. 7
( 9)


to this doctrine of transcendental idealism that I now turn in closing my
general argument.

 °©  ©   ¬    ¬©    ® ¤    ®    ® ¤ ®   ¬ © ¤ ¬ © 
In the Transcendental Aesthetic, Kant maintains that space and time
are unavoidable representations. He makes the point clearest with
respect to space: ˜˜we cannot represent to ourselves the non-existence of
space.™™ He maintains that the existence of space and time, while not
logically necessary, is necessary to any experience that is intelligible to
us, and necessary to any of our e¬orts to distinguish particular objects.
Kant was inclined to draw far-reaching conclusions from his thesis that
there is a distinctive non-logical and non-conceptual necessity involved
in our capacity to distinguish inner from outer within experience. He
saw the intuitive necessity in question as the basis for distinguishing what
belongs to our experience from the way things are independently of our
experience. This gives rise to what he calls a transcendental distinction
between the inner and the outer, where everything that we experience is
to be regarded as inner relative to what is completely independent of the
sensible pre-conditions governing our experience.
Kant™s conception of transcendental philosophy as an analysis of the
pre-conditions of experience is closely associated with the thesis that we
can know theoretically only what appears to us in accordance with the
± Kant and the demands of self-consciousness
sensible pre-conditions of experience ( ©). He refers to this thesis as
transcendental idealism, which he also glosses as the claim that objects
of experience are inherently subjective. Here subjective does not mean
private, but rather dependent on absolutely general conditions govern-
ing at least all human experience. Transcendental idealism is the thesis
that ˜˜such properties which belong to things in themselves can never be
given to us through the senses™™ ( / µ), where we are to understand
things in themselves as things in themselves in the transcendental, and
not the empirical sense. Here the senses refers not only to the ¬ve senses,
but also to the a priori forms of our sensibility. This is thus not the
traditional claim of representational realism that we do not perceive the
primary qualities of things, such as mass, force, wavelength, and exten-
sion, while we do perceive secondary qualities such as heaviness,
warmth, color, smell, taste, and sound. The traditional distinction
between secondary and primary qualities corresponds roughly to the
distinction between appearances and things in themselves within experi-
ence, that is to the empirical distinction between appearances and things
in themselves. The transcendental claim is rather that our notion of
mass, wavelength, and force are themselves restricted to the way things
must appear to us. They presuppose the notions of extension, space, and
time, and these notions in turn depend on our a priori forms of
Human beings can have sensory disabilities that make them unable to
experience certain kinds of sensations and hence certain secondary
qualities. The properties that objects have in virtue of being represented
by us in terms of the speci¬c make up of our sensory apparatus are
contingent. They are not representable by all human beings ( “).
The way things appear to us from a certain position in space and time is
also contingent. Thus an appearance in the empirical sense is the way
something looks, tastes, sounds, feels, or is experienced, from a particu-
lar spatio-temporal standpoint. By contrast, the properties that empiri-
cal objects have in themselves are those which ˜˜in universal experience
among all di¬erent positions relative to the senses, are determined thus
and in no other way in intuition™™ ( µ/ ). These standpoint-
independent properties that objects of experience must have for all
observers regardless of their standpoint in space and time are not ones
that can be directly perceived through the senses, since sense perception
is inherently perspectival. Empirical objects regarded in themselves
have a spatial and temporal position that is the same for all observers.
They also have standpoint-independent properties that allow us to
Empirical realism and transcendental idealism
locate them in a space and time that is the same for all observers, even
though they appear di¬erently to di¬erent observers depending on the
spatio-temporal standpoint and psycho-physiological make up of those
The transcendental distinction between appearances and things in
themselves is a distinction we draw between the ways in which we know
objects a priori ( ±/ µ). The only properties of objects that are
available to us are ones that are essentially spatial or temporal. We
would not know what the objects or their properties would be like in
abstraction from space and time. The key thesis of transcendental
idealism is that even the standpoint-independent spatial and temporal
properties and relations that characterize objects of experience are
themselves dependent on the way absolutely all human beings must
experience objects, and have no existence that is independent of the
forms according to which we and creatures relevantly like us must
experience the world. Thus, from a transcendental point of view, even
the seemingly standpoint-independent properties of objects turn out to
be dependent on the standpoint that we must occupy as human experi-
encers. Things taken in themselves in the transcendental sense would by
contrast be things that are radically independent of the way we as
human beings must experience the world.
The claim that space and time are ideal when things are regarded
from the transcendental point of view is a function of the fact that we
represent space and time by means of the forms of our sensibility. Space
and time cannot be represented as features that belong to things when
they are re¬‚ected on by reason alone in abstraction from the sensible
conditions that govern our experience, for our concepts of space and
time are supposed to be parasitic on our capacity to represent objects in
terms of the forms of our sensibility:
Our expositions teach therefore the reality (i.e. the objective validity) of space in
respect to everything that can occur to us as an external object, but also the
ideality of space in respect to things, insofar as they are regarded by reason in
themselves, i.e. without regard to the character of our sensibility. ( )±

The a¬rmation of the reality of space and time with respect to all
objects of our experience, the assertion of what Kant calls their objective
reality, is connected with the denial of their absolute reality. For it is
precisely because space and time are necessary structures of the mind that
spatial and temporal objects must have observer-independent properties
that, in principle, are cognitively accessible to us. To the extent that
° Kant and the demands of self-consciousness
objects in space and time are themselves parts of space and time, they
must also be conceived of as objects that cannot be represented by reason
alone. Kant wants to argue that they cannot have absolute reality to the
extent that they cannot be represented by reason alone. If we identify
absolute reality with being an object of reason alone, then we can
understand the thesis that spatio-temporal objects are mere appearances
for us, but are not represented by us as things as they exist in themselves.
But still we are inclined to wonder why we should think that things with
absolute reality can only be objects accessible to reason alone.

 ¤©   §µ ®  ¦   ®  ®¤ ® ¬ © ¤¬ ©
Kant presents the argument for the transcendental ideality of space and
time as a conclusion from the arguments developed in the Metaphysical
Expositions of Space and Time ( “µ/ “°;  °“/ “), as
well as the Transcendental Expositions of Space and Time ( °“±;
“). He thus assumes that once one has accepted these arguments
one will also accept transcendental idealism ( / ;  µ/ µ). It is
in the Conclusions from these Concepts (referring to the concepts of
space and time developed in the Metaphysical Exposition) and the
Transcendental Exposition that Kant explicitly commits himself to the
mere subjectivity of space and time ( / ;  / ).
The Metaphysical Exposition defends Kant™s thesis that space and
time are a priori intuitions, because space and time are immediate and
non-conceptual representations of a form according to which we must
order objects. The Transcendental Exposition is an addition of the
second edition which shifts the discussion of geometry out of the Meta-
physical Exposition, develops it in more detail, and articulates the
dependence of change and hence the laws of motion on the structure of
Kant articulates his worries about the legitimacy of the idea that
space and time might exist completely independently of the mind by
attacking the absolute and relational theories of space and time as
incoherent. Following Clarke™s exposition of the absolute theory in his
famous exchange with Leibniz on the nature of space and time, Kant
takes the adherent of an absolute theory of space and time to regard
space and time as properties of things in themselves (in Clarke™s theory,
properties of God), whereas, for the relationalist, space and time are
relations between things in themselves. From the premise derived from
the Metaphysical Exposition that space is an a priori intuition, together
Empirical realism and transcendental idealism
with the premise that one cannot intuit determinations a priori and
hence prior to the existence of things in themselves, Kant then infers
that space is not an absolute or relative determination applying to things
as they are in themselves and thus, by implication, that neither the
absolute nor the relational theory of space could be true.

Space does not represent any property of things in themselves, nor does it
represent them in their relation to one another. That is, space does not
represent any determination that attaches to the objects themselves, and which
remains even when abstraction has been made from all the subjective condi-
tions of intuition. For no determinations, whether absolute or relative, can be
intuited prior to the existence of things to which they belong, and none can be
intuited a priori. ( / µ)

