1. Nonconceptualist Readings

Proponents of NCR agree with many other readers of Kant that the following commitment is central to his overall position:

By saying that a representation is about substantial objects I mean that the representation possesses objective purport and that the objects it purports to be about are substantial objects.Footnote ⁸ By ‘substantial object’ I mean, roughly, objects that are capable of persisting through changes of their properties and that stand in causal relations; objects, that is, which instantiate Kant’s categories. Exactly what a concept is for Kant and what it takes to apply concepts need not concern us for now. The point is simply that concepts must in some way be involved for a representation to be about substantial objects.

Given Kant’s commitment to Concepts, we can ask whether intuitions are about substantial objects. Proponents of NCR deny this, but unlike many other commentators they do not conclude that there is no sense in which intuitions are about objects.Footnote ⁹ Nor do they conclude, as yet other commentators do, that intuitions possess objective purport only in the context of judgement.Footnote ¹⁰ The key to appreciating Kant’s thinking here, according to advocates of NCR, is to realize that he distinguishes between two different kinds of objective purport – one which is characteristic of concepts and judgements and one which is characteristic of intuitions. This, they claim, is the real upshot of Heterogeneity.

According to NCR, then, we need to draw a distinction between what I will call strong objective purport and weak objective purport. A representation has strong objective purport just in case it purports to be about objects in a weighty sense of ‘object’. Objects in this sense necessarily instantiate the categories. To say that a representation has strong objective purport, therefore, is to say that it represents its object as instantiating the categories; put differently, it is to say that the representation exhibits the ‘unity that is thought in the categories’ (Kant Reference Kant1998: 256; CPR, B151). That the object of such a representation instantiates the categories is thus part of the representation’s content.

By contrast, weak objective purport concerns objects in a very thin sense. Something is an object in the thin sense just in case it is (in principle) perceptible and it occupies (or at least apparently occupies) a determinate position in space and time. A representation has weak objective purport just in case it is about an object only in the thin sense of ‘object’. Although in fact the object picked out by such a representation may be an object in the weighty sense, this is not something that is part of the content of the representation.

According to NCR, only judgements possess strong objective purport.Footnote ¹¹ But again, proponents of the view argue that it does not follow from this, nor is it the case, that intuitions do not possess objective purport. For they possess weak objective purport. For the purposes of this paper, then, a nonconceptualist reading of Kant is any view that ascribes weak objective purport to intuitions and strong objective purport to judgement.Footnote ¹²

Different advocates of NCR have different ways of fleshing out the abstract notion of weak objective purport I have introduced.Footnote ¹³ But they all share the commitment that, if a representation has weak objective purport, it does not represent its object as instantiating the categories. Consider, first, the view of Allais, who draws the distinction between two kinds of objective purport in terms of ‘empirical particulars’ and ‘objects in the full blown sense’. The former are ‘spatially continuous and unified individuals existing outside the subject’ (2009: 405) and ‘uniquely located in an oriented and egocentrically-centered, three-dimensional framework’ (2009: 404). An object in the full blown sense, on the other hand, is

Hanna likewise holds that representing something as having determinate spatio-temporal location is sufficient for weak objective purport:

Clearly, the implication is that locating what Hanna calls a material object here and now is independent of representing it as, for instance, capable of persisting through changes of its properties or capable of existing unperceived and thus independent of representing it as instantiating the categories.Footnote ¹⁴

McLear, finally, characterizes weak objective purport as ‘a presentation of something in a spatially and temporally contiguous manner – e.g. the perceptual presentation of shape and coextensive color’ and distinguishes this from ‘the presentation of some particular thing as colored and shaped’, which is what strong objective purport comes to (McLear Reference McLearforthcoming: 41).Footnote ¹⁵

According to nonconceptualists, recognizing that weak objective purport constitutes a distinctive way in which a representation may be about objects is the key to doing justice to both Heterogeneity and Concepts. In their view, part of what it means to say that intuitions and concepts are two irreducibly distinct kinds of representation is that the objective purport of intuition cannot be understood in terms of the objective purport of judgement. But in contrast to many other commentators they also hold, characteristically, that one ought not to infer from this that intuitions do not have objective purport independently of the application of concepts in judgement.Footnote ¹⁶

