening between them: A203/B248). If the least
is a series of two such relations, then they are at that distance from one another,
that is, one cannot be reached from the other without passing through at least one
other space; if the least is a series of three, then two intervening adjacent spaces
must be traversed; if a series of four, then three intervening adjacent spaces must
be traversed; and so on for least series of every number. In this way, there emerges a
universal measure of distance applicable to all final subjects in space.
Dimension, as explicated via the synthesis speciosa of cause and effect and
possibility-impossibility, is somewhat subtle. Given the inexhaustible infinities of
concepts and judgments concerned in this synthesis, each space (final subject) will
have infinitely many adjacent spaces. Now, the natural thing is to consider these as
consisting entirely of different directions in three dimensions since that is the only
infinity of directions radiating from a single point we can "imagine." This, however,
is merely a limitation of human psychology, not transcendental productive imagi
nation. The latter's synthesis is purely discursive-philosophical in the sense specified
in the Methodology (A712/B740-A738/B766), as is evident from the fact that it
consists of nothing but irreversible ground-consequence relations connecting judg
ments formed from concepts of the manifold of pure space. This is why products of
transcendental synthesis speciosa "can never be met with in any intuition," "be itself
intuited through the senses," or be "contained in appearance" (Al37-8/Bl76-7),
and so "can never be brought into any image whatsoever" (Al42/Bl81). As a result,
transcendental imagination is immune to the limitations of the human psyche's
image-producing capacities (non-transcendental figurative synthesis and empirical
reproductive synthesis).'5 Although determinative of the manifold of pure-formal
space itself (which makes it imagination rather than mere pure understanding), its
synthesis consists entirely of concepts, the judgments formed from them, and the
determinations of these judgments conformably to the categories. The synthesis
speciosa of cause and effect and possibility-impossibility confers on each space in
14 This notion of number will be examined in section E as the product of the synthesis speciosa
of the categories of quantity.
15 It is equally immune to the limits of the human capacity for symbolic construction, which,
despite being symbolic, still counts as mathematical-intuitive rather than philosophical-discursive
synthetic a priori cognition, e.g. algebra (A717/B745 and A734/B763). See ch 6.
426 Kant's Anatomy of the Intelligent Mind
that manifold an inexhaustible infinity of adjacent spaces and so the same number
of directions radiating out from it into adjacent spaces. This infinity can no more
be exhausted by the infinity of directions radiating out from it in three dimensions
than it can be exhausted by the infinity of directions radiating out from it in two.
Instead, the spaces ordered by this synthesis can accommodate an infinity of direc
tions in an infinity of dimensions, and so are perfectly capable of supporting the
most recondite geometrical and topological constructions.
This is not of course to say that Kant supposed that our minds are endowed with
a transcendental capacity for spatial representation in infinite dimensions in order
that our species might one day produce experts in the mathematics of n-dimensional
spaces. Rather, I think he would have us view it just the way we view capacities like
the senses and ordinary imagination: they make possible many more representa
tions than we actually have or ever will have. Just as we all have the capacity, say,
to savor the flavor of a fruit growing on a planet in a different part of the galaxy
but never actually will, our transcendental imaginations endow us with powers of
cognition neither we nor our farthest progeny can hope ever to fully realize. It is, to
be sure, essential to Kant's project that some capacities impose their limits on oth
ers (e.g., sensibility on discursive understanding and the scope of its cognition). But
it should also now be clear, in light of what we have determined about how tran
scendental imagination functions, that our incapacity to represent more than three
dimensions visually, tactually, or in any other ostensive manner does nothing to
prevent transcendental imagination from yielding infinitely many more dimensions
as well since no imaging capacity is required, only the philosophical-discursive
capacity to represent the infinite totality of determinations of the infinite totality
of final subjects in causal lineages. The "extra" dimensions produced by the synthe
sis speciosa of cause and effect and possibility-impossibility are no less (or more)
really and objectively present in formally intuited pure space than the first three.
Our inability to image in more than three dimensions has no objective implications;
it simply means that beyond three spatial dimensions, the subjective limitations of
human cognition oblige mathematicians to construct such spaces purely symboli
cally, unaided by ostension.
While cause and effect and possibility-impossibility synthesis speciosa may be key
to conferring the objective unity of apperception on the manifolds of formally
intuited space and time, it falls short on one score: it cannot necessitate their
actuality. Their synthesis intellectualis can determine nothing except irreversible
ground-consequence relations in judgments, and ground-consequence relations are
possible only through the determination of their components as possible, not actual
judgments, via the category of possibility-impossibility (chs 10-D and 11-B). This
A Category-by-Category Elucidation 427
means that the categories of cause and effect and possibility-impossibility are only
capable of conditioning one actualization of a possibility upon another, but cannot
itself actualize anything.
Formal intuition is simply a case in point: cause and effect and
possibility-impossibility synthesis speciosa by itself is incapable of actualizing any
of the possible connections it forges between distinct final subjects of determina
tion in space. Something must actualize them and must do so in their inexhaust
ibly infinite totality (i.e., completely a priori) if the objective spatial and temporal
16
Kant made clear that the portion of the text cited here is a principle of determination estab
lished in the Transcendental Analytic and so concerns the category (community) derived from
disjunctive judgment rather than the transcendental ideal of reason.
428 Kant's Anatomy of the Intelligent Mind
infants. Nevertheless, it is a long way from saying this to motivating the notion that
infants occupy themselves with representing the infinite manifold of pure space and
time via an infinite manifold of concepts, represent these concepts as formed into
as many judgments, and then determine the infinite totality of those judgments
conformably with the categories. Moreover, since the entire process presupposes
an acquaintance with the synthesis intellectualis of the categories, this raises the
further question of how infants attain these concepts and become sufficiently profi
cient in their use as to be able to execute transcendental synthesis speciosa.
Many will no doubt be inclined to respond to such queries with a variation of
Berkeley's ridicule of Lockean abstract general ideas (though see UU ch 10-A):
[S]urely the great and multiplied labour of framing abstract notions will be
found a hard task for that tender age. Is it not a hard thing to imagine that
a couple of children cannot prate together of their sugar-plums and rattles
and the rest of their little trinkets, till they have first tacked together number
less inconsistencies, and so framed in their minds abstract general ideas, and
annexed them to every common name they make use of ? (PHK Intr § 14)
Kant, at least as presented here, might seem liable to the same ridicule, not for sup
posing infants to occupy themselves with tacking together numberless inconsisten
cies, but for supposing that they tack together infinite numbers of concepts and
judgments by means of the most abstract, purely intellectual representations pos
sible: representations of the logical functions themselves, whose only possible use is
to determine the logical positions of discursive representations in judgment. How
then is Kant's psychologism to be defended, especially against those who think the
notion of infants engaging in so complex and highly intellectualized an activity as
transcendental synthesis so preposterous as to cinch the case in favor of an anti
psychological, normativist Kant?
Curiously, Berkeley's own theory of vision provides the best model for defend
ing Kant's psychologism (UU ch 14). On Berkeley's view, visual sensations are as
devoid of intrinsic spatial content as sounds, smells, and flavors are; and were it not
for touch, we would be as ignorant of space and the spatial-figure, size, depth, dis
tance, orientation, etc.-as those born blind are of color and light. Consequently,
the use of visual inputs to represent spatially is a capacity our minds must acquire
from scratch, piecemeal, by utilizing tactual experience of space to separate spa
tially relevant visual information from spatially insignificant "noise" until we have
attained sufficient psychological command of visual-tactual correlations to be able
to "read" spatial information from nonspatial visual data with the same fluency we
speak and read our native language. And since fluency in the language of visual
spatiality is even more essential to survival than verbal language, it is a language our
minds must be equipped to master well before we are out of our cribs.
Berkeley's theory of vision is an explanation of how the infant mind is able to
do this. It distinguishes three stages in the acquisition of the language of visual
spatiality. It starts with the infant's exercise of its native abilities to copy and retain
430 Kant's Anatomy of the Intelligent Mind
complex suites of mental operations developing and using this language involves,
the Berkeleian response is that custom conditions us to perform them with such
facility and rapidity as to exceed our capacity to discern or attend to these doings.
And to the question why operations we cannot even discern or attend to should
nevertheless be regarded as conscious, the Berkeleian response is that we can no
more discern spaces through visual data without being conscious of them and their
associations with tactual data at some level than we can read Don Quixote through
the covers of a closed book (an extension of the Lockean conception of the differ
ence between perception and higher levels of mentation considered in ch 8-B and
UU ch 6).
Kant's transcendental synthesis speciosa is similar. It is manifestly highly com
plex. Yet as the only means whereby to confer objective unity of apperception on
the "scattered and single" manifold apprehended in the pure intuitions of space
and time, it is absolutely necessary that infants be equipped to execute the synthesis
since their mental life could otherwise never rise above "a blind play of representa
tions, less than a dream" (Al 12; ch 13). Its explanation, like that given by Berkeley
of visual spatiality, must consequently involve no capacity not within the power
of infants. Kant met this requirement by developing a completely new model of
the mind according to which sensibility requires pure intuitions, discursive under
standing requires the analytic unity of the I think , judgment requires the twelve
logical functions, and cognitive understanding requires the twelve categories. The
first are needed in order to have a unified sensibility (chs 3 and 4). The second
provides a means for converting intuitions directly into concepts, with or without
the intervention of reflection comparison and abstraction or customary associa
tion (ch 9). The third ensures the categorical relatability of all concepts as subjects
and predicates and the hypothetical and disjunctive relatability of all judgments
in complex judgments and inferences (ch 10). Finally, insofar as the judging sub
ject, in operating with logical functions, cannot fail to recognize the logical leeway
they permit (Bl28-9), it also cannot help forming the concepts that provide the
means to eliminate that leeway and thereby equips itself with precisely the con
cepts needed to perform transcendental synthesis (ch 11). To the objection that we
never have the least awareness of performing the operations necessary to produce
the intuitions, concepts, and judgments involved in this synthesis, Kant could reply
that it is enough to attain objective unity of apperception merely to exercise one's
capacity to form and employ these representations; one does not also need to make
A Category-by-Category Elucidation 431
them clear and distinct ("the first thing [the understanding] does is not to make
the representation of objects distinct but to make the representation of an object
in general possible," A199/B244-5). And to the objection that there is no warrant
for considering operations we are not aware of performing to involve conscious
ness, Kant could respond by noting that concepts, however weakly perceived, and
judgments, however confused and indistinct, are nevertheless, by their very nature,
conscious representations (premised on the analytic unity of the I think: B133--4n,
ch 9-C) and so must actually be thought (A103-4) before they can perform their
role of bringing the manifold of formally intuited space and time to the objective
unity of apperception.
The synthesis speciosa of the relational and modal categories suffices to synthesize a
universal nexus (community) of spaces and times, each uniquely differentiated from
and completely determined in relation to every other. Saliently lacking are quantita
tive and qualitative relations between these spaces and times. They are not points or
instants, not one, few, or many, not parts composing wholes or wholes composed of
parts, not continuous or discrete, and not numbers. Nor is this surprising: the cate
gories have only logical functions for content, and transcendental synthesis speciosa
differs from synthesis intellectualis only by the narrowed scope-to the manifold of
our pure-formal intuitions in particular, space and time, rather than intuitions in
general-of the concepts they determine with respect to the logical functions. Yet
one could equally well say that the difference between these two categorial synthe
ses is vast: only in synthesis speciosa does pure understanding pass directly to objec
tifying determinations of the appearances themselves insofar as all are juxtaposed
or successive. And so our question now is this: how are the logical relations of
quantity and quality between concepts that the categories of quantity and quality
render irreversible realized objectively through transcendental synthesis speciosa as
determinations of pure juxtaposition and succession in formal intuition?
In the case of quantity, the answer emerges when we recall the difference between
applying the logical functions of judgment purely logically (ch 10-B) and apply
ing them transcendentally (ch 11-A). Universality being the form of concepts in
Kantian pure general logic, all concepts as such are universal. But this does not
mean "universal" in the sense of being inherently specifiable (into species concepts,
subspecies concepts, etc.). Concepts are specifiable only insofar as content is omit
ted (abstracted) from the completely determinate intuitions from which they derive
by a process of comparison, reflection, and abstraction (ch 9-B). If no content is
omitted, the concepts necessarily remain unspecifiable, for they are then, quite lit
erally, fully as individually unique as the intuitions from which they are formed.
Since logical universality, according to Kant's account (ch 9-C), consists simply
432 Kant's Anatomy of the Intelligent Mind
and solely in attaching the analytic unity of the I think to any non-universal repre
sentation (intuition, sensation, desire, etc.), concepts of the manifold of formally
intuited space and time can be formed without any need for comparison, reflec
tion, and abstraction, much less language. As shown in Section C, this (1) makes
them impossible to multiply instantiate (intrinsically unspecifiable), (2) makes their
determinations of juxtaposition and succession just as ineluctably singular, and
(3) makes all the categorical judgments formed from the former as subjects and the
latter as predicates singular in logical quantity rather than universal or particular.
It is these categorical judgments that are represented in transcendental synthesis
speciosa as final subjects in space with determinations ("accidents") formed into
can be formed of a series of such relations that does not incorporate concepts
(AUA-universals) of all its members, represented in the order that reflects how each
immediately grounds or is grounded by the other members of the series. Thus, the
concept C1 of the series a--->b--->c necessarily contains concepts of a--->b and b--->c
represented in the order a--->b--->c; the concept C2 of the series a--->b--->c--->d necessar
ily contains concepts of a--->b b--->c and c--->d represented in the order a--->b--->c--->d;
and so on. The understanding can then compare its concepts of different series with
regard to the quantity of concepts ordered in them and represent each as a quantity
of the other in judgments. And this is where the synthesis speciosa of the categories
of quantity comes in: it irreversibly fixes the relation of their quantities in judg
ment so that cl is determined as a quantity of c2 (never the reverse), c2 (and so too
C1) as a quantity of C3 (never the reverse), etc ., thereby establishing a quantitative
order (ranking) comprising the infinite totality of concepts of series of irreversible
ground-consequence relations.
The spatial and temporal significance of this synthesis must be understood in
relation to that of the synthesis speciosa of the relational and modal categories.
The latter, as we have seen, orders all spaces from least to most remote from one
another, not directly but rather by connecting their determinations in irrevers
ible ground-consequence relations, and by doing so in as many ways as each is
thereby determinable in relation to the infinite totality of all the others. The unidi
rectional, never intersecting causal lineages leading linearly to each determination
from its grounds in order of relative immediacy and from that determination to
its consequences also in order of relative immediacy has the temporal significance
of a succession of unidirectional, one-dimensional times leading to and from the
time of every determination, thereby determining its position in formally intuited
time: which times have to be traversed in what order before its time is reached and
which times are traversed in which order after that (Section C). Insofar as deter
minations in these lineages belong to different spaces (i.e., distinct final subjects of
determination in space), they have the additional spatial significance of ordering
all spaces as determinately adjacent or not in relation to each of the others, which
spaces have to be traversed in what order before any of nonadjacent spaces can be
reached and which spaces in what order are traversed after that.
Still, the order introduced by relational and modal synthesis speciosa into the
manifold of pure-formal space and time does nothing to determine spatial or tem
poral quantity. The advance through a series of irreversible ground-consequence
relations in and of itself neither increases the quantity of spaces and times tra
versed nor establishes an order of increase. In the case of space, this can be
rectified only by means of a series of quantitative concepts representing all of
the spaces traversed to reach each successive space in a series of spaces (consti
tuted, as described earlier, by concepts Cl' C2, C3, etc. of the series of irreversible
ground-consequence relations between determinations of the final subjects con
cerned) so that each concept is distinguished from all the others by the quantity
of spaces it contains, while their serial order, reflecting as it does the order of
434 Kant's Anatomy of the Intelligent Mind
What this means in the case of formally intuited time can be understood by con
sidering a successive (alphabetical) advance from final subject (space) A to final
subject (space) Z. Though the route from A to Z is irreversible because the ground
consequence relations from which it is forged are irreversible, a direct return from
Z to A via all the same subjects (spaces) is nevertheless possible through a series
of irreversible ground-consequence relations connecting different determinations
of the same subjects in reverse order (Section C). But although the advance from
the first series to the second results in retracing the same spaces originally tra
versed, it has no such implication with respect to the time, which instead con
tinues always through new times in its advance from the first series (forward) to
the second (reverse). Since the determinations involved in the irreversible ground
consequence relation connecting Y to Z are necessarily different from the deter
minations involved in the irreversible ground-consequence relation connecting
Z to Y, there is no moving backwards through the same times. On the contrary,
insofar as the two series are thought to fall in non-overlapping times, the times
comprised in the later will all be in advance of the times making up the earlier.
Moreover, since the series that retraces the order of spaces (final subjects) can
always be thought to begin with a determination simultaneous with the determina
tion of the same final subject that terminates the first series, the advance of times
can be represented as continuous and, insofar as the tracing and retracing can be
thought to repeat without end (though always through different determinations
of these final subjects), as both continuous and perpetual. So when this or any
other similar continuous perpetual advance is represented as a continuously and
perpetually increasing quantity of times and is then determined conformably to the
synthesis speciosa of the categories of quantity as irreversible both in respect of its
contents (all preceding times) and the order of their occurrence in the series, the
result is none other than number: "a representation that binds together the succes
sive addition of (homogeneous) one to one .. . whereby I produce time itself in the
apprehension of an intuition" (Al42-3/Bl82).
This binding together, however, always requires the counting of a number of
spaces equal to the number of times counted:
To understand why space is as necessary to number as time, recall that the successive
addition of homogeneous one (time) to one (time) bound together in the representa
tion of number is, in Kant's transcendental synthesis speciosa, nothing but a series
of irreversible ground-consequence relations. Once such a relation is determined as
irreversible, it becomes necessarily the case that the consequence can only be actualized
after its ground, impossible for it ever to be actualized together with it, "simultane
ously" as it were (i.e., at the same rather than a subsequent position in the irreversible
order of grounds and consequences). In order for it to be actualized together with
its ground, the consequence would have to be actualizable outside its order in the
ground-consequence series, which, once the order has been determined as irreversible
by cause-effect synthesis speciosa, is objectively impossible (Section C). Since the times
concerned in the synthesis speciosa of the categories of quantity directly correspond to
irreversibly related grounds and consequences, they are therefore precluded from being
the plurality drawn together in a representation of number. Thus, in order for the bind
ing together essential to the synthesis speciosa of number to be achieved, each ground
and consequence in the series must be correlated one-to-one with items completely
distinct from them that are connected neither in irreversible ground-consequence rela
tions nor in any other kind of irreversible subordinative relation that would prevent
them from all being given together, simultaneously.17
The manifold of spaces determined according to the synthesis speciosa of com
munity meets this condition. For although their differentiation from and relation to
one another is just as much a function of irreversible ground-consequence relations
as that of times (Section C), it is as the final subjects of the determinations connected
in these relations and not these determinations themselves, 18 that transcendental syn
thesis speciosa uniquely differentiates each space from and completely determines its
relations to every another. This difference is crucial to number as a binding together
of the homogeneous, for it permits final subjects of determination in formally intu
ited space to retain their spatial character as a permanently co-actualized manifold
of coordinates (Section B) instead of (like times) subordinates actualized in irrevers
ible, never-intersecting linear order.19 Since this is just to say that the homogeneous
17 This, I believe, underlies Kant's assertion that "it can easily be proved that the possibility
of things as quantities, and so the objective reality of the categories of quantity, can only be
displayed in outer intuition, and by its means alone can afterwards be applied to inner sense"
(B293). It also explains why he repeatedly insisted that time cannot be exhibited in intuition oth
erwise than by drawing a line: only thus can time be apprehended even as a quantity (number),
much less a quantum (extensive magnitude). See ch 6-C.
8
1 This appears to inform Kant's distinction of space from time in the following: "the one
properly concerns the intuition of an object, while the other concerns its state, especially its rep
resentative state. Thus, space is also applied as an image to the concept of time itself, representing
it by a line and its limits (moments) by points" (ID 405).
19 Extended to extensive magnitudes, this is just to say that "the parts of space are not subor
dinated to one another but rather coordinate" (A412/B439). Space itself is "the intuitively given
possibility of universal coordination" (ID 407). "Space contains the form of all coordination in
intuition, time of all subordination" (AA 18 § 5886 [late 1770s or early 1780s]); "space and time
A Category-by-Category Elucidation 437
manifold set "next to one another in space" not only can but "must be thought as
given simultaneously, i.e. as drawn together into one representation," only spaces, not
times, can constitute the successively enumerated plurality "draw[n] ...together in the
unity of a number" (AA 18 § 6314). Thus, on Kant's account of number, the spaces
and times generated by transcendental synthesis speciosa are both equally indispens
able: times to continuously and perpetually enumerate the quantity of spaces ("the
successive addition of (homogeneous) one to one ...whereby I produce time itself"),
and spaces to allow what thereby is successively enumerated to constitute a continu
ously and perpetually increasing objective plurality that can always (i.e., at any step in
the enumeration) be drawn together in the objective totality of a number ("a repre
sentation that binds together the successive addition of (homogeneous) one to one,"
Al42-3/B182).20
The concept of number incorporates the synthesis speciosa of all three catego
ries of quantity. Its fundamental unit is given by the individual immediate irrevers
ible ground-consequence relations that compose the number series, so that each
such relation in the sequence measures (counts) off the next number ("measure"
being the alternative denomination for the category of unity at PFM 303). But it is
the synthesis speciosa of totality that is most essential ("number ...belongs to the
category of totality," B111). To represent the number five, for example, is not just
to proceed five steps in the series and represent the result as a plurality. To think
this plurality ("magnitude": PFM 303) as the number five, one must combine it
with the category of unity to represent it as incapable of being further added to,
and so as a totality ("whole": PFM 303). The addition of further units as the series
advances beyond five must then be conceived either as a number (or numbers)
added to five or as constituting a new totality (whole), that is, a higher number in
the series (ch 11-A).21
contain merely the form of composition (coordinatio, et subordinatio) of the manifold of intu
ition" (AA 22 413), and are "the form (Formale) of placing together (coordination et subordi
natio) in an aggregate of objects of the senses that in accordance with a principle of complete
determination [has] the tendency to a system of the manifold, i.e. for the sake of the possibility
of experience (which can only be one)" (42). In the case of the manifold of formally intuited
space and time, community/necessity synthesis speciosa is the principle of complete determina
tion (Section D).
20
That space and time are both necessary for quantitative representation was of the first
importance for Kant: "The necessity of the connection of both sensible forms, space and time, in
the determination of the objects of our intuition, so that time, when the subject makes itself into
an object of representation, must be represented as a line in order to cognize it as a quantum, as
conversely a line can only be thought as a quantum by constructing it in time-this insight as to
the necessary connection of the inner sense with the outer even in the time-determination of our
existence seems to me to provide assistance for the proof of the objective reality of the represen
tations of outer things (as against psychological idealism)" (letter to Rehberg, before September
25, 1790). Clearly, empirical psychological idealism is meant: chs 7 and 8.
21
This can explain why Kant regarded all numbers as finite: "the concept of a number .. . is not
always possible when the concepts of multiplicity (Menge) and unity are combined, e.g., in the
representation of the infinite" (Bl 11).
438 Kant's Anatomy of the Intelligent Mind
22
Also: "The origin of concepts from sensible representations or the understanding belongs to
psychology and transcendental philosophy" (AA 15 § 1697 [1770s]). See discussion in Chapter 2.
2
3 These considerations explain why Kant was in a position to insist on the sensible character
of all number concepts while conceding to a mathematician correspondent that "time has no
influence on the properties of numbers ... and the science of number, without regard to the suc
cession that every construction of magnitude requires, is a pure intellectual synthesis that we
represent in thought" (letter to Schultz, November 25, 1788). Transcendental synthesis speciosa is
just as intellectual as synthesis intellectualis inasmuch as both exemplify philosophical-discursive
synthetic a priori cognition rather than the properly intuitive sort (A712-38/B740-66). What
A Category-by-Category Elucidation 439
Are concepts of number possible given formal intuitions other than our own?
Although Kant refused to deny (or affirm) the possibility of such intuitions, he said
nothing I know of to indicate whether, if they were possible, the synthesis speciosa
of the categories of quantity would issue in representations of number or something
else that, with the present constitution of our sensibility, can be neither intuited nor
conceived. All that can be said, I think, is this: it could very well be that number
concepts, although not of spaces and times, are only possible given the coordinative
character unique to space and the subordinative character unique to time. Even
a small departure from these, or the addition of third formal intuition (or more),
might very well preclude number concepts and set something else quite different in
their place, perhaps with the same or even greater mathematical potential.
Like the transcendental schema of the categories of quality and the Anticipations
of Perception principle of pure understanding, the synthesis speciosa of formal
intuition conformably to these categories must be understood purely quantitatively
since, according to Kant, qualities can never be given except empirically. However,
with the material side of intuition here excluded, categorial quality cannot manifest
itself as intensive magnitude. How then does the synthesis speciosa of the categories
of quality proceed in relation to formally intuited space and time?
One might be tempted to answer that the realities are the uniquely differentiated,
completely determinate, quantifiable spaces and times produced by the synthesis
speciosa of the other categories. But then what would be the negations? Empirical
realities like light, sound, or impenetrability have theirs, but what is the negation
of a space or a time? A hole in space, a gap in time? Since transcendental synthesis
speciosa determines that the entire manifold of formal intuition is comprised of
uniquely differentiated, completely determinate spaces and times, no negations, no
differentiates them is that no manifold is actually given in synthesis intellectualis, and so there is
as yet nothing outside the understanding to which to attach the analytic unity of apperception
which can then be determined categorially. Only actual formal intuitions furnished by pure sensi
bility can make good this want, and so only transcendental synthesis speciosa can yield number.
