. 3
( 9)


Such representations are called obscure . . . Thus the ¬eld of obscure represen-
tations is the largest in human beings. (Anthro, Ak. ©©, pp. ±µ“±)
While Kant denies Locke™s thesis that all representations involve con-
sciousness and even self-consciousness, he does assert that there is an
indirect connection between representation and consciousness. We can
know of a representation not only directly, but also indirectly in virtue of
its role in explaining overt verbal and other behavior. Since Kant does
not require that all consciousness involve self-consciousness, there is no
reason to think that he would not concede that I may sometimes become
conscious of a representation as mine only after the fact. And this
suggests that even babies and comatose individuals could become con-
scious after the fact of representational states that they were in while in a
state of unconsciousness. However, unless they were able to become
conscious of their present or past experiences as babies or comatose
individuals, those representations would have no cognitive signi¬cance
for them.
Unlike the A-Deduction, the B-Deduction seems to allow for the
possibility of representations in me or in someone else that have no
cognitive signi¬cance for anyone. Such representations would be no-
The demands of judgment in the B-Deduction
thing for anyone. While we would have no reason to assume the
existence of such representations, Kant now seems to rightly eschew the
veri¬cationist thesis with respect to representation that he advocates in
those passages in the A-Deduction that rule out the very existence of
representations that are nothing for me or anyone else.

°µ  ° °   ° ©  ®  ®¤  ® ©  µ ®   ¦
 ¬¦ - © ¤  ®  ©  
Kant refers to the self-consciousness of the ˜˜I think™™ as pure or original
apperception, a notion he explicitly contrasts with empirical appercep-
tion (section ±,  ±). This pure apperception is said to be original since
it is the source of the representation ˜˜I think™™. This ˜˜I think™™ represents
a logically basic subject of representation, since it refers to whoever may
be the ultimate bearer of that representation. As a ˜˜spontaneous™™
representation ( ±), the occurrence of the representation ˜˜I think™™
cannot be understood solely in terms of the causal history of the person
who has it. Thus in thinking about my representations, I must be repre-
senting those representations of mine in a way which is underdeter-
mined by any empirical facts about myself. I could have had the same
thought in a whole range of di¬erent causal circumstances. So far, it
would seem that the spontaneity of the ˜˜I think™™ might merely mark the
fact that my I thoughts are independent of the particular causal circum-
stances in which I have such thoughts. But Kant goes on to make
stronger and more interesting claims. He links the spontaneity of
thought, its originality, to a consciousness of self that is context-indepen-
dent. The ˜˜I think™™ is supposed to be ˜˜one and the same in all
consciousness™™ ( ±). It must therefore also be the same in the di¬erent
states of consciousness that characterize di¬erent persons. This context-
independence of the ˜˜I think™™ re¬‚ects its independence from any
particular facts about the causal history of particular agents.
Perhaps the most important feature of the ˜˜I think™™ is that it is ˜˜one
and the same in all consciousness.™™ This identity of the ˜˜I think™™ in
di¬erent psychological contexts has implications for an understanding
of the notion of mineness that is linked to cognitive signi¬cance for an I
thinker. The content of my representational states is determined by
conditions which make it possible for those representations to co-occur
in a single consciousness of self. The representations would not other-
wise be my representations. But such mineness is, in a certain sense,
general. It is true of each and every individual; as such it is subject to
· Kant and the demands of self-consciousness
whatever general conditions apply to all consciousness in virtue of the
dependence of such consciousness on a self-consciousness that has the
same general self-referential structure for each individual instance of
self-consciousness. The possibility of my becoming self-conscious of my
representations puts constraints on whatever is represented by me that
turns out to be general in the same way that the ˜˜I think™™ is general:
[A]s my representations (although I may not be conscious of them right away)
they [my representations] must necessarily accord with the condition under
which they can alone stand together in one universal :in einem allgemeinen9 self-
consciousness because they would not otherwise belong to me through all
transitions :durchgangig9. (section ±,  ±-±)

Minimally a general (˜˜allgemeines™™) self-consciousness must involve a
consciousness that the members of a set of representations belong to
oneself. However, since Kant also maintains immediately prior to this
claim that the self-consciousness in question is transcendental because it is
the basis for a priori knowledge (that is, universal and necessary knowl-
edge), he must mean that the self-consciousness in question is an
absolutely general, or universal, consciousness that di¬erent representa-
tions can necessarily belong together. So Kant is not just claiming that
one™s ability to become conscious of a group of representations is a
constitutive feature of what makes them one™s own; he is also claiming
that, because such representation is possible for each of us who can
become self-conscious, we must be able to think of all of these possible
contents of self-consciousness belonging to di¬erent possible individuals
as belonging to one possible global self-consciousness. This is a conscious-
ness that any representation which is of any cognitive relevance to any
person can belong to any other representation that is of cognitive
relevance to that person. The representation in question may then be
represented by that person as a representation of that very person. This
very abstract notion of ownership can only be the capacity to say ˜˜I think™™
with respect to any arbitrary set of representations that one might have.
Kant links ownership of representations to a potential consciousness
of one™s self-identity. One is supposed to be able to become conscious of
the diverse states of consciousness that belong to one. The capacity to
say ˜˜I think™™ with respect to diverse possible representations that I could
regard as mine is the basis for our consciousness of self-identity. But the
self-identity in question is, ¬rst of all, that of any subject in general. It is
only on the basis of that general representation of self-identity that we
are then able to represent our individual self-identity.
The demands of judgment in the B-Deduction
Kant™s account of the kind of self-identity that we attain through
self-consciousness has anti-Humean implications, although he agrees
with Hume that individual states of empirical consciousness give one no
consciousness of self-identity. Individual disconnected representations,
Hume™s perceptions, do not even have the resources to express beliefs or
any form of recognitional awareness, since this requires a distinction
and, hence, also a connection between subject and predicate. The
capacity to connect representations in one self-consciousness is what
makes consciousness of the connection between representations poss-
ible. And the same capacity for self-consciousness is the basis of all
consciousness of identity.
Kant™s concern with self-identity, as in the A-Deduction, is with the
enabling role that a representation of self-identity plays in our ability to
understand and use concepts. Self-consciousness is of crucial import-
ance in concept formation and, hence, in all thought. There is a unity to
the way in which what we represent is connected for us (a synthetic unity
in Kant™s terminology). This synthetic unity logically precedes the
formation of any concepts:

A representation that is to be thought as common to di¬erent ones, is thought
as belonging to ones that have apart from it something di¬erent in them,
therefore they must previously be thought to be in synthetic unity with others
(if only possible representations) before I can think the analytic unity of
consciousness that makes them into conceptus communis [common concepts]
in them. And thus the synthetic unity of apperception is the highest point on
which I must support all use of the understanding, even the whole of logic,
and after it, transcendental philosophy, yes, this faculty is the understanding
itself. ( ±n)

In order to represent anything in terms of a sortal or attributive
concept I must grasp what it is that distinguishes the class of objects to
which that concept applies from other possible classes. I must therefore
be able to compare and contrast di¬erent items that are represented by
me. In his logic lectures, Kant divides the procedure of concept forma-
tion into three steps: (±) comparison, through which I compare represen-
tations to one another in a single consciousness, () re¬‚ection, in which I
re¬‚ect on how to grasp di¬erent representations in terms of a certain
unity of consciousness provided by the features that those representa-
tions have in common, and () abstraction, in which I abstract from all
of the features that those representations do not have in common (Logic,
ed. Jasche, Ak. ©, p. ).
· Kant and the demands of self-consciousness
The important point in the present context is that, in order to be able
to engage in the procedure of comparison, re¬‚ection, and abstraction, I
need not only a consciousness in which those items are compared and
contrasted, but also the capacity to represent myself as being the same
subject of representations that represents the one as that which repre-
sents the other items. From such observations, Kant can legitimately
conclude that the analytic unity of any conceptual consciousness, any
ability to grasp what is represented by me in terms of a one over many,
such as a sortal or attributive term, requires a synthetic unity of self-
consciousness, that is, a capacity to represent the contents of representa-
tions together as my representations ( ±). It is this synthetic unity
which is a condition for the formation of concepts and at the same time
the condition for their applicability to objects of experience. It is import-
ant to notice that the synthetic unity in question is one that connects
di¬erent items represented in a universal self-consciousness. It is then
our capacity to represent ourselves as one and the same subject with
respect to all the diverse representations of such a universal self-con-
sciousness that allows us to represent things in the universal and stand-
point-neutral way demanded by Kant™s conception of a concept.
Kant commits himself more explicitly to a link between the stand-
point-neutral identity of the I that serves as subject of self-consciousness,
and the kind of standpoint neutrality involved in having a concept, in his
lectures on Anthropology:

[E]xperience is empirical cognition, but cognition requires re¬‚ection (re¬‚exio),
and hence consciousness of the activity in putting together the manifold of a
representation according to a rule of unity for that manifold, that is, a concept
and thought in general (distinct from intuition). The I of re¬‚ection contains no
manifold in itself and is always one and the same in all judgment, since it is
merely this formality of consciousness. (Ak. ©©, p. ±±)