2. The Objective Validity of the Categories

I wish to argue that nonconceptualist readings of Kant fail because they cannot make good sense of the Transcendental Deduction of the categories. In particular, I will argue that proponents of NCR have either to say that the Deduction contradicts the teaching of the Transcendental Aesthetic or, if they see the Deduction as consistent with the Aesthetic, they face a dilemma. Either they must construe the Deduction as being concerned with a question of fact rather than a question of legitimacy, despite Kant’s explicit avowals to the contrary; or they must ascribe to Kant a position he criticizes as a form of scepticism.

According to Kant’s own official characterization the goal of the Deduction is to establish the objective validity of the categories, or pure concepts of the understanding, and he seeks to do this by showing that these concepts are ‘conditions of the possibility of experience’ (Kant Reference Kant1998: 225; CPR, A94/B126). Roughly, this means that the goal is to show that all objects of which we can in principle have empirical knowledge instantiate the categories. In making his case for this, Kant appears to argue that intuitions depend on synthesis and that synthesis is guided by, and therefore involves an application of, the categories.Footnote ¹⁷ But this amounts to saying that intuitions have strong objective purport, which is just what NCR denies.

At this point, a nonconceptualist has two options. Either she accepts that this is indeed what Kant claims in the Deduction or she does not. If she opts for the first alternative, she ascribes to Kant an inconsistent position, saying that he both affirms and denies that intuitions have weak objective purport. If she opts for the second alternative, she must say that it merely appears, but is not in fact the case, that Kant argues in the Deduction that intuitions depend on category-guided synthesis. I wish briefly to comment on the first alternative before discussing the second in greater detail.

The first alternative attributes to Kant a blatant contradiction on an issue of central importance to the Critique. Considerations of charity suggest that an interpretation of this sort must meet an especially high standard of evidence. I will argue that the evidence that has been given for this version of the nonconceptualist reading does not meet that standard.Footnote ¹⁸

This evidence consists primarily in a number of passages in which Kant seems to suggest that intuitions can present objects to the mind independently of any involvement of the understanding; hence that intuitions have weak objective purport (see CPR, A89/B122, A 90/B122, A90–1/B122–3, B132, B145, and JL, 9: 33).Footnote ¹⁹ However, as e.g. Griffith (Reference Griffith2012) and Grüne (Reference Grüne2011) show in detail, three of these passages (CPR, A89/B122, A 90/B122 and A90–1/B122–3) belong in a context in which, for dialectical purposes, Kant describes a scenario that he wants to show does not obtain.Footnote ²⁰ Of the remaining three passages, the one from JL implies that it is possible to have an intuition of a house without possessing the concept of a house. By itself, however, this does not show that Kant ascribes weak objective purport to intuitions. All it shows is that Kant rejects a very strong version of conceptualism about intuition, according to which having an intuition requires possession of all concepts that describe the content of that intuition. But rejecting this version of conceptualism is compatible with ascribing strong objective purport to intuitions.Footnote ²¹

The remaining two passages are as follows:

The first of these speaks generically of the manifold of intuition. It does not specify that the manifold at issue is an empirical manifold. For all the passage says it could be a pure manifold. However, the requirement that a pure manifold of intuition be given prior to the synthesis of the understanding is compatible with the view that empirical intuition depends on the synthesis of the understanding (and therefore possesses strong objective purport). So the passage certainly does not establish Kant’s commitment to nonconceptualism. Indeed, absent further argument it is not even clear that it provides any evidence in favour of ascribing this commitment to him.