The expression ˜˜for™™ in the ¬nal sentence indicates that the claim
that one cannot intuit a priori determinations prior to the existence of
things is being appealed to as a premise in the argument. The non-
spatiality of things in themselves is supposed to follow from the nature of
space as an a priori intuition. Kant seems to place weight on the idea
that neither space nor time could be properties or relations that depend
for their existence on things in themselves. He talks of them as being
determinations that cannot be ˜˜fastened™™ (˜˜anhaften™™  / ) or
˜˜attached™™ (˜˜anhangen™™  / ) to things in themselves. The impli-
cation is that space and time cannot be a priori conditions while also
being properties or relations that owe their existence to the things in
themselves that have them. For Kant™s notion of the a priori requires
that these properties or relations be necessary to all of the objects that
have them. Clarke would not have wanted to say that space or time are
necessary properties of God, and Leibniz would not have wanted to say
that space or time are necessary relations of the things which induce
them. This argument thus has some historical force. Kant also antici-
pates the possibility of thinking of time as something that exists for itself,
but rejects this possibility because he thinks that time cannot be an
object in its own right and so would have to be real without any real
object ( / ). This seems to be based on the then universally held,
but now widely rejected, assumption that time could not be a kind of
thing or object with dynamic properties of its own.
Even if one accepts the arguments of the Metaphysical Exposition,
one is entitled to be skeptical of the premise that one cannot represent
absolute or relative determinations of things in themselves prior to the
existence of those things. Kant™s claim that one cannot intuit determina-
 Kant and the demands of self-consciousness
tions prior to the existence of things seems to rule out an a priori
intuition of the structure that any objects have. It thus seems to preclude
us from having an a priori intuition of objects that we experience
(appearances) as well. Since logical priority is part of the meaning of
being a priori, a priori intuition must involve an immediate representa-
tion of objects that is logically prior to the existence of those objects.
Kant must mean that a thing that has properties or relations that we can
intuit a priori would not be a thing in the absolute sense. It would not be
fully independent of the mind.
Kant™s line of thought seems to go as follows: things as they exist in
themselves, in principle, could have spatio-temporal features that are
necessary to either their possible or their actual existence. But they
would not have such features in virtue of being represented by us
according to a certain intuitive form. An intuitive form is distinguishable
from a logical form precisely because its non-existence entails no contra-
diction. This is why we cannot prove that there cannot be other forms of
intuition than the ones which human beings must have ( ·). Since
there is no contradiction in the non-existence of a form of intuition, its
existence is metaphysically contingent. This metaphysical contingency
of our form of intuition is connected with its capacity to present us with
objects that actually exist, as opposed to mere conceptual connections.
Something whose very existence depends on the contingent fact that
there are representers like us who must represent the world in a certain
way cannot be a thing which exists in itself because it cannot be a fully
independent thing. It is, rather, an essentially relational thing which
cannot exist independently of being representable by us. Spatio-tem-
poral properties are essentially relational in the sense that we cannot
form a concept of them independently of the way objects must intuitively
appear to us. One might object that, while some objects may be necessar-
ily spatio-temporal in the sense of being objects of a priori intuition,
this need not preclude other objects from being spatio-temporal in the
sense that they fall under concepts of space and time. But to a¬rm this
possibility would be just to deny Kant™s claim that our concepts of
space and time necessarily depend on a priori intuition. The only
concepts of space and time which we have are ones that are (putatively)
based on a priori intuition. There are then no concepts of spaces and
times in terms of which we could make sense of spaces and times that
are completely independent of us. Thus we cannot make sense of
spaces and times which conform to the way things appear to us only
Empirical realism and transcendental idealism
The argument for transcendental idealism in the Transcendental
Exposition provides support for this reading. From the assumption that
we have a priori knowledge of geometry, it is inferred the objects of
geometry are necessarily mind-dependent ( ). For it is said that an
external intuition that precedes its object can only belong to the mind
and determine the concept of an object a priori, if it exists only in the
subject. Kant demands that this intuition be the formal condition for
being a¬ected by objects and thus an immediate representation of them. The
same argument is developed more explicitly at the end of the Transcen-
dental Aesthetic in section  ( ¬./ ¬.). First Kant argues that
geometry is synthetic. He notes that no purely conceptual analysis can
show the impossibility of a closed geometric ¬gure which consists of two
straight lines. He insists, more tendentiously, that for this one has to go
to intuition. Such intuition must be a priori if it is to justify propositions
that are necessary and strictly general. He then argues that geometry
can only be regarded as objective, as applying to the objects of geometry
in themselves, if these objects cannot also exist independently of being
intuited by us. The only way we have (according to Kant) of represen-
ting the objects of geometry is by means of the particular form of
externality which forces itself upon us. This form of externality is
assumed by him to be Euclidean and three-dimensional in structure,
although this assumption depends on his views about the possibility of
constructing mathematical objects in space. The key assumption, how-
ever, is that the only representation one could have of objects with
spatial (or temporal) properties is the one which is provided by a priori
intuition. Kant was convinced that, if the objects which we know as thus
and such with necessity and strict generality could exist in abstraction
from the conditions imposed on them by our experience, then all our
claims to make objective assertions about these objects are in doubt.

Now if a faculty of intuiting a priori did not lie in you; if this subjective condition
were not also as regards its form the general condition a priori under which
alone an object of this (external) intuition is possible; if the object (the triangle)
were something in itself without relation to your subject: how could you say that
what lies necessarily in the subjective conditions for constructing a triangle,
must also apply to the triangle itself? ( /  µ)

Only by taking space and time to be inherently subjective can we be
certain that the claims that we make on the basis of how possible objects
must appear to us spatio-temporally are in fact objectively true of
spatio-temporal objects. It might be thought that Kant™s claims depend
 Kant and the demands of self-consciousness
on the general nature of necessity claims. This is the way Kant™s
argument is understood by Paul Guyer with whose reconstruction I am
otherwise in considerable sympathy. But then the implication would
seem to be that the notion of a mind-independent object would be either
completely vacuous or incoherent. For any conceivable object must
conform to the demands of logic. This possibility can be avoided if we
take the argument to depend on the kind of non-logical necessity
provided by intuition.

  ® ¤ ¬  ®  µ  § ¬   °  ¬
There appears to be a gap in Kant™s argument, generally known as the
Trendelenburg loophole after the nineteenth-century German philos-
opher who ¬rst brought it to the general consciousness of the philosophi-
cal public. It is not obvious why objects as they are intuited by us
spatially and temporally cannot correspond to things in themselves that
are themselves spatio-temporal. Henry Allison has done the most to
rehabilitate Kant™s transcendental idealism and is responsible for re-
vived interest in the Trendelenburg loophole. Allison appeals to his
principle of formal idealism as the key to Kant™s argument against the
possibility that the absolute or relational theories of space and time can
account for the function of space and time as forms or conditions of
human experience.µ The purported principle of formal idealism forms
the basis for what Allison takes to be Kant™s strategy for dealing with
Newton™s and Clarke™s theories of absolute space and time as well as
Leibniz™s relational theory of absolute space and time: ˜˜regarding space
as an ontological condition (is) incompatible with also regarding it as an
epistemic condition.™™
However, in fairness to Allison, his speci¬c argument for the tran-
scendental ideality of space and time is supposed to depend on a priori
conditions of the mind which are distinctively sensible. In transcenden-
tal re¬‚ection, we may think of something in terms of the forms according
to which objects are given to us (the forms of our sensibility), or in terms
of the forms according to which we must interpret objects (the forms of
our thought). Something which merely appears to us is something which
we represent according to the forms of our sensibility, whereas some-
thing which exists in itself is something which can be understood
without reference to the forms of our sensibility. Unfortunately, when
Allison tries to close Trendelenburg™s loophole, he uses the general
notion of a form of representation. The speci¬cally sensible or intuitive
Empirical realism and transcendental idealism
character of our representation of space and time plays no obvious role
in his argument.
There may be things that are logically possible, which we cannot
make sense of in terms of the speci¬c ways in which objects are given to
us by our forms of sensibility. This raises the question whether some-
thing which we represent according to our forms of sensibility cannot
also be understood without reference to those forms of sensibility.
Allison considers two possibilities. There might be objects that are
independent of our forms of sensibility but numerically identical with
the objects which must conform to the a priori forms of our sensibility.
Alternatively, there might be objects that are qualitatively, but not
numerically, identical with the objects which must conform to our forms
of sensibility. Allison dismisses the possibility that space or time might be
numerically identical with what is independent of the mind, although
the double-aspect view of transcendental idealism, which he defends,
requires that one and the same thing be available under radically
di¬erent aspects or descriptions, indeed as a mere appearance and as
something that may be represented as it is in itself. He then argues that it
would be incoherent or at least utterly vacuous to maintain that mind-
dependent features could be qualitatively identical with mind-indepen-
dent features of things. Since the notion of space and time as forms of
intuition entails that what exists in these forms is something represented
or mind-dependent, the notion of things in themselves having intuition-
independent and hence mind-independent features is meaningless.·
Unfortunately, this line of thought would seem to lead one to a form of
solipsistic phenomenalism that is inconsistent with Kant™s empirical
realism. For we could raise the same objection with respect to any
purportedly mind-independent features. Allison could appeal for sup-
port to a remark from Kant in which he distances himself from Ber-
keley™s idealism:

I should be glad to know what my assertions must be in order to avoid all
idealism. Undoubtedly, I should say that the representation of space is not only
conformable to the relation which our sensibility has to objects “ that I have
said “ but that it is completely like the object “ an assertion in which I can ¬nd
no meaning anymore than if I had said that the sensation of red has a similarity
to the property of cinnabar which excites this sensation in me. (Prolegomena, Ak.
©, section ±, pp. “°)

Kant is assuming the thesis of representational realism that empirical
objects (objects as they are to be regarded in themselves from the
 Kant and the demands of self-consciousness
empirical standpoint) are not colored. Colors are mind-dependent sec-
ondary qualities, to which there are corresponding dispositional pri-
mary qualities in physical objects. The analogy to the relation between
primary and secondary qualities to which Kant himself appeals in
rejecting any relation of similarity between space and its objects suggests
a plausible alternative account of the relation between space and time as
features of things in themselves and space and time as forms of what is
represented. The sensation of red and cinnabar are not similar with
respect to the predicate red according to Kant. While the distinction
between primary and secondary qualities does not strictly preserve
qualitative identity, it does provide us with an analogical use of predi-
cates which is neither vacuous nor incoherent. For we can speak of a red
sensation, and a red sample of cinnabar, even though the sensation is
not red in the same sense that the cinnabar is. It is di¬cult to understand
how there could be a coherent use of analogical predicates across the
empirical distinction between appearances and things in themselves
while the use of analogy across the transcendental distinction would be
illicit. Thus this passage gives some support to Allison™s reading of Kant,
but not to the substantive claim he wants to make.
There is a deeper di¬culty with Allison™s general line of argument, as
well. For it rules out the possibility that things in themselves could be
grasped in purely conceptual terms as well as in terms of our intuitions of
spatio-temporal objects. If it never makes sense to say that an object can
have the same features when represented as when it is understood in
abstraction from being represented, then the whole notion of a thing as
it is in itself threatens to become incoherent. Thus we could say, if space
and time are forms of what is represented in intuition, things in them-
selves cannot coherently be represented spatio-temporally. But this
would only be because things in themselves cannot be coherently
represented at all. For this would be to represent something via a form of
representing that is de¬ned as independent of any form of representing.
What is missing in Allison™s account is a way of distinguishing the
necessary mind-dependentness of objects subject to our forms of intu-
ition from the mind-independence of objects that are subject only to
forms of understanding. He tries to provide such an account by distin-
guishing logic from epistemic conditions, but the consequence of that
view seems to be that we then have no representation of a thing in itself
at all.
More recently, Allison has come to accept much of the force of this
objection to his position. While he earlier rejected Jill Buroker™s
Empirical realism and transcendental idealism
argument for transcendental idealism from right- and left-handed ob-
jects (incongruent counterparts) as the basis of Kant™s idealism, he now
concedes that a generalized version of this idea is just what Kant needs.
For he now argues that something like Lorne Falkenstein™s distinction
between a presentational order for spatial and temporal properties of
the kind provided by intuition and a comparative order for properties
that can be grasped purely conceptually would be required in order to
establish Kant™s point that there is a sharp distinction between ap-
pearances and things in themselves.±°
Falkenstein argues that if space and time were properties or relations
of things as they exist in themselves, they would have to be based on a
comparative order of internal properties such as that characteristic of
the quality ˜˜space™™ of our sensations rather than the presentational
order which Kant believes to be characteristic of the form of objects
represented by us spatially (and temporally). The notion of a thing in
itself is taken to be de¬ned as allowing only relational properties that are
reducible to non-relational properties. Objects that appear to us spatio-
temporally would be the kind of objects that we must grasp in terms of
an essentially relational, comparative order (right, left, earlier, later),
while things in themselves would be the kind of things that can be
grasped in a non-relational, purely conceptual manner.
Such a view seems clearly to have played a role in Kant™s transcen-
dental idealism, as his appeal to incongruent counterparts in support of
his idealism indicates.±± However, the idea that distinctions between
right, left, earlier, later cannot be exhaustively expressed in terms of
purely conceptual relations seems to be on rather shaky ground, since
they can be given a rigorous mathematical formulation. And, even if we
accept the distinction between a comparative and conceptual order, we
must still demonstrate that only a comparative order of internal proper-
ties can exist independently of our sensibility. Otherwise, Kant, at best,
will have established that we can only conceive of a presentational order
by appeal to our sensibility, he will not have established the stronger
claim that space and time cannot exist independently of our sensibility.
Kant assumes that because space and time are forms of our sensibility
they cannot be forms which also exist independently of the way in which
we order objects. But this claim, that because space and time are forms
of our sensibility they can only be mere forms of our sensibility, raises
precisely the neglected alternative suggestion made by Trendelenburg.
Perhaps the most direct approach to the loophole, that taken by
Guyer, has a di¬erent, although related di¬culty. Guyer argues that the
 Kant and the demands of self-consciousness
loophole is closed by Kant™s assumption that the necessity and universal-
ity involved in the a priori must be coextensive ( ), so that, if something
is known to be necessarily thus and such, it must belong to the same
domain with respect to which we are able to make universal claims.±
This provides an answer as to why Kant would rule out the idea that
space and time might turn out independently of us contingently to have
the features that we necessarily represent them as having; it does
nothing, however, to show why they cannot necessarily have the proper-
ties that we represent them as having independently of us. And it raises
the question as to whether the contrast between the way things must
appear to us and things as they are in themselves is itself intelligible to us.
For the very generality of the (necessary) a priori knowledge that we
have, including, as it does, purely conceptual knowledge as well, seems
then to preclude us from having any notion of an object that is not
subject to necessary constraints.

