Marcus Willaschek

The Sensibility of Human Intuition

Kant’s Causal Condition on Accounts of Representation

1 Introduction
According to Kant, human intuition is sensible: “It comes along with our nature
that intuition can never be other than sensible” (A 51, B 75).1 Even though this
claim is central to Kant’s critical philosophy, Kant does not give an explicit argu­
ment for it. In what follows, I will offer such an argument, built out of elements
explicitly or implicitly accepted by Kant.
The claim that human intuition is sensible is an integral part of Kant’s distinc­
tion between sensibility and the understanding, of which he briefly “reminds” us
at the end of the Introduction to the first Critique (A 15, B 29)2 and from then on
takes for granted without any argument.3 For what follows, it will prove helpful to
present Kant’s distinction between sensibility and understanding in some detail
before we turn to the Kantian argument for the sensibility of human intuition.
This distinction amounts to a complex and highly original view about the struc­
ture of human cognition. Its central elements are the following claims:

1 References to the Critique of Pure Reason are to the page numbers of the first (A) and second (B)
original editions; all other references to Kant’s works are to the volume and page numbers of the
Academy Edition (Kant’s gesammelte Schriften, de Gruyter: Berlin 1900ff.). Translations follow
the Cambridge Edition of the Works of Immanuel Kant (Cambridge University Press), sometimes
with minor revisions.
2 The fact that Kant speaks of a “preliminary reminder” (“Vorerinnerung”) at A 15, B 29 may
suggest that he is referring back to something he had said – or published – before, e. g. his
inaugural dissertation De mundi, where Kant had distinguished between sensibility and intellect
as two different sources of representations in §§ 3–12. But first, the distinction drawn there differs
from the one in the critical works in that it treats sensibility and understanding independently
of each other as sources of cognition. Second, even in the dissertation, no explicit argument for
that distinction is given. Still, Kant may have thought that he had sufficiently established that
distinction in the earlier work and that therefore he could take it for granted in the first Critique.
3 Among recent commentators, some have tried to make up for this lack by offering such an ar­
gument (e. g. Allison 22004) or by defending Kant’s distinction against possible objections (e. g.
Engstrom 2006), while others have argued that we must accept Kant’s distinction between sen­
sibility and understanding as a fundamental assumption on which his theory rests – an assumption
that may be regarded as indirectly justified if the theory which is built on it is successful in
explaining the possibility of synthetic knowledge a priori (e. g. Heidemann 2002).
130 Marcus Willaschek

(SU1) Human beings can come to entertain mental representations in one of two
ways: either (a) as a result of an object’s causal impact on our minds (an
affection of our “Gemüt”) or (b) as a result of some “spontaneous” activity
of “uniting” various representations into a new one (cf. A 68, B 93).
(SU2) The capacity to come to represent something as a result of (SU1a) is a kind
of “receptivity” that Kant calls “ sensibility” (A 19, B 33).
(SU3) The capacity to come to represent something as a result of (SU1b) is a kind
of “spontaneity” called “understanding” (A 19, B 33).
(SU4) There are two basic kinds of “objective” representations (i. e. represen­
tations that purport to represent objects other than a subjective state of
mind), namely intuitions and concepts (A 19, B 33; cf. A 320, B 377).
(SU5) Intuitions are singular representations (that is, representations of par­
ticulars as such); through intuitions our minds do not refer to objects by
means of general marks and therefore refer immediately (A 19, B 33).4
(SU6) Concepts are general representations (that is, they represent objects only
indirectly insofar as they exhibit “marks” potentially shared by other
objects) (A 19, B 33).
(SU7) All intuitions in humans are sensible (A 51, B 75, cf. A 68, B 93); that is,
they arise from affections of our “sensibility” (A 19, B 33).5 Thus, human
intuitions essentially involve a moment of passivity; through them, objects
are “given” to us (A 19, B 33, cf. A 68, B 93).
(SU8) All concepts are intellectual; that is, with respect to concepts, our minds
are spontaneously active. Through them, objects are actively thought by us
by uniting various representations of them under a common one (A 19, B
33, cf. A 68, B 93).
(SU9) Human cognition requires both intuitions and concepts (A 51, B 75). (Very
roughly, concepts provide cognition with a content that can be true or false
and stand in rational relations; intuition provides the link to reality or, as
Kant puts it in the Critique of Judgment, to “objects” corresponding to our
concepts; cf. 5:401.)

4 Note that singularity, in this sense, is compatible both with a manifold of partial representations
and a multitude of represented objects. What matters is only that the objects in question are
represented not as falling under general concepts, but as particulars. While a concept represents
whichever particular happens to exhibit the general characteristics required for falling under that
concept, an intuition represents particulars as such or, as one might say, in their particularity.
5 This may seem to rule out the possibility of pure intuition; I will return to this issue below.
Note, however, that when Kant introduces the term “intuition” (A 19, B 33), he explicitly claims
that it “takes place only insofar as the object is given to us; but this in turn, is possible if it affects
the mind in a certain way.”
 The Sensibility of Human Intuition 131

These nine claims together constitute Kant’s distinction between sensibility and
understanding. The claim I will be primarily concerned with is SU7, the sensibility
of human intuition, since Kant never seems to give an explicit argument for this
central claim. In one place, Kant says that he has “proven” that human intuition
can only be sensible: “To be sure, above we were not able to prove that sensible
intuition is the only possible intuition, but rather that it is the only one possible
for us” (A 252). Unfortunately, beyond saying that it occurs “above,” Kant does
not tell us where this proof was given. However, it is important to see that SU7
does not stand on its own, but is an integral part of a complex conception of
human cognition. Only SU7 (in conjunction with SU8) allows Kant to treat the
distinctions between sensibility and understanding and between intuitions and
concepts as strictly parallel distinctions in the way he does. On the other hand,
only its place in the general framework gives SU7 a clear and precise meaning.6
In section 2, a causal condition on accounts of mental representation will be
introduced, according to which we can understand how a representation repre­
sents some object only if there is a causal connection between them. As will be
shown in section 3, this condition works as an implicit background assumption in
Kant’s thought from at least 1772 on and forms a central step in the argument for
the sensible character of human intuition. Given this assumption, it follows from
Kant’s definition of sensibility and the finitude of the human mind that human
intuition can only be sensible. Section 4 addresses the problem of how to recon­
cile the causal condition with Kant’s account of a priori cognition and with the
possibility of thoughts about non-sensible objects. Finally, section 5 discusses
some objections to this defence of Kant’s claim that human intuition can only be
sensible.