As universal representations, concepts involve consciousness of fea-
tures common to a possible plurality of particulars ( °/ ··). They
are representations that can be contained in a number of numerically
distinct individual representations or intuitions (Logic, ed. Jasche, Ak. ©,
section ±, p. ±). But what makes such representations universal is that
they represent things in a standpoint-neutral way. This is possible
because, as I thinkers, we can think of ourselves and other things in a
way that is completely independent of any particular facts about us or
the world.
The demands of judgment in the B-Deduction
 ®   © ®¤  ¬¦ - ® © µ® 
At this point, I need to say something more about the nature of synthetic
unity. Kant is careful to note that the need for synthesis is a feature of
representations that are mine in the sense that they are cognitively
signi¬cant for me. This is true regardless of whether we combine what
we are representing under a concept, or combine it in one spatio-
temporal experience (intuition). Indeed, Kant is committed to our
capacity to conceptually represent any perceptual (intuitive) contents,
since concepts are involved in all thought, and he has already claimed
that the only intuitions that are of any cognitive signi¬cance for us are
ones that we can think of potentially as our own. In order for a
representation to have cognitive signi¬cance for us, we must be able to
distinguish its di¬erent contents. But in order to be able to distinguish
those contents, we must already be able to connect. For Kant, analysis
always presupposes synthesis (section ±µ,  ±°). Representing something
as connected, regardless of whether there is anything already complexly
characterizable there to begin with, involves an activity that one can
refer to as a self-activity. For it is through the activity of connecting and
distinguishing information that the subject establishes the connected-
ness of the object for itself :

[W]e cannot represent anything as connected in the object unless we have
previously connected it ourselves and among all representations connection is the
only one which is not given through objects, but can only be performed by the
subject itself, since it is an act of its self-activity. (section ±µ,  ±°)

The di¬erent bits of information that we experience are only accessible
to us as what they are insofar as we can compare them with each other.
But to compare them we must connect them to each other. It is this fact
about us that makes us discursive intellects. Indeed, although Kant
introduces the notion of a purely intuitive intellect (a God™s eye point of
view) by way of contrast with our discursive intellect, he argues that we
cannot even make sense of such an intellect, so that the theocentric
perspective on the world that has been popular with rationalist philos-
ophers is not even really coherent for him:

That understanding through whose self-consciousness a manifold of intuition
would be given, an understanding through whose representation the objects of
this representation would also exist, would not require a particular act of
· Kant and the demands of self-consciousness
synthesis of the manifold for the unity of consciousness, needed by human
understanding which merely thinks and does not intuit. But it is the ¬rst
principle for human understanding, so that it cannot form the least concept of
an other possible understanding, either one that itself intuits or even of a sensible
intuition, but of another kind than the one grounded in space and time. (section
±·,  ±“±)

For a being with a discursive intelligence, there is a fundamental
distinction to be drawn between the way objects are given to it (receptiv-
ity), and the way it represents those objects as given to it (spontaneity).
An intuitive or non-discursive understanding would not, according to
Kant, require a connection or synthesis that is distinguishable from the
way objects are given to it. In the Critique, Kant accuses the philosophi-
cal tradition of con¬‚ating understanding and intuition. So he cannot
take the idea that ours is a discursive intellect in his sense as uncontested,
for otherwise his philosophical critics can accept the conclusion that he
draws from what is required for experience by a discursive understand-
ing and simply deny that ours is a discursive intellect.
Although the Critique as a whole can be regarded as a defense of the
claim that ours is a discursive intellect, at  ±° Kant adduces the fact
about us that he thinks directly supports the claim that ours is a
discursive intellect. This synthetic fact links cognitive signi¬cance for us
to our being able to make cognitive connections for ourselves. A repre-
sentation is cognitively signi¬cant for me only if I can think of that
representation as a representation that could be mine in the sense that it
is connectible to other representations that I ascribe to it myself. Kant
can then appeal to the claim that it is analytic to all representations that
are mine that they are ascribable by me to me. This mineness of
representations gives even intuitions, including those of space and time
as a whole, their unity ( ±n).
Now the A-Deduction maintains that there is a synthetic a priori
connection between all empirical consciousness and a possible self-
consciousness ( ±±·n). There Kant is concerned with the implications of
self-consciousness for the empirical content of thought. He argues that
an empirical consciousness is only consciousness of an object insofar as it
can be connected to other contents of consciousness in one possible
self-consciousness. In the B-Deduction, Kant takes the need for syn-
thesis to be analytic to self-consciousness. There is, however, an import-
ant distinction between the A-Deduction and the B-Deduction. For the
B-Deduction claim concerns the need for synthesis relative to represen-
tations that are mine:
The demands of judgment in the B-Deduction
(±) The thought: these representations given in intuition all belong to me means
as much as I connect them in one self-consciousness, or can at least connect
them there, and even if it is not itself the consciousness of the synthesis of
representations, it presupposes the possibility of the latter, that is, only insofar
as I can grasp that manifold in one consciousness, do I call all those representa-
tions my representations; for otherwise I would have such a multicolored
di¬erent self, as I have representations of which I am conscious . . . now this
principle of the necessary unity of apperception is identical and hence an
analytic proposition, but explains a synthesis of the manifold that is given in one
intuition as necessary without which that pervasive identity of self-conscious-
ness could not be thought. (section ±,  ±µ)

It is analytic of a representation being my representation that it is one
that I can ascribe to the identity of my self-consciousness. But I can think
of di¬erent representations as belonging to my self-consciousness only
insofar as I can think of them as ones that are linked together by my
consciousness of self. This might suggest that these representations must
therefore have a content that is purely subjective. But Kant rightly
resists this conclusion. To think of representations as my representations
is to think of them as representations that belong together in a certain
distinctive unity of consciousness, that is, in a certain history. But they
can only belong to a certain distinctive unity of consciousness, in a
certain history, to the extent that they could be connected together with
the representations belonging to other distinctive unities of conscious-
ness, that is, to other distinctive histories, histories that are yours, his,
hers, and its. Moreover, representations can only be regarded as belong-
ing to these distinctive unities of consciousness, or points of view, insofar
as they can be taken to have cognitive signi¬cance for me and my
thought. It is my capacity for such I thoughts that allows me to represent
things from di¬erent possible points of view. I can do this precisely
because I thoughts have a conceptual content that is independent of any
particular point of view.
() The synthetic unity of consciousness is therefore an objective condition of all
cognition, not which I merely require in order to know an object, but under
which any intuition must stand in order for it to become an object for me, because in
any other way and without this synthesis this manifold would not become
uni¬ed in consciousness.The last proposition is, as was said, itself analytic,
although it makes synthetic unity the condition of all thought; for it says nothing
more than that all my representations in any given intuition must stand under
the condition that I can alone ascribe them as my representations to an identical
self, and therefore can grasp them together synthetically connected in one
apperception by means of the universal expression I think. (section ±·,  ±)
· Kant and the demands of self-consciousness

 ° °    ° ©  ®  ® ¤  ®   °    ¦      
After arguing that self-consciousness places impersonal constraints on
all my representations, Kant moves to the claim that these impersonal
constraints are the basis for concepts of objects. These concepts of
objects are then interpreted as ones that we use to form judgments and
to articulate cognition or knowledge in judgments. Initially, we expect to
see Kant derive the conditions for concept use, judgment, and knowl-
edge from the conditions governing self-consciousness; he seems instead
merely to shift from talking about conditions on self-consciousness to
talk of conditions on conceptual cognition and judgment without
clarifying how concepts or judgment depend on self-consciousness. It
seems as if, instead of arguing from the a priori enabling conditions for
the unity of self-consciousness to the a priori enabling conditions for
concepts and for knowledge of objects, he argues from the a priori
conditions governing concepts and knowledge of objects to the a priori
enabling conditions for the unity of self-consciousness.
An argument to a priori enabling conditions for self-consciousness
based on the existence of knowledge will only be convincing to the
reader who is already prepared to accept the existence of knowledge as a
given. Initial appearances to the contrary, Kant really wants to argue
that enabling conditions for the unity of self-consciousness are enabling
conditions for judgment and knowledge. His line of thought in arguing
for the idea that the uni¬ability of representations in self-consciousness
is a necessary condition for cognition may be reconstructed as follows:
knowledge involves judgment. Judgment involves the possibility of
agreement or disagreement between di¬erent persons about some pur-
ported state of a¬airs. As such, judgment involves an implicit commit-
ment to the idea that there is some normative ground that allows one to
determine whether the judgment is right or wrong, whether it is true or
false. But such a normative basis for agreement or disagreement is only
present where there is at least the possibility of a standpoint that is
outside of the standpoints of those who articulate judgments that either
agree or disagree. It is this standpoint that is implicitly presupposed
when we interpret a certain object in terms of certain concepts. Such a
standpoint is precisely that provided by Kant™s idea that di¬erent
representations must be uni¬able in a possible self-consciousness that is
able to abstract from any particular point of view within experience.
Without self-consciousness, I do not have a consciousness of myself as
one person among others. I am thus unable to represent my point of
The demands of judgment in the B-Deduction
view as one which is either distinguishable or indistinguishable with
respect to some subject-matter from that available to some other being.
But if I have self-consciousness, then I also have the capacity to repre-
sent myself in a way that is indistinguishable from the way any other
self-conscious being represents itself or the world. I exercise this capacity
when I represent myself in abstraction from what distinguishes my point
of view from other points of view and when I represent things in a
manner that is, in principle, available to anyone. Concepts represent
contents in terms of representations that can be had in altogether
di¬erent circumstances in experience.
Kant captures the implicit objective commitments of subjective ex-
perience, the implicit universal intelligibility of even subjective experi-
ence, by characterizing his notion of object in terms of representations
that are uni¬ed under concepts: ˜˜Object is that in the concept of which
a given manifold is uni¬ed™™ (section ±·,  ±·). Since concepts are
representations of items in a standpoint-independent or universal man-
ner, the concept of an object is something the content of which is
represented in a way that does not depend on a standpoint. If concepts
are to be applicable to experiences, and if those experiences are to
become an object for me, then those experiences must be uni¬able in
consciousness under concepts. But the consciousness in question must
have universal signi¬cance if we are to think of it as a consciousness that
could be right or wrong about what it is representing.
Failure to pay attention to Kant™s somewhat technical conception of a
concept can lead one to think that he simply shifts from a notion of
uni¬ability in consciousness that applies to anything that can be an
experience in any sense at all to a notion of uni¬ability that is restricted
to representations that have objective validity. For Kant moves directly
from his de¬nition of an object to the claim that representations uni¬ed
in that way in consciousness have objective validity and count as
cognition or knowledge (section ±·,  ±·).
As in the A-Deduction, Kant introduces the concept of an object in a
context in which he is also willing to talk of cognition (knowledge) of an
object that has objective validity. The idea seems to be that we only have
a bona ¬de concept of an object if we are able correctly to use the
concept in judgments that provide us with knowledge of an object.
Indeed, Kant is not satis¬ed with the assertion that cognition has unity
of consciousness as its necessary condition; he claims that unity of
consciousness is su¬cient for cognition.±° The important thing to note
here is that the unity of consciousness is su¬cient for cognition only
° Kant and the demands of self-consciousness
when it includes both empirical and a priori conditions for the unity of
self-consciousness. Kant does not suggest that the a priori conditions are
themselves independently su¬cient, nor does he suggest that empirical
conditions on the unity of consciousness would be su¬cient.
Allison tries to ¬ll the apparent gap in the argument generated by
Kant™s shift from treating the unity of consciousness as a necessary
condition for knowledge to treating it as a su¬cient condition for
knowledge by weakening the notion of object involved here to that of a
purely logical notion.±± This is tempting, but it will not work, since Kant
does not restrict cognition (Erkenntnis) to logical knowledge in the ¬rst part
of the B-Deduction as Allison maintains. Kant de¬nes knowledge as the
determination of representation in relation to an object (section ±·,  ±·).
In his summary of the argument in the ¬rst part of the Deduction, Kant
claims that unitary empirical intuition is determined with respect to the
forms of judgment (section °,  ±). He thus takes himself to be showing
in the ¬rst part of the B-Deduction that we have knowledge of objects
belonging to experience as well as purely formal objects.
Treatment of the unity of consciousness as su¬cient for cognition is
licensed by a further premise, that something is a cognition if and only if
it involves the uni¬cation of empirical representations under concepts.
Kant™s de¬nition of a cognition as the determinate relation of given
representations to an object supplies this missing premise. He does not
o¬cially introduce the notion of judgment until section ±, but he does
think of cognition as involving judgment. So the determinate relation in
question is a relation for judgment, since judgment is ˜˜the representa-
tion of the unity of consciousness of di¬erent representations or the
representation of their relationship insofar as they constitute a concept™™
(Logic, Ak. ©, section ±·, p. ±°±). Judgment represents items in a way that
commits one to those items being the same for everyone, that is to their
uni¬ability in an ˜˜I think p™™ that could be anyone™s. This capacity to
abstract from what is the case for me as a particular individual and to
take things as they would be represented by anyone else is what is
expressed by the ˜˜is™™ of assertion:
[A] judgment is nothing but the way given cognitions are brought to the
objective unity of apperception. That is the target of the little relational word
˜˜is™™ in them [judgments] to distinguish the objective unity of given representa-
tions from the subjective. (section ±,  ±±“±)