The second passage also fails to be conclusive. Kant might well be using the term ‘thinking’ here in a narrow sense, in which thinking is equivalent to judging (in the Kantian sense of that word). But from the fact that intuition can be given prior to all judging it does not follow that intuitions possess weak objective purport. For intuitions could depend on a kind of exercise of the understanding that Kant calls the transcendental synthesis of the imagination and that he is at pains to distinguish from judging (see CPR, A101, A119, B151ff). If that were the case, intuitions would still have strong objective purport. The context of the passage arguably supports such a reading. For it makes clear that Kant’s talk of thinking is intended to pick out the act of accompanying a representation with ‘I think’. Whatever the precise nature of this act, it is certainly plausible to think that not every act of spontaneity takes the form of accompanying a representation with ‘I think’: in particular, that the synthesis of the imagination does not.Footnote ²² I conclude, then, that the textual evidence provided by the passages we have considered does not by itself establish that Kant ascribes weak objective purport to intuitions and therefore fails to meet the standard of proof that this variant of NCR needs to satisfy.

I turn now to the second alternative a proponent of a nonconceptualist reading might opt for in responding to the fact that Kant appears to argue in the Transcendental Deduction that intuitions depend on synthesis and that synthesis is guided by, and therefore involves an application of, the categories. This alternative is to claim that it merely appears that Kant argues this. I will argue that this strategy also fails. My discussion will focus on the articulation of this alternative due to Allais, but is intended to be general and apply to any nonconceptualist reading opting for this alternative.Footnote ²³

According to this alternative, what Kant is concerned with in the relevant parts of the Transcendental Deduction is not intuition and its intentionality, but what he calls experience and the kind of intentionality that pertains to it (see Allais Reference Allais2009: 402). Experience in Kant’s technical sense is empirical cognition, which takes the form of judgements. But from the fact that the categories are required to account for the intentionality of judgement, nonconceptualists argue, it does not follow that they are required to account for the intentionality of intuition.

What then, according to nonconceptualist readings, is Kant’s strategy for showing that the categories are valid of all objects of which we can have empirical knowledge? The following passage from Allais gives the flavour of the position such readings take:

The categories, the passage says, are necessary for the ascription of properties to objects in empirical judgement. The idea is that in representing an object as a bearer of properties I ipso facto represent the object as instantiating the categories. But properties are ascribed to objects in thought (i.e. judgement, in Kant’s terminology). Accordingly, the categories are conditions of the possibility of experience because they are conditions of the possibility of judgement. The Transcendental Deduction is intended to show that this is so.Footnote ²⁴

I now wish to argue that this account of the Deduction faces two problems. First, consider that Kant’s own characterizations of the goal he pursues in the Deduction strongly suggest that, for him, showing that the categories are conditions of the possibility of experience involves more than showing that they are necessary conditions of judgement. This comes out especially clearly in the B-Deduction.Footnote ²⁵ Here the argument is famously presented in two distinct steps. And when he explains why the second step is needed, Kant says that the goal of the Deduction will only be ‘fully attained’ if the validity of the categories for ‘all objects of our senses’ is demonstrated (Kant Reference Kant1998: 253; CPR, B145). This, he says, will be done by ‘[showing], from the way in which empirical intuition is given in sensibility, that its unity is none other than that which the category prescribes to the manifold of a given intuition in general’ (Kant Reference Kant1998: 253; CPR, B144–5). Clearly, the suggestion here is that the categories have a role to play in intuition itself, not just in judgement.

This suggestion is confirmed when Kant characterizes the conclusion he has reached at the end of the Deduction, in §26, where he says:

The claim here seems to be twofold: first, that perception depends on an act of synthesis which ‘stands under’ the categories; second, that the categories are conditions of the possibility of experience (at least in part) because this act of synthesis is one that is required for the possibility, not of judgement, but of perception and is thus an act of a specifically sensible synthesis, which according to CPR, B151–2, depends not merely on the capacity for judgement but also on the productive imagination.Footnote ²⁶ If this is right, nonconceptualists are simply wrong to see the main thesis of the Deduction as one exclusively concerning the role of the categories in judgement.Footnote ²⁷ Rather, that thesis concerns the role of the categories in perception, which I take to be equivalent to empirical intuition.Footnote ²⁸