To grant this, however, raises the further question why number and numbers are transcendental
and not mathematical. They would be mathematical if actual spatial and temporal synthesis were
involved in their production, for then, like arithmetical determinations, they would be construc
tions in pure intuition (albeit still not concepts of spaces or time, i.e. not extensive magnitudes: ch
6-C). However, it should now be clear that the synthesis speciosa of the categories of quantity
responsible for generating numbers is a species of philosophical-discursive synthetic a priori cog
nition, not the mathematical-intuitive kind. This changes only with pure and empirical applica
tions of the transcendental ordering scheme provided by the synthesis speciosa of the categories
of quantity, e.g., to count apprehended spaces and times, apples and oranges, mental actions and
passions, or the fully abstracted objects of quantification theory (ch 6-E).
440 Kant's Anatomy of the Intelligent Mind
absence, can be met with in the manifold of formally intuited space and time. This
leaves nothing for the second category of quality, negation, to signify ("Negation
is the lack not of things but of the determination of things," AA§ 5815 [1783--4]).
Moreover, given that the third category, limitation, is "reality combined with nega
tion" (Bl 11), no synthesis speciosa of negation would make limitation impossible
as well. One must therefore look elsewhere for the locus of the synthesis speciosa of
the categories of quality.
Nothing can limit space except other space. This suggests the following con
strual of the synthesis speciosa of the category of limitation: a space X that contains
the ground of the necessity that another space Y be situated next to it objectively
limits space Y to X's immediate proximity, and space Y limits space X in the same
way if it also contains the ground of the necessity that X be situated next to it.
Given the basis of this necessity in irreversible ground-consequence relations, this
is just to say that final subject X contains a determination that immediately grounds
a determination of final subject Y and Y contains a (different) determination that
immediately grounds a (different) determination of X. The same synthesis also
determines a space X that contains the ground of an entire series of consequences,
each involving a different final subject of determination in space, as setting limits
to the remove from it in space of each of these subjects (Section C). For X limits
the number of removes at which any space related to it in an irreversible ground
ing series can be to the number of removes at which X grounds it: the subject next
in the irreversible series of consequences can be no nearer or farther from X than
one remove; the one after no nearer or farther than two removes; and so on. At the
same time, X itself is limited at all removes by the spaces that precede it in irrevers
ible grounding series. And since the synthesis speciosa of community and necessity
relates each final subject of determination in space to every other through such
grounding series, the synthesis speciosa of limitation necessitates that every space
limit and be limited by every other at some determinate remove.
In addition to determining grounds as !imitative of their consequences in series of
irreversible ground-consequence relations, the synthesis speciosa of limitation also
determines their final consequence as the limit toward which all prior consequences
in the series approach. The last in a series of irreversible ground-consequence rela
tions is the limit of the series because it is there that the grounding initiated by the
first relation in the series inevitably (necessarily) ends up, there that the grounding
extended by the second in the series must end up, the third and so on for every rela
tion in the series leading to this last. In temporal terms, the final consequence is
the time at which the series of times preceding it inevitably ends up, and, in spatial
terms, it is the space at which the series of spaces inevitably arrives.
Yet the full significance of the synthesis speciosa of the categories of quality
only becomes clear with negation. The serial passage through the limits of adja
cent spaces and times to reach a terminating limit is a succession composed of
a sequence of irreversible ground-consequence relations. Yet the same thing that
requires the judgment in the consequence position of the first relation in such a
A Category-by-Category Elucidation 441
series to be referred to an antecedent ground necessitates that its ground too, albeit
stipulated first in the series, be referred to a ground antecedent to it upon which
it necessarily follows: the subordination of the infinite totality of determinations
of final subjects to the synthesis speciosa of the categories of cause and effect and
possibility-impossibility, making all merely possible until the actualization of a
grounding determination necessitates their own actualization (Section C). Since
this is just to say that there is no determination whose actualization is not con
ditional upon the actualization of another, the possibility of an absolute first in
the series of grounds is precluded. The first ground of every series of irreversible
ground-consequence relations must therefore itself be grounded, thereby negating
its status as a limit of everything subsequent in its series. The first in any such series
is therefore always only a stipulated first, never objectively, in its own right, first.
And this means that there is and can be no objectively first space from which entry
into any other space (much less all spaces) must proceed, or a first time from which
more than one time (much less all times) must begin.
The same is true of the limiting final consequence approached by other members
of a series of irreversible ground-consequence relations. Given the inexhaustible
infinity of determinations of final subjects present in formally intuited space and
time, it can never be impossible to form a hypothetical relation connecting every
determination to some other as its consequence. Since this includes determinations
that, as final consequences, are approached by every member in the series leading
up to them, they too are related as ground to some other determination as their
consequence. Because the immediate ground of each is determined by the synthe
sis speciosa of cause and effect as its unique ground and that consequence as its
unique consequence, this means that all final consequences of series are necessarily
determined by this synthesis as themselves the immediate ground of another deter
mination that falls outside its series, thus negating its status as the terminating limit
approached by every member of its series. In other words, if it is last in its series, it
is only a stipulated, not an objective, end . And this means that there is and can be
no final space to which other spaces (much less all spaces) lead but no further, and
no final time toward which other times (much less all times) proceed but no further.
By determining all limits within formally intuited space and time as stipulated
rather than objective, the synthesis speciosa of negation perfectly captures the prop
erty of these intuitions as conditions that precede and make possible all spaces
and times within them. This property is prominently featured in the metaphysical
expositions of the Transcendental Aesthetic:
[I]f one speaks of many spaces, one thereby understands only parts of
a single space that unites them all. These parts also cannot precede the one
all-encompassing space as if they were its constituents and space were pos
sible only through their composition, but instead are only thought in it. Space
is essentially one; the manifold in it, along with the general concept of spaces,
rest solely on limitations. (A25/B39)
442 Kant's Anatomy of the Intelligent Mind
The infinity of time signifies nothing further than that every determinate
magnitude of time is possible only through limitations of a single underlying
time. (A32/B47-8)
Since "it is only when both infinite space and infinite time are given that any deter
minate space and time can be specified by limiting" (ID 405), it is easy to see that
space and time themselves can have no real limits in them (what precedes limits and
first makes them possible can never be given with those limits already in it, and so
cannot really and objectively have these limits). It is this purely aesthetic (predis
cursive) feature of having no real, objective limits in them that is given objective
force by the synthesis speciosa of negation: it ensures that there can be (1) no first
space from which entry into any other space (much less all spaces) must proceed,
(2) no first time from which more than one time (much less all times) must begin,
(3) no end to the series of spaces (no final space), and (4) no end to time (no last in
the series of times). Like formally intuited space and time, all limitation within the
universal network (community) comprising the inexhaustible infinity of uniquely
differentiated, completely determinate spaces and times produced by transcenden
tal synthesis speciosa is never a real, objective limitation in that network itself, the
network that precedes and makes limitation itself possible. Thus, the community of
spaces and times produced by transcendental synthesis speciosa from the manifold
of pure-formal space and time is infinite in precisely the sense prediscursive for
mally intuited pure space and time themselves are.24
What then is the synthesis speciosa of the category of reality in respect to for
mally intuited space and time? In empirical-material intuition, sensation confers
objective reality on space and time through the matter (of appearance) correspond
ing to it, which fills and thereby delimits a space and time within the space and
time wholes. It seems to me that the only possible counterpart in pure-formal intu
ition is the reality conferred on a space or time by non-transcendentally stipulating
its defining boundaries (points, lines, planes, etc.), that is, ostensive and symbolic
mathematical construction (ch 6). Constructions do not of course amount to real
objective limits in the community of formal spaces and times since one can just as
easily re-stipulate or unstipulate whatever boundaries one defines. But the same is
true of the material boundaries defined by the matter corresponding to sensation in
the community of empirical-material spaces and times (ch 16-C), as every example
of movement demonstrates in regard to spatial boundaries and every delayed or
preempted effect demonstrates in regard to its temporal boundaries. In both cases,
the objective reality of spatial and temporal limits does not consist in any necessity
24 This feature will figure crucially in the Analytic of Principles in justification of Kant's
claim that the infinity of space and time is present in the field of appearances even though
space and time, as pure and so imperceptible (not giveable through sensation), can never them
selves be present there. The categories, as the basis of an infinite community, are determined
by their schematism as exponents (conceptual surrogates) of the infinity of space and time in
empirical-material intuition (A215-6/B262-3). Detailed analysis is provided in Part V.
A Category-by-Category Elucidation 443
they impose on the presupposed community of spaces and times constituted by the
synthesis speciosa of the relational and modal categories, but rather in the necessity
they impose on the quantitative ordering of those space and times.
The quantitative significance of the synthesis speciosa of the categories of quality
can now be described. If an arbitrary first and last limit are stipulated at an infinite
remove from one another in an infinite series of irreversible ground-consequence
relations (i.e., there are infinitely many such relations between them as well as before
the first and after the last), not only will the spaces (or times) corresponding to these
limits ipso facto be situated at an infinite remove in the series but so too will all those
that follow the first by a finite serial (successive, one by one) progression and pre
cede the last by a similarly finite serial regression.25 Of those spaces (or times) at an
infinite remove from the first, some will also be at an infinite remove from the last;
and, of these, some will be at infinite removes from one another as well. Given the
inexhaustible infinity of spaces (and times), the number of intermediates at infinite
removes from one another will likewise be infinite. It is therefore possible to utilize
the synthesis speciosa of quality to form a concept of a new kind of magnitude: a
bounded (i.e., delimited) space (or time) that consists entirely of spaces (times) that
are each infinitely removed from the bounding limits and from one another. As with
the transcendental schemata, this concept is at once sensible because it represents
spaces and times and purely discursive in that it "can never be met with in any
intuition," "be itself intuited through the senses," or be "contained in appearance"
(AI37-8/BI76-7) and so "can never be brought into any image whatsoever" (AI42/
BISI). Yet it is precisely the concept needed to make possible constructible and
empirical intuitions of continuous magnitudes
Continuity is the "property of magnitudes in virtue of which no part is the
smallest possible (no simple part)" (AI69/B2II). It applies to pure space and time
25 "The true (transcendental) concept of the infinite is that the successive synthesis of a unit
can never be completed in measuring off (Durchmessung) a quantum" (A432/B460). To this Kant
attached a footnote: "This quantum thereby contains an amount (Menge) (of a given unit) that is
greater than any number, which is the mathematical concept of the infinite." In the transcenden
tal concept of infinitude yielded by synthesis speciosa, the unit space and unit time correspond to
a "unit" irreversible ground-consequence relation and infinity to an unlimited (never beginning
and/or never ending) series of irreversible ground-consequence relations. The mathematical con
cept, by contrast, presumably concerns the number of a given (i.e. intuitable) unit-appearance of
some kind (see note 21 above).
444 Kant's Anatomy of the Intelligent Mind
26
The grounding of continuity in infinite series of irreversible ground-consequence relations
in which limits are stipulable seems to underlie Kant's claim that "just as time contains the sen
sible a priori condition of the possibility of a continuous advance from an existent to the existent
which follows it, the understanding, by means of the unity of apperception, is the a priori condi
tion of the possibility of a continuous determination of all positions for appearances in this time
through the series of causes and effects, wherein the first lead inexorably to the next, and thereby
render empirical cognition of the time-relations for each time (universally) objectively valid"
(A210-l l/B255-6).
27
In his discussion of Kant's conception of logic, MacFarlane espouses a reading that goes
back to Frege and beyond: "For Kant, the only way to represent denseness [(A relation R is a
dense ordering just in case (Vx)(Vy)(Rxy--> (3z)(Rxz & Rzy))] is to model it on the infinite divis
ibility of a line in space. As [Michael] Friedman explains, 'denseness is represented by a defi
nite fact about my intuitive capacities: namely, whenever I can represent (construct) two distinct
points a and bon a line, I can represent (construct) a third point c between them.' What Kant can
represent only through construction in intuition, Frege can represent using vocabulary he regards
as logical." John MacFarlane, "Frege, Kant, and the Logic in Logicism," Philosophical Review
111, no. I (2002): 64. By contrast, on my interpretation, mathematically constructed intuitions
are irrelevant because transcendental (philosophical-discursive) synthesis speciosa generates con
tinuous magnitudes that are capable of fully capturing what the mathematical-logical (which, as
I have argued, is mathematical rather than logical by Kantian criteria) notion of "denseness" is
supposed to convey. See also chs 6-C, 10, 14-E, and UU ch 4-B.
A Category-by-Category Elucidation 445
G. Results
Kant seldom features prominently in histories of the philosophy of space and time,
and is often overlooked entirely. But that is the dogmatic Euclidean nativist of lore,
not the Kant presented here. According to the former, the space given via mere
receptivity to the infant mind is (1) completely differentiated and determinate prior
to and independently of categorial synthesis and (2) in perfect conformity with
all the parameters of Greek geometry prior to and independently of all properly
mathematical synthesis (i.e. the formation of geometrical concepts and their exhibi
tion in intuition via construction). That a view which limits space to a Euclidean
space spawned fully formed by human sensibility would elicit little interest from
philosophers of space is not surprising. But it has an even greater shortcoming. So
long as it is believed that that space and time are fully differentiated and determined
by mere receptivity, prior to and independently of the categories, one will naturally
look elsewhere to comprehend the role these concepts play in Kant's philosophy.
So if their primary role is indeed to differentiate and determine the manifolds of
space and time, not only will that be overlooked, it will be irremediably obscured by
whatever is set in its place, while the latter, given the centrality of the categories to
Kant's philosophy, may have the effect of vitiating one's entire interpretation of it.
Last, but by no means least, the exclusion of the categories from Kant's account of
the space and time presupposed by mathematics and mathematical science cannot
but deprive the latter of precisely that which is most likely to excite the interest of
historians of the philosophy of space and time: a detailed, category-by-category
elucidation of the constitution of objective space and time that shows them to com
patible with mathematics of every kind, non-Euclidean geometry not excepted.
In the version presented in this chapter, objective unity of apperception is cen
tral. In Part II, I argued that if pure space and time may be said to bring objective
unity of apperception to their manifolds, it is only to the extent that all appearances
have a necessary relation to them as the conditions of their possibility (ch 5-D). Yet
in conforming to the objective unity of space and time, not only do appearances
lack the necessary relation to one another (interconnection, Zusammenhang) that
Kant deemed essential for cognitive objectivity (ch 13-G and -H), they are devoid
of relation and order of every kind ("scattered and single in themselves," Al20).
The properly cognitive objective unity of apperception of the manifolds of formally
intuited pure space and time thus depends entirely on categorial synthesis and is
philosophical-discursive in nature.
In Chapter 13, I argued that the objective unity the categories bring to the mani
fold of pure space and time consists precisely in making good the want of dif
ferentiation and determination in pure space and time, at least to the point where
actual intuitions of the objects thinkable in non-transcendental concepts, both pure
(mathematics and metaphysics of nature) and empirical (ordinary and scientific,),
become possible. I also contended that it was precisely on this ground that Kant
could claim that the categories are valid a priori not only of our experience of
446 Kant's Anatomy of the Intelligent Mind
objects but also of the objects themselves, their laws, and nature itself as their sys
tematic totality.
However, as noted in this chapter's introduction, there is an obstacle in the way
of applying the categories to the manifold of intuition, pure (formal) or empiri
cal (material): no content is thought in the categories other than logical functions
of judgment, which restricts their application to representations that have already
been made universal (concepts and judgments) and so seems to preclude the pos
sibility of their serving to differentiate and determine the manifold apprehended in
pure-formal intuition. A solution to the problem emerged when we factored in the
account of concept acquisition that receives its most extended elaboration in§ 16 of
the B edition Transcendental Deduction of the Categories. As shown in Chapter 9,
a nonconcept forthwith becomes a concept, carrying universal scope, as soon as the
analytic unity of the I think is incorporated into its representation. Since the way in
which a representation acquires the form of a concept on this conception bypasses
the need for comparison, reflection, and abstraction, it becomes possible for the
understanding to represent the inexhaustible infinity of both the representations
possible in pure intuition and the possibilities of their juxtaposition or succession
with one another in an equal number of transcendentally singular concepts sim
ply by attaching the representation of the identity of the I think to their intuition
(for "the I think must be able to accompany all my representations" or they would
"either be impossible or at least be nothing for me," B131-2). Once this is done, the
logical functions make it possible to represent these concepts as combined in an
inexhaustible infinity of singular judgments, thereby providing the categories with
precisely the inputs they require if they are to be able to confer objective unity of
apperception on the manifold of pure intuition: (1) concepts and judgments whose
logical position they can determine conformably to the logical functions of judg
ment and (2) a manifold of transcendentally singular concepts and singular judg
ments capable of representing the totality of intuitions possible within pure space
and time.
In the remainder of this chapter, the proposal was put to the test of whether it
could supply the means for a full, category-by-category elucidation of the transcen
dental synthesis speciosa of the manifold of the pure-formal intuitions of space and
time, thereby showing that and how this synthesis is able to non-amphbolously yield
an inexhaustibly infinite manifold of spaces and times that are at once uniquely
differentiated from and completely determined in relation to one another. To the
extent my attempt may be judged successful, it will have made clear just how the
categories can be understood as objective in the fully unattenuated sense of being
constitutive not merely of our judgments regarding the objects we meet with in
experience but of these objects themselves. But even to the extent it falls short,
there is one thing it nevertheless seems to me to put beyond doubt: an interpreta
tion of Kant can be judged successful only to the extent it is capable of solving
the heterogeneity problem by producing a category-by-category elucidation of the
transcendental synthesis speciosa of the manifold of pure space and time. For only
A Category-by-Category Elucidation 447
451
452 Kant's Anatomy of the Intelligent Mind
particular regard to the Analogies of Experience, which both build on and super
sede the Humean precedent. Finally, in Chapter 18, I examine the conceptions of
mind and world that emerge from the objective principles (Axioms of Intuition,
Anticipations of Perception, and Analogies of Experience), ending with a consid
eration of the three subjective principles of pure understanding, the Postulates of
Empirical Thought.
{ 16 }
The Analytic of Principles is the second book of the Transcendental Analytic. Its
focus is judgment ( Urteilskraft), "the capacity to subsume under rules, that is, to dis
tinguish whether something stands under a given rule or not (the case of the given
law)" (A132/Bl 71). Kant treated judgment as a faculty distinct from understanding,
the capacity for rules (A126), on the ground that the act of subsuming under rules
cannot itself be subjected to rules. To appreciate why, it suffices to recognize that any
attempt to formulate a second tier of rules to aid in judging which cases are and which
are not instances of a given rule ultimately ends up depending on judgment to descry
which of the resulting judgments instantiate the second-tier rules and which do not.
Clearly, it is not understanding (i.e., third-tier rules) that is wanted but some
thing else. In the case of empirical concepts, where good judgment "is a particular
talent that cannot be taught but needs to be exercised" (A133/B172), there is no
universally valid answer to the question of what that something is. But matters are
different when it comes to pure concepts of empirical objects:
[T]ranscendental philosophy has this peculiarity: that, outside the rule (or
rather the universal condition for rules) given in the pure concept of the
understanding, it can, at the same time, indicate the case a priori to which it is
to be applied. The cause of the advantage which it has over all other instruc
tive sciences in this regard (except mathematics) lies in this: it deals with
concepts that are to relate to their object a priori, which consequently can
not have their objective validity established a posteriori, . . . but must rather
set forth in universal but, at the same time, sufficient criteria the conditions
under which objects in agreement with those concepts can be given, failing
which they would be altogether devoid of content, hence mere logical forms
and not pure concepts of the understanding. (A136/Bl 75)
The mathematics exception is telling. Without calling upon the understanding for
guidance in the form of special subsumption rules, judgment can utilize the figurative
synthesis of productive imagination to construct objects corresponding to mathe
matical concepts in pure intuition that can then be employed by empirical imagina
tion as templates ("universal but sufficient criteria") for determining which empirical
454 Kant's Anatomy of the Intelligent Mind
objects do and do not fall under the concept. This means that any complex empirical
object (appearance) into which this pure object enters as a formal constituent is ipso
facto an instance of the concept and that even the (materially) least resembling phe
nomena may still instantiate it provided they incorporate that object (as the formal
mathematical component of the concept of gravity is instantiated by phenomena as
materially dissimilar as the fall of an apple, tides, and the orbits of planets).
But it is being embedded in a network of definitions, axioms, and postulates that
most signally sets mathematical concepts apart. While other concepts may form part
A. Transcendental Judgment
Judgment (subsumption) should not be confused with truth. The former is sim
ply the operation of comparing objects encountered in experience to the contents
thought in previously acquired concepts, subsuming them under whichever concept
or concepts they most closely approximate in their properties, relations, uses, and so
on. If one may speak here of "truth" at all, it is not the agreement of thought with
reality but of reality with thought. By contrast, if objective truth is the aim, then
new experiences may not only lead to changes in belief but also oblige us to replace
or modify the concepts used to frame them, resulting in changes in judgment as
well. For example, after discovering that the objects one judges to be gold always
resist rust, one may add this characteristic to the concept and thereafter subsume
accordingly. Even the application of (non-transcendental) pure concepts is answer
able to experience. For while empirical objects will be subsumed under the pure
concept of, say, a triangle to the extent that their synthesis incorporates the object
constructed in pure intuition in accordance with that concept, this applies only to
the images of these objects, not their reality. The road ahead, for example, may form
a visual triangle terminating at the horizon, but that does not mean the road itself
is triangular. To judge its true shape, causal information is required, particularly as
to how roads act on one's sense of sight from various standpoints.
Subsuming Reality 455
1 Hume came to a similar conclusion regarding applied mathematics (THN 413-4/265-6). See
HTC ch 4-C for discussion.
456 Kant's Anatomy of the Intelligent Mind
evidence, much less proof, of even so much as the possible truth of these judgments.
Yet without proof of their truth, the most one could claim is that concepts of the
necessary relation between the distinct (mathematical concepts included) are norma
tively indispensable to our thought of empirical realities, but not that these realities
themselves are actually connected together in a manner corresponding to these con
cepts (PFM 258-9). To assume otherwise is sheer question begging.
T he Schematism and System of Principles of Pure Understanding are Kant's
attempt to meet Hume's challenge by providing a priori proof that transcendental
judgment is equipped to subsume empirical realities under pure concepts of the
necessary synthetic unity of the manifold (the necessary relation of the distinct).
He did this by applying the transcendental synthesis speciosa established in the
Transcendental Deduction of the Categories to the special case of empirical reali
ties, understood as that in appearances which corresponds to sensations (i.e. , the
matter of appearances).2 To "apply" in the case of judgment, unlike that of under
standing, means to subsume empirical realities under this synthesis. Transcendental
synthesis, it will be recalled from Part IV, serves to transform an indeterminate,
undifferentiated manifold into an objectively united "community" of spaces and
times wherein each space and each time is uniquely differentiated from and com
pletely determined in relation to every one of the infinitely many others possible in
formally intuited space and time. Because empirical realities originally lack spatial
and temporal differentiation and determination, it falls to transcendental judgment
to form them into just such a community by subsuming them under transcendental
synthesis speciosa by means of transcendental schemata ("the sensible concept of
an object in agreement with the categories," Al46/B186). Since non-transcendental
judgments regarding empirical reality invariably presuppose that the objects to
which they relate are spatially and temporally differentiated and determined, the
possibility of their truth (agreement with perceptible reality) depends on this. T hus,
transcendental judgment becomes the source of "the transcendental truth that pre
cedes all empirical truth and makes it possible" (Al46/B185).
Even given transcendental idealism (ch 7), how can it be known, and known entirely a
priori, that realities, which can only be given a posteriori in perception, lack all spatial
2 "Sensation (here, outer) equally expresses the merely subjective of our representations of
things outside us, but properly their material (real) (whereby something is given as existent),
just as space expresses the mere a priori form of the possibility of their intuition; nevertheless,
sensation too is employed in the cognition of objects outside of us" (CJ 188). Unless otherwise
indicated, when I refer to "empirical-material intuition," "material" does not signify the physical
or corporeal but simply the matter of appearances corresponding to affections of sense without
regard to whether the appearances concerned are spatial in form. Regarding Kant's treatment of
sensation as the basis of the reality/actuality of appearances, see chs 3 and 8.
Subsuming Reality 457
never be perceived)" (B207); "time cannot itself be perceived" (B219; also B225, A183/
B226, B233, B257); "time is not regarded as that wherein experience immediately
determines for each existence its position ...since absolute time is no object of percep
tion whereby appearances could be held together" (A215/B262); and "we can never
take up a manifold as such in perception without doing so in space and time but ...do
not intuit the latter for themselves" (AA 23 LXX at Al 64). Indeed, their imperceptibil
ity is the raison d'etre of transcendental principles as such:
5 Also: "[T]his space itself, together with that time, and, at the same time, all appearances, are
not in themselves things but rather nothing else than representations, and can never exist outside
our mind, and even the inner and sensible intuition of our minds (as object of consciousness),
the determination of which is represented through the succession of distinct states in time, is also
not the genuine self as it exists in itself-that is, the transcendental subject-but only an appear
ance that is given to the sensibility of this, to us unknown, being" (A492/B520); "Space is not
an object of intuitions (an object or its determination) but the intuition itself which precedes all
objects . ..a pure intuition a priori." (AA 17 § 4373 [1773-5]); "Space is not an object of intuition
but intuition itself" (AA 22 5).
Subsuming Reality 459
[S]pace and time, as pure as these concepts7 are of everything empirical, and
as certain as it is that they are represented completely a priori in the mind,
would nevertheless be without objective validity and without meaning and
significance (Sinn und Bedeutung) if their necessary employment were not
manifested in objects of experience ...Although we therefore cognize a great
deal in synthetic a priori judgments regarding space in general or the figures
that the productive imagination delineates in it, so that we are in no need
of experience for this; yet, this cognition would never be anything but an
occupation with a mere figment (Hirngespinst) if space were not viewed as a
condition of the appearances that constitute the material of outer experience.