The universal commitments of judgment and claims to knowledge
expressed in judgment are already implicit in the concepts involved in
The demands of judgment in the B-Deduction
judgment. But in judgment concepts are related to each other in such a
way that a claim is made that can be either true or false. The unity of
consciousness in a judgment thus makes the relation of representations
to each other determinate in a way that makes agreement or disagree-
ment possible.
Kant ¬nally makes the distinction between the objective unity to be
found in shareable self-consciousness and the merely subjective unity of
consciousness we are accustomed to in our introspective self-conscious-
ness explicit in section ±. He identi¬es empirical apperception in
section ± ( ±°) with an associative connection between representa-
tions that is valid for me or for you, in contrast with an objective unity of
consciousness that is universally valid. In general, how I happen to
connect di¬erent words or other representations with objects in my
consciousness is a contingent matter that depends on the circumstances
under which I have come to connect those words with those objects.
Such accidental connection by association is not su¬cient for an objec-
tive unity of consciousness, that is, for a consciousness of what we
represent that can be the same for all of us. Kant identi¬es this objective
unity of consciousness with the transcendental unity of apperception
( ±).
Kant™s reference to transcendental apperception as having an objec-
tive unity to what it represents provides indirect support for the way I
have been reading his claim that representations must be able to belong
together in a self-consciousness. The universality of the unity of tran-
scendental self-consciousness that Kant invokes in the second para-
graph of section ± cannot be restricted to the representations in my
individual consciousness alone, for then it could not support the objec-
tive validity that he identi¬es with transcendental apperception in sec-
tion ±.
To be sure, the very introduction of a notion of ˜˜empirical appercep-
tion™™ ( ±°) in contrast to the transcendental unity of apperception
seems at ¬rst to wreak havoc with the argument. It is tempting to take
the view that consciousness of self-identity should then be possible
without the more ambitious conception of transcendental apperception.
Another response is to reject the notion that empirical apperception
really is a form of self-consciousness. On either of these interpretations,
the role of transcendental self-consciousness as enabling condition for
empirical self-consciousness drops out. Kant™s suggestion is crucial that
empirical apperception which is only subjectively valid is derived from
transcendental and objective apperception ˜˜under speci¬c conditions in
 Kant and the demands of self-consciousness
concreto™™ (section ±,  ±°). Empirical self-consciousness is character-
ized by a unity of the contents of consciousness that depends for its
character on the spatio-temporal context of a particular individual.
Such subjective unity of spatio-temporal experience is derived from an
objective unity based on facts about spatio-temporal objects that hold
for any arbitrary observer. Empirical apperception may be subject to
the unity required by transcendental self-consciousness without directly
displaying that unity.

 ¬ ¬©  ® ® ° °    °  ©  ®  ® ¤ µ     ©   µ ®©  
Henry Allison identi¬es apperception with the narrowly logical powers of
understanding and with the role of judgment in the making of objective
claims.± He worries that by assigning subjective validity to empirical
apperception Kant contradicts the principle that empirical consciousness
is subject to the transcendental conditions of unity. As a consequence of
Allison™s assimilation of self-consciousness to judgments with objective
import, self-consciousness threatens to drop out of subjective experience,
and subjective experience threatens to disappear altogether.
Allison insists that a subjective unity in consciousness is not something
through which even subjective states could be represented since it is not
something through which anything could be represented at all: ˜˜There
is in fact only one thing that could count as a subjective unity in the
Kantian sense: a unity or connection of representations through which
nothing is represented, not even our subjective states.™™± Since Allison
does not think that what Kant calls a subjective unity of consciousness
can be a bona ¬de unity of consciousness, he concludes that it cannot be
a fortiori a unity of self-consciousness either, although it can become the
topic of consciousness.± As objecti¬ed for self-conscious thought, the
subjective unity of my experience would be an object of judgment and
empirical knowledge. Thus on Allison™s interpretation all (empirical)
consciousness of oneself as a particular individual is knowledge of
oneself as an object. And this knowledge is knowledge by a non-empiri-
cal self. The implication is that all self-consciousness is knowledge of the
self through transcendental self-consciousness.
Although Allison initially maintains that Kant ˜˜refers to the subjec-
tive unity as a unity of consciousness and to the objective unity as a unity of
self-consciousness,™™ he soon is forced to concede that he also ¬nds Kant
˜˜treating the empirical unity of apperception as equivalent to the
subjective unity of consciousness. The problem is that Kant also seems
The demands of judgment in the B-Deduction
to regard empirical apperception as equivalent to empirical self-con-
sciousness, that is, as the mode of consciousness through which we
represent ourselves to ourselves as objects of inner sense.™™±µ As Allison
admits, the view that a subjective unity of consciousness is not a unity of
self-consciousness con¬‚icts with Kant™s reference to empirical appercep-
tion both as self-consciousness and as a subjective unity (B-Deduction:
section ±,  ±°). Allison concedes that, even though Kant notes that
the original unity of consciousness that is based on the relation of
intuition to one I think is ˜˜alone objectively valid,™™ Kant also insists that
˜˜the empirical unity of apperception which we are not considering here
and which is only derived from the former under given conditions in
concreto, has only subjective validity™™ ( ±°). Allison can ¬nd only
confusion in this remark of Kant™s because Allison™s interpretation of the
Kantian notion of apperception and self-consciousness only allows for a
self-consciousness of objective states of a¬airs.
Allison™s restriction of self-consciousness to objective states of a¬airs
forces not only him to ascribe confusion to Kant, it also has a more
devastating consequence. Our experience of ourselves as distinct indi-
viduals is based on what Kant calls inner sense. Allison argues plausibly
that ˜˜Kant™s theory of inner sense is best understood in terms of the
account of the subjective unity of consciousness which we have already
considered.™™± Because Allison™s reading of inner sense and the subjec-
tive unity of consciousness makes it independent of self-consciousness,
he argues that ˜˜Kant™s account of inner sense explains how the mind
can become aware of its own representations as ˜subjective objects,™ but
it does not explain how it can represent itself as an object.™™±· In contrast
to spatial objects, that is, the objects of outer sense, Allison thinks that
Kant is forced to adopt a ˜˜substratum™™ or ˜˜bare particular™™ theory of
predication when he deals with judgments about inner states.± One
might ask why the self cannot be experienced. Allison™s answer is: ˜˜the
important point is simply that, as non empirical, the I cannot know itself
through the (empirical predicates) representations which it refers to itself
in judgments in the same way in which it knows outer objects through
the predicates which it attributes to such objects in judgments of outer
The self ceases to be something that we can experience at all in
Allison™s reconstruction of Kant™s account of experience. For according
to Allison the self can only be experienced as an object of judgment in
which its character as a point of view is no longer in play. But worse, the
self cannot even be experienced as an object of judgment, for then its
 Kant and the demands of self-consciousness
character as self would elude us. This consequence of Allison™s interpre-
tation seems quite unappealing. In general, his failure to do justice to the
Kantian account of empirical self-consciousness and to the fact that
having self-consciousness involves having a particular point of view
makes his account of transcendental self-consciousness irrelevant to the
phenomenon of self-consciousness as it is generally understood. For
most consciousness of self is not a consciousness of a proposition that has
a truth value. To the extent that the self is something of which we can be
conscious in being conscious of judgments that have objective validity,
an account must be o¬ered of how our own subjective take on things
could be involved in a consciousness of something that on Allison™s view
must have a truth value.° We need to be able to understand how
empirical self-consciousness could require transcendental self-con-
sciousness. Allison™s interpretation of transcendental self-consciousness
precludes him from o¬ering such an account.±
Now I wish to argue that, in the logical space of reasons opened up by
transcendental apperception, transitions from one representation to
another are governed not only by the kind of causal connections that
underwrite habits of association, but by normative principles of rational-
ity as well. Since the causal connections between representations in-
volved in the empirical psychologist™s description of the regularities in
our individual representations themselves depend on our ability to make
normative claims about what came before what and where and what
caused what, these empirical connections are not really free standing. As
the A-Deduction makes abundantly clear, our ability to make sense of
associative patterns itself depends on our ability to form the concept of
an object that is independent of the way it appears to us at any given
moment, and this requires the possibility of a point of view that is able to
abstract my own particular take on things.