The second problem can be approached by asking what the problem is for which the Transcendental Deduction is intended to provide a solution. At one point Kant characterizes it as the problem of how ‘subjective conditions of thinking should have objective validity, that is, function as conditions of the possibility of all cognition of objects’ (Kant Reference Kant1998: 222; CPR, A89–90/B122). The issue here is that, because the categories are not derived from objects, by way of experience, but are rather possessed by a thinker independently of experience, simply in virtue of being a thinker, it is not clear that there is any reason for believing that objects of experience instantiate the categories. And since our cognitive access to objects of experience (what Kant calls ‘appearances’) is through intuition, this worry can also be expressed as a worry about the right kind of ‘fit’ between the pure concepts, on the one hand, and what is given in intuition (i.e. appearances), on the other. Absent such a fit, ‘appearances may well be so constituted that the understanding would not find them to be in accordance with the conditions of its unity’ (Kant Reference Kant1998: 223; CPR, A90/B123).Footnote ²⁹ If this were the case, the categories would be ‘empty, null, and without meaning’ (ibid.). It is the task of the Transcendental Deduction to demonstrate that this is not the case.

Now Kant famously connects this issue with a distinction between a question of fact (quid facti) and a question of legitimacy (quid iuris) (see CPR, A84–5/B116–17).Footnote ³⁰ Roughly, the question of fact is whether or not we possess certain concepts. The question of legitimacy is whether or not it is legitimate to apply these concepts in judgements that purport to be true of objects. It will be legitimate so to apply them if the objects to which they are applied in fact instantiate them.Footnote ³¹ Clearly, the question of fact does not by itself settle the question of legitimacy.

If we connect this point with the previous point, it looks as if for Kant the quid iuris question regarding the categories has to be answered, at least in part, by showing that there is the right kind of fit between the pure concepts, on the one hand, and what is given in intuition, on the other. We can now put the second objection to the kind of nonconceptualist proposal articulated by Allais by saying that it addresses the question of fact but not the question of legitimacy. Recall that according to that proposal Kant argues that the categories are conditions for the possibility of experience because they are necessary conditions of judgement.

It might be objected that in fact Kant’s argument (so construed) is concerned with a question of legitimacy rather than a question of fact. For it is concerned not with the question whether we in fact possess the categories, but whether the categories are necessary for representations of a certain kind. And one might think that showing that the categories are necessary in this sense confers legitimacy on their employment. In fact, however, it does nothing of the kind. For as far as the problem motivating the Deduction is concerned, an act of judgement is just as much a ‘subjective condition of thinking’ as a pure concept is. What needs to be shown is that it is legitimate, or appropriate, to make judgements about appearances (and thereby apply categories to them); that appearances have the kind of unity that makes category-involving judgement the appropriate manner of representing them. Claims about what it takes to make a judgement do not address this issue. The second objection to this kind of nonconceptualist reading, then, is that it misconstrues the task of the Transcendental Deduction.Footnote ³²

3. Objective Validity and Necessary Agreement

I have raised two objections for the version of NCR that sees Kant’s view in the Deduction as consistent with his position in the Aesthetic, but insists that the Deduction is not aimed at establishing that intuitions have strong objective purport. I now wish to consider possible responses. As regards the first objection, a proponent of NCR could respond that I infer from Kant’s claim that the aim of the Deduction is to demonstrate the validity of the categories for ‘all objects of our senses’ that the categories must be applied in perception itself (specifically, in perceptual synthesis). But while Kant’s claim is that the objects we perceive must be shown to instantiate the categories, the objection ascribes to him the view that, in perceiving them, we must represent these objects as instantiating the categories. Clearly, however, the former does not imply the latter. So the textual evidence does not in fact support the objection.

However, this response to my first objection is not convincing. What it shows is that passages such as B161 are open to different readings, each of which possesses at least some prima facie plausibility. This suggests that it will not be possible to resolve the dispute by textual considerations alone. For this reason I think our focus should be on the second objection because it is here that the philosophical issues at stake come to the fore.