6 The claim that all human intuitions are sensible is closely related to what Henry Allison has
called “the discursivity thesis” (Allison 22004, 12), which is the thesis that human cognition
requires both sensible intuition and discursive concepts. According to Allison, the argument for
the discursivity thesis is “based on three bedrock epistemological assumptions: (1) that cognition
of any kind requires that an object somehow be given (this applies even to the problematic
intellectual or archetypal intuition); (2) that since a finite mind like ours is receptive rather than
creative, its intuition must be sensible, resting on affections by objects; and (3) that sensible
intuition, of itself, is insufficient to yield cognition of objects and requires the cooperation of
the spontaneity of the understanding” (ibid., 77; my emphasis). But it is hard to see how Kant
could have thought that his anti-rationalist assumption (2) is “relatively non-controversial,”
given that (2) is a direct denial of the central tenet of rationalism. Thus, even though I will agree
with Allison that Kant’s claim that all intuition is sensible rests on assumptions about the object-
relatedness and the finitude of human cognition, I think that much more needs to be said in
order to understand why Kant thought that he could rely on assumptions as controversial as
these.
132 Marcus Willaschek

2 The Causal Condition on Accounts

of Representation
Kant’s claim that human intuition can only be sensible follows from his defini­
tion of sensibility and the finitude of the human mind in conjunction with one
fundamental background assumption Kant had employed at least since 1772.
The background assumption is that we can understand how a representation
can represent anything at all only if there is some causal connection between the
representation and what it represents. This causal connection can go either from
the object to the representation (here the paradigm is perception) or from the
representation to the object (here the paradigm is the archetypal or productive
intuition of a divine mind). Let’s call this the Causal Condition on Accounts of
Representation, or Causal Condition, for short.

Causal Condition: We can account for the fact that something r is a represen­
tation of some object o only if there is a causal connection between r and o
such that either o causally depends on (is caused, at least in part, by) r or vice
versa.

Given this background assumption, it follows from Kant’s definition of intuition

as singular representation that we can understand how an intuition represents
its object only if it stands in a causal relation to the particular object it represents.
This means that either the object causally depends on the intuition or the intui­
tion depends on the object. For human cognition, the first option is clearly absurd,
since, at least generally, we cannot bring about the object of an intuition (or even
contribute to bringing it about) merely by representing it. This means that the
second option must hold: If we are to be able to give an account of how intuitions
represent their objects, we must regard intuitions as caused by the objects they
represent. But this is just to say that they are sensible, given Kant’s definition of
sensibility (SU2) as the “capacity (receptivity) to acquire representations through
the way in which we are affected by objects” (A 19, B 33). Hence, given the Causal
Condition, the claim that any finite intuition can only be sensible follows from
Kant’s definitions of intuition and sensibility (assuming that representation is
not an inexplicable phenomenon).
Before I go on to argue that Kant indeed accepted the Causal Condition,
some clarificatory remarks are in order about what exactly that condition
requires and how it relates to other aspects of Kant’s philosophy. First, we must
take into account the difference between concepts and intuitions. Kant clearly
did not hold that concepts as such must stand in causal relations to the objects
that fall under them. But this can easily be reconciled with the Causal Condition
 The Sensibility of Human Intuition 133

if we keep in mind that, according to Kant, concepts relate to their objects only
indirectly, because they represent them through general marks: “In whatever
way and through whatever means a cognition may relate to objects, that through
which it relates immediately to them, and at which all thought as a means is
directed as an end, is intuition. […] all thought whether straightaway (directe) or
through a detour (indirecte), must ultimately be related to intuitions, since there
is no other way in which objects can be given to us” (A 19, B 33). This means that
if Kant in fact accepted the Causal Condition, then the causal relation required to
understand how concepts can represent their objects need not consist in a causal
relation between concepts themselves and their objects, but rather in their being
related to intuitions, which in turn stand in causal relation to their objects.
Second, no specific conception of causality is required for the Causal Condi­
tion. In particular, it does not require that representation and object be related
as cause and effect by a universal law. Rather, any conception of causation will
do, as long as it supports asymmetrical counterfactuals of the following kind: “If
there had not been object o, there would not have been representation r, but not
vice versa“ and “If there had not been representation r, there would not have been
this object o, but not vice versa.”7 Given that, according to Kant, the only causal
relations we can have knowledge of consist in lawlike conjunctions of spatio-tem­
poral events (cf. A 189ff., B 233ff.), it follows that we cannot have any knowledge
of the causal relations constitutive of re­presentations unless they are of this kind.
As the example of Kant’s conception of spontaneous agency and “noumenal”
causality shows (cf. A 532ff., B 560ff.), however, this does not prevent Kant from
considering, and indeed from positing, causal relations of a different, non-spatio-
temporal kind. Since no knowledge of particular causal relations between rep­
resentations and their objects is required by the Causal Condition, it is possible
here to leave open the precise character of the causal relations in question.
Third, it is important to keep in mind that the Causal Condition does not say
right away that representation requires a causal connection, but says only that
this is a condition for our understanding of how the representation represents
its object. Let’s call the stronger claim that representation requires causation the
Causal Assumption:

Causal Assumption: If something r is a representation of some object o, then

there is a causal connection between r and o such that either o causally
depends on (is caused, at least in part, by) r or vice versa.