     © ¬ ©¤© ¦ µ¤§ ®  ®¤   µ® © ¦
  ®   ©  µ  ®  
It is worthwhile to contrast the account of judgment in the B-Deduction
with that of the Prolegomena. In the Prolegomena, Kant distinguishes be-
tween subjective judgments of perception that do not require the appli-
cation of categories and necessarily intersubjective judgments of experi-
ence involving application of the categories. In the case of judgments of
perception, no issue of disagreement or agreement can arise, since they
merely express the subjective take that an individual has on his or her
The demands of judgment in the B-Deduction
own experience, whereas in the case of judgments of experience agree-
ment or disagreement between persons is necessarily possible:
[J]udgments are either merely subjective when representations are referred to a
consciousness in one subject only, and are united in it, or they are objective
when they are united in consciousness in general, that is necessarily.
(Prolegomena, section , Ak. ©, pp. °“µ)

In a review of Ulrich™s Institutiones, Johann Schultz argued that the
theory of perceptual judgment in the Prolegomena was inconsistent with
the claim in the ¬rst edition of the Critique that all perception is subject to
the categories. Kant refers to the review in a famous footnote to the
Metaphysical Foundations (Ak ©, p. ·n) near the date of the second
edition of the Critique, so he clearly gave the problem some thought.
Perceptual judgments in the Prolegomena are independent of any applica-
tion of the categories. This is what distinguishes them from judgments of
experience. The independence of perception from the categories is not
something that Kant can advocate without giving up the universal scope
of categories with respect to our experience, and that would be to give
up the claim that the categories can make to being a priori enabling
conditions of experience.
In the B-Deduction, Kant responds to this objection by restricting
judgments to the objective states of a¬airs expressed by judgments of
experience in the Prolegomena. Such judgments involve the use of catego-
ries. For categories are the basic concepts that underwrite claims to
objectivity. In section ± of the B-Deduction, the distinction between
judgments of perception and of experience in the Prolegomena becomes a
distinction between judgments of experience and associative connec-
tions between perceptions. There are now statements that seem prima
facie to express judgments which turn out merely to express associ-
ations. Such statements are to be understood as a mere evocation of
inner states, comparable to cries of pain. In the cases of such subjective
statements, agreement or disagreement is inappropriate. It no more
makes sense for me to call into question the associations that you have
than it does for me to reject your pain. Judgments, by contrast, are now
taken by Kant to make claims that presuppose the possibility of agree-
ment or disagreement.
Judgments are based on what Kant calls the necessary unity of
self-consciousness, that is, the unity that representations must have if
they are to be able to belong together in any self-consciousness. By
contrast, associative connections of the kind established by largely
 Kant and the demands of self-consciousness
accidental causal circumstances are based on empirical self-conscious-
ness (section ±,  ±). The fact that judgment involves an implicit
commitment to the existence of necessary connections between what is
represented by the concepts that one uses in the judgment seems at ¬rst
to rule out any possibility of expressing a subjective standpoint. How-
ever, such a commitment is always taken on from a particular subjective
standpoint. Thus, judgment is compatible with the existence of di¬eren-
ces between persons in what they experience. The important part is how
what is experienced is interpreted. A judgment should express relations
between the objects and properties or relations attributed to those
objects that purport to be true for anyone in the circumstances stated by
the judgment.
The assimilation of objecthood to objectivity, and the idea that
judgment expresses objectivity lead to several problems. The ¬rst prob-
lem concerns the relation between judgment and knowledge. Kant
sometimes treats judgment as a relationship between concepts that is
objectively valid (section ±,  ±). Of course, not all judgments are
objectively valid even if they purport to be so. But, on the whole and in
the same context, he implicitly concedes that judgments have the kind of
objective unity that yields objective validity merely as their target. Not
every judgment is an instance of knowledge. Kant is interested in the
fact that a judgment is a claim to knowledge. As a judgment it must
purport to be true. He also treats all objects of experience as publicly
accessible because he wants to deny that there are any private objects
that are not available to us under some public description. Any object of
which I can be conscious ought to be an object of which someone else
could have indirect consciousness. Kant thus ¬nds himself compelled to
defend the view that all judgment purports to state objective facts.
Restricting judgment to objective facts seems to leave no room for
judgments concerning my inner states, at least insofar as these judg-
ments are about those inner states as I experience them. Since knowl-
edge or cognition requires judgment, the implication seems to be that
there is no self-knowledge.
Kant tries to ¬nd a way of accommodating self-knowledge in the
second step of the proof in the B-Deduction, which I will concern myself
with in the next chapter. I will show how the second step of Kant™s proof
resolves the di¬culties posed by his theory of objective judgment for his
account of inner experience. In the process, I will try to throw some light
on the manner in which subjective experience is dependent on objective
° µ

Self-consciousness and the unity of intuition:
completing the B-Deduction

Kant develops and supports the claim to objectivity implicit in judgment
by ¬rst arguing that all judgment that is dependent on a speci¬c
subject-matter dependent judgment derives its content from spatio-
temporal experience and then by arguing that we can represent all
objects in space and time together in a manner that is standpoint-
neutral. This standpoint-neutral manner of representing objects in
space and time is due to their relation to a possible self-consciousness. As
contents of consciousness, objects in space and time are representable in
a manner that depends on the spatio-temporal standpoint of the observ-
ing consciousness. However, this standpoint-dependent perspective is
itself only intelligible relative to a possible standpoint-independent per-
spective from which the standpoint of the observer becomes cognitively
By appealing to the standpoint-neutral constraints on representing
standpoint-dependent truths, it is possible to justify the objectivity claim
made by judgment. Objectivity then consists in the way things must be
represented in space and time so that they are the same for all observers
at all spatio-temporal locations. Kant seeks to make it comprehensible
how even subjective experiences can be regarded as subject to objectiv-
ity constraints. The key thesis here is that subjective experiences are
inherently dependent on the way things present themselves to the
spatio-temporal point of view of some consciousness. But this particular
point of view is only intelligible as a speci¬c perspective that one can
take as a self. In self-consciousness one is then, in principle, able to
combine the di¬erent possible spatio-temporal points of view in a single
encompassing objective point of view. It is because the individual
perspectives of diverse subjective experiences themselves presuppose a
single comprehensive point of view of which they are the perspectives,
that subjective experiences are subject to objectivity constraints.

 Kant and the demands of self-consciousness

  °   ¬   ¦    °   ¦    µ   µ  
Initially, Kant™s argument for the objective validity and empirical reality
of the categories seems to be complete by section ° at the end of the ¬rst
step in his two-step proof. Thus, his transcendental deduction seems to be
complete. He takes himself to have shown that categories act as con-
straints on the empirical contents of judgments made by any possible
self-consciousness. But he soon claims in section ± that he has only just
begun his proof. This has given rise to the much debated problem of the
proof structure of the B-Deduction. The question of the proof structure is
of some importance to my general argument. For I wish to argue against
the leading interpretations that Kant does not need to appeal to an ad hoc
assumption of spatial and temporal unity in order to support his argument
from the self-ascribability of representations to their being subject to
conceptual constraints that allow us to form judgments about them.
Kant summarizes what he takes himself already to have proven in the
form of a syllogism (section °,  ±). The major term of the syllogism
and initial premise of the argument states that the content of intuition
must be subject to the unity of self-consciousness, since this unity of
self-consciousness makes the unity of what we sensibly experience poss-
ible. This is what he takes himself to have established by section ±·. The
middle term of the syllogism injects the idea that data are brought under
self-consciousness in general through logical functions of judgment.
This is supposed to be established in section ±“±. The notion of a
logical function is not mentioned explicitly in section ±, but Kant has
developed the idea that the objective import of judgment is based on its
relation to impersonal self-consciousness, and he has already argued in
the Metaphysical Deduction of the Categories that categories are de-
rived from the logical functions underlying thought by applying such
logical functions to objects. Assuming that the way objects may be given
to us is through empirical intuition, he then concludes that anything
given through a unitary empirical intuition is determined with respect to
the forms of judgment. For it is through being determined in respect to
the logical functions of judgment that contents are brought into a
˜˜consciousness in general™™ (Bewusstsein uberhaupt) (section °,  ±). The
important thing to note here is that every cognitively signi¬cant (spatio-
temporal) intuition is given to us as part of an intuition that is unitary in
the sense that it can be represented by us as belonging to our own
consciousness. But this need not yet involve a representation of how the
object of such an intuition would ¬gure in a consciousness in general,
The unity of intuition: completing the B-Deduction
that is, in an impersonal consciousness. The task of showing what the
objective position in consciousness and intuition is of such an object falls
to the category:

A manifold that is contained in an intuition that I call mine is represented as
belonging to the necessary unity of self-consciousness by means of the synthesis of
the understanding, and this occurs by means of the category [ ±]. The basis
for the proof depends on the represented unity of intuition through which an
object is given that always includes a synthesis of the manifold given to an
intuition and already contains the relations of the latter to the unity of apper-
ception. ( ±n)

According to Kant™s de¬nition of the categories, categories are not
just forms of judgment, but ways in which data are determined by
thought to ¬t forms of judgment (section ±). Application of the category
of substance to experience determines, for instance, whether informa-
tion provided by sensibility is to be represented by a singular or an
attributive term. Given that anything represented by us in a single
empirical intuition is supposed to be subject to the categories, the
conclusion of the argument must be that we have knowledge whenever
we have a unitary empirical intuition, that belongs to a consciousness in
general, that is, an empirical intuition that is the same for each of us.
The initial conclusion in section ° limits the domain of application that
categories have to information in unitary empirical representations,
thus seeming to allow for the possibility of non-unitary intuitions. But
the ¬nal conclusion in section ° seems to close o¬ the possibility of
non-unitary intuitions by lifting the restriction: ˜˜Therefore the manifold
of a given intuition necessarily stands under categories™™ ( ±). The
possibility of non-unitary intuitions is at any rate di¬cult to take very
seriously in Kant™s epistemology, since intuitions are de¬ned as unitary
and immediate representations.±
Although it is now generally agreed that the B-Deduction argument
consists of a proof in two steps, with the argument from sections ±µ to ± as
one step and the argument from sections  to · as the other, what is
supposed to be shown in each step is very much a matter of debate. There
are ¬ve di¬erent basic approaches that have been taken to the proof
structure in the more recent literature: (±) according to the Henrich
interpretation,the argument of the ¬rst step is synthetic, but is restricted to
the applicability of categories to unitary intuitions, while in the second
step the applicability of categories to all of our sensible intuitions is
demonstrated. () According to the interpretation ¬rst suggested by
° Kant and the demands of self-consciousness
Tenbruck and developed by Bernhard Thole and Henry Allison, the ¬rst
step shows that categories apply only to possible objects of intuition,
relying on a second step for a proof that they apply to actual empirical
objects. The ¬rst step is then analytic, while only the argument of the
second is synthetic. () The approach defended by Brouillet, Wagner,
and, more recently, by Howell, takes the ¬rst step to show that categories
apply to intuitions in general and the second to show that they apply to
objects of our empirical intuition. () According to Baum™s interpreta-
tion, the second step of the proof is required in order to demonstrate that
space and time are intuitions in the sense required by the ¬rst part.µ (µ)
McCann argues that the ¬rst step of the Deduction is based on the analytic
principle of apperception that all my representations can be represented
by me as mine. The concept of an object in general and the a priori
concepts that specify it are derived from the unity of this apperception. In
the second step, Kant then purportedly argues that one cannot so much as
think of oneself as an individual thinking thing without having determi-
nate knowledge of oneself. From this it follows that the existence of a
determinate self-consciousness entails the validity of categories that apply
to objects of sensory experience.
Henrich attributes some confusion to Kant about his own proof-
intentions in an e¬ort to explain away those passages (especially in
section ±) in the ¬rst step which appear to claim that any manifold of
our intuition belongs to a unitary self-consciousness (see  ±,  ±µ“
±). Kant is supposed to slide from ascribing unity to any representa-
tion which is mine in some restricted sense to ascribing unity to all
representations in me. But, if he is guilty of this error, why should he
then attempt to prove the latter claim in an additional step? The
principle of charity discourages one from accepting Henrich™s proposal
as a reconstruction of Kant™s own intended purpose for the second half
of his proof.· It is also di¬cult to see how an appeal to space and time
would help to establish the fact that all intuitions must be unitary, if that
unity depends, as Henrich claims that it does, on powers of synthesis by
the understanding, the scope of which are themselves in doubt.
One can object to the second interpretation of the ¬rst step as an
analytic argument that this makes it implausible to refer to the ¬rst step
as a bona ¬de ¬rst step in a two-step proof, since the burden of proof has
then been shifted to the second step. The decisive di¬culty for the
second view seems to me however to be that Kant insists in his summary
of the ¬rst step in section ° that ˜˜any manifold insofar as it is given in
one empirical intuition is determined in respect to one of the logical
The unity of intuition: completing the B-Deduction
functions through which it is namely brought to one consciousness in
general™™ (section °,  ±). Kant thinks he has already established the
possibility of empirical judgment and empirical knowledge in the ¬rst
step. In fact, he introduces cognitions (˜˜Erkenntnisse™™) into the ¬rst step
in the B-Deduction through the claim that they ˜˜consist in the determi-
nate relation of given representations to an object™™ (section ±·,  ±·).
The speci¬c contribution of the second step cannot therefore be a proof
that categories are objectively real. The notion of determination is used
by Kant to characterize a feature of the object of judgment which is
precisely not determined by logic alone. For instance, which concept
will serve as predicate and which as subject in a judgment will be
determined by which concept is held to refer to a substance and which
to its accident. This point is made in section ± ( ±“±), a passage to
which Kant himself refers in the next sentence of section °. Kant must
therefore believe himself already to have proven that the categories have
not only objective, but also empirical reality by the time he has com-
pleted the ¬rst step of the proof. This objection also seems to be enough
to reject the third interpretation. The conditional claim that the Baum
interpretation ascribes to the ¬rst step in the proof is hard to identify in
the text. It is also di¬cult to see why Kant would not think that the
Aesthetic already simply supplies the antecedent premise that we have a
priori knowledge of space and time to the conditional claim without an
additional step in the proof.
McCann is certainly right that in the second step of the Deduction,
Kant wants to show that all states of my empirical self are subject to the
categories. However, a central point of Kant™s is also that not all
experiences that I have as a particular empirically knowable individual
are instances of objective judgment or knowledge of the kind requiring
the application of categories. How do the categories apply to such
subjective experiences if they are constraints on objective judgment?
McCann misses this important dimension of Kant™s argument, and thus
leaves the nature of subjective experience a mystery for Kant, because
he con¬‚ates all empirical consciousness of self with empirical knowledge
of the self. McCann argues that by making self-knowledge a necessary
condition for consciousness of oneself as a thinking being one is able to
meet a Cartesian skeptic who endorses the cogito argument, but rejects
our possession of any knowledge of the content of our mental or physical
states. But, curiously enough, McCann quotes a passage from the
Refutation of Idealism in which Kant expressly disconnects the certain-
ty of the cogito from any claims of self-knowledge:
 Kant and the demands of self-consciousness
Certainly, the representation ˜˜I am,™™ which expresses the consciousness that
can accompany all thought, immediately includes in itself the existence of a
subject; but it does not so include any knowledge of that subject, and therefore
also no empirical knowledge, that is, no experience of it. ( ··)