 © ®§  © ®    ¬   ® ¤ ®  µ ® 
There are a priori conditions for Kant governing even the concept of an
object in general. These a priori conditions are provided by formal
logic. If an object were necessarily mind-dependent in virtue of being
subject to a priori conditions for thought in general, then the very notion
of an object which exists independently of being represented by us
would entail a logical contradiction. But then the distinction between an
object which appears to us and an object as it is in itself would amount to
just the distinction between an intelligible and an unintelligible object.
In this case, the very claim that we can know only appearances but not
things as they are in themselves would collapse into the trivial point that
we cannot know what is unintelligible. Thinking of things as they are in
themselves as unintelligible would con¬‚ict with Kant™s tendency to
identify things as they are in themselves with noumena or intelligible
objects, objects of understanding alone. The problem with such objects
is that they are vacuous for us, not that they are unintelligible. Kant™s
view, by contrast, is that the range of the logical a priori determines
what is intelligible.
Since Kant does not think that there can be alternative logics to the
term logic of his time, he does not think that there can be alternative
logics with more or less equal claims to plausibility. It is thus reasonable
for him to take the a priori of general or, rather, universal, logic
(˜˜allgemeine Logik™™) to determine what can intelligibly be said to be. It
Empirical realism and transcendental idealism
thus has a general criteriological character for ontology. This assign-
ment of a general criteriological function to logic in ontology is reason-
able even from a more contemporary point of view that is bound to be
more skeptical about the neo-Aristotelian and Stoic term logic which
Kant thought to be largely, but not wholly, immune to revision. Some-
thing resembling ¬rst-order quanti¬cation theory with identity seems to
be very close to any adequate way of ¬‚eshing out basic assumptions
governing thought. At the very least, logic is that part of our corpus of
beliefs which we must hold to be most resistant to revision in the face of
It seems reasonable to look to general logic for the concept of an
object that can be thought independently of any sensible intuition. And,
given the completely general character of logic, we might hope that
logical notions would provide us with some representation of a thing
that is independent of sensibility. But here we run up against the limits
of what reason can provide us with. We do have the general idea of an
object that could be grasped by reason without appeal to experience.
Kant introduces the noumenon as a thing ˜˜that ought to be thought of
not at all as an object of the senses, but rather as a thing in itself (merely
through the pure understanding)™™ ( µµ/ ±°). The identi¬cation of
what would be represented by a pure understanding in abstraction
from sensibility with a thing in itself ¬ts his view that it is because space
and time and their objects are given to us by our forms of sensibility that
they cannot be things as they exist in themselves. This notion of a
noumenon is not itself self-contradictory, Kant maintains, since one
cannot say of sensibility that it is the only possible form of intuition. But
he does argue that our concepts are restricted in their meaning and
application to the domain of sensibility. Even general logic can only be
applied to objects if those objects are somehow given to us. This means
for him that a noumenon, or a thing represented as it is in itself, cannot
be an intelligible object for our understanding. We are to think of a
thing in itself as something which could only be thought by a radically
di¬erent kind of intuition than our own, a non-sensible intuition (
µ/ ±). For such an intuition, there is no distinction to be drawn
anymore between the way objects are given and the conditions under
which they can be thought. Indeed, the very notion of objects being
given to such an intuition loses any meaning. Kant admits that the very
possibility of such an intuition is problematic. This leads Kant to
conclude that the notion of a noumenon involves the representation of
a thing of which we can neither say that it is possible nor that it is
° Kant and the demands of self-consciousness
impossible, since we do not have the kind of intuition required to make
sense of it ( ·/ ).
Kant denies in many places that we do or can know whether there
could be objects which are not spatial or temporal. This is the dominant
view in the sections on Phenomena and Noumena ( µµ/ ±°) and the
Amphiboly of Concepts of Re¬‚ection ( ·“··/ “). He thus
leaves the question open whether there could be things which exist as
they are in themselves independently of the way we represent things
spatially and temporally. So interpreted, the thing in itself is a limiting
concept for theoretical reason of which we have no theoretical knowl-
edge. The question of whether things in themselves are spatio-temporal
or not is meaningless, since it requires resources outstripping the en-
abling conditions of meaningful discourse. Space and time cannot be
taken by us to exist independently of the mind because we have no
concept of them that would make their mind-independent existence
intelligible. If we cannot have a concept of space and time that is
independent of the way objects must be given to us by the contingent
fact that we must represent the inner“outer distinction as we do, then we
cannot make sense of the idea that space and time themselves exist
completely independently of us. Kant is clearly right that, even though
space and time may be necessary to the way we experience the world,
we cannot regard their existence as logically necessary. Still, even if this
is true, it is too weak a claim to establish that space and time cannot be
things as they exist in themselves or properties or relations of those
things. The best that Kant can hope for is that he has established that
space and time are not the kind of things that can be regarded by us as
things in themselves or their properties or relations. This interpretation
¬ts Kant™s tendency to refer to the distinction between things in them-
selves and appearances as two ways that we have of regarding things,
rather than as a distinction that applies to things completely indepen-
dently of our re¬‚ection.
Most of our beliefs might turn out to be false, and the best standards
of knowledge we have might turn out to be inadequate by the lights of
the noumena presented to an intellectual intuition. But we do not even
know that the perspective of intellectual intuition is possible. It is not
really intelligible to us and might turn out to be unintelligible in a more
general sense. Here we ¬nd the basis for a modest version of transcen-
dental idealism, according to which there is a way in which together we
must understand and experience the world, although we cannot be
certain that this way of experiencing the world is compelling for abso-
Empirical realism and transcendental idealism
lutely all creatures. This anodyne notion of transcendental idealism
leaves it an open question whether there can be things that are intelli-
gible independently of the conditions of our spatial and temporal repre-
sentations of them. This epistemically modest version of transcendental
idealism is a view which should be very attractive for anyone who thinks
that there can be transcendental arguments in the theory of knowledge
which compel us to think of the world as structured in a certain way.
Those passages in which Kant pushes the idea that the notion of a
noumenon is necessarily empty for us, make it di¬cult to understand
how Kant can think that ˜˜it is indubitably certain and not just possible
or probable™™ that space and time are forms only of appearances rather
than things in themselves ( / ). The epistemically modest notion
of transcendental idealism allows for the logical possibility of something
that is distinct from the way objects must appear to us, and thus provides
us with a purely negative notion of a thing in itself, but Kant often wants
something much stronger:

In fact, when we regard objects of the senses as mere appearances, as is
appropriate, then we admit that a thing in itself is their ground, although we
know it not as it is in itself, but only its appearance, i.e. the way our senses are
a¬ected by an unknown something. The understanding therefore in assuming
appearances also admits the existence :das Dasein9 of things in themselves,
and in this respect we can say that the representation of such beings that are the
ground of appearances, that is, mere beings of the understanding, is not only
admissible, but also unavoidable. (Prolegomena, Ak. ©, section , pp. ±“±µ)

