Sei sulla pagina 1di 9

North American Philosophical Publications

Kant on Apperception and "A Priori" Synthesis


Author(s): Paul Guyer
Source: American Philosophical Quarterly, Vol. 17, No. 3 (Jul., 1980), pp. 205-212
Published by: University of Illinois Press on behalf of the North American Philosophical Publications
Stable URL: http://www.jstor.org/stable/20013865 .
Accessed: 23/07/2013 13:36

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

University of Illinois Press and North American Philosophical Publications are collaborating with JSTOR to
digitize, preserve and extend access to American Philosophical Quarterly.

http://www.jstor.org

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
American Philosophical Quarterly
Volume 17, Number 3, July 1980

IV. KANT ON APPERCEPTION AND A PRIORI


SYNTHESIS
PAUL GUYER

I objective affinity on nature, for the underlying


are we to make of Kant's theory of motivation of the former's postulation of a single act
WHAT synthesis? One popular approach has been of imagination antecedent to all particular experience
simply to reject it all as "an essay in the imaginary and the latter's suggestion of an ongoing constitution
subject of transcendental psychology," having no of the objects of nature is the same, and the distinction
necessary connection to any of the Critique of Pure between them, given Kant's doctrine of the transcen?
Reason's true epistemological objectives.1 It seems dental ideality of time, is as little significant as that
more however, to grant that certain theses between occasionalism and the of
temperate, theory pre-estab?
about mental processes, though by subject matter in lished harmony. Second, both the bowdlerization of
the realm of psychology, may nevertheless be proper Kant's theory of a priori synthesis as a harmless
consequences of Kant's epistemology, insofar as
expression and the false restriction of it tomathematics
"every science has a conceptual apparatus of necessary obscure a real problem in Kant's deduction of the
truths" accessible even to the philosopher without applicability of the categories to any possible repre?
On this view, Kant's sentation, which must be before any recon?
experimentation.2 postulation exposed
of various processes of empirical synthesis is an struction of Kant's theory of empirical knowledge
acceptable correlate of his analysis of the principles can succeed. For Kant's theory of a priori synthesis is
of human knowledge, and two stratagems may be no accident, but the consequence of a mistake in.his

employed to avoid any problem with regard to the a conception of the very keystone of the deduction, the
priori or transcendental synthesis to which the first transcendental unity of apperception itself. The
approach really objects. First, references to a priori theory of the a priori synthesis or constitution of nature
synthesis can be interpreted as obscure but harmless arises from the fact that Kant does not begin his
expressions of the theory that empirical syntheses are argument from an
analytic connection between the
conducted according
to certain a
priori constraints ; concept of the self-ascription of experiences and the
or, second, it can be conceded that Kant does postulate existence of synthetic unity among so
experiences
a priori syntheses which are actually independent of ascribed, but instead commits himself to the a priori
but as a correlate of his of a thoroughly connection be?
ordinary experience, only certainty synthetic
philosophy of mathematics,3 which may readily be tween consciousness and the self-ascription of expe?
separated from his basic theory of knowledge. rience, or
^^-consciousness. Unless we realize that
But while these stratagems may appear to rescue the theory of a priori synthesis of empirical reality
what is otherwise respectable in the theory of cannot be neatly removed from Kant's philosophy
synthesis, they are in fact doubly misguided. First, with the excision of his philosophy of mathematics,
they are false to the text, for Kant postulates not but can be avoided only be avoiding an error about
merely empirical syntheses in accord with a priori transcendental apperception itself, we shall have no

rules, but also an a or transcendental clear of what we can assume


priori synthesis conception properly
of all manifolds of intuition, empirical as well as pure about apperception, and thus no
prospect for a
or mathematical.4 This is evident not only in Kant's successful reconstruction of Kant's deduction of the a
detailed discussion of threefold synthesis but also in priori constraints on empirical knowledge.
his "Copernican" insistence that the mind imposes I will not actually provide such a reconstruction
1P. F.
Strawson, The Bounds of Sense (London, 1966), p. 32.
2Graham
Bird, "Logik und Psychologie in der Transzendentalen Deduktion," Kant-Studien 56 (1966), p. 382.
3
This is the solution suggested by Bird, ibid., p. 375.
4
H. J. Paton's interpretation of trancendental synthesis as a synthesis employing the categories but exercized only on the pure
manifolds of the a priori intuitions of space and time is thus false to Kant's text as well as incoherent. See Kant's Metaphysic of Experience
(London, 1936), vol. I, p. 474.