Those pure synthetic judgments therefore relate, although only mediately, to
possible experience, or rather to the possibility of experience, and upon this
alone is the objective validity of its synthesis grounded. (A156-7/B195-6;
also B146-7 and A239/B298)
The thesis that pure, imperceptible space and time lack objective empirical real
ity in their own right might be thought to be contradicted by the empirical real
ity accorded to them in the Transcendental Aesthetic. Yet as we saw in Chapter 7,
Kant's point there was that appearances do not merely present an illusion of the
spatial and temporal but actually are, in their formal constitution, spatial and
6 What distinguishes pure space and time from fictions of imagination is that although they
are "mere thought entities and creatures of the imagination," they are, at the same time, "the
essential form of our sensibility and the receptivity of its intuitions whereby in general objects are
given to us, and because the universal conditions of sensibility must at the same time necessarily
be a priori conditions of the possibility of all objects of the senses as appearances and so agree
with these, they are not fictitiously invented by the imagination but underlie all its compositions
and creations" (Discovery 203). See Part II and KMM Part I.
7 This seems to be another instance of Kant's use of the term "concept" in a secondary sense,
i.e., the (second-order) concepts of space and time are concepts of nondiscursive intuitive rather
than (first-order) conceptual representations (see ch 3-C).
460 Kant's Anatomy of the Intelligent Mind
temporal. The issue of the objective reality of space and time posed by their imper
ceptibility, however, relates not to the formal but to the material constitution of
appearances. It is the question whether and how their pure, spatio-temporal formal
constitution can partake of their material reality corresponding to sensations and,
through them, to the "transcendental matter," "facticity, reality" of things in them
selves (Al43/B182).
The incommensurability between sensations and space and time should stanch
any temptation to imagine the matter corresponding to the former as filling the
latter the way light fills a dark room when the switch is flicked to the "on" position
or a silent hall is filled with music when the conductor's baton gives the signal to
play. Given how remote Kantian space and time are in every sense from the vis
ible, tangible, and everything else a posteriori in our representation (including inner
perception), the matter that fills appearances in them must not be confounded with
what fills the visual, auditory, or other "fields" of sensation (ch 3-D). Appearances
are simply the way sensations, however heterogeneous in quality, are exhibited
("appear") in the form of a homogeneous manifold, a manifold that has and can
have no content whatsoever in common with these affections and which therefore
can and must be represented completely a priori (ch 3). Until actual experience
enables us to represent their matter as gravity, solidity, or any other real efficac
ity, we are limited to understanding its correspondence to sensation in terms of
the only two features of sensations that admit of being represented a priori: their
manifoldness and their reality (actuality, existence) (ch 3-A). Accordingly, the mat
ter of appearances contrasts with their form as comprising a manifold of realities
as devoid of spatial and temporal quality and relations as of sensational qualia
(ch 3-B). And it is this that creates the problem formulated at the start of this sec
tion: since space and time are precisely that in appearances which do not exhibit or
correspond to sensations, and so lack the reality that is the only feature represented
in the manifold of empirical-material intuition, it is impossible for these manifold
realities to exist in (occupy, contain) space and time, and so too, it seems, impossible
for them to be subjected to transcendental synthesis speciosa.
Or is it? Unity of sensibility, as the original (prediscursive) synthetic unity of
all the manifold in one apperception, is the consequence of a reciprocal mediation
between appearances and sense affections: the appearances apprehended through
synopsized affections exhibit these same affections, howsoever heterogeneous, in the
form of a homogeneous manifold, immediately contained within the prediscursive
unity of a pure intuitive consciousness (pure space and time) (chs 3--4). Since the
matter of these appearances-the perceptible realities apprehended through sense
affections in empirical-material intuition-is as indelibly part of these appearances
as their constitutive forms-the manifold spaces or times apprehended in pure for
mal intuition-that matter too is encompassed in the original unity of sensibility.
This unity is none other than the prediscursive original synthetic unity of appercep
tion that precedes and makes possible the analytic unity of apperception (ch 9-B).
Since this is just to say that this prediscursive original synthetic unity is just as much
Subsuming Reality 461
The schema of substance is the permanence of the real in time, i.e. the rep
resentation of the real as the substratum of empirical time-determination in
general, which thus remains while everything else changes. ... To time, which
is itself unchangeable and remains, there corresponds in appearance what is
unchanging in existence, i.e. substance, and only in it can the succession and
simultaneity of appearances be determined conformably to time. (Al44/B183).
The successive and simultaneous accidents of substantial final subjects are those
realities whose concepts have been subjected to irreversible determination regard
ing their logical position in categorical judgments and must always be considered
as predicates (changes in a substrate) in relation to those realities whose concepts
have been determined as final subjects (substrates of change), never themselves
Subsuming Reality 463
predicable. In this way, the schema of the category of substance and accident
extends the transcendental synthesis speciosa of the manifold of pure space and
time to the matter of appearances, conferring empirical material reality on final
subjects of determination (ch 15-B) in the field of appearance and, with them,
by proxy as it were, time itself "as substratum, as permanent form of inner intu
ition" (B224).
2. The schema of cause and effect subjects all changeable empirical realities
to a rule according to which the positing of the existence of one in time neces
sitates that the existence of some other immediately follows it in time (Al44/
B 183). With all change determined as necessarily effected change, and everything
changeable irreversibly determined as the consequence of another changeable as
its ground, change itself constitutes an irreversible temporal succession in the
field of empirical-material intuition exactly mirroring the serial ordering of pure
times effected by cause and effect synthesis speciosa of the manifold of formally
intuited time (ch 15-C). At the same time, the schema serves to spatially differ
entiate and determinately relate the manifold of empirical realities that, under
the schema of substance, count as the final subjects of these causally connected
determinations.
3. Community subjects empirical realities to a rule of simultaneity according to
which substantial final subjects (the ultimate reality/matter of appearances) caus
ally determine one another's accidents (Al44/B183--4). Since causally related acci
dents always exist as a succession, only the substances in which they inhere can be
determined as simultaneous by this rule. This, however, raises the question of why it
should be necessary for substances that have already been determined as permanent
also to be determined as simultaneous; how can permanent existents possibly fail to
be simultaneous?
Permanent substances would be simultaneous if all were situated in a sin
gle, materially real time that, a la Newton, exists prior to and independently of
them (Section B). But it is precisely because no such time is possible when time
is construed as a pure, merely formal intuition that exists only in and through the
imagination's pure synthesis of apprehension that it becomes necessary for pure
understanding to synthesize permanent substances from the manifold of empirical
realities to function as surrogates (exponents)8 in empirical-material intuition for
the permanence of pure time. For precisely the same reason, however, the schema
of substance is incapable by itself of precluding the possibility that each empirical,
materially real permanent constitutes a distinct time unto itself, the substratum of a
succession completely alien to and isolated from the succession constituted by some
or all of the others. Hence there is the need for a schema of the category of cause
and effect: insofar as an accident that comes to exist in one permanent causes an
8 "Exponent" is Kant's term for the role of schematized categories as proxies of pure time in
the field of appearance. See A216/B263 and ch 18n8.
464 Kant's Anatomy of the Intelligent Mind
9 More precisely, the cause and effect synthesis speciosa can only ensure that the direct or
indirect interaction between all substances must ultimately come to pass. In the case of pure for
mal time, where the totality of times is given a priori and there is no question of their successive
elapse, this poses no problem. In the sphere of empirical-material intuition, however, a synthesis
that cannot necessitate that all substances permanently coexist in one and the same time, with
respect to the past no less than to the future, is incapable of determining all as belonging to (going
proxy for) a single, all-encompassing empirical materially real time. For unity of time in the field
of appearances requires that, at any point in time, all substances always already belong to one
and the same community, not that they may (or may not) subsequently come to do so, perhaps
even at a time infinitely remote from the present.
Subsuming Reality 465
are not just spatially and temporally but altogether, in every way, undifferentiated
and indeterminate (for they are not sensations, but that to which sensations corre
spond: chs 3-A, -B, and 8-E). Yet being even so much as a manifold is not nothing.
Insofar as all empirical realities are necessarily united in a single unity of sensibility
made possible by pure space and time, they are available for synthesis conformably
with the schemata of the categories of relation, which, by ordering them as deter
minations of substantial final subjects causally united in a single, all-encompassing
community, determine them exactly as if they existed in an absolute space and time.
As we saw in Chapter 15-E, it is precisely this all-encompassing network of causal
connections (irreversible ground-consequence relations) that the synthesis speciosa
of the categories of quantity requires, and their schematism is no exception. The
"manifold of a homogeneous intuition in general" (Al43/Bl82), including the
manifold of perceptible realities, is, minimally, neither two spaces nor two times
but the causal connection between two accidents of substantial final subjects in
space that leads necessarily from the space and time of one (final subject) into the
(adjacent) space and time of the other, together with the regressive and progressive
causal series continuing therefrom. The unit of number-the schema of unity, the
first category of quantity-is, accordingly, a single causal connection between acci
dents of substances in a series of such connections. These causally ordered series
themselves, which link substantial final subjects in space together in a community
and serve as the proxy for pure formal temporal succession in empirical reality, are
the schema of plurality. The schema of totality, or number proper (B111, ch 11-A),
is the conception of a finite causally ordered series (plurality) as incapable of being
added to (unity), thereby making it possible to represent the series as a continuous
increase in number (from one total to the next higher total to the next higher, etc.).
It thus gives empirical-material reality to the pure formal character of continuous
increase conferred by transcendental synthesis on the manifold of formally intuited
pure time.
5. The schemata of the categories of quality represent by far the greatest depar
ture from their synthesis speciosa counterpart in respect of the manifolds of pure
formal space and time, for they specifically concern sensations. Kant well appreci
ated how surprising it is that sensations should figure in a priori cognition of qual
ity at all:
[S]ensation is what is really empirical in our cognition, and the real of the rep
resentations of inner sense in contrast to their form, time. Sensation therefore
lies outside all a priori cognition. (AA 23 LX, marginal notation at Al43).
That particular sensations vary in multifarious ways is not something that can be
known a priori. That light and heat may intensify; sounds blend into harmonies
and fade away; tactual objects smooth out, warp, moisten and desiccate, and so on
are paradigmatically empirical judgments. That yellow is lighter than purple, that
a certain sweet taste pleasant but another bitter taste unpleasant, that one scent is
fragrant and another acrid are purely empirical judgments as well. So what remains
466 Kant's Anatomy of the Intelligent Mind
10
Once this sum-total is subjected to the categories, it becomes the all-encompassing unity of
experience (A581-2/B609-10) and indeed a sum-total in time: "Reality, in opposition to nega
tion, can only be explained if one thinks time, as the sum-total of all being, as either filled with
being or empty" (A242/B300). The reader should be sure to consult A581-2/B609-10, discussed
in connection with the synthesis speciosa of community in ch 15-D; for though it occurs in the
Transcendental Dialectic, Kant explicitly related it to the Analytic. Consequently, the notion of a
qualitative sum-total may-indeed must-be factored into one's interpretation of Kant's theory
of the understanding.
Subsuming Reality 467
u The result of negating all of the constituent empirical realities of an appearance is "pure
intuition = O," that is, "a merely formal (a priori) consciousness of the manifold in space and
time" (B208). Also: "The real is opposed to the merely formal in intuition as well. Therefore
empty space, where no being of anything (Sein von etwas) is represented" (AA 18 § 5822 [1785-8])
and "That (in representation) which is related to the object of the senses in itself (Gegenstand der
Sinne an sich) is sensation; but since the representation is related merely to the subject (according
to its quality), the object is a mere something in general. If I eliminate this something (sensation)
and, at the same time, the composition, there then remains the form of intuition and of the object
as appearance" (AA 18 § 6314 [1790-1]).
468 Kant's Anatomy of the Intelligent Mind
Now, every sensation has a degree or magnitude whereby it can fill one
and the same time more or less, i.e. fill inner sense with respect to the same
12
There are, however, two caveats:(!) we cannot anticipate a priori whether there is a cause
capable of altering the intensive magnitude of any particular quality given through sensation
(Al 71/B212-3), although (2) we can know a priori from the permanence of substance that the
sum-total of reality in the universe can neither diminish nor increase (Al 82, Al84-6/B227-9).
Subsuming Reality 469
The schema of limitation is the quantum of reality that always has to remain
after negation. For if absolutely all reality were to be eliminated from a space or a
time (i.e., a vacuum), it would become a pure, merely formal space or time and so
would cease to be a possible object of experience at all (Al 72-3/B214, A229/B281,
A521/549, and note 11 above).
Two final points before proceeding to the modal schemata. F irst, insofar as sen
sations correspond to "the transcendental matter of all objects as things in them
selves(facticity, reality)" (Al43/Bl82), the intensive magnitudes of the appearances
corresponding to these sensations inherit their relation to that reality. So even if
these magnitudes belong only to the constitution of objects as appearances, not
as they are in themselves, they nevertheless lead straight back to the subjectively
unconditioned, transcendentally objective reality of the in itself (ch 8). Second,
although the intensive magnitude of appearances can be apprehended in intuition
only via sensation, not pure space or time, it is as objectively independent of sen
sation as their extensive magnitude, substance, and causal connections are. For
example, insofar as the schema of existence(considered below) extends actuality to
unperceived appearances if they are connected by any of the schemata of the cat
egories of relation to perceived appearances, the schemata of reality, negation, and
limitation will apply no less to the former than to the latter. This means that inten
sive magnitudes qualify appearances with the same indifference to whether they are
present to consciousness in sensation as is true of their extensive magnitudes, their
substance, and their causal connections. Intensive magnitudes are thus objective
in the full empirical sense of being completely independent of the standpoint of
the observer, the structure of the observer's senses, and everything else that distin
guishes one observer's perceptions from those of another (A45/B62-3 and A393).13
13 This is one of the reasons I made such a point in Chapter 3-B of the need to distin
guish the matter of the appearances apprehended in intuition (perception) from the sensations
through which they are given: if the former were not completely distinct, both existentially
and qualitatively, from the sensations to which they correspond, it would have been impossi
ble for Kant to apply the schema of existence to intensive magnitudes, with the consequence
that these magnitudes would fall short of the objectivity of extensive magnitudes, substances,
and causal connections. Indeed, since the reality of substances is the same reality accorded to
intensive magnitude under the schemata of the categories of quality, failure to distinguish the
matter of appearances from sensations would extend the transience of the latter to the for
mer, and so nullify the application of the schema of substance (permanence) to appearances.
470 Kant's Anatomy of the Intelligent Mind
The schemata confer the form of objectively real times (and spaces) on the empiri
cal realities that constitute the matter of apprehended appearances and thereby give
existence in the field of appearance to the objective unity conferred by transcenden
tal synthesis speciosa on the manifold apprehended in formal intuition. Because of
this, it is entirely to their schemata that the categories owe their objective meaning
as conditions of the possibility of experience and its objects, their laws, and nature
itself as their systematic totality. Since transcendental schematism is impossible
apart from either the capacity of pure imagination to comprehend infinite spaces
and times within its a priori synthesis of apprehension or the precategorial ana
lytic unity of apperception that makes categorially determinable concepts of all
these spaces and times possible (ch 15), the categories straightaway lose all objective
meaning the moment their scope is supposed to exceed the scope of those faculties.
And since they therefore have meaning only in respect to objects (appearances)
produced by a priori imagination and thought, their scope is limited to transcen
dentally ideal phenomena (ch 7)-excepting only that in these objects which cor
responds (via sensation) to "the transcendental matter of all objects as things in
themselves (facticity, reality)" (Al43/B182).
It is productive imagination's schematism of the categories, not pure under
standing's categorial synthesis intellectualis, that furnishes the predicates for syn
thetic a priori principles of pure understanding: "what must be kept in mind with
all synthetic principles ...[is] that appearances must be subsumed not under the cat
egories unqualifiedly but only under their schemata" (A181/B223; also A136/Bl75
and Al48/B187-8).14 What are the subjects of which the schemata are predicated
in transcendental judgments? The answer cannot be appearances, for appearances
are objects intuited through sensation, not subjects of judgments.The only subjects
principles of pure understanding can have are concepts of appearances. How then
could Kant claim that these judgments apply objectively to the appearances them
selves and not merely to our thought of them (A158/B197)? In other words, what
makes them not merely normative laws of our thought of nature but genuine laws
of nature itself, indeed the most fundamental of all, presupposed by every other
(A159/B198) and constitutive of nature as such (A216/B263)?
The answer, Kant leaves no doubt, is the doctrine of original apperception estab
lished in the Transcendental Deduction of the Categories:
[T]he supreme principle of all synthetic judgments is that every object stands
under the necessary conditions of the synthetic unity of the manifold of intu
ition in a possible experience.
14 Elsewhere, the relation between the schematism and the principles is formulated slightly
differently: "The schematism indicates the conditions under which an appearance is determined
in respect of the logical function and thus stands under a category; the transcendental principles
indicate the categories under which the schemata of sensibility stand." (AA 18 § 5933 [1783-4]).
See also letter to Tieftrunk, December 11, 1797.
Subsuming Reality 473
In this way, synthetic a priori judgments are possible if the formal condi
tions of a priori intuition, synthesis of imagination, and the necessary unity
of that synthesis in a transcendental apperception are related to a possible
experiential cognition in general. We may then say that the conditions of the
possibility of experience in general are at the same time conditions of the pos
sibility of the objects of experience, and for this reason have objective validity
in a synthetic a priori judgment. (Al58/Bl97)
For an object to be given, if this is not again to mean only mediately, but
to exhibit it immediately in intuition, is nothing else than for its representa
tion to relate to experience (be it actual or still possible). Even space and
time ... [would] be without objective validity and without meaning and
signification (Sinn und Bedeutung) if their necessary employment were not
manifested in objects of experience. Indeed, their representation is a mere
schema which is always related to the reproductive imagination that calls
up (herbeiruft) the objects of experience, without which they would have
no meaning; and so it is for all concepts without distinction.... Thus, since
experience, as empirical synthesis, is, in its possibility, the one kind of cogni
tion that gives reality to all other synthesis, the latter too, as a priori cogni
tion, has truth (agreement with the object) only by its containing nothing
further than what is necessary for the synthetic unity of experience in gen
eral. (Al55-7/Bl94-7)
15 Also: "Philosophy relates only to concepts of being in general, and so to what corresponds
to sensation, and thus cannot make its concepts intuitive" (AA 18 § 5277 [early 1770s]).
{ 17 }
[A]lthough proof of a principle that underlies all cognition of its object can
not be pursued any further objectively, this still does not preclude the pos
sibility of supplying a proof from the subjective sources of the possibility of
a cognition of an object in general, indeed it would also be necessary to do
so, since the proposition otherwise would lie under the greatest suspicion of
being a merely fraudulent claim. (Al48-9/B188; also A216-7/263--4)
Kant drew his proofs of the principles of pure understanding from the doctrines
of the subjective transcendental deduction of the categories, which comprises vir
tually the entirety of Sections 2 and 3 of the Transcendental Deduction of the
Categories in the A edition and Section 2 (§§ 15-27) of B (ch 12). What this deduc
tion demonstrates is the extent to which it is subjectively necessary to proceed from
478 Kant's Anatomy of the Intelligent Mind
precisely the a priori consciousness of the objective reality confronting the senses
that transcendental synthesis produces (chs 13-15). (1) Without it, not only would
we be unable to cognize objects through appearances, their want of affinity would
make it impossible to experience recurrent sequences and co-occurrences of per
ceptions (regularities) at all. (2) Without experience of regularities, no customary
associations based on them could be formed between distinct perceptions. (3) And
without these experience-based associations of perceptions, our perceptions would
be nothing but "a blind play of representations, less than a dream" (Al 12), hence
The principles of pure understanding are transcendental judgments that affirm the
perfect agreement between empirical reality and the schematized categories, and
so are constitutive of the transcendental truth that precedes and makes possible
empirical truth (Al46/B185 and A221-2/B269). They include principles of the intu
ition of appearances-necessary relations of both their form (Axioms of Intuition)
and their matter (Anticipations of Perception)-and principles of their existence
different relations of dependence both on one another (Analogies of Experience)
and on different faculties of the mind (Postulates of Empirical Thought). Since
the latter are essentially relational whereas the former concern what belongs essen
tially to each appearance individually (form and matter), the principles of intuition
and perception come first. The principles of existence concern only the position of
appearances in the overall unity of nature (experience) and the relation of that unity
to the faculties of the mind, and so come second. The first group Kant designated
mathematical principles, the second dynamical principles, not because of the differ
ence in their content but because of the special principles of, respectively, mathe
matics and dynamics made possible by the subsumption of appearances under their
predicates in transcendental judgment (A162/B202). Nevertheless, their difference
in content does have an important consequence. While both sorts of principle of
pure understanding are necessary objectively and completely a priori (A160/B199-
200), and are alike "completely certain" (A162/B201; also A180/B223), mathemati
cal principles are "unconditionally necessary, i.e. apodeictic" (A160/B199), and
"capable of intuitive (intuitiven) certainty" (A162/B201), whereas the dynamical are
"in themselves merely accidental" (A199/B299), necessary "only under the condi
tion of empirical thought in an experience, hence only mediately and indirectly, and
so without containing the immediate evidence of the former" (A160/B199-200),
and therefore have "merely discursive certainty" (A162/B201).
Kant titled the principles so as to give prominence to these differences in evi
dence, as well as to highlight what they apply to in appearances (A161/B200). As
given in empirical intuition, appearances have a form, a matter, and an existence.
Since both the form and the matter are cognizable mathematically in pure intuition,
the principle of the form is termed Axioms of Intuition because it both grounds
and has the same certainty (intuitive) as the axioms and other mathematical prin
ciples and theorems whose possibility it grounds, while the principle of the mat
ter of appearances is termed Anticipations of Perception because it is concerned
exclusively with the only non-intuitive (non-spatial, non-temporal) quality given
a posteriori that is capable of being constructed a priori and cognized with intui
tive certainty: intensive magnitude (A161-2/B201; ch 16-C-5). Since existence, by
contrast, is not constructible in intuition, the most that can be cognized a priori in
a transcendental judgment are the relations in which any existence given in empiri
cal intuition must stand, never the existences themselves. Consequently, even if the
480 Kant's Anatomy of the Intelligent Mind
existence of something could somehow be known a priori, this, together with the
a priori transcendental cognition that it must stand in such-or-such a relation to
something else, cannot yield a priori cognition of what that existence is to which it
is related. In Kantian nomenclature, this is just to say that the dynamical principles
are merely regulative in relation to what can be given in empirical intuition, whereas
the mathematical principles are constitutive since, in their case, it is always possible
to construct an empirically unknown object in a priori intuition given the a priori
cognizable relations and constructions of everything else standing in that relation
the other in the space adjacent to it or at that moment of time. So far as the exis
tence of either is concerned, there might as well be a square, pentagon, a series of
tangent circles, or anything else existing in the adjacent space at that time. Since
extensive magnitudes in empirical intuition are therefore always "in themselves
merely accidental" (A160/B199), if there is a necessity to these triangles existing
there and then, it can be intuited neither in the form nor the matter of the appear
ances themselves.
Similarly, it is apodeictic and intuitively certain under the Anticipations of
Perception principle that the difference in intensive magnitude between lunar and
482 Kant's Anatomy of the Intelligent Mind
solar illumination given in empirical intuition can be constructed a priori and that a
multitude of the one can coalesce to constitute the other (Al 79/B221). Yet, none of
this multitude of empirical realities in the moon or the sun necessitates the existence
of any of the others. As far as their existence is concerned, these realities "are in
themselves merely accidental" (A160/B199). So, again, if there is a necessity to their
relation, it cannot be met with in the matter of appearances.
Since appearances are constituted solely by a form and a matter, there is noth
ing else intuitable in them a priori whereby to connect their existence, and so no
occasion for additional constitutive, intuitively certain, unconditionally necessary
transcendental principles of mathematical composition. Nevertheless, there is still
one respect in which the existence of appearances can and must be subjected to
transcendental principles: their situation in relation to the whole (totality) of exis
tence in space and time. It is only insofar as appearances have existence there that
they have an objective existence empirically distinguishable from their existence in
our perception (in empirical intuition through sensation). This is the empirical dis
tinction presupposed in all cognitive experience between "what depends essentially
on the intuition of appearances and is valid for all human sense in general" and
"what belongs to appearances only accidentally because valid only in relation to a
particular position (Ste/lung) or organization of this or that sense, not in relation to
sensibility in general" (A45/B62). The former is the objective existence we accord to
the appearance itself, the latter its accidental relation via sensation to our cognition;
and however the details of their objective existence are filled in over the course of
experience, our representation of that whole of existence in space and time, with
a uniquely differentiated, completely determinate place therein for every reality
apprehensible in empirical-material intuition, must already be in place completely a
priori, ahead of all actual experience.
This transcendental cognition, however, involves more than simply the relation
of appearances to "what depends essentially on the intuition of appearances and
is valid for all human sense in general." The pure intuitions of space and time on
which all appearances "essentially depend" are, as we have seen, completely undif
ferentiated and undetermined (ch 4). For this reason, a transcendental synthesis
speciosa of the manifold of pure space and time is also necessary for transcen
dental cognition (ch 15), as well as a schematism to extend this imperceptible
spatially-temporally differentiated and determined manifold to the manifold per
ceptible realities apprehensible in empirical-material intuition (ch 16-B). It is thus
under the predicates of transcendental schematism alone that we can cognize
appearances a priori in relation to the whole of existence in space and time. And it
is the schematized categories of relation and modality in particular that alone are
capable of constituting a genuinely objective whole of existence in space and time
that is neither dependent on nor limited by the accident of its relation to empirical
cognition via sensation (ch 16-C and -D).
I use the word "constituting" advisedly. For even if transcendental cognition of
appearances in relation to the totality of existence in space and time corresponds
Time Out of Mind 483
In the Analytic of Principles, Kant maintained that "since the existence of appear
ances cannot be constructed," principles of existence "can only concern relation
of existence, and can yield none but merely regulative principles" (Al 79/B222).