If space and time are merely subjective in contrast to things in them-
selves, then we can conclude from the existence of space and time as
mere subjective forms that there must be things in themselves. For
without the existence of things in themselves there would be no justi¬ca-
tion in claiming that space and time are merely subjective. The notion of
a ˜˜mere appearance™™ requires some form of real distinction between
appearances and things in themselves, and some thing that exists in itself
to which what appears merely appears to be thus and such. To establish
a real distinction between appearances and things in themselves one
would need some knowledge or at least some conception of how things
are or might be in themselves apart from the pre-conditions of our
knowledge. This would commit one to a substantive concept of a thing
as it exists in itself.
The claim that things as they appear to us cannot be things as they
exist in themselves is the basis for Kant™s rejection of transcendental
 Kant and the demands of self-consciousness
realism ˜˜which regards space and time as something given in them-
selves, independently of our sensibility™™ ( ·°). The contrast between
transcendental realism and idealism thus depends on a distinction
between things in themselves and appearances that seems only to be
articulable by rejecting the very epistemological criterion of reality upon
which that distinction might be founded. Kant maintains that we cannot
have bona ¬de theoretical knowledge of things as they are in themselves,
except in the manner in which they must appear to us. But then the
worry is that we have no way of knowing that they must, indeed, be
distinct from the way that things must appear to us. If we cannot know
that things in themselves must be distinct from the way things must
appear to us, we cannot know that transcendental idealism is true.
There is an inherent ambiguity in Kant™s notion of a thing in itself
even when one disregards what he calls the physical or empirical notion
of a thing in itself. In one sense, a thing in itself is an object of an
intellectual intuition, a kind of intuition which we cannot have as ¬nite
rational beings. In this sense, Kant consistently denies that we can have
a concept or cognition of things in themselves. Even our understanding
of an intellectual intuition is purely negative. It has powers of compre-
hension that our intellect and our intuition do not have. On the other
hand, Kant also thinks of moral properties and the individuals to which
they belong as existing in and of themselves. The Aesthetic, for instance,
contrasts the notion of space, belonging only to appearances, with the
notion of right, that belongs to things in themselves; the contrast with
space makes it clear that Kant does not have in mind the claim that right
is a property of things in themselves in the empirical sense: ˜˜right cannot
appear at all, its concept lies rather in the understanding and represents
a property (the moral one) of actions which belongs to those actions in
themselves™™ ( / ±).
In the Critique, Kant no longer allows for any grasp of things in purely
conceptual terms by theoretical reason, as he seems to in the Dissertation
where he identi¬es purely conceptual knowledge of things with meta-
physics (Ak. ©©, pp. µ“). But he continues to think that, from a
purely practical or moral point of view, we can say that there are real
things that are not spatial and temporal. Thus the passage in the
Aesthetic is very much in the spirit of a similar passage in the Dissertation
where Kant also notes that our moral concepts may be confused, that is,
not such that we could pick out their objects by means of necessary and
su¬cient conditions, but that their objects are known by the pure
intellect itself (Ak. ©©, p. µ). Kant tends to regard the practical notion of
Empirical realism and transcendental idealism
a thing in itself as something that we merely postulate as ultimately real
in order to express the primacy of moral agency. Things in themselves in
this sense are not objects of an intellectual intuition, but rather a way of
looking at the sensible world in abstraction from what is sensibly
determined about it and, hence, as an object of pure practical reason (
°/ ). When we relate to the world in moral terms, we relate to it
in a way which abstracts from the sensible conditions which govern
what must merely appear to us. This purely intellectual conception of
the world also grasps the world as it is in itself in abstraction from
sensibility. But the notion of a thing in itself is not, in this case,
independent of our practices, although it is something substantive.
The assumption that we understand the way the world really is from
the moral point of view has some force. It is hard to ¬nd it completely
compelling, since this primacy of the moral point of view is something
that can only be understood once one has already accepted the moral
point of view. Also, its bearing on most objects of theoretical re¬‚ection is
indirect. Nature is connected to rights and the moral point of view only
indirectly. For instance, when we take a natural object to be something
we regard as property, we are conceiving of nature in the framework of
rights. But some sense can be given to the initially implausible assump-
tion that moral agency must be somehow independent of the spatial and
temporal conditions otherwise governing agency. The categorical im-
perative articulated in Kant™s moral theory spells out a procedure for
universalizing the maxims that we derive from everyday deliberations.
This promises a way of understanding what we are doing that applies
regardless of our spatio-temporal context. To the extent that we adopt
and act according to those universalized maxims solely because of their
universality, we may be said to act from motives that abstract from
spatio-temporal content. Even here there is the problem that moral law
has content only through the way it generalizes from our experience. It
is thus somewhat doubtful that we have succeeded in representing an
action in a way that is genuinely independent of our sensibility. But we
can take this to be a representation of the sensible in terms of purely
intellectual concepts, as Kant does in the Critique.
Kant™s transcendental idealism is most plausible when it presents itself
in its modest form. Otherwise it requires a compelling reason to believe
that the concept of an object which could be grasped in abstraction from
a priori intuition is not altogether empty of real content. Kant has
admittedly an initially plausible motivation for denying spatiality and
temporality to things as they exist in themselves. Our ability to know
 Kant and the demands of self-consciousness
necessary facts about space and time, and the objects that are to be
found in space and time, is based on the way the mind is constrained to
represent the world. If the features that those objects must have in virtue
of the way we must represent the world were features that those objects
might not have independently of being represented by us, then we
would be left with what might seem to be a subjective necessity of the
kind Hume attributed to our fundamental beliefs. We could not be
certain that the structures in terms of which we make the world intelli-
gible to ourselves are not mere projections on to the world of what we
must believe about the world. But this challenge could also be met by
rejecting the subjectivist interpretation of necessity in favor of the idea
that we have no reason to doubt that the necessities in question hold
independently of what we as humans must believe. Kant seems to have
been prevented from endorsing such a move by worries that ascribe
metaphysical necessity to things other than God, but this does not really
seem to be a compelling worry, even for someone with religious beliefs.
Even if the legitimacy of what we must believe about the world is, in
principle, subject to challenge, this challenge will have to come from
inside what we must believe. The worry must be taken seriously that
what we take to be facts might not be genuine facts about anything but
what we must believe. But this worry is much less serious than the more
pressing worry that we will fail altogether to identify anything that we
all, even as human beings, must believe. This is the place where Kant
needs to take a stand against Hume™s notion of psychological necessity.
Genuinely a priori knowledge is enough, if it can be had. We do not
need to worry about the way things might appear to a point of view that
we cannot even make intelligible to ourselves. Thus it seems that Kant™s
concern with how to draw the distinction between what is inner, in the
sense of private, and what is outer, in the sense of public, is ultimately
more fruitful than his attempt to work out how we are to understand the
further distinction between what is inner, in the sense that it is internal
to the experience that we all share in common, and what is outer, in the
sense that it is completely independent of us. However, despite the
reservations that are in order concerning the substantive negative claims
that Kant makes about the nature of things as they are in themselves, he
seems to be right that our conception of the world must provide for both
of these inner“outer distinctions.