205

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
206 AMERICAN PHILOSOPHICAL QUARTERLY

here. Rather, I will be chiefly concerned to demon? is clearly committed to the existence of a creative
strate only the following points: first, that Kant is synthesis imposing order on the manifold of empirical
committed to the existence of an a priori synthesis or intuition, whether it be conceived of as a single act of
constitution of objects in all cases of knowledge, not transcendental imagination preceding all empirical
just inmathematics; second, that this theory ofa priori synthesis or as an ongoing activity of constitution
synthesis is a consequence of the certainty which underlying the objective affinity of the objects of
Kant ascribes to the transcendental unity of apper? nature. Evidence for this claim is not hard to find. It
ception; but, third, that what Kant assumes about may be found in Kant's assertion that there is a
apperception can only be construed as a synthetic "synthesis of reproduction in imagination" which
and empirical rather than analytic a priori proposi? must be counted among the "transcendental actions
tion; and, finally, that this proposition about apper? of the mind" and which is the "0 priori ground of a
ception is not justified by Kant, but rests on a mere necessary synthetic unity of appearances," for "this
conflation of the concepts of consciousness and self synthesis of imagination," he holds, "is grounded
consciousness. Before concluding, however, I will prior to all experience on an a priori principle, and
suggest that while one of Kant's strategies for the one must assume a pure transcendental synthesis of
deduction of the categories is doomed by his confusion imagination which grounds even the possibility of
about the principle of apperception, another is not. experience" (A 101-102). And if this is not clear
enough evidence of Kant's postulation of a pure
synthesis independent of, and in some sense prior to,
II
any empirical synthesis, then perhaps such evidence
Some of Kant's references to a
priori
or transcen? can be seen in these statements from the "systematic"
dental synthesis do commit him only to the a priori (A 115) exposition of the deduction in the first edition
synthesis of the pure intuitions of mathematics, or else of the Critique:
to the not patently implausible theory that empirical ...
pure a of the synthetic
apperception supplies principle
manifolds are synthesized, in the ordinary course of in all possible intuition.
unity of themanifold
experience, in accord with certain a priori constraints. This a synthesis, or
synthetic unity presupposes
A clear example of the first sort of reference is this : includes one, and if the former is to be necessary a
priori,
the "synthesis of apprehension must also be exercised then so must the latter be an a priori synthesis. ... But
a priori, that is, in respect of representations which are only the productive synthesis of the imagination can take
not empirical. For without it we should never have a a priori', the rests on
place reproductive empirical
the either of space or of time" conditions. Thus the principle of the necessary unity of
priori representations
the imagination to apperception is the ground of
prior
(A 99-100).5 Examples which suggest that synthesis the possibility of all cognition, of experience
must have a priori rules, but not that itmust actually particularly

occur a priori, may be found in Kant's claim that "all (A 117-118).

appearances, insofar as
objects
should be given
to us And in the second edition of the deduction, from
through them, must stand under a priori rules of their which somuch of the detailed theory of synthesis has
synthetic unity, according to which alone their been removed, the assertion of transcendental synthe?
relation in empirical intuition is possible" (A no), sis prior to any particular experience is still allowed
and in his assertion that since the identity of to stand : "Synthetic unity of the manifold, as given6
apperception "must necessarily come into the synthe? a priori, is therefore the ground of the identity of
sis of all the manifold of appearance, so far as this apperception itself, which precedes a priori all my
synthesis should become knowledge, the appearances determinate thought" (B 134).
are subjected to a priori conditions, with which their Even if this evidence for Kant's postulation of a
... must thoroughly accord" (A 113). single act of transcendental synthesis antecedent to
synthesis
But not all of Kant's references to a priori or all particular experiences is dismissed, surely the
transcendental synthesis can be so interpreted, for he evidence of his commitment to a transcendental

5All references to the


Critique ofPure Reason are made by reference to the pagination of the first (A) and second (B) editions. (This paper
cites no passages included in both editions.) The translations are often my own, though all are made with the translation by Norman
Kemp Smith [Immanuel Kant's Critique of Pure Reason, (Second Edition) with corrections, (London, 1933)] ?Pcn before me.
6Hans this to "generated" to an a priori
Vaihinger proposed emending (hervorgebracht), which would have made Kant's commitment
act of synthesis even clearer [see the edition of the Kritik der reinen Vernunft, Raymund Schmidt., ed., (Hamburg, 1930), p. 144b]. This point
has also been discussed by Malte Hossenfelder, Kants Konstitutionstheorie und die Transzendentale Deduktion (Berlin and New York, 1978),
P. na