That said, I do think that the reading of B161 I have sketched on NCR’s behalf is a stretch. For just a few lines further on in the text, at B162, Kant says that the synthesis of apprehension ‘must be thoroughly in accordance with’ the category of quantity. To this claim he appends a footnote in which he says that the capacity responsible for the synthesis of apprehension is spontaneity ‘under the title of imagination’ (Kant Reference Kant1998: 262; CPR, B162n.) and this strongly suggests that B161 should be read in my way rather than Allais’s.Footnote ³³

Let me now turn to my second objection. A proponent of NCR might respond to it as follows: the objection is right, the response goes, to draw attention to the fact that Kant is concerned not only with showing that the categories are necessary for judgement, but also with showing that the objects which we represent in judgement in fact instantiate the categories. However, his argument in support of the latter claim does not pivot on the idea that perception depends on a category-guided synthesis. Rather, the argument is roughly as follows. As demonstrated in the Transcendental Aesthetic, we know a priori that objects of experience are in space and time. But space and time have a certain kind of unity, on account of which everything that is in space and time stands in spatio-temporal relations to everything else. If this is right, then objects of experience have the kinds of properties and relations, whatever they are, which are necessary for standing in spatio-temporal relations to other objects of experience. And now the idea is that these are just the properties and relations that are represented by the (schematized) categories. Consequently, the objects given in sensibility instantiate the categories and do so in virtue of their spatio-temporal form. Moreover, it is open to a nonconceptualist to insist that this argument fits the description of ‘[showing], from the way in which empirical intuition is given in sensibility’, that the unity of objects given in sensibility is ‘none other than that which the category prescribes to the manifold of a given intuition in general’, namely, categorial unity (Kant Reference Kant1998: 253; CPR, B144–5). It would follow that it is simply not true that NCR cannot account for the fact that the Deduction seeks to answer a quid iuris question.

I wish to argue that this response to my second objection runs into a problem that Kant himself discusses, i.e. that a position of the kind sketched in the response amounts to a form of scepticism. More precisely, the position is an instance of a type of view which Kant dubs ‘a kind of preformation-system of pure reason’ (Kant Reference Kant1998: 265; CPR, B167) and against which he argues as follows: another way of saying that the categories are valid of all objects of experience is to say that there is a ‘necessary agreement’ between experience and the categories. I take this to mean that the fact that experience conforms to the categories is non-accidental, not simply a brute fact. Kant then argues that there are only two possible ways for experience and the concepts of its objects to stand in necessary agreement with each other: ‘either the experience makes these concepts possible or these concepts make the experience possible’ (Kant Reference Kant1998: 264; CPR, B166).

Notice that ‘experience’ here cannot simply mean ‘empirical judgement’ (as the nonconceptualist would likely argue). For if that were the case, then showing that there is a necessary agreement between a concept, on the one hand, and a judgement in which this concept is employed, on the other, is either trivial or bizarre.Footnote ³⁴ Rather, what must be at issue here is the agreement between the concepts of the objects of experience and these objects themselves; in other words, what must be at issue is the claim that the objects of experience instantiate the relevant concepts, not just the claim that they are represented as instantiating these concepts.

Having put forth this alternative, Kant then considers a potential objection, which is that there is a third option. The third option is that there is no dependence in either direction. Rather, experience and the categories are independent of one another, yet nonetheless in agreement. What this means is that on the one hand there are ‘the laws of nature along which experience runs’ (Kant Reference Kant1998: 265; CPR, B167) and on the other hand there are dispositions for thinking in certain ways, which capture those laws. However, to this proposal Kant objects that, although there is agreement here, the agreement is accidental. It is accidental because the ways of thinking would remain the same even if the laws of nature were different. This follows from the fact that what explains the way we think (in this scenario) is the fact that we have a psychological disposition to think this way. But that is just to say that we would think in this way whether or not it agreed with the way nature is. And if this is so, then it is only an accident that there is agreement rather than disagreement.