7 The asymmetry condition is meant to rule out that Leibnizian pre-estblished harmony and
Malebranchean occasionalism satisfy the Causal Condition.
134 Marcus Willaschek

As we will see below, there are places where Kant seems to accept this stronger
claim. However, the weaker Causal Condition (on accounts of representation) will
prove sufficient as a basis for an argument for the sensibility of human intuition.8
I will return to this point below.
Finally, there is an obvious exegetical problem with attributing the Causal
Condition and/or the Causal Assumption to Kant, namely that this seems to con­
flict with the possibility of a priori representations – first, in the case of pure
intuitions of space and time, and second, in the case of transcendental ideas,
since both kinds of representation seem to represent their objects without being
causally dependent on them. I will return to these issues in section 4.

3 Textual Evidence for Kant’s acceptance of the

Causal Condition and the Causal Assumption
I will now turn to some textual evidence that Kant in fact accepted both the Causal
Condition and the Causal Assumption. Let us start right at the beginning of Kant’s
critical project, with his famous letter to Marcus Herz from February 21, 1772. After
asking the seminal question:

What is the ground of the relation of that in us which we call ‘representation’ to the object?,

8 Stephen Engstrom attributes to Kant a condition on cognition that is very similar to the Causal
Assumption: “Specifically, there must be a relation of causal dependence connecting the
actuality of cognition and the actuality of its object” (Engstrom 2006, p. 9). And further: “This
condition, Kant notes, may take one of two forms, since there are two different directions the
dependence on which this connection consists may have – a difference on which is based the
division of finite cognition into its two types, theoretical and practical” (Engstrom 2006, p. 11–2).
My approach in this paper is in general agreement with Engstrom’s in this respect. The differences
mainly concern the aims: First, Engstrom uses his causal condition on cognition to explain why
cognition, even though spontaneous, requires receptivity for its exercise. By contrast, I will use
the Causal Condition to explain why, according to Kant, intuition in finite minds can only be
sensible. Second, Engstrom is primarily concerned with cognition, whereas I am interested here
in representation in general, including sub-judgemental representations. If there is a causal
condition on representation in general, this implies a causal condition on cognition, but not vice
versa. Finally, Engstrom does not provide any textual evidence that Kant indeed accepted a causal
condition on cognition. The two passages he refers to (Bix–x and 5:46) explicitly require a causal
connection only for practical, but not for theoretical cognition. To my knowledge, Kant nowhere
explicitly endorses a causal condition either on representation in general or on cognition in
particular. However, I will discuss below various passages in which Kant does so implicitly.
 The Sensibility of Human Intuition 135

Kant continues:

If a representation comprises only the manner in which the subject is affected by the object,
then it is easy to see how it is in conformity with this object, namely, as an effect accords with
its cause, and it is easy to see how this modification of our mind can represent something,
that is, have an object. […] Similarly, if that in us which we call “representation” were active
with regard to the object, that is, if the object itself were created by the representation, […]
the conformity of these representations to their objects could also be understood. Thus the
possibility of both an intellectus archetypus (an intellect whose intuition is itself the ground
of things) and an intellectus ectypus, an intellect which would derive the data for its logical
procedure from the sensuous intuition of things, is at least comprehensible. However, our
understanding, through its representations, is neither the cause of the object (save in the
case of moral ends), nor is the object the cause of our intellectual representations in the real
sense (in sensu reali) (10:130).

This raises the very question Kant answers in the Transcendental Analytic of the
first Critique, namely how pure concepts of the understanding can have “objec­
tive reality,” that is, how they can represent objects. In the letter to Herz, Kant
does not yet envisage the solution offered in the first Critique, so he merely formu­
lates questions that remain unanswered in the letter:

But by what means are these things given to us, if not by the way in which they affect us?
And if such intellectual representations depend on our inner activity, whence comes the
agreement that they are supposed to have with objects – objects that are nevertheless not
possibly produced by them? (10:131).

Kant’s reasoning in the letter to Herz has the form of a dilemma: In order to under­
stand how a “determination of our mind can represent something”, either the
“determination of our mind” must be caused by the object or the object must
be caused by the “determination of our mind.” Since it seems that intellectual
representations are neither causes of nor caused by their objects, we cannot
understand how they represent something. This way of reasoning clearly presup­
poses the Causal Condition: In order for us to understand how a representation
represents an object, there has to be some kind of causal connection between the
representation and its object.9

9 Béatrice Longuenesse, too, detects a causal account of representation at work in Kant’s letter
to Herz (cf. Longuenesse 1998, p. 18ff.), but goes on to claim that Kant gave up this account in
favor of an account that treats the relation between representation and object “as internal to
representation” (Longuenesse 1998, p. 20). In what follows, I will argue that Kant did not give up
the Causal Condition, but accepted it throughout the critical period.
136 Marcus Willaschek

Let us now turn to a note Kant made at the margin of his copy of the first
Critique, at the beginning of the “Transcendental Aesthetic,” where he seems to
accept even the stronger Causal Assumption. In the published text, Kant had said
that an intuition takes place “only insofar as the object is given to us; but this, in
turn, is possible only if it affects the mind in a certain way” (A 19). To this, Kant
adds at the margin: “If the representation is not in itself the cause of the object”
(23: 21). If we put the published text and the note together, we get something like
the following claim: “If the representation is not in itself the cause of the object,
an intuition is possible only if its object affects the mind in a certain way.” This
is not just an application, to the case of intuitions, of the Causal Condition on
how we can understand representations to be related to their objects. Rather, it
amounts to the stronger claim about what is constitutive of representations (the
Causal Assumption): For an intuition to represent an object, either the object
must be caused by the intuition or the intuition must be caused by its object.10
Next, let us take a look at Kant’s explanation, in the Introduction to the Tran­
scendental Dialectic, of why it is problematic, and indeed “paradoxical” (some­
thing sehr Widersinnisches), to expect principles of pure reason to have objective
validity. As a contrasting case, Kant considers the attempt to simplify a body of
legal norms by deriving all norms from some fundamental principles in accor­
dance with the rational concept of right (cf. A 301, B 358). According to Kant, we
can understand how this might be possible because here the principles “apply to
something that is wholly our own work, and of which we can be the cause through
that concept” (A 301, B 358; my emphasis). Kant continues: “But that objects in
themselves, as well as the nature of things, should stand under principles and be
determined according to mere concepts is something that, if not impossible, is
at least very paradoxical in what it demands” (A 302, B 358). Why should this be
paradoxical? Because the only non-paradoxical way of explaining how a priori
principles and concepts of pure reason can have objective validity would be on
the model of simplifying a legal system according to the idea of right, that is, on
the model according to which the object is caused by, or caused according to, the
concept a priori. Again, the Causal Condition stands in the background according
to which there are only two ways in which we can understand how a representa­
tion can “determine,” that is, adequately represent, its object: Either because the