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
KANT ON APPERCEPTION AND A PRIORI SYNTHESIS 207

constitution of the objective affinity of nature as presupposes a synthetic unity among them (cf. B 133
underlying all particular empirical investigation i34n). Kant's commitment to the first premise is
cannot be. Kant's theory is not merely that nature made particularly clear at B 130, where he maintains
must demonstrate a certain
unity if we are to succeed "that we cannot represent to ourselves anything
as
in comprehending it in empirical syntheses con? connected in the object without having previously
strained by certain rules, but rather that "the order combined it."8 His statements of the second premise
and regularity
... in appearances, which we entitle are less obvious, but it seems to be implied in both
nature, we ourselves introduce." And this is because versions of the deduction : in the first, by the claim
"this unity of nature must be a necessary, that is, a that "the original and necessary consciousness of the
priori certain unity of the connection of appearances. identity of the self is at the same time a consciousness
But how should we be able to get such a synthetic of an equally necessary unity of the synthesis of all
unity started a priori, if subjective grounds of such appearances according to concepts" (A 108); in the
unity were not contained in the original sources of second, perhaps more distinctly, by the assertion that
cognition in our minds [?]"(A 125). Or as Kant "the thoroughgoing identity of apperception in a
bluntly concludes the first edition deduction, "con? manifold given in intuition contains a synthesis of
nection and unity (in the representation of an object) representations, and is possible only through the
are to be found solely in ourselves, and therefore consciousness of this ... This
synthesis. relation,

precede all experience" (A 130). therefore, does not occur because I accompany each
with consciousness, but rather because
representation
I conjoin each to the others and am conscious of the

Ill synthesis of them" (B 133). From these two premises


it follows that apperception with regard to amanifold
Whatever the details of the machinery, then, it is of representations, or consciousness of their ascription
clear that Kant does not argue just that objects must to a continuing, identical self, requires a synthesis of
conform to certain constraints if experience is to be these representations in accord with some concepts
possible, but that by a productive power of the mind distinct from the mere concept of the self.
we actually impose certain rules on the objects of The justification of the second premise and the
or that our has a connection of it to a particular set of are
experience, experience necessarily categories
unity which can only be explained by the productive the key to any reconstruction of Kant's deduction.
unity of the mind. We should not just reject this claim, My present concern, however, is only to ask why
but must ask first why Kant insists upon it. Kant should take this argument to imply that there
Kant's insistence upon the mind's imposition of must be an a priori synthesis of the manifold, or an
order on nature is derived from his conception of the actual imposition of order on nature? Why should his
transcendental unity of apperception. His argument conclusion not be merely that since apperception
on at least these two : first, that all a to certain constraints,
depends premises requires synthesis according
consciousness of combination in the manifold of consciousness of self-identity will occur only when
intuition requires an act of synthesis by the mind ; the objects of our experience are, independently of us,
and, second, that apperception, or the consciousness such as to allow the successful synthesis of their
that various thoughts belong to the same continuing representations according to these rules?

self,7 requires a consciousness of a synthetic unity


or The answer to this question is that Kant takes the
combination of these thoughts by means of some representation of apperception itself to be a priori, or
concepts other than the bare concept of the self maintains that since self-consciousness is a "transcen?
itself?that the mere "analytic unity" of predicating dental representation," "numerical identity is insep?
"I think" of various thoughts as a "conceptus communis" arable [from it], and a priori certain" (A 113). If we
7The senses in which Kant uses the term "apperception" are multiple, : the
but clearly include at least three distinct meanings
(necessary) representation of the numerical identity of a self in its different states (e.g., A 107, A 117-118, A 122) ; the power of the mind
which produces this representation or the conditions for its occurrence (e.g., A 108, A 112, A 115); and, possibly, the bare thought or
expression "I think" (e.g., B 133, end). But since it isKant's view that the mere representation "I think" has no significance apart from
an actual synthetic unity of mental states, we can ignore it as a sense of "apperception" ; and since there are alternative terms available
for the power which produces apperception in its first sense?such as "transcendental imagination"?I feel justified in using
"apperception" in a sense corresponding most closely to the first of the three here distinguished :as a representation (or judgment) of the
ascription of a given mental state or states to a continuing self.
8Kant's commitment to this premise is also made clear in several reflections, e.g. R 6359, Kant's gesammelte Schriften (Akademie edition),
vol. XVIII, p. 687 ;R 6360, Akademie ed. vol. XVIII, p. 689; R 6362, Akademie ed. vol. XVIII, p. 691.