I have traced a path that began with the idea that a distinction between
what is internal to one™s point of view and what is external to one™s point
of view is constitutive of the kind of consciousness that we have of
ourselves as ¬nite rational beings. I have argued that the most funda-
mental way in which we distinguish what is internal to our point of view
from what is external to our point of view is by appeal to the idea that we
experience what is internal to our point of view successively, and hence
temporally, while we experience what is external to our point of view in
spatial relations that are only contingently successive. Spatial relations,
and the objects that occupy them are ultimately the basis upon which we
are able to ascribe a determinate position even to the successive repre-
sentations that are constitutive of our individual point of view. At the
same time, we cannot make sense of objects that we experience as
outside of us without representing them temporally.
The idea that experiencing objects spatially and temporally is un-
avoidable for us, indeed the more general idea that the only grip that we
have on existence is in terms of existence in space, or at least time, leads
Kant to claim that space and time are necessary forms of our experi-
ence. The logical possibility that there might be objects that are not
spatial or temporal leads him to argue that space and time constitute
necessary forms of our experience only. In this way, we come to have the
idea of objects that might be outer to us in a very radical way; they might
be outside of the conditions governing our experience. This, in turn,
leads to the conception of objects that we experience as objects that are
inner in a new, transcendental sense; they are internal to our experi-
ence. Kant then, unfortunately, succumbs to the temptation to argue
that these objects that are completely outside of the conditions govern-
ing our experience would have to be neither spatial nor temporal. This
is an idea without much plausibility to it. And, fortunately, it is not a
view that he consistently espouses. Sometimes he opts for the more
 Kant and the demands of self-consciousness
modest view that we cannot make sense of real objects that are not
spatial or temporal. It is more fruitful to think of such non-spatial or
temporal objects as mere logical possibilities, rather than as correct
descriptions of objects as they would have to be completely indepen-
dently of our experience.
We distinguish the inner from the outer within experience, and the
outer as what is completely independent of the way we must experience
the world, on the basis of the consciousness that we have of ourselves.
The consciousness that we have of ourselves has two aspects to it. On
the one hand, we are conscious of ourselves as individuals who have
distinctive experiences that distinguish us from all other individuals,
each of whom has his or her own distinctive point of view. These
experiences are distinguished on the basis of their di¬erent spatial and
temporal content. On the other hand, we are also conscious of ourselves
as individuals who have a point of view that might have been a di¬erent
point of view had our experience gone di¬erently than the way it
actually has gone. This gives us some understanding of what it is to have
a point of view in a more general sense. In the end, our only grasp of our
own distinctive point of view seems to be based on what is distinctive
about it and our ability to grasp that depends on our capacity to grasp
other possible points of view.
The capacity to grasp di¬erent points of view as possible ways in
which one might experience the world is what Kant thinks of as the
capacity for certain representation of one™s numerical identity. One
must be able to represent oneself as identical through di¬erent experien-
ces and representations in a more general sense, for this is an a priori
condition for the ascription of determinate and communicable content
both to our representations and to the representations of any other
sentient being. To regard another creature as sentient, at least in
principle, is to recognize the representations of that creature as ones that
could have belonged to one™s own experience had that experience only
been su¬ciently di¬erent. It is to treat the states of that creature as
potentially connectible together in a system of self-ascribable represen-
tations in which one™s own representations would also have a deter-
mined position. This system of di¬erent self-ascribable representations
is our best way of making sense of the notion of an object of experience
that is independent of my, or even our, particular take on what we
experience. The system in question takes its starting-point from what we
experience, so it is not completely arbitrary, but the only way we have of
getting on in understanding what we are experiencing.
Our grasp of concepts is based on the distinctive ability for self-
conscious abstraction that we have. For concepts are just capacities that
we have to represent things in ways that make sense not only to each of
us individually, but also potentially to each and all of us, collectively. We
develop and exercise our understanding of concepts in terms of our
understanding of a whole battery of concepts that we systematically
apply to our experience. And it is this systematic unity of conceptualized
reality that provides us with the very distinction between a correct and
an incorrect use or application of a given concept in a judgment. It is this
systematic unity in what we experience that thus provides the only way
we have of distinguishing what is true from what merely appears to us to
be thus and such from our own distinctive point of view.
Our capacity to use concepts to interpret what we experience as inner
or outer to us is based on our capacity for self-consciousness. But the
self-consciousness that allows us to do this must be thought of as an
impersonal one. For it is based on the capacity to stand back from or
abstract from the way things are for each of us individually. Such a
subject of thought is the condition for the possibility of the kind of
neutrality with respect to di¬erences in point of view that characterize
our conception of objectivity. The impersonal point of view gives
judgment its target of truth. We conceive of what is true as holding
independently of what each of us may happen to believe or take to be
true. But, in order for this idea of independence to be coherent, we must
assume the possibility of a point of view from which what we happen to
take to be true might turn out to be false. We can only do that to the
extent that we are willing to allow for the possibility of alternative takes
on reality, and for us to do that we must be able to consider the
possibility that our own distinctive take on things might have been, or
could be, a di¬erent one.
The very abstractness of our representation of self, expressed in
tokenings of ˜˜I,™™ makes it capable of representing both the impersonal
point of view and a particular point of view as a point of view that shares
something in common with other points of view, but also di¬ers in some
respect or other from those di¬erent points of view. The impersonal
consciousness of self to which I have been referring is an enabling
condition of personal identity, but it is not to be con¬‚ated with one™s
empirical personal identity. Acknowledgement of the possibility of being
conscious of oneself in such a way that one thinks of oneself as a
consciousness that is yet not identi¬ed with any spatio-temporal location
is an important dimension to our ability to understand and use concepts.
 Kant and the demands of self-consciousness
For the very impersonal way in which one represents the self allows the
impersonal representation of self to serve as a pre-condition of com-
municable experience and communicable representations. In particu-
lar, it underwrites our capacity for judgment and to come up with
concepts that articulate what we experience and otherwise grasp in ways
that we can communicate to others. Such a self-consciousness allows
one to conceive of space and time and the objects of experience in space
and time as a unitary system of permutations on potential standpoints
that I might assume with respect to spatio-temporal experience.
It might seem that only our capacity to form concepts and make
judgments with them that have objective import is based on our capac-
ity for abstractive self-consciousness. But, in fact, not only our grasp of
concepts, but our very grasp of the distinction between inner and outer
and the distinctive temporal and spatial form that it takes for us, is based
on the kind of capacity for abstraction that comes with the capacity for
self-consciousness. For this is a structure that is abstract in the sense that
each of us attributes it to him- or herself, and yet it is also a general
condition under which each of us individually is capable of experiencing
him- or herself, as well as the world. The self can represent itself spatially
and temporally in abstraction from the empirical objects that present
themselves to various particular spatial and temporal locations. This is
to represent oneself as having a point of view, but in abstraction from
the particular context in experience which makes one the distinctive
person one is. Any arbitrary self-conscious individual may think of him-
or herself as existing here and now, and in each case something di¬erent
will present itself to him or to her in accordance with the change in
context within spatial and temporal experience. Thus the abstract
representation of self as the possessor of a spatio-temporal experience
provides the most general notion of the self as an object of self-con-
sciousness which relates to itself as a distinctive object and to other
possible objects as distinguishable from it. The di¬erent objects that
present themselves to di¬erent spatio-temporal subjects, together with
those subjects, each of which may, in turn, be regarded as a spatio-
temporal object, may be ordered in a standpoint-independent way by
each self-conscious individual in a way that is consistent with each of
their di¬erent spatio-temporal perspectives. In order for them to order
episodes in this standpoint-independent way, subjects set up a common
Cartesian co-ordinate system with four di¬erent axes corresponding to
the three dimensions of space and a fourth dimension for time. They
can then assign tenseless relations of earlier, later, simultaneous, and
spatial co-ordinates to all objects. To do this, they, and indeed each one
of us, must be able to apply the concepts of causation, interaction, and
substance to experience, for only if we can legitimately apply these
concepts to all of our experiences will we be able to determine the
temporal relations and positions of all episodes within experience.
The unity of spatio-temporal experience manifests itself in the pos-
sibility of our having determinate beliefs and representations about
spatio-temporal objects. Beliefs about objects experienced in space and
time are determinate because there is a determinate procedure for
con¬rming and discon¬rming (verifying and falsifying) those beliefs. We
can make out something permanent in spatio-temporal experience that
provides the basis for a determinate procedure for con¬rming and
discon¬rming our beliefs. The very notion of an object as something
permanent in experience, or as something that is numerically identical
through space and time, turns out to be a function of our need to assume
a self-conscious standpoint that allows us to regard our experience as an
experience that can be captured and expressed in intersubjectively
communicable concepts.
The communicability of beliefs presupposes the possibility of a con-
sciousness from which we may distinguish ourselves. Our possession of
determinate beliefs also requires the existence of objects existing outside
of us. For only such objects can provide us with the permanence which
we need as a reference-point for determining the content of our beliefs
in some way which promises potential criteria for success and failure in
our self-ascriptions of beliefs and cognitions. Without permanent objects
whose existence is not at the mercy of our immediate representations,
we would fail to have any basis for believing that we had correctly or
incorrectly ascribed a belief or other representation to ourselves or to
someone else.
Kant™s transcendental idealism suggests that we can only give deter-
minate content to our beliefs at the cost of subjecting them to substan-
tive restrictive constraints, such as that every event must have a cause or
that every event must be the state of a permanent substance. These are
constraints on the self-ascription of spatio-temporal representations.
Thus Kant ¬nds himself defending the claim that all non-logical theor-
etical knowledge is restricted to spatio-temporal objects. These objects
might, however, in principle be subject to some other more fundamen-
tal mode of description that would not be spatio-temporal at all. But,
even where Kant seeks to ¬nd such a description in the dimension of
action motivated by impersonal reasons, he does not give up the basic-
° Kant and the demands of self-consciousness
ness of the idea of a numerical identity of the self that is a priori. For this
notion is the very basis for his defense of the impersonal point of view.
Nor does he give up the idea that the immediacy of experience is
something that we must conceptualize in a manner that sustains the
publicity of what we represent. The publicity of what we represent
continues to express the numerical identity of an impersonal point of
view. And this impersonal point of view, in principle, is available to us in
any arbitrary context of experience.
While the constraints that self-consciousness imposes on experience
may be necessary to any experience that we may conceive of as such, it is
not obvious that we can dictate that the world must be such as to
conform to conditions on the ascription of belief. This is the sense in
which Kant is entitled to defend his thesis that objects of experience are
transcendentally ideal, that is, inherently mind-dependent. We cannot
know what they would be like independently of the conditions under
which we must ascribe beliefs about them to us and to other thinkers.
The thesis that our only access to the world is constrained by pre-
conditions on experience governing that access, is a weaker construal
than Kant often gives to his transcendental idealism. For he often insists
that we can know that things as they exist are not spatial or temporal,
and sometimes even claims that we can be certain of the non-spatio-
temporality of things as they exist in themselves. The most that the
purely epistemic dimension of Kant™s transcendental idealism allows
him is that we can only make sense of the world in terms of the
spatio-temporal conditions governing our experience. It does not li-
cence the further claim that space and time could not exist independent-
ly of the role that they play in allowing us to identify and reidentify
objects of experience, and our cognitive states as well.
Kant is most successful in spelling out a framework in terms of which
we can think of how to make sense of our position in the world. It is this
framework that I have tried to develop. I have sought to restore the
standpoint of an impersonal self to the status it deserves as the crucial
starting-point for understanding our capacity to understand our individ-
ual identity and our place in the world relative to other selves and to
material things. In the process, I have argued that Kant™s transcendental
self-consciousness is not just a logical capacity that we associate with the
capacity for judgment, although it is crucial to our capacity for judg-
ment. Nor does transcendental self-consciousness simply collapse into
consciousness of personal identity. But transcendental self-conscious-
ness is the key to drawing the distinction between inner and outer
experiences and recognizing inner experiences as inner and outer ex-
periences as outer. It is on the basis of this capacity to distinguish
between inner and outer experiences that we are then able to apply
concepts to our experience in judgment.

± © ®    ¤ µ  © ®
± The objection that Kant must appeal to an unintelligible notion of subject in
his conception of transcendental self-consciousness is well articulated by
Rudiger Bittner, ˜˜Transzendental,™™ in H. Krings et al. (eds.), Handbuch
philosophischen Grundbegri¬e (Munich: Kossel Verlag, ±·), vol. µ, pp. ±µ“
 Interpreters who have treated Kant™s theory of the identity of apperception
as a theory of personal identity include: Graham Bird, Robert Paul Wol¬,
Dieter Henrich, Paul Guyer, and Patricia Kitcher. I discuss their interpreta-
tion of the representation that we have of our numerical identity in chapter
 I will argue in chapter four that the interpretation that Henry Allison
provides of transcendental apperception collapses all self-consciousness for
Kant into knowledge of objective states of a¬airs, or at least into judgments
that purport to be objective; see H. Allison, Kant™s Transcendental Idealism (New
Haven: Yale University Press, ±), see esp. p. ±µµ: ˜˜In other words, a
subjective unity of consciousness is not a unity of self-consciousness, although
it can (as objecti¬ed) become a unity for self-conscious thought.™™
 In his classic discussion of Kant, The Bounds of Sense (London: Methuen, ±),
P. F. Strawson is careful not to run together the identity of the subject of
transcendental self-consciousness, which is concerned with the possibility of
experience, with empirical personal identity. A priori identity of self-con-
sciousness is supposed to be the basic condition for the possibility of empirical
self-ascription without being actually identical with that possibility (p. ±°).
Strawson rightly notes that, for Kant, objectivity has something to do with
transcendental self-consciousness and that personal self-consciousness has
something to do with how each of us experiences the world in our own
distinctive ways. Strawson is also clearer than many contemporary commen-
tators about the dependence of empirical self-consciousness on transcenden-
tal self-consciousness that Kant emphasizes in the same context. But Straw-
son does not spell out how transcendental self-consciousness is supposed to
help us to understand the connection and distinction between the subjective
and objective (Bounds of Sense, p. ±°·). While he notes that transcendental

Notes to pages “µ 
self-consciousness is supposed to be the core of empirical self-consciousness,
he has nothing to say about how it might serve as the core of empirical
self-consciousness. Indeed, he has almost nothing substantive to say about
what transcendental self-consciousness might be.
µ Kant™s general line of thought is articulated in an interesting way by Theodor
Lipps, Leitfaden der Psychologie (Leipzig: Engelmann, ±°), pp. ¬.: ˜˜I know . .
. immediately only of my own consciousness or of ˜me™. But this consciousness
is not in itself something individual, but rather simple consciousness; and this I
is not in itself ˜my™ I or ˜this™ I, but it is simply I . . . Not until I know of other I™s
does this I become ˜this™, ˜mine™, one among many, in short individual.™™ A
similar claim is developed in Theodor Lipps, Inhalt und Gegenstand: Psychologie
und Logik in Sitzungsberichte der philosophischen-philologischen und der historischen
Klasse der Koniglichen Bayerischen Akademie der Wissenschaften  (±°µ), p. µ. Kant
does not demand that I actually know that there are other egos in order to
ascribe beliefs or desires to myself. But he does think that one must have the
concept of other possible individuals endowed with self-consciousness in
order to have a grasp of the individuality of one™s own self-consciousness.