10 What Kant has in mind when he speaks of the representation being the cause of the object is
of course the possibility of an intellectual intuition, already mentioned in this connection in the
letter to Herz. – Incidentally, Kant did not include the marginal note itself in the B-edition, but
rather restricted the claim that an object is given only if it affects the mind by adding “at least for
us humans” (B 33), thus excluding from consideration a possible intellectual intuition.
 The Sensibility of Human Intuition 137

representation is causally dependent on the object or because the object causally

depends on the representation. If neither of these two possibilities obtains, as in
the case of the principles and ideas of pure reason, the claim that they nevertheless
are adequate representations of “objects in themselves” becomes paradoxical.11
Besides these passages where Kant seems to presuppose either the Causal
Condition or the Causal Assumption directly (and one passage – A 92, B 124 –
where he comes close to accepting the latter explicitly, more on which below),
there are two groups of passages in which Kant seems to rely on one or both of
them at least indirectly. On the one hand, there are the passages where Kant moti­
vates the transcendental deduction of the categories; on the other, there are those
where he contrasts finite and infinite minds and their respective kinds of cogni­
tion. I will briefly discuss these two groups, but can do so here only quite sum­
marily.
The whole problematic of the “objective reality” (or the “sense and reference”)
of pure a priori concepts – the problem to which the transcendental deduction of
the categories is meant to be the solution – starts from the recognition that we
need a special explanation of how a priori concepts can refer to objects. We need
this explanation precisely because their object-relatedness cannot be traced back
to “experience,” i. e. to some causal impact by the represented object.12 This, of
course, is the problem Kant mentions for the first time in the letter to Herz quoted
above. That Kant still sees the problem in this light in the critical period becomes
evident from numerous passages such as this one:

Among the many concepts, however, that constitute the very mixed fabric of human cog­
nition, there are some that are also meant for pure use a priori (completely independently
of all experience), and these always require a deduction of their entitlement, since proofs
from experience are not sufficient for the lawfulness of such use, and yet one must know
how these concepts can be related to objects that they do not derive from any experience.
I therefore call the explanation of the way in which concepts can relate to objects a priori
their transcendental deduction (A 85, B 117).

Even though Kant does not frame the problem in terms of causal relations between
representations and their objects here (in part because he is talking about con­
cepts and not about intuitions), it is clear that something like the Causal Condition
must stand in the background once we see that experience, for Kant, essentially
involves a causal impact of the experienced object on our sensibility: Experience is
empirical cognition (B 147), and what makes cognition empirical is that it contains

11 For a similar passage, cf. A 306, B 362.

12 For a similar point in a similar context, cf. Engstrom (2006, Fn. 13).
138 Marcus Willaschek

“sensation” (A 50, B 74), which in turn is the effect of an object on our sensibility (A
19f., B 34). Hence, a priori concepts require a transcendental deduction precisely
because their relatedness to objects cannot be accounted for by their being cau­
sally dependent on the object they represent. Since they are not the causes of their
objects (at least “as far as their existence is concerned,” cf. A 92, B 125; more on
this below), we need an account of how they can relate to objects at all.
On the other hand, there are all those passages where Kant contrasts human
with divine cognition and the cognition of finite with that of infinite minds. As
Kant himself emphasizes in various places, he is using this contrast exclusively
to bring out what is special about human cognition, since apart from the contrast
with human cognition we do not have any positive conception of a divine mind
(cf. e. g. B 307f.; 5:405; 5:408). Now according to Kant, the finitude of our minds
has the consequence that in order for its representations to relate to objects (for
them to have “objective reality”), our minds are dependent on something external
to them.13 As we have seen, this external factor must take the form of a causal
impact on our sense organs. Now it is striking that the contrasting conception
Kant works with is not that of a finite being that gets its input in some other (non-
sensible) way, but rather that of an infinite being that doesn’t require any external
input at all. Even though Kant does not say so explicitly, he seems to assume that
only an infinite or divine mind can have non-sensible intuitions or, conversely,
that all finite minds need some sensible input, even if they may have other forms
of sensibility than ours (space and time).
This becomes apparent, for instance, in the B-deduction, where Kant first
characterizes an intuitive understanding as one “through whose self-conscious­
ness the manifold of intuition would at the same time be given” (B 138), that is, an
infinite mind that does not require any “external” input into its cognitive system.
Kant continues: “the human understanding cannot even form for itself the least
concept of another possible understanding, either one that would itself intuit
[i. e. the infinite mind just mentioned] or one that, while possessing a sensible
intuition, would possess one of a different kind than one grounded in space and
time” (B 139). A little later, Kant then explains that an intuitive understanding
would be one “through whose representations the objects would themselves at
the same time be given, or produced” (B 145). Hence, it seems that for Kant the
distinctions between finite and infinite minds, between discursive and intuitive
understanding, and between minds that do and minds that don’t require sensible