Relations of existence in space and time are all that remain when everything con
structible in intuition is excluded: the relation of the changeable in time to the per
manent, of what begins to what preceded it in time, and of the coexistent. The
dynamical principles are constitutive of these relations among appearances but not
of the appearances in them. Nevertheless, in regulating appearances, they constitute
an experience in which every appearance has its own objective existence in space
and time, distinct from the accident of its relation to cognition via sensation. Their
existence thus becomes objective in the fullest sense: independent of standpoint, of
everything that distinguishes humans or any other species of sensibly conditioned
understanding from any other, and of every other epistemological limit (the math
ematics yet to be developed, experiments yet to be conceived or made performable,
conceptual apparatus the need for which is as yet unrecognized, etc.).
The constitutive role played by Kant's dynamical principles bears a striking resem
blance, almost certainly intentional, to the role played in Hume's system by "cus
tom, or if you will, by the relation of cause or effect," as the "principle, which
peoples the world, and brings us acquainted with such existences, as by their
removal in time and place, lie beyond the reach of the senses and memory" (THN
108/75; also 74/ 53, EHU VIII/i if 5, UU ch 18-D). There is another resemblance
too. Just as for Hume the whole of existence in space and time is merely "the uni
verse ...I paint ...in my imagination, ...nothing but ideas [that] by their force and
settled order, arising from custom and the relation of cause and effect, ...distin
guish themselves from the other ideas, which are merely the offspring of the imagi
nation," so too for Kant "the order and regularity in appearances that we call nature
we ourselves introduce (hineinbringen)" through "subjective grounds" such as the
categories (A125). Everything-space, time, the order of existence in them, and the
484 Kant's Anatomy of the Intelligent Mind
For it is it alone that makes pure mathematics, in all its precision, applica
ble to objects of experience ...Appearances are not things in themselves.1
Empirical intuition is possible only through pure intuition (of space and
time). So, what geometry asserts of the latter also holds incontrovertibly of
the former, and such subterfuges (Ausfiuchte) as that objects of the senses
may not be in conformity with the rules of construction in space (e.g. of the
infinite divisibility of lines or angles) fall away ...The synthesis of spaces and
times, as the essential form of all intuition, is what at the same time makes
possible the apprehension of appearance, hence every outer experience, and
consequently all cognition of objects of such experience. (A165-6/B206-7).
Clearly, the claim is not that Hume and other deniers of infinite divisibility are
wrong about the presence of a minimum of perceptibility in the visual, tactual, and
inner-sense fields. Kant's affirmation of infinite divisibility relates not to sensations
but to the appearances intuited through them (ch 3-B). Since neither the form (ch
3-C) nor the matter (ch 3-B) of appearances has any features in common with sen
sations, their intuition is not subject to the limits of sight, touch, or any other sense.
Consequently, when transcendental judgment, in the Axioms of Intuition princi
ple, provides intuitive certainty that the manifold of appearances apprehended in
1 Kant took Hume to have affirmed things in themselves: CPrR 52-3. See chs 2-E and 7-A.
Time Out of Mind 485
simply and solely in facile, idea-enlivening transitions from presently sensed impres
sions to their customary associates in thought, associations formed by the experienced
constant conjunction of the perceptions concerned, whether in immediate sequence,
as in relations of cause and effect, or as co-occurrent, as in the simplicity at a time
and identity over time of complex individuals (UU chs 16 and 17, HTC Parts II and
111). Cause and effect relations, however, hold a unique place in Hume's system: as the
only relations that enable the mind to "go beyond what is immediately present to the
senses, either to discover the real existence or the relations of objects" (THN 73/52-3;
also 195-8/130-2, EHU VIII/i if 5, analyzed in UU ch 18-D and HTC ch 4), no other
relation, including the other associative relations of contiguity and resemblance, can
be extended beyond the senses unless founded on some causal relation.
Hume singled out three cases as especially important. (1) Whenever objects sep
arated in space and together in time exhibit a fixed, unalterable2 spatial or temporal
relation, we suppose their relation to be founded on a relation of existence involving
"some secret cause, which separates or unites them" (THN 74/53). (2) This includes
the fixed, unalterable relation of the parts that make up any whole, such as a hand,
a stone, a tree, a mountain range, or the solar system, for a cause of the constant
concomitance in space of distinct spatially contiguous elements and qualities is just
as essential to ideas of bodies as spatial contiguity itself (16/16 and 237/155). From
such causes, I infer the hidden interior and sides of the house from seeing its front;
I infer the presence of fig flavor in the fig I see on the table; and I infer that the per
son sees me who looked in my direction when I called to her. Finally, and perhaps
most important, (3) it includes the fixed, unalterable relations of identity over time
whereby an object is conceived to continue existing after ceasing to be perceived:
We readily suppose an object may continue individually the same, tho' sev
eral times absent from and present to the senses; and ascribe to it an identity,
2 Here and in what follows, I have opted to emphasize part of Hume's formulation on THN
110/76: "The objects [the relation of cause and effect] presents are fixt and unalterable.... [E]
ach impression draws along with it a precise idea, which takes its place in the imagination, as
something solid and real, certain and invariable." Although not said specifically in reference to
the capacity of this relation alone to extend the scope of what we regard as real beyond the senses
and memory, it is the sequel to Hume's assertion a few paragraphs previously that the "force and
settled order " of ideas standing in causal relations underlies the ability to "paint the universe
in my imagination, and fix my attention on any part of it I please." On 74/53, Hume explains
"invariable " relations between distinct perceptions in term of causal connections.
486 Kant's Anatomy of the Intelligent Mind
Insofar as it is natural to conceive the existence of sensations as (i) the effect of (ii)
the object that appears through them on (iii) the sensibility of the mind that is con
scious of the object through them, sensations of that object are equally naturally
conceived to depend on its presence to the senses. Since its causation of sensations
is the only feature we have whereby to conceive the object's existence apart from
the sensations themselves, it is necessarily its efficacy with respect to sensations that
must be represented in order to conceive it as having an existence distinct from the
sensations of itself it causes in one's mind and so as capable of continuing after it
has ceased to cause them. In other words, by conceiving our senses as sometimes
present to the object's causal influence and sometimes absent from it (so "that if we
had kept our eye or hand constantly upon it, it wou'd have convey'd an invariable
and uninterrupted perception"), we can use the existence implied by its efficacy
as a cause of sensations to conceive its continuance in existence even when none
are actually being caused by it. 3 And while this way of conceiving the existence of
objects suffices even to represent their existence as permanent (backward as well
as forward in time), we normally find empirical reason to ascribe other causations
(besides its causation of sensations) to its existence, including its connection as
effect to causes capable of beginning or ending it.4
To explain how the mind goes from (1)- (3) to an idea of an unbounded whole
of existence in space and time, Hume supposed that it posits causal series extending
to causes situated in the remotest regions of space and time. Since knowledge of
causes is limited to our own and other's past experience, Hume also supposed that
3 On the supposition that the sensations it causes do not alter in the interim, the object is
thought as constant; and by conceiving its causation of sensation to change in accordance with
well-known causes, we conceive its coherence (e.g., when we return five years later, we expect to
find a child, not the infant we knew to begin with): THN 194-9/130-2. This is important because
the basis of constancy in causal association is less obvious than in the case of coherence. It is
thus not the qualitative but the causal character of visual and tactual sensations-their having a
specifically spatial and temporal constancy and coherence lacking in all other (non-spatial) sen
sations-that explains why it is the objects of these senses that are the prime beneficiaries of the
associative fiction of distinct, continued existence (207-8/138; see also HTC ch 7).
4 Hence the sequel to the passage cited earlier: "Whenever we discover such a perfect resem
blance [between present and past objects whose appearance to the senses has been interrupted],
we consider, whether it be common in that species of objects; whether possibly or probably any
cause cou'd operate in producing the change and resemblance; and according as we determine
concerning these causes and effects, we form our judgment concerning the identity of the object"
(THN 74/53).
Time Out of Mind 487
the mind is naturally able to equip itself with two general causal principles: the uni
formity of nature and the general causal maxim. According to the first, causes, from
the remotest past to the most distant future, always necessitate the same effects, and
effects are always necessitated by the same causes. According to the second, noth
ing begins to exist without a cause: no existence, no action, no motion, no quality,
in short, no matter of fact (THN 12/13, 94/65). Both principles are products of
experienced constant conjunction and customary association, and so derive from
human nature (UU ch 19-D, HTC ch 5). This means that "long before I was out of
my cradle" (EHU IV/ii if 23), they had come to inform all my reasoning in matters
of fact. By reasoning in accordance with the uniformity principle, we determine
that causes sufficiently like those we have previously encountered in experience will
operate the same everywhere and everywhen, even in places and times no human
has experience of; and if they fail of their usual effect or usual cause, it is because of
some hidden factor (this and other "natural principles of our understanding ...by
which to judge of causes" described in THN III/iii/§15 enable us "to distinguish
the accidental circumstances from the efficacious causes," 149/101). By reasoning
in accordance with the general causal maxim, we extend the scope of the causal
nexus to include the as yet unknown causes and effects of the objects (existents)
that experience has so far disclosed, the causes and effects of these unknown causes
and effects, their causes and effects, and so on. Moreover, by extending causal series
in this way, the maxim also makes it possible to extend the three noncausal rela
tions that have to be underpinned by causal relations: fixed unalterable relations
in space and time, fixed unalterable relations of contiguity between the parts of
bodies, and fixed unalterable relations of simplicity at and identity over time. In
this way, Hume's two fundamental causal principles enable causal inference, both
in its own right and as underpinning other kinds of reasoning, to yield a genuine
concept of the whole of existence in space and time, one vast and diverse enough
to accommodate scientific and ordinary consciousness alike. And thus do associa
tive relations become "the only links that bind the parts of the universe together, or
connect us with any person or object exterior to ourselves," being "really to us the
cement of the universe" (662/417).
Of all the problems with Hume's associationist account of the whole of existence in
space and time, none is more serious from a Kantian standpoint than the quandary
concerning personal identity described in an appendix to the second volume of the
Treatise (1740) (UU ch 3):
While personal identity may seem a different question from that of the whole of
existence in space and time, in Hume's system, both are at their core the same: iden
tity relations constituted by uninterrupted sequences of causal and other associative
relations connecting a succession of distinct existents. Since Hume's psychologis
tic explications of identity and cause and effect show that both alike are nothing
more than associative relations connecting ideas of these successive existents in
thought-"uniting principles in the ideal world" whose "very essence ...consists
in their producing an easy transition of ideas" (THN 260/169; also 99/69, 204/135,
220/145}-and that they are incapable of holding, or even being conceived to hold,
of the successive existents themselves, prior to or independently of the association
of their ideas in imagination, it follows that these existents neither have nor can be
conceived to have any real causal or identity relation connecting them, either to
one another or to anything else (222/146-7, 233-4/153--4, 244/160, and 252/164-5).
Thus, the same quandary that confronted Hume in connection with the causal
associative system of impressions and ideas that yields "the true idea of the human
mind" (261/170) also arises in connection with the causal-associative system that
yields the idea of the whole of existence in space and time in the manner described
in Section B.
What Hume realized by the time he penned this appendix was that he lacked
the means to explain the fixed, unalterable relation our successive perceptions all
have to "our thought or consciousness." Normally in his system, such a relation
would imply the existence of a causal relation (necessary connection between dis
tinct existents) underpinning it. Causal association, however, cannot explain it for
two reasons. First, the customs essential to such association can be formed only if
one has experienced, at many different times, similar perceptions in similar relations
of immediate succession, and this is possible to do only if the perceptions experi
enced at these scattered times are all already united in one and the same "thought
or consciousness." Second, even if the experiential conditions (the frequency and
constancy of resembling sequences of perceptions) for customary association are
set aside and we limit our consideration of causal relations to their associationist
Time Out of Mind 489
essence as a facility felt in the transition from an impression to an idea (UU chs 16
and 17), the successive perceptions from and to which such transitions are made
must already be united in our thought or consciousness if the transition from one to
the other is to be perceived at all, much less feel easy , difficult, or indifferent (UU ch
3). So no matter how Humean causal association is viewed, it presupposes, and so
cannot explain, the fixed unalterable relation of our successive perceptions to our
thought or consciousness. And since both the world of external objects to which a
continued, distinct existence relative to the mind is attributed and the mind itself
are, for Hume, nothing but fictions of associative imagination that take their start,
as imperfect identities, from one and the same succession of perceptions united
in consciousness, his accounts of the internal world of the mind and the external
world of space and time are both equally put in question of his inability to explain
this unity.5
Hume could conceive of no nonassociative principles capable of explaining this
relation except principles of real connection between distinct existents of precisely
the kind he foreswore in his original account of personal identity ("something that
really binds our successive perceptions together," so that "in pronouncing concern
ing the identity of a person, we observe some real bond among his perceptions,"
THN 259/169). Yet Hume ascribed so much certainty to the two principles that
underlay his original account-"that all our distinct perceptions are distinct exis
tences, and that the mind never perceives any real connexion among distinct exis
tences" (636/400)-that even after discovering it to be fatally flawed, he was unable
to renounce either of them and so had no option but to "confess, that this difficulty
is too hard for my understanding." The first principle prevented him from deny
ing the distinctness of successive perceptions by supposing them all to inhere in a
single (at a time), identical (over time) substrate consciousness. The second avoids
this hurdle because it is compatible with the uniting consciousness being no less
manifold (nonsimple) and fleeting (nonidentical) in its existence than the successive
perceptions united in it. But since such a consciousness can only be supposed to
unite the perceptions that preceded it if it is perceived to be their pre-associatively
real effect (along lines similar to Kant's thought experiment at A363), the second
principle precludes this way of explaining their unity in that consciousness as well.
5 In THN I/iii, before addressing questions of identity, Hume treated the internal mental
world and the external corporeal world as consisting of exactly the same perceptions, connected
by memory in the case of the former and by causal association in the case of the latter (I 08/75).
In THN I/iv, however, it emerges that the internal world of the mind, memory included, is also
formed by causal association (the "true idea of the human mind," 261/170; also 265/173), so
that what distinguishes external objects (bodies) from the mind is the fiction of continued, dis
tinct existence attributed to its most constant sensations (visual and tactual) as well as to those
most closely associated with them (olfactory, auditory, and gustatory sensations) (207-8/138 and
237/155-6). Thus, mind and bodies emerge as fictitious imperfect identities conjured up by asso
ciative imagination from the very same succession of perceptions pre-associatively united in our
thought or consciousness (UU ch 3 and Part III).
490 Kant's Anatomy of the Intelligent Mind
Where did this final reaffirmation of his rejection of all principles of real con
nection leave Hume? With the successive perceptions "thus loosen'd," he presum
ably returned to the starting point not just of his account of personal identity but
of all of his associationist explications, including of identities and necessary con
nections generally. That might seem to exempt the causal relations Hume relied on
in these explications: ideas copied from, and so causally dependent on, antecedent
impressions; impressions of reflexion occasioned by, and so causally dependent on,
antecedent ideas of sensation; secondary ideas dependent on, because copied from,
primary; ideas conceived and believed because of the appearance of their associates
in an impression or believed idea; and so on.However, in utilizing the causal system
formed by these relations to explicate "the true idea of the human mind" (THN
261/170), Hume made clear even in the first volume of the Treatise that these causal
relations too are to be understood not as real bonds among perceptions but only
as associations of "their ideas in the imagination," which we "only feel," and so as
products of "uniting principles in the ideal world" whose "very essence" consists in
"an easy transition of ideas" (259-60/169; UU chs 16 and 17, HTC ch 6-C). Since
this is just to say that not only understanding (reasoning in matters of fact) but the
senses and memory as well are "all of them ...founded on the imagination, or the
vivacity of our ideas" (265/173), the given from which association takes its start
can consist of nothing except the successive perceptions themselves, devoid of all
simplicity at a time, identity over time, and necessary connection of existence either
to one another or to anything else:
All these are different, and distinguishable, and separable form each other,
and may be separately consider'd, and may exist separately, and have no need
of any thing to support their existence.... I never can catch myself at any time
without a perception, and never can observe any thing but the perception ....
The mind is a kind of theatre, where several perceptions successively make
their appearance; pass, re-pass, glide away, and mingle in an infinite variety of
postures and situations. There is properly no simplicity in it at one time, nor
identity in different; whatever natural propension we may have to image that
simplicity and identity. The comparison of the theatre must not mislead us.
They are the successive perceptions only, that constitute the mind; nor have
we the most distant notion of the place, where these scenes are represented,
or the materials, of which it is compos'd. (252-3/164-5)
by the perceptions of the next. In other words, it is not merely succession but change
that Hume conceived in his characterization of the pre-associative given. For even if
we lack "the most distant notion of the place, where these scenes are represented," it
is still impossible to conceive the succession he described without conceiving it as
changes taking place in some persistent. A nd how can change be conceived without
also conceiving something distinct from the successive perceptions that is altered by
their change without itself changing, something that is the same before and after
every change of perceptions and so persists identically through the duration of
their succession?
It was Kant who remarked that "[w]e can only notice change in that which per
sists; if everything is in flux, then the flux itself cannot be perceived" (A A 18 §
5871 [early1780s, possibly late1770s]).6 To conceive the perceptual succession in
the theater of the mind as a flux of perceptions that succeed one another, and so
as a changing scene, Hume would have to have had a means of according persis
tence (continued existence, identity over time) directly to the consciousness that
is spectator to this scene. Being unable to explain how an idea of something that
persists in existence is possible given that nothing is ever present to us but "differ
ent perceptions, which succeed each other with an inconceivable rapidity, and are
in a perpetual flux and movement" (THN 2521
/ 65), Hume should also have recog
nized that the absence of a persistent precludes his portrait of perceptions in the
theater analogy as succeeding one another and so forming in themselves a changing
scene ("all change in time can be regarded only as a mode of the existence of what
remains and persists," A183B
/ 227). For in the absence of persistence and change,
they can only be supposed to exist as a succession in the barest sense conceivable: a
flux of the fleetingly existent perceptions, each an absolute beginning as well as an
absolute end with respect to the rest.
From a Kantian point of view, it is this that most clearly exposes the limits,
and ultimate inadequacy, of empirical psychologism. For in order to conceive dis
tinct perceptions as existing in the same mind (unity of consciousness), or even
to conceive them as existing in any mind at all (which the separability principles
does not require: THN 207/ 138), they must be conceived as existing in succession
to one another and as forming in themselves a changing scene. Once it is recog
nized that empirical psychologism is incapable of explicating the latter, the correct
conclusion, from Kant's perspective, is that it is equally incapable of explicating
the former and that the quandary that ultimately stymied Hume can be solved
only under the aesthetic and intellectual principles of transcendental psychology
(UU ch 3).
6 Here and throughout it is important not to equate the flux of successive appearances in
inner intuition with the smooth flow of visual, auditory, or other sensations. Even if they were
not experienced as a flow but like a film run below proper speed (a series of "stills"), they would
still be apprehended in inner intuition in the form of a succession.
492 Kant's Anatomy of the Intelligent Mind
and -D). Yet with no perceptibly existent persisting time in the field of appearance
to contain all nonpersisting appearances, how can the latter be supposed to exist
in succession to one another, as changing scenes of consciousness, or in objective
temporal relations of any kind? The imperceptibility of pure time means that sen
sibility alone cannot satisfy the demand for something that persists in existence
right in the field of appearances itself (ch 16-B). This requirement must there
fore be met by the understanding if even so much as a continuant as otherwise
completely indeterminate as the spectator in Hume's theatre of the mind is to be
representable.
But how? On the evidence of passages like the following, it might be supposed
that the existent that persists in all consciousness is the I:
[T]ime, the sole form of inner intuition, has nothing abiding, and so gives
cognition only of change of determinations but not of a determinable object.
For in that which we call the psyche, everything is in continuous flux and
nothing abiding, outside perhaps (if one insists) of the simple I since this
representation contains no content, and so no manifold. (A381)
The reason for the qualifications "perhaps" and "if one insists" is that the I in ques
tion is the I think, not the I that intuits itself as an appearance (continuant) in inner
empirical intuition. The significance of this is made clear elsewhere:
That the selfsame I think may accompany all of my representations tells me noth
ing about the temporal appearance I will make in inner intuition (the I that intuits
itself). It makes no difference to my existence as an I think whether my existence
as an appearance in inner intuition is that of a continuant that persists through all
changes in my perceptions or that of a series of fleetingly existent consciousnesses,
each of which transmits the representations of its temporal moment, in the form
of new memories (together with the memories transmitted to it by previous equally
fleetingly existent l's), to succeeding, no less fleeting l's. Both temporal scenarios are
equally compatible with the logical unity of the I "which expresses the conscious
ness that can accompany all thinking" and "immediately includes the existence of
494 Kant's Anatomy of the Intelligent Mind
the self, but still no cognition thereof, and so no empirical cognition, i.e. experience"
(B276).7 Being an I that can, at any moment in time, think its identity in relation to
all the manifold (analytic unity of apperception: ch 9-B) therefore has nothing to do
with my nature as an existent in time, apprehended empirically in inner intuition via
self-affections (in the same way bodies are apprehended in outer intuition via sensa
tions), and so cannot be what determines the flux of appearances in inner intuitions
as changes taking place in an enduring existent.
Indeed, any uncertainty regarding the temporal character of the I arises insofar
as it is considered strictly a priori vanishes when one takes into account what actu
ally is apprehended in inner empirical-material intuition. Like Hume before him,
Kant held that since "time, and so all that is in inner sense, is in constant flux"
(B291), "there can be no standing or abiding self in this flux of inner appearances"
(A107). No representation, no consciousness, exists in inner empirical intuition
with even the slightest duration, including the imperceptible pure time intuition that
"as substratum, as permanent form of inner intuition" (B224), precedes and makes
possible all simultaneity and succession of appearances. Even the empirically real
substances whose permanent existence is affirmed in the Analogies of Experience
to be given with our every perception (inner included: B275) are in no sense per
manently intuited appearances, for "the representation of something permanent in
existence is not the same as a permanent representation" (Bxli n). In all such cases,
one must be careful to differentiate what it is for a representation to be a transcen
dental constant that must either be incorporated into appearances a priori (pure
time into their form, permanent substances into their matter) or accompany them
a priori (the I think as analytic unity of apperception) from what it is for such
a representation to be perceived, and so exist, as an appearance apprehended in
inner empirical intuition. In the latter regard, transcendental representations are no
different from everything else in inner intuition: part of a flux wherein "existence
is always only vanishing and recommencing, and never has the least magnitude"
(A183/B226). For want of anything that persists even from one instant to the next,
the existence of successive appearances in inner intuition neither changes anything
nor leaves anything the same so that each is in itself both absolute beginning and
absolute end-succession without change, a perfect Heraclitean flux. Thus, Kant's
account of the existence of appearances in temporal relations takes its start at pre
cisely the point where Hume's account collapsed: the search for a principle of real
existence (rather than a mere formal intuition) that enables the fleetingly existent
appearances in empirical inner intuition to be represented as changes in something
that, by persisting in existence, can "unite our successive perceptions in our thought
or consciousness" (THN 636/400).
7 The existence represented in the I think is "something real that is given, given indeed only to
thought in general, and so neither as appearance nor as a thing (Sache) in itself (noumenon), but
as something that in fact exists and in the proposition I think is designated as such an existent"
(B422-3n). See chs 4-C and 8-F.
Time Out of Mind 495
D. Permanent Substances
(B224). This principle converts the flux of appearances into successive changes in
the permanent, "transition[s] from this state into the other and from nonbeing into
being" (A188/B231). One appearance now takes the place of another to become the
substance's present determination (state) until replaced by another. The substance
is not this or that reality (quality) present in perception via sensation but what
supports the existence of these realities and continues existing before, during, and
after each changeable reality is displaced by a successor reality. Its unchangeability
therefore concerns only its existence, never its qualities, all of which, as empirically
determinable (A204-6/B250-l), are changeable.
Affections correspond to the "transcendental matter of all objects as things in
themselves (facticity, reality)" (Al43/Bl82), the matter of appearances exhibits
affections, and appearances are apprehended in the immediacy of intuition as a
succession of fleeting existents. Under the First Analogy, each such appearance
constitutes a change in how the permanent exists; and since the permanent, as
that with "an inner necessity of persisting" (Al85/B229), cannot be identified with
anything changeable, none of the changeable qualities met with in the intuition of
either the matter or the form of these appearances-"the way[s] the existence of
substance is positively determined" (A187/B230)-can be in any way essential to
the permanence of the substance. Even the existence of perceptions themselves,
the consciousnesses of affections through which substances are apprehended in
empirical-material intuition, cannot be deemed essential to their existence since the
appearances that exhibit them in inner intuition are one and all fleeting, and so
mere changes in the permanent.
The permanent thus plays much the same role in Kant's conception of existence
in time that causation does in Hume's (Section B): it extends the scope of the real
beyond the senses and memory. It changelessly extends into the infinite past (A185/
B229), before my mind and its representations existed, and extends changelessly
into the endless future, after I am long gone. During the limitless eons stretching
before and after me, substance can support (as its determinations) outer appear
ances as well as the inner appearances produced by earlier minds and, moreover,
can do so during my existence as well as before and after it. The affirmation of
permanent substances in the First Analogy thus obliges me to conceive the entire
succession of appearances I apprehend in intuition, and so my entire existence, as
I would a single moment in relation to the whole of infinite time; and insofar as
permanence is possible only in space, their affirmation also obliges me to conceive
the space of my experience as I would a single point in relation to the whole of
infinite space.