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
2o8 AMERICAN PHILOSOPHICAL QUARTERLY

can be certain a
priori that we can represent our self-ascribing them, does it follow that one can know
continuing identity in any empirical manifold, then in advance of any particular experiences that one

we must be certain a priori that we can synthesize it in will be capable of synthesizing them in accord with
accordance with the rules that are the conditions of the conditions of apperception. And only if the
our consciousness of apperception. This will be certainty of apperception is so interpreted must the
possible only if our attempts at empirical synthesis, advance guarantee for the possibility of empirical
the success of which must otherwise wait upon synthesis represented by a priori syntheses or the
experience,
are
preceded by
an act of transcendental transcendental constitution of nature be invoked. In
synthesis, or if we otherwise actively impose order on other words: the purely hypothetical and analytic
nature. Thus the key step of Kant's argument lies in assertion that if the ascription of a representation to
his claim that "the mind could not possibly think its an identical self is possible, then it must also be
own identity in the manifold of its representations, possible
to connect that representation
to others

and indeed think it a priori, if it did not have before its ascribed to that self by a certain kind of synthesis, as
eyes the identity of its act, which subjects all synthesis the condition of self-ascription, implies nothing about
of ... to a transcendental the actual occurrence of the
apprehension unity" (A any syntheses; only
108), or in this sequence of claims: thoroughly existential and synthetic claim that all of
one's can be ascribed to oneself
We are conscious a priori of the thoroughgoing of representations
identity
which can ever requires the advance postulation of any form of
the self in respect of all representations,
to our knowledge ... these can synthesis or a theory of imposition.
belong ([but]) represent
in me if they belong with all others to That Kant is committed to the interpretation of
something only
one consciousness, and must therefore at least be the principle of apperception as a synthetic rather
capable
of being connected (A 116-117). This synthetic unity than analytic certainty is especially obvious in his
theory of affinity, where he simply denies that it can
a and if the former is to be
presupposes synthesis,
a priori, then so must the latter be an a
necessary priori be "quite contingent that appearances chance to fit
synthesis (A 118). into a nexus of human knowledge." That is, Kant
as Kant asserts in the second to say that asserts the impossibility of what I have claimed to be
Or, edition,
"I am conscious of identical self in respect of the compatible with a strictly analytic interpretation of
manifold of representations given to me in an the certainty of apperception?"that though we had
intuition ... is to say as much as that I am conscious the power of associating perceptions, it would yet
to myself a priori of a necessary synthesis of them, remain in itself quite undetermined and contingent
whether were in fact associable ; and in case
which is called the original synthetic unity of they they
were not, that there might exist a multitude of
apperception" (B 135).
perceptions, and indeed an entire sensibility, inwhich
much consciousness would be encountered
IV empirical
inmy mind, but separated, and without belonging to
But does the a priori certainty of the numerical a consciousness of myself"?precisely because he

identity of the self actually imply the suspicious takes the principle of apperception to entail that "all
doctrine of a priori synthesis? It does so only on a appearances must so enter the mind or be appre?
certain interpretation. If the a priori certainty of self hended that they accord with the unity of appercep?
identity is understood only as the conceptual truth tion" (A 121-122).
that whatever one ascribes to oneself But isn't the interpretation of Kant's principle of
representations
must be ascribed to the same continuing set of apperception as synthetic precluded by his own
to which belong all other represen? repeated assertion that the "principle of the necessary
representations
tations ascribed to oneself, in accordance with the unity of apperception is itself indeed an identical,
rules for constructing such sets, it would not imply therefore analytical proposition" (B 135; cf. B 138)?
any a priori synthesis. For one could wait upon No, for his assertion is false. Even on a generous
the principle which Kant
experience and its empirical synthesis to determine conception of analyticity,
what are ascribed to oneself, and thus both relies on and characterizes as
representations actually analytic
to determine how far the identity of one's appercep? is, in fact, synthetic. By allowing Kant a generous
tion actually extends. Only if the content of the a conception of analyticity, I mean that I am not
priori principle of apperception is interpreted as the denying the analyticity of his inference from the self
very different proposition that no matter what ascription of representations "to that condition under
alone I can ascribe them to the identical self as
representations one has, one will always be capable of which