13 This formulation is meant to capture Engstrom’s point that the role of receptivity in cognition
is not to constrain the workings of spontaneity, but rather to enable them (cf. Engstrom “Under­
standing and Sensibility,” 17 et passim).
 The Sensibility of Human Intuition 139

input coincide. Only an infinite mind (that produces the objects it represents by
representing them) can have a non-sensible intuition (or an intuitive understand­
ing), whereas all finite minds require some kind of sensible input, even though
their forms of sensibility may vary. In this way, the possibility of a finite mind with
a non-sensible intuition does not come into view at all.
I think we can explain this by attributing to Kant acceptance of the Causal Con­
dition (and/or Causal Assumption). If representations (or philosophical accounts
thereof) require a causal connection between representations and represented
objects, then all finite minds will require some sensible input, because (i) their
representations must be either caused by or causes of the represented objects, but
(ii) due to their finitude, at least generally, their representations are not the causes
of the represented objects, so that (iii) the objects must be the causes of their rep­
resentations, which means, according to Kant’s definition of sensibility (cf. SU2
above), (iv) that at least some of their representations must be “sensible” repre­
sentations. Conversely, a mind that does not require sensible input can only be
an infinite or creative mind (an intellectus archetypus; cf. 5:408), since, on Kant’s
definition of sensibility, any kind of external input – that is, any representation
caused by an object that exists independently of its being thus represented – will
count as sensible, so that only an infinite mind that does not require any input at
all will not require sensible input. (I will return to the question of whether a critic
can, and should, object to this definition of sensibility below.) So it seems that
Kant’s specific way of contrasting human and divine cognition, too, presupposes
either the Causal Condition or the Causal Assumption, because otherwise Kant
would have had to allow for the possibility of finite but non-sensible minds.
So much then for direct and indirect evidence that Kant indeed accepted
both the Causal Condition and the Causal Assumption. Given Kant’s definitions
of intuition as singular representation and of sensibility as the capacity to receive
representations through being causally affected by objects, Kant’s claim that all
intuition in humans is sensible follows from the obvious fact that we don’t have
the power to bring objects into existence merely by representing them. If we don’t
have that power, the only way to understand how our intuitions can represent
objects is to hold that the intuitions are caused by their objects, which means that
the intuitions must be sensible.
140 Marcus Willaschek

4 Exegetical Problems with the Causal Condition

and Causal Assumption
Ascribing to Kant acceptance of the Causal Condition and/or Causal Assumption
raises a number of exegetical questions concerning the internal consistency of
Kant’s position. In particular, if Kant held that mental representation requires
a causal connection to the object represented, we have to ask how this is com­
patible (1) with his own transcendental-idealist account of representations a
priori (space, time, categories) and (2) with his view that, even though we cannot
cognize non-sensible objects (such as God and immortal souls), we can at least
think about them.

(1) Concerning the first question, the problem is how there can be representations
a priori, and how they can relate to objects, even though qua a priori they are
not caused by the objects they represent. The solution to the problem consists in
a refinement of the general claim that finite minds cannot cause objects just by
representing them. Here is what Kant says in the “Transition to the transcenden­
tal deduction of the categories”:

There are only two possible cases in which synthetic representation and its objects can
come together, necessarily relate to each other, and, as it were, meet each other: Either if
the object alone makes the representation possible, or if the representation alone makes the
object possible (A 92, B 124f.).

If we may take Kant’s talk of “making something possible” to have causal impli­
cations (as the words “causality” and “produce” in the sentences that follow
suggests)14, then this is a straightforward application of the Causal Assumption
to the case of “synthetic representation” (by which Kant here, as the context
makes clear, means synthetic cognition). In order for the representation to relate
to the object “necessarily” (i. e. so as to constitute a representational relation),
either the representation has to depend counterfactually on the object (if there
had not been that object, there would not have been that representation) or vice
versa. Kant continues: “If it is the first, then this relation is only empirical and
the representation is never possible a priori” (A 92, B 125). Assuming that a finite
mind cannot “make possible” an object merely by representing it, this raises the
question how we can ever have representations a priori. Kant’s response is to dis­

14 “[R]epresentation in itself (for we are not here talking about its causality by means of the will)
does not produce its object as far as its existence is concerned […]” (A 92, B 125).
 The Sensibility of Human Intuition 141

tinguish between making an object possible as far as its existence is concerend

and as far as its form is concerned:

But if it is the second, then since representation in itself (for we are not talking about its cau­
sality by means of the will) does not produce its object as far as its existence is concerned,
the representation is still determinate of the object a priori if it is possible through it alone
to cognize something as an object. But there are two conditions under which alone the cog­
nition of an object is possible: first, intuition […]; second, concept […]. […] the first condition
[i. e. intuition] in fact does lie in the mind a priori as the ground of the form of objects. […]
consequently, the objective validity of the categories, as a priori concepts, rests on the fact
that through them alone is experience possible (as far as the form of thinking is concerned)
(A 92f., B 125f.; second and third emphasis mine).

Setting aside many difficulties raised by this passage, I think we can discern in it
Kant’s general strategy of how to reconcile the Causal Condition with the possi­
bility of a priori representations: Even though we cannot “produce” an object, as
far as its existence is concerned, merely by representing it, our a priori represen­
tations can determine the form of that object in so far as we represent it. Accor­
ding to Kant, there are two kinds of such forms, namely those of intuition and
those of thought. Whereas space and time are a priori forms of intuition und thus
make possible the sensible form of any object we can cognize, the categories are
the forms of thought that constitute the conceptual structure of anything we can
think of as an object. These forms are not imposed on our cognition by the object
itself, but rather imposed by our minds on the object of cognition. In this sense,
we make the object possible – not in its existence, but in its sensible and intellec­
tual form – by representing it.
That Kant’s talk of “making an object possible” (and related talk of “condi­
tions of possibility” etc.) should really be understood as having causal implica­
tions can be confirmed if we consider the famous passage from the B-preface that
announces Kant’s Copernican Turn: “Up to now it has been assumed that all our
cognition must conform to the objects […]. Let us once try whether we do not
get farther with the problems of metaphysics by assuming that the objects must
conform to our cognition, which would agree better with the requested possibility
of an a priori cognition of them” (B xvi). Later in the Critique, the sense in which
the objects, according to Kant, must conform to our cognition, is captured by the
slogan that the conditions of the possibility of experience are also conditions of
the possibility of the objects of experience (A 158, B 197). Presumably, Kant does
not want to claim that we actually create the objects of experience by cognizing
them; but still, the talk of objects conforming to our cognition has causal implica­
tions at least in the weak sense that it implies the following counterfactual claim:
If our cognition were different in relevant ways (in particular, if it had different
a priori forms), then the objects of our cognition would be relevantly different,
142 Marcus Willaschek