Why a permanent? Why is the affirmation of a continuant not sufficient for
change to be apprehended in inner intuition, as some suppose? And even if per
manence alone suffices, why must it be permanently existing individuals rather than
a permanent series of overlapping continuants, or simply permanence in the guise
of universal constants and/or laws of nature? The answers to all of these questions
start from the same ground: the imperceptibility of time (cited both in the A and
Time Out of Mind 497
B edition proofs of the principle: Al83 and B225). Being perceptible is Kant's
criterion of existence in the field of appearances: what cannot be perceived can
not be accorded existence there. Time, as a pure intuition, can never be perceived
and therefore cannot be accorded existence in the field of appearances (Section C
and ch 16-B). This means that there can be no all-encompassing empirically real
time in which every appearance has its existence before, during, or after this, that,
and every other, much less an empirically real time in which they all exist deter
minately so much before, so long during, or so much after one another. To exist in
time relations, therefore, something must exist that "represents time in general ...in
the objects of perception, i.e. the appearances" (B225), something that is able to
"represent the unity of time" (Al86/229). This no mere continuant, however endur
ing or unchanging, can do. Only a genuine permanent, with "an inner necessity of
persisting" (Al85/B229), can go proxy for pure time in the field of appearances as
the unchangeable substratum of all that is changeable and so confer time's "thor
oughgoing unity" on "all change" (substance as the exponent of the unity of time
in appearance: A216/B263).8
Why does the permanent required to represent the unity of time have to
consist of one or more permanent individuals rather than an infinite series of
impermanents of imbricating finite durations? For different finite durations of
perceptible realities to overlap in this way, the impermanent substrates of change
concerned would already have to exist in one and the same time. However, if time
is imperceptible not only in the Newtonian sense of being real in itself yet never
in any affection of the senses (our own or any other possible sense) but also in
the Kantian sense of being an ideal representation that exists, and can exist, only
in and through the capacity of our sensibility for pure formal intuition, then it
is quite simply unintelligible to suppose impermanent perceptible realities with
distinct beginnings, durations, and/or ends to exist in the same time. In the con
text of Kant's idealism, the imperceptibility of time implies that each of these
impermanents would be a distinct unity of time unto itself, with no absolute
time distinct from them to comprehend their different time-unities in a single
all-encompassing unity of time. Since this means that appearances inhering in
different impermanent substrates of change would ipso facto fall into distinct,
incommensurable unities of time, it would make their temporal imbrication, and
so too any unity of time their imbrication might be supposed to form, quite
impossible:
Substances (in appearance) are the substrate of all time determinations. The
coming to be of some and the ceasing to be of other substances would annul
the unique condition of the empirical unity of time, and appearances would
then relate to two different times in which, one alongside the other, existence
would elapse, which is absurd. For there is only one time in which all distinct
times must be posited, not simultaneously but one after the other. (A188-9/
B231-2)
If the ultimate substrates of change that represent (go proxy for, exponentiate) time
"in appearance" were impermanent, each its own distinct unity of time alongside
the others, then, in relation to different such substrates, each appearance would
simultaneously exist in different unities of time; and since "simultaneity is not a
mode of time itself, the parts of which are never simultaneous, but all successive"
(A183/B226), existence in distinct yet non-successive times is impossible. If to avoid
the absurdity of simultaneous unities of time caused by a permanent series of
impermanents of imbricating duration, the durations of impermanent substrates
of change were instead posited as non-imbricating (each ending exactly when
another begins, none simultaneous), we would still have to think another time in
appearances spanning their infinite succession, and so an underlying permanent
substratum of all change, fully as concrete and as individual as they.9 There is there
fore no way for the permanent required for the unity of time to be synthesized by
combining the times of impermanent individual substrates of change.
Could the affirmation of permanent individuals be avoided by positing universal
constants and/or permanent laws of nature instead? The First Analogy converts
the succession of fleetingly existent appearances in inner intuition into continuous
changes in the way the permanent exists ("its mere determination, i.e. the manner
how the objects exists," A182; also A186-7/B229-30). How could this be possible
if, instead of existing just as the appearances do (perceptible existence, existence
apprehensible via sensation) and being just as individual as they are, this permanent
were a constant, a law, or something else general in nature, akin perhaps to Plato's
eternal, unchanging forms or Descartes's eternal truths? Universal validity can only
be equated with the permanent insofar as it has validity in all times. But the prob
lem posed by the imperceptibility of time is the absence from the field of appear
ance of anything but the flux of inner intuition, in which each appearance is both
an absolute beginning and absolute end. How can anything be valid at all times if
everything that exists ends at the very instant it begins, without the least connection
to what exists before or after? There must be a true unity of time in existence before
there can be validity in respect to all times; and this means that only what has exis
tence in the field of appearances itself, with the same perceptible reality imperma
nents have, can go proxy for time as the substrate of all succession and simultaneity.
Before there can be laws of change, constants through change, or any other abstract
general permanents, there must already be permanent individuals existing in the
field of appearance, underlying all relations of succession and simultaneity among
9 This is just an extension to the substance that represents time in appearance of what Kant
argued in respect to pure formal time itself: "Were one to ascribe a succession of one after the
other to time itself, then one would still have to think another time in which this succession was
possible" (A183/B226).
Time Out of Mind 499
perceptible realities ("all existence in time and all change in time can be regarded
only as a mode of the existence of what remains and persists," Al83/B227). The
First Analogy therefore well deserves its place "at the head of pure laws that obtain
completely a priori" (Al84/B228): far from being explicable in terms of validity at
all times, it affirms that all laws of nature, all constants in nature, and everything
else capable of such validity presupposes individual permanent existents as real
substrates (final subjects) of the changeable.
As the First Analogy converts the flux of inner appearances into changes (con
tinuously shifting scenes) in the permanent, the Second converts these changes into
effected changes (events, happenings, occurrences). Each appearance in every mani
fold apprehended in inner intuition is necessitated to exist by something that, by
becoming existent in the immediately preceding time, produced a sufficient con
dition for its existence. Since these are themselves appearances that have already
been determined as changes in the permanent, and the permanent extends beyond
appearances actually apprehended in intuition, the possibility always exists that the
sufficient condition for their existence includes objects that are not now and may
never have been present in intuition via sensation and that some, or even all, of their
other effects may be unperceived as well. Thus, the Second Analogy integrates inner
appearances into a wider causal nexus extending into the remotest reaches of the
past, the future, and (together with the Third Analogy) everything simultaneous
with present appearances but not itself appearing. And the concept of that nexus
is simply the schema of the category of cause and effect that extends the cause and
effect synthesis speciosa of pure-formal intuition (ch 15-C) to empirical-material
intuition (ch 16-C-2).
The causal nexus of the Second Analogy situates all existents in the unity of
time defined by the permanent, determining at what time they begin to exist, for
which duration of time they continue to exist, and at what time they cease to exist,
just as if there was an absolute time in which they all existed. They thus constitute
precisely the single linear objective time-order presupposed for experience of the
frequency and constancy of their conjunctions, and so make customary association
possible (ch 13):
Understanding belongs to experience and its possibility. To this end, the first
thing it does is not to make the representation of objects distinct but to make
the representation of an object in general possible. This happens when it
carries time-order into appearances and their existence. It does this a priori
by conferring on every one of them, as a consequence, a determinate place
in time with respect to the preceding appearances. Without this time-order,
they would not be in agreement with time itself, which determines the posi
tion of all its parts a priori. Now, this determination of position cannot
be borrowed from the relation of appearances to absolute time (for time is
not an object of perception), but, quite the contrary, the appearances must
determine one another's place in time and make this time-order necessary,
i.e. that what follows, or happens, must follow upon what the preceding state
contained state according to a universal rule, from which there arises a series
of appearances that, by means of the understanding, produces, and makes
necessary, precisely the same order and constant connection in the series
of possible perceptions as that found a priori in the form of inner intu
ition (time) wherein all perceptions have to have their position. (A199-200/
B244-5)
Time Out of Mind 501
Appearances under the Second Analogy determine one another's place in time, and
thereby make their time order necessary. Since Kant analyzed objectivity in terms
of precisely such a necessary synthetic unity (ch 13), this is just to say that the
Second Analogy converts these appearances into objects of experience. It is accord
ingly no surprise that Kant regarded the law under which appearances determine
one another's place in time as the decisive difference between his explanation of
cause and effect and its explanation by empiricists:
It does indeed seem to contradict all that has hitherto been observed regard
ing the procedure of our understanding, according to which we are first led
to discover a rule only through agreeing sequences of many events that we
have perceived and compared, a rule in conformity to which certain events
always follow certain appearances and are first occasioned by them, from
which we form a concept of cause. On such a basis, the concept would be
merely empirical, and the rule that it confers, that everything that happens
has a cause, would be quite as contingent as experience itself: its universality
10
As elucidated in chapter 15-C, "indirect" refers to causal series involving different deter
minations of the same final subject (here, substance) than the determination involved in the
causal series in question. The series here in question is that of my own inner appearances, i.e., the
appearances whose succession constitutes the duration of my empirical self, where that self, as
an impermanent continuant, is simply a further determination ("accident") of the same under
lying substantial final subject in space to which these appearances belong. Through its other
determinations, this substance connects up with all other substances in the infinite community of
permanent existents, thus indirectly connecting my empirical self and its constituent appearances
to the whole of existence in space and time (nature as phenomenon).
502 Kant's Anatomy of the Intelligent Mind
and necessity would then be only fictitious, and would have no true universal
validity because it would be grounded not a priori but only on induction. Yet,
matters here are as they are with other pure a priori representations (e.g .space
and time), clear concepts of which we are only able to extract from experience
because we put them into it, and through these concepts first brought experi
ence into being. To be sure, the logical clarity of this representation of a rule
determining the series of events as a concept of cause is possible only when
we have made use of it in experience; but having it in view, as a condition of
the synthetic unity of appearances in time, was still the ground of experience
itself and thus preceded it a priori. (A195-6/B240-l)
The empiricist supposes that the law of cause and effect originates in the many
regularities to which we are witness in experience, particularly those so common
and predictable that the understanding generalizes from them to the causal law
that whatever happens in experience is caused by something that preceded (UU
ch 19-D). Against this, Kant argued that we must already take cognizance of the
law if we are to experience regularities of any kind in the first place. It is a prin
ciple in accordance with which appearances-whether given in empirical intuition
or as causal conditions of what is given-not only fix their own positions in time
but do so irreversibly, since causal series connecting existents in time, just like the
time series itself, can proceed only in one direction, and so exhibit "precisely the
same order and constant connection ...as that found a priori in the form of inner
intuition (time) wherein all perceptions have to have their position" (A199-200/
B244-5).11 In other words, rather than time itself carrying perceptible existents ever
forward into new times, like debris swept along by the temporal current, perceptible
existents carry themselves forward as events causally necessitated by and causally
necessitating other events and so, in their series, give reality to the necessary order
of pure, imperceptible time in the sphere of perceptible existence.12 Thus, the mis
take committed by Hume and other empiricists was to suppose that the successive
existence of appearances in time, and so too experience of their frequent and con
stant conjunction, can be represented independently of cause and effect when, in
truth, given the nature of time as a pure, imperceptible intuition with no existence
in in the field of appearances, their succession to one another must originally be
constituted by a causal law of transcendental judgment as effected changes.
Kant did not pretend that we have a clear conception of the causal law of the
Second Analogy "as a condition of the synthetic unity of appearances in time"
u The order can, however, also be thought to proceed temporally reversewise when the schema
is omitted, as with purposes (CJ 372).
12 "Principium rationis is the principle of the determination of things in time-sequence; for
that cannot be determined through time, rather time must first of all be determined through the
rule of the existence of appearances. Principium of the possibility of experience." (AA§ 18 5202
[late 1770s]) It is the schema of cause and effect (ch 16-C-2), on the basis of the synthesis speciosa
of the manifold of pure formal intuition according to this category (ch 15-C), that accounts for
time being "first of all determined through the rule of the existence of appearances."
Time Out of Mind 503
13 What was said regarding the obscure representation of the causal system in infancy in ch
15-D applies here (mutatis mutandis) as well.
504 Kant's Anatomy of the Intelligent Mind
start of our experience and may extend beyond our perceptions now. The awareness
whereby alone reality can be marked off from dreams as the succession a perfect
memory would present therefore requires that all perceptions in inner intuition be
cognized a priori as belonging to a greater whole of existence that includes unper
ceived as well as perceived existents. And the only representation capable of doing
this is the schema (transcendental synthesis speciosa) of cause and effect predicated
of all perceptions in the Second Analogy by transcendental judgment.
If this is an accurate reflection of how Kant's conception of causal efficacy dif
fers from those of his empiricist forbears, then it can be viewed in relation to the
efficacy criterion of existence in substances (discussed in connection with Locke in
UU ch 7-C). Pre-Humean philosophers who adopted this notion used it to distin
guish not only existent from inexistent substances but also different fundamental
kinds of substance, above all material and immaterial. Matter is a substance pre
cisely because it is passively and actively efficacious. For material objects can be
conceived as existent only insofar as they are able to move and be moved, while
anything that can neither move, be moved, or move anything else has no reality
in space and must either be able to act and be acted upon in some other way (e.g.,
thinking and its modes) or cannot be conceived to exist at all (space itself must
therefore either be treated as dependent on matter, as Descartes and Leibniz did,
or on infinite substance, as Spinoza and Newton did). Thus, in the view of some
pre-Humeans, causal efficacy is not just criterial for the real existence of substances
but essential to the very concept of their existence, so that to conceive a substance
to exist JS to incorporate some variety of causal efficacy into its concept, and to
define its nature (material, immaterial, etc.) JS to specify the kind(s) of efficacy
it possesses (the causation of movement in the case of matter, the causation of
thought in the case of thinking being).
This causal efficacy conception of the real existence of substance is, how
ever, vulnerable to Hume's critique of causal relations. First, he showed that
the concept of cause required to conceive real causal efficacy and impute real
(imagination-independent) existence to substances can be derived neither from
pure reason nor from any empirical source. From this, he concluded that there is
no extra-linguistic concept of real causal efficacy at all and traced the existence
accorded to substances to idea-enlivening facile transitions of thought (the prin
ciple that vivacity always follows facility: UU ch 17-C)-the merely subjective
necessity of customary association that Kant branded "a bastard of imagination"
(PFM 258). Hume then adapted this conception of real existence to the case of the
self, using it to distinguish the actual succession of remembered perceptions from
any fictitious one such that, without the facile, idea-enlivening causal associations
that relate present perceptions to their predecessors, "we cou'd only admit of those
perceptions, which are immediately present to our consciousness, nor cou'd those
lively images, with which the memory presents us, be ever receiv'd as true pictures
of past perceptions" (THN 265/173). Yet it is precisely here, as we saw in Section C,
that Hume's account breaks down, in that association presupposes and so cannot
Time Out of Mind 505
explain "the principles, that unite our successive perceptions in our thought or con
sciousness" (636/400). His appendix regarding personal identity in effect concedes
that there must be something in the perceptions themselves, prior to and indepen
dently of experienced constant conjunction and customary association, that objec
tively marks out one succession of perceptions as objectively real from all other
possible successions of those same perceptions, and so as a succession that marks
out the existence of an equally objectively real enduring consciousness in which
these perceptions are united. However, with facile, idea-enlivening associative tran
(Prinzip) of affinity, which has its seat in the understanding and expresses nec
essary connection, a rule of association that is found merely in the copying
imagination and can exhibit only contingent combinations that are in no way
objective. (A766/B794)
14 Some find it curious that Kant made no mention of Hume in the Second Analogy and that
there is little, if anything, in it that seems to have been penned with Hume specifically, or even
primarily, in mind. To understand why, one should keep in mind that the Critique of Pure Reason
is built on the expansion of what Kant sometimes termed "Hume's problem" from the law of
cause and effect to synthetic a priori judgments generally (ch 2-D). T his means that the refuta
tion of Hume is not to be found in any particular part of the Transcendental Analytic, but in
the Analytic as a whole (in conjunction with the Transcendental Aesthetic). In the Prolegomena,
where the problem of the Critique is framed with specific reference to Hume's treatment of cause
and effect, Kant had little choice but to emphasize the importance of Hume's "crux metaphysi
corum" (PFM 312). In the Critique, by contrast, where Kant framed the problem without privi
leging cause and effect, it should come as no surprise that he opted not to introduce Hume into
the Second Analogy, where he sought merely to resolve part, not the whole, of the problem of
the possibility and validity of synthetic a priori judgments. T his, however, does not mean that
the Second Analogy should not be read as a refutation of Hume's doctrine of causality. Quite
the contrary: by showing how the category of cause and effect makes possible experience of the
Time Out of Mind 507
The F irst Analogy adds the existence of the permanent to the fleetingly existent
appearances furnished by sensibility in inner intuition. But what of the existents
between these two extremes-continuants with no "inner necessity to persist"
(Al 85/B229)? The empirical subject (self, person, mind, psyche, soul), its more or
less enduring acts and states, as well as everything else in space and time with which
we have to do in ordinary and scientific experience fall into this category. Hume
dealt with continuants with great depth and originality in the Treatise (perhaps
unknown to Kant when he penned the first Critique). For him, the question of con
tinued existence arises in connection with his view (shared by Kant) that nothing
is ever present to inner sense but a flux of perceptions. Since for every continuant
there is a possible succession of fleetingly existent appearances from which it is
qualitatively indistinguishable, it is in existence only, not quality, that continuants
can be distinguished from the fleeting. But how can an idea of duration of exis
tence be acquired at all if "all impressions are internal and perishing existences, and
appear as such" (THN 194/129)? With no impression of duration of existence, there
can be no idea of it (37/30); and with no idea of duration, there also can be none of
continuance of existence. Believing that there really is such an idea, over and above
linguistic convention, Hume saw no alternative to explicating ideas of continued
existence as fictions of associative imagination (UU ch 17 and HTC Part III).
For Kant, matters were different. He viewed the concept of duration of exis
tence as unproblematic because, in addition to succession, pure intuition affords
us a notion of time "as permanent form of inner intuition," and so the "substra
tum" wherein "alone simultaneity as well as succession can be represented" (B224).
frequent and constant co-occurrence and sequence of perceptions and so how association itself is
possible (ch 13-D), it cannot but be seen as a direct answer to Hume. Moreover, its explanation of
objective succession completes the account begun with the explanation of subjective succession
in terms of the pure time intuition, thus fully and finally surmounting the challenge posed by the
Humean quandary concerning personal identity (as Hume himself was unable to do). I therefore
cannot agree with those like Eric Watkins who infer from the absence of any mention of Hume
in the Second Analogy that the text was neither structured nor intended to reckon with the spe
cifics of the challenge posed by Hume's account of cause and effect (see my review of Watkins'
Kant and the Metaphysics of Causality (Cambridge: Cambridge University Press, 2004) in the
Times Literary Supplement (fall 2005) [but note: due to an editorial error, "Hume's problem"
was changed to "Kant's problem," obliterating the point made in the conclusion of the review]).
508 Kant's Anatomy of the Intelligent Mind
The first thing to note in this regard is that the Second Analogy does not imply
that our successive perceptions are caused by their predecessors. Indeed, if the sub
jective succession of one's perceptions is thought of as a rising vertical line, then it
can perfectly well be the case under this law that the causal series leading up to each
perception takes the form of a diagonal line intersecting it at a single point so that
none of one's preceding perceptions are proximately or remotely involved in the
sufficient causal condition for the existence of any of one's succeeding perceptions.
The only restriction the principle imposes is that their causes, whether apprehended
in empirical intuition or not, be themselves determinations (changes) effected in the
permanent (as substantia phamomenon).
Nevertheless, what commands our attention in cognitive experience are the
objects actually present to us in perception, particularly the outer appearances
apprehended in empirical intuition conformably to pure space. Inner appearances
are nothing except the means whereby the mind exhibits inner affections that first
arise through its active and passive reflection on the representations of the outer
senses ("the representations of the outer senses constitute the actual material with
which we occupy our mind," B67), including its affection by its own activity: pure
and empirical apprehensions, productive and reproductive imaginings, and syn
thesis of every kind, including judgments and inferences (ch 4-C and -D). Outer
appearances, by contrast, are the means whereby the prerefiective spontaneity of
the mind exhibits sensations, and so express the existence of the subjectively uncon
ditioned "transcendental matter of all objects as things in themselves (facticity,
reality)" (Al43/B182; chs 4-A, -B, and 8-E). Since outer appearances never exhibit
self-affections, their existence must be distinguished from the inner appearances
that represent the perceptions of these outer appearances and so exhibit not only the
affection exhibited in the outer appearance itself but also the self-affection involved
in synthesizing that outer appearance (viz. the imagination's synthesis of apprehen
sion in intuition: chs 3-E, 4-C and -D, 6-C, and 8-B). Under the Second Analogy,
this means that the outer appearance and the corresponding inner appearance nec
essarily have different empirical causes because the sufficient causal condition for
the existence of the former can never be sufficient for the latter. Indeed, since the
existence given in outer appearances enters into the causally sufficient condition for
the corresponding inner appearances, these inner appearances are always causally
dependent on the outer appearances apprehended in them while those outer appear
ances are always causally independent of the inner appearances that exhibit their
apprehension.
This is the judgment we all make throughout our lives, regarding, say, the tree
we perceive as the cause of our perception of it and not vice versa. Hume perhaps
expressed it best when describing what he termed "the single existence view of the
vulgar": "there is only a single existence, which I shall call indifferently object or
perception, according as it shall seem best to suit my purpose, understanding by
both of them what any common man means by a hat, or shoe, or stone, or any
other impression, convey'd to him by his senses" (THN 202/134). The vulgar see
Time Out of Mind 509
not an appearance of a hat but a hat, not an appearance of a stone but a stone,
and so on; and since it is only because a hat, a stone, etc. really exist externally that
these objects are internally seen there, i.e., perceived via sensation, the vulgar judge
their perceptions (inner appearances) to be causally dependent on their objects.
They also attribute continued existence to these objects insofar as they cause a suc
cession of perceptions of themselves. For insofar as the perceptions continuously
changing in inner intuition are changes effected by the outer appearances presented
in them, continuous series of perceptions of qualitatively indistinguishable outer
The Second Analogy applies to the whole of changeable existence in time, that
is, everything except the permanent, and so holds for all that, unlike substances
(A205-6/B250-1), does not exist by "inner necessity" (A185/B229). Thus, nothing
changeable exists unless its causally sufficient condition precedes it, which, as itself
changeable, exists only if it is preceded by its causally sufficient condition and so
on without end. The outer appearances that cause perceptions therefore fall within
causal series and may have inner appearances among their causal antecedents (as
when the appearance exhibiting a self-affection coinciding with volition is a mem
ber of the causal series leading to the existence of Michelangelo's David). But no
matter how varied the appearances in any given causal series, all are ipso facto alter
ations of the permanent, and so changes in its state (though Kant also recognized
that there are alterations in spatial relations that are not changes of state of the
permanent and so do not require a cause, e.g., inertial motion: A207/B252n).
The restriction to succession in Kant's causal principle does, however, pose a
problem:
and the succession in time of the latter is occasioned only by the inability of
the cause to perform its entire action in an instant. (A202-3/B247)
One must here be sure to note that it is the order of time and not the lapse of
time that is to be considered; the relation remains although no time elapses.
The time between the causality of the cause and its immediate effect may be
vanishing, so that they are simultaneous, but the relation of the one to the
other still always remains determinable in accordance with time. (A203/B248)
The schema of cause and effect expresses only the irreversible order in time of cause
with respect to effect and nothing else (Al45/Bl 84 and A203/B248; ch 16-C-2), and
so determines cause and effect as adjacent in time, where "adjacent" means that
no time intervenes (elapses) between them (ch 15-C). Only with the addition of
the effect's effect-that is, a second irreversible ground-consequence relation whose
ground is the consequence of the first-is an interval created, and so too the possi
bility of time magnitude (i.e., only then can the synthesis speciosa of the category of
plurality synthesize a number of times: ch 15-E). Thus, under the Second Analogy
512 Kant's Anatomy of the Intelligent Mind
The consequence is posited after the ground as regards time even if this time
is vanishing. But if there is a series of consequences and, indeed, if this series
is infinite, then the succession becomes noticeable (merklich). The immedi
ate ground is the starting point of a series of consequences. (AA 18 § 6378
[1790-1])
15 As the example is presented, Kant enters an already-heated room and locates the cause of
its warmth in the stove, the causality of which is not only coexistent with its effect but enduringly
so. But the sequel makes clear that his real concern is with causes that do and causes that do not
require the elapse of time to complete their effect so that every cause is coexistent either with the
whole or, at least, the beginning of the production of its effect.
Time Out of Mind 513
of the cause that precedes it," A203/B249). And so it is for all Second Analogy
causes: their causality can have immediate effects on more than one time scale-a
coexistent effect, an effect one second hence, another requiring a day, another a
year, another a millennium, and yet another a million years.The Big Bang may have
effects so protracted as to require countless billion years beyond the present before
the entire effective action required to produce them will have been performed; yet,
ordinally, they still qualify as its immediate effects.
This is where the purely ordinal character of causal relations under the Second
Analogy becomes important in relation to continuants. No non-coexistent ordi
nally immediate effect can be conceived to ripen into wholeness if the causality
of its cause does not continue in existence through the entire period of its ripening
("every alteration has a cause that demonstrates its causality in the entire time in
which the alteration advances ... [and] is only possible through a continuous action
of causality," A208/B253--4). This means that, under the Second Analogy, every
ordinally immediate, quantitatively protracted effect presupposes some continu
ously exercised efficacy and so the continued existence of the causality of its cause.
Its beginning coincides with its instantaneous immediate effect and its end with the
cessation of its most protracted yet still immediate effect. In the case of the heated
room, it is the identity over time of a particular heating event that begins with the
ignition of the contents of the stove, continues through the time of their combus
tion, and ends with the extinguishing of the last ember. Similarly, the identity of
the stove is a function of the protracted immediate effects of the continuing efficacy
of some16 cause, all of which must eventually exist in their completeness unless the
efficacy of another cause intervenes to prevent it from performing its entire effec
tive action (the complete effect coinciding with the latest time the stove remains
functional). The identity of an organism is a function of the protracted immediate
effects of the continuing efficacy of some cause, all of which must eventually exist
in their completeness unless the efficacy of another cause prevents it from perform
ing its entire effective action (the complete effect coinciding with the latest time
the organism remains able to sustain life). The identity of the earth is a function
of the protracted immediate effects of the continuing efficacy of some cause, all
of which must eventually exist in their completeness unless the efficacy of another
cause intervenes to prevent it (the complete effect coinciding with the latest time the
materials that make up the earth remain able to cohere). And so on for all continu
ants, be they events, individual things, systems of things, social systems, or anything
else. For whenever causes are supposed to have protracted immediate effects, the
same efficacy is deemed to continue in existence until its entire effective action has
been performed. It thus yields a concept of itself as a continuant, something that
16 Remember that the Second Analogy only gives us a priori assurance of the relation; actual
experience, filtered through the finite fallible minds of actual human beings, is always required to
determine what existent occupies the position of cause in the relation.