Kant calls this a ‘kind of preformation system of pure reason’ because the dispositions for thinking in certain ways are envisaged as being implanted in us by our maker, who sees to it that they agree with the laws of nature. But the crucial point is that our forms of thinking are independent of the laws of nature. Adding the maker into the mix does not alter that.

Now of the preformation system Kant says that it is ‘what the sceptic most desires’ (Kant Reference Kant1998: 265; CPR, B168). Why is that? Here is what Kant says:

The point Kant makes here is this: a causal judgement to the effect that A causes B asserts a necessary connection because it says that, given A, B cannot fail to obtain. In saying this the judgement represents the necessary connection between A and B as an objective fact about them. If, however, the concept of cause reflected merely a psychological disposition to combine representations of As with representations of Bs, then we could not legitimately claim that a causal judgement represents a necessary connection as obtaining objectively. All we would be entitled to assume is that we cannot help but think this way. And this would fail to establish the legitimacy of our concept of cause.

It is Kant’s view, then, that a ‘preformation’ account of the applicability of the categories to objects of experience leads directly to scepticism. I have discussed this account because I wish to claim that the position I attributed to the nonconceptualist in response to my second objection above is of exactly this kind. That position construes Kant’s argument in the Transcendental Deduction as resting on an appeal to the character of the forms of intuition, space and time. The claim is that it is in virtue of certain facts about space and time that the categories apply to what is given in space and time: specifically, in virtue of the fact that space and time have a certain kind of unity such that everything that is in space stands in determinate spatial relations to everything else that is. But these facts obtain in complete independence of the categories. They would obtain even if we possessed no categories or completely different ones. Conversely, the categories would constitute our forms of thinking whether or not the relevant facts about space and time obtain. This, however, is just the kind of scenario that Kant thinks is characteristic of a preformation system. Accordingly, the scepticism charge made against the latter applies here as well.

If this is right, then a view on which the objective validity of the categories ultimately rests on independent facts about space and time is one that is committed to regarding the categories as having merely subjective necessity. On such a view, the categories turn out to be mere psychological dispositions that we are bound to actualize whether or not they agree with their objects. For this reason, there is no necessary agreement between the categories and their objects. As a consequence, the categories do not ‘make experience possible’ and fail to be objectively valid. I conclude, then, that the nonconceptualist’s response to my second objection fails.

Above I presented this response as a reaction to the objection that versions of NCR such as Allais’s and McLear’s are unable to account for Kant’s claim that the Transcendental Deduction addresses a question of legitimacy rather than of fact. I can now summarize the discussion so far by saying that these versions of NCR face a dilemma: either to construe the Deduction as being addressed to a question of fact rather than one of legitimacy or to attribute to Kant a preformation view of reason. Neither horn is acceptable.

4. The Heterogeneity of Understanding and Sensibility

If my argument so far is on the right track, all versions of NCR should be rejected. This invites the question what an alternative reading should look like. In the remainder of this paper I would like to present an outline for one.

From my discussion of the preformation charge we can extract a requirement that a more promising alternative must meet. It must make room for a robust dependence of the objects of experience on the categories, so that the latter ‘make possible’ the former in the sense required for the necessary agreement Kant wants. How could this requirement be met? Questions about transcendental idealism loom here, which I cannot address in this paper. But the following is relatively uncontroversial: Kant wants a kind of idealism according to which what is mind-dependent is not the existence of objects but only their form. The form of objects is mind-dependent in the sense that it is constituted by the form of representation.Footnote ³⁵ Since there are two kinds of representations that are relevant here, i.e. concepts and intuitions, the form of objects is constituted jointly by the form of conceptual representation and the form of intuitive representation. The question then is what this entails, and it is here that we enter more contested territory. My own view is that the preformation charge makes it clear that Kant holds the following position: if the categories make possible the objects of experience, then not just the form of conceptual representation depends on the categories, but also, in part, the form of intuitive representation. Accordingly, the categories must be understood to be required not just for judgement, but also for intuition.