too. Thus, Kant is committed to the claim that if we had different a priori forms of
intuition, then the objects of our experience would not be in space and time (cf.
e. g. A 27, B 43; A 34f., B 50).
In this way, Kant’s own transcendental idealist solution to the problem raised
in the letter to Herz confirms, rather than contradicts, his acceptance of the Causal
Condition on accounts of representation. While the empirical aspects of our cog­
nition causally depend on the represented object either directly (as in the case of
empirical intuitions) or indirectly (as in the case of empirical concepts), the non-
empirical aspects of cognition – that is, a priori intuitions and a priori concepts –
can represent something in the object of our cognition only in so far as they make
possible the very features they represent. As Kant puts it in the B-preface: “we can
cognize of things a priori only what we ourselves have put into them” (B xviii). To
be sure, talk of putting something into objects has to be taken metaphorically.
But in whichever way one wants to cash out the metaphor, one must retain the
idea that the possibility of a priori cognition is explained by a dependence of the
objects as cognized on the conditions of cognizing them. In claiming that there
is such a dependence, Kant makes sure that his transcendental idealism satisfies
the Causal Condition on accounts of representation.
This general strategy of reconciling the possibility of a priori cognition with
the Causal Condition also works for the case of our a priori representations of
space and time. These are what Kant calls “pure intuitions,” that is, intuitions
that do not contain sensation (A 20, B 24f.) and thus do not seem to require affec­
tations of our senses. As Kant argues in the “Transcendental Aesthetic,” space
and time are primarily a priori forms of intuition. As Kant acknowledges in a foot­
note to the B-version of the Transcendental Deduction, however, their status as
forms of intuition as such does not give us representations of space and time as
objects (as they are needed in mathematical thinking) (B 160). The forms of intu­
ition provide us with a manifold of intuition (namely points in space and time),
but as such they do not unite this manifold into intuitive representations of space
and time. This latter kind of representation Kant calls “formal intuition” (B 160),
which results from actively synthesising the manifolds of space and time into
unified representations of space and time as objects (in the widest sense). Hence,
space and time as objects of our representations are not mind-independent
objects, but products of acts of synthesis. The same is true about mathematical
objects such as a line or triangle (cf. B 154). Hence, Kant’s account of mathemati­
cal objects, including space and time considered as objects (as opposed to forms
of intuition), conforms to the Causal Condition by making these objects causally
depend on acts of synthesis.
 The Sensibility of Human Intuition 143

(2) A second exegetical problem concerns Kant’s distinction, most prominent in

the B-preface to the first Critique, between cognizing something and merely thin-
king it: “Yet the reservation must also be well noted that even if we cannot cognize
these same objects [i. e. objects of experience] as things in themselves, we at least
must be able to think them as things in themselves” (B xxvi; cf. B 166 fn.). And
Kant goes on to apply this distinction not only to objects of experience (which
we can also think of as things in themselves), but to non-sensible objects such
as God, a simple soul and free will (B xxix) which we can only think of as things
in themselves. This leads up to Kant’s famous assertion that he “had to deny
knowledge in order to make room for faith,” which implies that, even though we
cannot know that God exists and that our souls are immortal, we can believe in
(and hence think of) God and an immortal soul.
Now the question is how this is compatible with the Causal Condition. Since
God and souls are non-sensible objects, they cannot be the causes of the repre­
sentations through which we think (of) them. (On Kant’s definition of sensation,
this would make them sensible objects.) But neither do we “make possible” these
intelligible objects by representing them – either concerning their existence or
concerning their form. Hence, it seems that the Causal Condition would imply
that thoughts of God and souls are either inexplicable or impossible. And indeed,
this is the consequence Kant draws – as long as we restrict ourselves to theoreti­
cal cognition alone.
The radical consequences of Kant’s theory of human cognition for the possi­
bility of thoughts about God and other non-sensible objects are often overlooked.
Kant’s account of metaphysical thinking is highly complex, and I cannot begin
to discuss it adequately here, so I will restrict myself to three brief remarks. First,
Kant takes the transcendental deduction of the categories to show that a priori
concepts have “objective reality” (do relate to possible objects) only insofar as
their objects can, at least in principle, be given in experience. As Kant repeats
several times, without sensible intuition, a priori concepts would be “empty” (B
149; cf. A 51, B 75) and “without sense and reference” (cf. B 149; 8:133; A 239, B
298): “The merely transcendental use of the categories [i. e. a use not restricted to
objects of a possible experience; A 238, B 298] is thus in fact no use at all, and has
no determinate object, nor even an object that is determinable at least as far as its
form is concerned” (A 247, B 304f.).15

15 There is room for controversy here, since in many other passages Kant only claims that
without intuition we cannot cognize objects. For my purposes, it will suffice that Kant, in the
passage just quoted and various other places, clearly commits himself to the stronger claim.
144 Marcus Willaschek

Second, Kant offers a broadly “subjectivist” account of how we come to have

concepts such as that of God and of an immortal soul – concepts Kant calls “tran­
scendental ideas” – which explains our having such concepts in a way that leaves
entirely open whether there are any objects corresponding to them and even how
they might relate to these objects if there should be any (cf. A 312, B 377–A 338,
B 396). Our concepts of God and of an immortal soul are not derived from any
“objective” feature of the world, but rather respond to a “subjective” need of our
own reason.16
Third, Kant does not deny that our thoughts of God and an immortal soul do
have a content that suffices to specify determinate objects. But they receive this
content only indirectly, through their relation to the moral law and the uncondi­
tional obligation it lays upon us. Consider the following footnote Kant adds to the
sentence in the B-preface where he distinguishes between cognizing objects and
thinking them:

To cognize an object, it is required that I be able to prove its possibility (whether by the
testimony of the experience from its actuality or a priori through reason). But I can think
whatever I like, as long as I do not contradict myself, i. e. as long as my concept is a pos­
sible thought, even if I cannot give any assurance whether or not there is a corresponding
object somewhere within the sum total of all possibilities. But in order to ascribe objective
validity to such a concept (real possibility, for the first sort of possibility was merely logical),
something more is required. This ‘more,’ however, need not be sought in theoretical sources
of cognition; it may also lie in practical ones (B xxvi).