514 Kant's Anatomy of the Intelligent Mind
preserves its identity through a period of time, but without any "inner necessity of
persisting."
To cognize appearances as continuants is to represent them as subjects of deter
mination (effected changes). However, it is equally clear from the Second Analogy
that continuants are not final subjects in their own right but mere changes under
gone by subjects more final than they, which are either themselves substantial final
subjects (permanent) or mere transitory states in which such subjects exist. This is
important to bear in mind when considering what does and does not have genuine
agency (and patiency). In Kant's stove example, setting the wood alight alters the
state of the stove, while opening a vent to feed more air into the flame alters the
state of the fire. Does this make continuants that exert their efficacy on other con
tinuants causal agents in their own right? We certainly speak of non-ultimate sub
jects (nonpermanent continuants) as doing this and having that done to them. Yet
Kant refused to accord objective sense and meaning to such talk, instead ascribing
all agency and patiency to substantial final subjects:
Where there is action, and so activity and force, there is substance as well,
and it is in substance alone that this fruitful source of appearances must be
sought . . . . Action already has the meaning of a relation of the subject of
causality to the effect. Now, since all effect consists in that which happens,
and so in the changeable that designates time according to succession, the
final subject (letzte Subjekt) of the changeable is the permanent, as the sub
stratum of all that changes, i.e. substance. For according to the principle of
causality, actions are always the first ground of all change of appearances,
and thus cannot lie in a subject which itself changes, since otherwise further
actions would be required and another subject which determined this change.
(A204-5/B249-50)
If efficacious action is the "first ground" of all change, then continuants, being
changeable, can never, in strict transcendental truth, act or be acted on (only in our
intelligible character can selves be so much as conceived as agents and patients, even
if cognition of ourselves as such is impossible: A544-57/B572-85).
This is not to suggest that the Second Analogy proscribes the ascription of
actions to people, stoves, rain clouds, the sun, or any other nonfinal subjects of
predicates of causal efficacy. On the contrary, by providing a foundation for contin
ued existence in causal efficacy, the principle necessitates it. Nevertheless, because
nonfinal subjects, under the First Analogy, are all mere accidents inhering in final
subjects, all genuinely efficacious actions must originate in substantial final subjects
as their "first grounds." For to conceive continuants through pure concepts as non
final subjects simply means that they are subjects only in a relative sense (e.g., the
stove in relation to a heating event) but, in and of themselves, are mere determina
tions predicated of more fundamental subjects that persist before, during, and after
their existence. These more fundamental subjects, in turn, are either themselves per
manent final subjects, in which case the regress ends, or they are mere continuants,
Time Out of Mind 515
How are we to conclude from the action of what acts directly to its per
manence, this being a feature so essential and peculiar to substance (phe
nomenon)? . . . [A]ction is a sufficient empirical criterion to demonstrate
substantiality, without my needing first to search for its permanence by
comparing perceptions, which could not happen with the exhaustiveness
demanded for the magnitude and strict universal validity of the concept. For
that the first subject of the causality of all coming to be and ceasing to be
cannot, in the field of experience, itself come to be or cease to be is a certain
conclusion, which issues in empirical necessity and permanence in existence,
and so in the concept of a substance as appearance. (A205-6/B250-1)
17 Recent experiments (Benjamin Libet's being the most famous) indicate that the brain has
already signaled the muscles to perform an action well before the agent is conscious of deciding
to do it, and this temporal gap is often taken to prove that consciousness (here, the exercise of
"free" will) cannot be the action's true cause. But what if Kant is right and the conscious psyche
is the author of time (duration, succession, and simultaneity)? In that case, it is not consciousness
itself that occurred at a later time than the brain state but its appearance in inner intuition that
did-a situation that slots easily into Kant's general theory of inner appearances as determina
tions of final subjects in space causally necessitated by outer appearances that are themselves
determinations of such subjects. The questions which came first-the neural determination to act
or the conscious decision-and which caused which have no meaning in relation to the conscious
psyche studied in Kant's a priori psychology.
516 Kant's Anatomy of the Intelligent Mind
these subjects are mere predicates, and so on in the direction of substantial final
subjects (permanent existents). Once the limits of discernibility are reached, tech
nological aids, the application of mathematics, and other cognitive enhancements
can extend empirical inquiry to otherwise indiscernible (though still nonfinal) sub
jects and predicates.
The object of such inquiry is, accordingly, twofold. On the one hand, it seeks
to resolve the time in which the causality of a given cause elapses into its tempo
ral constituents (a succession of states), which, given the continuity and infinite
divisibility of the time-series constituted by the Second Analogy, admits of being
represented even where the time between cause and effect is vanishingly small (as
in calculus). On the other hand, it seeks to resolve the causality of a given cause
into a state of something that persists before, during, and after it, to resolve this
similarly, and so on until continuants are discovered whose reality (intensive mag
nitude) gives every indication of being a conserved, unchanging quantity (without,
of course, our ever being able to arrive empirically at a cognition of them as neces
sarily conserved and as impossible to increase or decrease in quantity, which are
features uniquely of permanent substances cognizable transcendentally or not at
all). In neither direction does our search for new efficacies, actions, and forces ever
come to an end. For even if science were someday to develop to the point where
all known phenomena could be explained (whether by means of a quantum theory
of gravity such as M-theory or something as yet unthought-of), these cannot be
known to coincide with all phenomena accessible and comprehensible under the
Second Analogy. Other creatures, with superior senses, psychology, mathematics,
technology, and/or significative techniques, might regard the phenomena posited
as ultimate in our science as comparatively superficial, as a misleading sampling,
as ultimate but too crudely represented, or as otherwise inadequate to yield a gen
uinely all-encompassing theory of nature. Nevertheless, the Second Analogy of
Experience affirmed by transcendental judgment gives us complete assurance that
the.first ground of all actions, and so the only true causal agents (and patients), are
substantial final subjects: genuinely permanent existents.
Suppose that it is night, and on a nearby hill a beacon is shining . To one side of the
beacon, a gleaming star can be discerned and to the other is a galaxy that is per
ceived obscurely but cannot be clearly perceived (discerned) without a telescope. All
three are perceived simultaneously, which is just to say that the observer is affected
by light from all three sources simultaneously. The question is whether these objects
themselves exist simultaneously. Science tells us that the light from the beacon is
only the tiniest fraction of a second old, while the light from the star originated
centuries ago, and that from the galaxy originated millions of years ago. It thus is
Time Out of Mind 517
perfectly possible that the galaxy may have ceased to exist before the star began to
exist and that the star ceased to exist before the beacon was erected. So what would
have to be the case for all three objects to be simultaneous?
For Kant, the key to answering such questions lies in the recognition that objec
tive simultaneity is possible only through a principle of pure understanding.18
Nothing less than such a principle suffices-here as in all other matters bearing
on the magnitude and relation of distinct existents in time-because of the imper
ceptibility of time (B257, A215/B262; ch 16-B). If time were perceptible, then, in
perceiving the beacon, star, and galaxy, I would immediately perceive that I am
seeing objects that exist in widely separated times. Conversely, in cases where two
perceptions are never had at the same time, such as the front and rear exteriors of
a building, I could perceive their simultaneity. But since time is nothing but a pure,
hence imperceptible intuition of sensibility, entirely lacking the existence only that
can have which is capable of affecting the senses, the difference between simultane
ous and successive existence must be grounded in something other than time itself.
The ground required must not only suffice to distinguish simultaneous from
successive existence objectively but also make their difference cognizable through
perceptions (inner appearances). Given the imperceptibility of time, the only way
simultaneous existents can be distinguished from successive through perceptions
is by means of the order and relation of those perceptions. Since perceptions are
always apprehended successively, their order in intuition, whether regarded as
effected changes under the first two Analogies or not, cannot suffice. Instead, they
must be subjected to a rule determinative of their ordering in reproductive imagina
tion in such a way as to enable us to distinguish simultaneous from successive (tem
porally non-overlapping) existents through them. The question is what that rule is.
It is not a third Analogy of Experience, but a rule that Kant will prove to be
possible only under such a principle: if the order of perceptions is a matter of
indifference, then the existents perceived in them are simultaneous, whereas if it is
impossible to reverse their order, then they are successive (B257, A21 l/B258). If it is
impossible to perceive A after perceiving B or B after perceiving C, then A, B, and
C are successive, temporally non-overlapping existents. But if A can be perceived
after B and C, B before A and after C, and C before both A and B, then A, B, and
C are simultaneous. Of course, to apply this rule to the example of the beacon, star,
and galaxy, one would have to postulate a medium of perception so much faster
than the speed of light that it could communicate information across the distances
between these objects or between any in the universe in times vanishingly small; and
Kant was not in the business of grounding principles of transcendental judgment
on empirical media-light, gravity, sound waves, etc .-much less any as speculative
18 I believe Kant would have understood and, as I will proceed to show, been able to deal with
the example. The speed of light had already been determined to within an order of magnitude
in the seventeenth century, and while galaxies were still a novel notion, Kant himself was among
the first to propose their existence.
518 Kant's Anatomy of the Intelligent Mind
is the recipient of another's influence, then I say that their simultaneity would
not be an object of a possible perception and that the existence of one could
not lead, by any path of empirical synthesis, to the existence of another. For
bearing in mind that they would be separated by a completely empty space,
the perception that advances in time from one substance to the other would
indeed determine that this latter has existence by means of a succeeding per
ception but not be able to distinguish whether the appearance objectively fol
lows the former or is rather simultaneous with it. (A212/B258-9)
had a determination causally dependent on the first, then it would be impossible for
them never to have existed simultaneously ("only by means of their reciprocal influ
ence can they establish their simultaneity," A213/B260). But what if the two were
causally completely isolated from one another so that no determination in either
depended for its existence, directly or indirectly, on any determination of the other?
If time were perceptible, and so had the same existence they do-the kind that
pertains only to what cannot be apprehended in intuition except through sensation
then, notwithstanding their causal isolation, they would be objectively and really
either simultaneous or successive. But given that time is pure, hence imperceptible, and
Time Out of Mind 519
The word "community" is ambiguous in our language and can mean communio
but also commercium. I employ it here in the latter sense, as a dynamical com
munity without which even local community (communio spatii) could never be
cognized.. . . Without community every perception (of appearance in space) is
broken off from the others, and the chain of empirical representations, i .e. expe
rience, would commence anew with each new object, without the preceding per
ception being able to connect up (zusammenhiinge) with it in the least, or stand
in time relation with it. (A213-14/B260-1)
19 Their status as exponents (proxies) for the manifold of pure space in the field of appear
ance is simply an extension to space of their status as exponents of the permanence of time
accorded to the relations predicated of appearances in the Analogies at A216/B264 (analyzed in
Chapter 18). It is justified by the necessity for intuition in space as well as time in order to estab
lish the objective reality of the principles of pure understanding (B291-3).
520 Kant's Anatomy of the Intelligent Mind
21
By "matter" here, I do not mean physical matter but the matter that Kant distinguished in
appearances from their spatial or temporal form (ch 3-B). Of course, in beings constituted like
ourselves, the only kind of substantia phamomena possible is corporeal, since given time alone,
permanence is impossible in the field of appearances: B29 l-3.
522 Kant's Anatomy of the Intelligent Mind
With its conversion into a subject of experience, the subject's appearance in inner intu
ition becomes an object of experience like any other, uniquely differentiated from and
completely determined in relation to all others under Kant's principle of community:
It is easy to recognize from our experiences that only the continuous influ
ences in all positions of space can lead our senses from one object to another,
that the light which plays between our eyes and cosmological objects effects
a mediate community between ourselves and them and thereby proves the
simultaneity of the latter, that we cannot alter our place empirically-per
ceive this alteration-unless matter everywhere makes possible the percep
tion of our place for us, and can only show us their simultaneity by means of
524 Kant's Anatomy of the Intelligent Mind
their reciprocal influence, and, through this, the coexistence of even the most
remote objects (albeit only mediately). (A213/B260)
1 Kant held that space can be filled even by the reality of the least possible non-zero value
of intensive magnitude: Al 72-3/B214. It should also be noted that he was careful to remark
that the reasoning at A213-4/B260- l does not "disprove empty space," which may exist where
empirical cognition of simultaneity is not possible, even if such a space is for that reason "never
an object for all our possible experience" (A214/B261; also A228-9/B281, A487/B515, and A521/
B549). Does this mean that he was prepared to allow that space can still exist as an object with
the same reality/existence possessed by objects of experience even though it fails to conform
to conditions of possible experience? In other words, is Kant's conception of space closer to
Newtonian absolute space than I maintained in the equivalent case of time in chs 16-B and
17-D? The answer, it seems clear, is no. In his proof of the Anticipations of Perception principle,
Kant equated the absence of all reality corresponding to sensation (i.e., negation = 0) with pure
intuition (B208) so that empty (because imperceptible) space is none other than the pure space
of sensibility. Elsewhere, he stated that where both sensation and the composition contributed
by the understanding are eliminated, nothing remains except "the form of intuition and of the
object as appearance" (AA 18 § 6314 [1790-1]; ch 16-C). And if Kant was unequivocal about
anything, it is that space as pure intuition, the form of the object as appearance, must never be
confounded with Newtonian absolute space, e.g., A429/B457n and AA 18 § 5377: "There is no
absolute time or space. Pure intuition does not here mean something that is intuited, but the pure
formal condition that precedes appearance."
2 In its character as pure consciousness, however, this subject is not subject to deterministic
natural law, and the way remains open for the transcendental freedom of the Antinomy chapter
and the positive, though purely practical, freedom of the Critique of Practical Reason.
Our Place in Nature and Its Place in Us 525
(ch 17-C).3 Finally, under the Third Analogy, the empirical subject is cognized a
priori as an object of nature, no less or differently subject to natural law in all its
being and doings than stones, trees, or the sun (the ground of the subject's empiri
cal character as described at A539-40/B567-8).
Of course, being a continuant, the empirical subject is also nothing more than
a contingent (causally conditioned), changeable state (empirical predicate) of the
permanent (substantial final subject in space). When there is a sufficient causal
condition for it to exist, it comes to be, a state into which the existence of the per
manent has passed; it then endures for as long as the efficacy whose ordinally imme
diate, quantitatively protracted effects maintain its existence; and it ceases to exist
the moment that efficacy ceases, ceding its place in the permanent to a successor
state. Moreover, since the permanent is only possible in space (B291-3), the empiri
cal subject, together with all of its states (inner appearances), can be nothing other
than a state of material substance (i.e., of what is permanent in corporeal mat
ter), making it heir to all "that flesh is heir to." The empirical subject of the Three
Analogies of Experience therefore exists in the figure of a human being or, at any
rate, in some other equally embodied corporeal creature with sensibility and under
standing akin to ours:
If, under the Third Analogy of Experience, empirical subjects cannot be represented
otherwise than as inhabitants of the realm of material nature, this transcendental
3 That the nature of the unity of the mind in its empirical character as a continuant is that
of a (nonfinal) subject of determinations (ch 17-H) is particularly evident here: "The question
whether, with distinct inner alterations of the mind of a person (of his memory or of principles
accepted by him), if that person is conscious of these alterations, one can still say he is one and
the same (as to his psyche) is absurd: for he can only be aware of these alterations because he
represents himself as one and the same subject in the distinct states" (Anthropology 134n).
4 This leaves open the question of precisely how the mind, as object of perceptions (inner
appearances), relates to the corporeal: "On the seat of the psyche: whether it is determined as a
material substance in respect of space [externally] in the body or only through the commericium
with the body in respect of the remainder of the world. It has its seat, i .e., the first and immedi
ate connection with the nervous system. if Everything whereof a universal rule can be cognized
through experience is natural. Accordingly, the commercium is natural. Immateriality is a genu
inely problematic concept that cannot be confuted" (AA 18 § 5457 [late 1770s]).
526 Kant's Anatomy of the Intelligent Mind
judgment also separates them from the rest of nature by equipping them to experi
ence and understand that realm. Earlier in life than we can remember (ch 15-D), we
represent everything according to the principle of dynamical community as a single
whole of existence in space and time: ourselves and everything that has been, is,
will, or could ever be present to our senses, as well as all that can, either directly or
through an extended course of reasoning, be inferred to exist from what is present to
the senses. Bodies have effects on other bodies, including the human body, through
the sense organs of which they influence human minds. Reciprocally, human minds
influence human bodies, and, through them, other bodies (if one considers the light
reflected into space by bodily actions occasioned by empirical psyches at the time
our species first evolved, human mental influences have by now reached quite far).
This community extends to objects unperceived, including those difficult or impos
sible for humans to access otherwise than through their sensible effects (e.g., par
ticle collisions at CERN registered by detectors and exhibited in perceptible form
such as a printout). Science, by replacing ordinary notions of material nature with
theories superior to these in almost every way, has extended the community still
further. Why then did Kant regard this "community of apperception" as merely
subjective, a communio rather than a commercium, not a compositum reale but only
ideate (A214-5/B261-2)?
It is subjective because it is defined relatively to human knowledge. We conceptu
alize ourselves as continuants as we do outer appearances that exhibit the requisite
(to use Humean terminology) "constancy and coherence" to be accorded a "con
tinued, distinct existence." This does not mean that we explicitly regard everything
in the community of material nature, particularly what seems most permanent to
us (the earth, sky, sun, and stars), as impermanent nonsubstances that can cease to
exist at any time and must at some time in the past have begun to exist. Implicitly,
however, we do, since we never conceive the continuants that populate the world of
human experience as existing with the "inner necessity" (A185/B229) of the genu
inely permanent. Indeed, it is precisely when we reason most carefully that we real
ize it is entirely from our human perspective that earth, sky, sun, and stars and even
the Big Bang Universe itself seem everlasting, but that on another scale of space
and time, they may be merely ephemeral, local phenomena and so a communio
rather than a true commercium of substances:
The most common way of watering down Kant's iterated claim that the under
standing is nature's lawgiver and, as such, its creator (ch 13-F and -G) is to treat
it as an epistemological rather than an ontological thesis. Accordingly, where he
seems to say that pure concepts of the understanding are constitutive of objects
of experience, their laws, and nature itself as the totality of existence in space and
time, he is interpreted as meaning that these concepts supply the framework that
enables us to rise above subjective, impressionistic, anthropomorphic, and other
idiosyncratic ways of regarding the natural world to attain the kind of genuinely
objective thought capable of spawning sciences. It is thus, on this view, always
merely from a subjective, human standpoint, characterized by limited data and
conceptual resources, that material nature and everything in it are grasped, even
by transcendental philosophy. Because of this, it is impossible ever to transcend
this perspective to attain a perfectly objective cognition of even the most humdrum
objects of everyday experience. Material nature and everything in it are prior to
and independent of the subject, and it is only in relation to our knowledge of them,
and in particular the conditions under which alone such knowledge is possible, that
the understanding stamps its innate transcendental structures on the world much
as donning pink lenses stamps pink on everything in the visual world. And anyone
convinced that Kantian transcendental philosophy is, at its heart, a renunciation
of ontology in favor of epistemology can then point for confirmation to Kant's
insistence that the principles of pure understanding are "merely principles of the
exposition of appearances, and the proud name of an ontology that takes itself to
yield synthetic a priori cognitions of things in general in a systematic doctrine must
give way to a mere analytic of pure understanding" (A247/B303).
From our consideration of the difference between community as a commercium
of substantial final subjects and a communio of nonsubstantial nonfinal subjects
in the previous section, however, it should be clear that Kant's Third Analogy of
Experience affirms precisely the kind of objective, subjectivity-transcending per
spective on material nature that epistemologizing interpreters suppose him to
have forsworn. Community qua commercium transcends the subjective human
knower-relative sphere of the epistemologist, which, even as represented in the most
Our Place in Nature and Its Place in Us 529
advanced science of the day, may, on some scale of space and time, be local and
ephemeral and so a mere communio rather than a commercium. Instead of relating
appearances to knowers subjectively, commercium relates them objectively to mate
rial substances that exist permanently in their own right, prior to and independently
of everything having to do with the accident of their presence in the experience of
knowers via sensation and that, together, constitute the objective whole (totality)
of existence in space and time, i.e., nature. From a purely epistemological stand
point, such substances are neither necessary nor justifiable. That they are unneces
sary is clear, for insofar as persistence through change is necessary for knowledge of
material nature, continuants with de facto permanence-the absence of all evidence
of impermanence or local restriction-will do quite nicely (e.g., for conservation
principles). That they are unjustifiable is equally clear: since permanence implies an
"inner necessity" to exist that makes it quite literally impossible for anything that
has it not to always exist in the future and to have always existed in the past (ch
17-D), not even the most advanced science can be supposed capable of adducing
evidence sufficient to establish the permanence of anything encounterable in experi
ence. The relation of appearances to a commercium of permanent substances must
consequently be shown to be necessary and justifiable ontologically or not at all, and
it is precisely this that Kant believed himself to have achieved in the Third Analogy.6
When Kant renounced "the proud name of ontology" in favor of transcendental
analytic, it was to reject a "systematic doctrine" of "synthetic a priori cognitions of
things in general" -including things in themselves-rather than an ontology of things
as appearances (A247/B303). A systematic doctrine of synthetic a priori cognitions
of "appearances as substances" (A214/B261) is precisely what the Transcendental
Analytic's system of principles is, and, in contexts other than the distinction between
phenomena and noumena, Kant did not hesitate to term it ontology:
6 As we saw in Chapter 17, once permanent substances are affirmed in transcendental judg
ment, actions are then able to serve as an adequate empirical criterion for their existence.
7 Also: "[O]ntology is nothing else than a transcendental logic (subjective)" (AA 17 § 4152
[1769-70]); "It was in ontology that we discussed concepts of the understanding, the employment
of which is possible in experience, because they themselves make experience possible.... Ontology
530 Kant's Anatomy of the Intelligent Mind
is the sum-total of all concepts and principles of pure thought" (AA 18 § 5603 [late 1770s, early
1780s]); and "Ontology is the science of things in general, i.e. of the possibility of our cognition
of things a priori, i.e. independently of experience. It can teach us nothing of things in themselves
but only of the a priori conditions under which we can cognize things in experience in general, i.e.
principles of the possibility of experience" (AA 18 § 5936 [early to mid-l 780s]).
Our Place in Nature and Its Place in Us 531
former the same boundless mathematical determinability that pertains to the latter
by means of its predicate, the schema of community (e.g., a space of potentially
infinite dimensions: ch 15-C).
The second species of subjectivity is the l's a priori representation of itself as
existing outside and independently of time as the analytic unity of apperception
that accompanies and is common to all representations: "the representation I am
which expresses the consciousness that can accompany all thinking is what imme
diately includes the existence of the self, but still no cognition thereof, and so no
empirical cognition, i.e. experience" (B277). As the subject of both time and the
categories, it cannot determine itself by their means (A346/B404 , A362-3 , A401-2,
B422) and so, unlike its appearance in time (the empirical subject), is not part of
material nature at all (A546-7/B574-5). This I is the subject that judges and so is
the subject that predicates the schemata of appearances in transcendental judg
ments (A341-2/B399--400 and A348/B406). As such, it is, quite literally, nature's
creator and lawgiver, at least to the extent that all empirical laws are simply specifi
cations of the transcendental laws whereby the pure I constitutes the commercium
of nature. So, however paradoxical it may seem, the ontologically perfect objectiv
ity attained when material nature is raised from a knower-relative communio to a
knower-transcending commercium derives entirely from the perfect subjectivity of
the pure I as subject of transcendental judgment. The proofs on which the ontologi
cal validity of its judgments are founded thus derive all of their premises from the
nature and workings of its psychology: pure intuitions of sensibility, pure concepts
of the understanding, the transcendental synthesis of productive imagination, and,
finally, transcendental judgment.
The upshot is that nature, considered purely ontologically as the commercium
of the Third Analogy, and so prior to and independently of all knowledge of it by
means of empirical concepts and data (including the empirical inputs essential to
Kant's metaphysics of natural science: Al 71-2/B213, MFPNS 469-70), is noth
ing over and above the original synthetic unity of apperception in its final, fully
developed form:
possible and indeed do so a priori. Empirical laws can obtain and be found
only through experience and indeed in consequence of those original laws
according to which experience itself first become possible. Hence, our
Analogies actually exhibit the unity of nature in the interconnection of all
appearances under certain exponents, which express nothing else than the
relationship of time-insofar as all existence is comprehended within it-to
the unity of apperception, which can occur only in a synthesis according to
rules. Thus, together they say that all appearances lie in one nature, and must
lie therein, since without this a priori unity no unity of experience, and so too
no determination of objects in experience, would be possible. (A216/B263;
also Al 77-8/B220 and A210-11/B256)8
follows: "[A]ll laws of nature, without distinction, stand under higher principles of the under
standing, because they only apply these principles to special cases of appearance. Thus, these
principles alone yield the concept that contains the condition and the exponents, as it were, for
a rule in general, while experience gives the case that stands under the rule" (Al59/Bl 98). Kant
seems to have adapted the notion from syllogistic logic (A331/B387, A414/B441, A416/B443, L
142-3), but also used it in the sense of producing concepts where it is impossible to construct
intuitions: "We must exponentiate (exponieren) concepts if we are unable to construct them.