 © ®  ¤ µ  © ® § ° °    ° ©  ®
± J. Tetens, Philosophische Versuche uber die menschliche Natur und ihre Entwicklung of
±··± reprint: (Berlin: Reuther and Reichard, ±±), identi¬es inner sense with
apperception. In his Psychologia Empirica of ±· reprint (Hildesheim: Georg
Olms, ±), section µ, Christian Wol¬ argued that ˜˜apperception is at-
tributed to the mind insofar as it is conscious for itself of its perception.™™ Kant
criticizes the identi¬cation of inner sense with pure apperception in Anthro,
Ak. ©©, p. ±±, and with apperception in the B-Deduction at  ±µ.
 Andrew Brook, Kant and the Mind (New York: Cambridge University Press,
±), p. ±µ. Brook interprets Kant™s distinction between actual self-con-
sciousness and the possibility of self-consciousness and that between empiri-
cal and transcendental apperception as the di¬erence between optional and
non-optional acts of synthesis. The mind only sometimes engages in acts of
attentive awareness, however the mind is supposed to represent itself to the
self through the self™s activity of synthesis all the time. Now Kant does
distinguish between acts of attentive awareness and non-attentive awareness,
but he also distinguishes between self-consciousness and consciousness, and
there is little of no textual support for the claim that the distinction between
attentive and non-attentive awareness corresponds to that between self-
consciousness and consciousness. Con¬‚ating the two distinctions also seems
to be intrinsically implausible. The view that apperception is not about
self-awareness is also shared by Patricia Kitcher, Kant™s Transcendental Psychol-
ogy (New York: Oxford, ±°), esp. p. ±.
 Graham Bird, Kant™s Theory of Knowledge (London: Routledge and Kegan Paul,
±), pp. ±¬., sees Kant as responding to Hume™s worries about personal
identity. R. P. Wol¬, Kant™s Theory of Mental Activity (Cambridge, Mass.:
Harvard University Press, ±), pp. ±¬. Wol¬ argues that Kant came to
 Notes to pages µ“°
know Hume™s critique of self-identity through Beattie™s Essay on the Nature
and Immutability of Truth in Beattie™s Works (Philadelphia: Hopkins and Earle,
±°), translated into German in ±··. The argument is presented in detail
in R. P. Wol¬, ˜˜Kant™s Debt to Hume via Beattie,™™ Journal of the History of
Ideas ± (±°), ±±·“±. More recently it has been articulated with vigor by
Patricia Kitcher, Kant™s Transcendental Psychology, pp. ·¬. Kant™s conception
of self-identity a priori is said by Patricia Kitcher to be a response to Hume™s
worries about self-identity. She argues for this in ˜˜Kant on Self-Identity,™™
The Philosophical Review ± (±). ±“·, and in Kant™s Transcendental Psychol-
ogy, p. ·. Paul Guyer also maintains that Kant™s conception of the numeri-
cal identity of the self a priori is intended as a reply to Hume in Kant and the
Claims of Knowledge (Cambridge University Press, ±·), pp. ·, ±·, ±·.
Other interpreters, such as Strawson, note similarities and di¬erences
between the two views of self-identity, but are more circumspect in their
claims of in¬‚uence, cf. Bounds of Sense, pp. ±“±·°. Wolfgang Carl, Die
Transzendentale Deduktion der Kategorien in der ersten Au¬‚age der ˜˜Kritik der reinen
Vernunft™™: Ein Kommentar (Frankfurt: Klostermann, ±), is an exception to
this approach to the identity of self-consciousness; Carl sees the German
philosopher and psychologist Tetens as the target of Kant™s critique, pp.
 David Hume, A Treatise on Human Nature, L. E. Selby-Bigge (ed.) (Oxford:
Clarendon Press, ±), Book ©, Part ©, Section ©, pp. µ°“µ±.
µ Graham Bird, Kant™s Theory of Knowledge, pp. ±°¬., thinks that Kant runs
together apperception as involving personal identity and apperception as
providing a unity for objective judgment and conceptual rules.
 Strawson, The Bounds of Sense, pp. ±“±µ. Dieter Henrich, ˜˜Identity and
Objectivity,™™ in The Unity of Reason (Cambridge, Mass.: Harvard University
Press, ±.
· The in¬‚ated view of a priori consciousness of self-identity is defended by
Dieter Henrich, ˜˜Identity and Objectivity,™™ in his The Unity of Reason, pp.
±·¬., and Peter Rohs, ˜˜Uber Sinn und Sinnlosigkeit von Kants Theorie
der Subjektivitat,™™ Neue Hefte fur Philosophie ·/ (±), esp. ¬. Henrich
¨ ¨
argues that we have Cartesian certainty of our numerical identity over a
series of states. Rohs maintains that self-identi¬cation across time is criteria-
less and certain. He has nothing to say about the possibility of delusive
quasi-memories which threaten such a claim.
 Paul Guyer™s critique of Henrich is ¬rst presented in his review of Identitat ¨
und Objektivitat (the original German version of ˜˜Identity and Objectivity™™),
The Journal of Philosophy · (±·), ±µ±“±·. Henrich™s response may be found
in ˜˜The Identity of the Subject in the Transcendental Deduction,™™ in E.
Schaper and W. Vosenkuhl (eds.), Reading Kant: New Perspectives on Transcen-
dental Arguments and Critical Philosophy (Oxford: Basil Blackwell, ±), p. ·.
 Guyer, Kant and the Claims of Knowledge, pp. ±“±·.
±° Ibid., p. ±.
±± Ibid., p. °.
Notes to pages °“ µ
± A variant on Guyer™s interpretation is defended by Lorne Falkenstein, he
argues that, even though intuition is itself a manifold, when it is though it is
then collapsed into an absolute momentary spatial unity, i.e. that of a
particular shape. ˜˜The mind seems to distinguish automatically the mo-
ments of time over which the intuitions occur, and so, whereas it collapses
the manifold in space at any instant as an ˜absolute unity,™ it continues to
distinguish the di¬erent times over which the intuition occurs as so many
successive and distinct representations.™™ Kant™s Intuitionism (University of
Toronto Press, ±µ), pp. ·“··. Apart from the fact that Kant never claims
that thought transforms the manifold into an absolute unity, and that it is
hard to see how a particular shape could be an absolute unity, this view has
the di¬culty that it also requires powers of temporal discrimination of the
kind rightly criticized by Parsons. Falkenstein™s interpretation also seems to
be inconsistent with the position that Kant maps out in Re¬‚ection µ°:
˜˜All appearances stand as representations in time and are determined in
time. As a part of a whole appearance it [sic] cannot be determined
(genetically apprehended) in an instant, but only in a part of time.™™
± Ibid., pp. ±“±.
± P. Guyer, ˜˜Placing Myself in Time: Kant™s Third Paralogism,™™ in G. Funke
(ed.), Akten des µten Internationalen Kant-Kongresses (Bonn: Bouvier, ±µ), pp.
±µ Guyer, Kant and the Claims of Knowledge, p. ±°.
± Ibid.
±· Ibid., pp. ±µ“±.
± Jonathan Bennett, Kant™s Analytic (Cambridge University Press, ±), pp.
± A certain capacity for recognitional awareness also characterizes the intelli-
gence of higher animals. Kant maintains that, while subhuman animals,
such as dogs and hares are not able to develop bona ¬de concepts, they do
learn from the animals which they are pursuing or who are pursuing them
(Metaphysics Lecture K (±·/µ), Ak. ©, p. ). This is a longstanding view
of Kant™s. In a ±· paper, Kant notes that animals can distinguish objects
based on their sensations and desires, ˜˜the dog distinguishes the roast from
the bread, because he is di¬erently a¬ected by the roast than by the bread™™
(Ak. ©©, p. °). The distinctive feature of (subhuman) animal awareness is, for
him, its lack of the capacity for universalization and full-blown concept use:
˜˜Animals are acquainted [kennen] with objects, but they do not know
[erkennen] them.™™ (Logic, ed. Jasche, Ak. ©, p. µ). Knowledge or cognition
(˜˜erkennen,™™ ˜˜cognoscere™™) involves consciousness while the acquaintance
(˜˜kennen,™™ ˜˜gnoscere™™) requires only the associative capacity to represent
something in comparison with other things. Norman Kemp-Smith™s brief
discussion of Kant™s views on human and animal intelligence is quite
helpful, A Commentary on Kant™s ˜˜Critique of Pure Reason™™ (London: Macmillan,
±), pp. x±vii“±. There is also a useful discussion on the contrast between
animal and human intelligence in J. Michael Young, ˜˜Kant™s View of the
 Notes to pages µ“
Imagination,™™ Kant-Studien · (±), esp. ±µ°, and S. Naragon, ˜˜Kant on
Descartes and the Brutes,™™ Kant-Studien ± (±°), ±“.
° Thomas Nagel seems to me to be right when he argues in ˜˜What is it like to
be a bat?,™™ in Mortal Questions (Cambridge University Press, ±·), pp. ±¬.,
that it makes sense to talk of what it would be like to have the experience of
non-self-conscious being, such as a bat. However, I would argue that we can
make sense of this idea that there is something that it is like to be such a
creature and hence of its having a point of view only by thinking of such a
being as if it had self-consciousness. We understand representational con-
tent as such only from the vantage-point of the kind of creature that can
ascribe representational content to itself and to others.
± It is not entirely clear whether Kant thought he had a completeness proof
for the forms of judgment and the categories available or not. In a famous
Rostock dissertation, The Completeness of Kant™s Table of Judgments (Palo Alto:
Stanford University Press, ±), Klaus Reich argued that Kant had
thought through such an argument. Reich™s argument moves from consti-
tutive features of the unity of apperception (self-consciousness) to the indi-
vidual logical functions and from there to the categories. A more recent
book by R. Brandt, Die Urteilstafel (Hamburg: Meiner, ±), attempts a new
reconstruction based on the nature of judgment and the di¬erent manner in
which the nature of judgment manifests itself in the structure of concept,
judgment, and inference. The argument from the nature of judgment to a
completeness proof has been further developed by Beatrice Longuenesse,
Kant et le pouvoir de juger (Paris: Presses universitaires de France, ±) and
Michael Wol¬, Die Vollstandigkeit von Kants Urteilstafel (Frankfurt: Kloster-
mann, ±µ).
 In the second edition of the Critique, Kant admits that we would not give a
reason why the understanding requires the kind and number of categories it
supposedly does, or even why we have these and no other functions of
judgment, or, for that matter, why space and time are the only forms
according to which we can perceive objects ( ±). Great weight is given to
the passage from section ± in the critical examination of Reich™s argument
for the completeness of the table of judgment forms and categories under-
taken by Lorenz Kruger, ˜˜Wollte Kant die Vollstandigkeit seiner Urteils-
¨ ¨
tafel beweisen?,™™ Kant-Studien µ (±), “µµ.
 Wilfried Sellars, ˜˜this I or he or it (the thing) which thinks,™™ Proceedings and
Addresses of the American Philosophical Association  (±·±), ±±.
 Ibid., ·“.
µ Sellars, ˜˜Metaphysics and the Concept of a Person,™™ in K. Lambert (ed.),
The Logical Way of Doing Things (New Haven: Yale University Press, ±),
pp. “.
 Sellars, ˜˜This I or he or it (the thing) which thinks,™™ ±n.
· Ibid., ±.
 Kitcher, Kant™s Transcendental Psychology, esp. pp. ±±±“±±. Ralf Meerbote
generally endorses Kitcher™s claim that Kant™s account of representations is
Notes to pages “µ± ·
functionalist because it makes representations depend on each other by
virtue of their content, although he insists that sensations cannot be under-
stood by Kant purely functionally, see R. Meerbote, ˜˜Kant™s Functional-
ism,™™ in J.-C. Smith (ed.), Historical Foundations of Cognitive Science (Dordrecht:
Kluwer, ±°), esp. pp. ±, ±·°.
 Ibid., pp. ±“±.
° Kitcher, Kant™s Transcendental Psychology, pp. ±±±¬.
± Ibid., p. ±.
 A somewhat di¬erent version of the functionalist interpretation is defended
by Andrew Brook, Kant and the Mind, esp. pp. ±¬. Like Kitcher, Brook
emphasizes the importance of synthesis, the process of connecting represen-
tations, in the constitution of representations for Kant. Brook meets the
problem posed for Kitcher by the unity of the mind by identifying the mind
with the unity of its representations (Ibid. p. °). Where Kitcher insists on
the synthetic character of the claim that we can become conscious of our
representations, Brook takes it to follow trivially from the self-intimating
character of representation.
 Sellars, ˜˜This I or he or it (the thing) which thinks,™™ ±.
 Sellars™s functionalist interpretation of the Kantian notion of the self has
been developed and defended by Jay Rosenberg, The Thinking Self (Philadel-
phia: Temple University Press, ±), and C. Thomas Powell, Kant™s Theory
of Self-Consciousness (New York: Oxford University Press, ±°). In particular,
Powell has noted the tendency to misinterpret Kant™s claims for conscious-
ness of self-identity as claims to knowledge that we are identical particulars.
But Powell pushes this insight too far. He interprets the self of which we are
conscious in self-consciousness as an illusion. We must represent di¬erent
states only as if they belonged to a single consciousness (Ibid., p. µ). This
prevents him from ascribing a constitutive role to self-consciousness which
merely becomes a stand-in for objective constraints on experience that are
themselves intelligible independently of self-consciousness.