Nothing in the Deduction does anything to undermine a Cartesian
skeptic who accepts the existential claim embodied in the conditions
governing the assertion of the proposition ˜˜I think,™™ but rejects the
existence of external bodies. However, section  of the Deduction does
argue with the Refutation that knowledge of inner states is parasitic on
knowledge of outer states.
It seems to be speci¬cally the problem of self-knowledge that leads
Kant to divide the argument of the B-Deduction into two steps. Since he
de¬nes knowledge in terms of judgment in the B-Deduction, and judg-
ment has an objective force that the so-called perceptual judgments of
the Prolegomena do not, the implication of Kant™s new theory of judgment
appears to be that no knowledge of inner states is possible. For inner
states seem to be precisely states that are inherently subjective. This
suggests that there is no such thing as self-knowledge and that there are
no inner objects of judgment. Since categories are de¬ned in terms of
that with respect to which logical functions of judgments are deter-
mined, the implication is that categories do not apply to inner states. A
whole dimension of experience appears to resist use of the categories.
Kant needs to block this implication of his new theory of judgment,
without undermining the link he has established between judgment,
objectivity, and self-consciousness.
At the same time he must meet the serious objection that his theory of
self-knowledge renders transcendental idealism as a whole incoherent.
The clever critic Pistorius had argued in his ±· review of the
Prolegomena that transcendental idealism is incoherent because the thesis
that things in themselves are unknowable cannot apply to self-knowl-
edge without undermining the possibility of anything genuinely appear-
ing at all. Pistorius could not:
convince himself that the sensations given in time would be merely phenomena
just as the intuitions given in space, because he cannot help himself past the
di¬culty that because our inner sensations or representations would not then
be things in themselves, but appearances, there would be nothing but mere
illusion :Schein9and no real object would remain to which something would
appear. How is one to think it possible that representations which one must
after all presuppose as real [reell] or as things in themselves, if one wants to
explain at all how appearing :Scheinen9 is possible, can themselves be an
The unity of intuition: completing the B-Deduction
illusion :Schein9, and what is it through which and in which this illusion
:Schein9 exists?±°
Consciousness of oneself must be of a bearer that is not itself a mere
appearance, if there is to be any representer to which inner and outer
objects of any kind are to appear. Otherwise, appearance becomes total
illusion. At the same time, self-knowledge and knowledge in general
must be restricted to objects as they must appear to us spatially and
temporally, if the categories are to apply necessarily to all the objects to
which they do apply.
In contrast to the di¬erent reconstructions of the argument in the
Deduction, I wish to argue that the ¬rst step attempts to show that all
cognitively signi¬cant contents of our representations are candidates for
judgments determined by categories. Cognitively signi¬cant contents
are candidates for judgment, because they are potential candidates for
the kind of impersonal self-consciousness expressed by the proposition
˜˜I think.™™ One might worry that the restriction of the argument to
cognitively signi¬cant representations would leave the argument in the
¬rst step incomplete. However, there is no reason to think of the
argument for objectivity in the ¬rst step as essentially incomplete, since
Kant has no need to be concerned with the status of representations that
are completely beyond anyone™s ken. Why then is a second step needed
in Kant™s proof? Kant needs to show how all sensible representational
contents can be candidates for judgment requiring categories. In this
way, Kant can sustain the claim of judgment to objectivity and still
provide room for subjective experiences.
Before the second step, Kant claims to have abstracted the second
step from the manner in which data belonging to an empirical intuition
are given. He now wants to explain how it is possible for the categories
to apply to all objects:
In what follows (section ), it will be shown through the way that empirical
intuition is given in sensibility that its unity is none other than that which the
category prescribes according to the previous section ° to a given intuition in
general and by thus explaining their validity a priori with respect to all objects
of our senses the Deduction will be attained for the ¬rst time. (section ±,  ±µ)
Since Kant has already concluded that any manifold given in an
empirical intuition will have to be subject to the categories, in the second
step, he can only be concerned with demonstrating how the validity of
the categories with respect to all empirical objects is to be explained,
that is, he wants to show how a priori knowledge of all objects of
 Kant and the demands of self-consciousness
experience is possible. Thus in section  he notes that he ¬rst argued for
the categories as a priori cognitions of objects belonging to an intuition
in general (sections °“±), and now wants to explain how categories can
apply to all objects that may present themselves to our senses.
Kant thinks that an a priori claim must be necessarily true of all the
objects of which it is true. And he thinks that this implies that the objects
of which an a priori claim is true can only be objects as they must appear
to us. This is why he invokes his conception of transcendental idealism
to show that the categories must apply to all objects of experience and
only to objects of experience. First, he argues that judgment and
cognition, including mathematical knowledge, is restricted to the form
of objects belonging to our experience (sections “). He then argues
that judgment and hence the categories that provide contentual con-
straints on judgment necessarily apply even to our inner experience,
even though not every inner experience is per se an object of judgment
(section ). While inner states are necessarily subject to a pre-judgmen-
tal temporal synthesis that makes them candidates for judgment, the
judging subject is conscious of itself in a way that is independent of such
temporal synthesis (section µ). Kant concludes his argument by noting
that any object of empirical consciousness in any sense, and hence all
inner states, belong to space and time that have their unity in virtue of
their cognitive signi¬cance for self-consciousness. Thus, even the subjec-
tive experiences that belong to my empirical self must be subject to the
laws that allow one to form a single empirical space and time in which
every object has a determinate position for an objective judgment.