Although making a proper case for this contention would require more than I can provide here, the following passage offers some textual support:

Kant says here that the synthesis responsible for the unity of an intuition, which is an act of the imagination, itself depends on the understanding. Clearly, this suggests that intuition itself depends for its unity on the understanding. It seems, therefore, that the way in which the understanding ‘makes possible’ the objects of experience is by being responsible for the unity of intuition. What this suggests is that at least part of the reason why nothing can be given in sensibility that does not conform to the categories is that the categories themselves account for the unity of intuition.

A proponent of NCR will likely object that this proposal runs afoul of Heterogeneity. For I have argued that the categories are responsible for the unity of intuition. And since the categories derive from the logical forms of judgement, it is hard to see what this could mean if not that intuition is not after all independent of judgement. But to say that seems to amount to denying that intuitions are distinct in kind from conceptual representations, that is, to denying Heterogeneity. Call this the Objection from Heterogeneity.

It will be useful to give an alternative formulation of this objection in terms of the distinction employed by Allais, between representing an x that is F and representing an x as an F.Footnote ³⁶ According to her, one applies a concept F to an object x just in case one represents x as being F. This contrasts with representing an object that is F, but is not represented as being F; no application of concepts is required here. If now we think of the categories collectively as constituting the concept of an object, we get the thought that to apply the categories to an object is to represent this object as an object. But in perception, according to Allais, we do not represent objects as objects. Rather, we simply perceive objects, much like the savage in Kant’s famous example sees a house without seeing it as a house.Footnote ³⁷ From Allais’s point of view, then, to say that perception involves the application of the categories is to say that in perception we represent objects as objects. But this seems to imply that perception exhibits the structure characteristic of predicative judgement. And this is, in effect, to turn perception into a species of conceptual thought and thus to deny Heterogeneity.Footnote ³⁸

The Objection from Heterogeneity deserves serious consideration. Still, I shall argue that the reading I have sketched has the resources to deflect it. In a first step I will specify in the abstract what it would take to deflect the objection and to do so I will deploy the set of inconsistent claims I discussed in the introduction to this paper. These are:

The Objection from Heterogeneity is motivated in part by Incompatibility, the claim that a representation that is sensible cannot depend on acts of spontaneity. For my proposed reading is clearly incompatible with this claim: since the categories are products of spontaneity, intuitions would depend on spontaneity if they depended on the categories. To make my reading viable, therefore, I need to show that we can reject Incompatibility. And this can be done if, contrary to what is presupposed by the objection, it can be shown that Incompatibility is not entailed by Heterogeneity. For recall that I was able to generate an inconsistency among the three basic Kantian commitments with which I began this paper only by adding Incompatibility. But Kant’s alleged commitment to Incompatibility depended on his commitment to Heterogeneity. Simply adding Heterogeneity to our three basic commitments, however, does not generate any inconsistency. It follows that if I can show that Heterogeneity does not entail Incompatibility, the Objection from Heterogeneity poses no threat to my proposed reading.

To take a step towards this goal I would like to return to the way Allais fleshes out the Objection from Heterogeneity. As she presents it, the objection that an involvement of the categories in intuition itself would undermine Heterogeneity is premised on the idea that to apply categories is to make judgements. This is the thought Allais expresses by saying that the categories are required for representing an object as an object and that representing-as requires the predicative structure of judgement. If she is right about this, the objection stands. To block the move from Heterogeneity to Incompatibility, therefore, I need to argue that there is a way of applying categories that is not tied to judgement.Footnote ³⁹ This can be done by showing that the categories are required for a distinctive kind of self-consciousness, namely one that is necessarily involved in exercises of the capacity to have intuitions.Footnote ⁴⁰ Although this would require considerable work to be developed properly, the basic idea is as follows. Footnote ⁴¹

The first point to note is that the categories are formal rather than material concepts. They characterize a form of representation and do not serve to distinguish one instance of this form from another, as a material concept would. As a consequence, applying a category does not paradigmatically take the form that applying a material concept takes, namely that of the explicit ascription of a predicate to an object in an act of judgement. Rather, categories are applied in every judgement, not in virtue of figuring in its material content (as the concept of being F figures in ‘a is F’), but rather in virtue of its form, that is, in virtue of its being a representation of this kind, a judgement.Footnote ⁴²

To spell this out just a bit, the idea is that a judgement is by its nature a kind of representation that has objective purport (see CPR, B140–2).Footnote ⁴³ But since the categories jointly constitute the concept of an object in general, this amounts to saying that a judgement by its nature represents its content as instantiating the categories – as having the kind of unity that is constitutive of elements of the world of appearances.