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
KANT ON APPERCEPTION AND A PRIORI SYNTHESIS 20Q

my representations" (B 138).9 Even if one insists that all particular experience), namely the consciousness
the conditions for its occurrence are not "contained of myself, as original apperception" ;he also says that
in the concept" of apperception, this proposition still "the synthetic proposition, that all diverse empirical
has the ring of self-evidence. My objection is rather consciousness must be connected in a
single self

that Kant does not confine himself to the inference consciousness, is the absolutely first and synthetic
as I
from the self-ascribability of particular representa? principle of our thought in general." And as far
tions to the satisfaction of the conditions for their self can see, there is simply no difference between what is
com? here said to be synthetic and the claim that all of my
ascription by those representations, but instead
mits himself to the quite distinct, synthetic proposition representations
must be self-ascribable, which the

that whatever is to count as a


representation
at all must second edition calls analytic.
be fit for self-ascription.
This is not evident in the first formulation of the V
principle of apperception which Kant calls analytic,
the statement at B 134 that "the thought [that] these Kant does not attempt to reconcile his two
one and all belong characterizations of the principle of apperception,
representations given in intuition
to me accordingly means the same as [the thought] but instead simply suppresses the first edition's
that I unite them in one self-consciousness, or at least admission that it is synthetic in favor of the second
can so unite them." This can be taken to mean merely edition's claim of analyticity. And in a way, this is
that the thought that various representations,
sev?
appropriate. For even in the first edition, Kant does
not put forth his principle as brute fact. Rather, he
erally, belong tome is equivalent to the thought that
constitute the thoughts of a single attempts to support it by a manipulation of the
they, collectively,
self, and to rest on nothing but the concept of a concept of consciousness itself. He does not, however,
continuing self. But this is not the principle from succeed in deriving his principle by a strictly
which Kant deduces the applicability of the categories analytical procedure, but instead attains it only by
to all possible representations. This principle emerges conflating the concepts of consciousness and self

only in what Kant calls analytic at B 138, the consciousness.

proposition that "all my representations in any given Kant's argument might be thought to require two
intuition must be subject to that condition under steps :first, the demonstration that all representations
which alone I can ascribe them to the identical self as involve consciousness, and second, the proof that all

This, however, does not state the consciousness involves self-consciousness. The first
my representations."
analytical truth that whatever representations I can step is easy enough, for representations just
are

ascribe to
myself
as my own are
subject
to whatever modifications of consciousness. This is asserted several
conditions govern such ascription.
It asserts the very times : thus, preceding his assertion that all empirical
different claim that I cannot have a representation consciousness is related to original apperception,
which is not subject to these conditions. To put it Kant first reminds us that to
deny that "all represen?

Kant asserts that I cannot have a represen? tations have a necessary relation to a
bluntly, possible empirical
tation which I cannot as my own. This is consciousness... would be to say as much as that
recognize
certainly a synthetic claim, and it is only to explain they do not exist at all" (A i i7n; cf. A 116) ; and in
how this proposition can be known to be true a priori introducing his discussion of affinity, he maintains
that Kant must invoke a of transcendental that from its relation to an at least
theory "apart possible
consciousness, would never be an
synthesis. appearance object
In one passage which has not received the press of knowledge for us, and would therefore be nothing
accorded to his assertions of analyticity, Kant does for us, and since appearance has in itself no objective
appear to admit precisely what I have just argued. reality, and exists only in cognition, it would be
This passage is the footnote on A117, and occurs right nothing at all" (A 120). And thatrepresentationsjust
in the midst of the first edition's systematic exposition are
impingements
on consciousness, and thus cannot

of the deduction. In it,Kant not only asserts his claim exist except as states of consciousness, is incontesta?

that "all empirical


consciousness has a necessary ble.10 But Kant's attempt to use this fact to prove the
relation to a transcendental consciousness (preceding second step, that there can be no perceptions which
9This move
sort of inference, of course, is a typically Kantian ; it is central, for instance, in the Second Edition's Refutation of Idealism
(see B 276).
10One in the Anthropology from a Pragmatic Point of View, where Kant considers "obscure" ideas (?5, Akademie ed. vol. VII, pp.
passage
may appear to cast doubt on Kant's subscription to this principle. But it does not, for it really makes only the two points that
135-137),

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
210 AMERICAN PHILOSOPHICAL QUARTERLY

cannot fit into the "nexus of human knowledge" (A and of consciousness. It does
third-person ascriptions
121), is not. For he derives the impossibility of seem indisputable that whichever of one's own states
consciousness without self-consciousness thus, one can as states of consciousness must also
(and, recognize
synthetic unity), solely from this further claim: "it is involve self-consciousness, for to
recognize one's state