   ®   °  , ¬ · ,  ® ¤      §® © ©  ®  ¦       
± Richard Aquila, Matter in Mind: A Study of Kant™s Transcendental Deduction
(Bloomington, Indiana: University of Indiana Press, ±), pp. ±¬.
 Christian Wol¬ also claims in his Psychologia Empirica of ±· (Hildesheim:
Olms, ±), section ±·, that a reproduced idea is recognized when we are
conscious of already having had it. This is said to be memory. I owe the
references to Wol¬ to Wolfgang Carl, Die Transzendentale Deduktion, p. ±n.
The same point is made by Baumgarten in section µ· of the Metaphysics
(±·µ·) handbook that Kant used as the basis for his lectures (Ak. ).
 Kant connects the notion of a concept as a universal representation with its
function as a rule. Concepts are universal representations in two di¬erent
senses. They can be shared by di¬erent individuals with di¬erent intuitive
states, and they represent intuited particulars in repeatable terms. Kant
does not simply identify concepts with rules. In his discussion of the
 Notes to pages µ“·
subjective deduction, R. P. Wol¬ maintains that rules are concepts. This
leads him to detect ˜˜a con¬‚ict which runs through all the Critique and which
Kant never successfully resolves. On the one hand, his rationalist orienta-
tion and concern with the conditions of knowledge incline him toward an
emphasis on concepts, judgments, reasoning, and the other conscious
processes of cognition. On the other hand, his discovery of the problem of
consciousness, and his distinction between appearances and reality, force
him to assign the generative processes of the mind to a pre-conscious limbo.
Kant obscures this ambiguity to a certain extent by attributing the non-
conscious functions to faculties of the mind whose operations are customar-
ily considered conscious, by distinguishing synthesis itself from the bringing
of synthesis to concepts ( ·). This won™t do, however, for the concept is
simply the rule according to which the synthesis is performed, and hence
must precede, now follow it. ™™ Kant™s Theory of Mental Activity, p. ±±. Wol¬
confuses the concept with its schema. It is no surprise that Wol¬ cannot
make sense of Kant™s notion of a schema, since it is the schema rather than
the concept corresponding to it that precedes the synthesis to be performed.
 A helpful discussion of the di¬erent uses of the term ˜˜object™™ to be found in
Kant may be found in Charles Parsons, ˜˜Objects and Logic,™™ The Monist µ
(±), ±“µ±.
µ Henry Allison, ˜˜Transcendental A¬nity “ Kant™s Answer to Hume,™™ in L.
Beck (ed.), Kant™s Theory of Knowledge (Dordrecht: Reidel, ±·), esp. p. ±.
 Further remarks about the relation of Goodman™s problem to Kant™s
position may be found in Ralph Walker, Kant (Routledge and Kegan Paul,
±·), pp. ±·¬. as well as Gordon Brittan, Kant™s Theory of Science (Princeton
University Press, ±·), pp. ±¬. Neither author relates the Goodman
problem speci¬cally to Kant™s notion of the a¬nity of nature.
· Further discussion of the notion of a projected systematic unity of nature
may be found in Gerd Buchdahl, Metaphysics and the Philosophy of Science
(Cambridge Mass.: MIT Press, ±), chapter , as well as in the papers
collected in his Kant and the Dynamics of Reason (Oxford: Blackwell, ±).
Philip Kitcher, ˜˜Projecting the Order of Nature,™™ in R. Butts (ed.), Kant™s
Philosophy of Physical Science (Dordrecht: Reidel, ±), pp. ±°“µ, and a
discussion by Paul Guyer and Ralph Walker of ˜˜Kant™s Conception of
Empirical Law,™™ Proceedings of the Aristotelian Society, Supplementary Volume 
(±°), ±“, and “µ, are also extremely interesting. Systematic
unity has been discussed more recently by Susan Neiman, The Unity of
Reason (New York: Oxford University Press, ±), pp. ·°¬. and especially
Michael Friedman, Kant and the Exact Sciences (Cambridge, Mass.: Harvard
University Press, ±), esp. pp. ¬.

  ¬¦ - ®   ©  µ ®    ® ¤   ¤   ® ¤  ¦  µ¤ §  ® 
© ®    - ¤  ¤ µ  ©  ®
± Robert Howell, Kant™s Transcendental Deduction (Dordrecht: Kluwer, ±),
pp. ±°¬.
Notes to pages “ 
 Allison, Kant™s Transcendental Idealism, p. ±·.
 D. Henrich, ˜˜The Proof Structure of Kant™s Transcendental Deduction,™™
in R. C. Walker (ed.), Kant on Pure Reason (Oxford University Press, ±), pp.
 P. Guyer, ˜˜Kant on Apperception and A Priori Synthesis,™™ American Philo-
sophical Quarterly ±· (±°), ±°, Kant and the Claims of Knowledge, p. ±±.
µ Hector-Neri Castaneda, ˜˜The Role of Apperception in Kant™s Transcen-
dental Deduction of the Categories,™™ Nous  (±°), ±µ¬.
 Thomas Nagel, ˜˜Brain Bisection and the Unity of Consciousness,™™ in
Mortal Questions (Cambridge University Press, ±·), pp. ±·“±.
· Terence Wilkerson, ˜˜Kant on Self-Consciousness,™™ Philosophical Quarterly °
(±°), µ¬.
 Manfred Baum, Deduktion und Beweis in Kants Transzendentalphilosophie
(Meisenheim: Konigstein, ±).
 Guyer, Kant and the Claims of Knowledge, p. ±±·.
±° P. Guyer, ˜˜The Transcendental Deduction of the Categories,™™ The Cam-
bridge Companion to Kant (New York: Cambridge University Press, ±), p.
±µ±, sees Kant™s claim that the unity of consciousness is su¬cient for
cognition as a blatant error.
±± Allison, Kant™s Transcendental Idealism, p. ±µ.
± Ibid., p. ±.
± Ibid., p. ±µ.
± Ibid., p. ±µµ.
±µ Ibid., p. ±µµ.
± Ibid., p. °.
±· Ibid., p. ±.
± Ibid., p. .
± Ibid., p. .
° Ibid., p. ·.
± According to Allison, the only things of cognitive signi¬cance for Kant are
objective states of a¬airs. This is because objective states of a¬airs are the
only things of which we can be conscious in apperception. Allison™s restric-
tion of self-conscious experience to objective states of a¬airs also seems to
encourage him to give an absolute reading to the spontaneity or indepen-
dence from causal determination that Kant ascribes to self-consciousness in
epistemic contexts. For Allison, self-consciousness is more completely inde-
pendent of sensible information than it is for functionalists and other
philosophers who interpret the theoretical subject as only relatively sponta-
neous. See H. Allison, ˜˜Autonomy and Spontaneity in the Self,™™ in his
Idealism and Freedom, pp. ±“±.
 By treating objective validity as the target of judgment, I avoid the di¬cul-
ties that beset the interpretation that is defended by G. Prauss in his
Erscheinung bei Kant (Berlin: De Gruyter, ±·±), pp. ¬. and by H. Allison,
Kant™s Transcendental Idealism, pp. ·“·, according to which objective validity
consists only in the possibility of a judgment having a truth value. This
µ° Notes to pages “°
would hardly be enough to provide the truth with which Kant repeatedly
identi¬es objective validity.