 ¬ ¦- ®  ©µ ®  ,  ¬ ¦ -« ® · ¬ ¤ § , ® ¤
 µ       
Kant does not just restrict the role of the transcendental self-conscious-
ness expressed in the proposition ˜˜I think p™™ to concepts, judgments,
and inferences. He argues that any representational content that is to
have any cognitive signi¬cance for any one of us, must be a representa-
tional content that is a potential content of the proposition ˜˜I think p.™™
The world is not simply a construction of the self. But the world is
something which the self must construct for itself on the basis of informa-
tion it receives. The self must be able to view information provided by
receptivity as information for it. In order for the self to view information
as its own, it must be able to interpret and hence to conceptualize
experience. This means that even those representations that have a
The unity of intuition: completing the B-Deduction
content that depends on our receiving information from the world must
be potential contents of the proposition ˜˜I think p.™™ Such representa-
tional contents are potential candidates for judgments of the form ˜˜I
think p™™ because they are experienced by us in a manner that is already
concept-laden. However, the representational contents provided by
receptivity are not themselves concepts, judgments, or inferences.
Rather they are contents that are concept-laden because they are
potential candidates for the kind of self-consciousness expressed in the
proposition ˜˜I think p.™™
The most fundamental distinction within our capacity for receiving
information is that between information that is internal to our own
particular point of view as particular individuals endowed with self-
consciousness and that of information that is external to our own point of
view as particular self-conscious beings. The distinction between what is
external and what is internal to representation is not one that is entailed
by thought in its most general sense. However, any ¬nite rational being
endowed with self-consciousness will have to be able to draw a distinction
between what is internal to its own point of view and what is external to
that point of view. Due to the nature of ¬nite self-consciousness there
must be objects that present themselves to a being that is conscious of
itself in such a way that it represents them as internal to its point of view
and in such a way that it represents others as external to its point of view.
For without the distinction between the internal and the external, there is
no determinate consciousness of oneself as a distinct individual at all.
Such consciousness is implicitly contrastive. For there to be a contrast
there must be some distinction between the way things are for me and the
way they are externally to me.
Creatures that have a sense of self that they can articulate have the
ability to attribute experiences to themselves and thereby to distinguish
themselves from other objects and other selves. The ability to ascribe a
multiplicity of di¬erent episodes belonging to experience to oneself as
opposed to attributing it to some other individual, accounts for the
general sense in which what is experienced by a self is inner to the self. In
order for the self to be conscious of facts about itself as a distinctive
individual, the self must be given to itself in a way which distinguishes it
from other selves. The particular experience which distinguishes one
individual self from another may be referred to as what is inner to that
self. What is inner is juxtaposed to what is outer in experience. What is
outer in experience is just what is outside of the self as representer while
what is inner is what makes the representer™s point of view what it is.
 Kant and the demands of self-consciousness
The distinction between the internal and the external would seem to
lapse for a non-¬nite rational being, at least for a being that was
non-¬nite in every respect. But unless a ¬nite rational being can draw a
distinction between the inner and the outer, it will not be able to
distinguish the way things appear to it from its own point of view from
the way they might appear to some other possible point of view. Such a
being would have no grip on the idea that it itself has beliefs. Unless one
can allow for the possibility that one™s own beliefs might be di¬erent
from those of someone else, one does not have the notion that one holds
them for either the right or the wrong reasons. But then there is no
reason to think that one is a (¬nite) rational being at all. Even a
non-¬nite rational being would need to be able to represent things, but it
would not need to regard its point of view as distinct from the point of
view of the whole universe.
Now, even if the very notion of ¬nite rationality requires some
distinction between the inner and the outer, this does not give any
indication how such a distinction must be drawn. Kant makes two
rather controversial moves here. First, he argues that there is a form of
inner and a form of outer experience, then he maintains that the only
distinction we have between the inner and the outer is one we make in
terms of time and space. He concludes that the form according to which
we must order our inner states is time, while the form according to
which we must order our outer states is space.
Kant insists that the di¬erence between what is internal and what is
external to a certain ¬nite point of view is to be expressed in terms of the
distinction between states that are essentially tied to a certain point of
view and those that are, in principle, independent of a certain perspec-
tive. Past, present, and future are essentially dependent on a temporal
point of view, while temporal properties or relations that are indepen-
dent of a particular point of view seem to depend for their existence on
spatial relations. From this, Kant concludes that time and space are the
only forms in terms of which we can make sense of objects given to the
self. He refers to these forms as the forms of receptivity and, more
speci¬cally, as the forms of our sensibility. He maintains, quite plausibly,
that the content of our experience is essentially spatial and temporal
because the faculty through which we receive information about the
world represents the world to us spatially and temporally.
Kant argues for the claim that there must be a form of inner and outer
sense in the following way: in order to be able to relate sensations to
objects that are external to each other or internal to our point of view
The unity of intuition: completing the B-Deduction
one must already have the disposition to represent objects as outside of
or inside of our point of view in terms of correlated sensations. The way
in which what is sensed is ordered is distinguishable from what is sensed.
It is not a property of what is sensed in isolation but depends on the
relations between what is sensed: ˜˜That in which sensations alone order
themselves and can be put in a certain form cannot itself be in turn a
sensation™™ ( °/ ).±± The way in which sensations array themselves
is a function in part of those sensations themselves. This suggests the
idea that sensations might have positions relative to each other that
would be su¬cient to induce an order in inner and outer sense. How-
ever, sensations taken on their own lack intentionality, that is, they lack
object-directedness. They are what Kant calls ˜˜subjective representa-
tions™™ that merely express the state of their bearer rather than represen-
ting an object. Since sensations do not represent objects, they do not
represent outer or inner objects. To represent outer or inner objects we
need a representation that does have intentionality, and this representa-
tion will have a content that is derived from sensation, but a form that is
independent of such sensation. Here Kant distinguishes a form accord-
ing to which we represent outer objects, and a form according to which
we represent inner objects. The former is the form of outer sense, the
latter is the form of inner sense.
The notion of outer sense is relatively straightforward, since we
have a good rough-and-ready understanding of what it is for an object
to be experienced as outside of our representations. It is somewhat
more di¬cult to understand just what it means for an object to be
internal to our representations and yet logically distinct from those
representations. Thus, Karl Ameriks distinguishes three di¬erent the-
ories of inner sense: (±) the re¬‚ection, () the independent stream, and
() the act theory.± According to the re¬‚ection theory, inner sense
consists only of re¬‚ection on past acts of consciousness. This theory
seems to draw its support from earlier texts in which Kant did not
distinguish between inner sense and self-conscious re¬‚ection. It has the
di¬culty that it fails to account for how representations are given to us
in the ¬rst place which is one of the avowed tasks of Kant™s conception
of inner sense. According to the independent stream theory, there are
two streams of consciousness, one representing spatial contents, and
the other representing contents that may have no direct reference to
spatial contents. This two-stream theory fails to do justice to the de-
pendence of both spatial items and mental events on the (single) tem-
poral stream of our successive representings. Finally, according to the
 Kant and the demands of self-consciousness
act theory, anything that is a representing, rather than a represented,
belongs to inner sense. The notion of an act suggests a process of
taking that is not wholly appropriate to the way representations are
immediately given to us in inner experience. I want to defend a modi-
¬ed version of the act theory according to which anything belongs to
the form of inner sense in virtue of being something represented by a
representing. Thus, on my interpretation, sensations belong to inner
sense, although they do not have an object that is logically distinct
from the sensing itself. But inner sense also includes the representa-
tions of all other objects insofar as they are potential objects of con-
The identi¬cation of the form of inner sense with time and the form of
outer sense with space raises obvious questions. For one thing, things
outside of us seem to be as much in time as our inner states. And other
creatures have distinctive experiences that are internal to their distinc-
tive points of view, but external to each of our own. It is not an analytic
truth about self-conscious beings in general that they must experience
things temporally or spatially. Di¬erent temporal series of representings
corresponding to di¬erent spatial standpoints do, however, distinguish
di¬erent self-conscious beings for us. These di¬erent series are orderable
in relation to each other in time and space, giving di¬erent empirical
meanings to the notion of the inner and the outer. Now, in order for one
to have a way of thinking of a certain set of representations as belonging
to a certain point of view, there must be some further way of character-
izing the relation of that point of view to another point of view. This
function is performed by linking time as the form of inner sense to space
as the form of outer sense.
Without an outer sense there would be no inner sense. Due to the
diaphanous character of consciousness and its self, the feature that Kant
refers to as the emptiness of consciousness, there would not be anything
to be represented by an experiencer. If the self as a determinate self is
essentially associated with a temporal point of view undergoing a suc-
cessive shift, then that which is outside of the self must be thought of as
not essentially temporal. We can make sense of the notion of a givenness
to us that is not essentially temporal through the notion of spatiality.
Space provides us with a way of organizing those items that we experi-
ence as outside of us in a system of di¬erences common to di¬erent
points of view. Time distinguishes di¬erent experiences that belong to
the same individual, and thus also provides a way of attributing di¬erent
mutually exclusive states to the same individual regardless of whether
The unity of intuition: completing the B-Deduction
those states are themselves experiences or not. We then bring space and
time together when we think of di¬erent individuals, each of which is
outside of the other, as being in a sequence of di¬erent states.
Kant takes the plausible position that the inner experience which is
essential to a distinctive notion of mineness does not have a content
which is uniquely inner (section ,  ±µ; section ,  ·). The content of
our representations of our inner states, inner experience, is derived from
outer experience. This thesis is intimately connected to his argument
against psychological idealism. In his Refutation of Idealism ( ·¬.),
he argues that I could not have determinate mental states if there were
no physical objects (no objects existing outside of my mind). The upshot
of this is that there is no self as a particular empirically knowable
individual that can exist independently of a certain body. The necessary
relationship between determinate mental states and embodiment avoids
the bare substratum view of the self that has sometimes been ascribed to
Kant. By committing Kant to a form of externalism concerning the
content of representations, the dependence of mental content on em-
bodied experience also provides the self with a determinate point of view
from within experience.± It is, however, true that the self is never itself
directly present as an object of empirical inquiry. This is because the self
is essentially the point of view from which any inquiry can take place.
The self can thus be characterized as an object of knowledge only by
objectifying the set of empirical representations that distinguish one
point of view in the totality of all experience from another point of view.
It is possible to derive a kind of no-ownership theory of the self from this
view, as Allison does. But such a view fails to do justice to the fact that we
are able to think of ourselves as a distinct individual, with a distinctive
point of view, by identifying the point of view of the self in general with
the history of a particular body. It is this body and its states that are then
the appropriate objects of self-knowledge. They have the kind of accessi-
bility to intersubjective scrutiny that Kant demands for bona ¬de
judgments and cognitions. Once we have such self-knowledge or knowl-
edge of others, then the spatio-temporal location and hence standpoint
dependence of my self-consciousness accounts for the discrepancy be-
tween the merely subjective validity of what is given to self-conscious-
ness empirically and the intersubjectivity possible on the basis of an
impersonal self-consciousness. We can thus allow a place for subjective
experience in our account of subjectivity without giving up on the claim
that objective experience is a condition for the very intelligibility of
subjective experience.
±°° Kant and the demands of self-consciousness
In section  Kant makes it clear that all non-empirical concepts have
content only in respect to appearances which are given to us through
our inner sense and outer sense, that is through our experience of
objects in time and space. He speci¬cally argues that self-knowledge
requires a ¬gural synthesis of transcendental imagination. The import-
ant thing about this synthesis of the sensible contents of experience is
that it is contrasted with the kind of ˜˜intellectual synthesis™™ involved in
applying categories to objects of an intuition in general. While it takes
places ˜˜in accordance with the categories,™™ and depends on the ˜˜orig-
inal synthetic unity of apperception,™™ it is independent of the actual
application of the categories in the forming of judgments ( ±µ±).
What Kant is concerned with here is a perceptual synthesis that is
independent of actual perceptual judgment, but guided through its cogni-
tive signi¬cance for us by the possibility of judgment. ˜˜Figural synthesis™™
or synthesis speciosa ( ±µ±) is the construction of shapes and sizes in
productive imagination. There is a double potential for paradox about
the idea of ¬gural synthesis. First, it is surprising to ¬nd our conscious-
ness of our inner states and even self-knowledge linked to the construc-
tion of shapes, and, second, it is odder still to connect self-knowledge
with a priori construction of shapes. Kant™s motivation for these two
claims has to do with his need to make a place for subjective experience
in the kind of objective experience that we can have with respect to
spatial objects.
The identi¬cation of the process through which we conceptualize
perceptual information with the tracing of a priori structures in space
points up the way empirical perceptions of particular shapes in space
depend on a priori constraints on how spaces may be connected. This is
relevant to self-knowledge because of the way in which our beliefs
depend for their content on what we perceive outside of ourselves.
Figural synthesis is involved even in our beliefs about our inner states,
since Kant wishes to argue that all our beliefs are ultimately about
external objects in space. Even the truths of logic and mathematics are
truths only because they tell us about possibilities that constrain the
existence of things in space. If, as Kant claims, the content of even our
beliefs about our own inner states is parasitic on objects outside of us in
space that are intersubjectively accessible, and if, as he also claims, our
very ability to order those inner states in a determinate way is parasitic
on the existence of a determinate order in objects that exist outside of us
in space, then even objects that we represent in a subjective fashion will
turn out to be available in objective terms. We will be able to have
The unity of intuition: completing the B-Deduction
knowledge of our own inner states, albeit not under the description
under which they immediately present themselves to our minds.
A subject of experience can only become a determinate object of
knowledge insofar as it is able to grasp itself as something that is
characterizable by more than the bare idea of having a temporal
perspective. In order to provide an adequate characterization of oneself
as an object of knowledge, one must also be able to characterize this
temporal perspective from outside of it. One must grasp it as a stand-
point within a certain system of standpoints. Thus, the only way we have
of locating ourselves in time is parasitic on our ability to locate ourselves
spatially. Space provides us with the only conception we have of the way
in which individuals exist outside of each other. Our knowledge of the
duration and temporal position of introspective objects depends on our
knowledge of the changes in external objects (section ,  ±µ). Indeed,
Kant wants to argue that the very successiveness of inner episodes is
based on our ability to connect di¬erent spaces together:
Motion, as an act of the subject (not as a determination of an object), hence the
synthesis of the manifold in time, when we abstract from it [space], and merely
pay attention to the act through which we determine inner sense according to
its form, actually generates the concept of succession. ( ±µµ)

Kant interprets the idea that self-knowledge is based on ¬gural
synthesis by means of the claim that self-a¬ection is parasitic on outer
a¬ection. Self-a¬ection is the way in which we a¬ect ourselves when we
have an experience of our inner states. Self-a¬ection may also be
described as the process of connecting the nows of consecutive aware-
ness in a second-order consciousness of one™s identity through those
successive states ( ±µµ). There is an obvious disanalogy between the
outer a¬ection that supplies us with information about external objects
and self-a¬ection.± Inner or self-a¬ection is the way in which we
connect bits of information that are already part of consciousness (inner
sense). Self-a¬ection allows us to interpret who we are as well as what
the objects are that exist outside of us. By contrast, outer a¬ection is
supposed to be the source of information about as yet uninterpreted
(undetermined) appearances ( ±“°/ “), and self-a¬ection is
about the interpreting (determining) of these appearances. The analogy
between inner and outer a¬ection seems to depend on an equivocation
in the notion of determination. Outer a¬ection determines the content
of our beliefs by providing us with speci¬c information, while inner
a¬ection determines the content of our beliefs by interpreting that
±° Kant and the demands of self-consciousness
information in a speci¬c way. The problem is less serious than it seems
to be. Kant does not claim that we are to understand self-a¬ection on
the model of outer a¬ection. These are, in fact, simply two di¬erent
ways in which we are determined. There is indeed a connection be-
tween self-a¬ection and outer a¬ection. But this connection is based on
the thesis that self-interpretation depends on the interpretation of what
is outside of us. The interpretation of objects outside of us does depend
on information about objects inside of us, as well, but the argument here
is based on the relation between self-interpretation and interpretation of
what is outside of us. Working from the assumption that outer a¬ection
is being a¬ected by external objects that are mere appearances, Kant
attempts to show that self-a¬ection must concern objects merely as they
appear to us to be. The content of what we experience when we have
determinate representations of our inner states is based on our represen-
tations of what exists outside of us, and what exists outside of us is itself
represented by us only as it must appear to us to be. From this he
concludes that we have self-knowledge only of the way in which we must
appear to ourselves. We do not have knowledge of how we really are.
To meet the worry about incoherence posed by his idealist interpreta-
tion of self-knowledge, in section µ Kant attempts to establish a sense in
which one can be conscious of oneself without knowing who one is. If I
am conscious of myself then it is pragmatically necessary that I exist,
whoever I might be. This existence is not restricted to how I appear to
myself or to anyone. In thinking about myself I am thinking about
someone and hence referring to a particular individual. However, I do
not know who that individual is unless I know some further self-locating
facts about myself. In self-consciousness I am conscious of someone who
is a basic rather than a dependent particular, since I am conscious of
whoever it is who is the bearer of that self-consciousness. Thus, in
self-consciousness I am conscious of myself as a thing in itself in the
transcendental sense, although any description I have of myself will
apply only to the way I must appear to myself spatially and temporally.