only because I ascribe all perceptions to one conscious? as conscious (or anything else, for that matter) first
ness (of original apperception) that I can say of all it as one's own state. But from
requires recognizing
perceptions that I am conscious of them" (A 122). But this it does not follow that one could not have a state
this, while perhaps true, proves Kant's point only if of consciousness without being able to ascribe it to
it is already assumed that I can say I am conscious of oneself, or without
self-consciousness, for there might
whatever states I am in fact conscious. But this is just be cases inwhich one is conscious without recognizing
to assume that consciousness is self-consciousness, or it. There might be cases, in other words, in which
that I must be capable of self-ascription whenever I other persons could ascribe consciousness to one
truly
am
capable of consciousness at all?the very propo? without one being able to so oneself.11 For :
example
sition that Kant must prove. In other words, Kant if we dreams as modifications of consciousness,
regard
has no argument at all for the thesis that all but also regard the occurrence of rapid eye movements
consciousness is self-consciousness, but only
a confla?
(REMs) as good empirical evidence for the simulta?
tion of the two neous occurrence of then we
concepts. dreaming, may regard
Perhaps Kant's error stems from tacitly thinking our own observation of REMs in another as evidence
of consciousness only from a first-person point of that the other is conscious at a
given moment, even

view; at any rate, the error I am attributing to him though it later turns out that the other person himself
can be made obvious by distinguishing between first has no memory of his dream, and thus cannot ascribe

(i) in the case of Leibnizian obscure ideas we may not be immediately aware of all the features of the objects we refer those ideas to and
(ii) there may be sensations in us of which we are not conscious at a given time, though we may become so. Neither of these points amounts
to an assertion that we may have ideas without consciousness tout court.
11
(a)My use of the distinction between first- and third-person ascriptions of consciousness to explain the confusion underlying Kant's
assertion that we are certain a priori of our continuing identity may be objected to on the ground that Kant himself introduces this very
distinction in the third Paralogism of the first edition of the Critique. There, Kant writes that
In my own consciousness, therefore, identity of person is unfailingly met with. But if I view myself from the standpoint of another
person (as object of his outer intuition), it is this outer observer who first represents me in time, for in the apperception time is represented,
strictly speaking, only in me. Although he admits, therefore, the "I" which accompanies, and indeed with complete identity, all
representations at all times in my consciousness, he will draw no inference from this to the objective permanence of myself (A 362-363).
If this ismeant that while it is analytically true that anything which / call a state of myself is ascribed to the same continuing subject as
any other state so ascribed, it is at most synthetically true and knowable only a posteriori that whatever another calls a state of me must be
ascribed to such a subject, it would make the very point which I accuse Kant of failing to notice in his Transcendental Deduction. But
(i) Kant's notice of this point in the Paralogisms would, in any case, not be incompatible with his failure to notice it in the Deduction, and
(ii) the thesis that I have just expressed is not, I think, the point of A 362-363. Rather, this passage ismeant to argue only that a subjective
awareness of identity or continuing consciousness is not in itself direct and immediate evidence for the continuous and co-terminous
existence of a self-identical substratum for that consciousness, such as a body which might be observed by another. And this point neither
a
requires nor entails a retraction of what I have argued to be the unjustified claim that we can never be conscious of mental state at all
without also ascribing it to a continuing self.
(b) In his article on "The Proof-Structure of Kant's Transcendental Deduction" (The Review ofMetaphysics 22 (1969), pp. 640-659),
Dieter Henrich attempts to make a point like that for which I have just argued, but does not make it as clearly as I hope I have. His
suggestion is that Kant's commitment to "the extensive thesis that all*representations which arise in the sensibility of a being are to that
also already potentially conscious representations" is connected to an equivocation or "shift of meaning in the expression 'mine' "(p.
degree
654). This is less than completely clear for two reasons, (i)While I believe that Henrich does mean by "sensibility" and "conscious" much
what I mean by "consciousness" and "self-conscious", respectively, his use of "conscious" to mean "self-conscious" obscures how Kant
actually conflates consciousness and self-consciousness, (ii) While Henrich does not make fully explicit exactly what the two senses of
"mine" are which Kant is supposed to confuse, it seems that they are supposed to be, on the one hand, whatever is "only available to be
taken up into [my] consciousness" or merely "in relation to me," and, on the other hand, what I actually recognize to belong to me or to
be my own state of mind. But I do not find it obvious that there are any such two senses of "mine". While there may be certain
circumstances in which Imay use "mine" indeterminately, to refer to whatever belongs or pertains to me without specifying particulars,
it seems generally true of "mine" not only that itmeans something like "belonging to or pertaining to the speaker" but also that it can only
be usedhy the speaker to mean this, and thus that it cannot be used to refer to anything that pertains to a speaker, such as any state of which
he merely could be conscious. So it can be used to refer only to what a speaker recognizes as his own. But to say this is precisely to say that
a first-person expression like "mine" is contextually opaque in a way that a third-person expression like "his" or "hers" is not, or that
though when another uses "his" in reference to me, "his" and "mine" may have the same intension?what belongs to me?their
extensions may not be the same, for the other may recognize me to be in a state which I do not or cannot recognize as mine. Thus, what
Kant fails to see is not that "mine" is ambiguous, but that "his" and "mine" are not always co-extensional, or first- and third-person
ascriptions of consciousness always congruent.