µ   ¬¦ -  ®   ©  µ ®    ® ¤   µ ® ©    ¦ ©® µ © ©  ® :
  °¬  © ®§    -¤  ¤ µ   ©  ®
± The point that intuition involves unity or rather singularity of what is
represented by de¬nition is emphasized by Hoke Robinson, ˜˜Intuition and
the Manifold,™™ Southern Journal of Philosophy  (±), °µ.
 Henrich, ˜˜The Proof Structure of Kant™s Transcendental Deduction,™™ ·°“
·±. J. Claude Evans has recently defended Henrich in ˜˜Two-Steps-in-One-
Proof: The Structure of the Transcendental Deduction of the Categories,™™
Journal of the History of Philosophy  (±°), µ¬.
 Friedrich Tenbruck seems to be the ¬rst interpreter to have emphasized the
two-step character of the proof in the B-Deduction in ˜˜Die transzendentale
Deduktion der Kategorien nach der zweiten Au¬‚age der ˜Kritik der reinen
Vernunft,™ ™™ Ph.D. Thesis, Marburg (±). Tenbruck sees the ¬rst step as
establishing that the categories apply to all representations of which we are
conscious, since all such representations must be unitary. He also thinks
that there is some question whether ˜˜all the manifold must be given in One
empirical intuition™™ (p. µ°). This is because he takes the ¬rst step in the
proof to be analytic, expressing a merely hypothetical necessity. The second
step must then provide a premise from which a categorical necessity may be
derived. Tenbruck traces this second premise back to the a priori unity of
intuition defended in the Aesthetic. According to Bernhard Thole, ˜˜Die¨
Beweisstruktur der transzendentalen Deduktion in der zweiten Au¬‚age der
˜Kritik der reinen Vernunft,™ ™™ in G. Funke (ed.), Akten des µten Internationalen
Kant-Kongresses (Bonn: Bouvier, ±±), pp. °“±, and Kant und die
Gesetzmaßigkeit der Natur (Berlin: De Gruyter, ±±), the ¬rst step (sections
±µ“°) demonstrates only the possibility of thinking (not of knowing) objects
through the categories. A second step is required (sections “µ) in order to
show that there is a synthesis of empirical intuition corresponding to the
intellectual synthesis of thought. A third step then shows that the function of
synthesis in judgment is identical with that performed by the imagination in
the synthesis of perception and unity of time. Henry Allison divides the
argument into a proof of the objective validity of the categories in which
categories are only shown to apply to objects in a logical sense (step ±) and a
proof of the objective reality of categories in which they apply to real
empirical objects (step ), Kant™s Transcendental Idealism, p. ±µ; cf. also his
˜˜Apperception and Analyticity in the B-Deduction.™™
 Raymond Brouillet, ˜˜Dieter Henrich et ˜The Proof-Structure of Kant™s
Transcendental Deduction.™ Re¬‚exions critiques,™™ Dialogue ± (±·µ), ·“
; Hans Wagner, ˜˜Der Argumentationsgang in Kants Deduktion der
Kategorien,™™ Kant-Studien ·° (±°), µ“; and Robert Howell, Kant™s
Transcendental Deduction, pp. ±±¬., and ·n.
µ Manfred Baum, ˜˜The B-Deduction and the Refutation of Idealism,™™ ±°,
and especially Deduktion und Beweis, p. °±.
Notes to pages °“ µ±
 Edwin McCann, ˜˜Skepticism and Kant™s B-Deduction,™™ History of Philos-
ophy Quarterly  (±µ), esp. ·, ·, ·, ·.
· J. Claude Evans has recently defended Henrich in this way in ˜˜Two-Steps-
in-One-Proof,™™ esp. µ¬.
 The objection that an analytic ¬rst step would give the ¬rst step no
substantive role in the argument is forcefully made by Patricia Kitcher,
Kant™s Transcendental Psychology, p. ±·.
 McCann, ˜˜Skepticism and Kant™s B-Deduction,™™ ·.
±° A. Pistorius, ˜˜Review of Kant™s Prolegomena zu einer jeden kunftigen Metaphysik,™™
Allgemeine Deutsche Bibliothek µ (±·), µ. It has seemed, even to some recent
commentators, that self-consciousness for Kant must be a form of self-
knowledge. Edwin McCann argues, for instance, that one cannot even
think of oneself as a self without having the determinate self-knowledge that
is the basis for the argument of the Refutation of Idealism; McCann,
˜˜Skepticism and Kant™s B-Deduction,™™ . But this reconstruction of the
argument of the Deduction would leave it wide open to the objection made
by Pistorius.
±± Kemp-Smith translates ˜˜das, worinnen sich die Emp¬ndungen allein ord-
nen, und in gewisse Form gestellt werden konnen™™ ( ±/ µ) as ˜˜that in
which alone the sensations can be posited and ordered in a certain form,™™
Kant™s ˜˜Critique of Pure Reason™™ (New York: St Martin™s Press, ±), p. ,
treating ˜˜sich ordnen™™ as if it were not a re¬‚exive expression referring back
to ˜˜sensations.™™
± Karl Ameriks, Kant™s Theory of Mind (Oxford: Clarendon Press, ±),
p. ±.
± Henry Allison appeals to Re¬‚ection µ, Ak. ©©©, p. ±, in support of the
ascription of Kant of the bar substratum view of the self when it comes to
judgments concerning inner sense: ˜˜All inner experience is a judgment in
which the predicate is empirical and the subject is I. Independently of
experience, therefore, there remains merely the I for rational psychology;
for the I is the substratum of all empirical judgments.™™ Allison is clearly right
when he says that Kant is referring here to the I qua subject of judgments as
something non-empirical. It is also true for Kant that the I as such cannot be
known empirically. Kant is left with the bare substratum view Allison
ascribes to him if one denies the empirical existence of the self as anything
but an object of re¬‚ection. Allison does claim that ˜˜a subjective unity of
consciousness is not a unity of self-consciousness, although it can (as objecti-
¬ed) become a unity for self-conscious thought,™™ Kant™s Transcendental Idealism,
p. ±µµ. On this view all (empirical) consciousness of oneself as a particular
individual is knowledge of oneself as an object. Put more bluntly all self-
consciousness that is not self-knowledge seems to be pure apperception for
Allison. This is di¬cult to reconcile with Kant™s references to empirical
apperception as self-consciousness. Kant™s reference to that object of re¬‚ec-
tion as self-intuiting indicates that the non-empirical or logical I of (pure)
apperception may refer to itself as the possessor of a certain individuating
spatio-temporal history.
µ Notes to pages ±°±“±°
± Paton, Kant™s Metaphysic of Experience, pp. “; cf. also Allison, Kant™s
Transcendental Idealism, p. .
±µ Robert Howell, ˜˜Apperception and the ±· Transcendental Deduction,™™
Synthese · (±±), µ, ¬., also Kant™s Transcendental Deduction, pp. ±¬.
± ˜˜Kant takes the I to be an a priori representation of thought, a representa-
tion that gives me no awareness of any particular object called the subject of
thought, but instead gives me only the awareness that there is an entity “
whose nature is not further revealed to me “ that we can call the subject of
thought. Hence by an appeal simply to his revisionary view of knowledge,
Kant takes the I to yield me no (empirical) intuition or intellectual represen-
tation of any particular entity at all, let alone any representation that
displays to me the nature of such an entity. And Kant takes the I to yield me
no such intuition or representation despite the fact that, as he sees it, the
relevant type of I think-related unity does obtain among my representa-
tions.™™ R. Howell, ˜˜Apperception and the ±· Transcendental Deduc-
tion,™™ .
±· Ibid., .
± Ibid.
± The attribution of the re¬‚ection theory of self-consciousness to Kant has
enjoyed a certain vogue in the German philosophical literature. Henrich
develops this view of Kant™s theory of self-consciousness at greatest length in
˜˜Fichtes ˜Ich™,™™ in his Selbstverhaltnisse (Stuttgart: Reclam, ±), pp. ±“µ.
But he also makes some remarks about Kant™s purported re¬‚ection theory
in ˜˜Die Anfange des Subjekts (±·),™™ in A. Honneth et al. (eds.), Zwischen-
betrachtungen im Prozeß der Aufklarung (Frankfurt: Suhrkamp, ±), pp. ±°“
±·°. The re¬‚ection theory is developed and discussed in more detail in his
essay ˜˜Self-Consciousness: A Critical Introduction to a Theory,™™ Man and
World  (±·±), “. Henrich™s interpretation of Kant has been followed in
this regard by Ulrich Pothast, Uber einige Probleme der Selbstbeziehung (Frank-
furt: Suhrkamp, ±·±), p. ±; Ernst Tugendhat, Self-Consciousness and Self-
Determination (Cambridge, Mass.: MIT Press, ±), p. ±; and Manfred
Frank, ˜˜Intellektuelle Anschauung,™™ in E. Behler and J. Horisch (eds.), Die
Aktualitat der Fruhromantik (Paderborn: Schoningh, ±·), pp. “±. Tugen-
¨ ¨ ¨
dhat devotes most of his attention, nota bene, to criticizing Henrich™s and
Pothast™s theory of self-consciousness (pp. “·). Charles Larmore notes in
his review of Tugendhat™s book, Philosophical Review  (±), ±°“±°, that
Tugendhat does not escape the circularity di¬culties in the analysis of
self-consciousness which he diagnoses in the re¬‚ection theory. Dieter Hen-
rich argues the same point in his response to Tugendhat, ˜˜Noch einmal in
Zirkeln,™™ in C. Bellut and U. Muller Scholl (eds.), Mensch und Moderne
¨ ¨
(Wurzburg: Konigshausen and Neumann, ±), pp. “±.


. 7
( 9)