  °©  ©µ °  ¬ 
It is now time to see how one can meet the objection raised by Pistorius
by looking at a contemporary version of it articulated by Robert Howell.
Howell argues that Kant gets trapped between the necessity claims he
wants to make based on de dicto properties of self-consciousness and the
need for self-knowledge to anchor those claims. One must be in a
The unity of intuition: completing the B-Deduction
position to be conscious that this representation is a representation of
(belonging to) the entity that one in fact is. According to Howell, Kant is
then seduced into making con¬‚icting demands on self-consciousness not
only by the deceptive surface grammar of his claims, but because he
runs together ˜˜the traditional theory of self-awareness™™ with a revision-
ary theory of knowledge.±µ Kant™s theory of knowledge commits him to
restricting knowledge in general, and self-knowledge in particular, to
appearances. A conception of knowledge which is restricted to ap-
pearances will not support the traditional conception of self-knowledge
and self-awareness. It will not support knowledge of the subject de re.±
On the other hand, such de re knowledge is needed, if self-ascriptions to
particular selves are to have any purchase a priori. We need to know
that we do in fact have knowledge, but ˜˜if we are to know ourselves
really to have knowledge, then that entity “ our self “ to which all
appearances appear must be known to be real.™™±· But, according to
Howell, we can only know this by violating the strictures on knowledge
associated with Kant™s claim that we can know only appearances.
Howell™s worry is not compelling. For the bearer of representations
that we know to have knowledge need not be known by us under the
description under which it is the ultimate bearer of thought. We may
know that the bearer of self-reference is real insofar as self-reference
could not take place without something which is real. This does not
imply that we have any knowledge of the speci¬c character of that
bearer. It does imply, according to Howell, that ˜˜one veridically grasps
the fact that the self in itself really exists.™™± This conclusion is, however,
premature. While it is true that there must be some bearer for one™s
representations, it simply does not follow that this bearer must be
characterized in itself as a self.

 ¬ ¦- « ®  ·¬ ¤ §   ® ¤    µ   “         
Kant treats self-knowledge on the model of a subject that represents
itself as an object: ˜˜I as intelligence and thinking subject know myself as
thought object™™ (section ,  ±µµ). This has suggested to many inter-
preters that empirical self-consciousness or even transcendental self-
consciousness is to be understood on the subject“object model, or the
re¬‚ection theory as it is generally called.± According to the re¬‚ection
theory, self-consciousness is capable of a reductive analysis into a two-
termed relation between the subject of consciousness and the object of
consciousness. It should be noted that Kant does not seem to use the
±° Kant and the demands of self-consciousness
term re¬‚ection to characterize the activity of self-consciousness in the
way described by the re¬‚ection theory.° According to the section on the
Amphiboly of Concepts of Re¬‚ection, re¬‚ection ˜˜is the consciousness of
the relationship of our di¬erent sources of cognition through which
alone their relationship to one another can be correctly determined™™ (
±/ ±). Re¬‚ection thus refers to the capacity to distinguish and
properly connect the contributions of the di¬erent faculties of represen-
tation. In order for one to be able to evaluate the contributions of the
various faculties of cognition, one must be able to see them as contribu-
tions to cognition that one makes as the subject of cognition. But there is
no obvious reason that re¬‚ection of this kind has to be construed as an
identi¬cation of a subject of consciousness with an object of conscious-
Those who have interpreted Kant as a defender of the subject“object
model and re¬‚ection theory of self-consciousness have come to the
conclusion that Kant™s conception of either empirical or transcendental
self-consciousness is incoherent. While I shall argue that Kant did not
hold a subject“object or re¬‚ection theory of self-consciousness, I do
think that he held a re¬‚ection theory of self-knowledge. The re¬‚ection
theory of self-knowledge is defensible so long as it is based on a non-
reductive theory of self-consciousness. The subject“object schema ap-
plies to self-knowledge because self-knowledge is constrained by criteria
governing the recognition of oneself as an individual person distinct
from other persons. The criteria for identifying and reidentifying per-
sons are parasitic on the criteria for identifying and reidentifying ma-
terial bodies, since the only way we have of identifying and reidentifying
di¬erent times is in relation to material objects that occupy spaces.
However, since self-knowledge requires self-identi¬cation, the object to
be identi¬ed under a certain description must not only be identi¬ed in
spatio-temporal terms but also be thought of as oneself. It is here that the
subject-model needs supplementation if it is to provide a coherent
account of self-knowledge.
If we try to extend the subject“object model to self-consciousness we
get the following paradox. The self must be able to identify itself (subject
= object) in order to be conscious of itself, but in order to identify itself it
must already have some form of knowledge of itself as a thinking subject
and object of that thinking subject™s thought. Self-knowledge involves
knowledge of the identity of the knower qua subject of self-consciousness
and the known qua object of self-consciousness (and subject of the
conscious state of which one is conscious in self-consciousness). But this
The unity of intuition: completing the B-Deduction
knowledge of the identity of knower and the known actually presup-
poses an immediate re¬‚exive awareness of self. The subject“object
re¬‚ection theory already tacitly presupposes the self-conscious aware-
ness it attempts to explain. This is a fatal ¬‚aw if the subject“object theory
of self-knowledge is extended to include self-consciousness, or thought
to be a stand-alone theory.
The subject cannot be thought of as its own object without already
having some direct access to itself. Kant has two bases for such direct
access. As thinkers, we have an immediate representation of an imperso-
nal self, and, as sentient creatures, we have an immediate spatio-
temporal representation of ourselves. Kant sometimes seems to suggest
that there is no empirical subject or self at all. But he also carefully
distinguishes the I which thinks from the I or self-consciousness which
intuits itself (and therefore has data concerning) itself (section ,  ±µµ).
The empirical aspect of the self does not reduce the self to a mere object
of re¬‚ection. In fact, being an object of such re¬‚ection, the possibility of
self-knowledge depends on the ability to refer to oneself as having a
distinctive point of view. Empirical apperception must express a par-
ticular point of view with respect to experience from within experience
without necessarily involving re¬‚ection on oneself as an object of knowl-
edge. This is the force of the remark concerning self-intuition. Such
self-presentation from a particular point of view within experience is
what makes my representations mine, as opposed to yours. In this way,
Kant can explain how one can represent oneself as object. The directly
self-referential aspect of self-consciousness needs to be extended to
include a self-descriptive aspect. Self-description requires experience.
Although self-reference is possible without self-description, no self-
description is possible without self-reference. This descriptive aspect
distinguishing empirical self-consciousness is expressed in the Kantian
doctrine of self-intuition and self-a¬ection.
There are strong reasons to resist interpreting the direct access of the
self as representer to the self as represented on the subject“object model.
On the other hand, self-consciousness appears to presuppose a certain
amount of self-knowledge. The knowledge that I (who am the subject of
self-consciousness) am the subject of the consciousness I am ascribing to
myself is presupposed in self-consciousness. Such self-knowledge might
seem to preclude a sharp distinction between self-consciousness and
self-knowledge. A question arises as to how the self can know or be
conscious of itself, without already being conscious of the fact that it is of
itself that it is conscious. Indeed, we must already think of our experien-
±° Kant and the demands of self-consciousness
ces as re¬‚exively self-referential in order to make sense of the possibility
of self-consciousness with respect to them. Re¬‚exive self-referring to x
involves not only the use of a term that actually refers to x, but reference
to the referee doing the referring as internal to the act of referring.±
Ascription by me of representations to my own particular conscious-
ness must conform to the conditions under which access to any self-
consciousness is possible. But, in self-ascription to my consciousness,
reference to the fact that I am thinking the proposition in question is an
essential part of the statement. In such self-ascription of states, I claim
that this is the way representations are connected in my consciousness,
as opposed to someone else™s consciousness. As Kant sees it, intersubjec-
tivity and objectivity are attained when such self-reference is no longer
relevant to the empirical truth of a statement. Then we have what he
would call a judgment (or, more perspicuously, a judgment of experi-
ence). Looking at things in this way helps to resolve the paradoxical
status of self-knowledge in the B-Deduction. Statements expressing
one™s own propositional attitudes cannot be instances of self-knowledge
for Kant, because they involve an essential reference to the context in
which they are formulated. It is, however, possible to form judgments
concerning the having of such propositional attitudes. Such thoughts
are judgments because there is no essential reference to the point of view
of the empirical consciousness in which the thoughts in question occur.
We must therefore distinguish three di¬erent things: (±) consciousness
of myself as the potential subject of any of a potential in¬nity of di¬erent
representational contents, () empirical consciousness of inner states in
which I associate one representation of mine with another, and ()
knowledge of myself as a particular empirical individual (based on
evidence accessible to a third-person point of view). In pure appercep-


. 3
( 9)