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
KANT ON APPERCEPTION AND A PRIORI SYNTHESIS 211

consciousness to himself with respect to the moment valid judgments by which one asserts the continuing
in question?or accompany what we know to have of one's own consciousness and ascribes
identity
been his consciousness with self-consciousness. determinate positions to particular thoughts or
representations within this consciousness (cf. B 139
VI
142). That is, apperception does not consist merely of
These remarks need not disprove that we are self protocols that various representations
seem to
belong
conscious whenever we are conscious. I need argue together, nor does it consist of judgments directly
only that there is a burden of proof on whoever would about the objects of those representations. Instead, it
assert that proposition, which burden Kant does not consists of considered judgments about the continuing
sustain. Thus, the premise of his theory of transcen? identity of one's consciousness and the particular
dental synthesis or the subjective constitution of the positions of its thoughts in an objectively or publicly
affinity of nature is unfounded, and that "transcen? valid order of time or events. Apperception consists
dental psychology" may be removed from "what is of genuine knowledge-claims about one's identity. As
living" in Kant's philosophy. But if the a priori Kant characterizes them, these claims are assertions

certainty of the principle of apperception is denied, that one's representations "belong to one another in
what remains viable in Kant's program for the virtue of the necessary unity of apperception in the
deduction of the categories? synthesis of intuitions, that is, according to principles
To answer this question,
we must
recognize that of the objective determination of all representations,
Kant suggested (at least) two quite distinct strategies in so far as knowledge can be acquired by means of
for this deduction, and that only one of them is these representations" (B 142). The deduction of the
undermined by the argument I have presented. This categories would then proceed by the demonstration
is the strategy on which it is actually the primary that the contents of apperception, these objectively
premise of the deduction that apperception is com? valid judgments about oneself, are possible, or,
prised by "Cartesian evidence" for or a priori certainty epistemologically, can be justified, only if certain
of our continuing self-identity, an identity presum? categories, or certain principles employing those
ably meant to extend to any situation in which we categories, are judged to obtain of the objects of the
have consciousness at all. With the additional premise representations with respect to which apperception is
that we have no a priori knowledge of the particular asserted. The force of this argument is not that the
contents of as yet unexperienced thoughts, such a psychological phenomenon of consciousness, or per?
deduction then concludes that we must have a priori haps even a sense of self, can occur only if the
knowledge of the rules by which individual thoughts, categories apply to objects, a position which would
whatever their contents, may be connected.12 These require some "transcendental psychology" ; rather, it
rules, of course, are the categories. But the a
priori
is that objectively valid judgments about one's identity
certainty of apperception is not possible without a and the position of one's thoughts within it can be
priori syntheses, and without such a priori certainty an justified only if certain categorial principles apply to
argument of this form cannot get off the ground. the objects of one's thoughts.13
However, the exclusion of apriori synthesis from his Kant provided materials for or sketches of this
argument leaves room for Kant's alternative concep? argument in a number of places?not in such
just
tion of apperception, and for his alternative strategy obvious ones as the first edition deduction's discussion
for the deduction of synthetic a priori principles of the of three-fold synthesis (A 99-107), the Analogies of
understanding. On this alternative account, apper? Experience, and the second edition's Refutation of
ception is not comprised of any a priori insight into Idealism (B 275-279), but also in a number of private
one's own continuing identity, but is instead consti? notes from periods both antedating and postdating
tuted by the collection of empirical yet objectively the two editions of the Critique of Pure Reason itself.14
12An of this form has been held to be the key to Kant's deduction
argument by Dieter Henrich in Identit?t und Objektivit?t: Eine
Untersuchung ?berKants transzendentale Deduktion (Heidelberg :Karl Winter Universit?tsverlag, 1976), pp. 86-92 ; the expression "Cartesian
evidence" is introduced on p. 59, and "Cartesian certainty" is used on p. 87. For further discussion of this book, see my review in The
Journal of Philosophy, vol. 76 ( 1979), pp. 151 -167.
13For further discussion of
the Kantian idea of objectively valid judgments about subjective states, see Lewis White Beck's article, "Did
the Sage of K?nigsberg Have no Dreams?" now published in English in his collection Essays onKant andHume (New Haven and London,
i978))PP-38-6o
14 I some of the reflections
have in mind particularly included in the so-called Duisburg Nachlass of about 1775 (R 4674-4684, Akademie
ed. vol. XVII, pp. 643-671), and another group of reflections from 1790 in which Kant continues his efforts at a refutation of idealism
(R 6311-6316, Akademie ed. vol. XVIII, pp. 607-623).