As Kant makes clear in this footnote, we must distinguish between the logical
possibility of a concept, the real possibility of a concept, and the actuality of the
object corresponding to that concept. Whereas the logical possibility of a concept
requires only that the marks that are united in that concept do not contradict each
other, its real possibility requires that some possible object correspond to it (cf. A
596, B 624 Fn.). In the case of empirical objects, their possibility consists in their
conformity with the “formal conditions of experience” (cf. A 218, B 265). While it
remains somewhat unclear what the corresponding possibility of a non-empirical

16 Cf. e. g. A 309, B 365; A 336, B 393. An idea, according to Kant, is “a necessary concept of
reason to which no congruent object can be given in the senses” (A 327, B 383). Neither, we may
add, can its object be given in pure intuition. But then it follows from Kant’s views about concepts
without intuition that transcendental ideas as such do not have any determinate, nor even a
determinable, object. And how could they, if we can account for our having these concepts purely
from “within” our own thinking, quite independently of their having any relation to an object
at all? Consequently, Kant says of the transcendental ideas that “no object can be determined
through them” (A 329, B 385). – On Kant’s “subjective deduction” of the transcendental ideas
from the structure of human reason (cf. Klimmek 2005).
 The Sensibility of Human Intuition 145

object (such as God or a soul) would consist in, Kant insists that, even in the case
of our concepts of God and soul, their real possibility goes beyond their logical
possibility in requiring “objective reality” (cf. A 596, B 624 Fn.; 20:325). As Kant
points out in the first sentence of the quoted footnote, if there is an actual object
corresponding to the concept, this guarantees the concept’s objective reality (and
thus its real possibility). But, as Kant indicates in the last sentence, there is a
further way to guarantee its real possibility (i. e. that there is a possible object
corresponding to it), namely from “practical sources of cognition.”
What Kant has in mind here becomes fully clear only in the second Critique,
where, in the context of his doctrine of the postulates of pure practical reason,
Kant explains how our ideas of God and immortality receive “objective reality”
through their relation to the moral law. This relation, to put it very briefly, consists
in the fact that the moral law requires us to realize the highest good, which consist
in a necessary congruence of complete virtue with complete happiness (5:110f.),
but we can think of the highest good as realizable only if we presuppose freedom
(5:114f.), immortality (5:122f.) and God (5:124f.). Only in this way, Kant argues, do
these ideas get any content determinate enough to specify possible objects:

The abovementioned three ideas of speculative reason [of freedom, immortality and God]
in themselves are no cognitions; but they are (transcendent) thoughts in which there is
nothing impossible. Now they receive, through an apodictic practical law […], objective
reality, i. e. it [the law] indicates to us that they have objects, without being able to show how
their concept can refer to an object, and that, too, is not yet cognition of these objects […].
But nevertheless theoretical cognition […] has been thus extended insofar as, through the
practical postulates objects were still given to these ideas by lending objective reality to a
merely problematic thought (5:135; my emphasis).

It will not be possible here to do full justice to the complexity of Kant’s reaso­
ning in this passage and its context.17 I only want to highlight three points that
become sufficiently clear in this passage: First, as far as speculative reason is
concerned, the ideas of freedom, God, and immortality have only logical, but not
real possibility. Second, the transcendental ideas receive objective reality – that
is, a relation to some possible18 object – only through their relation to the moral
law and the postulates based on it. And third, this does not suffice to explain
how they relate to objects (“without being able to show how their concept can

17 For a more detailed interpretation cf. Willaschek (2010).

18 Kant leaves out “possible” in the passage quoted, but he had it in two sentences before:
“Since hereby nothing further has been achieved by practical reason than that those concepts
are real, and really have their (possible) objects […], no synthetic sentence is possible through
their acknowledged reality” (5: 134); my emphasis).
146 Marcus Willaschek

refer to an object”). In this way, Kant’s account of the transcendental ideas, even
though it may ultimately conflict with the Causal Assumption, at least respects
the Causal Condition insofar as Kant admits that we cannot explain how our ideas
of freedom, God, and immortality relate to their objects, even though the moral
law assures us that they do have (possible) objects and at the same time warrants
our belief in their reality.

5 Conclusion
Given the Causal Assumption, the sensibility of human intuition follows from
Kant’s definition of sensibility plus the fact that we, as finite beings, cannot
produce objects (with respect to their existence) simply by representing them.
Even if this assumption should turn out to be problematic, Kant is clearly com­
mitted to the Causal Condition, which means that if we are to have an account
of representation at all, we will have to think of human intuition as sensible.
Now this defence of Kant’s claim that human intuition is sensible will only be as
convincing as its crucial premises: the Causal Assumption/Condition and Kant’s
definition of sensibility (assuming that the finitude of the human mind is indis­
putable). I will close by looking very briefly at some objections to these premises.
Of course, it is possible simply to deny the Causal Assumption. A response
on behalf of Kant could then be to just drop the more demanding Causal Assump­
tion and restrict the argument to the Causal Condition. In this way, the burden of
proof is shifted to the critic: Either the critic will have to admit that the possibil­
ity of non-sensible intuitions is strictly inexplicable or she will have to offer a
non-causal account of how non-sensible intuitions are supposed to relate to their
objects.
Now at least the traditional rationalist of the Cartesian kind may indeed
have available a non-causal account of the representational character of non-
sensible intuition, namely the so-called resemblance theory of representation.
Very roughly, on this theory a representation represents its object due to the fact
that it resembles it in some relevant way. However, even setting aside the internal
problems of this theory, it will not do as an account of Kantian intuitions, since
Kantian intuitions are supposed to be singular, essentially picking out one partic­
ular object. If what constitutes the representational relation is resemblance alone,
there is no way in which a representation can pick out a particular object as such,
since the representation would then represent whichever object it resembles.
Even if it should happen to represent just one single object, this would not suffice
to make the representation an intuition in the Kantian sense. Another way to put
this would be to say that, since resemblance is a potentially many-place relation,
 The Sensibility of Human Intuition 147