We cannot construct appearances, although we can intuitions. But we must have rules of their
exposition" (AA 17 § 4678 [1773-5]). This puts one in mind of the discussion in the Methodology
of the difference between philosophical-discursive and mathematical-intuitive synthetic a priori
judgments, particularly A719-23/B747-51, where transcendental concepts are said to designate
neither pure nor empirical intuitions, but merely "the synthesis of empirical intuitions." By com
bining this with the remark at CJ 343 that "bringing a representation of imagination to concepts
means the same as exponentiating it," one can understand the Analogies as exponents in their
capacity as conceptual surrogates in the field of perceptible existence for the imperceptible pure
time produced by the imagination's pure synthesis of apprehension (it is worth noting that the
remark at A216/B263 is prefaced with a reminder that absolute time is not an object of per
ception). Since time itself is in its own right an original synthetic unity of apperception (chs
4-C, -D, and 5), as is the objective time produced by transcendental synthesis speciosa from the
manifold of the pure time intuition (chs 13-H and 15), the Analogies are not just exponents of
time in the field of appearance but also of the original synthetic of apperception shown in the
Transcendental Deduction to be presupposed by all experience and its objects (ch 13). And since
the Analogies, as constitutive of nature and its laws, are presupposed by and implicit in all more
particular, less fundamental laws (both laws of nature qua commercium and, in respect to empiri
cally discoverable laws, laws of nature qua communio), this reading also captures the sense of
"exponent" at Al 59/Bl98.
Our Place in Nature and Its Place in Us 533
AM the world. This is certainly true, but with one caveat: while everything requisite
for existence to become material nature derives from dynamical and mathemati
cal principles of transcendental judgment, and so from subjective conditions of
apperception, existence per se is subjectively unconditioned (ch 8). Since existence
in the field of appearance alone distinguishes actual from merely possible experi
ence, material nature and apperception are not the same insofar as the former is
empirical-material and not simply pure-formal. It would therefore be more accurate
to characterize their relationship thusly: material nature gives actuality to apper
necessitate the order of reproductive synthesis and thereby distinguish one repro
duction of an apprehended manifold as necessary from all others. Yet because every
possible reproduction of such a manifold can be reflected in a concept and there
after necessitated in this way, this merely transforms the problem of distinguishing
the "true" or "objective" reproduction of apprehended appearances from all others
into a problem of distinguishing the "true" or "objective" recognitive synthesis of
apprehended appearances from all others (ch 13-C, -F, and -G).
The only way out of this predicament is an all-encompassing original synthetic
unity of apperception in which all representations are represented in necessary rela
tion to one another, that is, objective unity of apperception (chs 13-H and 14). In
an objective unity of apperception, each possible reproduction of appearances is
determined as having its own unique condition so that with the actualization of that
condition, it, rather than all possible alternatives, is actualized, and the recognitive
concept that represents it is determined as the only genuinely objective concept.
The Transcendental Deduction of the Categories proves such a unity is possible
only by means of the categories and that the transcendental synthesis speciosa of
productive imagination is the means whereby this unity can be conferred on the
appearances apprehended conformably with the conditions of our sensibility, pure
space and time. In particular, transcendental synthesis speciosa requires us to repre
sent the inexhaustibly infinite manifold of pure space and time through an equally
inexhaustible manifold of concepts combinable into an inexhaustible infinity of
judgments. These judgments can then be represented as all determined conform
ably to the categories, with the result that the spaces and times represented through
the concepts combined in these judgments are themselves brought to the objec
tive unity of apperception, with each uniquely differentiated from and completely
determined in relation to every other (ch 15). Finally, by means of transcendental
schematism, this objective unity of apperception is extended from the manifold
of pure-formal intuition to that of empirical-material intuition, thus ensuring that
everything apprehensible in empirical intuition is determined a priori as belonging
to a single whole of existence in space and time that perfectly conforms to the objec
tive unity of pure space and time presupposed by pure mathematics (ch 16).
Since transcendental judgment can relate the schemata of the categories to
appearances only according to their form as a manifold in pure space and time
and according to their matter as a manifold of empirically apprehended (perceived)
realities, a great deal of sensible content must fall by the wayside. The first to go is
everything apprehended in them having to do with feelings (pleasure and pain) and
sensations, each supplanted by the objective reality of the intensive magnitude of
the matter of the appearances that correspond to them. Next to go is everything in
apprehended appearances that is peculiar to a particular standpoint or standpoints
(rainbows, mirages, sensory illusions, etc.), supplanted by extensive magnitude as
that in them which alone is completely indifferent to observational standpoint.
Finally, nothing may be admitted that is in any way specific to humans or other
cognition-capable species but is not necessarily valid for all sensibly conditioned
Our Place in Nature and Its Place in Us 535
understandings that share our pure intuitions and categories (no matter how dif
ferent they may otherwise be from us). Thus, the epistemological, together with all
things scientific, must in the end be supplanted by all and only that which passes
muster with ontology as immanent thinking before empirical subjectivity can be
completely transcended and perfect objectivity attained (sections A-C and ch 17).
If the Axioms of Intuition, Anticipations of Perception, and Analogies of
Experience mark the difference between that in appearances which is and is not
perfectly objective, the Postulates of Empirical Thought specify the ways in which
appearances depend on pure subjectivity to transcend empirical subjectivity and
attain the status of objective phenomena:
The principles of modality are not objectively synthetic, because the predi
cates of possibility, actuality and necessity do not in the least increase the
concept of which they are affirmed, so that still something more is added
to the representation of the object. But since they are nonetheless synthetic,
they are so only subjectively, i.e. they add to the concept of a thing (a real),
whereof they otherwise say nothing, the cognitive faculty in which it origi
nates and has its seat, so that if it is connected merely with the formal condi
tions of experience in the understanding, its object is called possible; if it is
linked to perception (sensation, as the matter of the senses), and determined
through this by means of the understanding, then the object is actual; and
if it is determined through the interconnection of perceptions according to
concepts, then the object is called necessary. The principles of modality thus
affirm of a concept nothing else than the action of the capacity of cognition
whereby the concept is produced. (A233--4/B286-7)
According to the First Postulate, "what agrees with the formal conditions of
experience (according to intuition and to concepts) is possible" (A218/B266). The
possible therefore includes both that which is conceived by means of the transcen
dental schemata-"the phenomenon, or sensible concept of an object in agreement
with the category" (Al46/Bl86)-as predicated of appearances in the objective
principles (Axioms, Anticipations, and Analogies) and that which is conceived by
means of all concepts of appearances to the extent they are capable of incorpo
rating these schemata into their content (i.e., those contents that remain after the
aforementioned subjective factors have all been excluded) (A220/B267-8).There are
two further cases that Kant also saw fit to consider: mathematical concepts and
concepts invented by putting together elements derived from experience in combi
nations or relations in which they have never been encountered.
(l) To construct a concept in pure intuition does not yet show that its object is a
possible phenomenon or object of experience. A constructed object is possible in this
sense only insofar as the pure space in which it is constructed "is a formal a priori
condition of outer experiences, that the very same formative synthesis whereby we
construct [the object] in imagination is entirely one with that which we exercise in
536 Kant's Anatomy of the Intelligent Mind
9 Also: "What is determined in time and space is actual. Against idealism." (AA 23 XCII at
A218f.). And: "Something whose relation (to everything possible) is determined in absolute space
and time is actual" (AA 18 § 6290 [1783-4]).
10
How is it known, and knowable a priori, that the community (commercium) affirmed in
the Third Analogy comprises an inexhaustible infinity of permanent substances? The ground of
proof of the existence of this community is the demand for an exponent of the synthetic unity
of the manifold of pure space and time in the field of appearances, and this unity consists of
an inexhaustible infinity of spaces and times each uniquely differentiated and completely deter
mined in relation to every other.
Our Place in Nature and Its Place in Us 537
[T]he perception that provides material to the concept is the sole characteristic
of actuality. But we can also cognize an existence prior to its perception, and
so comparatively a priori, if only it is interconnected with some perception in
accordance with the principles of their empirical connection (the Analogies).
For the existence of the thing is then interconnected with our perceptions in
a possible experience, and we can reach the thing from our actual perception
in the series of possible perceptions according to the guiding thread of these
Analogies.... [W]here perception and its supplementation by empirical laws
reaches, so too does our cognition of the existence of things. If we do not
start from experience, or if we do not advance according to the laws of the
empirical interconnection of appearances, then we make an idle pretense of
wishing to guess or inquire into the existence of any thing. (A225-6/B273-4;
also A600-l/B628-9)11
Appearances that are possible as phenomena under the First Postulate can
only become actual under the Second if there is a causally sufficient condition that
necessitates their existence and precludes anything else from existing in their stead.
In other words, since actuality can be predicated of appearances only insofar as
they are subsumable as phenomena under the Analogies of Experience, the Second
Postulate not only sets their existence in the space and time these principles consti
tute in empirical-material intuition but, through the Second Analogy in particular,
also determines all phenomena as conditioned existents. Accordingly, if a causally
sufficient condition for a phenomenon to become actual at a certain place and time
"It is here, as a corollary to the Second Postulate, that Kant saw fit to add a Refutation of
Idealism in the B edition of the Critique of Pure Reason. The reason the treatment of this princi
ple was the natural place for him to insert the argument is that, in his view, the issue of Cartesian
problematic idealism turns on the conditions requisite for empirical cognition of the cogito (the
I think) as a phenomenon existing together (in commercium) with all other existents in the objec
tive time constituted by the Analogies in empirical-material intuition. Since all cognition presup
poses the existence of the permanent in the field of appearance, the truth of problematic idealism
comes down to the question whether the cogito is or is not capable of serving as the permanent
to which the determination of time in inner sense can be correlated (B278). Kant argued that it
cannot, not only because "we can deal with time-determination only through change in outer
relations (motion) in relation to the permanent in space (e.g. the movement of the sun in respect
to objects on earth)," but also because "we never have anything permanent that, as intuition,
could support our concept of a substance except only matter,· and this permanence itself is not
drawn from outer experience but is rather presupposed a priori as a necessary condition of all
time-determination, and so too as a determination of inner sense through the existence of outer
things in respect to our own existence" (B277-8). Kant fleshed this out in the general note to the
principles, arguing that none of the modes of time-permanence, succession, and simultaneity
can be given in intuition except by means of outer intuitions (B291-3).
538 Kant's Anatomy of the Intelligent Mind
exists, that phenomenon necessarily comes to exist there and then and is impossible
otherwise. Thus, the Third Postulate of Empirical Thought determines all imperma
nent appearances subsumable under the First and Second Postulates as necessary
existents: "That the connection of which with the actual is determined in accor
dance with universal conditions of experience is (exists) necessarily" (A218/B266).
Under the Third Postulate, natural necessity is determined as intelligible neces
sity: no fate, no mystery, nothing whatsoever can intervene to prevent, delay, or in
any way hinder nature from taking its course once the condition for any phenom
enon to exist is met. Since this is true of the condition's own condition, and so on
regressively without end, any possibility of causes not determinable conformably
to the Second Analogy, including supernatural causes, intervening or overriding
the course of nature under that principle is precluded (A228/B280-l). Nature is a
causally closed system in which everything that happens is entirely an affair of per
manently existing material agents and patients.12 And since nature itself is simply
the unity of apperception produced by the understanding through its pure concepts
(A216/B263), the true meaning of the Third Postulate is that nature-at every scale,
in every dimension, under all conditions, and in every set of circumstances-neces
sarily conforms to and can never diverge from the laws of pure understanding:
These four propositions-in mundo non datur hiatus, non datur saltus, non
datur casus, non datur fatum [in the world there are no gaps, no leaps, no
chance happenings, no interventions of fate]-.. . are all united solely in this,
in admitting nothing in the empirical synthesis that could inflict a rupture or
do detriment to the understanding and the continuous interconnection of all
appearances, i.e. to the unity of its concepts. For it is in the understanding
that the unity of experience in which all perceptions must have their place
becomes possible. (A229-30/B282)
The conception of material nature affirmed in the Third Postulate may seem to
sit uncomfortably with quantum theory, which on certain interpretations involves
chance happenings, causal leaps (action at a distance), entangled fates, physically
efficacious conscious free will, and, potentially at least, gaps in space and time
themselves (a possibility in an eventual quantum theory of gravity). Similarly, the
principle might be thought to set Kant in inflexible (and futile) opposition to any
future physics that diverged still more radically from Newtonianism than quan
tum theory and relativity. There is, however, a problem with such objections: Kant's
transcendental account of nature can only come into conflict with contemporary
mathematical physics if he intended it to explain the same things in the same
kind of way the latter does, and there is every reason to doubt that this was so.
12
This may still leave the door open to property dualism, although for Kant, designating
certain appearances "inner," as conforming only to time but not space, is not a metaphysical des
ignation that implies that inner appearances are mental and not physical and so cannot be used
as evidence that he was any kind of property dualist. Indeed, as far as I can tell, the classificatory
scheme favored in philosophy of mind today has little if any relevance to Kant.
Our Place in Nature and Its Place in Us 539
in their quest for a "theory of everything," physicists often speak of someday arriving
at definitive measures. For example, if such a theory were to succeed in establishing
a quantum theory of gravity that reduced general relativity to an effective (merely
approximate) theory, empirical space and time might have to be understood as emer
gent in relation to more fundamental, Planck-scale quanta, thus in effect establishing
an absolute minimum beyond which their further division is impossible. The problem
with such claims is that a theory of "everything" really only explains the forces and
phenomena with which humans are acquainted. For how can we-existing at the
13 This is why it is so important not to confound the infinite space in the field of appearance
constituted through transcendental laws, especially the Analogies of Experience in their capacity
as conceptual exponents (see note 8 above), with the empirical space of the Metaphysical First
Principles, defined as "the sum total of all objects of experience and itself an object of experi
ence" (MFPNS 481). The former is the presupposition of the latter, which is merely a finite
extensive magnitude constituted through non-transcendental concepts of motion that always
admit of enlargement: "such space ... is itself movable," and "a movable space, if its motion is to
be capable of being perceived, presupposes again another enlarged material space in which it is
movable, and this enlarged space presupposes just as well another, and so to infinity.. .. The space
in which motion is perceived is a relative space, which itself moves again."
Our Place in Nature and Its Place in Us 541
and their empirical specifications (i.e., natural laws generally). The upshot is this.
On the one hand, the Analogies of Experience, as exponents of space and time in
the field of appearance, set appearances in a nature fully as infinite as pure space
and time themselves and, in so doing, confer on them an objectivity that, as a priori,
frees their existence as phenomenal substances from any dependence on the empiri
cal accident of their presence to consciousness in sensation. On the other hand, the
Third Postulate of Empirical Thought determines that nowhere in nature's infinite
realm can anything except phenomenal substances exist so that nature itself and
everything in it are nothing apart from that most subjective representation of all,
original apperception.14
14 Insofar as Kant's psychologism purports to show that our notions of material substance,
the inwardly experienced self, space and time, cause and effect, number, and everything else
objective are essentially bound up by content with pure self-consciousness, only nonsense can
result if they are supposed to correspond to the reality of objects as they exist in themselves,
independently of any relation to original apperception. There is thus no escaping the conclusion
that for Kant, the understanding, through the unity of its apperception, is the author of both the
material universe and the inner life of the mind (though not of existence as such, as given in sense
affection). One should not, however, be too quick to condemn him. We live in a time when dual
ism, panpsychism, and neutral monism are each enjoying a resurgence, while physicists speculate
about extra dimensions, the universe being a hologram, and multiverses of various kinds. So why
should psychological idealism continue to be such a bugbear? In Kant's hands, at any rate, it
deserves to be brought back within the pale and treated as a respectable philosophical position.
{ Conclusion }
I stated in the preface that the purpose of this book was to convince readers that
much of what they think they know about Kant's transcendental analytic of the
understanding needs to be unlearned. I have attempted to deliver on that promise by
showing that the very analyses and proofs whereby Kant is commonly supposed to
have demarcated the objective from the subjective are in fact psychologistic explica
tions that exhibit space, time, nature, mathematics, and even logic itself to be nothing
more than diverse expressions of pure self-consciousness (apperception), and so spe
cies of subjectivity. All that remains, therefore, is to reassess Kant's legacy accordingly.
Interpreters who take apperception to presuppose the categories are only reasoning
consequently when they deny that Kant should be ranked with Hume and other
exponents of a sensibilist theory of ideas that traces all representational content to
sensible origins. Yet if the understanding comes by its conceptual bounty entirely
independently of sensibility, what then is the origin of the categories, both of the
contents thought in them and how they come to be represented in our minds?
As far as I can see, there are only two ways their origin can be explained on any
such anti-sensibilist reading: either they arise in an act of spontaneous creation or
they originate objectively (mind-independently) and are innately implanted in the
understanding (intellectualism). The first option is, of course, a nonstarter: far from
explaining their origin, creation ex nihilo is just another way of calling it inexpli
cable. But does the second fare any better? Implantationist scenarios oblige one to
distinguish the source of the categories themselves from their source in us (how they
become part of the innate representational endowment of human understanding).
In the former regard, the categories might be traced either to some physical source
544 Kant's Anatomy of the Intelligent Mind
tive manifold of space and time, it is the essential precondition for the objectivity
of intuitive cognition, be it mathematical or empirical (ordinary or scientific). For
without the framework provided by the infinite, universally differentiated, com
pletely determined manifold of transcendental space and time, objectivity-the
necessary relation of distinct intuitions-would be impossible. Thus, when Kant's
account of space and time is interpreted in genuinely sensibilist fashion, his strict
demarcation of the transcendental from the mathematical and empirical can not
only be preserved, but the former emerges in precisely the grounding relation with
respect to the latter that he affirmed.
T hanks to the psychologistic method that he adapted from Hume, Kant's sensi
bilism also preserves his demarcation within the transcendental sphere of the vari
ous contributions each faculty makes to the constitution of transcendental space
and time. Before the imagination can synthesize their manifolds in accordance with
the categories, they must be apprehended. But given receptivity alone (the synopsis
of the manifold), the manifold of sense is too heterogeneous to be represented as
the homogeneous manifold of a single representation, all united in the conscious
ness of that representation (unity of sensibility). To make good this want, intuitions
of space and time must be formed through a pure synthesis of apprehension, a
synthesis that depends as much on the (prediscursive) spontaneity of imagination
and apperception as on receptivity. At the same time, sensibilist psychologism dis
closes what this synthesis is incapable of doing by itself: conferring objective unity
on the manifold of pure space and time by uniquely differentiating and completely
determining each space and time with respect to every other. It thereby demarcates
the contributions of sensibility (including not only synopsis but also the predis
cursive spontaneity of imagination and apperception) from those of (discursive)
understanding. It further shows that the objective unity of the manifold of space
548 Kant's Anatomy of the Intelligent Mind
and time is impossible given non-transcendental concepts alone, and that only pure
concepts of the understanding, explicated (in their transcendental deduction) as
conditions of the objective unity of apperception, are capable of making good this
want. Psychologism then reveals how this happens: only if pure understanding is
supplemented by transcendental judgment can the objective unity founded on the
categories be extended from the manifold of pure-formal space and time to the
manifold of realities apprehended in empirical-material intuition. And last, but
by no means least, it shows that transcendental judgments suffice only for purely
philosophical-discursive synthetic a priori cognition, thereby marking the precise
point where transcendental cognition leaves off and properly intuitive (mathemati
cal and empirical) cognition begins.
What this makes clear is that a primary recommendation of Kant's psychologistic
method is its deftness at exposing and eliminating free riders. For by disclosing that
the transcendental faculties of the mind contribute essential elements of the con
tent of our most fundamental representations, what each contributes and, equally
importantly, what none of them do (i.e., without additional non-transcendental
empirical or mathematical-input), psychologism shows that
(1) the unity of the manifold in space and time is not something receptivity (the
synopsis of sense) alone is capable of furnishing, because the prediscursive
spontaneity of imagination and apperception are required in addition;
(2) the differentiation and determination of the manifold of pure intuition
are something that prediscursive spontaneity is incapable of furnishing by
itself, because discursive spontaneity, with its categorial synthesis speciosa, is
required as well;
(3) Euclidean space, or indeed any properly mathematical object, is something
discursive understanding is incapable of furnishing by itself, because actual
construction in pure intuition, whether ostensive or symbolic, is required
as well.
(4) truth functional connectives, set theoretical operators, and other
constituents of mathematical logic involve more than a mere capacity
to judge equipped with logical forms suitable only for combining
ADA-universals and ensuring the synthetic unity of the judging subject,
because actual symbolic constructions of the manifold of pure intuition are
required in addition; and
(5) Newton's, Einstein's, or any physical theories and models that express more
than simply the conformity of appearances to conditions of the objective
unity of apperception are not something transcendental judgment alone is
capable of furnishing; applied mathematics, involving actual experience in
the form of empirical concepts and observations, is required as well.
The results of reversing the frame from Kant the anti-psychological critic of psy
chological philosophy to Kant the psychologistic critic of anti-psychological phi
losophy have significance not only for the philosophy of mind but for the mind
sciences as well, where there is also much that is "anti-psychological" in the relevant
sense to criticize. A case in point is their well-nigh total exclusion of consciousness
from the nature and workings of intelligent mind. Although notions vary widely,
consciousness is generally viewed as, at best, an optional extra, along for the neural
ride but inessential to virtually all mental functioning, including everything that
enters into making our minds what we call "intelligent." Indeed, so elusive is the
mind-brain's need for consciousness that one of the greatest puzzles in the sciences
of mind today is why it exists at all (e.g., what possible evolutionary advantage
could it have conferred?).
If there is anything for which consciousness still tends to be considered indis
pensable it is "qualia": not the experience of such-or-such, much less the such-or
such experienced, but "what it's like" experiencing it, that special, "je-ne-sais-quoi"
qualitative dimension of experiences. For those scientists and philosophers inclined
to this way of thinking, qualia pose a "hard problem": mental phenomena that
cannot exist without consciousness and so seem incapable of being explained
purely physically. But even if that were so, when it comes to intelligent mind, the
hard problem becomes easy: since qualia contribute nothing to making our minds
intelligent, they, and whatever ontological conundrums they may pose, are entirely
ignorable. One thinks here of Locke's surmise that even in creatures as primitive as
oysters, "there is some small dull Perception, whereby they are distinguished from
perfect Insensibility" (ECHU II/ix/§14). I see no reason to preclude that percep
tions in the simplest creatures possessed of consciousness have qualia, so that there
is an answer to the question of what it is like to be that creature, whether an oyster
or some other. And this is just to say that qualia, whether in humans or anything
else, are an extremely primitive species of mentation, which therefore seem not only
irrelevant to intelligence but also perfectly dispensable for its existence.3
While Kant would probably have agreed that qualia have no place in the investi
gation of intelligent mind, he certainly would not have regarded that as a reason to
exclude consciousness from it. Of course, everything hinges on the source from which
one draws the relevant notions (consciousness, intelligence, conscious intelligence,
intelligent consciousness; attention, attentive consciousness, conscious attention;
3 If it is objected that qualia are at least not irrelevant to the knowledge of qualia themselves,
it first has to be asked whether qualia are fit "objects" for inquiry at all. Since they are supposed
merely to "color" an experience but never to define it, the very attempt to turn an inquiring gaze
upon them constitutes a different experience, with its own coloration of qualia displacing the
qualia of the experience one sought to scrutinize. Similarly, they are inaccessible to memory since
the quale of the remembrance itself displaces any other. Thus, intelligent mind seems just as use
less for explaining qualia as qualia are for explaining intelligent mind.
Reversing the Frame 551
referenced phenomena from the best physical science of the day, psychology too must
be prepared to cut loose from the vernacular. Vernacular meaning and use are human
artifacts shaped by history, subject at every turn to cultural forces and governed by
conventions rather than natural laws, whereas consciousness and intelligence are natu
ral phenomena that not only are independent of societal norms but also precede and
make them possible. Word meaning and use should therefore never be mistaken for
psychological knowledge of consciousness or intelligent mind.
The same is true the other way around. Much goes on in the minds of speakers and
their listeners, some of it conscious and some of it not, some of it intelligent and some
of it not. But only empty babble can result if the sounds one utters are not underwrit
ten by any conventional meaning and use in language, much less if they conflict with
it. Not even a complete psychological knowledge of the contents of the mind-brains of
those engaging in dialogue can guarantee that linguistic conventions exist to determine
a meaning and prescribe a use for their utterances. Even if they do, that psychological
knowledge probably still would not suffice for a linguistic knowledge of what those
meanings and uses are, at least not so as to permit scientifically sophisticated aliens
to use the former to comprehend the latter if they had no previous acquaintance with
human language or anything much like it. For the locus of linguistic conventions is
not individual minds or collections of minds, much less any sort of collective mind.
Following W ittgenstein, they may be compared to games of a kind where the rules do
not need codifying and the practice consists in conforming one's actions to conven
tions that became established in the distant past without being designed, sanctioned,
or assented to (this applies to the bedrock conventions of language, in contrast to the
more ephemeral, specialized uses one finds in the sciences, religious rites, oath taking,
etc.). This is just to say that the conscious and unconscious mental phenomena inci
dent to language can never be equated with the meanings, references, or uses of words,
and this is just as true of the psychological vernacular as the rest. Mental phenomena
are, for all vernacular intents and purposes, transparent to language and might as well
be absent (as in automatons and zombies) for all they contribute to meaning and use.4
4 Wittgenstein: "It is of course thinkable that two people from a tribe unacquainted with
games seat themselves before a chess board and execute the moves of a game of chess, and even
do so with all the accompanying psychical phenomena. And if we saw this, we would say they
played chess" (Philosophical Investigations§ 200). But can it be said that their intention is to play
a game of chess? Elsewhere, Wittgenstein explained why it would be nonsensical to say yes: "An
intention is embedded in its situation, in human customs and institutions. If the technique of the
552 Kant's Anatomy of the Intelligent Mind
Thus, consciousness, attention, memory, and other goings-on in the mind-brain cor
relate no more reliably or informatively with the psychological vernacular than the
latter does with them.