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions
212 AMERICAN PHILOSOPHICAL QUARTERLY

The detailed discussion of this argument is beyond concept of apperception. The conclusion of Kant's
the compass of the present, chiefly propaedeutic argument, however, that judgments of apperception
paper. But a hint of Kant's intentions can be given. are possible only if categorial principles apply to our
The argument begins with a conception of appercep? objects of thought, would result from neither a mere
tion as "the consciousness of thinking, that is, of analysis of self-consciousness nor an inductive inves?

representations, as they are placed in [one's] mind [so tigation of actual occurrences of self-consciousness.
wie sie im Gem?the gesetzt werden]." Such apperception Rather, it would derive from the fundamental
or positioning of one's thoughts requires three kinds requirements for making judgments of temporal
of temporal determination : a determination of the duration and position, and could thus be regarded as
temporal magnitude of the continuing subject to the source of synthetic a priori principles. The
which particular thoughts are assigned; determina? prospects for a Kantian deduction of the categories,
tions of the temporal positions in which particular therefore, need not depend on a theory of a priori
successions of thoughts occur; and determinations
apperception and a priori synthesis. That self-con?
that various thoughts may be regarded as held sciousness obtains with respect to any given intuitions
simultaneously.15 In Kant's various versions, the
might be known only a posteriori, and thus that the
argument is then continued by the introduction of categories apply to any particular objects may also be
such premises as that time itself cannot be perceived contingent. But that self-consciousness is possible only
(B 219) or that time does not itself have, as the mere when the categories do apply to the external objects
form of inner sense, a sufficient number of dimensions of such consciousness might still be a synthetic
to allow for all of these different determinations of proposition known a priori. And this would mean that
temporal relations.16 And by such lemmata, the the objective validity of the categories would be as
argument concludes that the objectively valid deter? well-entrenched as claims of self-consciousness them?
minations required by apperception can be made or selves?claims which, while contingent and aposteriori,
justified only if the objects of the thoughts positioned may still, as a class, be beyond the reach of any serious
in such apperception are regarded as enduring skepticism.
external substances with states in relations of both So the implausibility of apriori synthesis undermines
causal succession and simultaneous interaction. But the idea of apperception as an a priori certainty of
this is just to say that apperception itself is possible one's continuing identity which could directly entail
only if the of thought are subjected to what the a priori certainty of categories for its objects. But
objects
Kant always regarded as the chief among his it leaves ample room for a conception of apperception
as an valid consciousness of one's own self
categories.17 objectively
Obviously, this deduction would work only with a identity, which comprises assertions of the temporal
persuasive demonstration that judgments of self duration, succession, and simultaneity of thoughts,
consciousness do
require
correlative judgments about and which, because of the conditions of the possibility
the states of external objects, and such a demonstration of such assertions, is itself possible only when certain
has not been provided here. But supposing it could be categories apply to objects. Seen in this light, the
supplied, we might then describe the results of Kant's notion of apperception cannot prove that no object of
as follows. Particular assertions of consciousness could ever fail to conform to the
argument apper?
but can
ception or self-identity would not be known a priori, categories, still provide insight into
as they could be only if the theory of a priori synthesis the fundamental conditions of the possibility
were true, but would instead be synthetic truths of self-consciousness, and thus of self-conscious

known a
dependent
on actual expe? claims to knowledge in general. Freed of the
posteriori?truths
rience. And Kant's three claims about the temporal encumbrance of a priori synthesis, the concept of
structure of self-consciousness could be known a priori, apperception may still be of profound importance
for epistemology.18
perhaps, but only because they merely analyze the

Received
University of Illinois January 26, igjg
at Chicago Circle

15R
4674, Akademie ed. vol. XVII, p 647.
16R
6315, Akademie ed. vol. XVIII, p. 618.
17 and 5286 Akademie ed. vol. XVIII,
See, for instance, R 5284 p. 143.
18
Work on this paper has been supported by a National Endowment for the Humanities for Independent and
Fellowship Study
Research for 1978-79.

This content downloaded from 128.197.27.9 on Tue, 23 Jul 2013 13:36:48 PM


All use subject to JSTOR Terms and Conditions

Potrebbero piacerti anche