Moreover, the fact that the categories are implicated in a judgement in this way is something that is known to the subject making the judgement. And, again, this is so in virtue of the nature of this manner of representation and thus a formal aspect of it: a judgement is essentially such that the subject making it knows that she is making a claim about how things are objectively, a claim that is truth-evaluable.Footnote ⁴⁴ This kind of knowledge does not typically take the form of a conscious thought to the effect that the judgement lays claim to saying how things are. It is rather a kind of implicit understanding, which manifests itself, for instance, in the subject’s readiness to withdraw her judgement when presented with countervailing evidence, etc.Footnote ⁴⁵

Now in the case of judgement, application of the categories is tied to the employment of the logical forms of judgement and thus to the presence of predicative structure. As a consequence, so is the consciousness of objective purport that I have just tried to sketch. However, it seems that it is possible in principle to divorce this consciousness from the presence of predicative structure. If we can form the idea of a kind of capacity whose exercises include, in virtue of their form, the kind of consciousness of objective purport just sketched, but without tying it to the presence of predicative structure, then we will have identified a way of applying the categories that is sufficiently independent from judgement to present no threat to Heterogeneity.Footnote ⁴⁶

With regard to intuition the idea would be that the application of the categories manifests itself in the fact that the subject has an understanding of the kind of representation that sensibility provides her with. In particular, she understands that, when all goes well, sensibility provides her with representations of mind-independent objects; objects that instantiate the categories. Again, this will be a kind of understanding that manifests itself not in the explicit ascription of certain properties but, for instance, in the disposition to treat an intuition of, say, a red ball as a (defeasible) reason for judging that there is a red ball in front of her.Footnote ⁴⁷ Or to give another example, one that brings to the fore the link between the categories and the notion of synthesis, possessing this understanding will dispose the subject to treat what from a certain perspective can be described as a momentary impression of a red-facing surface as the perception of an enduring three-dimensional material object; say, a red ball.Footnote ⁴⁸

While the kind of grasp that a subject has of her capacity to have intuitions is typically implicit in this way, it can be articulated. The pure principles of the understanding, in which the categories are related to the form of inner sense, will figure in a fully reflective articulation. Thus one aspect of what a subject capable of intuiting objects implicitly understands is that, when a perceived object undergoes a change of qualities, there is a cause by reference to which this change can be explained (see CPR, B232–4). The crucial point, however, is that while the categories figure explicitly in the articulation of the subject’s understanding of her capacity for intuition, this understanding can be (and typically is) operative in intuition without an explicit deployment of the categories. The idea is that this opens the way for recognizing an application of the categories in intuition, which does not require us to attribute judgemental structure to intuition and is therefore fully compatible with Heterogeneity. In other words, sensible representations exhibit their own distinctive structure, which is different from the discursive structure of judgements, while also being dependent on the categories. Using the notion of representing-as the point can be put as follows: because the categories are part of the implicit understanding of her own sensible capacity that a subject possesses, the act of representing something as an object need not take the form of making a judgement. Rather, objects are represented as objects in intuition itself. Instead of forcing us to attribute to intuition the predicative structure of judgement, what this requires is that the capacity for intuition essentially involves the kind of implicit understanding of what it is that an intuition is, in general, a representation of that I have elaborated. This idea allows one to preserve Kant’s commitment to Heterogeneity, while also ascribing to him, without inconsistency, all three of Sensible Intuition, Synthesis Dependence, and Spontaneity.Footnote ⁴⁹