Yet it is precisely this insight that current philosophical lore denies to early mod
ern theorists of ideas when they are supposed to have treated ideas (conscious men
tal phenomena) as the meanings of words and the principal determinants of their
use. Thus, one finds Locke frogmarched to center stage in the historical introduc
tion to the typical undergraduate analytic philosophy of language course as a signal
representative of now discredited early modern private language theories because
he supposedly situated meanings in the mind, prior to and independently of every
thing intersubjective (conventions) or objective (Fregean senses and references,
etc.). Kant's Lockean propositional subjectivism could just as easily be used for the
same purpose because it defines judgments as analytic only if the judger actually
thinks the predicate when forming the subject concept, regardless of what vernacu
lar language and/or science prescribes that he ought to think in it , so that concepts
are no less subjective and variable from individual to individual than associations
(ch 2-A and -B). Applied to the psychological vernacular, it becomes the view that
words like "consciousness" and "intelligence" are distinguished from meaningless
noises purely psychologically by the ideas (Vorstellugen) in the minds of their users.
Yet, as I have been at pains to show in both volumes of this book, Locke, Kant, and
their peers were under no illusion that meaning could be conferred on words prior
to and independently of instituted conventions ("rules of propriety"), rendering
any equation of the meanings of words with ideas in the mind untenable. Instead,
they espoused the view that the mind's actual psychology, even in its most essential
aspects, may leave little if any trace of itself in the vernacular and that it is therefore
essential that its psychology be approached without taking any vernacular concep
tions of consciousness or intelligence for granted. Kant, most radically, theorized
about human understanding in a manner almost totally divorced from the vernacu
lar reality of consciousness and intelligent mind. We therefore must look elsewhere
for the source of the concepts of consciousness and intelligent mind that informed
the psychological philosophy of Kant and other early modern theorists of ideas.
Current philosophical lore also identifies introspection as the empiricists' pri
mary source of psychological information, including their conceptions of con
sciousness and intelligent mind. Thus, Locke, Berkeley, or Hume are regularly
game of chess did not exist, I could not intend to play a game of chess"(§ 337). The point is that
language-games are just as dependent in their way on human customs and institutions as real
games are, be the accompanying psychical phenomena what they may; and the psychological ver
nacular is no exception. An informative account of how linguistic conventions arise, develop, and
are superseded can be found in The Unfolding of Language, by Guy Deutscher (New York: Henry
Holt and Company, 2005). Unfortunately, the statistical methods available today can yield evi
dence about the history of language that goes back no further than 20,000 years. They therefore
can tell us nothing about the conventions originally responsible for getting language up and run
ning somewhere between 60,000 and 200,000 years go.
Reversing the Frame 553
no impression constant and invariable ... [so that] when I enter most inti
mately into what I call myself, I always stumble on some particular percep
tion, of heat or cold, light or shade, love or hatred, pain or pleasure. I never
can catch myself at any time without a perception, and never can observe any
thing but the perception. (THN 251-2/164-5)
5 The point is clearer still if one accepts that qualia cannot distinguish the intelligent I from
the sensing or sensing/imagining but non-intelligent consciousness. However, since Descartes,
unlike the sensibilists and most other intellectualists, rejected the notion of non-intelligent minds,
i.e., minds with no cogito (e.g., nonhuman animal consciousness), this line of reasoning was not
open to him.
Reversing the Frame 555
appearance it may be found to undergo, and (3) through every change in appear
ance it may be imagined to undergo. And even then, that is not sufficient: to per
ceive the wax or any body is to perceive something that admits of an immeasurable
number of changes in appearance, far beyond anything mere imagining can encom
pass. Since only intellect is capable of perceiving such variability, it therefore falls to
this faculty to perceive the wax: "bodies are not strictly perceived by the senses or
the faculty of imagination but by the intellect alone, and ...this perception derives
not from their being touched or seen but from their being understood" (Second
of predication at all, but "only something real that is given, given indeed only to
thought in general, and so neither as appearance nor as a thing (Sache) in itself
(noumenon), but as something that in fact exists and in the proposition I think is
designated as such an existent" (B422-3n). The only representable meaning that can
be extracted from this otherwise completely "indeterminate perception" is, accord
ing to Kant, purely logical.6
The logical character of the I think is easily discerned in Descartes's analysis. His
affirmation of the l's objective reality as a finite thinking substance was based on
the unique epistemological status he accorded it. That status derives from his proof
that the I of intellectual self-consciousness is known to exist with a certainty that
stands even if one supposes that the intellect is determined to err by its very nature
so that even the simplest, most intuitive truths of mathematics and metaphysics
are not merely false but also objectively incapable of being true ("chimerical"). To
Descartes, this meant that the cogito brings an element of truth to every thought it
accompanies; and since it must be able to accompany every possible thought, from
the most primitive sense perceptions to the most complicated inferences, it can serve
both as a touchstone against which to assess the certainty of any other thought
and, insofar as its certainty admits of being extended, as a foundation for further
indefeasible knowledge. But to Kant, the only genuinely indefeasible truth present
in these considerations is that the I think, being the only representation necessarily
able to accompany all others, is a consciousness logically common to them all and
so is the one and only representation with the intrinsic logical value of a univer
sal. Being otherwise empty, devoid of all intuitable or thinkable content, he saw
fit to equate it with the form of universality itself: the representation that by being
added to any non-universal representation ipso facto converts it into a concept.
Thus, from the perspective of Kant, what Descartes failed to recognize is that this
purely logical meaning completely exhausts the properly intellectual contribution
to self-consciousness, leaving no warrant for any further intellectualization of the
I think.
Beyond logical universality, the sole and entire properly intellectual meaning of
the representation I think, Kant could find nothing in it beyond the bare fact of
consciousness itself, which he regarded as "nothing more than a feeling of an exis
tence without the least concept" (PFM 334n). This is the same fact of consciousness
evident even in the most primitive, least intellectualized representations of imagi
nation and sense (including qualia and introspective internal perception), which
Kant, pace Descartes, was perfectly content to attribute to lesser creatures (perhaps
even Locke's oysters). Since this is just to say that the properly intellectual meaning
6 "The I in every judgment is neither an intuition nor a concept and not at all a determination
of any object, but an act of the understanding of a determinative subject in general and the con
sciousness of itself, pure apperception itself, thus belongs merely to logic (apart from all matter
and content)." (from Kant's unpublished addition to the remark at Anthropology§ 7 141). See
chs 4, 8-F, 9, 10-A, and 18-C.
Reversing the Frame 557
This, then, is the concept of consciousness that Kant offered in proof of his psy
chologistic thesis that consciousness is essential to intelligent mind. The implication
is not merely that anything that lacks consciousness ipsofacto lacks all of the features
of intelligent mind, starting with the capacity for propositional thought (concep
tion, judgment, and inference), but, more specifically, that we either contradict our
selves or talk without a meaning if we attribute intelligence to anything-biological,
electronic, or whatever-that lacks it in the form of intellectual self-consciousness.
I doubt whether an accurate understanding of the concept of consciousness that
Kant utilized in his anatomy of the intelligent mind will improve the reception of
his a priori psychology among philosophers and scientists of mind. But it should, at
the very least, make clear that the current practice in books and courses of present
ing the cogito as if it were synonymous with metaphysical dualism or epistemologi
cal foundationalism is, post-Kant, both negligent and misleading. The I think needs
to be assessed in Kant's more rigorously analyzed form, freed from all of the excess
epistemological and metaphysical baggage with which Descartes and his successors,
both empiricist and rationalist, encumbered it. And this, at the very least, means
that instead of following the usual procedure of skipping from Descartes straight
to Mill or James, with perhaps a brief stopover at Locke's anti-innatism or Hume's
skepticism, philosophers and scientists of mind need to recognize that one cannot
even begin to do justice to early modern conceptions of intelligent mind without
making Kant's a priori psychology one's central focus.
Nor is that all. In Kant's view, the I think not only validates the a priori psycho
logical investigation of thought but also constitutes its sole and entire explanandum.
For as soon as it is asked how the representation of an I think able to accompany
all other representations is possible, a further question poses itself: are there other
equally a priori conscious representations it presupposes as its explanans? This may
not seem like a question that need concern scientists of mind, but it may actu
ally fall closer to their remit than the cogito. For if the original synthetic unity
of apperception that Kant held to precede and make possible the representation
I think (analytic unity of apperception) must be in place ahead of all conceptual
thought, then this synthetic unity must be understood prediscursively, and so in
terms ultimately of sensibility and its unity: which psychological conditions must
be in place a priori if sensibility is to constitute an original unity of consciousness?
As far as I am aware, this has hitherto no more been a question for psychologists
than it has been for post-Kantian philosophers. But shouldn't it be? It might at
558 Kant's Anatomy of the Intelligent Mind
first seem no different from the question of the possibility of consciousness itself,
since that too involves some notion of sensibility as a unity, at least as involving the
compresence in consciousness of manifold, heterogeneous sensory data. Yet Kant's
question of the unity of sensibility is subtly and importantly different. For him, it
must be framed specifically with the possibility of thought defined by the analytic
unity of apperception in mind: not how is just any unity of sensibility possible, but
rather that specifically required if the I think, and so thought itself, is to be pos
sible? It is the answer to this that leads directly to the notion of pure yet sensible
7 E.g., "_ is sitting near_" is a flexible propositional form that can take two or more object
arguments, as in "John is sitting near Beverley," "Jim is sitting near Jody and Judy," and "Louis
is sitting near the Settembrini Sextet."
Reversing the Frame 559
the unity of the judging subject and for nothing else. Thus, Kant's analysis of the
I think promises a new, more modest psychological lease on life to logical forms
that post-Fregean mathematical logic and modern propositional analysis have long
since driven into desuetude.
The remainder of Kant's a priori psychological account of thought is concerned
with how original apperception functions not just as a judging subject but also as
an experiencing, cognizing one. At its core is the quest for categories, that is, pure
concepts of objects that are underivable from any others and from which all other
concepts of objects, pure and empirical, must derive. The metaphysical deduction
whereby Kant derived his table of categories from his table of judgments has gen
erally not been well received. Not only does it inherit the decrepitude commonly
attributed to the derivation of the table of judgments from Aristotelean logic, but
Kant's failure to spell out how each category originates in a particular logical func
tion also leaves many questions unanswered and feeds the suspicion that the con
cepts in the table of categories cannot, in any meaningful sense, be "deduced" from
propositional forms, much less be accorded "transcendental content." The result,
in practice, is that interpreters read whatever meaning into the concepts of "sub
stance," "cause and effect," etc. they think Kant's text can bear, so that the doctrine
of categories we find in the literature is typically encumbered with an assortment of
free riders and seldom amounts to more than an ungainly hodgepodge.
What is almost invariably overlooked is that Kant subjected his doctrine of cat
egories to the same restriction he applied to his doctrines of pure sensibility and
general logic: since a priori psychology has but one datum on which to draw, the
I think , its inquiry can never have any other focus. This means that there is no other
source from which to derive fundamental concepts of objects but a priori psychol
ogy itself; and since Kant stripped the I think of the objective meaning Descartes
attributed to it as res cogitans, this leaves only its logical meaning as the one rep
resentation with the intrinsic worth of logical universality, the aesthetic forms that
make it possible, and the logical forms that make possible the discursive synthetic
unity of the I think as a judging subject. The question is whether any concepts
derivable from these purely a priori psychological sources amount to fundamental
concepts of objects. The addition of the analytic unity of the I think to pure sensi
bility yields a variety of concepts (ADA-universals): concepts of sensibility as an a
priori unity, of pure intuitions and their manifold, of pure space, pure time, succes
sion, and juxtaposition. But where everything is so completely undifferentiated and
indeterminate, and so scattered and single that nothing can emerge beyond "a blind
play of representations less than a dream," no concepts of objects are possible,
much less concepts of objects so fundamental as to qualify as categories. In the case
of logical functions, the only concepts the AUA makes possible are those whereby
given concepts and judgments can be determined conformably to the logical func
tions, such as the concept that annuls the logical reversibility whereby a concept can
be used either as subject or predicate. And it is, of course, here that Kant found the
concepts that constitute his table of categories.
560 Kant's Anatomy of the Intelligent Mind
When one first encounters Kant's claim that concepts which do nothing more
than annul logical reversibility are categories, one cannot help but be skeptical. They
are completely devoid of objective content, both empirical and mathematical, and,
having none but the logical meaning they derive from logical functions, also lack the
objective reality that intellectualists surreptitiously introduce through the backdoor
of implantationist innatism or illuminationist insight. They are indeed what Kant
termed them: mere "pure concepts of the understanding," containing nothing that
cannot be fully accounted for a priori psychologically, in terms solely of the I think.
There is, to be sure, an element of necessity introduced into concepts and judgments
by Kant's "categories" insofar as they annul our freedom to reverse their logical posi
tions in judgments. Yet the astonishing thing is that this necessity is all he required to
justify according them the status of fundamental concepts of objects. For as we saw
in Chapters 13-16, when the analytic unity of apperception is attached to the mani
fold of intuition, converting it into a manifold of concepts, the concepts derived from
logical functions make possible a transcendental synthesis able to uniquely differen
tiate and completely determine each of the infinitely many spaces and times made
possible by pure space and time. The resulting necessary synthetic unity of the mani
fold of pure-formal space and time and its extension to the field of appearance in the
form of a community of causally interacting permanent substances thus count as
genuinely objective unities of apperception. Kant's use of this paradoxical-seeming
expression, combining the quintessentially subjective notion of "unity of appercep
tion" with its antithesis, "objective," was therefore justified. For without drawing on
concepts that depend in any way on actual experience, mathematical construction,
or anything else beyond (1) the representation I think, (2) its basis in pure sensibility,
and (3) its unity as a judging subject, Kant defined an objective framework that he
could credibly argue must inform all non-transcendental intuition and propositional
thought before such thought can take on objective sense and significance (Sinn und
Bedeutung, in Kant's phrase). Indeed, he could use it to take matters to their idealist
extreme by arguing that the objective unity of apperception makes possible not only
an experiencing, cognizing subject but the objects it experiences, their fundamental
laws, and nature itself as their all-encompassing commercium, as well. Thus, in what
is surely a psycho-philosophical tour de force genuinely worthy of the Guardian's
accolade of "heavyweight philosophy champion of the world," Kant developed an
entire "ontology as immanent thinking" by premising nothing more than the inde
feasibility of the Cartesian ego in its purely logical sense, stripped by critique of all
the epistemological and metaphysical doctrines added by what he termed "dogmatic
philosophy."
Kant's theory of experience remains beset by misunderstanding. It is almost
universally taken for granted that it introduces elements of Euclidean geometry and
Newtonian physics into the innate endowment of the human psyche. As a result,
Kant's attempt to render the distinction between succession and simultaneity objec
tive is often supposed to be refuted by special relativity, while his law of cause and
effect is commonly taken to be undercut by quantum uncertainty. F inally, and
Reversing the Frame 561
perhaps most important, given how fundamental the unity of nature was to Kant,
Everett's many worlds interpretation of quantum mechanics and other concep
tions of the multiverse may also be thought to expose Kant's insistence on this
unity as mistaken, or at best misguided. But once one has developed a sufficient
appreciation of Kant's a priori psychology and what demarcates it from all other
kinds of knowledge, such criticisms are easily shown to be unfounded. For it then
becomes clear, first, that transcendental synthesis, rooted entirely in the I think and
what it and the other representations of a priori psychology suffice by themselves
to validate, can neither imply nor entail any possible pure or applied mathematical
definition, axiom, postulate, theorem, or postulate, and, second, that this is just as
true of the infinity of uniquely differentiated, completely determined spaces and
times produced by transcendental synthesis speciosa of the manifold of pure-formal
intuition as it is of the infinite community of causally interactive community of
permanent substances produced by the transcendental schematism of the mani
fold of empirical-material intuition. Transcendental cognition, as Kant repeatedly
made clear, can yield none but philosophical-discursive synthetic a priori cogni
tion and so leaves the field completely open for mathematical-intuitive synthetic a
priori cognition, both pure and applied, even while furnishing the conditions-dif
ferentiated, determinate pure space and time, community as comercium-that make
the latter possible. Conversely, all possible pure and applied mathematical-intuitive
synthetic a priori cognition is preceded and made possible by transcendental syn
thesis founded on the identical set of innate aesthetic and logical forms: the latter
underwrites the former indiscriminately and in general, without distinguishing one
kind from any other (Riemannian versus Euclidean, Newtonian versus Einsteinian,
Bohrian versus Everttian, etc.). Thus, insofar as there is "only so much genuine
science as there is mathematics capable of application" (MFPNS 470), it is quite
impossible for the purely philosophical-discursive objectivity synthesized under the
aegis of Kantian a priori psychology to enter into conflict with relativity theory,
quantum mechanics, or any other properly scientific cognition of nature.
None of this is to say that Kant would not have been profoundly shocked by some
of the innovations in the natural sciences since his time. Nor is it to deny that he
believed he could close the gap between apperception-based transcendental laws of
nature and Newtonian physics with a metaphysics of nature that draws on empirical
concepts and fundamental experiences (Al71-2/B213) to provide "examples (cases
in concreto) that realize the concepts and theorems of the former ..., i.e. give mean
ing and signification (Sinn und Bedeutung) to a mere form of thought" (MFPNS
478). And here one might object to my claim that there is no conflict between Kant's
a priori psychology and the physical sciences by asking how transcendental judg
ments, mere forms of thought, can be given concrete meaning and signification in
a manner compatible with relativity and quantum theory. For example, how can
the irreversible ground-consequence relations of Kant's transcendental causal law
be mapped onto a world characterized by quantum vacuum energy, entanglement,
tunneling, and uncertainty? The obvious reply is that, from the standpoint of a priori
562 Kant's Anatomy of the Intelligent Mind
psychology, it is a feature, not a bug, that transcendental judgments have nothing what
soever to say about the world of actual experience investigated by science beyond the
role of original apperception in its constitution. The most likely reason Kant believed
he could and should construct a metaphysics of nature linking transcendental laws of
nature to Newtonian science is that so little was known about matter in his day that
there could not have seemed to be any essential difference between his transcendental
concept of substantial final subjects in space and Newton's empirical concept of mat
ter (the inclusion of empirical marks in the latter such as impenetrability and mobil
ity would not have seemed essential, given what was then known). By contrast, there
can be no mistaking the empirical community (communio) of impermanent fermions
and bosons, strings, branes, black holes, galaxy clusters, the Big Bang Universe, and
multiverses of the physics of our time with Kant's transcendental community (com
mercium) of empirically unknowable, mathematically unconstructible substances that
are permanent with respect to both earlier and later time in the sense of having "an
inner necessity of persisting" (Al85/B229). My surmise is that Kant would not have
felt any need for a metaphysics of nature to connect the commercium of transcendental
philosophy with the communio of today's physics, and would instead have been content
to treat them as complementary, non-overlapping magisteria of objective cognition.
Having identified the concept of consciousness on which Kant based a priori
psychology, I can now answer the question posed earlier regarding whether he could
admit the independence of intelligence vis a vis consciousness in the mind-brain. The
transcendental synthesis posited by Kant's cogito-centered a priori psychology leaves
empirical psychology just as inviolate as it does science generally. It has no conflict
with the supposition that intelligence can operate independently of consciousness in
every sense in which scientists of mind conceive it, be it consciousness as dreaming
sleep, wakefulness, alertness, qualia, introspective internal perception, what experi
mental subjects report when asked to describe their experiences, behavior patterns,
states of the brain detected by fMRI, and so forth. It is only specifically with regard to
the I think (pure self-consciousness qua original apperception) that the supposition of
intelligence without consciousness comes into conflict with Kantian a priori psychol
ogy, where it is tantamount to saying there could be propositional thought without
concepts (ADA-universals) or the means to logically unite them (logical functions of
judgment). A priori psychology views neural activity as intelligent not in and for itself
but only insofar as it subserves the aesthetic logical and/or objective unity of pure
self-consciousness. Since, in Kant's scheme, the latter are pure, merely formal repre
sentations, which may, but need not,8 manifest themselves empirically as inner appear
ances and states of the brain, they are best regarded not as targets of investigation
8 As remarked in Chapters 4 and 6, there is no necessity that every expression of the sponta
neity of the mind coincide with an affection of inner sense (self-affection) or that this affection
be exhibited by an appearance in inner intuition or that this appearance ever attract the kind
of attention that would raise it from an obscure to a clear perception. All of this would have to
happen before the exercises of spontaneity responsible for transcendental judgments and other
transcendental representations would become available for empirical inquiry.
Reversing the Frame 563
Allen, John S. The Lives of the Brain: Human Evolution and the Organ of Mind.
Cambridge: Harvard University Press, 2009.
Deutscher, Guy. The Unfolding of Language. New York: Henry Holt and Company, 2005.
Evans, Nicholas and Stephen Levinson. "T he Myth of Language Universals: Language
Diversity and Its Importance for Cognitive Science." Behavioral and Brain Sciences, 32
(2009): 429-48.
Friedman, Michael. Kant and the Exact Sciences. Cambridge, MA: Harvard University
Press, 1992.
Guyer, Paul. Kant and the Claims of Knowledge. Cambridge: Cambridge University
Press, 1987.
Longuenesse, Beatrice. Kant et le pouvoir de juger. Paris: Presses Universitaires de
France, 1993.
MacFarlane, John. "Frege, Kant, and the Logic in Logicism." Philosophical Review, 111,
no. 1 (2002): 25-65.
McCulloch, Alison. "Freedom and Reason: Kant's Construction of Morality." PhD diss.,
University of Colorado, Boulder, 2003.
Parrini, Paolo, ed. Kant and Contemporary Epistemology. Dordrecht: Kluwer Academic
Publishers, 1994.
Prauss, Gerald. Kant und das Problem der Dinge an sich. Bonn: Bouvier Verlag Herbert
Grundmann, 1974.
Rescher, Nicholas. The Primacy of Practice; Essays Towards a Pragmatically Kantian
Theory of Empirical Knowledge. Oxford: Blackwell, 1973.
Stabler, Eugen. George Berkeley's Auffassung und Wirkung in der Deutschen Philosophie bis
Hegel. Zeulenroda: Bernhard Sporn, 1935.
Stewart, Ian. Taming the Infinite. The Story of Mathematics. London: Quercus, 2008.
Strawson, P. F. Entity and Identity. Oxford: Clarendon Press, 1996.
Watkins, Eric. Kant and the Metaphysics of Causality. Cambridge: Cambridge University
Press, 2004.
Winkler, Kenneth. The Cambridge Companion to Berkeley. Cambridge: Cambridge
University Press, 2005.
Wittgenstein, Ludwig. The Philosophical Investigations. Translated by G. E. M. Anscombe.
Oxford: Blackwell Publishers Ltd., 1953.
This page intentionally left blank
{INDEX}
geometry,161-70,455 space,92,112n l 0
apriority of,164-6 subject,224-5
empirical antecedents of,163 theater of the mind,490-2,495
Euclidean,146,161-2,168,454-5,548 time,92
extended to motion and pure mechanics,169 uniformity of nature,26,30,342,487
Kant not Euclidean dogmatist,17,63-4, whole of space and time,342,483-9
117-8,158,162-3,445,546-7
Kantian space can accommodate infinite idealism,20,33
spatial dimensions,425-6 conceptual not epistemic notion,193-4
Kantian space can accommodate non distinction between appearance and sensation
Euclidean,160-3,445 essential to Kant's,187-8,194,213-5
least abstract form of mathematics,163-4 dogmatic,20 l n l
CJod,61,192,222,239n2,301-2,342 empirical,187-8,190-1,342
gravity,86,408,454,460,516-7,519-20,538, esse is percipi (Berkeleyan),201-8
540 formal contrasted with material,214-5,
CJuyer,Paul,338n10 222
how far empirical and transcendental agree,
Hamann,Johann,68n21 190-1
Hegel,CJeorg,149nl l, 193,223,254 Hume's transformed conception of,185
Henrich, Dieter,32 l n l and illusion,189-95
heterogeneity problem,6 153-4,308-9,311,339, Kant's transformation of Hume's conception,
403-5,411,446 185-6
holism,47-9 problematic (Cartesian),196-7,201,537nl l
Hume,David,vii,3,4n,17,25-7,30,156n l , psychologistic,64-5,198-200. See also Hume,
216n l 0,224n l 5,230nl 8,253,455n l ,553 psychologism.
associationist psychology of,36,62-3, l 12n10, pure intuition essential to Kant's, 185-6,
342-3,349,351-2,358-9,361,483-8, 190-3
504-5,555 regarding the material world,182-9,196-200,
causal succession,511 211-2
continuants,485-6,507-9 regarding the sensible world,182-9
custom substituted for necessary connection, transcendental,61,168-9,186-8,190-3,
56-7,485,487,504-6 196-8,212-3,342,368
customary associations extends purview why unavoidable,193-4
beyond sense and memory,485-6 identity
general causal maxim,30,56-7,68,342,487, subjective vs. objective conceptual,43-6,49,
506 57,59,415
importance of treatment of cause and effect illuminationism,37,194
for Kant,35,55,57-8 illusion,122n l 7,189-95
inconceivability of necessary relations between imagination. See also apprehension,intuition,
the distinct,55-6 reproduction,synthesis.
on language and universals,64n l 6,239-40, aesthetic synthesis of,41,142,174-5,310
255-6,352,544,549 apprehending (faculty of intuition),101-5,
psychologistic idealism of,184-5,192,488, 113-4,123
490 apprehending necessary for perception,102-4,
psychologism of,36,61-6,255-6,488 120,209-10
quandary concerning personal identity,64, blind until concepts supplied by
129n24,186,213,225,487-94,504-5 understanding,310,350-2,356,358,
quandary concerning personal identity holds 370
of whole of space and time,487-9 empirical,164
on relations,351-2 figurative synthesis of,372. See also synthesis
skeptical challenge posed by,36,56-7,59, speciosa.
255-8,303,309-10,312-3,326,335-6, objectivity of categories bound up with,28.
455-6,544 See also metaphysical deduction,synthesis
source of problem addressed in Critique speciosa, schematism
of Pure Reason, 57-60,67-8,73,256-7, productive,38-9,49n8,113-4,166
506n l 4,557 reproductive,31,347-50,533
572 Index