all representation through resemblance is representation through general marks

and thus does not establish any direct relation to represented objects.19
Here the traditional rationalist may respond that there is a different way to
satisfy the Causal Condition, namely by claiming that some of our representa­
tions are innate ideas and that these ideas were placed in our minds by God in
such a way as to guarantee that they adequately represent their objects. As Des­
cartes famously argues in the third of his Meditations, the only way to account for
our having the idea of God in our minds is to assume that God himself caused us
to have this idea. Notice that this account satisfies the Causal Condition, since it
establishes a causal connection between our innate ideas and the objects they
represent. In the case of our idea of God, God himself is the cause of that idea.
In the case of the other innate ideas, God establishes at least an indirect causal
connection by causing us to have these ideas so as adequately to represent their
objects.
Kant, of course, doesn’t have any sympathy with this kind of theory. In the
letter to Herz, after ascribing to Plato, Malebranche, and Crusius the view that
the adequacy of our non-sensible representations is due either to our immedi­
ate intuition of a deity (Plato, Malebranche) or to the deity’s causing us to have
adequate representations (Crusius and others), Kant writes: “However, when it
comes to determine the origin and validity of our cognition, the deus ex machina
is the most incongruous thing one might possibly choose and has, apart from the
deceiving circle in the line of inferences of our cognitions, the disadvantage that
it abets any whim or either pious or brooding figment of the brain” (10:131). The
“circle” Kant is thinking of may be the so-called Cartesian circle quite generally,
but it may also be, more specifically, the circle that would result from presup­
posing, in an account of how we manage to represent non-sensible objects, that
one succeeds in representing God (who, of course, would have to be a non-sen­
sible object himself). Particularly when backed by Kant’s “subjectivist” account
of how we come to have transcendental ideas and by Kant’s critique of possible
proofs of the existence of God, this charge of circular reasoning presents a serious
challenge to the traditional rationalist. It means that the rationalist can neither
presuppose the existence of God nor even presuppose that we succeed in repre­
senting Him. Rather, the Cartesian critic would first have to explain how our idea
of God can have “objective reality” in the Kantian sense (i. e. represent a pos­

19 This response may not seem to work in the case of our concept of God, which (on Kant’s as
well as on all traditional accounts) has singularity built into it. But note that in this case then, it
is not similarity that constitutes the relation between representation and its object.
148 Marcus Willaschek

sible object) before she can try to account for the objective reality of our ideas in
general by appeal to God.
Next, one might object that Kant’s definition of sensibility is either too wide
or inadequate. It is too wide, and trivializes Kant’s denial of non-sensible intu­
itions, if it is read in such a way as to allow for affections of our minds that do not
engage any of our (internal or external) senses (e. g. divine inspiration). It is inad­
equate, and effectively begs the question whether human intuition is sensible, if
it presupposes that the only way for an object to affect our minds is by exciting
one or more of our senses. Faced with this choice, it is clear that Kant must opt
for the second possibility: The only way an object can cause a representation by
affecting our minds is to excite one or more of our senses. Admittedly, this claim,
although it will seem highly plausible to most modern readers, would have been
disputed by many of Kant’s contemporaries. Again, the best strategy on behalf of
Kant may consist in shifting the burden of proof to the critic: Granted that there
is a logical possibility of non-sensible affections of our minds, it is the critic who
must come up with a plausible account of how this kind of non-sensible affection
might work and offer us reasons for attributing to the human mind a receptivity
for non-sensible affections. Moreover, in light of subjectivist accounts of our rep­
resentations of non-sensible objects, it seems that the critic cannot simply rely
either on our having representations of the kind in question or on there being
non-sensible objects for them to represent. Even if this strategy does not exclude
the possibility of non-sensible affections of our minds, it considerably strength­
ens the Kantian position in this respect.
I conclude that even though Kant doesn’t offer any explicit argument for his
claim that human intuition must be sensible, his philosophy contains the ele­
ments for a powerful defence of this central Kantian claim.20

Literature
Allison, Henry (22004): Kant’s Transcendental Idealism. An Interpretation and Defense.
2. Edition. New Haven/London: Yale University Press.
Engstrom, Stephen (2006): “Understanding and Sensibility”. In: Inquiry 49 (2006), p. 2–25.

20 Thanks for helpful comments on a predecessor of this paper to Steve Engstrom, Thomas
Höwing, James Messina, Daniel Warren and Eric Watkins, as well as audiences in Chicago, San
Diego, Konstanz, Pisa and Tübingen. Thanks also to Carolyn Benson for correcting my English.
 The Sensibility of Human Intuition 149

Heidemann, Dietmar (2002): “Anschauung und Begriff. Ein Begründungsversuch des Stämme-
Dualismus in Kants Erkenntnistheorie”. In Engelhard, K. (ed.): Aufklärungen. Festschrift für
Klaus Düsing zum 60. Geburtstag. Berlin: de Gruyter 2002, p. 65–90.
Klimmek, Nikolai F. (2005): Kants System der transzendentalen Ideen. Berlin: de Gruyter.
Longuenesse, Béatrice (1998): Kant and the Capacity to Judge. Princeton: Princeton University
Press.
Willaschek, Marcus (2010): “The Primacy of Pure Practical Reason and the Very Idea of a
Postulate”. In: Reath, A./Timmermann, J. (eds.): Kant’s Critique of Practical Reason. A
Critical Guide. Cambridge: Cambridge University Press, p. 168–196.